here - IRRI books - International Rice Research Institute

The International Rice Research Institute (IRRI) was established 

in 1960 by the Ford and Rockefeller Foundations with 

the help and approval of the Government of the Philippines. 

Today IRRI is one of the 15 nonprofit international research 

centers supported by the Consultative Group on International 

Agricultural Research (CGIAR – www.cgiar.org). 

IRRI receives support from several CGIAR members, 

including the World Bank, European Union, Asian Development 

Bank, International Fund for Agricultural Development, 

International Development Research Centre, 

Rockefeller Foundation, Food and Agriculture Organization 

of the United Nations, and agencies of the following countries: 

Australia, Austria, Belgium, Brazil, Canada, Denmark, 

France, Germany, India, Iran, Japan, Malaysia, Netherlands, 

Norway, People’s Republic of China, Republic of Korea, 

Republic of the Philippines, Spain, Sweden, Switzerland, 

Thailand, United Kingdom, United States, and Vietnam. 

The responsibility for this publication rests with the 

International Rice Research Institute. 

Suggested citation: 

Toriyama K, Heong KL, Hardy B, editors. 2005. Rice is 

life: scientific perspectives for the 21st century. Proceedings 

of the World Rice Research Conference held in Tokyo 

and Tsukuba, Japan, 4-7 November 2004. Los Baños (Philippines): 

International Rice Research Institute, and Tsukuba 

(Japan): Japan International Research Center for Agricultural 

Sciences. CD-ROM. 590 p. 

Copyright International Rice Research Institute 2005 

Mailing address: DAPO Box 7777, Metro Manila, 

Philippines 

Phone: +63 (2) 580-5600, 845-0563, 

844-3351 to 53 

Fax: +63 (2) 580-5699, 891-1292, 

845-0606 

Email: irri@cgiar.org 

Home page: www.irri.org 

Riceweb: www.riceweb.org 

Rice Knowledge Bank: 

www.riceknowledgebank.org 

Courier address: Suite 1009, Pacific 

Bank Building 6776 Ayala 

Avenue, Makati City, Philippines 

Tel. (63-2) 891-1236, 891-1174, 

891-1258, 891-1303 

CD-ROM cover design: Juan Lazaro IV 

CD-ROM production coordinator: George R. Reyes 

Layout and design: Ariel Paelmo 

Figures and illustrations: Ariel Paelmo 

ISBN 971-22-0204-6

Contents 

FOREWORD 

ACKNOWLEDGMENTS 

xiv 

xv 

Keynotes 

The power of settled life: rice farming as a lifestyle 2 

Ken-ichi Matsumoto 

The role of rice in the Japanese diet 3 

Yoshiko Kagawa 

A greeting 4 

Yoshinobu Shimamura 

Statement by Jacques Diouf 4 

Address by Imperial Highness Crown Prince Naruhito 5 

New technologies for rice production 6 

Gurdev S. Khush 

The changing economics and politics of rice: implications for food security, 7 

globalization, and environmental sustainability 

Joachim von Braun and María Soledad Bos 

Feeding the world: How much more rice do we need 21 

Vaclav Smil 

Development of sustainable agriculture from rice, water, 23 

and the living environment 

Riota Nakamura 

Research strategy for rice in the 21st century 26 

Ronald P. Cantrell and Gene P. Hettel 

SESSION 1 

The genus Oryza, its diversity, and its evolution 

Molecular phylogeny and divergence of the rice tribe Oryzeae, with special 40 

reference to the origin of the genus Oryza 

Song Ge, Ya-long Guo, and Qi-hui Zhu 

Eco-genetic diversification in the genus Oryza: implications for sustainable 44 

rice production 

Duncan Vaughan, Koh-ichi Kadowaki, Akito Kaga, and Norihiko Tomooka 

Toward a global strategy for the conservation of rice genetic resources 47 

Ruaraidh Sackville Hamilton and Ruth Raymond 

Genetic architecture and complexity in wild and cultivated rice 49 

Y. Sano 

Who was the mother of cultivated rice Differentiation of chloroplast genome 52 

structure among cultivated rice and ancestral wild species 

Koh-ichi Kadowaki 

Diverse mechanisms of low-temperature stress response in rice 54 

Ryozo Imai, Jiangqi Wen, Kentaro Sasaki, and Kiyoharu Oono 

Genetic diversity of Myanmar rice landraces 56 

Ye Tint Tun, K. Irie, T. Nagamine, Jhon Ba Maw, F. Kikuchi, and H. Fujimaki 

Contents 

iii

QTL analysis for eating quality in japonica rice 60 

H.G. Hwang, J.P. Suh, Y.C. Cho, S.J. Kwon, I.S. Choi, H.C. Hong, Y.G. Kim, 

M.K. Kim, H.C. Choi, and Y.T. Lee 

SESSION 2 

Structure and function of the rice genome 

The complete rice genome sequence and its application to breeding 66 

and genetics 

Takuji Sasaki 

Exploitation and use of naturally occurring allelic variations in rice 69 

Masahiro Yano, Yasunori Nonoue, Tsuyu Ando, Ayahiko Shomura, 

Takehiko Shimizu, Izumi Kono, Saeko Konishi, Utako Yamanouchi, 

Tadamasa Ueda, Shin-ichi Yamamoto, and Takeshi Izawa 

Allelic and functional diversity of stress-tolerance genes in rice 73 

Hei Leung, Hehe Wang, Jianli Wu, Ma. Elizabeth Naredo, 

Marietta Baraoidan, Alicia Bordeos, Suzette Madamba, Gay Carrillo, 

Jatinder Sangha, Zenna Negussie, Jill Cairns, Bin Liu, Yolanda Chen, 

Darshan Brar, Il Ryong Choi, Cassiana Vera Cruz, Renee Lafitte, 

Luca Comai, and Kenneth L. McNally 

Functional genomics by reverse genetics 76 

Gynheung An 

Tissue culture-induced mutations and a new type of activation tagging 78 

as tools for functional analysis of rice genes 

Hirohiko Hirochika 

Toward genome-wide transcriptional analysis in rice using MAS 80 

oligonucleotide tiling-path microarrays 

Lei Li, Xiangfeng Wang, Xueyong Li, Ning Su, Viktor Stolc, Bin Han, 

Jiayang Li, Yongbiao Xue, Jun Wang, and Xing Wang Deng 

Candidate gene characterization at the Pup1 locus: a major QTL increasing 83 

tolerance of phosphorus deficiency 

Matthias Wissuwa, Kristy Gatdula, and Abdelbagi Ismail 

Wrap-up of Session 2 85 

SESSION 3 

Opportunities and challenges of transgenic rice 

Overproduction of C 4 

enzymes in transgenic rice: an approach for improved 88 

photosynthesis and crop yield 

Mitsue Miyao-Tokutomi and Hiroshi Fukayama 

The uptake and translocation of minerals in rice plants 90 

Naoko K. Nishizawa 

Improving drought and cold-stress tolerance in transgenic rice 94 

Kazuko Yamaguchi-Shinozaki and Kazuo Shinozaki 

Broad-spectrum disease resistance in transgenic rice 97 

Motoshige Kawata 

Golden Rice and improvement of human nutrition 99 

Swapan Datta, Vilas Parkhi, Mayank Rai, Jing Tan, Niranjan Baisakh, 

Lina Torrizo, Editha Abrigo, Norman Oliva, Md. Alamgir Hossain, 

Russel Julian, Anindya Bandyopadhyay, and Karabi Datta 

Health-promoting transgenic rice suppressing life-related disease 102 

and type-I allergy 

Fumio Takaiwa 

Producing rice plants with a site-specific base change in the acetolactate 105 

synthase gene by chimeraplast-directed gene targeting 

A. Okuzaki and K. Toriyama 

Transgenic rice plants expressing wheat catalase show improved 108 

tolerance for chilling-induced damage in membranes 

Haruo Saruyama, Hidenori Onodera, and Matsuo Uemura 


iv 

Rice is life: scientific perspectives for the 21st century

SESSION 4 

Improving the rice yield potential 

Photosynthesis improvement in rice: Rubisco as a target for enhancing 114 

N-use efficiency 

Amane Makino 

Carbon metabolism to improve sink and source function 116 

R. Ohsugi 

Improving radiation-use efficiency: the acclimation of photosynthesis 118 

to high irradiance in rice leaves 

E.H. Murchie, S. Hubbart, S. Peng, and P. Horton 

Sink-source relationship and yield potential of rice: effect of ethylene 120 

on grain filling of late-flowering spikelets 

P.K. Mohapatra and Rashmi Mohapatra 

Impact of increased source capacity on rice yield: a case study 124 

with CO 2 

enrichment 

Toshihiro Hasegawa, Kazuhiko Kobayashi, Mark Lieffering, Han Yong Kim, 

Hidemitsu Sakai, Hiroyuki Shimono, Yasuhiro Yamakawa, Mayumi Yoshimoto, 

and Masumi Okada 

Possibilities for and constraints to improving canopy photosynthesis 127 

J.E. Sheehy, A.V. Elmido, P. Pablico, M.J.A. Dionora, and A.B. Ferrer 

Progress in breeding the new plant type for yield improvement: 130 

a physiological view 

Shaobing Peng, Rebecca C. Laza, Romeo M. Visperas, 

Gurdev S. Khush, and Parminder Virk 

Current status and prospects of rice breeding for increased yield in China 132 

Wan Jianmin 

A large-grain rice cultivar, Akita 63, exhibits high yield and N-use efficiency 135 

for grain production 

Tadahiko Mae, Ayako Inaba, Yoshihiro Kaneta, Satoshi Masaki, 

Mizuo Sasaki, and Amane Makino 

Using heterosis and hybrid rice to increase yield potential in China 138 

Xuhua Zhong, Shaobing Peng, Feng Wang, and Nongrong Huang 

Breeding and prevalence of japonica hybrid rice variety Mitsuhikari 140 

Atsushi Nakamura 

Dry-matter production and nitrogen distribution in a female-sterile line of rice 142 

Morio Kato, Sachio Maruyama, and Masao Yokoo 

Cytokinin as a causal factor of varietal differences in the reduction in leaf level 145 

of ribulose-1,5-bisphosphate carboxylase/oxygenase during senescence 

in rice plants 

Taiichiro Ookawa, Yukiko Naruoka, Ayumi Sayama, and Tadashi Hirasawa 

Measurement and evaluation of rice plant type by means of an image 148 

analysis method and a 3-D digitizer 

Masaaki Oka and Takashi Ogawa 


SESSION 5 

Broadening the gene pool and exploiting heterosis in cultivated rice 

Developing aerobic rice in Brazil 154 

B. da S. Pinheiro, E. da M. de Castro, O.P. de Moraes, and F. Breseghello 

Broadening the gene pool of rice through introgression from wild species 157 

D.S. Brar 

Mutation in seed reserves and its use for improving grain quality in rice 160 

Hikaru Satoh, Ken-ichi Ohtsubo, and Yasunori Nakamura 

Heterosis in rice for increasing yield, production efficiency, 162 

and rural employment opportunities 

Sant S. Virmani 

Contents 

v

Harnessing molecular markers in hybrid rice commercialization in the Philippines 166 

E.D. Redoña, L.M. Perez, L.R. Hipolito, V.E. Elec, I.A. Pacada, L.M. Borines, 

R.O. Solis, S.A. Ordoñez, and J. Agarcio 

Genetic evolution of Rf1 locus for the fertility restorer gene of BT-type CMS rice 170 

Tomohiko Kazama and Kinya Toriyama 

A polygenic balance model in yield components as revealed by QTL analysis 172 

in rice 

Wilhelm E. Hagiwara, Kazumitsu Onishi, Itsuro Takamure, and Yoshio Sano 


SESSION 6 

Trends in crop establishment and management in Asia 

Trends in crop establishment methods in Asia and research issues 178 

Sushil Pandey and Lourdes Velasco 

Direct seeding and weed management in Korea 181 

Soon-Chul Kim and Woon-Goo Ha 

Direct-seeding cultivation of rice in Japan: stabilization of seedling establishment 184 

and improvement of lodging resistance 

Satoshi Yoshinaga 

Direct seeding of aerobic rice in China 186 

Guang Hui Xie, Jun Yu, Jing Yan, Huaqi Wang, and Xiurong Zhu 

An overview on direct seeding for rice crop establishment in the Philippines 189 

Jovino L. de Dios, Evelyn F. Javier, Myrna D. Malabayabas, 

Madonna C. Casimero, Alex J. Espiritu 

Rice establishment in drought-prone areas of Bangladesh 193 

M.A. Mazid, M.A. Mortimer, C.R. Riches, A. Orr, B. Karmaker, 

A. Ali, M.A. Jabbar, and L.J. Wade 

Emerging issues in weed management of direct-seeded rice in Malaysia, 196 

Vietnam, and Thailand 

M. Azmi, D. V. Chin, P. Vongsaroj, and D.E. Johnson 

Changing from transplanted rice to direct seeding in the rice-wheat 198 

cropping system in India 

Y. Singh, Govindra Singh, David Johnson, and Martin Mortimer 

Seedling recruitment in direct-seeded rice: weed biology and water management 202 

A.M. Mortimer, O. Namuco, and D.E. Johnson 

The crop protection industry’s view on trends in rice crop establishment 205 

in Asia and their impact on weed management techniques 

Jean-Louis Allard, Kee Fui Kon, Yasuo Morishima, and Ruediger Kotzian 

Improved anchorage and bird protection with iron-coated seeds 209 

in wet direct seeding of rice crops 

Minoru Yamauchi 

Issues for integrated weed management and decision support 211 

in direct-seeded rice 

D.E. Johnson and A.M. Mortimer 

Control of Leptochloa chinensis (L.) Nees in wet-seeded rice fields in Sri Lanka 215 

Anuruddhka S.K. Abeysekera and U.B. Wickrama 

Village-level modeling of environment-friendly and appropriate 217 

technologies and practices for direct seeding 

Rowena G. Manalili, Bernard D. Tadeo, Emmanuel R. Tiongco, 

Wilfredo B. Collado, Rodolfo V. Bermudez, Constancio A. Asis, 

Jovino L. De Dios, Marvin F. Adap, Mario dela Cruz, Ulysses G. Duque, 

Leonardo V. Marquez, Cheryll B. Casiwan, Roy F. Tabalno, 

Placida C. Lanuza, and Belen C. Tejada 


vi 


SESSION 7 

Improving efficiency through innovations in mechanization 

Rice as a key resource for saving our planet 224 

Nobutaka Ito 

Mechanizing paddy rice cultivation in Korea 225 

Woo-Pung Park and Sang-Cheol Kim 

The status and prospects of rice production mechanization in China 228 

Li Yaoming 

The PhilRice-JICA rotary rice reaper: redesigning a technology for Filipino 229 

farmers and manufacturers 

Eulito Bautista, Manuel Jose Regalado, Arnold Juliano, Shuji Ishihara, 

Hiroyuki Monobe, Joel Ramos, and Leo Molinawe 

Long-mat seedling culture and the transplanting system: an innovative 232 

one-person operational technology for mechanical rice transplanting 

Hisashi Kitagawa, Akio Ogura, Kouhei Tasaka, Hiroyuki Shiratsuchi, 

and Mikio Yashiro 

High-precision autonomous operation using an unmanned rice transplanter 235 

Yoshisada Nagasaka, Yutaka Kanetani, Naonobu Umeda, and Takuo Kokuryu 

Precision-drilling methods for the direct sowing of rice in flooded paddy fields 238 

Yoh Nishimura, Kazunobu Hayashi, Takashi Goto, and Mitsuhiro Horio 

The rice direct-seeding system using multiple seed pellets in northern Tohoku 241 

Hiroyuki Sekiya, Hitoshi Ogiwara, Shoichi Kimura, Ryuji Otani, 

Yukio Yaji, Satoshi Morita, Tatsushi Togashi, and Hiroaki Watanabe 


SESSION 8 

Improving rice quality 

The chemical basis of rice end-use quality 246 

Kshirod R. Bhattacharya 

Improving rice grain quality in Thailand 248 

Kunya Cheaupun, Sunantha Wongpiyachon, and Ngamchuen Kongseree 

New tools for understanding starch synthesis 251 

Melissa Fitzgerald, Jeffrey Castro, Rosa Paula Cuevas, and Robert Gilbert 

Textural differences between indica and japonica varieties in cooked rice 253 

Keiko Hatae, Sonoko Ayabe, and Midori Kasai 

Biological efficacy of consuming rice biofortified with iron 256 

J. Haas, J. Beard, A. Del Mundo, G. Gregorio, L.M. Kolb, and A. Felix 

Radical-scavenging activity of red and black rice 256 

Tomoyuki Oki, Mami Masuda, Saki Nagai, Miwako Take’ichi, 

Mio Kobayashi, Yoichi Nishiba, Terumi Sugawara, Ikuo Suda, 

and Tetsuo Sato 

A rice mutant with enhanced amylose content in endosperm derived 260 

from low-amylose variety Snow Pearl: isolation and characterization 

Yasuhiro Suzuki, Hiro-Yuki Hirano, Yoshio Sano, Kazuo Ise, 

Ushio Matsukura, Noriaki Aoki, and Hiroyuki Sato 

The role of the water-soluble fraction in rice pasting behavior 261 

Tadashi Yoshihashi, Eizo Tatsumi, Vipa Surojanametakul, 

Patcharee Tungtrakul, and Warunee Varanyanond 


SESSION 9 

Developing new uses of rice 

Overview of rice and rice-based products 268 

Bienvenido O. Juliano 

Sterilization effect of electrolyzed water on rice food 270 

Seiichiro Isobe, Chang-yong Lee, and Kyoichiro Yoshida 

Current status of varietal improvement and use of specialty rice in Korea 272 

Hae Chune Choi 

Contents 

vii

Processed novel foodstuffs from pregerminated brown rice 275 

by a twin-screw extruder 

Ken’ichi Ohtsubo, Tomoya Okunishi, and Keitaro Suzuki 

High-pressure food processing of rice and starch foods 278 

Rikimaru Hayashi 

Developing novel processes for incorporating the unique nutritional 280 

and functional properties of rice into value-added products 

Elaine T. Champagne, Harmeet S. Guraya, Frederick F. Shih, 

and Ranjit S. Kadan 

Dehydrin proteins in rice bran 283 

Michiko Momma 

Physicochemical properties of modified rice flour and its use 285 

for processed food 

T. Takahashi, M. Miura, N. Ohisa, K. Mori, and S. Kobayashi 

Introducing soybean β-conglycinin genes into rice to improve nutritional 288 

and physiological value 

Takayasu Motoyama, Nobuyuki Maruyama, Takahiko Higasa, 

Masaaki Yoshikawa, Fumio Takaiwa, and Shigeru Utsumi 


SESSION 10 

Postharvest technology for efficient processing and distribution of rice 

Postharvest technology for rice in India: a changing scenario 294 

Pallab Kumar Chattopadhyay 

Development of a far-infrared radiation dryer for grain 296 

Yasuyuki Hidaka, Kotaro Kubota, and Tomohiko Ichikawa 

Status of rice milling and use of by-products 299 

Naoto Shimizu, Yuji Katsuragi, and Toshinori Kimura 

Advanced application technology of rice bran: preparation of ferulic acid 301 

and its applications 

Hisaji Taniguchi, Eisaku Nomura, and Asao Hosoda 

New movements regarding the safety of rice and residual agrochemical inspection 304 

Yukio Hosaka 

Impact of infrastructure on profitability and global competitiveness 307 

of rice production in the Philippines 

Rowena G. Manalili and Leonardo A. Gonzales 

Development of an on-farm rice storage technique using fresh chilly air and 310 

preservation of high-quality rice 

Shuso Kawamura, Kazuhiro Takekura, and Kazuhiko Itoh 

Life-cycle inventory analysis of local parboiling processes 313 

Poritosh Roy, Naoto Shimizu, Takeo Shiina, and Toshinori Kimura 

Using genetics to create changes in rice cooking, processing, storage, 315 

and health-beneficial properties 

Christine Bergman 

The role of proteins in textural changes in aged rice 318 

Toshihisa Ohno, Takahiro Kaneko, and Naganori Ohisa 

Postharvest technology of rice: the role of farm women in storing grains 320 

with different storage practices 

P. Sumathi and M.N. Budhar 


SESSION 11 

Enhancing the multifunctionality of rice systems 

Multifunctional roles of paddy irrigation in monsoon Asia 324 

Takao Masumoto 

Accounting for culture in paddy cultivation: toward a broader definition 327 

of “livelihood” 

David Groenfeldt 

viii 


The need to keep irrigated rice culture sustainable 329 

Nyoman Sutawan 

Multifunctional roles of irrigation water and rice fields in Dujiangyan, China 331 

Liu Yulong, Yamaoka Kazumi, and Ren Yonghuai 

Decommissioning of paddy lands in the Wet Zone of Sri Lanka: 334 

some effects on food security and ecosystems 

Kusum Athukorala and Missaka Hettiarachchi 

True agro-biodiversity depending on irrigated rice cultivation as a multifunction 337 

of paddy fields 

Kazumasa Hidaka 

Enhancing the multifunctionality of floating-rice farming in the Chao Phraya 340 

delta of Thailand 

Yuyama Yoshito, Ogawa Shigeo, and Ueda Tatsuki 

Demonstration program on multifunctionality of paddy fields 343 

in the northeastern region of Thailand 

Chatchai Boonlue 

Remediation effect of rice terraces for strong acidic and nitrate-rich 346 

effluent water from tea bush areas in the Makinohara plateau 

of Shizuoka, Japan 

Kiyoshi Matsuo, Yuhei Hirono, and Kunihiko Nonaka 

Where are the hot spots of RDB plants in rice paddy fields A GIS-based 347 

analysis using information accumulated at a natural history museum 

Takuya Mineta, Kenji Ishida, and Takashi Iijima 


SESSION 12 

Conservation of soil, water, and environment in rice culture 

Paddy soils around the world 352 

K. Kyuma 

Sustainability of paddy soil fertility in Vietnam 354 

Ngo Ngoc Hung, Nguyen Bao Ve, Roland J. Buresh, Mark Bayley, 

and Takeshi Watanabe 

Managing soil fertility for sustainable rice production in northeast Thailand 357 

Kunnika Naklang 

Site-specific nutrient management and the sustainability of phosphorus 360 

and potassium supply in irrigated rice soils of Asia 

C. Witt, A. Dobermann, R. Buresh, S. Abdulrachman, H.C. Gines, 

R. Nagarajan, S. Ramanathan, P.S. Tan, and G.H. Wang 

Ecological engineering for sustainable rice production and the restoration 363 

of degraded watersheds in West Africa 

Toshiyuki Wakatsuki, Md. Moro Buri, and Oluwarotimi O. Fashola 

Nitrogen cycling under the rice-wheat rotation and environmental effects 367 

Jian-guo Zhu, Xiaozhi Wang, Zu-cong Cai, Ren Gao, and Yasukazu Hosen 

Urea deep placement increases yield and saves nitrogen fertilizer 369 

in farmers’ fields in Bangladesh 

W.T. Bowen, R.B. Diamond, U. Singh, and T.P. Thompson 

New fertilizer management to maximize yield and minimize environmental 372 

effects in rice culture 

Masahiko Saigusa 

Does anaerobic decomposition of crop residues impair soil nitrogen 374 

cycling and yield trends in lowland rice 

D.C. Olk, K.G. Cassman, M.M. Anders, K. Schmidt-Rohr, and J.-D. Mao 

Influence of the paddy-upland rotation on soil physico-chemical properties 377 

and crop growth in the Honam Plain of Korea 

Lee Deog-Bae, Yang Chang-Hyu, Ryu Chul-Hyun, Lee Kyeong-Bo, 

and Kim Jae-Duk 

A decrease in soil fertility and crop productivity by succession of 379 

the paddy-upland rotation 

Hirokazu Sumida 

Contents 

ix

Promising technologies for reducing cadmium contamination in rice 381 

Satoru Ishikawa 

N uptake inhibition of the rice plant in flooded soils receiving wheat straw 384 

Fukuyo Tanaka 

Stable isotope ratios of hydrogen and oxygen in paddy water affected 386 

by evaporation 

Yohei Hamada, Shiho Yabusaki, Norio Tase, and Ichiro Taniyama 

Erosion control by sawah in comparison to other land-use systems 388 

Fahmuddin Agus and Irawan 


SESSION 13 

Farmers’ participatory approaches to facilitate adoption of improved technology 

How does a farmer accept a new technology 394 

Fujihiko Tozawa 

Farmer participatory evaluation of nitrogen management technology: the case 395 

of the leaf color chart in West Bengal, India 

B. Bagchi, M.Z. Abedin, and S.K.T. Nasar 

A farmer participatory approach in the adaptation and adoption of controlled 397 

irrigation for saving water: a case study in Canarem, Victoria, Tarlac, 

Philippines 

F.G. Palis, M. Hossain, B.A.M. Bouman, P.A.A. Cenas, R.M. Lampayan, 

A.T. Lactaoen, T.M. Norte, V.R. Vicmudo, and G.T. Castillo 

Companion modeling and multi-agent systems for collective learning 401 

and resource management in Asian rice ecosystems 

F. Bousquet and G. Trébuil 

Participatory approaches for improving rice breeding in the Mekong Delta 404 

of Vietnam 

Nguyen Ngoc De and Kotaro Ohara 

IRRI’s approach to participatory research for development: advances 408 

and limitations 

Thelma R. Paris and M. Zainul Abedin 

A participatory approach for building sustainable rice-farming systems 411 

in the reclaimed farmland of Ogata, Japan 

Yoshimitsu Taniguchi and Satoru Sato 

Rice farmers’ participatory research has played a key role 412 

in implementing the System of Rice Intensification 

Dandu Jagannadha Raju 


SESSION 14 

Potentials for diversification in rice-based systems to enhance rural livelihoods 

Agricultural diversification in Asia: opportunities and constraints 420 

Prabhu Pingali 

Consequences of technologies and production diversification for 422 

the economic and environmental performance of rice-based farming 

systems in East and Southeast Asia 

Huib Hengsdijk, Marrit van den Berg, Reimund Roetter, Wang Guanghuo, 

Joost Wolf, Lu Changhe, and Herman van Keulen 

Rural poverty and agricultural diversification in Thailand 425 

Alia Ahmad and Somporn Isvilanonda 

Sustaining higher efficiency in rice production 428 

Amelia S. delos Reyes, Arelene Julia B. Malabayabas, 

and Mercedita A. Sombilla 

Determinants of agricultural diversification in Vietnam: changes 432 

at the farm level in the Mekong and Red River deltas 

Magnus Jirström and Franz-Michael Rundquist 

x 


Growth of the rural nonfarm economy in Bangladesh: determinants and impact 436 

on poverty reduction 

Mahabub Hossain 

SESSION 15 

Challenges to expanding rice production in unfavorable environments 

In vitro selection of somaclonal and gametoclonal variants for salt tolerance 442 

in rice 

Nguyen Thi Lang, Dang Minh Tam, Hiromi Kobayashi, and Bui Chi Buu 

Submergence damage in rice and challenges in expanding the crop’s 445 

adaptability to submerged conditions in West and Central Africa 

Koichi Futakuchi 

Ecological, morphological, and physiological aspects of drought adaptation 448 

of rice in upland and rainfed lowland systems 

Shu Fukai and Akihiko Kamoshita 

Managing iron toxicity in lowland rice: the role of tolerant genotypes 452 

and plant nutrients 

Kanwar L. Sahrawat 

Soil acidity and related problems in upland rice in the tropics 454 

Kensuke Okada and Matthias Wissuwa 

The physiological foundation of crop breeding for stress environments 456 

A. Blum 

Expression of a serine protease during microsporogenesis in rice 458 

Kentaro Kawaguchi, Naoshi Dohmae, Shuichi Matsuba, 

Hideyuki Funatsuki, Yutaka Sato, and Masao Ishimoto 

Responses to chilling temperature at the early stage of development 460 

in rice: geographical clines and genetic bases as revealed by QTL analysis 

Kazumitsu Onishi, Noriko Ishigoh-Oka, Mieko Adachi, and Yoshio Sano 

QTL analysis on plasticity in lateral root development in response 463 

to water stress in the rice plant 

Hong Wang, Yoshiaki Inukai, Akihiko Kamoshita, Len Wade, 

Joel Siopongco, Henry Nguyen, and Akira Yamauchi 


SESSION 16 

Pest management with minimal environmental stress 

Integrated biodiversity management (IBM) in rice fields 468 

Keizi Kiritani 

Habitat manipulation in sustainable pest management in the rice ecosystem 470 

of the Yangtze River Delta 

Xiaoping Yu, Jianming Chen, Zhongxian Lu, Xusong Zheng, Hongxin Xu, 

Juefeng Zhang 

Evaluating augmentative releases of the mirid bug Cyrtorhinus lividipennis 473 

to suppress brown planthopper Nilaparvata lugens in open paddy fields 

Masaya Matsumura, Satoru Urano, and Yoshito Suzuki 

Developing a rice production system for sustainable pest management 475 

T.W. Mew 

Evaluation of a leaf blast simulation model (BLASTMUL) for rice multilines 477 

in different locations and cultivars, and effective blast control using 

the model 

T. Ashizawa, M. Sasahara, A. Ohba, T. Hori, K. Ishikawa, Y. Sasaki, 

T. Kuroda, R. Harasawa, K.S. Zenbayashi, and S. Koizumi 

Association of candidate defense genes with quantitative resistance to rice blast 479 

and in silico analysis of their characteristics 

G. Carrillo, J. Wu, B. Liu, N. Sugiyama, I. Oña, M. Variar, B. Courtois, 

J.E. Leach, P.H. Goodwin, H. Leung, and C.M. Vera Cruz 

Transferring resistance genes among different cereal species 482 

Bingyu Zhao, Shavannor Smith, Jan Leach, and Scot Hulbert 

Contents 

xi

Screening of allelopathic activity from rice cultivars by bioassay and field test 484 

Yoshiharu Fujii, Hiroshi Araya, Syuntarou Hiradate, and Kaoru Ebana 

Evaluating near-isogenic lines with QTLs for field resistance to rice blast 487 

from upland rice cultivar Sensho through marker-aided selection 

Norikuni Saka and Shuichi Fukuoka 


SESSION 17 

Rice supply and demand 

International trade in rice: recent developments and prospects 492 

Concepcion Calpe 

Rice in the world at stake 494 

Shoichi Ito 

Rice consumption in China: Can China change rice consumption 497 

from quantity to quality 

Chien Hsiaoping 

Surplus rice supply in Asia 499 

Vichai Sriprasert 

Review of existing global rice market models 502 

Eric J. Wailes 

Household rice consumption in Japan: quantity and price by income 505 

while controlling for household types 

Kimiko Ishibashi, Yoshinobu Kono, and Yuuji Ooura 


SESSION 18 

Impact of globalization on rice farmers 

Impacts of distorted trade policies on rice productivity 510 

Manitra A. Rakotoarisoa 

A new strategy for group farming in Japan 513 

Kyoichi Miyatake 

The role of the rice economy after the implementation of agricultural 515 

policy reform and trade liberalization from the perspective of farmers: 

the case of rice farmers in Java, Indonesia 

Roosiana 

El Niño sensitivity, resource endowment, and socioeconomic characteristics: 517 

the case of wetland rice in Java, Indonesia 

Shigeki Yokoyama and Bambang Irawan 

Global competitiveness of medium-quality Indian rice: a PAM analysis 520 

B.V. Chinnappa Reddy, M.S. Raghavendra, and Lalith Achoth 

Behavior and strategies of Japanese rice producers under globalization 523 

Masaki Umemoto 

The future perspective of upland rice farmers in Indonesia in the era of 525 

globalization 

Yusman Syaukat and Sushil Pandey 

Impact of globalization on rice farmers in Thailand 527 

Rangsan Pitipunya 

The rice economy and rice policy in China 530 

Li Ninghui 


SESSION 19 

Climate change and rice production 

Effects of elevated atmospheric CO 2 

concentration and increased temperature 536 

on rice: implications for Asian rice production 

T. Horie, H. Yoshida, S. Kawatsu, K. Katsura, K. Homma, and T. Shiraiwa 

Monitoring rice growth and development using a crop model: the case 539 

of northern Japan 

Masaharu Yajima 

xii 


Coping with climate variability and change in rice production systems 542 

in the Philippines 

Felino P. Lansigan 

Effect of elevated CO 2 

on nutrient uptake and nutritional conditions of rice 545 

Jianguo Zhu 

Effect of soil properties, substrate addition, microbial inoculation, and 546 

sonication on methane production from three rice topsoils and subsoils 

in the Philippines 

Sudip Mitra and Deepanjan Majumdar 

Modeling the effects of farming management alternatives on greenhouse 551 

gas emissions: a case study for rice agriculture in China 

Changsheng Li, Steve Frolking, Xiangming Xiao, Berrien Moore III, 

Steve Boles, Yu Zhang, Jianjun Qiu, Yao Huang, William Salas, 

and Ronald Sass 

Impact of rising atmospheric CO 2 

on CH 4 

emissions from rice paddies 555 

Weiguo Cheng, Kazuyuki Yagi, Kazuyuki Inubushi, Kazuhiko Kobayashi, 

Hidemitsu Sakai, Han-Yong Kim, and Masumi Okada 

The detrimental effect of tropospheric O 3 

on lowland rice is ameliorated 557 

by elevated CO 2 

T. Ishioh, K. Kobori, and K. Imai 

Effects of topdressing on grain shape and grain damage under high temperature 560 

during ripening of rice 

Satoshi Morita, Osamu Kusuda, Jun-ichi Yonemaru, Akira Fukushima, 

and Hiroshi Nakano 


SESSION 20 

Improving rice productivity through IT 

Data mining using combined yield and quality maps of paddy fields 566 

Tadashi Chosa 

A wireless sensor network with Field-Monitoring Servers and MetBroker 568 

in paddy fields 

Masayuki Hirafuji, Tokihiro Fukatsu, Hu Haoming, Takuji Kiura, 

Tominari Watanabe, and Seishi Ninomiya 

Prediction of airborne immigration of rice insect pests 571 

Akira Otuka, Tomonari Watanabe, Yoshito Suzuki, Masaya Matsumura, 

Akiko Furuno, and Masamichi Chino 

Distance education and eLearning for sustainable agriculture: 574 

lessons learned 

Robert T. Raab and Buenafe R. Abdon 

The Rice Knowledge Bank 576 

Mark Bell and David Shires 

Integrated information systems for crop research and improvement 579 

Graham McLaren, Arllet Portugal, Alexander Cosico, William Eusebio, 

Teri Ulat, May Ann Sallan, Victor Jun Ulat, and Richard Bruskiewich 

Using APAN for content delivery: possibilities for the CGIAR 583 

Paul O’Nolan 

A decision support system for site-specific nitrogen management 587 

of paddy rice 

Ryouji Sasaki, Kazunobu Toriyama, and Yoichi Shibata 


Contents 

xiii

Foreword 

Just after World War II, rapid population growth with limited rice production led 

experts to predict starvation in Asia. On its own, the Food and Agriculture Organization 

(FAO) of the United Nations had declared 1966 the Year of Rice and numerous 

countries took measures to improve production, marketing, milling, and nutrition. 

Conferences were organized and scientific research and technology were stimulated. 

Rice is now the staple food for over half of the world’s population but its production 

faces many constraints under conditions of increasing world population and diminishing 

resources of water and land. 

The United Nations launched the International Year of Rice 2004 on 31 October 

2003. This is the second time that the United Nations has paid such a special 

tribute to rice, the only food crop honored twice. Rice is the single most important 

employment and income source for the rural poor. Rice will play a significant role in 

meeting the important UN Millennium Development Goal of poverty reduction in 

the world. Besides being an essential food, rice is also an important factor in enriching 

culture, lifestyles, and ecosystem functions. It is thus fitting that the United Nations 

pronounced 2004 the International Year of Rice to emphasize the important 

roles rice plays in the livelihoods and culture of humankind. Rice is a symbol of 

cultural identity, global unity, and life. 

To mark the Year of Rice 2004, the Ministry of Agriculture, Forestry, and Fisheries 

(MAFF) of Japan, research organizations affiliated with MAFF, and the International 

Rice Research Institute (IRRI) jointly financed and organized the World 

Rice Research Conference (WRRC) that was held in Tokyo and Tsukuba, Japan, 4-7 

November 2004. The WRRC had two parts: the Tokyo Opening Ceremony and Symposium 

in the Akasaka Prince Hotel graced by His Imperial Highness Crown Prince 

Naruhito of Japan on 4 November and the International Rice Research Conference at 

the Tsukuba Congress Center on 5-7 November. 

The WRRC attracted 1,274 participants from 43 nations, who presented 190 

scientific papers and 302 posters in 20 sessions and 6 workshops. Scientific themes 

covered genomics to climate change and involved scientists with expertise in genes 

to ecosystems. Such a broad range of subjects discussed during the conference makes 

the WRRC one of the most significant scientific events of the last few decades. Scientists 

presented their latest concepts, research findings, and products, which are 

captured in the topics presented in this proceedings. The proceedings contains the 

state of the art in rice science and production that we hope will be useful to rice 

scientists, extension specialists, development agents, and policymakers who will use 

this to better the lives of all humans, but especially those of poor farmers and consumers. 

AKINORI NOGUCHI 

Vice president of the Japan International Research Center 

for Agricultural Sciences and chairman 

of the WRRC 2004 Organizing Committee 

xiv 


Acknowledgments 

The success of the World Rice Research Conference (WRRC) was due largely to the 

constant endeavor of many individuals and organizations. The framework of this 

conference was set up by the executive committee composed of the Division of International 

Research in the Secretariat of Agriculture, Forestry, and Fisheries Research 

Council of the Ministry of Agriculture, Forestry, and Fisheries (MAFF) and the presidents 

of co-sponsoring research institutes as well as the delegate from the International 

Rice Research Institute (IRRI). 

The organizing committee consisted of A. Noguchi and Y. Morokoka (Japan 

International Research Center for Agricultural Sciences, JIRCAS) as chairs, with 

seven representatives from the co-sponsoring research institutes: T. Imbe (National 

Agricultural and Bio-oriented Research Organization, NARO), K. Okuno (National 

Institute of Agrobiological Sciences; NIAS), M. Oka (National Institute for Agro- 

Environmental Sciences, NIAES), Y. Tsutsui (National Institute of Rural Engineering, 

NIRE), T. Imai (National Food Research Institute, NFRI), Y. Yunoki (MAFF), 

and K.L. Heong (IRRI). Special thanks are due to M. Iwamoto, former president of 

JIRCAS, and E. Miwa, president of NARO, for their valuable suggestions and constant 

encouragement. 

The secretariat was successfully handled by JIRCAS personnel headed by O. 

Ito, with R. Ikeda, T. Kumashiro, K. Toriyama, K. Yasunobu, T. Uetani, H. Fujii, S. 

Tsuchiya, R. Saito, H. Tanaka, T. Hatta, Y. Hamada, and H. Miura. For Web management, 

E. Hettel (IRRI) and H. Miura (JIRCAS) contributed significantly. Special 

appreciation goes to IRRI personnel: K.L. Heong for coordinating with IRRI and the 

JIRCAS secretariat and giving much useful advice for the success of the WRRC, D. 

Macintosh for press releases, and B. Hardy for editorial work on the manuscripts of 

this large proceedings. 

The conveners of each session organized and provided a first review of the 

papers from their respective session contributers. Without their strong efforts, the 

WRRC could not have succeeded. These are K. Okuno and R.S. Hamilton for S1; T. 

Sasaki, H. Hirochika, and H. Leung for S2; F. Takaiwa, S. Oka, and S. Datta for S3; 

M. Kondo, H. Ikehashi, and S. Peng for S4; Y. Fukuta, D. Mackill, and R. Ikeda for 

S5; M. Yamauchi and D. Johnson for S6; N. Itokawa and J. Rickman for S7; U. 

Matsukura for S8; K. Otsubo for S9; T. Kimura and J. Rickman for S10; Y. Tsutsui 

and K. Yamaoka for S11; M. Saito, S. Ito, and T. Nozoe for S12; T. Paris and J. 

Caldwell for S13; M. Hossain for S14; S. Tobita and R. Lafitte for S15; C. Vera Cruz 

and Y. Suzuki for S16; O. Koyama and D. Dawe for S17; M. Umemoto and D. Dawe 

for S18; T. Imagawa, T. Hasegawa, Y. Hayashi, and J. Sheehy for S19; and S. Ninomiya 

and M. Bell for S20. 

This conference was supported by different academic societies and it attracted 

many participants of different disciplines. These were the Japan Society of Breeding, 

Crop Science Society of Japan, Japanese Society of Soil Science and Plant Nutrition, 

the Japanese Society of Irrigation, Drainage, and Reclamation Engineering, the Japanese 

Society of Agricultural Machinery, the Japanese Society of Applied Entomology 

and Zoology, the Phytopathological Society of Japan, Pesticide Science Society 

of Japan, the Farm Management Society of Japan, the Japanese Society for Food 

Contents 

xv

Science and Technology, the Society of Agricultural Meteorology of Japan, Japanese 

Society for Tropical Agriculture, and Science Council of Japan (National Committee 

for Rural Planning). 

Financial support was given by the Japan Science Promotion Society. Special 

thanks are also due to the Tsuchiura Tsukuba Convention Bureau and the International 

Congress Center (Epochal Tsukuba). 

xvi 


Keynotes

The power of settled life: rice farming as a lifestyle 

Ken-ichi Matsumoto 

Folklorist Yanagida Kunio (1875-1962), known as the father 

of Japanese folklore study, once raised the question, “What is 

Japan” In his view, insularity and rice cultivation are the two 

key factors composing Japanese ethnicity. In other words, 

Yanagida sees our Japanese way of life as the result of being 

rice farmers on an island. 

I would like to add another factor to these two: settled 

life, rooted in a community and a plot of land. 

Some may argue that rice cultivation, either paddy rice 

or dry rice, is naturally coupled with a settled life and is too 

obvious to mention as a new factor. However, this is not so. 

If you see rice cultivation as a part of life deeply connected 

to ethnicity, being a rice farmer means living a settled 

life on specific parcels of land. However, if you see rice farming 

as an industry that generates profits, farmers may choose 

to plant strawberries instead of rice when profits are low, as 

happens in California. Furthermore, they may decide to sell 

their land to Disneyland and move to other areas. In this case, 

rice cultivation is not inevitably linked to a settled lifestyle. 

California rice farmers are not the only ones who choose 

alternate lifestyles. Another good example is Indonesia. Indonesia 

achieved 100% rice self-sufficiency in 1997. Soon, the 

farmers recognized that shrimp cultivation would triple their 

income and they turned their paddies into shrimp cultivation 

ponds. As a result, the self-sufficiency rate of rice dropped to 

80%. 

Now, let us focus on settled rice farming as an ethnic 

lifestyle. The reward of your life as a farmer is to sustain your 

paddies and paddy terraces passed down from your ancestors. 

Issho-kenmei, a Japanese idiom for “do your best,” literally 

means “maintain a piece of land for all your worth.” This idea 

has been supported as an ethnic ethos of Japanese. This describes 

why Japanese pay more respect to the process of “doing 

your best” than to getting good results or high profits. 

Sharing the ethos of issho-kenmei, people living in an 

area work together to dig drainage ditches, build irrigation systems, 

and weed the whole village. The working group unavoidably 

encourages the development of a community system. 

The Japanese call the community system work kou. Similar 

systems developed in Okinawa are called yui, and ke in 

Korea and ban in South China. 

In these community work systems, innovations and processes 

to increase productivity develop to improve ancestral 

farming land. In other words, the whole community is encouraged 

to exert an effort to breed crops and control the crops’ 

quality, as well as construct drainage and irrigation systems 

for the community and weed the rice paddies. 

As another example, to harvest more crops on the same 

areas of land that their ancestors reclaimed, farmers must breed 

rice stocks more resistant to extreme temperature, drought, and 

disease. They also carry out thorough quality control by removing 

pests and weeds from the rice field, improving the 

rice paddy to create efficient irrigation and drainage conditions, 

and figuring out what the most suitable spaces are between 

rice plants. 

This lifestyle is very different from that of Western countries, 

where productivity increased through territorial expansion. 

The Western ranchers’ way of adding new pastures outside 

the towns, in my opinion, demonstrates the principle of 

the “power to expand.” This naturally extends expansion to 

products and innovations in fields such as traffic technology, 

transportation, communication, and military affairs. 

On Earth today, we do not have any new land to expand 

onto. The natural environment is worsening, the population is 

exploding, and a food crisis is just around the corner. 

If we recognize this reality, isn’t this the time for us to 

remember “the power of inner accumulation,” traditionally cultivated 

by rice farming It is the wisdom of various ethnic 

groups to live on the same land for generations, thus increasing 

wealth in a village, and in a rice paddy, in a single rice 

grain. Rice farming will provide the core motive for the settled 

lifestyle we now seek. 

Notes 

Author’s address: Professor, Reitaku University. 

2 Rice is life: scientific perspectives for the 21st century

The role of rice in the Japanese diet 

Yoshiko Kagawa 

Rice is the starting point of Japanese food culture. In a traditional 

Japanese diet, the carbohydrates contained in rice are 

the main energy source. The average Japanese eats about 594 

kcal of rice and rice products per day, which is 31% of the 

average energy intake. Rice contains 6.1 g of protein per 100 

g, and, though this level is not so high, rice protein is superior 

in quality to wheat protein. Therefore, rice alone can support a 

person if that person eats enough of it. Even today, the Japanese 

receive 12% of their daily protein from rice, and decades 

ago many sustained their lives by eating only rice. That is why 

the Japanese call it shu-shoku, the principal food. 

The conventional Japanese diet also contains a variety 

of fuku-shoku, the side dishes served along with rice. These 

wholesome foods are economical, too. Rice at breakfast reduces 

the day’s food expense compared to popular Westernstyle 

breakfasts. Furthermore, a rice-based diet is healthier in 

nutritional balance. 

Although rice is an excellent food, a large problem involving 

a rice-rich diet exists: beriberi, a thiamine deficiency. 

This disease emerged with the grain polishing that produces 

white rice, intended to improve the digestion and taste of 

cooked rice. The process removes the rice germ that contains 

most of the naturally occurring vitamin B 1 . Some 0.42 mg of 

vitamin B 1 is required for the metabolism of 1,000 kcal of 

carbohydrates. Therefore, for people who try to sustain their 

diet mainly by rice, the relative shortage in vitamin B 1 develops 

into an absolute deficiency and causes beriberi. 

An adult male needs 1.1 mg of vitamin B 1 per day, and 

an adult female needs 0.7 mg. This is why young men who 

consume more food and energy are more likely to develop 

beriberi. In the 1920s and ’30s, when the Japanese diet depended 

much more on rice, beriberi killed more than 10,000 

people and terrified Japanese called it a “nation-ruining disease.” 

Beriberi is common to ethnic groups that eat rice as 

their principal food, and has often occurred in Southeast Asian 

countries, including Indonesia. 

It is easy to solve the thiamine problem: just change to 

“rice with the embryo,” produced by polishing the grain almost 

white but leaving the embryo. If one prefers white rice, 

eating vitamin B 1 -enriched rice and/or accompanying dishes 

rich in vitamin B 1 prevents beriberi. Getting enough vitamin 

B 1 is the first and major issue in a rice-based diet. 

The next point we must keep in mind in a rice-based diet 

is the imbalance in nutrition when one eats rice too much. White 

rice is most appetizing when accompanied by salty dishes. 

Because of this, one is more likely to develop unhealthy eating 

habits—consuming a small amount of salty food and eating 

a lot of white rice, thinking that white rice is the principal 

food. This eating pattern results in an overconsumption of salt. 

Sodium causes high blood pressure and a higher chance of 

stroke. In addition, this diet pattern often lacks food made from 

animal products that supply a certain amount of high-quality 

protein necessary to avoid a range of diseases. 

In the 1930s, physicians Shozo Kagawa and Aya Kagawa 

treating beriberi patients found a method to produce “rice with 

the embryo.” They promoted “rice with the embryo” to prevent 

beriberi, and founded an educational institute that later 

became the Kagawa Nutrition University. To help people remember 

the benefits of a balanced “rice with the embryo” diet, 

they made a campaign slogan: the principal food is “rice with 

the embryo,” accompanied by one fish, one bean, and four 

vegetables. This reminded people to eat 100 g of fish, 100 g of 

bean and bean products, and 400 g of vegetables a day. 

The idea of a food guideline based on food groups was 

new at the time. This eating method was widely accepted by 

the educated homemakers of Japan. After World War II, milk 

and milk products were introduced to school lunch programs. 

Children’s better health proved its effectiveness and milk was 

added to the list. This is how the well-known and popular “four 

food groups” method we use today developed. 

A rice-based diet has another advantage: rice is eaten as 

a form of grain. Grains are digested and taken into the body 

more slowly than milled cereals. Therefore, in spite of low-fat 

and low-calorie meals, one has a long-lasting feeling of satiety 

and less insulin is consumed. The key to the lower rate of diabetes 

in the Japanese population in the past seems to be rice. 

Rice is a delicious and handy food. I would like to emphasize 

that it is an excellent food if you avoid too much polishing and 

pay close attention to good nutritional balance. 

Notes 

Author’s address: President of Kagawa Nutrition University, Japan. 

Keynotes 3

A greeting 

Yoshinobu Shimamura 

The United Nations has designated 2004 as the International 

Year of Rice (IYR) and countries around the world are joining 

hands to hold events to raise awareness about rice, which plays 

a major role in eliminating hunger and poverty in developing 

nations. Japan kicked off this special year with the International 

Symposium to Celebrate IYR, which was sponsored by 

the Ministry of Agriculture, Forestry, and Fisheries in January. 

A variety of other events, spearheaded by the Japanese Committee 

for the International Year of Rice 2004, are being staged 

throughout the nation this year. 

Rice is not only the staple food of monsoon Asian; it 

also plays an important role in regional culture because of the 

idyllic agricultural villages that form around the beautiful paddies. 

The Japanese people’s lifestyle and society are based on 

rice and rice cultivation and have advanced thanks to the tireless 

efforts of farmers and researchers. Rice research has produced 

many technologies and varieties of rice and contributed 

greatly to the furtherance of our nation’s agriculture. 

Rice is the staple food of more than half of the world’s 

population, and about 1 billion households depend on rice cultivation 

for employment and their main source of livelihood. 

At the same time, approximately 800 million people in developing 

nations suffer from hunger. Most of the world popula- 

tion that suffers from hunger and poverty lives in Asia and 

Africa. New research findings related to rice are expected to 

help eradicate hunger and poverty among farmers and others 

in developing nations. 

To address these issues, Japan has provided cooperation 

to improve rice varieties and establish cultivation technologies 

in Asia. Our nation also plays a leadership role in rice 

research. In addition to working alongside ten other regions 

and countries to sequence the rice genome, Japan is engaged 

in research on New Rice for Africa (NERICA) in cooperation 

with international research institutes. To achieve the UN Millennium 

Development Goals, Japan will continue to promote 

research to meet international expectations. 

By holding the World Rice Research Conference during 

this memorable year, I am confident that we will strengthen 

cooperation among rice researchers throughout the world and 

further stimulate research to eradicate world hunger and poverty 

and solve environmental problems. 

Notes 

Author’s address: Minister of Agriculture, Forestry, and Fisheries, 

Japan. 

Statement by Jacques Diouf 

It is a privilege to address the distinguished participants of the 

World Rice Research Conference. I wish to thank the Ministry 

of Agriculture, Forestry, and Fisheries in Japan for organizing 

this important event and for its efforts to promote the International 

Year of Rice, 2004. 

It is a pleasure to be with you today as we draw the attention 

of the world community to the vital role of rice in food 

security and poverty alleviation. I am confident that this opportunity 

will be used by the many individuals and representatives 

of countries and organizations and all those dedicated to 

sustainable rice production here today to share their thoughts 

and make further commitments toward enhancing sustainable 

rice-based production systems. 

On 16 December 2002, the United Nations General Assembly 

declared the year 2004 the International Year of Rice 

(IYR). The dedication of an International Year to a single crop 

is unprecedented in the history of the UN General Assembly. 

This reflects not only the role of rice in food security and poverty 

alleviation but also the significance of rice in almost every 

culture and indeed in our world’s cultural heritage. 

About 842 million people, including millions of children 

worldwide, are undernourished. Hunger impedes development 

through malnutrition and disease. Hunger and poverty 

lead to increased susceptibility to illness and reduced capacity 

for work and concentration. The end result is even deeper poverty 

and hunger. Food security is essential to sustainable livelihoods 

and to maintain a peaceful environment. In this regard, 

rice plays a vital role in fighting hunger and promoting 

stability. Rice is the staple food for over half of the world’s 

population. Almost a billion households in Asia, Africa, and 

the Americas depend on rice-based systems as their main source 

of employment and livelihood. Rice is the most rapidly growing 

food source in Africa. About four-fifths of the world’s rice 

is produced by small-scale farmers in developing countries, 

and is consumed locally. 

Rice systems are hubs for biodiversity, including aquatic 

and terrestrial organisms, wildlife, livestock, and other crop 

varieties. This biodiversity is crucial for the environment, and 

contributes to sustainable livelihoods through increased rural 

income and improved human health. Rice-based production 


systems have been designed to support biodiversity, enabling 

intensive rice cultivation systems that include fisheries, livestock, 

and other plant species. Diversified agriculture and a 

more diversified diet promote improved nutrition, help to improve 

sustainable livelihoods, and provide protection for agricultural 

genetic resources. 

In declaring the International Year of Rice, the UN General 

Assembly reaffirms the need to focus world attention on 

the role that rice can play in providing food security and eradicating 

poverty in the attainment of the internationally agreedupon 

development goals, including those contained in the 

Millennium Declaration and the World Food Summit. 

A major objective of FAO and the IYR Steering Committee 

is to promote sustainable rice production and increased 

access to this vital food crop. Through the International Rice 

Commission, FAO works closely with member governments, 

the international centers of the Consultative Group on International 

Agricultural Research (CGIAR), and national agricultural 

research and extension systems (such as the Japan International 

Research Center for Agricultural Sciences) to promote 

the sustainable development of rice-based production 

systems in order to alleviate poverty and hunger worldwide. 

In addition, FAO actively pursues partnerships with UN agencies, 

nongovernmental organizations, civil society, and the 

private sector to raise awareness, encourage information generation 

and exchange, and promote improved production and 

access to food. 

FAO recognizes, and appreciates, the role played by Japan 

in helping developing countries by promoting sustainable 

rice production during the Green Revolution in Asia and more 

recently in developing and disseminating NERICA rice in sub- 

Saharan Africa. The International Year of Rice and this World 

Rice Research Conference both recognize the potential for rice 

and rice-based production systems to improve human nutrition 

and food security around the globe. 

As the rice-consuming population continues to grow, and 

the land and water resources needed for rice production diminish, 

we face a potential crisis. World rice production has 

been less than rice consumption since 2000. This insufficiency 

has been addressed by drawing on rice from buffer stock. In 

this context, advances in science and technology, as well as 

rice research, are increasingly critical to enhance rice production 

and sustainable agricultural development. The successful 

mapping of the rice genome in 2002 has opened the way for 

potential scientific breakthroughs, but it has also created new 

issues related to biosafety, intellectual property rights, and 

access. 

Ensuring an increase in sustainable rice production will 

require innovation and cooperation within the scientific community, 

as well as commitment and shared responsibility among 

all stakeholders. The Global Plan of Action for the Conservation 

and Sustainable Use of Plant Genetic Resources for Food 

and Agriculture and the International Treaty on Plant Genetic 

Resources for Food and Agriculture provide a framework for 

global collaboration in the dissemination and use of rice genetic 

resources and protection of biodiversity. Capacity building 

within nations is urgently required to ensure that the innovations 

benefit local people and do not trigger long-term costs 

to the environment. 

Collaboration among institutions and stakeholders, from 

policymakers to farmers in the field, is imperative in order to 

promote scientific understanding and achieve effective, longterm 

improvements in rice production. This is the greatest challenge 

ahead. We must ensure that the fruits of our research, 

and the knowledge thereby gained, reach those hardest hit by 

hunger and poverty. 

In this regard, the World Rice Research Conference is 

very timely. It is an opportunity for those gathered here to share 

their experience and knowledge, formulate initiatives, and 

confirm their commitment toward achieving global food security. 

It will address the challenges and opportunities that face 

us today. 

I would like to thank the government of Japan once again 

for hosting this conference, as well as the many leaders, 

policymakers, scientists, researchers, and other partners for 

their participation and efforts in the fight against hunger and 

poverty. I wish you a very successful and fruitful conference. 

Notes 

Author’s address: Director-General, Food and Agriculture Organization 

of the United Nations. 

Address by Imperial Highness Crown Prince Naruhito 

I am very pleased that the World Rice Research Conference is 

being held with rice researchers and other participants from 

Japan and overseas. 

Rice is an excellent crop that is the staple food for more 

than half of the world’s population. Rice has been cultivated 

in the Japanese archipelago since ancient times and has provided 

the Japanese people with the basis for a stable life. Every 

Japanese holds dear the idyllic scene of placid rice paddies, 

which have worked as dams by storing water, preserved 

the soil, and nurtured a great variety of living things. I am told 

that both at home and abroad the Japanese-style diet based on 

rice has come to be highly regarded in recent years as a very 

healthy diet. It is no exaggeration to say that rice cultivation 

has been the basis for the development of Japanese society 

and culture. 

The United Nations designated this year as the International 

Year of Rice, and activities are under way throughout 

the world to raise awareness of the important role of rice in 

Keynotes 5

eliminating hunger and poverty. I believe that this conference 

is very significant as an opportunity for experts to discuss the 

production and use of rice and the many roles of rice paddies 

from the perspective of science and technology, and to consider 

the best avenues for future rice research. 

By bringing together the wisdom from all over the world 

and promoting active discussions from a wide range of perspectives, 

I hope that this conference will make a great contribution 

to solving the world’s food problems and protecting 

the global environment. 

New technologies for rice production 

Gurdev S. Khush 

Major advances have taken place in rice production during the 

last four decades because of the adoption of Green Revolution 

technology. Rice production increased 130% from 257 million 

tons in 1966 to 598 million tons in 1999. Average rice 

yield increased from 2.1 to 3.9 t ha –1 during the same period. 

In 2000, average per capita food availability was 18% higher 

than in 1966. Whereas rice production increased at 3.0% and 

2.5% per annum during the 1970s and ’80s, respectively, the 

increase was only 1.5% during the 1990s. During the last four 

years, production has been stagnant. According to the International 

Food Policy Research Institute, rice production must 

increase 38% by 2025 to feed 4 billion rice consumers. 

The area planted to rice is declining because of the pressure 

of urbanization and industrialization. Availability of water 

for agriculture is declining and labor is moving to industry. 

To meet the challenge of producing more rice under those constraints, 

we need new technologies. These include rice varieties 

with higher yield potential, greater yield stability, and 

adapted to changing global climate, as well as more efficient 

management practices. 

In addition to conventional hybridization and selection 

procedures, other strategies for developing rice varieties with 

higher yield potential include ideotype breeding, exploitation 

of heterosis, wide hybridization, molecular breeding, genetic 

engineering, and apomixis. Similarly, marker-aided selection 

and genetic engineering approaches are useful for developing 

rice varieties with more durable resistance to diseases and insects 

and tolerance of abiotic stresses such as drought, submergence, 

and salinity. 

Precision agriculture should lead to more sustainable increases 

in rice production, It is important to improve the efficiency 

of all production inputs such as nutrients and water and 

fine-tune the strategies for managing natural resources. Fertilizer 

is the most important input in crop production. Yet, areaspecific 

or regional fertilizer recommendations are based on 

short-term trials, which do not reflect the long-term impact of 

fertilizer use on the resource base. The chlorophyll meter is an 

excellent technique to measure the leaf N status of rice and to 

synchronize N application with crop demand. Water efficiency 

can be improved through its judicious use and zero tillage. 

Vast lands are unsuitable for crop production or have poor 

productivity because of soil problems such as salinity and alkalinity. 

Gypsum can be used to reclaim acidic and saline soils. 

Crop diversification is necessary to improve the productivity 

of intensively farmed soils with cereal crops. Crop rotations in 

addition to improving the soil structure show allelopathic effects 

through the release of chemical compounds directly in 

the soil or indirectly through microbial decomposition of residue, 

which affect the growth of another species. Integrated 

pest management (IPM) approaches have led to a reduction in 

the overuse and misuse of pesticides in rice production in a 

few countries. Universal adoption of IPM should lead to a reduction 

in the cost of rice production and more healthy rice. 

Notes 

Author’s address: Adjunct Professor, University of California, Davis. 


The changing economics and politics of rice: 

implications for food security, globalization, 

and environmental sustainability 

Joachim von Braun and María Soledad Bos 

For centuries, rice has been one of the world’s most important 

food crops. For the people of Asia in particular, rice has been 

the main source of calories and an important source of income 

and employment throughout most of their history. Furthermore, 

rice has shaped societies and cultures. Today, the cultivation, 

marketing, and consumption of rice are changing faster than 

ever before, yet there are also strong forces working to stabilize 

and conserve rice systems. The economics and politics of 

rice are changing because of globalization, technological transformation 

of agriculture driven by science, and changing diets 

and tastes due to prosperity and urbanization. At the same time, 

rice remains the lifeline of many poor producers and consumers. 

Deeply embedded in established food systems, rice remains 

local and traditional and appears to be partially excluded 

from globalization trends. Thus, rice is a most modern food 

and a very traditional one at the same time. 

As 2004—the International Year of Rice, as declared by 

the United Nations—comes to a close, it is pertinent to address 

key issues about the emerging changes in the economics 

and politics of rice and to look into the future of rice and the 

key role it plays in the livelihood of millions of people. 

Rice consumption plays a crucial role in assuring food 

security to people in Asia, and parts of Africa and 

Latin America. However, diets are diversifying and 

other crops are emerging as a result of growing urbanization 

and rising incomes. Moreover, demand for 

rice is shifting from lower-quality rice to higher-quality 

rice. What will be the patterns of demand for rice 

in the future 

Rice production and consumption remain closely 

linked. Only a small proportion of the rice produced 

is traded. Furthermore, production of different qualities 

of rice remains tied to the tastes and preferences 

of the region where it is produced. What are the rice 

production patterns across the world and how will 

they change 

Rice systems are highly sustainable, if well managed. 

However, it remains a permanent challenge to maintain 

the environmental sustainability of the diverse 

rice ecosystems. The environmental externalities related 

to rice production are complex. Moreover, rice 

is increasingly grown in irrigated ecosystems; thus, 

as water becomes increasingly scarce, priority should 

be given to research aimed at producing water-saving 

rice varieties. What is the future of rice technology 

 

 

Rice markets are generally characterized as thin, concentrated, 

and volatile. Because most rice is consumed 

where it is produced, only a small share of production 

is traded internationally. In the fastest globalizing 

region of the world, Asia, it is surprising that the 

most important crop, rice, is not globalizing—its integration 

into the international traded economy remains 

limited, lagging far behind other commodities. 

What are the reasons behind this phenomenon and 

can we expect this situation to change in the future 

Rice policies are partly shaped by culture, in particular 

as they affect consumer and producer behaviors, 

but for how much longer will that be the case as societal 

transformations progress The political economy 

of rice is changing, and that will shape rice production 

and consumption in the future. Even though the 

long history of state engagement in rice stockholding 

and trading is coming to an end, rice remains a strategic 

food security crop for policymakers and voters. 

What will be the future role of governments and the 

private sector in rice systems 

We approach these questions with deep respect for rice and 

what it means to the people and cultures in Asia and around 

the world, and are conscious that our brief review can only 

address selected aspects of the above-mentioned issues and 

themes. 1 

The culture of rice 

In many societies, when people consume or produce rice, they 

are not just consuming calories or producing grain, they are 

also engaging in practices that have an intrinsic cultural value. 

While prices and incomes are important drivers of demand, 

other factors such as traditions are also significant. For centuries, 

rice has shaped Asian societies and their cultures. Asian 

cultures are partly cultures of rice, and many Asian societies 

relate to rice beyond the satisfaction of basic needs. Rice is 

mentioned in the scriptures of the ancient civilizations of Asia. 

Its cultivation was considered as the basis of the social order 

and occupied a major place in Asian religions and customs 

1 We gratefully acknowledge comments on an earlier draft by Ron Cantrell, 

Susumu Matsuoka, Rajul Pandya-Lorch, and Keijiro Otsuka. 

Keynotes 7

(Hossain 1998). Hamilton (2003) describes the integral and 

vital role of rice as follows: “A key tenet of rice culture is that 

rice is a sacred food divinely given to humans that uniquely 

sustains the human body in a way that no other food can.” 2 

Rice is so embedded in Asian cultures that even the languages 

reflect the special nature of rice as the primary food for humans. 

In many Asian languages, there is no general word for 

“food” other than the word for rice and an invitation to “eat” 

implies the eating of rice. Moreover, growing rice used to be 

considered an ideal form of human labor and reflective of a 

well-ordered moral society (Hamilton 2003). Beyond the cultural 

influence of rice, its cultivation defines the landscape in 

many Asian countries. In China, Korea, and Japan, there is a 

well-established concept of an ideal landscape, consisting of 

rice fields spread out over a valley floor at the base of a mountain. 

Furthermore, in Japanese art, there is a strong tradition of 

depicting rice plants or rice agriculture as part of an idealized 

natural environment (Hamilton 2003). Japanese “rice culture” 

is based on a social infrastructure accumulated over centuries. 

Just half a century ago, Japanese agriculture and rural areas 

were not much different from those in present-day Vietnam or 

Thailand or southern China (being subsistent to a considerable 

degree), and, even today, many Japanese farm families 

eat their own rice, a practice unheard of in the U.S. or Europe. 

Changing demand for rice 

Rice is crucial for food security, especially in Asia, where more 

than 500 million people are undernourished and where rice 

provides on average more than 30% of total calories. The importance 

of rice is even starker in countries such as Cambodia, 

Vietnam, and Bangladesh, where rice represents more than 70% 

of the daily calorie intake (Fig. 1). 

Although the average share of total calories originating 

from rice has remained fairly constant during the last three 

decades for the world as a whole, in Asia this share has fallen 

from 38% to 31%, primarily due to diversification of diets 

caused by urbanization and increasing incomes. Within Asia, 

it is notable that the share of calories from rice has declined in 

China and Japan but this trend is not observed in other Asian 

countries; instead, the share of rice has remained fairly stable. 

In Africa, however, the share of rice in total food consumption 

is growing (Table 1). 

The world’s biggest consumer of rice is China, followed 

by India and Indonesia. While Asia has been and continues to 

be the main consumer of rice, consumption in other regions of 

the world is increasing. Outside Asia, Nigeria and Brazil are 

the most significant consumers; in fact, Nigeria is a leading 

rice importer. 

2 In some Asian cultures, the traditional understanding goes further: “Since 

humans live by eating rice, the human body and soul are regarded as being 

made of rice. Therefore, it is by eating rice that humans are defined” (Hamilton 

2003). 

US$ per ton 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

1970 1980 1990 2001 

Year 

Cambodia Indonesia Japan Brazil 

Bangladesh Philippines India Nigeria 

Vietnam China Côte d’Ivoire 

Fig. 1. Evolution of the share of calories originating from rice, selected 

countries, 1970-2001. Source: FAOSTAT (September 2004). 

Table 1. Share (%) of calories originating from rice, by region, 

1970-2001. 

Region 1970 1980 1990 2001 

Asia 38 36 35 31 

South America 11 11 12 11 

Africa 5 7 7 8 

United States 1 1 2 2 

European Union 1 1 1 2 

World 20 19 21 21 

Source: World Rice Statistics, International Rice Research Institute (September 

2004). 

The key factors that affect demand for rice are incomes, 

prices, population growth, and urbanization, but they do so in 

different ways. As incomes rise, consumers tend to diversify 

their diet, with the share of rice declining and with consumption 

of wheat initially increasing and of livestock and other 

products later increasing (Barker and Dawe 2002). Further, 

once calorie needs have been met, people express a preference 

for higher-quality rice as their incomes rise. In Japan, 

even though overall consumption of rice has declined in the 

past few decades, the type of rice demanded by consumers has 

changed from standard-quality rice to high-quality rice. 

Outside Asia, patterns of rice consumption vary considerably. 

In African countries, such as Côte d’Ivoire or Nigeria, 

rice consumption is increasing both in quantity as well as share 

of total food consumption, partly because of savings in time 


for preparation relative to roots and tubers and coarse grains. 

European consumers have traditionally chosen to consume 

round- to medium-grain japonica rice, but in recent years their 

consumption of long-grain indica rice has increased. Furthermore, 

their demand for aromatic rice varieties has increased 

significantly since the early 1990s. European consumers are 

increasingly interested in special rice varieties such as organic 

rice, waxy rice, jasmine rice, wild rice, and colored pericarp 

(Ferrero and Nguyen 2004). In the United States, rice consumption 

has more than doubled in the last 20 years, mainly 

because of increased concerns about diet and health but also, 

as in Europe, because of increases in Asian and Hispanic populations 

who prefer rice. Rice imported into the United States 

accounts for 11% of direct food consumption, and is mostly 

aromatic Thai jasmine and Indian and Pakistani basmati consumed 

by ethnic Asians (Maclean et al 2002). 

There is tremendous variation in tastes and preferences 

for rice across the world. Hence, the market for rice is segmented 

by type and quality, with little substitution in consumption. 

Geographically, the production of different qualities of 

rice follows the tastes and preferences of the region where it is 

consumed. Among traded rice varieties, indica rice accounts 

for the bulk of global rice trade (75–80%), followed by japonica 

(10–12%) and aromatic rice varieties such as basmati and jasmine 

(10%), with glutinous rice accounting for the rest (Gulati 

and Narayanan 2002). The variation in tastes and preferences 

partly depends on historical and socio-cultural factors. Topquality 

rice in one region may be considered low-quality in 

another (Kaosa-ard and Juliano 1991). For example, in East 

Asia, the preferred grain is short and round “sticky rice” that 

can be eaten easily with chopsticks. In Southeast Asia, most 

consumers prefer medium- to long-grain rice with little aroma. 

In South Asia, consumers prefer parboiled long-grain rice with 

a strong aroma, with basmati being the most popular rice of 

this type. Production patterns tend to follow consumption patterns. 

Among the available rice varieties, traditional varieties 

are generally considered to be of higher quality and thus get 

premium prices in the market. Thailand, for example, still extensively 

grows low-yielding traditional varieties, because the 

export market is mostly for high-quality rice (Pingali et al 

1997). In many traditional rice markets such as India, Pakistan, 

and Thailand, fragrant rice—considered a national treasure—fetches 

the highest prices and is therefore the most profitable 

type of rice to cultivate under rainfed conditions in these 

countries. 

Rice and poverty 

Rice consumption and production are closely connected to 

poverty. Rice is the main staple food in Asia and the most important 

source of employment and income for rural people 

(Hossain 1998). Because of the large size of the rice economy 

and the importance of rice in the Asian diet, productivity gains 

in rice have a greater impact on poverty reduction than gains 

in other agricultural commodities. The impact on poverty 

mainly operates through lower prices for consumers, driven 

by increased production, which reduces spending on food by 

the poor, including the urban poor, rural landless people, and 

nonrice farmers. Rice farmers also benefit from higher rice 

productivity, which helps them to diversify into high-value 

agriculture. In a later section, we will describe how rice research 

in India and China has increased rice production, which 

in turn has decreased the number of rural poor. 

Rice consumption typically varies among different income 

groups. In Bangladesh, expenditure on rice alone accounts 

for more than half of the weekly per capita expenditure 

for all income groups. For example, in the lowest income 

quintile, rice represents 65% of total food expenditure and in 

the highest income quintile this share drops to 50%, which is 

still high. Moreover, studies indicate that calorie intake in 

Bangladesh increases by almost 40% from the lowest to the 

highest income group, with increases in rice consumption accounting 

for two-thirds of this increase (Bouis and Novenario- 

Reese 1997). In low-income countries such as Bangladesh, 

rice consumption is not only the main source of energy but 

also of essential micronutrients such as iron. Many poor people 

consume poor-quality diets that provide insufficient micronutrients. 

Fortification of rice is one option for addressing this 

situation. The introduction of biofortified crops, that is, varieties 

bred for increased mineral and vitamin content, complements 

existing nutrition interventions and provides a sustainable 

and low-cost way of reaching people who have poor access 

to formal markets and health care. The high levels of per 

capita consumption of rice in the developing countries of Asia 

mean that increasing the nutrient value of rice can have significant 

positive health outcomes. For example, food consumption 

studies suggest that doubling the iron content of rice could 

increase the iron intake of the poor by 50% (HarvestPlus 2004) 

and could thus have a positive effect on health. 

A study of the two main rice-producing and -consuming 

countries, China and India, showed that improvements in rice 

yields significantly helped to reduce rural poverty (Fan et al. 

2003, Fan and Chan-Kang 2005). Investments in rice research 

have increased rice production in both these countries over 

the past four decades. From 1961 to 2001, rice production 

grew at an average rate of 2.7% per year in India and 2.6% per 

year in China. Fan et al (2003) estimated that rice research 

reduced poverty in India by about 2% per annum and in China 

by about 4.5% per annum in the 1990s (Table 2). Rural poverty 

is reduced through three main channels. First, the release 

of new and better rice varieties enables farmers to produce 

more rice at the same cost, which in turn improves their income 

(Kerr and Kolavalli 1999). Second, the diffusion of 

modern rice varieties can result in lower food prices, which is 

crucial to most poor people, who spend a large proportion of 

their income on food (Datt and Ravallion 1997). Third, the 

productivity consequences of improved varieties results in 

greater demand for labor (David and Otsuka 1994, Hossain 

1988). 

The growing economic prosperity in Asia is reducing 

the incentives of farmers to engage in rice cultivation. The 

Keynotes 9

Table 2. Poverty impact of rice research in India and China. 

India 

China 

Year Number of poor Reduction as % Number of poor Reduction as % 

reduced by rice of total poor reduced by rice of total poor 

research(million) 

research(million) 

1991 4.95 2.1 5.20 5.5 

1995 4.81 1.9 2.85 4.4 

1999 3.06 1.9 1.53 4.5 

Source: Fan et al (2003). 

Table 3. Total rice production, percentage of world production, and percentage increase of production by region, 1970- 

2002. 

Item World Asia South America North and Africa Europe 

Central America 

Production (1,000 mt, paddy), 576,280 523,030 19,846 11,880 17,034 3,194 

2002 

Percentage of world production, 100 91 3 2 3 1 

2002 

Percentage increase in production, 82 80 90 126 133 79 

1970-2002 

Source: World Rice Statistics, International Rice Research Institute (September 2004). 

expansion of the nonfarm sector and rising labor productivity 

have pushed up nonfarm wage rates, which motivates migration 

of labor—particularly of the young generation and children—from 

rural areas to cities and from farm to nonfarm activities 

within rural areas. Since traditional rice farming is a 

highly labor-intensive activity, increases in wages have pushed 

up the cost of rice production and reduced the incomes and 

profits of farmers (Hossain and Narciso 2004). In response, 

mechanization of harvesting and threshing has occurred in 

Malaysia, Thailand, and, in recent years, China. Direct seeding 

has also replaced transplanting to further save on the labor 

costs of crop establishment. These labor-saving techniques help 

to restore profitability in the face of rising wages. 

Rice production and technology 

Developing countries account for 95% of global rice production. 

Asia alone produces 90% of world rice, with China and 

India accounting for over half of the world’s output. Worldwide, 

rice production has increased by more than 80% in the 

last three decades (Table 3). The most striking increase is observed 

in Africa, where rice production increased by 133%, 

albeit from very low levels, reflecting the growing importance 

of rice in this region. Other nontraditional rice-producing areas, 

including South, North, and Central America, have increased 

rice production at a higher rate than traditional rice 

producers. 

World rice production increased steadily during the 

1990s and started to decline only in the past four years (Fig. 

2). The decline in rice prices may be a main factor behind 

these recent developments in world rice production. Stocks of 

rice followed patterns similar to those of production, increasing 

during the 1990s and declining more recently. Dawe (2002) 

offers three reasons for the plunge in world prices: the significant 

reduction in Indonesia’s imports after 1998, the devaluation 

of the Thai baht during the Asian financial crisis, and the 

increase in rice production in Bangladesh, Vietnam, Pakistan, 

and India. 

Productivity and sustainability issues 

around rice ecosystems 

Rice ecosystems vary widely across countries and regions. In 

Europe, the United States, and Australia, rice is cultivated solely 

under irrigation. In Africa, on the other hand, the majority of 

rice is cultivated upland, with less than 20% of rice cultivated 

under irrigation. A similar situation is found in Latin America. 

In Asia, more than 50% of rice is irrigated, with the majority 

of the remaining 50% being cultivated in rainfed lowlands (Fig. 

3). 

Successful agricultural development requires the diversification 

of agriculture away from staple crops, for which 

demand gradually declines. In the case of Asia, however, diversification 

away from rice production is difficult because 

the surface irrigation systems have been designed to provide 

an adequate water supply for rice but not for other crops. 

Moreover, because of the monsoon climate, other crops are 

not easily adopted. 

The transformation to irrigated rice production has varied 

considerably in Asian countries during the past three decades. 

Countries such as Pakistan and Japan have always cul- 


US$ per ton 

280 

260 

240 

220 

200 

180 

160 

140 

120 

100 

1983 

White broken rice, Thai A1 

Super, f.o.b. Bangkok 

Production (milled) 

1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 

Year 

000 tons 

420,000 

400,000 

380,000 

360,000 

340,000 

320,000 

300,000 

Fig. 2. Milled rice production and world price trends. Source: USDA PS&D Database (2004) and 

FAO Commodities and Trade Division (2004). 

% 

% 

100 

Deepwater 

Upland 

Rainfed lowland 

Irrigated 

100 

Bangladesh 

China 

India 

Japan 

Pakistan 

Philippines 

Thailand 

80 

80 

60 

60 

40 

20 

0 

Asia 

Latin America 

Africa 

Australia 

Region 

Fig. 3. Distribution of rice ecosystems, by regions and selected 

countries, 2001. Source: World Rice Statistics, IRRI (September 

2004). 

USA 

Europe 

World 

40 

20 

0 

1970 1980 1990 2001 

Year 

Fig. 4. Evolution of the share of rice in the irrigated 

ecosystem, selected countries, 1970- 

2001. Source: World Rice Statistics, IRRI (September 

2004). 

tivated rice under the irrigated ecosystem and continue to do 

so, largely because of the limited rainfall in these countries. 

Another group of countries that includes the Philippines, India, 

and Bangladesh has experienced a steady increase in the 

use of irrigation for rice cultivation in the last three decades. 

Most surprising is the case of Thailand, however, where the 

proportion of rice cultivated under irrigation declined from 

the 1970s to the 1990s to a low of less than 10% before turning 

around, although it has still not returned to the level of the 

1970s (Fig. 4). 

A move toward irrigated rice cultivation is desirable not 

only because it increases rice yields but also because it decreases 

the vulnerability of rice cultivation to weather condi- 

tions (David and Otsuka 1994). In addition, the main technological 

breakthroughs in rice cultivation, which occurred as 

part of the Green Revolution, were in the cultivation of irrigated 

rice; as a consequence, the yield of rice cultivated under 

this ecosystem is now much higher than yields that can be 

achieved under other conditions. 

Concerns regarding the adverse environmental effects 

of irrigation and flood control projects on waterlogging, salinity, 

fish production, and the quality of groundwater have been 

growing (Hossain and Narciso 2004). Water is becoming increasingly 

scarce and there is growing competition for it between 

agricultural and nonagricultural uses. The water-use efficiency 

of irrigated rice is low. For example, rice production 

Keynotes 11

equires about twice as much water as other crops such as maize 

and wheat. In Asia, irrigated rice consumes 150 billion m 3 of 

water, which corresponds to a water-use efficiency of approximately 

20,000 m 3 ha –1 . Assuming an average yield of 5 t ha –1 , 

the water productivity of irrigated rice is only 0.15 kg of milled 

rice per m 3 of water. This level of productivity is low compared 

with that of other crops and makes rice noncompetitive 

compared to other uses of water in the face of growing water 

scarcity (Sohl 2002). 

To improve the productivity of water use, it is important 

to identify sources of water loss and minimize them. For example, 

reducing seepage and percolation will increase water 

efficiency on farms. 3 Given the growing constraints to water 

use for rice, the future of rice production relies on the development 

and adoption of water-saving technologies, a research 

area currently being pursued at the International Rice Research 

Institute (IRRI). 4 

Beyond the negative externalities associated with inefficient 

use of water in rice cultivation, population increases are 

putting increasing pressure on the land to be more productive. 

In marginal areas, intensification of land use may lead to degradation 

of resources through loss of biodiversity, deforestation, 

pest infestations, depletion of natural soil fertility, and 

soil erosion. Furthermore, the use of pesticides and fertilizers 

to increase the output of a given piece of land will likely result 

in degradation, environmental pollution, and adverse effects 

on human health (Cantrell and Hettel 2004). The lands that 

are most threatened in Asia are the fragile rainfed and upland 

environments where the poor are forced to use whatever resources 

are available to produce the food they need. 

Further, rice ecologies interact with regional and global 

ecosystems. A case in point is the interaction among climate 

change, land use, and production systems. Some lowland rice 

systems in coastal zones and river deltas are potentially threatened 

by rising sea levels caused by global warming. At the 

same time, irrigated rice contributes to greenhouse gases. These 

complex interactions, as well as the appropriate responses in 

terms of ecological, technological, and policy measures, need 

further study. 

3 If this water is recovered for crop consumption at some point downstream, 

then there is no loss in the irrigation system and it does not affect water efficiency 

at the basin level (Guerra et al 1998). 

4 Guerra et al (1998) suggest four strategies to increase farm irrigation productivity: 

(1) increasing production per unit of evapotranspiration, which involves 

developing new varieties with higher yields and better fertilizer and 

weed management; (2) reducing water use in land preparation; (3) adopting a 

water-efficient method of rice establishment, mainly by moving from transplanting 

to direct seeding to reduce irrigation inflow during land preparation; 

and (4) reducing surface runoff, that is, reducing seepage and percolation 

during the crop growth period. The authors also suggest some strategies at the 

irrigation system level; in brief, they include changing the crop and irrigation 

schedule to use rainfall more effectively, improving water distribution strategies 

and rehabilitation, and modernizing irrigation infrastructure, among others. 

Table 4. Rice yields in different countries and ecosystems. 

Country Ecosystem Year Rice yield 

(t ha -1 ) 

Bangladesh All 2000 3.6 

Burkina Faso Rainfed 1987-90 2.5 

India Irrigated 1995-96 5.2 

Rainfed 1995-97 2.3 

United States Irrigated 2001 7.0 

Japan Irrigated 1999 6.4 

Philippines All 1999-2000 3.1 

Thailand Irrigated 2000 4.2 

Rainfed 2000 2.2 

South Korea Irrigated 1999 6.6 

Vietnam Irrigated 2000 4.2 

Source: Hossain and Narciso (2004). 

Yield patterns and changes 

Rice yields vary enormously across ecosystems and countries. 

Yields of 4–6 t ha –1 are common in irrigated settings, as are 

yields of 2–3 t ha –1 in rainfed ecologies (Table 4). The potential 

to increase yields in rainfed ecosystems is still high. This 

ecosystem is dominant in the low-income countries of South 

Asia, Southeast Asia, West Africa, and Central America. Rice 

farming in rainfed ecosystems is subject to natural hazards such 

as droughts, floods, and typhoons. Where rainfall is unreliable 

and drainage is poor, farmers still grow traditional varieties 

and use fertilizers in suboptimal amounts because of the uncertainty 

of obtaining adequate returns from investment in inputs. 

This is one of the main reasons for the low yield and 

large yield gap in countries with predominantly rainfed ecosystems 

(Hossain and Narciso 2004). 

The scope for further conversion of rainfed ecosystems 

into irrigated ecosystems is becoming increasingly limited. In 

addition, the cost of introducing irrigation to new areas has 

increased substantially, partly because areas that are amenable 

to irrigation tend already to have irrigation systems in place, 

and partly because of drastic declines in investment in the development 

and maintenance of large-scale irrigation projects 

in many Asian countries (Hossain and Narciso 2002). Moreover, 

energy costs for water lifting are rising. 

Scientific breakthroughs have enabled less-favored lands 

to improve their productivity. For instance, research at IRRI 

has produced rice varieties that are more drought-tolerant, and 

the Chinese agricultural research system has produced hybrid 

rice strains with higher yields. Another good example is the 

development of the “New Rice for Africa” (NERICA) variety, 

a cross of African and Asian rice cultivars, by the West Africa 

Rice Development Association (WARDA). NERICA promises 

more tolerance of most African stresses, including weeds and 

drought, shorter growing cycles, and higher yield potential. 

This technological development provides an opportunity for 

farmers to stabilize and intensify low-input upland farming 

systems. Faster adoption of these varieties may lead to a substantial 

increase in production in Nigeria, Guinea, Côte 

d’Ivoire, Sierra Leone, Liberia, and Uganda, where upland and 


Table 5. Growing yield differences of rice within countries 

in the 1980s and 1990s. 

Average annual growth rates (%) in national average cereal 

yield and subnational variation. 

Country Average growth Standard deviation 

per annum (%) (subnational variation) 

Bangladesh 1.8 5.1 

Brazil 4.6 2.3 

China 2.1 1.7 

India 2.3 2.4 

Source: Schreinemachers (2004). 

rainfed lowland ecosystems predominate (Hossain and Narciso 

2004). 

Recent analyses suggest that growth rates of yields are 

increasingly differing between and within countries. In three 

out of four major rice-producing countries, the growth in average 

yields over the past 20 years is higher than the standard 

deviation of yield growth by provinces in the respective countries 

(Table 5). Rising inequality in yield growth within countries 

is not to be interpreted as a “growing yield gap,” which is 

understood as the difference between actual and potential yield. 

In fact, rising inequality in growth of rice yields within a country 

(e.g., by region or as uplands versus lowlands) would be 

expected where rice is cultivated more in line with the respective 

ecologies, and where scarcities of production factors determine 

the costs and input uses more strongly, as is the case in 

countries whose governments limit their interference in input 

and output markets. This rising inequality can also be found 

where earlier low-yield regions are catching up because of 

improved technology. 

Closing the actual yield gap will not only increase production 

per se but will also improve the efficiency of land and 

labor use, reduce production costs, and, most importantly, 

improve food security. 

In Asian countries, rice is mostly grown on small family 

farms. The average size of a rice farm is typically less than 

half a hectare in China, Indonesia (Java), and the Red River 

Delta in Vietnam; less than 1 ha in Bangladesh, eastern India, 

and the Mekong River Delta in Vietnam; and 1 to 2 ha in most 

other countries in Asia. Only in Thailand, Myanmar, Cambodia, 

and the Punjab in India is the average farm size over 2 ha 

(Hossain and Narciso 2004). Hayami and Kikuchi (2000) have 

presented an excellent analysis of the dynamics and perspectives 

of small-farm rice production in Asia, taking a small village 

in the Philippines as an illustrative example. The “saga” 

of this rice village demonstrates that improving rice yield is a 

necessary but not sufficient condition to increase rice production. 

Irrigation, appropriate policies, and investment in infrastructure 

are also needed in this process. 

Political economy of rice, markets, and trade 

Rice has played a key role in the historical conception of the 

state in many Asian countries. Today, most Asian governments 

still view rice as a strategic commodity because of its importance 

to the diet of the poor and to the employment and income 

of farmers. Given this importance, fluctuations in rice 

prices are considered a threat to political stability, and this 

may be one reason why governments tend to intervene in their 

country’s rice market (Hossain and Narciso 2004). Historically, 

governments in the main rice-producing and -consuming 

countries have favored policies that maintained stable prices 

for consumers in urban centers and provided subsidies to farmers. 

Dorosh and Shahabuddin (2002) explain that, in 

Bangladesh, policy instruments are designed to minimize both 

annual price fluctuations and seasonal price variations. Before 

1994, public imports and, to a lesser extent, drawdown of 

stocks were the main policy instruments used to achieve 

interannual price stability. After trade liberalization in 1994, 

however, the private-sector import trade has been the dominant 

factor in successfully keeping prices within acceptable 

boundaries. In addition, during this period, the main instruments 

used to achieve intra-annual price stability were domestic 

procurement and open market sales. 

In Indonesia, a central strategy for providing affordable 

rice to the poor has been a targeted rice subsidy, termed “Rice 

for the Poor” (Raskin). A ration of 20 kg per month per family 

is provided as part of the social safety net program initiated 

after the crisis in 1997 (Sidik 2004). In addition, the Indonesian 

government provides incentives to diversify consumption. 

Probably the most notable trend in rice-related policies 

over the past years has been the deregulation of postharvest 

activities and efforts to shift some functions traditionally carried 

out by the public sector, such as stock holding, to the private 

sector (FAO 2003a). Governments are less dominant today 

in international rice trade than they were in the late 1970s, 

although the magnitude of the change is difficult to quantify 

precisely. Slayton (1999) calculated that, from 1995 to 1999, 

government-to-government contracts, including food aid, averaged 

less than 7% of world rice trade, whereas, in 1980, 

total government-to-government contracts were at least 19% 

of world trade. 

Indonesia is an example of a country where government 

intervention in the rice market and international trade is declining. 

BULOG, initially a parastatal organization whose main 

function was to stabilize the prices of basic foods in line with 

the government’s food policy, has shaped Indonesia’s rice 

policy and market for years. In the late 1990s, the Indonesian 

government started releasing the right to import to other private 

importers and BULOG lost its monopoly. For the past 

several years, the private sector has been responsible for about 

75% of Indonesia’s imports (Dawe 2004). In early 2003, 

BULOG became a state-owned enterprise (SOE) with dual 

Keynotes 13

400 

350 

300 

250 

200 

US$ t –1 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 

150 

100 

50 

White broken rice, Thai A1 

Super, f.o.b. Bangkok 

White rice, Thai 100% B 

second grade, f.o.b. Bangkok 

0 

1983 

Year 

Fig. 5. Evolution of Thailand’s (main rice exporter) rice price, 1983-2004. Source: FAO Commodities 

and Trade Division (September 2004), World Rice Statistics, IRRI (2004). 

functions, to conduct public-service activities and to conduct 

business as a commercial organization. Under the former function, 

BULOG still distributes “Rice for the Poor” (Raskin) and 

maintains national stock reserves to counter food shortages 

during emergencies. As a commercial organization, BULOG 

can enter into trading activities as a profit-making organization. 

The role of governments in the international rice trade 

has also diminished, as has the influence of large trading companies. 

Today, numerous smaller trading companies are engaged 

in the international trade of rice, which has increased 

competition and eroded trading margins. However, the degree 

and type of policy changes that have occurred in recent years 

vary greatly from country to country. We consider two examples: 

Vietnam and India. 

Vietnam’s rice market is characterized by a high level of 

commercialization and a large role of the private sector in the 

production and marketing stages. The private sector plays a 

key role in procuring rice from farmers, supplying rice to exporters, 

and distributing rice within the country. At the same 

time, however, the private sector is highly underdeveloped and 

regionally imbalanced (Golleti and Minot 1997). Until recently, 

all rice exports from Vietnam were conducted by SOEs, but 

over the last few years private firms have gradually been allowed 

to export rice. Although SOEs buy from farmers, the 

importance of these transactions is not sufficient for SOEs to 

stabilize producer prices, which is considered one of the main 

functions of these organizations. Thus, although SOEs still play 

an important role in Vietnam’s international trade, gradual liberalization 

is occurring (Tolentino and Bruce 2003, Dawe 

2004). 

In India, the government still intervenes in trade transactions. 

The Food Corporation of India is one of the largest 

organizations procuring food grains on behalf of the government 

of India. One of the main roles of this organization is to 

buy and transport rice from surplus areas to deficit areas, under 

the Public Distribution System, to ensure food for the poor 

and to stabilize prices (Nigam 2004). Nevertheless, although 

the private sector negotiates and handles the export of rice, 

the government still makes key decisions that determine the 

quantity of rice exported (Dawe 2004). 

Developments in and determinants 

of the price of rice 

The world price for rice has shown a declining trend over the 

past 50 years. After adjusting for inflation, world rice prices 

today are 77% lower than the average from 1950 to 1981. The 

main reason for the decline in prices is the Green Revolution, 

which led to an increase in yields and a decrease in unit production 

costs, and consequently to an increase in worldwide 

supply (Dawe 2004). World prices for other grains have also 

declined in real terms, but to a lesser extent than has occurred 

for rice. Rice prices continue to fluctuate substantially, as can 

be seen in the plot of rice prices for Thailand over the period 

1983-2004 (Fig. 5). 

The price of rice fluctuates significantly and is sensitive 

to variations in supply and demand. The main cause for this is 

the small amount of rice traded internationally; less than 7% 

of the world’s total rice production is traded. Furthermore, six 

countries (Thailand, Vietnam, China, the United States, Pakistan, 

and India) supply 85% of the rice that is traded. This 

concentration in world rice markets implies that changes in 

production or consumption in major rice-trading countries have 

a strong effect on world prices. In recent years, there has been 

a diversification of world importers and exporters that could 

further stabilize rice prices. Before the Green Revolution, Asia 

(excluding the Middle East) accounted for 64% of world rice 

imports, and an identical share of world exports. Today, a more 

diversified group of countries accounts for rice imports. 


120 

110 

100 

90 

80 

70 

60 

1998 1999 2000 2001 2002 2003 

Year 

Indica high 

Indica low 

Japonica 

Aromatic 

Fig. 6. Evolution of rice price indices by quantity (1998-2000 = 

100). Source: FAO (2004, 2003b). 

Asian countries have been successful at stabilizing domestic 

prices relative to those in the world market. This has 

been achieved both under centralized (Malaysia and Indonesia) 

and less-centralized (Thailand) political systems. Rice price 

stabilization remains central to protecting poor consumers in 

many Asian countries. An interesting case to follow is that of 

China, which recently entered the World Trade Organization 

and committed to suppress rice price stabilization mechanisms. 

This policy shift gave rice producers an incentive to diversify 

production to other more profitable crops. Rice production in 

China has declined over the last three years, and, if this trend 

continues, China will become one of the world’s main rice 

importers. 

Another interesting feature of world rice prices is the 

evolution of prices of different qualities of rice. As mentioned 

earlier, rising income levels in many countries have been accompanied 

by a shift away from low-quality rice (defined as 

less than or equal to 20% broken) during the past 25–30 years 

(Dawe 2004). Dawe (2004) further observes that in recent years 

(1999-2003 versus 1986-95) prices of different qualities of 

indica varieties moved closer to each other (this is observed in 

Figure 6, where the price indices for low- and high-quality 

indica move almost in synchrony). There is still a lot of substitution 

between the qualities, as would be expected for a food 

item that is predominantly consumed by people with low income. 

However, relative price differences between different 

types of rice in 2002 and 2003 appear higher than five years 

earlier. 

International trade of rice and trade policies 

Global trade in rice expanded on average by 7% a year over 

the 1990s. Nevertheless, the international rice market remains 

thin, accounting for only 7% of global output compared with 

19% for wheat and 13% for maize. Developing countries are 

still the main players in the world rice trade, with a share of 

83% of world exports and 85% of world imports. 

Despite a trend toward governments limiting their interventions 

in the rice market, the long tradition of government 

intervention in rice markets (Siamwalla and Haykin 1983) continues 

across the world. Both developing and developed countries 

protect this sector, although the policies differ between 

these two groups of countries. Developed countries typically 

use a combination of domestic market interventions and border 

protection or export subsidies depending on whether they 

are net importers or exporters. Border protection through bans, 

high tariffs, and state trading enterprises (STEs) is very common 

in wealthy rice-importing countries and regions such as 

Japan, South Korea, and the European Union. Japan and South 

Korea, for example, control imports through monopoly STEs, 

with Japan’s Food Agency being one of the world’s largest 

importer STEs. Domestically, developed countries support rice 

producers with a combination of market price interventions 

and direct payments. Developed-country exporters also use 

export subsidies and export credit guarantees to promote their 

rice exports (Gulati and Narayanan 2002). 

Poorer developing countries in South and Southeast Asia 

have tended to tax rice producers by imposing support measures 

that keep the rice price at levels below international prices. 

Public procurement and government stocking are undertaken 

in almost all of these countries for public distribution of food 

grain (India) or for market interventions (Vietnam, Thailand, 

China, and Indonesia). Stocking limits on private traders and 

restrictions on the movement of rice are also common practices. 

STEs continue to play a dominant role in many of these 

countries, particularly those that are net importers (Gulati and 

Narayanan 2002). 

Given that developing countries protect their producers 

mostly by border support and other price-based policy measures, 

a measure of the protection of producers in these countries 

is given by the gap between domestic and international 

prices (Mullen et al 2004a,b). Mullen et al (2004a,b) estimated 

the market price support (MPS) and producer support estimates 

(PSE) 5 for rice in India, Indonesia, and Vietnam (Table 

6). 

The general pattern that emerges from these estimates is 

that current levels of support of rice by trade and market policies 

in Asian developing countries are considerably higher than 

the levels in the 1980s or early 1990s. India’s trade policies in 

the last decade can be followed by examining the PSEs. Before 

1994, India had in place a ban on exports of certain types 

of rice. In that year, the ban was lifted and India’s rice exports 

increased from less than 1 million tons per year to about 5 

5 MPS is calculated comparing the domestic farm-gate price and the adjusted 

reference price. The reference price at the border is the “world market” c.i.f. 

price for an importer and f.o.b. price for an exporter, adjusted by the costs of 

handling, transporting, and marketing and difference in quality. For more details 

on the calculations and selection of reference price, see Mullen et al 

(2004a,b). The PSE is calculated by adding the budgetary transfers to the 

MPS. 

Keynotes 15

Table 6. Producer support estimates for India, Indonesia, and Vietnam. 

Country 1990 1991 1992 2000 2001 2002 

India –10.3 –2.7 –25.1 28.3 11.2 40.1 

Indonesia 2.2 1.0 10.9 21.4 19.3 45.6 

Vietnam –28.4 –15.8 –10.7 53.0 60.4 32.4 

Sources: Mullen et al (2004a), Thomas and Orden (2004). 

million tons in 1995-96, making India the second-largest exporter 

of rice in that year. Since then, India has become a major 

supplier of common and basmati rice. Vietnam’s PSEs for 

rice show that rice farmers have tended to be protected, especially 

at the end of the 1980s and after 1996, although from 

1990 to the mid-1990s they were taxed. This trend was further 

accentuated by a dramatic decline in public stockholding and 

fertilizer prices. In general, as indicated by the PSEs, the trend 

has been toward increasing support for rice (Nguyen and Grote 

2004). 

We now return to one of the initial questions posed above: 

Why is the integration of rice into international trade so limited 

Of critical importance to this issue is inelastic supply 

and demand, that is, both supply and demand do not respond 

quickly and strongly to price changes. An additional factor is 

the vulnerability of production to occasional disruptions due 

to climatic phenomena. These characteristics lead to large price 

fluctuations. Market interventions are then used to stabilize 

prices and policies are applied to transfer income from consumers 

to producers. 

Future scenarios for rice: toward 2050 

In view of the close linkages among production, markets, and 

consumption of rice, any outlook on rice must take a holistic 

perspective, and the future of rice cannot be assessed without 

due consideration of the broader economic context and developments 

related to other foods. Global and regional rice consumption 

and production will be determined by changes in 

consumption behavior and policy actions related to technology 

(research) and trade. An analysis of the effects of these 

variables on rice consumption, production, and market developments 

can be useful for identifying appropriate policies that 

serve growth, sustainability, and poverty reduction. Below, we 

present three policy scenarios using the International Model 

for Policy Analysis of Agricultural Commodities and Trade 

(IMPACT) 6 developed by Mark Rosegrant of the International 

6 IMPACT covers 36 countries and accounts for almost all of the world’s commodities. 

The IMPACT model explores the potential implications of policies 

in key variables such as prices, demand, yields, production, and net trade. 

This model assumes competitive agricultural markets for crops and livestock 

(Rosegrant et al 2002). 

Food Policy Research Institute (IFPRI). The key policy assumptions 

for each scenario are briefly described below. 7 

Progressive policy scenario 

This scenario assumes a new focus on agricultural growth and 

rural development. Developing countries’ public investments 

and government expenditures on agriculture and rural development, 

appropriately supported by official development assistance 

(ODA), increase between 2005 and 2015 and stabilize 

thereafter. Investments in education, social services, and 

health increase. The rate of agricultural technology improvement 

is high owing to increased investment in agricultural research 

and development. Irrigation efficiency and water-use 

efficiency improve in this scenario, and the rate of irrigation 

expansion is moderate to high. Furthermore, producer support 

to farmers in wealthy countries declines substantially, dropping 

to half of current levels by 2010, and to half of the 2010 

level by 2020. 

Policy failure scenario 

This scenario assumes trade and political conflicts, with no 

progress on global agricultural trade negotiations and increased 

levels of protectionism worldwide. Today, we cannot exclude 

political-economic forces from producing such outcomes. The 

policy failure scenario assumes decreases in yield growth for 

all crops and fish, and decreases in numbers growth for all 

livestock. It also shows trade policies that lead to an increase 

in protection in many countries, contributing to stagnating 

world trade and slow growth in the net imports of developing 

countries. This impasse in agricultural trade liberalization further 

contributes to the growing food deficit in developing countries. 

As a result of political conflicts, investments in social 

services and agricultural R&D are low. Producer support to 

farmers in wealthy countries triples from current levels by 2020 

and remains steady through 2050. 

Technology and natural resource management 

failure scenario 

This scenario is characterized by water mismanagement, declining 

irrigation efficiency, lack of adaptation to climate 

change, and pest problems in agriculture. Low agricultural in- 

7 For further details, see von Braun et al (2004); for a detailed model description, 

see Rosegrant et al (2002). 


000 t 

kg ha –1 

590,000 

550,000 

510,000 

Progressive policy actions 

Policy failure 

Technology and resource 

management failure 

5,000 

4,000 





470,000 

430,000 

390,000 

3,000 

350,000 

1997 2015 2030 2050 

Year 

Fig. 7. World rice demand projections (1,000 mt, milled rice). 

Source: IFPRI IMPACT projections (September 2004). 

2,000 

1997 2015 2030 2050 

Year 

Fig. 8. World rice yield projections under different policy scenarios. 


vestments undermine the development of new agricultural technology 

and contribute to marginal levels of irrigation efficiency 

and lack of improvement in water-use efficiency. In addition, 

investments in many sectors, including education, social services, 

and health, are low in developing countries. The lack of 

growth in agricultural yields is the outcome of all of the above, 

and also partly a result of weak income growth in developing 

countries, and only moderate income growth in industrialized 

countries. 

Under the progressive policy actions scenario, world 

demand for rice is projected to increase by more than 50% 

from 1997 to 2050 (Fig. 7). The projected change in demand 

varies by region; for example, under this scenario, rice demand 

is predicted to rise by almost 200% in sub-Saharan Africa 

and by 110% in Latin America. However, if natural resources, 

especially water, are mismanaged, and investment in 

agriculture is low, as in the technology and natural resource 

management failure scenario, demand for rice is projected to 

remain flat. 

Appropriate policies can boost rice production under sustainable 

use of the natural resources on which rice production 

largely depends. Future trends in rice yields will be greatly 

affected by the degree of investment in research, resource 

management, and market and trade policy. In the progressive 

policy actions scenario, in which developing countries increase 

public investment in agriculture, rural development, and agricultural 

research and development, rice yields are projected to 

almost double from 1997 to 2050 (Fig. 8). However, if agricultural 

investment is low and natural resources are mismanaged, 

contributing among other things to marginal levels of 

irrigation efficiency, as in the technology and natural resource 

management failure scenario, yield growth is predicted to stagnate 

over the next 50 years. 

Projections under the progressive policy actions scenario 

suggest that, with increased public investment and government 

expenditure, rice production will increase in the two poorest 

regions of the world, South Asia and sub-Saharan Africa (Figs. 

000 t 

200,000 

180,000 

160,000 

140,000 

120,000 

100,000 





1997 2015 2030 2050 

Year 

Fig. 9. South Asia rice production projections under different policy 

scenarios. Source: IFPRI IMPACT projections (September 2004). 

9 and 10). On the other hand, if trade and political conflict rise 

and the level of protectionism increases, as in the policy failure 

scenario, production will not increase sharply. The worst 

outcome in terms of rice production stagnation arises when 

there is technology and natural resource mismanagement. The 

patterns of change differ by world region, as suggested by differences 

in the predicted trends for Asia and sub-Saharan Africa. 

Projections under the progressive policy actions scenario, 

in which producer support to farmers in developed countries 

declines substantially, suggest that the international price for 

rice will continue to decline in the future. By contrast, under 

the policy failure scenario, which assumes strong trade distortions, 

the price of rice will remain at 1997 levels (Fig. 11). 

In sum, the long-term future of rice is going to be very 

much influenced by policies related to technologies, water 

management, and trade. In addition, the incomes and tastes of 

consumers will remain fundamental drivers of change in the 

economics of rice. 

Keynotes 17

000 t 

US$ t –1 

40,000 

30,000 

20,000 





600 

500 

300 

200 





10,000 

100 

0 

1997 2015 2030 2050 

Year 

0 

1997 2015 2030 2050 

Year 

Fig. 10. Sub-Saharan Africa rice production projections under different 

policy scenarios. Source: IFPRI IMPACT projections (September 

2004). 

Fig. 11. Projections of rice price under different policy scenarios. 


Conclusions 

The cultivation and consumption of rice have shaped the lives 

of millions of people for centuries. Today, rice is facing new 

risks and new opportunities, but, for millions of people, every 

year is the Year of Rice, not just the year 2004, as declared by 

the United Nations. 

More than 500 million undernourished people live in 

Asia, accounting for 63% of the world’s undernourished people. 

The importance of rice in the Asian diet is expected to decline 

over time because of increases in migration to urban areas and 

increases in incomes to above the level at which people begin 

to diversify their diets. This pattern stands in contrast to the 

patterns observed in other parts of the world such as Africa 

and Latin America, where rice is becoming increasingly important. 

However, regardless of these mixed trends, rice will 

remain central to the diets of millions of people across the 

world. 

Rice is a “health food” with many characteristics and 

diverse flavors, and can be developed further to accentuate its 

desirable characteristics. For example, scientific research into 

the diet quality aspects of rice, including the contents of micronutrients 

such as iron and vitamin A, has the potential to 

transform rice into an even more important agent of change, 

especially for the poor. At the same time, high-income consumers 

are expected to increase their demand for specialty rice 

varieties. 

Under a scenario that assumed increased investment in 

rural development, improved technology, efficient use of natural 

resources, and increased trade liberalization—referred to 

as the progressive policy actions scenario—the global demand 

for rice was predicted to increase by more than 50% over the 

next 50 years. Most significant in relative terms is the prediction 

that the demand for rice in sub-Saharan Africa and Latin 

America will increase by almost 200% and 110%, respectively, 

over this period. 

Over the last three decades, global rice production has 

increased by more than 80%. North and Central America as 

well as sub-Saharan Africa increased their production by more 

than 120% during this time. This growth can be largely explained 

by the increase in yields accomplished with the Green 

Revolution. Feeding the ever-increasing world population poses 

a challenge that must be tackled by increasing yields. If this is 

to be achieved, increased investment in research and development 

is needed. Increases in rice yields have strongly contributed 

to reducing poverty in Asia, and such increases will continue 

to be much needed by the millions of people who live in 

poverty across the world. 

Efforts to increase rice production must not overlook 

the externalities associated with ecological effects. Water is 

an increasingly scarce resource, and current rice cultivation 

methods use water inefficiently; hence, new methods of cultivation 

and new technologies will be needed in the future to 

achieve sustainable and efficient use of this resource. Beyond 

the adverse externalities associated with the inefficient use of 

water in rice cultivation, population increases put pressure on 

the land to be more productive. The development of appropriate 

improved technologies represents an opportunity to enhance 

sustainable rice production. Scientific research aimed at improving 

the water-use efficiency of rice and the resistance of 

rice to pests and tolerance of drought has the potential to facilitate 

environmental sustainability, thereby increasing yields 

further. These objectives will need to remain high on the agenda 

of rice research for decades to come. 

Countries with long-standing experience in enhancing 

rice productivity and sustainability, such as Japan, may accept 

a certain “corporate” responsibility to share their know-how 

in research and management with low-income countries. Such 

knowledge sharing, which already occurs to a considerable 

extent, is probably as important as sharing opportunity through 

access to markets. 

Among its multiple functions, rice is a strategic commodity 

because it is the single most important element in the 

diet of the poor and one of the main sources of income and 

employment for farmers in Asia. Traditionally, governments 

have sought to maintain stable prices for consumers in urban 


areas and to provide input subsidies to farmers. Nowadays, a 

shift toward increased private intervention in the rice market 

is noted, and governments are increasingly delegating functions 

to private enterprises. Hence, appropriate policies need 

to be developed so that these new players can provide affordable 

food to the poor and do not exclude poor farmers from 

the transforming food and agricultural system. 

As a worldwide trend toward freer trade is observed, 

developing countries are likely to become increasingly important 

players in the world trade of rice. Developments in policy 

at the national, regional, and international levels are a decisive 

component of the state of rice in the world today and will also 

shape the future role of rice. 

Changes in policies may be driven by societal and technological 

changes. Rice culture has modernized but certainly 

not vanished. The roles of rice for improving food security, 

reducing poverty, and sustaining the environment will remain 

strong. Rice cultures show strong resilience against drastic 

changes. This aspect of rice may actually be a healthy force in 

a rapidly changing and globalizing world, where rice has been 

as much an agent of developmental change as it has been a 

conserving factor serving sustainability. 

References 

Barker R, Dawe D. 2002. The transformation of the Asian rice 

economy and directions for future research: the need for increased 

productivity. In: Sombilla M, Hossain M, Hardy B, 

editors. Developments in the Asian rice economy. Los Baños 

(Philippines): International Rice Research Institute. p 1-30. 

Bouis HE, Novenario-Reese MJ. 1997. The determinants of demand 

for micronutrients: an analysis of rural households in 

Bangladesh. Food Consumption and Nutrition Division Discussion 

Paper No. 32. Washington, D.C. (USA): International 

Food Policy Research Institute. 

Cantrell RP, Hettel GP. 2004. New challenges and technological 

opportunities for rice-based production systems for food security 

and poverty alleviation in Asia and the Pacific. Paper 

presented at the FAO Rice Conference, Rome, Italy, 12-13 

February 2004. 

David CC, Otsuka K. 1994. Modern rice technology and income 

distribution in Asia. Boulder, Colo. (USA): Lynn Rienner. 

Datt G, Ravallion M. 1997. Why some Indian states performed better 

than others at reducing rural poverty. Food, Consumption, 

and Nutrition Division Discussion Paper No. 26. Washington, 

D.C. (USA): International Food Policy Research Institute. 

Dawe D. 2001. How far the path of free trade The importance of 

rice price stabilization in developing Asia. Food Policy 

26(2):163-175. 

Dawe D. 2004. Changing structure, conduct and performance of the 

world rice market. Rome (Italy): FAO Rice Conference. 

Dorosh P, Shahabuddin Q. 2002. Rice price stabilization in 

Bangladesh: an analysis of policy options. Markets and Structural 

Studies Division Discussion Paper No. 46. Washington, 


Fan S, Chan-Kang C. 2005. National and international agricultural 

research and rural poverty: the case of rice in India and China. 

Agricultural Economics (Forthcoming) 

Fan S, Chan-Kang C, Qian K, Krishnaiah K. 2003. National and 

international agricultural research and rural poverty: the case 

of rice research in India and China. Environmental and Production 

Technology Division Discussion Paper No. 109. Washington, 


Fan S, Thorat S, Rao N. 2004. Investment, subsidies, and pro-poor 

growth in rural India. Development and Governance Studies 

Division Discussion Paper (Forthcoming). Washington, DC: 

IFPRI. 

FAO (Food and Agriculture Organization of the United Nations). 

2001. Review of basic food policies. Rome (Italy): FAO. 


2003a. Review of basic food policies. Rome (Italy): FAO. 


2003b. Rice market monitor. Rome (Italy): FAO. 


2004. Rice market monitor. Rome (Italy): FAO. 

Ferrero A, Nguyen NV. 2004. Constraints and opportunities for sustainable 

development of rice-based production systems in 

Europe. Paper presented at the FAO Rice Conference, Rome, 

Italy, 12-13 February 2004. 

Golleti F, Minot N. 1997. Rice markets, agricultural growth and policy 

options in Vietnam Market and Structural Studies Division 

Discussion Paper No. 14. Washington, D.C. (USA): International 


Guerra LC, Bhuiyan SI, Tuong TP, Barker R. 1998. Producing more 

rice with less water from irrigated systems. Discussion Paper 

Series No. 29. Los Baños (Philippines): International Rice 

Research Institute. 

Gulati A, Narayanan S. 2002. Rice trade liberalization and poverty. 

Markets and Structural Studies Division Discussion Paper No. 

51. Washington, D.C. (USA): International Food Policy Research 

Institute. 

Hamilton RW. 2003. The art of rice: spirit and sustenance in Asia. 

Los Angeles, Calif. (USA): UCLA Fowler Museum of Cultural 

History. 552 p. 

Harvest Plus. 2004. Breeding crops for better nutrition. Brief. Washington, 


Hayami Y, Kikuchi M. 2000. A rice village saga: three decades of 

green revolution in the Philippines. Lanham, Md. (USA): 

Barnes & Noble. 

Hossain M. 1988. Nature and impact of the green revolution in 

Bangladesh. Research Report No. 67. Washington, D.C. 

(USA): International Food Policy Research Institute. 

Hossain M. 1998. Sustaining food security in Asia: economic, social, 

and political aspects. In: Dowling NG, Greenfield SM, 

Fischer KS, editors. Sustainability of rice in the global food 

system. Davis, Calif. (USA) and Manila (Philippines): Pacific 

Basin Study Center and International Rice Research Institute. 

Hossain M, Narciso J. 2002. New rice technologies and challenges 

for food security in Asia and the Pacific. In: Sustainable rice 

production for food security. Proceedings of the 20th Session 

of the International Rice Commission. Bangkok, Thailand, 

23-26 July 2002. 

Keynotes 19

Hossain M, Narciso J. 2004. Long-term prospects for the global rice 

economy. Rome (Italy): FAO Rice Conference. 

Huang J, Bouis HE. 1996. Structural changes in the demand for food 

in Asia. 2020 Vision Discussion Paper No. 11. Washington, 


IRRI. World rice statistics, various years. Los Baños (Philippines): 


Kaosa-ard M, Juliano BO. 1991. Assessing rice quality characteristics 

and prices in selected international markets: rice grain 

marketing and quality issues. Selected papers from the International 

Rice Research Conference, 27-31 August 1990, Seoul, 

Korea. 

Kerr J, Kolavalli S. 1999. Impact of agricultural research on poverty 

alleviation: conceptual framework with illustrations from the 

literature. Environmental and Production Technology Division 

Discussion Paper No. 56. Washington, D.C. (USA): International 


Maclean JL, Dawe D, Hardy B, Hettel GP, editors. 2002. Rice almanac. 

Third edition. Los Baños (Phillipines), Bouaké (Côte 

d’Ivoire), Cali (Colombia), and Rome (Italy): IRRI, WARDA, 

CIAT, and FAO. 

Mullen K, Gulati A, Orden D. 2004a. Agricultural policies in India: 

producer support estimates 1985-2002. Markets, Trade, and 

Institutions Division Discussion Paper. Washington, D.C. 

(USA): International Food Policy Research Institute. (Forthcoming.) 

Mullen K, Sun D, Thomas M, Orden D, Gulati A. 2004b. Agricultural 

policy interventions in developing countries: mapping 

the nature, degree and progress in reforms. Selected paper 

prepared for presentation at the American Agricultural Economics 

Association Annual Meeting, Denver, Colorado, 1-4 

August 2004. 

Nguyen NV, Grote U. 2004. Agricultural policies in Vietnam: producer 

support estimates 1986-2002. Discussion Paper. Washington, 

D.C., (USA) and Bonn (Germany): IFPRI and Center 

for Development Research (ZEF). (Forthcoming.) 

Nigam S. 2004. Rice: challenges in production and marketing in 

India. Paper presented at the FAO Rice Conference, 12-13 

February 2004. 

Pingali PL, Hossain M, Gerpacio RV. 1997. Asian rice bowls: the 

returning crisis New York and Manila: CAB International 

and International Rice Research Institute. 

Rosegrant M, Meijer S, Cline S. 2002. International Model for Policy 

Analysis of Agricultural Commodities and Trade (IMPACT): 

model description. Washington, D.C. (USA): International 


Schreinemachers P. 2004. Yield gaps and their determinants. Bonn 

(Germany): Center for Development Research. (In mimeo.) 

Siamwalla A, Haykin S. 1983. The world rice market: structure, conduct, 

and performance. IFPRI Research Report No. 39. Washington, 


Sidik M. 2004. Indonesia rice policy in view of trade liberalization. 

Paper presented at the FAO Rice Conference, 12-13 February 

2004. 

Slayton T. 1999. The outlook for the rice trade in the new millennium. 

World Rice Commerce Conference. 

Sohl M. 2002. Issues and challenges in rice technological development 

for sustainable food security. Keynote Address at the 

20th Session of the International Rice Commission, Bangkok, 

Thailand, 23-26 July 2002. Rome (Italy): FAO. 

Thomas M, Orden D. 2004. Agricultural policies in Indonesia: producer 

support estimates 1985-2005. Markets, Trade, and Institutions 

Division Discussion Paper. Washington, D.C. 

(USA): International Food Policy Research Institute. (Forthcoming.) 

Tolentino V, Bruce J. 2003. Successful transition to the market 

economy: overview of the experience in Vietnam. Yangon 

(Myanmar): FAO Seminar on rice sector policy: neighboring 

countries’ experiences. 

USDA. 2004. Producer, supply and distribution online database. 

Washington, D.C.: United States Department of Agriculture. 

www.fas.usda.gov/psd/. 

von Braun J, Rosegrant MW, Cohen MJ, Cline S, Brown, Bos MS, 

Pandya-Lorch R. 2004. New risks and opportunities for food 

security. 2020 Discussion Paper. Washington, D.C. (USA): 

International Food Policy Research Institute. 

Notes 

Authors’ address: International Food Policy Research Institute 

(IFPRI), e-mail: j.vonbraun@cgiar.org. 


Feeding the world: How much more rice do we need 

Vaclav Smil 

The key question that underlies efforts to increase rice yields, 

to improve the crop’s nutritional quality, and to cut its 

postharvest losses is how much more of the grain will we need 

to assure global adequacy of its supply I will look just a generation 

(25 years) ahead, but, even then, a precise answer is 

impossible as well as unnecessary: what we need is a good 

approximation of where we are headed. Key determinants of 

future rice consumption are the rates of population growth, 

the process of population aging, the increased mechanization 

of labor, the need to improve inadequate diets, and the extent 

and pace of dietary transitions. The complex interplay of these 

factors will determine potential demand, the total that should 

give us a better appreciation of how high we should aim our 

efforts at future yield increases and agronomic improvement. 

Population growth 

The global population growth rate has been slowing ever since 

it peaked at just over 2.0% in 1967 (Fig. 1). As a result, consecutive 

long-range population forecasts issued by the United 

Nations have been repeatedly revised downward and the low 

variant in the forecast for 2025 is just 7.3 billion people, not 

even 20% above the 6.1 billion in 2000 (UN 2002). The medium 

variant is 7.8 billion and the high 8.3 billion, but the 

recent accelerated decline in fertility indicates that the most 

likely outcome will be closer to the low total and that virtually 

all of the net increase will take place in Africa, Asia, and Latin 

America. 

The medium forecast of population growth in the world’s 

ten most populous countries where rice is the staple grain and 

that consume about 85% of the world crop—Bangladesh, Brazil, 

China, Egypt, India, Indonesia, Myanmar, the Philippines, 

Thailand, and Vietnam—is for about a 26% increase from 2000 

to 2025, an addition of just over 800 million people. Everything 

else being equal, demand for rice would rise by about 

25% in 25 years compared with the 68% increase from 1975 

to 2000. But everything else will not be equal. Rapidly aging 

populations will actually require less food per capita because 

of the natural decline in metabolic needs (unless we assume 

universal obesity), further mechanization of labor will eliminate 

most of the hard physical tasks that are still common in 

many countries, and rising incomes will improve average diets 

and change their composition. 

Population aging and activity 

The process of population aging will be pronounced not only 

in Europe and Japan but also in China (thanks to its one-child 

policy) and in many countries in Asia (Fig. 1). In the world’s 

ten most populous rice-eating countries, the share of the popu- 

Declining population growth rates, annual growth rate (%) 

2.5 

2.0 

1.5 

1.0 

0.5 

0 

100 

80 

60 

40 

20 

World 

Asia 

Declining annual absolute growth rates, population change 

per year (miilions) 

Population aged 60+ (%) 

16 

14 

12 

10 

8 

World 

Asia 

World 

6 

1950 1975 2000 2025 

Year 

Fig. 1. Declining population growth rates and population aging: 

global and Asian trends, 1950-2025. 

Asia 

Keynotes 21

lation older than 60 years will increase by roughly 15% to 

about two-thirds, or nearly 600 million people, by 2025. Complexities 

of human metabolism preclude any accurate calculation 

of this aging effect (Smil 2000), but, assuming that, on 

average, people above age 60 consume 15–20% less food energy 

per day per capita than the mean for the adult population, 

then aging would reduce 2025 demand by no more than 2–3% 

compared to a population with a stationary age structure. 

The continuing mechanization of agricultural and industrial 

tasks in modernizing countries may have a similar effect 

on overall food demand. Again, complexities of human metabolism 

make it impossible to make any accurate estimates, 

but, assuming an average difference of at least 10–15% between 

moderately and highly active population groups and a 

shift of 20–25% of today’s labor force from more demanding 

tasks in agriculture and industry to mechanized manufacturing 

and services would lower overall food demand and hence presumably 

also rice consumption by 2–3% compared to the unchanged 

frequency of typical labor exertions. The combined 

effect of aging and mechanization could decrease the overall 

demand for rice by no more than 5% by 2025. 

Eliminating undernutrition 

In spite of major gains during the last quarter of the 20th century, 

the FAO (Food and Agriculture Organization of the United 

Nations) estimated the total number of malnourished people 

at about 800 million during the late 1990s (FAO 2002). Nearly 

60% of this total, or almost 480 million people, were in the 

world’s ten most populous rice-eating countries. India had 

about 233 million undernourished people (almost 29% of the 

world total) and China 119 million (about 15%). Undernourished 

people thus added up to about 15% of the population in 

countries where rice is the staple cereal. 

When assuming that, to eliminate these nutritional deficiencies, 

this disadvantaged segment of the population would 

have to increase its average food intake by 25%, the overall 

food supply in 2025 would have to rise by 3.75% above the 

level needed to provide larger populations merely with today’s 

inadequate diets. An excellent confirmation of this additional 

amount comes from the FAO’s estimates of the relative inadequacy 

of the food supply that was calculated for all of the 

world’s countries for the early 1990s and that ranged for the 

rice-eating countries from roughly 1% in Brazil and Egypt to 

nearly 9% in Bangladesh (FAO 1996). The populationweighted 

mean of this relative inadequacy was almost exactly 

4% for the ten countries and improvements during the 1990s 

would have reduced it by about 10%. 

Dietary transitions 

Predicting the course of dietary transitions is difficult because 

they can be greatly influenced by unexpected changes in a 

nation’s economic course: post-Mao China is the best example 

of such unpredictable shifts. As gross domestic product (GDP), 

and with it average income, rises beyond subsistence levels, it 

Annual rice consumption (kg capita –1 ) 

200 

150 

100 

50 

Myanmar 

Bangladesh 

Vietnam 

Indonesia 

Philippines 

Thailand 

India 

China 

0 

1,000 10,000 100,000 

PPP GDP capita –1 (2001 US$) 

may be accompanied by a temporary increase in demand for 

staple grain, followed by a prolonged decrease in consumption 

as more varied diets and higher shares of lipids (both plant 

oils and animal fats), meat, eggs, and dairy and aquatic products 

displace staple grains (Caballero and Popkin 2002). No 

other rice-eating country experienced such a precipitous decline 

in average intake as did post-World War II Japan: as annual 

per capita consumption dropped to less than 60 kg, rice 

ceased to be a staple food and became more of a luxury foodstuff 

(Chern et al 2003). The South Korean drop was also precipitous, 

from nearly 125 kg per capita in 1975 to about 85 kg 

in 2000, or more than 30% in 25 years. 

Many country-specific scenarios of future dietary trends 

are thus possible, but there is no doubt about the general decline 

in average per capita rice consumption with rising economic 

performance. Every tripling of purchasing power parity-adjusted 

per capita GDP appears to be accompanied by a 

50-kg decline in average annual per capita rice consumption 

(Fig. 2). Consequently, even when assuming a sluggish growth 

of average per capita GDPs in the world’s ten most populous 

rice-eating countries—on the order of 1.5% per year, or less 

than half of the rate they achieved during the 1990s—a nearly 

50% increase in that average would correspond to an annual 

per capita reduction of about 20 kg of milled rice. By 2025, 

this would translate into overall demand about 20% below the 

level that would result from unchanged dietary patterns. 

Overall demand for rice 

Brazil 

South Korea 

Japan 

Fig. 2. Log-normal relationship between per capita GDP (PPP) and 

annual per capita rice consumption at the beginning of the 21st 

century. 

This is not a forecast, merely an exploration of key 

countervailing trends that point us in the most likely direction 

and that should be useful in guiding our decisions regarding 

future agronomic efforts. By 2025, population growth in the 

world’s ten most populous rice-eating countries whose con- 


sumption dominates the global demand for rice will increase 

demand by about 25%; lower intakes because of aging and the 

physically less active population would reduce it by about 5%, 

and improvements in average diets needed to eliminate the 

existing undernutrition would boost it by about the same 

amount. The single largest factor determining the eventual 

outcome will be the extent and rapidity of dietary change and 

even very conservative assumptions point to a decline on the 

order of 20%. 

This would leave us with the need to produce only about 

5% more rice than we do now, or, bracketing the estimate by a 

factor of two, the range would be from no net increase at all to 

as much as a 10% higher output. Postharvest losses in the ten 

most populous rice-eating countries are on the order of 15% 

(Smil 2000) and cutting them in half during the next generation 

would only reinforce the conclusion that there may be no 

need for any net increase in rice production, or that the needed 

increase could be only a marginal addition of a few percent 

above the current level. 

Even in the unlikely case that these estimates err on the 

low side by as much as 100%, it is obvious that, during the 

first quarter of the 21st century, we will need increases in global 

rice production that will not be even remotely comparable 

to those of the past 25 years. Consequently, our research and 

development should concentrate primarily on the maintenance 

of existing yields, on improved nutritional quality, and on lowering 

the environmental effects of rice cultivation, particularly 

on reducing large (commonly in excess of 60%) losses of nitrogen 

from applied urea (Cassman et al 2002). 

References 

Caballero B, Popkin BM, editors. 2002. The nutrition transition: 

diet and disease in the developing world. Amsterdam (Netherlands): 

Elsevier. 261 p. 

Cassman KG et al. 2002. Agroecosystems, nitrogen-use efficiency, 

and nitrogen management. Ambio 31:132-140. 

Chern Wen S et al. 2003. Analysis of the food consumption of Japanese 

households. Rome (Italy): Food and Agriculture Organization 

of the United Nations. 88 p. 


1996. The sixth world food survey. Rome (Italy): FAO. 153 

p. 


2002. The state of food insecurity in the world 2002. Rome 

(Italy): FAO. 36 p. 

Smil V. 2000. Feeding the world. Cambridge, Mass. (USA): The 

MIT Press. 360 p. 

UN (United Nations). 2002. World population prospects: the 2002 

revision. New York (USA): UN. http://esa.un.org/unpp. 

Notes 

Author’s address: Faculty of Environment, University of Manitoba, 

Winnipeg R3T 2N2, Canada, e-mail: vsmil@cc.umanitoba.ca. 

Development of sustainable agriculture 

from rice, water, and the living environment 

Riota Nakamura 

The Green Revolution based on the development 

of irrigated agriculture 

Wisdom developed from soil and water has continuously been 

the fundamental element of all human activities since ancient 

civilization. The tragedies and ruins of the Mesopotamian/ 

Sumerian civilization remind us of the significance of building 

sustainable agriculture and society. The drama of the collapse 

of Mesopotamia started with a gradual rising of the saline 

groundwater level. The collapse advanced rapidly when 

this level exceeded a threshold. It is well known that this problem 

was caused by faulty irrigation. Can we be confident that 

the current threshold is high enough when we look at the global 

expansion of commercialized agriculture and hazardously 

exploited water resources, such as in the Aral Sea basin, the 

Ogallala aquifer (Nebraska) and Central Valley (California) in 

the United States, Punjab and Haryana in India, and Northern 

China 

A stable water supply was a strong driving force behind 

the Green Revolution. Irrigated agriculture provides an essen- 

tial environment for high yield, so that improved bred varieties 

of crops can be fully used. From 1961 to 2002, global irrigated 

agricultural land roughly doubled from 139 million to 

277 million ha, while total land for arable and permanent crops 

expanded slightly from 1,357 million to 1,534 million ha (Fig. 

1). Global population and cereal production have also doubled 

from 3.08 billion to 6.23 billion and from 877 million metric 

tons to 2.03 billion metric tons. Irrigated land, which accounts 

for about 18% of agricultural land area, produces about 40% 

of the food for the global population, contributing considerably 

to the alleviation of global poverty and starvation. 

Sound and sustainable irrigated agriculture is indispensable 

for humankind to survive in the future. Now, about 70%, 

or 2,504 km 3 , of the world’s annual freshwater usage of 3,572 

km 3 is for agriculture, and, of this, about 70% is used mainly 

for rice paddy agriculture in Asia. 

Our generation is primarily responsible for assuring sustainable 

and efficient agriculture through wiser governance and 

management of soil and water resources. 

Keynotes 23

Index (1961 = 100) 

250 

240 Total cereal production (1961 = 100) 

Population (1961 = 100 

230 

Irrigation area (1961 = 100) 

220 Land for arable and permanent crops (1961 = 100) 

210 Rainfed area (1961 = 100) 

200 

190 

180 

170 

160 

150 

140 

130 

120 

110 

90 

1961 

1963 

1965 

1967 

1969 

1971 

1973 

1975 

1977 

1979 

1981 

1983 

1985 

1987 

1989 

1991 

1993 

1995 

1997 

1999 

2001 

2002 

Year 

Fig. 1. World cereal production, population, and farmland area (1961-2002). 

World water issues and conversion of policies 

for irrigation—from expansion to increasing efficiency 

of existing systems 

Fresh water existing on the global land surface in readily usable 

forms such as lakes, swamps, and rivers accounts for only 

about 0.0075% (104,620 km 3 ) of all the water existing on our 

planet. This percentage is equivalent to the ratio of two teaspoonfuls 

of water (15 cm 3 ) in a typical household bathtub 

full of water. Of the annual global rainfall on land, which supports 

the hydrologic cycle of fresh water, only about 40% 

(45,000 km 3 ) becomes potential water resources after excluding 

evaporation. Humankind has to share this available water 

with other water uses such as industry, domestic use, and 

biodiversity while producing food for the more than 6 billion 

people living on the planet. 

In the 20th century, the century of fire and machines, we 

strived to develop water resources mostly through construction 

technologies such as large reservoirs, until the early 1980s. 

Not only drastic growth in the world human population but 

also worldwide trends of economic growth and expansion of 

cities, especially in developing countries, resulted in sharp increases 

in demand for domestic and industrial water use, and 

put strong and continuous pressure on the increase in water 

supplies. Water resource development by construction of “hardware” 

was promoted in many regions around the world. 

However, after the 1980s, this type of development met 

with certain limitations. Although the most effective way to 

increase water resources, speaking from an engineering viewpoint, 

was to build large reservoirs, appropriate construction 

sites for new reservoirs became limited. Moreover, governments 

of developing countries faced financial pressure for the 

operation and maintenance of overaged water facilities. 

In the 21st century, the century of water and life, another 

option has been recognized as a better solution. That is to increase 

the efficiency of the use of water in existing systems. In 

many countries, attempts to increase this efficiency have been 

made through (1) renovating irrigation water facilities such as 

lining canals with concrete, (2) introducing/reinforcing volumetric 

water pricing, and (3) introducing participatory irrigation 

management (PIM). 

Efficiency of agricultural water use in different regional 

conditions—arid and humid 

Improving water-use efficiency in agriculture is a key issue 

during international water discussions nowadays. Many experts 

have reported case studies and mentioned success stories. 

These reports are helpful for improving water-use efficiency 

in regions where it is reasonable for farmers to constantly 

use a minimum amount of irrigation water to secure 

good crop growth. The concept of water-use efficiency in this 

definition is typically applicable to agriculture in arid/semiarid 

regions. This concept comes from the idea that all water 

should be consumed in crop fields in the form of evapotranspiration, 

allowing no water to be lost elsewhere. 

However, in humid regions blessed with abundant precipitation, 

the shortcomings of this concept have now come to 

be widely recognized. These shortcomings are mainly caused 

by neglect of the following factors accompanying rice paddy 

farming in these regions: (1) the highly substitutable characteristic 

of water usage and labor investment, (2) the dynami- 


Socioeconomic externalities generated by irrigation 

and rice paddy agriculture in the Asian monsoon region 

Automatically provided by agricultural activities 

Multiple use of water by farmers and residents 

Aquaculture, duck raising, washing, cleaning, bathing, 

cooling, gardening, fire fighting, etc. 

Nonuse-value in cultural-religious activities 

Multifaceted socioeconomic benefits to the public 

Protect aqua-ecosystem, enhance water-related environment, 

form landscape, recharge groundwater aquifer, stabilize 

downstream river flow by return flow, etc. 

Intentionally provided by special consideration and actions 

l 

l 

l 

l 

Provide water from agriculture for domestic use during severe 

dry spells 

Increase performance of paddy fields while protecting reservoirs 

during extreme floods 

Create winter sanctuaries for migratory birds 

Restore groundwater level for downstream city, etc. 

Fig. 2. Externalities provided by activities of irrigation and rice paddy agriculture. 

cally fluctuating competitiveness among water users over the 

short term, and (3) the enormous value of socioeconomic elements 

other than food/fiber production; this value is defined 

as an “externality.” Factor 1 allows for reduced labor investment 

costs where farmers use much more than the minimum 

amount of water required for meeting crop water requirements. 

Factor 1 also allows for a drastic reduction in water use through 

farmers’ collective efforts during severe dry spells that occur 

unexpectedly under the conditions of factor 2. 

A considerable part of the water irrigated into rice paddy 

fields is not consumed but drained into the basin downstream 

during the rainy season from vast paddy fields in the Asian 

monsoon region. This is because (1) the value of water is quite 

low because of abundant precipitation and low competitiveness 

among water users, and (2) immersion cultivation with a 

larger amount of water than the equivalent to evapotranspiration 

reduces labor investment cost. 

This behavior seemingly decreases water-use efficiency, 

but, on the contrary, the affluent water use provides diverse 

and enormous value for socioeconomic externalities. The 

drained return flow also helps to preserve nature, resulting in 

any negative externality counting for nothing. Note that, in 

rice paddy agriculture in humid regions, the irrigation water 

automatically serves as an externality, whereas, in upland fields 

and relatively arid regions, the water has to be set aside and 

preserved, especially for externalities. 

Socioeconomic externalities in rice paddy agriculture 

in humid regions 

The enormous external value generated by irrigated rice paddy 

agriculture can be grouped in the following categories: (1) 

multiple use of water by farmers and residents for aquaculture, 

duck raising, washing, cleaning, bathing, cooling, gardening, 

and fire fighting; (2) multifaceted socioeconomic benefits 

to the public such as protecting the aqua-ecosystem, enhancing 

the water-related environment, forming landscape, 

recharging the groundwater aquifer, and stabilizing downstream 

river flow by return flow; and (3) various other nonuse-values 

in cultural-religious activities (Fig. 2). 

The socioeconomic situation of countries in the Asian 

monsoon region is diversified, including tropical developing 

and temperate developed countries. However, predominant rice 

paddy agriculture and its enormous external value are common 

to them. Note that the external values are not only important 

for farmers and the economy in developing countries but 

are also very valuable for citizens of developed countries in 

providing multifaceted socioeconomic benefits. 

Water productivity, often advocated by the catch phrase 

“more crop per drop,” is a newly conceptualized term as an 

indicator of water-use efficiency in rainfed agriculture, using 

about 16,000 km 3 of rainwater annually, as well as in irrigated 

agriculture. This concept, however, unjustly underestimates the 

water productivity of rice paddy agriculture in the Asian monsoon 

region. Therefore, it must be reevaluated after incorporating 

the high value of socioeconomic externalities. 

Traditional participatory irrigation management 

and governance (PIM/G) in rice paddy systems 

Recent challenges to establish PIM in Asian monsoon countries 

show difficulties in transferring the operation and management 

(O&M) of irrigation from government agencies to 

farmers. Lessons learned from these challenges include how 

to establish effective incentives for farmers. Initially, strong 

economic incentives superior to disincentives such as labor 

Keynotes 25

contributions of O&M and water charge are necessary to launch 

modernized operations. Afterward, the diverse values of socioeconomic 

externalities for farmers become driving forces 

to improve their performance. The diverse values are realized 

step by step through various group activities. Well-sustained 

traditional PIM shows many examples of built-in incentives 

related to externalities as well as agricultural production. 

When we look around the world, we find many current 

irrigation systems that have been fulfilling both requisites of 

food production and externalities, such as the Muang Fai in 

Thailand, Kanna in Sri Lanka, and Subak in Indonesia. The 

Dujiangyan irrigation system in Sichuan, China, was established 

in 250 B.C. and is still working for about 670,000 ha of 

farmland. In Japan, many old irrigation systems have lasted 

for centuries, with technology advancing in stages. It has to be 

noted that these systems always consist of weirs, canals, and 

ponds operated and managed by indigenous farmers’ organizations, 

namely, Land Improvement Districts (LIDs). Most of 

them are historically as old as their weirs and canals. The LIDs 

are responsible for O&M of nationwide canals extending a 

total length of about 400,000 km (equivalent to ten times around 

the globe). 

We found it very interesting that irrigation systems that 

achieve a good balance between externalities and food production 

are observed mostly in old systems. In presenting this 

paper at the WRRC, I wanted to mention representative canal 

systems that have completed the difficult task of satisfying both 

requisites. 

Renaissance of efficient and sustainable rice production 

with socioeconomic externalities generated by PIM/G 

Group activities through PIM/G can generate much more value 

for externalities than individually disconnected activities. A 

good example is shown by a full cultural and religious performance 

in Bali, Indonesia. Such a group activity works as a 

platform for consultation in a local community and sometimes 

develops new activities such as mutual farming aid, group 

composting, and joint ventures for purchasing and shipping. 

In Japan, agriculture has lost its substantial position in 

the country’s economy. However, farmer group activities 

through PIM/G create new values for socioeconomic externalities 

as a safety net against natural disasters and as a generator 

of resources for other sectors. At the WRRC, I can mention 

examples of providing water from agriculture to domestic 

use during severe dry spells, increasing the performance of 

paddy fields while protecting reservoirs during extreme floods, 

creating winter sanctuaries for migratory birds, and restoring 

groundwater levels for downstream water use in cities in winter. 

The value of socioeconomic externalities cannot be commercially 

exchanged, whereas agricultural products are widely 

traded in the international market. Irrigation systems and water 

accompanied by the value of externalities are social overhead 

capital and commons. Therefore, development programs 

for irrigated agriculture should take into account external value 

as a benefit vis-à-vis the cost of the investment. 

I believe that improving water-use efficiency by fully 

taking into account the value of socioeconomic externalities 

can create a future with more sustainable agriculture through a 

rice-based system. Sustainability of this agriculture will be 

strongly supported by interactions between human activities 

through PIM/G and broad-banded products, including various 

economic values, the living environment, and cultural-religious 

activities, as well as the agricultural product rice. 

The proposition I have maintained in this paper might 

be a claim for the renaissance of old systems because water 

management in the 21st century requires a holistic approach 

in harmony with various factors, including externalities, as 

some old irrigation systems have achieved. These old systems 

make a good guide for future sustainable development. 

I especially wish to acknowledge the contributions of 

Mr. Kazumi Yamaoka of the National Institute for Rural Engineering 

in co-authoring this document and preparing the presentation 

materials. 

Research strategy for rice in the 21st century 

Ronald P. Cantrell and Gene P. Hettel 

Thank you for this opportunity to speak 

to you as one of the keynoters of this 

week’s World Rice Research Conference, 

which—as the culminating scientific 

event of the International Year 

of Rice 2004—has brought to Tsukuba 

the planet’s leading rice scientists to exchange 

the latest research information 

on key rice-related issues. 

Why rice research must continue 

Before we go into the details of what we think should be the 

research strategy for rice in the 21st century, I would like to 

discuss, briefly, why indeed rice research for developing countries 

must continue and be reinvigorated. 

Since the dawn of the Green Revolution—which began 

in Asia with IRRI’s release in 1966 of IR8, the first modern, 

high-yielding semidwarf rice variety—the global rice harvest 

has more than doubled, racing slightly ahead of population 

growth. This increased production and the resulting lower 


prices of rice across Asia have been one of the most important 

results of the higher yields that rice research and new farming 

technologies have made possible. Around 1,000 modern varieties—approximately 

half the number released in 12 countries 

of South and Southeast Asia over the last 38 years—are linked 

to germplasm developed by IRRI and its partners alone—a 

large impact indeed (Cantrell and Hettel 2004a). These modern 

varieties and the resultant increase in production have increased 

the overall availability of rice and also helped to reduce 

world market rice prices by 80% over the last 20 years. 

Poor and well-to-do farmers alike have benefited directly 

through more efficient production that has led to lower unit 

costs and increased profits. Poor consumers have benefited 

indirectly through lower prices. This has brought national food 

security to China and India, not to mention Indonesia and other 

countries. 

So, why is more rice research needed Clearly, there are 

two major integral challenges—for now and well into the 21st 

century—involving rice—and particularly in Asia (Cantrell and 

Hettel 2004a). The first is the ability of nations to meet their 

national and household food security needs with an ever-declining 

natural resource base, two of the critical resources being 

water and land. How the current level of annual rice production 

of around 545 million tons can be increased to about 

700 million tons to feed an additional 650 million rice eaters 

by 2025 (D. Dawe, IRRI, 2004, pers. comm.) using less water 

and less land is indeed one great challenge in Asia. 

The second is—as has been stated so eloquently by the 

United Nations as one of its eight Millennium Development 

Goals (www.undp.org/mdg)—the eradication of extreme poverty 

and hunger. Rice is so central to the lives of most Asians 

that any solution to global poverty and hunger must include 

research that helps poor Asian farmers reduce their risks and 

earn a decent profit while growing rice that is still affordable 

to poor consumers. 

These new challenges that we face today may not be on 

the mammoth scale of those at the onset of the Green Revolution 

some 38 years ago when large parts of Asia were on the 

brink of famine and starvation. But perhaps they are even more 

difficult from a technology standpoint. Indeed, the challenges 

for rice farmers and researchers in 1966—although daunting— 

were fairly straightforward. Renowned economist Prof. Peter 

Timmer points out that the task of agricultural development 

was much easier back then when the need for greater cereal 

output to accomplish national 

food security was met by new 

seed-fertilizer technologies, 

which were already in fairly 

advanced stages of development 

(Timmer 2003). Today, 

it is important to elevate the 

awareness of key 

policymakers and donors that 

rice research must continue 

so that we all can face these 

new—and much more complicated—challenges 

of the 

21st century. 

Research priorities 

Bridging the yield gap 

So, now that we have made the case that rice research must 

continue, what are the priorities Clearly, one area of research 

involves the continuing effort to close the quite large rice yield 

gap between what is possible and what is actually achieved in 

farmers’ fields—especially in the more marginal environments 

such as the rainfed lowlands and uplands—across regions, 

ecological zones, and crop seasons in all rice-growing countries 

in the Asia-Pacific Region (FAO 1999, Pingali 2001). 

This yield deficit ranges from 10% to 60% between attainable 

and economically exploitable yields depending on the ecosystems 

and countries. The various factors currently contributing 

to the yield gap in different countries and regions include biophysical, 

technical/management, socioeconomic, institutional/ 

policy, technology transfer, and adoption/linkage problems 

(FAO 1999). 

Reducing the yield gap can be seen as the local solution 

to a global problem. It can lead to increased production with 

the additional incentives of cost reduction, poverty alleviation, 

social justice, and equity. The FAO suggests that reducing 

the yield gap could alone supply as much as 60% of the 

increased annual rice demand by 2025 that we mentioned earlier 

(FAO 1999). 

Annual changes in weather can cause yield variations of 

nearly 30% and this can mask gains made by plant breeders. 

Currently, we estimate that breeding new elite cultivars and 

Keynotes 27

using the best management techniques could lead to yield improvements 

of 18% in the coming decades if the negative effects 

of climate change can be avoided (Sheehy et al 2005). 

The incidence of weather extremes appears to be increasing 

and there is a need to breed new rice cultivars that can tolerate 

the worst aspects of climate change. 

Recent IRRI research has brought both good and bad 

news. The good news is: It appears that improved crop and 

nitrogen management has effectively closed the yield gap between 

simulated potential yield and actual yield of existing 

elite cultivars in long-term plots at IRRI (Kropff et al 2003). 

The bad news is: Even when these improved crop and nitrogen 

management practices are employed, there can still be yield 

fluctuations, which are largely associated with climatic factors 

(Dobermann et al 2000). Recent fluctuations in grain yield 

trials at IRRI have been associated with reduced solar radiation 

and increased seasonal mean minimum temperature, and 

not related to crop management (Kropff et al 2003, Peng et al 

2004b). 

In many countries, new technologies are not making the 

impact that they should be in closing the yield gap. To make 

any major breakthroughs in the immediate future will first require 

the bridging of another gap—the knowledge gap 

(Schiffrin 2001). Extension specialists and the farmers they 

serve are starved for information. They need new management 

skills and technical know-how as soon as possible. 

To help close this knowledge gap by getting technology 

off the shelf and into farmers’ hands, scientists and information 

and communication technology (ICT) specialists at IRRI 

have placed a Rice Knowledge Bank on the Web 

(www.knowledgebank.irri.org) aimed at extension workers and 

farmers. At the speed of a keystroke, it can provide the latest 

information—from diagnosing field problems via RiceDoctor 

(www.knowledgebank.irri.org/riceDoctor_MX/default.htm) to 

making crop management decisions via TropRice 

(www.knowledgebank.irri.org/troprice/default.htm). Usually, 

only more prosperous farmers are currently going online. But, 

sooner than some might expect, even the poorest farmers will 

be going online to access field problem and crop management 

information near where they toil in their rice fields. For example, 

in the Philippines, the Open Academy of Philippine 

Agriculture predicts that within 2 years it will be commonplace 

for Filipino farmers—through strategically located 

Internet kiosks in farming villages—to be diagnosing crop pests 

by accessing the Web and sharing information with peers anywhere 

in the country using electronic mail and cell phone technology 

(Marquez 2004). 

In addition, continued on-farm research will be needed 

to fine-tune new technologies if the yield gap is to be narrowed. 

For example, IRRI has 

a long tradition of listening to 

farmers (Hettel and Dedolph 

1996) and planning farmer 

participatory research (FPR). 

FPR is an approach that involves 

encouraging farmers to 

engage in experiments in their 

own fields so that they can 

learn, adopt new technologies, 

and spread them to other farmers. 

With scientists acting as 

facilitators, farmers and scientists 

work closely together 

from initial design of the research 

project to data gathering, 

analysis, final conclusions, 

and follow-up actions. 

A great example of this is the ongoing FPR in Bangladesh 

where farmers’ livelihoods are being improved by closing rice 

yield gaps caused by weeds. Farmers in Bangladesh grow rice 

over a wide range of rainfed and irrigated lands and weeds are 

a major constraint. In this project of the Irrigated Rice Research 

Consortium (www.irri.org/irrc/weeds/closing.asp), researchers 

are examining two contrasting farming systems in 

Comilla and Rajshahi districts to learn how different rice-cropping 

systems affect weed abundance and methods for their 

management. This work has shown the benefit of integrated 

management involving herbicides in combination with manual 

weeding. Application of herbicides or use of a hand-pushed 

weeder in transplanted rice results in similar yields—and a 


one-third reduction in weeding costs—compared to farmers’ 

current practice of weeding twice by hand. 

In one study, more than 100 farmers in both districts participated 

in measuring the yield gaps caused by weeds in their 

own fields under their own practices. The results are helping 

them to make management decisions about whether or not to 

invest in additional weeding and/or fertilizer. 

Maintenance research (MR)—exciting 

and award-winning and a jumpstart 

for a Doubly Green Revolution in rice! 

A second area of research that we must continue is maintenance. 

The tendency for some researchers is to continually 

chase the newest fad—organizations sometimes stop doing one 

thing and start doing another because it might be easier to “sell” 

the latest research trend that is getting the “buzz.” Indeed, research 

with the mundane label of “maintenance” is often difficult 

to promote to donors. But, even so, truly successful organizations 

will always have that appropriate mix of MR and 

new strategic directions. 

For example, although IRRI is embarking on novel and 

exciting strategic research on a number of fronts involving new 

tools and directions (mentioned later in this paper), it is still 

committed to MR. For example, during our recent external review, 

we emphasized to the review panel that our germplasm 

improvement work will continue to contribute to both maintaining 

yield gains and increasing yield potential—at least 1% 

per year (the average of the recent past). In light of what we 

are discovering in our long-term yield plots (mentioned above), 

our maintenance breeding will be placing particular emphasis 

on stress associated with anticipated environmental and climatic 

changes. 

Dana Dalrymple, senior research adviser and agricultural 

economist at USAID, who has been thinking and writing 

about MR recently, says we cannot take current yield levels 

for granted and that an increasing amount of MR will be needed 

to hold on to the current levels. He adds that MR, which is 

heavily oriented to plant and animal diseases and insect pests, 

tends to be taken for granted until there is a severe outbreak of 

some type and to get overlooked in priority setting (Dalrymple 

2004). 

One great challenge in agricultural MR is the development 

of natural resistance to reduce chemical usage in agriculture. 

The search for new chemical treatments in agriculture 

has gone on for over a century and is never-ending (Dalrymple 

2004). In rice, IRRI is countering this trend with significant 

projects in Asia—showing that MR is not necessarily mundane 

and it can even be exciting and award-winning and that, 

indeed, a Doubly Green Revolution in rice has already been 

jumpstarted (Cantrell and Hettel 2004b). 

China. IRRI researchers and collaborators achieved notable 

success with integrated pest management (IPM) in China’s 

southwestern province of Yunnan. There—in what the New 

York Times called “one of the largest agricultural experiments 

ever” (Yoon 2000)—we found that intercropping rows of different 

varieties of rice can almost completely control the devastating 

rice blast disease (Zhu et al 2000). Some farmers there 

were already using this technique, albeit in a haphazard way. 

Keynotes 29

We scientifically tested several variations of the concept and 

improved it. Now we are disseminating our findings with confidence 

that the practice not only reduces farmers’ reliance on 

chemical pesticides, thereby protecting the environment, but 

also improves yield and income to give farmers the options 

they need to break out of the poverty trap. Word-of-mouth is 

already leading to the technique’s wide adoption on nearly 1 

million hectares across 10 provinces in China. 

Vietnam. Our Vietnam study offers valuable lessons on 

how to reduce pesticide applications. We converted our findings 

into one simple rule: “Don’t spray for the first 40 days.” 

We launched a media campaign to deliver the message to farmers, 

stressing the cost savings and health benefits of reduced 

spraying. The result In the test area of 21,000 households, 

after an 18-month interval, we recorded a 53% reduction in the 

number of insecticide applications—without affecting yield! 

Many farmers reduce input costs by US$30–50 per season— 

equal to a month’s income in Vietnam. Eventually, this effort 

persuaded almost 2 million rice-growing households in the 

Mekong Delta to cut back on using harmful and unnecessary 

farm chemicals. Underscoring the significance of this work, 

the project received the St. Andrews Prize for the Environment 

from Scotland’s St. Andrews University in 2002 and the International 

Green Apple Environment Award from the U.K.-based 

Green Organization in 2003 (IRRI 2003). 

SSNM. Our past research has produced awareness of excessive 

use of chemical nitrogen fertilizer in farmers’ fields. 

So, we are now promoting a technique called site-specific nutrient 

management (SSNM) across 20 key sites in eight countries 

in tropical and subtropical Asia. The technique enables 

farmers to “feed” the rice plant with nutrients only as and when 

needed, with optimal use of existing indigenous nutrient 

sources, including crop residues and manures, and timely application 

of fertilizer (Wang et al 2004). 

More efficient use of nitrogen through the principles of 

SSNM in ratios to better match plant needs can reduce the 

incidence of rice disease and insect pests, which in turn further 

reduces the need for pesticides. The benefits from SSNM 

multiply when improved management of phosphorus and potassium 

is thrown into the mix with improved N management. 

Improved management of N, P, and K fertilizers through SSNM 

is now increasingly being shown in Asia to reduce disease and 

insect damage (Buresh 2004, Jahn 2004, Wang et al 2004)— 

yet another avenue on which farmers can reduce their need for 

pesticides and usher in the Doubly Green Revolution in rice. 

When does MR begin There are some interesting new 

twists on the importance of maintenance research and Dr. 

Dalrymple has been recently looking at the interface between 

MR and what he calls productivity-enhancing research (PER). 

He assumes that there is an area of overlap between the two, 

and that, in some cases, particularly significant advances in 

MR could actually contribute to increasing yield. This is likely 

to be true in the case of resistance to plant diseases and insects 

(D. Dalrymple, USAID, 2004, pers. comm.). Clearly, the line 

between MR and PER is fuzzy and it raises some definitional 

questions, such as, When does MR actually begin to kick in 

Otsuka et al (1994) concluded that the Green Revolution would 

not have been very revolutionary without the development and 

the diffusion of second-generation modern rice varieties with 

multiple pest and disease resistance. So, depending on where 

MR is placed on the rice-improvement breeding continuum, it 

can be argued that, without it, there might not have been a 

Green Revolution. 

Exciting strategic research—new directions 

and new tools to get us there 

New tools such as biotechnology are affording us the opportunity 

to follow some new strategic research directions. As we 

mentioned earlier, any successful agricultural research institution 

must have a harmonious mix of maintenance research and 

strategic research on its agenda. We believe that our strategic 

directions addressing such critical issues as the rice yield barrier, 

global warming, the water crisis, and human nutrition are 

the right blend for IRRI. 

Breaking the yield barrier. During the recent 4th International 

Crop Science Congress in Brisbane (26 Sept.-1 Oct.), 

IRRI researchers provided an update on progress in breaking 

the yield barrier using the ideotype—or idealized plant—approach 

(Peng et al 2004a). Great progress has been made with 

second-generation new plant type (NPT) lines developed by 


crossing elite indica plants with improved tropical japonica 

plants (Virk et al 2004). Our researchers believe that the recent 

success of “super” hybrid rice breeding in China combined 

with progress of NPT breeding at IRRI will contribute 

greatly to the goal of breaking the yield barrier in irrigated 

rice-based cropping systems. 

However, leaving no stone unturned in the quest for a 

higher yield potential in rice, we are also looking at the link 

among photosynthesis, yield, and radiation-use efficiency— 

or RUE. Some of our scientists have concluded that the upper 

yield limit of rice with its conventional photosynthetic pathway 

will go only halfway to a target of increasing rice yield 

50% by 2050. Improved crop photosynthesis would then seem 

essential. One proposal for increasing rice’s RUE is to incorporate 

the high C 4 photosynthetic capacity of a crop such as 

maize into rice, which is a less photosynthetically efficient C 3 

cereal (Sheehy et al 2000). An additional benefit of such a 

change would be a doubling in the efficiency with which the 

plant uses water for growth. 

Making the photosynthetic pathway of rice resemble that 

of maize would require a long-term genetic engineering project 

(10–15 years) to introduce genes for enzymes of the C 4 pathway 

and for leaf anatomy. If this were accomplished, the benefits 

would be enormous across the rice ecosystem spectrum. 

A C 4 rice plant would yield the same as a C 3 with half the 

transpirational water loss. It would also require significantly 

less N fertilizer, thus providing for a cleaner environment and 

lower input costs for farmers. In irrigated rice, yield potential 

would rise significantly. In drought-prone ecosystems (rainfed 

lowland and upland rice), yield could be maintained or increased 

with less water and less fertilizer, especially when 

coupled with the predicted rising atmospheric concentration 

of carbon dioxide that is associated with future world climate 

change (see below). Farmers living at the margins in these ecosystems 

would see improvements in yield and yield stability. 

It would be a new revolution in rice farming. 

Global warming. The September 2004 issue of National 

Geographic (http://magma.nationalgeographic.com/ngm/ 

0409) devoted 74 pages to a three-part series of stories on 

global climate change. Under the umbrella title of Global 

Warning: Bulletins from a Warmer World, the editors state that 

there’s no question that Earth is getting hotter—and fast. Warning 

signs discussed included retreating glaciers, rising seas, 

shrinking lakes, damaged coral reefs, and declining penguin 

populations in the Antarctic. Another warning sign that could 

have also been included comes directly from the irrigated field 

experiments at IRRI’s headquarters in the Philippines. 

Research at IRRI agrees with predictions from theoretical 

studies that global warming will make rice crops much less 

productive unless research to ameliorate its worst effects is 

undertaken (Sheehy et al 2005). Early work (Satake and 

Yoshida 1978) showed that, as day temperatures increased from 

34 to 40 °C, floret sterility increased to 100% and yield fell to 

zero. Later, Ziska and Manalo (1996) showed that very high 

nighttime temperatures could also reduce yield. More recently, 

Peng et al (2004b) suggested that increasing mean minimum 

nighttime temperatures were already decreasing yield in the 

field acting through previously unknown mechanisms. The 

Third Assessment Report of the Intergovernmental Panel on 

Climate Change (IPCC; www.ipcc.ch) forecasts that by 2100 

mean planet-wide surface temperatures will rise by 1.4 to 5.8 

°C. Global warming could threaten to erase the hard-won productivity 

gains that have kept the rice harvest in step with population 

growth. 

Our future studies will focus on investigating the differential 

effects of global warming on night versus daytime temperatures 

and searching for more heat-tolerant rice. We must 

also look at the effects of manmade ozone and other gaseous 

pollutants that could have serious effects on rice yield as Asia 

continues to develop. Climate and weather are becoming more 

complex because of these changing gaseous components in 

the atmosphere. An important tie-in will include the photosynthetic 

pathway research we mentioned earlier—a C 4 rice plant 

would fare much better in the warmer world of the future. These 

are certainly important issues to include on our strategic research 

agenda since—as the UK’s Department for International 

Development states—we cannot afford to ignore the impact of 

increasing climate risks to particularly the poor in Asia 

(www.dfid.gov.uk/pubs/files/climatechange/11asia.pdf). 

The water crisis. As put forth by the CGIAR Challenge 

Program on Water and Food, increasing water scarcity and 

competition for the same water from nonagricultural sectors 

point to an urgent need to improve crop water productivity to 

ensure adequate food for future generations with the same or 

less water than is presently available to agriculture 

(www.waterforfood.org). About 70% of the water currently 

withdrawn from all freshwater sources worldwide is used for 

agriculture and to grow rice requires about twice as much water 

as other grain crops such as wheat or maize. In Asia, irrigated 

agriculture accounts for 90% of the total diverted fresh 

water used, and more than 50% of this is used to irrigate rice. 

Keynotes 31

Until recently, this amount of water has been taken for granted, 

but this cannot continue. 

To meet the water crisis head-on for rice, the International 

Platform for Saving Water in Rice (www.irri.org/ipswar/ 

about_us/ipswar.htm) has been created as a mechanism to increase 

the efficiency and enhance the coherence of research 

on water savings in rice-based cropping systems in Asia. The 

overarching goal is to conserve water resources, which will in 

turn safeguard national and household food security and alleviate 

poverty. 

About 75% of the world’s rice is produced in irrigated 

fields, where farmers can control the amount of water applied. 

In the tropics, these systems can produce yields up to 9 t ha –1 , 

but they typically use from 1,000 to 2,000 mm of water per 

hectare. This means that from 1,000 to more than 3,000 L of 

water are used to produce each kg of rice. In contrast, rainfed 

rice can produce yields of up to 5 t ha –1 with a water input of 

only 800 mm of well-distributed rainfall. 

Approaches to improving water productivity in both irrigated 

and rainfed rice systems involve decreasing the water 

requirement of irrigated systems without affecting yield and 

increasing the yield of rainfed systems without increasing their 

water use. These approaches are complementary, with spillover 

benefits expected across the irrigated and rainfed systems. The 

recent advancement of crop sciences, especially in rice 

genomics (see below) and linkage of conventional breeding 

with molecular tools, has shown some promising possibilities 

for making such innovations a reality. We are already aiming 

our strategic research to attack the water problem from several 

angles, using modern tools of molecular biology, environmental 

monitoring, simulations, and database integration. 

Two types of water-saving systems can be used to replace 

the traditional irrigated rice production schemes that are 

now under threat. One is alternate wetting and drying (AWD). 

In this system, the field is irrigated with enough water to flood 

the paddy for 3 to 5 days, and, as the water soaks into the soil, 

the surface is then allowed to dry for a few days (usually from 

2 to 4) before getting re-flooded. 

Another alternative is dry-field or aerobic rice, where 

the rice is sown directly into dry soil, like wheat or maize, and 

irrigation is applied to keep the soil sufficiently moist for good 

plant growth, but the soil is never saturated. Both of these systems 

allow for substantial water savings of from 30% to 50%. 

Our strategic research on water will be looking closely 

at both of these systems. For example, we have already established 

an Aerobic Rice Working Group, involving breeders, 

physiologists, and water and soil scientists, that is striving to 

overcome the many difficulties in taking rice out of its natural 

aquatic environment. By developing a completely new management 

system, the new aerobic rice should be able to yield 

up to 6 t ha –1 using only half the water! 

Human nutrition. And finally, the world’s poor will 

achieve household food security only when—in addition to 

being available in sufficient quantity—the food is of ample 

quality as well. Although rice supplies adequate energy in the 

form of calories and is a good source of thiamine, riboflavin, 

and niacin, it is lacking as a source of vitamin A and other 

critical vitamins, iron, zinc, and other micronutrients and amino 


We are happy to announce some other exciting developments 

in our biofortification work to produce iron-enriched 

rice. First, IRRI and collaborators, including here in Japan, 

have introduced an iron-enhancing ferritin gene to indica rice 

in such a way that it expresses itself in the rice endosperm. 

Thus, after polishing, the rice grains contain three times more 

iron than usual (Vasconcelos et al 2003). This is indeed the 

most significant increase in iron ever achieved in an indica 

rice variety and it could have important benefits for the 3.5 

billion people in the world who have iron-deficient diets. 

Second, IRRI is part of a forthcoming, multi-institutional 

publication that will detail the results of a 9-month human biological 

efficacy study of the consumption of iron-enriched rice 

(Studdert 2004). It involved 317 religious sisters from 10 convents 

in the Philippines. The results of this large feeding trial 

demonstrated that there is indeed a significantly positive effect 

for women with iron-poor diets who consume biofortified 

rice. The high-iron rice, IR68144-3-3B-2-2-3, used in this ironuptake 

study is now a pre-release variety in the Philippines 

with the name Maligaya Special #13 (MS-13). 

acids that are essential to human health, especially the health 

of women and children. We believe the nutrient content of rice 

can be improved substantially by using both traditional plant 

breeding and new biotechnology approaches. 

IRRI is a major participant in the CGIAR Challenge Program 

called Harvest Plus (www.harvestplus.org), which is seeking 

to reduce the effects of micronutrient malnutrition by harnessing 

the power of plant breeding to develop staple food 

crops that are rich in micronutrients, a process called 

biofortification. IRRI has been active in research on enhancing 

micronutrient levels in rice through genetic engineering 

and leading the development at IRRI of tropical varieties of 

vitamin A-enriched golden rice (Datta et al 2003). It was only 

in early 2001 that the first seed samples were delivered to IRRI 

by Professor Ingo Potrykus and Dr. Peter Beyer, the inventors 

of this genetically modified rice, which could save half a million 

children each year from irreversible blindness. 

Our team of scientists has bioengineered several Asian 

indica varieties with genes for beta-carotene biosynthesis. Selected 

lines—including genotypes of IR64—show expression 

of beta-carotene, the precursor of vitamin A (Datta et al 2003). 

These IR64 golden rice events are now being evaluated for 

agronomic performance. IRRI has shown that very little of the 

beta-carotene in the golden rice is lost during milling and cooking. 

There is still a lot of work to do in this strategic research 

that involves bioavailability tests and working with national 

partners who will advance the leading materials following local 

biosafety regulations. Indica golden rice is still probably 

4–6 years away from release to farmers. 

Research consortia—the way of the future 

Significant progress in agricultural research—especially strategic 

research—cannot be accomplished in a vacuum. With 

today’s rapid accumulation of “mountains” of scientific data, 

how we conduct research is 

changing rapidly as well. 

Back in the 19th century, a 

genius such as Gregor 

Mendel—the Father of Genetics—could 

come up with 

brilliant and unprecedented 

theories of heredity in relative 

isolation while tending pea 

pod plants in his monastery 

garden. But today, researchers 

in our various public and 

private institutions—the geniuses 

of today—can rarely 

accomplish much in isolation—even 

in their sophisticated 

laboratories. 

The need for researchers to work together in collaboration 

to accomplish a goal is paramount—and the best mechanism 

we have seen for this is the concept of a consortium. A 

simple definition of “consortium” is a group of individuals or 

institutions/companies formed to undertake an enterprise or 

activity that would be beyond the capabilities of the individual 

members or difficult to perform effectively without partnership. 

International Rice Genome Sequencing Project 

The finest example of a consortium we can give is the International 

Rice Genome Sequencing Project (http:// 

Keynotes 33

gp.dna.affrc.go.jp/ 

IRGSP/index.html), a 

leading member of which 

is the Rice Genome Research 

Program (http://rgp.dna.affrc.go.jp), based right here 

in Tsukuba. The IRGSP, a consortium of publicly funded laboratories 

in both developed and developing countries (Japan, 

USA, China, Taiwan-China, Korea, India, Thailand, France, 

Brazil, and the United Kingdom), was established in 1997 to 

obtain a high-quality, map-based sequence of the entire 12- 

chromosome rice genome using the cultivar Nipponbare of 

Oryza sativa subsp. japonica. 

Five years ahead of the original schedule, the IRGSP, in 

December 2002, announced the completion of a high-quality 

draft sequence of the genome (Feng et al 2002, Sasaki et al 

2002). This great achievement could not have been accomplished 

in any one country or region. The rice genome sequence, 

now freely available on the Internet (see http:// 

rgp.dna.affrc.go.jp/cgi-bin/statusdb/irgsp-status.cgi) with data 

integrated from Monsanto and Syngenta, is already inducing 

innovative research worldwide in the new science of functional 

genomics. With 50,000 or more genes predicted in the DNA 

sequence, the task ahead requires the development of many 

genetic resources to unravel the function and interaction of 

these genes and their relationship with important traits. 

International Rice Functional Genomics Consortium 

Enter the International Rice Functional Genomics Consortium. 

Parallel to the initiation and activities of the IRGSP, IRRI recognized 

the need to bring together diverse expertise to capitalize 

on the wealth of genomic 

information for functional 

analysis. Informal 

meetings were held in 1999 

and 2000 that led to the formation 

of an International 

Rice Functional Genomics 

Working Group. This Working 

Group brought together 

research groups from around 

the world to discuss mutual 

interests and promoted collaboration 

in anticipation of 

the completion of the rice 

genome sequence. During 

this time, members of this research 

community developed useful materials to form a base 

for collaboration in functional genomics studies in rice involving 

the genome sequence itself, gene cloning and arrays, and 

mutants (Phillips et al 2004). 

After two years of discussion, the idea of creating a more 

structured consortium was discussed at a Working Group meeting 

in Beijing in September 2002, and at the November 2002 

meeting in Canberra, Towards building a global rice gene 

machine (www.pi.csiro.au/grgm02/home1.htm), organized by 

the Commonwealth Scientific and Industrial Research Organization. 

An interim steering committee for the consortium was 

formed at the Plant and Animal Genome Conference in San 

Diego in January 2003, leading to the formal inauguration of 

the International Rice Functional Genomics Consortium 

(www.iris.irri.org/IRFGC). 

Currently, the IRRI-coordinated consortium is guided by 

a 21-member steering committee, which includes 17 institutions 

from 10 countries—including Japan—and two Future 

Harvest Centers—CIAT and IRRI. The collective goals are to 

share materials, integrate databases, seek bilateral and multilateral 

partnerships, implement initiatives for the cooperative 

elucidation of gene function, and accelerate delivery of research 

results to benefit rice production (Phillips et al 2004). Again, 

this is an effort that no one institution, country, or region could 

take on in isolation. For example, through the IRFGC, IRRI is 

investing its USAID Linkage Program funds to support eight 

laboratories in the United States to identify genes for stress 

tolerance (www.iris.irri.org/irfgc/researchstrength.shtml). This 

unique consortium embraces and encourages the role of developing 

countries in gene discovery and applications of the 

new science (Leung and An 2004). IRRI will continue to work 

closely with the international community to pursue initiatives 

that will accelerate gene discovery and take advantage of the 

new knowledge to make a true impact. 

To celebrate the completion of the rice genome and learn 

about the latest developments in functional as well as structural 

and evolutionary genomics research, IRFGC members 

will gather during the 2nd Annual Symposium on Rice Functional 

Genomics (www.rfg2004.org) later this month (15-17 

November) in Tucson, Arizona. We’re truly excited about what 

the consortium will be coming up with in the years to come. 

Other important rice-related consortia 

Within the CGIAR, IRRI has long taken the lead using the 

consortium concept to link research capacity of national agricultural 

research and extension systems (NARES) with that of 

IRRI’s to solve important problems through multi-country collaboration. 

We are using consortia to provide ways to conduct 

primarily strategic research by sharing research responsibilities 

according to each partner’s interests and capabilities. The 

key to success here is that we have, in consortia, truly peer 

relationships among NARES and with IRRI for much of the 

strategic research we mentioned earlier, with a substantial commitment 

of staff and resources. The special feature of collaboration 

in these consortia is that IRRI scientists put substantive 

components of their own strategic research at the consortia 

sites with NARES partners. 

Some key consortia we would like to mention briefly 

are the Irrigated Rice Research Consortium (IRRC), the new 

Consortium for Unfavorable Rice Environments (CURE), and 

the Rice-Wheat Consortium (RWC). 

IRRC (www.irri.org/irrc). Some 2.2 billion Asian rice 

farmers and consumers depend on the sustainable productivity 

of irrigated rice ecosystems. The IRRC, composed of 

NARES from Bangladesh, Cambodia, China, India, Indonesia, 

Lao PDR, Malaysia, the Philippines, Thailand, and Viet- 


nam, is working with IRRI to promote research collaboration 

and integration of research, strengthen multi-institutional and 

interdisciplinary research, and facilitate technology delivery. 

Farmers’ knowledge and input are key ingredients as we illustrated 

earlier with the example of the ongoing weed research 

to close the yield gap in Bangladesh (www.irri.org/irrc/weeds/ 

closing.asp). 

CURE. Some of the greatest challenges facing rice researchers 

are found in the fragile rainfed and marginal environments 

where few of the modern technologies targeted to 

high productivity through high levels of inputs are appropriate. 

To make rice research more relevant under rainfed conditions, 

it is necessary to conduct the research on-site in a wide 

range of environments that present specific constraints to production. 

Since 1991, IRRI had worked in partnership with 

NARES to conduct rice research in these unfavorable environments 

through two research consortia—one for the rainfed 

lowlands and one for the uplands. 

A review in 2001 indicated that the consortium approach 

had notable successes in developing new technologies more 

suited to these environments. However, we decided to restructure 

the two consortia to consolidate the research programs 

and focus on increasing impact through wider dissemination 

of the research findings—hence, the June 2002 birth of CURE, 

the streamlined Consortium for Unfavorable Rice Environments 

(www.irri.org/cure/cure.htm). Consortium members— 

Asian NARES in Bangladesh, Cambodia, India, Indonesia, 

Laos, Myanmar, Nepal, the Philippines, Thailand, and Vietnam; 

advanced research institutes; and other international agricultural 

research centers—have adopted a holistic and systems 

approach to achieve a higher degree of integration of the 

disciplines in identifying and addressing problems relating to 

Keynotes 35

agricultural production and rural livelihoods involving—but 

not necessarily confined to—rice. We are confident that the 

new consortium will serve as a platform and forum for identifying 

and prioritizing the rainfed research needed to generate 

impact in farmers’ fields. 

RWC. Bangladesh, India, Nepal, and Pakistan have devoted 

to agriculture nearly half of their total land area of 402 

million hectares to feed and provide livelihoods for 1.8 billion 

people. Rice and wheat, the staple food crops for these people, 

contribute more than 80% of the total cereal production in these 

countries. 

Celebrating its 10th anniversary this year, the Rice-Wheat 

Consortium (www.rwc.cgiar.org/rwc), a NARES-driven initiative 

of the four countries, has truly emerged as an innovative 

model for regional and international collaboration. Operating 

as an inter-institutional and intercenter multidisciplinary network 

facilitating system-based farmer participatory research 

in the rice-wheat ecology of the Indo-Gangetic Plains, it has 

successfully addressed the issues of productivity enhancement 

of rice and wheat in a sustainable fashion. These issues include 

tillage and crop establishment, crop diversification, 

postharvest, and environmental benefits. Through the RWC’s 

efforts, tens of thousands of resource-poor farmers in South 

Asia have been able to increase their income by using practices 

that save water, fuel, and other inputs; facilitate timely 

planting; reduce tillage needs and crop residue burning; and 

allow them to diversify their cropping systems (Ladha et al 

2003). 

Last year, external reviewers pointed out that “the effectiveness 

of partnership among IRRI and CIMMYT and their 

NARES partners as well as within and between the four national 

systems is one of the most important achievements of 

the RWC”—and truly this is what consortia are all about and 

why they have the potential to achieve so much. 

Summary and conclusions 

We are happy to have been able to keynote this scientific meeting 

of the WRRC—this culminating conference of the International 

Year of Rice—to suggest what the strategy should be 

for rice research in the 21st century. As we have just outlined, 

we feel that strategy must include 

Renewing and invigorating efforts to close the yield 

gap in rice; 

Remembering the critical importance of maintenance 

research; 

Using cutting-edge tools to enhance strategic thrusts 

to (1) break the yield barrier, (2) face head-on the 

crises of impending global warming and water shortages, 

and (3) improve human nutrition; and 

 

Employing the consortium model whenever and wherever 

possible for researchers to work together to accomplish 

the important goals we just discussed. 

If you remember anything of what we’ve discussed today, 

we hope it is that, for any agricultural research institution 

to be successful, it must have a harmonious mix of maintenance 

research and strategic research on its agenda. There really 

can’t be one without the other in the long term. The dual 

challenges are producing new products to solve future problems 

caused by diminishing resources and climate change and 

maintaining and securing what we have already achieved. 

References 

Buresh RJ. 2004. Specific benefits: farmers earn more from their 

rice crop by scientifically optimizing fertilizer use. Rice Today 

3(4):32-33. www.irri.org/publications/today/pdfs/3-4/ 

Ricetoday_3-4.pdf. 

Cantrell RP, Hettel GP. 2004a. New challenges and technological 

opportunities for rice-based production systems for food security 

and poverty alleviation. In: Proceedings of the FAO 

Rice Conference, 13-14 Feb. 2004. Rome (Italy): Food and 

Agriculture Organization. www.fao.org/rice2004/en/pdf/ 

cantrell.pdf. 

Cantrell RP, Hettel GP. 2004b. The doubly green revolution in rice. 

Presentation at the World Food Prize Symposium: Rice, 

Biofortification, and Enhanced Nutrition, Des Moines, Iowa, 

14-15 October 2004. 

Dalrymple DG. 2004. Thoughts on context and reform in agriculture. 

Workshop on Technology Development in the Life Sciences 

sponsored by the Program on Science, Technology and 

Global Development, Center on Globalization and Sustainable 

Development, The Earth Institute, Columbia University, 

20-21 May. www.earthinstitute.columbia.edu/cgsd/events/ 

documents/dalrymple_000.doc. 

Datta K, Baisakh N, Oliva N, Torrizo L, Abrigo E, Tan J, Rai M, 

Rehana S, Al-Babili S, Beyer P, Potrykus I, Datta S. 2003. 

Bioengineered ‘golden’ indica rice cultivars with beta-carotene 

metabolism in the endosperm with hygromycin and mannose 

selection systems. Plant Biotechnol. J. 1(2):81-90. 

www.blackwell-synergy.com/links/doi/10.1046/j.1467- 

7652.2003.00015.x/abs/ 

Dobermann A, Dawe D, Roetter RP, Cassman KG. 2000. Reversal 

of rice yield decline in a long-term continuous cropping experiment. 

Agron. J. 92:633-643. 

FAO (Food and Agriculture Organization). 1999. Regional expert 

consultation on bridging the rice yield gap in the Asia and 

Pacific Region. FAO Regional Office for Asia and the Pacific, 

Bangkok, Thailand, 5-7 October 1999. www.fao.org/ag/ 

AGP/AGPC/doc/field/commrice/session/54.htm. 

Feng Q, Zhang Y, Hao P, et al. 2002. Sequence and analysis of rice 

chromosome 4. Nature 420:316-320. 

Hettel GP, Dedolph C, editors. 1996. Listening to the farmers. Los 

Baños (Philippines): International Rice Research Institute. 97 

p. www.irri.org/publications/annual/ar1996.asp. 

IRRI. 2003. Innovative response to pesticide misuse. IRRI press release. 

www.irri.org/media/press/press.aspid=79. 

Jahn GC. 2004. Effect of soil nutrients on the growth, survival and 

fecundity of insect pests of rice: an overview and a theory of 

pest outbreaks with consideration of research approaches. 

Multitrophic interactions in soil and integrated control. International 

Organization for Biological Control Bulletin 27 

(1):115-122. 

Kropff MJ, Cassman KG, Peng S, van Laar HH. 2003. Yields at 

IRRI research farm are still close to the climatic potential level. 

Int. Rice Res. Notes 28(2):19-21. www.irri.org/publications/ 

irrn/pdfs/vol28no2/IRRNmini3.pdf. 


Ladha JK, Hill JE, Duxbury JD, Gupta RK, Buresh RJ, editors. 2003. 

Improving the productivity and sustainability of rice-wheat 

systems: issues and impact. American Society of Agronomy 

Spec. Publ. 65. ASA, CSSA, SSSA, Madison, WI, USA. 211 

p. 

Leung H, An G. 2004. Rice functional genomics: large scale gene 

discovery and applications for crop improvement. Adv. Agron. 

82:55-111. 

Marquez C. 2004. First open academy for Filipino farmers goes hightech. 

Today, 24 August. p 4. 

Otsuka K, Gascon F, Asano S. 1994. ‘Second-generation’ MVs and 

the evolution of the Green Revolution: the case of Central 

Luzon, 1966-1990. Agric. Econ. 10:283-295. 

www.sciencedirect.com/science/article/B6T3V-45DNP4H- 

3M/2/a8a12f7c16ac9dfb644e024940332e02. 

Peng S, Laza RC, Visperas RM, Khush GS, Virk P, Zhu D. 2004a. 

Rice: progress in breaking the yield ceiling. In: Proceedings 

of the 4th International Crop Science Congress, 26 Sept.-1 

Oct., Brisbane, Australia. (In press.) 

Peng S, Huang J, Sheehy JE, Laza RC, Visperas RM, Zhong X, 

Centeno GS, Khush GS, Cassman KG. 2004b. Rice yields 

decline with higher night temperature from global warming. 

Proc. Natl. Acad. Sci. USA 101(27):9971-9975. 

www.pnas.org/cgi/content/full/101/27/9971. 

Phillips RL, Leung H, Cantrell RP. 2004. An international genetic 

platform for the assessment of gene function in rice. In: Proceedings 

of the 4th International Crop Science Congress, 26 

Sept.-1 Oct., Brisbane, Australia. (In press.) 

Pingali PL. 2001. Conventional research-based technology. In: 2020 

Focus 7, Appropriate Technology for Sustainable Food, Brief 

3. Washington, D.C. (USA): International Food Policy Research 

Institute. www.ifpri.org/2020/focus/focus07/ 

focus07_03.htm. 

Sasaki T, Matsumoto T, Yamamoto K, et al. 2002. The genome 

sequence and structure of rice chromosome 1. Nature 420:312- 

316. 

Satake T, Yoshida S. 1978: High temperature induced sterility in 

indica rice at flowering. Jpn. J. Crop Sci. 47:6-17. 

Schiffrin A. 2001. Of rice and men. The Industry Standard, 16 January. 

www.e-gateway.net/infoarea/news/news.cfmnid=1317. 

Sheehy JE, Mitchell PL, Hardy B, editors. 2000. Redesigning rice 

photosynthesis to increase yield. Makati City (Philippines): 

International Rice Research Institute and Amsterdam (The 

Netherlands): Elsevier Science B.V. 293 p. 

Sheehy JE, Elmido A, Centeno, G, Pablico P. 2005. Searching for 

new plants for climate change. J. Agric. Meteorol. 60(5): Special 

Edition. (In press.) 

Studdert L. 2004. Trials to gauge the effectiveness of boosting the 

iron content of rice are showing encouraging results that could 

drastically decrease micronutrient deficiencies. ADB Rev. 

36(3).www.adb.org/Documents/Periodicals/ADB_Review/ 

2004/vol36_3/rice_power.asp. 

Timmer CP. 2003. Agriculture and poverty. In: Mew TW, Brar DS, 

Peng S, Dawe D, Hardy B, editors. Rice science: innovations 

and impact for livelihood. Proceedings of the International 

Rice Research Conference, 16-19 September 2002, Beijing, 

China. Beijing (China): International Rice Research Institute, 

Chinese Academy of Engineering, and Chinese Academy of 

Agricultural Sciences. p 37-61. 

Vasconcelos M, Datta K, Oliva N, Khalekuzzaman M, Torrizo L, 

Krishnan S, Oliveira M, Goto F, Datta SK. 2003. Enhanced 

iron and zinc accumulation in transgenic rice with the ferritin 

gene. Plant Sci. 164(3):371-378. www.science.siu.edu/plant- 

biology/PLB425/Lit_files/set_7/ferritin%20- 

%20plant%20science%201.pdf. 

Virk PS, Khush GS, Peng S. 2004. Breeding to enhance yield potential 

of rice at IRRI: the ideotype approach. Int. Rice Res. Notes 

29(1):5-9. 

Wang R, Buresh R, Kam SP. 2004. Rice farming for sustainable resource 

use. Paper presented at the Workshop on Sustainable 

Use of Agricultural Resources and Environment Management 

with Focus on the Role of Rice Farming. Tokyo, Japan. 

Yoon CK. 2000. Simple method found to vastly increase crop yields. 

New York Times, 22 August. www.nytimes.com/library/national/science/082200sci-gm-rice.html. 

Zhu Y, Chen H, Fan J, Wang Y, Li Y, Chen J, Fan J, Yang S, Hu L, 

Leung H, Mew T, Teng P, Wang Z, Mundt C. 2000. Genetic 

diversity and disease control in rice. Nature 406:718-722. 

www.nature.com/cgi-taf/DynaPage.taffile=/nature/journal/ 

v406/n6797/full/406718a0_fs.html. 

Ziska L, Manalo P. 1996. Increasing night temperature can reduce 

seed set and potential yield of tropical rice. Austr. J. Plant 

Physiol. 23:791-794. 

Notes 

Authors’ address: International Rice Research Institute (IRRI), Los 

Baños, Laguna, Philippines. 

Keynotes 37

SESSION 1 

The genus Oryza, its diversity, 

and its evolution 

CONVENER: R. Okuno (NIAS) 

CO-CONVENER: R. Sackville Hamilton (IRRI)

Molecular phylogeny and divergence of the rice tribe 

Oryzeae, with special reference to the origin 

of the genus Oryza 

Song Ge, Ya-long Guo, and Qi-hui Zhu 

The rice tribe Oryzeae consists of 12 genera and more than 70 

species distributed in tropical and temperate regions worldwide 

(Tzvelev 1989, Vaughan 1994). Species in the genus 

Oryza and other genera closely related to Oryza have been 

extensively studied either because of their agronomically useful 

traits in rice genetic improvement (wild species in Oryza 

and Porteresia) or because of their economic value as part of 

cuisine (Zizania) and forage (Leersia) (Vaughan and Morishima 

2003). However, phylogenetic relationships among genera in 

this tribe have not been well studied, and the circumscription 

and taxonomic position of some genera have remained controversial 

for decades. In addition, the origin and diversification 

of this tribe, in particular the origin of the genus Oryza 

and its divergence, remain largely unclear. Recently, sequences 

of low-copy nuclear genes have been successfully used for 

addressing phylogenetic questions and, in combination with 

chloroplast DNA fragments, have provided especially powerful 

markers in terms of phylogenetic reconstruction and biogeographic 

inference (Sang 2002). 

The alcohol dehydrogenase (Adh) gene is the most widely 

used low-copy nuclear gene, whereas the chloroplast matK 

gene, trnL intron, and trnL-trnF spacer, as well as mitochondrial 

nad1 intron, are widely used in phylogenetic studies. 

Nuclear GPA1 that encodes a G protein α subunit is also a 

source of choice because it is a single copy in higher plants 

and well characterized in function and structure. To better understand 

the evolutionary and biogeographic history of the rice 

tribe, we sequenced two chloroplast (matK and trnL-trnF) fragments 

and one mitochondrial (nad1 intron) fragment and portions 

of two nuclear genes (Adh2 and GPA1) from 35 species 

representing 12 genera in the tribe, as well as Ehrharta and 

Phyllostachys (subfamily Bambusoideae) as the outgroups. In 

addition, to clarify the phylogenetic relationship among the A- 

genome species, which include cultivated rice (O. sativa), we 

have chosen to sequence introns of three anonymous nuclear 

single-copy genes (OsRFCD001283, OsRFCD017357, and 

OsRFCD009971 on chromosomes 1, 4, and 2, respectively) 

that evolve much faster than commonly used ITS and cpDNA 

fragments. Based on sequences of these introns, we reconstructed 

the phylogeny and dated the origin and divergence of 

the A-genome species. 

Molecular phylogeny based on multiple sequences 

from three genomes 

Phylogenetic analyses of the aligned sequences of the above 

fragments from three genomes were conducted using maxi- 

mum parsimony (MP), neighbor-joining (NJ), and Bayesian 

(BI) approaches. A comparison of the phylogenies inferred 

from two nuclear genes shows that they are largely congruent, 

with only one area of disagreement involving the placement of 

the monotypic genus Hygroryza (see below). The partitionhomogeneity 

test (PHT) indicated that two data sets (Adh2 

and GPA1) were statistically incongruent (P

2 

Oryza sativa 

O. glaberrima 

O. meridionalis 

O. punctata 

5 

O. officinalis 

O. rhizomatis 

O. australiensis 

8 

*Porteresia coarctata-7 

*P. coarctata-4 

O. brachyantha 

O. granulata 

14 

Leersia oryzoides-1 

L. perrieri-2 

L. hexandra-2 

L. perrieri-6 

L. hexandra-3 

L. oryzoides-4 

L. tisserantti 

20 

*Hygroryza aristata-Adh2 

*Prosphytochloa 

*Potamophila parviflora 

Chikusichloa aquatica 

*Rhynchoryza 

35 

–10 changes 

Zizania aquatica 

Z. latifolia 

Zizaniopsis villanensis 

Luziola leiocarpa 

L. fluitans 

*Hygroryza aristata-GPA1 

Ehrharta erecta 

Phyllostachys aurea 

Fig. 1. Strict consensus tree of the rice 

tribe Oryzeae generated from the combined 

Adh2 and GPA1 sequences. The 

thick lines indicate the nodes with the 

bootstrap support and Bayesian posterior 

probability over 95%, while the 

thin lines indicate the nodes with the 

bootstrap support and Bayesian posterior 

probability from 75% to 94%. The 

numbers on the node represent divergence 

times (approximate) in million 

years ago (MYA) as estimated by molecular 

clock approaches. Broken 

lines indicate the placement of 

Hygroryza on Adh2 and GPA1 trees, respectively. 

The numbers following the 

species name represent the clone sequenced. 

The asterisks indicate the 

monotypic genera. 

Session 1: The genus Oryza, its diversity, and its evolution 41

just “spurious” insufficient data or some other unknown artifact 

is unclear and needs to be further explored. 

Divergence of the rice tribe and the origin of Oryza 

Although the phylogeny and divergence time of Poaceae have 

been revealed by molecular approaches and estimated based 

on fossil evidence (GPWG 2001, Gaut 2002), little in the fossil 

record reflects upon the origin or diversification of the genus 

Oryza. A few studies have been made to estimate the divergence 

time of Oryza and its related genera (Second 1991, 

Vaughan and Morishima 2003). However, estimation of the 

possible dates of divergent events related to the evolution of 

Oryza was largely based on isozyme studies (Second 1991) 

and many contradictions exist regarding the divergence event 

proposed (see Vaughan and Morishima 2003). Despite several 

limitations to the use of clock-based sequence data (Gaut 2002), 

they are useful for estimating divergence time when the clock 

can be calibrated with some confidence, in particular when 

fossil evidence is not available or is inadequate. 

To test the molecular clock hypothesis in Oryzeae, we 

performed relative rate tests for all the data sets. Results show 

that no rate heterogeneity exists at synonymous sites for both 

Adh2 and matK data sets and so the molecular clock hypothesis 

cannot be rejected for the synonymous evolution at both 

loci. Using the synonymous molecular clocks of matK and Adh2 

and assuming that maize and rice diverged 50 million years 

ago (MYA) (Gaut 2002), we calculated approximate divergence 

times for various lineages within the tribe and within 

Oryza (Fig. 1). The estimates suggest that the rice tribe originated 

roughly 35 MYA, during approximately the transition 

from the Eocene to the Oligocene. Within the tribe, Oryza and 

Leersia separated from the rest of the tribe approximately 20 

MYA, and from each other 14 MYA. Within Oryza, the age of 

the deepest split between the most basal G genome and remaining 

genomes was estimated at about 8 MYA. 

Phylogeny, origin, and divergence 

of A-genome species in Oryza 

Evidence showed that the A-genome group that consists of eight 

diploid species is one of the most recently diverged lineages 

within the genus Oryza (Ge et al 2001). Their evolutionary 

relationships have long been controversial and are still not well 

studied at the sequence level. This is mainly because the commonly 

used sequences such as ITS and cpDNA fragments lack 

sufficient resolution in groups that have undergone rapid or 

recent radiations. Accordingly, we sequenced introns of three 

nuclear genes located on different chromosomes from multiple 

accessions of eight A-genome species. As expected, the 

intron sequences provide much higher (more than two times 

higher) informative characters than those of ITS (data not 

shown). Phylogenetic relationships and divergence times of 

the main lineages of the A-genome species were inferred based 

on these intron sequences using the similar approaches mentioned 

above. Figure 2 presents a phylogeny and divergence 

times among eight species, along with the B- (O. punctata) 

and E-genome (O. australiensis) species as the outgroups. All 

the species except O. rufipogon and O. nivara form a monophyletic 

group and the Australian endemic O. meridionalis is 

the earliest divergent lineage. However, accessions from O. 

rufipogon and O. nivara entirely mix together, supporting previous 

opinion that the two species should be treated as a single 

one. Moreover, two ecogeographic races of O. sativa (indica 

and japonica) form monophyletic groups along with some O. 

rufipogon and O. nivara accessions. 

Sequence divergence is low in all pair-wise comparisons 

among A-genome species, suggesting that the group radiated 

relatively recently. Using the similar molecular clock approaches 

mentioned above, we estimate that the A-genome 

species began to diverge roughly 2 MYA. The Asian cultivated 

rice (O. sativa) diverged from the African cultivated rice (O. 

glaberrima) at about 0.6 MYA, whereas two races of O. sativa 

(indica and japonica) separated approximately 0.4 MYA 

(Fig. 2). 

To put the divergence of the rice tribe and rice genus 

into a temporal framework raises an interesting question: How 

can the present distribution patterns in the rice tribe be reconciled 

with the apparent pantropical distribution for the tribe 

and many genera such as Oryza and Leersia In particular, 

many closely related species (e.g., the A-genome species) are 

currently geographically isolated from one another by thousands 

of kilometers of open ocean. This implies that oceanic 

dispersal would contribute to the evolution and divergence of 

the genus Oryza and its tribe. 

References 

Gaut BS. 2002. Evolutionary dynamics of grass genomes. New Phytol. 

154:15-28. 

Ge S, Li A, Lu BR, Zhang SZ, Hong DY. 2002. A phylogeny of the 

rice tribe Oryzeae (Poaceae) based on matK sequence data. 

Am. J. Bot. 89:1967-1972. 

Ge S, Sang T, Lu BR, Hong DY. 2001. Phylogeny of the genus Oryza 

as revealed by molecular approaches. In: Khush GS, Brar DS, 

Hardy B, editors. Rice genetics IV. Enfield, N.H. (USA): Science 

Publishers, Inc. and Los Baños (Philippines): International 

Rice Research Institute. p 89-105. 

GPWG (Grass Phylogeny Wording Group). 2001. Phylogeny and 

subfamilial classification of the grasses (Poaceae). Ann. Missouri 

Bot. Garden 88:373-457. 

Pyrah GL. 1969. Taxonomic and distributional studies in Leersia 

(Gramineae). Iowa State J. Sci. 44:215-270. 

Sang T. 2002. Utility of low-copy nuclear gene sequences in plant 

phylogenetics. Crit. Rev. Biochem. Mol. Biol. 37:121-147. 

Second G. 1991. Molecular markers in rice systematics and the evaluation 

of genetic resources. In: Bajaj YPS, editor. Rice biotechnology 

in agriculture and forestry. Vol. 14. Berlin (Germany): 

Springer-Verlag. p 468-494. 

Tzvelev NN. 1989. The system of grasses (Poaceae) and their evolution. 

Bot. Rev. 55:141-204. 


indica 

+ 

rufipogon 

+ 

nivara 

0.4 

0.6 

japonica 

+ 

rufipogon 

+ 

nivara 

glumaepatula 

glaberrima 

+ 

barthii 

2.0 

longistaminata 

meridionalis 

–5 changes 

punctata 

australiensis 

Fig. 2. Strict consensus tree of the A-genome species in Oryza generated from the 

combined intron sequences of three nuclear genes. The thick lines indicate the nodes 

with the bootstrap support and Bayesian posterior probability over 95%, while the thin 

lines indicate the nodes with the bootstrap support and Bayesian posterior probability 

from 75% to 94%. The numbers on the node represent divergence times in million 

years ago (MYA) as estimated by molecular clock approaches. Accessions were omitted 

from the figures. 


Vaughan DA, Morishima H. 2003. Biosystematics of the genus Oryza. 

In: Smith W, editor. Rice: origin, history, technology, and production. 

New York, N.Y. (USA): John Wiley & Sons, Inc. p 

27-65. 

Vaughan DA. 1994. The wild relatives of rice: a genetic resources 

handbook. Manila (Philippines): International Rice Research 


Eco-genetic diversification in the genus Oryza: 

implications for sustainable rice production 

Duncan Vaughan, Koh-ichi Kadowaki, Akito Kaga, and Norihiko Tomooka 

Notes 

Authors’ address: Laboratory of Systematic and Evolutionary Botany, 

Institute of Botany, Chinese Academy of Sciences, Beijing 

100093, China, e-mail: gesong@ibcas.ac.cn. 

Acknowledgments: We thank Tao Sang of Michigan State University 

(East Lansing, Mich., USA) and Bao-rong Lu of Fudan University 

(Shanghai, China) for their encouragement and assistance. 

We are also grateful to the International Rice Genebank 

at IRRI (Los Baños, Philippines) for providing seed samples. 

This study was supported by the Chinese Academy of Sciences 

(Kscxz-sw-101A), the National Natural Science Foundation 

of China (30025005), and the Program for Key International 

S & T Cooperation Project of China (2001CB711103). 

Rice is a component of some remarkably productive and sustainable 

farming systems. However, many factors threaten sustainable 

rice farming, such as economic development leading 

to labor, land, and freshwater constraints to rice-farming systems. 

In addition, abiotic stresses, including the consequences 

of climate change, and ever-changing biotic stresses require 

diverse approaches to maintain and improve rice yields. In this 

paper, the contributions that diversity in the gene pools of rice 

can make to rice improvement are considered. 

Primary gene pool 

Since Oryza species having the AA genome can hybridize naturally 

with one another, they can all be considered part of the 

primary gene pool of rice. AA-genome Oryza species have 

astonishing eco-genetic diversification in relation to environments 

with different hydrological regimes. During domestication, 

particular ecotypes of wild rice were selected for increased 

yield. Subsequently, because of selection and accompanying 

human migrations, diversification of the cultivated rice gene 

pool extended the climatic and geographic range of rice to 

areas where its wild relatives do not grow (Vaughan et al 

2004a,b). The wild relatives of rice in this gene pool have been 

the main source of useful genes for rice improvement from 

wild Oryza. Recent examples of the use of the wild species in 

this gene pool are given below. 

Abiotic stress resistance in O. rufipogon 

from the Mekong Delta 

In December 1990, a collaborative collecting mission for the 

wild relatives of rice was undertaken by the Vietnam Agricultural 

Sciences Institute and the International Rice Research 

Institute (IRRI) in the Mekong Delta. During that mission, areas 

of adverse soils where the wild rice O. rufipogon Griff. 

grows were targeted for collection. Many samples of O. 

rufipogon were collected from different parts of the Mekong 

Delta, including Phup Hiep and Thap Muoi districts, where 

the soils were bright red and known to be highly acid sulfate. 

Passport data of the collections reported this. Consequently, 

germplasm from these locations could be rapidly evaluated in 

standardized conditions to determine the type and degree of 

abiotic stress resistance it had. Three accessions (106412, 

106423, and 106424) 1 were found to be useful sources of resistance 

to aluminum toxicity. One accession (106424) was 

crossed to rice (cv. IR64) and recombinant inbred lines developed 

from this cross were analyzed. Using root length to evaluate 

the response to stress, five QTLs from O. rufipogon were 

found that explained aluminum tolerance and these were 

mapped to rice chromosomes 1, 3, 7, 8, and 9 (Nyugen et al 

2003). Since this accession (106424) has useful traits and is 

being used in rice improvement programs, it is one of the accessions 

chosen for analysis by the Oryza Map Alignment 

Project (www.genome.arizona.edu/BAC_special_projects/). 

Biotic stress resistance in O. rufipogon 

from the Central Plains of Thailand 

Tungro is the most serious virus disease of rice in South and 

Southeast Asia (Azzam and Chancellor 2002). Sources of resistance 

to tungro viruses are found in cultivated rice but some 

of these have broken down as new forms of these viruses have 

emerged. Generally, field collection notes cannot help guide 

1 Accession numbers mentioned are materials in the International Rice 

Genebank at IRRI and refer to germplasm collected directly by the first author 

with collaborating colleagues in national programs while a staff member 

at IRRI. 


evaluators in relation to biotic stress tolerance because biotic 

stresses are transient and do not affect germplasm in the field 

uniformly. Therefore, a logical approach to evaluation is the 

core collection approach. To find new sources of tungro resistance 

for use in rice breeding, a wild Oryza core collection 

was evaluated (Kobayashi et al 1997, Vaughan and Jackson 

1995). A few accessions of O. rufipogon from Thailand 

(105908, 105909, and 105910) collected by boat in a deeply 

flooded area 68 km north of Bangkok were tolerant of both 

the spherical and bacilliform virus types that constitute tungro 

(Kobayashi et al 1997). These O. rufipogon accessions tolerant 

of tungro were used to transfer tungro tolerance to advanced 

breeding lines (Brar and Khush 2003) and one variety, called 

Matatag 9, was released in the Philippines with tungro tolerance 

from accession 105908. 

These two examples of germplasm collected, evaluated, 

and used in rice improvement and basic research reinforce the 

importance of the wild relatives of rice for rice improvement. 

However, these examples also illustrate the potential problem 

associated with the loss of wild Oryza genetic resources in 

situ. The Mekong Delta and Central Plains of Thailand are 

developing rapidly and the extinction of populations of wild 

rice has been recorded in various parts of Southeast Asia 

(Akimoto et al 1999, Kuroda et al 2004). Therefore, the chance 

for evolutionary change in natural wild rice populations to 

enable adaptation to changing biotic and abiotic stresses, such 

as climate change, is diminished. 

Secondary and tertiary gene pools of rice 

It is debatable whether rice has a secondary gene pool since 

making crosses between rice and species with other genomes 

is very difficult and leads to rapid chromosome elimination in 

subsequent generations. However, in Thailand and Costa Rica, 

hybrid populations between AA-genome Oryza species and 

species having other genomes have been reported and these 

populations are reported to be vigorous but sterile. The secondary/tertiary 

gene pools of rice can be considered to be composed 

of Oryza species with genomes other than the AA genome. 

Eco-genetic diversification in non-AA-genome Oryza 

species reveals different types of environmental adaptation not 

found among AA-genome species. Shade tolerance is a feature 

of species with the CC, BBCC, GG, and HHJJ genomes. 

Among the three diploid CC-genome species (O. eichingeri 

Peter, O. officinalis Wall ex Watt, and O. rhizomatis Vaughan), 

there is diversification in relation to the degree of sun their 

habitats receive. O. eichingeri is a forest-dwelling species. O. 

rhizomatis is adapted to open hot and dry areas in Sri Lanka. 

O. officinalis has wide ecological adaptation but is usually 

found in open habitats and areas with higher precipitation than 

O. rhizomatis. GG diploid species (O. granulata complex) tend 

to be confined to upland forested habitats with pronounced 

wet and dry seasons. In contrast, the HHJJ-genome species 

(O. ridleyi complex) tend to be abundant in wet tropical lowland 

forests (Vaughan 2003). There has yet to be a comprehensive 

analysis of the photosynthetic and drought-tolerance 

systems of Oryza species not having the AA genome. Given 

the diverse eco-edaphic habitats of Oryza species, they likely 

contain much useful variation in a range of ecological traits. 

Beyond Oryza in the tribe Oryzeae 

The genus Oryza belongs to a small group of genera in the 

tribe Oryzeae in the subfamily Bambusoideae of the Poaceae 

(Graminae). The related genera have several characteristics 

that are not found in the genus Oryza. Among these, two are 

worthy of special mention because of their potential use in 

rice improvement. 

Separation of male and female florets: Zizania 

Several genera in the tribe Oryzeae have unisexual spikelets— 

Luziola, Zizania, and Zizaniopsis. In the case of Luziola, staminate 

and pistillate spikelets are usually on separate panicles. 

In Zizania and Zizaniopsis, they are both on the same panicle. 

Of these three genera, Zizania has been the best studied because 

of its economic importance in Asia (where Zizania 

latifolia [Griseb.] Turcz. ex Stapf is a vegetable) and in North 

America (where Zizania palustris Linn., “wild rice,” is a gourmet 

food). The genome map of Zizania palustris has been compared 

with that of rice (Kennard et al 1999, Hass et al 2003). 

This has revealed that the 15 haploid chromosomes of Zizania 

represent the 12 haploid chromosomes of Oryza, with duplication 

of chromosomes 1, 4, and 9. However, these studies 

have not been revealing regarding the genetic control of pistillate 

and staminate spikelets. Using protoplast fusion, Chinese 

researchers have succeeded in producing two somatic hybrids 

between rice and Z. latifolia (Liu et al 1999). These intergeneric 

hybrids had Z. latifolia fragments in the rice genome. 

Improvement in protoplast culture and fusion conditions may 

enable abundant and potentially useful somatic hybrids to be 

developed. 

Adaptation to saline conditions: Porteresia 

In the saline tracts at the mouth of the Ganges Delta, wild 

grasses are sometimes harvested for food. Myriostachya 

wightiana Hook., locally called Dhanshi, and Porteresia 

coarctata (Roxb.) Tateoka, locally called Uri-dhan, are two of 

these salt-tolerant grasses that are gathered. Porteresia 

coarctata is in the tribe Oryzeae and, because it is the most 

salt-tolerant relative of rice, it has been subject to much research 

attention. Porteresia is a monospecific genus that exhibits 

several unique morphological and anatomical features 

of the embryo and leaf (Balakrishna 2003). Porteresia 

coarctata has salt hairs on the leaves through which salt is 

secreted (Flowers et al 1990). Success has been reported in 

crossing Porteresia with rice (Jena 1994, Brar et al 1997) but 

transfer of the complex structural mechanism by which 

Porteresia has adapted to living in saline conditions has yet to 

be achieved. 


Conclusions 

The primary gene pool of rice, AA-genome Oryza species, offers 

gene sources from regions to which rice is adapted. The 

secondary and tertiary gene pools of rice, consisting of the 

other Oryza species, provide gene sources that include adaptation 

to areas to which rice is not adapted, such as forest shade. 

Beyond the genus Oryza in the tribe Oryzeae, gene sources for 

major morphological and structural change such as panicle 

structure or leaf anatomy can be found. Global and regional 

environmental changes suggest that, in the future, germplasm 

naturally exposed and evolving in response to these changes 

may be an essential source of genes to address these changes. 

Hence, there is a need for innovative approaches to in situ 

conservation of rice genetic resources to complement ex situ 

conservation efforts. 

References 

Akimoto M, Shimamoto Y, Morishima H. 1999. The extinction of 

genetic resources of Asian wild rice, Oryza rufipogon Griff.: 

a case study in Thailand. Genet. Resour. Crop Evol. 46:419- 

425. 

Azzam O, Chancellor TCB. 2001. The biology, epidemiology, and 

management of rice tungro disease in Asia. Plant Dis. 86:88- 

100. 

Balakrishna P. 2003. Porteresia coarctata: scope and limitations of 

utilization in rice improvement. In: Nanda JS, Sharma SD, 

editors. Monograph on the genus Oryza. Enfield, N.H. (USA): 

Science Publishers, Inc. p 359-373. 

Brar DS, Elloran R, Talag JD, Abbasi F, Khush GS. 1997. Cytogenetic 

and molecular characterization of an intergenic hybrid 

between Oryza sativa L. and Porteresia coarctata (Roxb.) 

Tateoka. Rice Genet. Newsl. 14:43-44. 

Brar DS, Khush GS. 2003. Utilization of wild species of genus Oryza 

in rice improvement. In: Nanda JS, Sharma SD, editors. Monograph 

on the genus Oryza. Enfield, N.H. (USA): Science Publishers, 

Inc. p 283-309. 

Flowers TJ, Flowers SA, Hajibagheri MA. 1990. Salt tolerance in 

the halophytic wild rice, Porteresia coarctata Tateoka. New 

Phytol. 114:675-684. 

Hass BL, Pires JC, Porter R, Phillips RL, Jackson SA. 2003. Comparative 

genetics at the gene and chromosome levels between 

rice (Oryza sativa) and wild rice (Zizania palustris). Theor. 

Appl. Genet. 107:773-782. 

Jena KK. 1994. Develoment of intergeneric hybrid between O. sativa 

and Porteresia coarctata. Rice Genet. Newsl. 11:78-79. 

Kennard W, Phillips R, Porter R, Grombacher A. 1999. A comparative 

map of wild rice (Zizania palustris L. 2n=2x=30). Theor. 

Appl. Genet. 99:793-799. 

Kobayashi N, Ikeda R, Vaughan DA, Imbe T. 1997. New gene sources 

for tungro resistance in wild species of rice (Oryza spp.). In: 

Chancellor TCB, Thresh JM, editors. Epidemiology and management 

of rice tungro disease. Chatham (UK): Natural Resources 

Institute. p 60-68. 

Kuroda Y, Appa Rao S, Bounphanousay C, Kongphanh K, Iwata A, 

Tanaka K, Sato YI. 2004. Diversity of wild and weedy rice in 

the Lao PDR. In: Schiller JM, Chanphengxay M, Linquist B, 

Appa Rao S, editors. Rice in Laos. (In press.) 

Liu B, Liu ZL, Li XW. 1999. Production of a highly asymmetric 

somatic hybrid between rice and Zizania latifolia (Griseb.): 

evidence for inter-genomic exchange. Theor. Appl. Genet. 

98:1099-1103. 

Nguyen BD, Brar DS, Bui BC, Nguyen TV, Pham LN, Nguyen HT. 

2003. Identification and mapping of the QTL for aluminium 

tolerance introgressed from the new source, Oryza rufipogon 

Griff., into indica rice (Oryza sativa L.). Theor. Appl. Genet. 

106:583-593. 

Vaughan DA. 2003. Genepools in the genus Oryza. In: Nanda JS, 

Sharma SD, editors. Monograph on the genus Oryza. Enfield, 

N.H. (USA): Science Publishers, Inc. p 113-138. 

Vaughan DA, Jackson MT. 1995. Core collections as a guide to the 

whole collection. In: Hodgkin T, Brown AHD, van Hintum 

ThJL, Morales EAV, editors. Core collections of plant genetic 

resources. Chichester (UK): Wiley and Sons. p 229-239. 

Vaughan DA, Miyzaki S, Miyashita K. 2004a. The rice genepool 

and human migrations. In: Werner D, editor. Biological resources 

and migration. Berlin (Germany): Springer. p 1-13. 

Vaughan DA, Sanchez, Ushiki J, Kaga A, Tomooka N. 2004b. Asian 

rice and weedy rice: evolutionary perspectives. In: Gressel J, 

editor. Crop ferality and volunteerism. Boca Raton, Fla. 

(USA): CRC Press. (In press.) 

Notes 

Authors’ address: National Institute of Agrobiological Sciences, 

Tsukuba, Japan, e-mail: duncan@affrc.go.jp. 


Toward a global strategy for the conservation 

of rice genetic resources 

Ruaraidh Sackville Hamilton and Ruth Raymond 

Twenty-two species of wild rice of the genus Oryza occur as 

natives in 81 countries, in all tropical and subtropical regions 

of the world. Rice is believed to have been first cultivated 

around 6,000 to 14,000 years ago, with probably at least two 

independent domestication events of O. sativa (with separate 

events for indica and japonica rice) in Asia and a third of O. 

glaberrima in Africa. Rice is now cultivated in 114 countries, 

also in all tropical and subtropical regions, with distinctive 

forms occurring in different parts of Asia, Africa, Europe, and 

Oceania. 

The genetic diversity encompassed by the genus is remarkable. 

More than 400,000 accessions are conserved ex situ 

in genebanks. The task of efficiently conserving and using all 

this diversity is formidable, and requires a globally integrated, 

multinational effort. Yet, rice genebanks are working largely 

independently and ignorant of the achievements of sister 

genebanks. The resulting scientific and economic inefficiency, 

with uncontrolled duplication and ineffective targeting of conservation 

and use, is clear. 

In this paper, we present the first steps toward a plan to 

rectify the situation, through development of a global strategy 

for the efficient conservation and use of rice diversity. We begin 

by describing the new Global Crop Diversity Trust, through 

which it has become possible to consider an efficient global 

strategy. We then consider the required elements of the strategy, 

and the process toward implementation of those steps. 

The Global Crop Diversity Trust 

The Global Crop Diversity Trust (www.startwithaseed.org) has 

been set up to establish and support a rational, efficient, and 

sustainable global system for the conservation of crop diversity. 

In the short term, it will assist in the development of rational 

and efficient systems, by providing both expert technical 

support and financial support. It will fund the upgrading of 

selected national and regional genebanks and the establishment 

of infrastructure for coordination. In the longer term, it 

seeks to provide a permanent source of funding for the maintenance 

of crop diversity collections around the world. 

It is developing the global system through two complementary 

strategies. It is developing a series of regional strategies, 

based on identification of regional priorities building on 

the economic efficiencies of multicrop genebanks. At the same 

time, it is developing a series of global crop strategies, which 

will be implemented within the context of the regional strategies 

and which will seek the genetic efficiencies of global coordination. 

In International Year of Rice 2004, we are celebrating 

the world’s most important crop for food security. It is highly 

appropriate to start developing a global rice conservation strategy 

as one of the early initiatives of the Global Crop Diversity 

Trust. 

The history and diversity of rice 

Cultivated rice has at least three centers of origin (Fig. 1). Oryza 

sativa spread from its origins in South and East Asia to the 

south, north, east, and west, eventually to cover all continents 

except Antarctica. The precise time line of the migration of 

rice remains obscure in many countries. 

Primary centers of diversity are clearly associated with 

the centers of origin. An additional primary center of diversity 

in the mainland and islands off Southeast Asia appears to be 

associated with the intercrossing of the indica and japonica 

forms of O. sativa, encompassing a wide range of intermediate 

forms and tropical japonica. 

On its westward migration, O. sativa reached Europe 

more than 2,000 years ago, and secondary centers of diversity 

can be recognized with distinctive forms in the Indian Ocean, 

western Asia, and Europe. More recent distinctive secondary 

centers of diversity are also apparent in Africa and South 

America. 

For wild rice, the centers of species diversity and genomic 

diversity are the islands from Southeast Asia into the 

Pacific Ocean. Nine of the 22 wild species occur in Indonesia, 

and 7 of the 10 genome types are found in the Asian-Pacific 

islands. In addition, distinctive sets of species assemblages are 

associated with South Asia, Africa, and the Americas. 

Given so much genomic, species, and genetic diversity 

within and between regions, a comprehensive strategy for conserving 

the diversity of rice must be truly global in extent. 

Ex situ conservation of rice 

According to the FAO report on the State of the World’s Plant 

Genetic Resources, more than 400,000 accessions of rice are 

conserved in genebanks around the world. However, the majority 

of the accessions are in only a small number of genebanks 

(Table 1). Over 25% are held in IRRI and more than 60% of 

the remaining accessions are held in just five national Asian 

genebanks in China, India, Japan, Thailand, and the Republic 

of Korea. 

This conservation effort is far from coordinated or global. 

The extent of duplication across the major genebanks is 

unknown. There are strong indications of gaps in coverage 


Fig. 1. History of the spread of rice agriculture. Each rice-producing country is shaded according 

to the date of the first cultivation of rice in the country. Key: dark green > 5,000 years ago 

→ yellow 2,000–3,000 years ago → dark red 100–200 years ago; hatched = rice is cultivated, 

unclear date of first cultivation; white = rice is not cultivated. The three centers of origin of 

cultivated rice are highlighted. 

Table 1. Summary of approximate numbers of accessions 

of rice held in major rice genebanks around the world. 

Genebank a 

Approximate number of accessions 

IRRI 107,000 

China-CAAS 64,000 

India-NBPGR 54,000 

Japan-NIAR 36,000 

Thailand-RRI 24,000 

Korea-RDA 23,000 

WARDA 20,000 

USA-NSGC 17,000 

Brazil-CENARGEN 14,000 

Laos-NARC 12,000 

IITA 12,000 

Russia-VIR 5,500 

Vietnam-PGRC 4,800 

Pakistan-PGRI 2,400 

Australia 1,400 

Indonesia-RIR 1,000 

Total 398,100 

a IRRI = International Rice Research Institute, CAAS = Chinese Academy 

of Agricultural Sciences, NBPGR = National Bureau of Plant Genetic 

Resources, NIAR = National Institute of Agrobiological Resources, 

RRI = Rice Research Institute, RDA = Rural Development Administration, 

WARDA = Africa Rice Center, NSGC = National Small Grains 

Center, CENARGEN = Embrapa Recursos Genéticos e Biotecnologia, 

EMBRAPA = Empresa Brasileira de Pesquisa Agropecuária, NARC = 

National Agricultural Research Center, IITA = International Institute for 

Tropical Agriculture, VIR = Vavilov Research Institute of Plant Industry, 

PGRC = Plant Genetic Resources Center, PGRI = PARC Plant Genetic 

Resources Institute, PARC = Pakistan Agricultural Research Council, 

RIR = Research Institute for Rice. 

outside South and East Asia, especially in wild species but 

also in cultivated rice. The absence of coordinated information 

makes it impossible to assess the extent of duplication or 

gaps, or to fully and rationally analyze the distribution of diversity 

or target the worst gaps. 

The global strategy outlined 

A major element of the rice conservation strategy must be a 

global network, through which to coordinate conservation activities 

worldwide. Yet, most existing rice networks are for rice 

breeding and research (e.g., INGER, the International Network 

for the Genetic Evaluation of Rice), not for genetic resources. 

Through the GCDT, it is planned to develop such a network. 

Yet, any such network would be ineffective without an effective 

information resource on which to base decisions and strategies. 

Indeed, the lack of coordinated information is the major 

hindrance to the development of an efficient and rational global 

strategy for the conservation of rice genetic resources. 

Therefore, the heart of the network must contain a global information 

system documenting, as far as possible, the relevant 

data of all accessions conserved in all participating genebanks. 

Next steps 

Development of a global information resource will be the first 

objective of the global rice conservation strategy. 

There is a risk, however, that we are planning to avoid, 

associated with assembling these data into a common system. 

Early attempts to integrate data from different genebanks were 

based on standardizing subsets of diverse data from different 

genebanks into a common format and collating them into a 

central database. Considerable effort was required to populate 

the central database with data. Therefore, data in the central 


database were updated only infrequently, if at all, so that the 

central database quickly became out-of-date. A rational, integrated 

decision-making process cannot be efficient or effective 

if it is based on outdated data. 

For the global rice database, an entirely different approach 

is being developed. We are working on developing a 

central search engine that will redirect incoming queries directly 

to the databases of participating genebanks. The only 

information kept in the central search engine will be metadata 

describing the data held by participating genebanks; these 

metadata are needed to redirect incoming queries correctly. 

Security mechanisms will keep the genebanks’ live databases 

safe behind firewalls and will ensure that only nonconfidential 

data are accessed. For each participating genebank, a map will 

be created that describes the relationship between the 

genebank’s own internal data structures and the common international 

standard. This map will be used to create a dynamic 

view of the data that has the appearance of the international 

standard but without changing the internal data structures 

or values. After creating the initial map and view, the 

global information resources will be kept automatically updated, 

with no additional investment by the contributing 

genebank. 

The outlook 

Once the test system is in place, we will be seeking additional 

partners to add new database nodes to the information network. 

The resulting global information resources will guide 

the establishment of responsibilities for conservation, identification 

of duplicates, and gaps in the global collection, and 

the rationalization of procedures for sharing germplasm and 

making it available to farmers and other users. The objectives 

of the International Treaty on Plant Genetic Resources for Food 

and Agriculture will be facilitated for rice in a way that will 

benefit us all. 

Notes 

Genetic architecture and complexity 

in wild and cultivated rice 

Y. Sano 

Authors’ addresses: Ruaraidh Sackville Hamilton, International Rice 

Research Institute, DAPO Box 7777, Metro Manila, Philippines; 

Ruth Raymond, International Plant Genetic Resources 

Institute, Via dei Tre Denari, 472/a, 00057, Maccarese, Rome, 

Italy. 

A major challenge to modern geneticists and breeders is to 

understand the nature of quantitative trait loci (QTL) and environmental 

factors causing variation in quantitative traits, since 

adaptive traits are not determined by a single mutation under 

changing environments. Genetic complexity results from the 

combined effects of multiple genes whose expression is greatly 

influenced by the environment. We present here our recent studies 

on the significance of genetic diversity in wild and cultivated 

forms of rice. 

Hidden variation in the primary gene pool 

The offspring of two plants can contain novel combinations of 

alleles that produce more extreme phenotypes than those of 

either parent. This phenomenon (transgressive segregation) has 

been shown to be common (Rieseberg et al 2003). This indicates 

the importance of gene flow for the emergence of novelty 

in crops under man-made habitats, possibly showing a rapid 

change in their genetic architecture responding to the environment. 

Recently, the wild ancestor of maize (teosinte) was shown 

to have a high potential to produce phenotypic diversity although 

epistasis had tended to conceal the expression of genetic 

differences in natural populations (Lauter and Doebley 

2002), suggesting that phenotypic diversity has not been sufficiently 

exposed to date. 

Generally, isolating barriers appear between species but 

not within species, raising the question of how such a reproductive 

barrier could develop between species. We detected 

cross-incompatibility in advanced generations of backcrossing 

between wild and cultivated rice strains although no crossing 

barrier had been reported within the primary gene pool 

(Matsubara et al 2003). Genetic analysis revealed that an alien 

segment of chromosome 6 from a wild strain caused a failure 

of early endosperm development and that the cross-incompatibility 

resulted from the sexual affinity between the cross-incompatibility 

reactions in the female and male. Further, the 

postfertilization barrier was controlled by at least three linked 

genes on chromosome 6, showing a different distribution of 

interacting alleles among taxa (Fig. 1). The results suggest that 

differences leading to species-specific barriers might be partly 

preexisting within the gene pool even though they are not observed 

by conventional methods. 

A profound isolating barrier tends to be species-specific. 

We analyzed a hybrid sterility gene, S1, on chromosome 6 between 

Asian and African cultivated rice species. The hybrid 

sterility gene has the strongest effect among plant hybrids. The 

S1 gene is called a “gamete eliminator” since it induces abor- 


Short arm of chromosome 6 

Centromere 

wx 

RM204 

RZ516 RZ398 

S1520 RZ588 OsC1 G200 

RG264 

Hd1 

C235 

R111 R32 

G2028 

6.8 

2.6 

2.0 

7.5 

3.5 

5.3 

7.6 

11.9 

2.1 

7.3 

3.0 

5.6 

cM 

Cif 

cim 

Su-Cif 

Female gamete 

Male gamete 

Gene Reaction Reaction Gene 

Cif 

Cif 

+ 

Cim 

Su-Cif 

cif 

+ 

cim 

cim 

Incompatible cross 

Compatible cross 

Fig. 1. (A) Three linked genes (Cif, cim, and Su-Cif) responsible for the unidirectional cross-incompatibility 

detected between wild and cultivated rice strains. (B) The genetic model proposed by 

these genes, all of which act sporophytically. Cross-incompatibility occurs when gametes with the 

cross-incompatibility reactions in the female (Cif) and the male (cim) are fertilized. Cim and cim are 

predominantly distributed in indica and japonica types, respectively. + indicates the cross-compatibility 

reaction. The locations of the three genes are shown by thick solid lines. 

tion of both the male and female gametes possessing its allelic 

alternative S1-a only in the heterozygote. Surveys in their distribution 

showed that the S1 and S1-a alleles were specific to 

African and Asian rice species, respectively, although the sensitivity 

to distortion in the female differed between indica and 

japonica types. In addition, fine mapping revealed that only 

the female gametes became viable when the flanking region of 

the S1 gene was deleted, showing a preferential transmission 

only in pollen grains. Such a meiotic drive (MD) system violating 

Mendel’s rules has been detected in natural populations 

of a variety of organisms. The results in rice species suggested 

that a gamete eliminator could be derived from the male MD 

by accumulating responder-like elements affecting the female 

gamete and that this system could be involved in reproductive 

barriers. 

Polygenic balance theory 

Phenotypic variation is central to evolutionary adaptation underlying 

natural and artificial selection. Polygenes, which are 

very sensitive to residual variation and the environment, might 

be partly built up on a chromosome by selection, and repulsion 

or coupling linkages along a chromosome have been demonstrated 

in Drosophila (Mather and Jinks 1982). Stabilizing 

or normalizing selection could encourage the development of 

hidden variation within and between populations. Such a linked 

group of polygenes could be reshuffled through recombination, 

resulting in the formation of newly recombined phenotypes 

or transgression. 

Grain characteristics were examined between the two 

subspecific rice strains regarding QTLs on chromosome 6. 

Grain characteristics are known to be quantitative traits and to 

be diagnostic between them. To enhance the power to detect 

QTLs with small effects, recombinant inbred lines (RILs) were 

made after introgression of chromosome 6, so the segmented 

fragments could be precisely compared, minimizing the residual 

variation. The resulting RILs showed a distinct transgressive 

segregation in the seed dimension. Multiple linked 

QTLs were detected on the segment, showing positive and 

negative values in additive effects as well as epistatic interactions 

between the detected QTLs, although such QTLs on chromosome 

6 were not reported before. The results confirmed 


that the transgressive segregation is driven from breakdown 

of linkage and that the detection of QTLs highly depends on 

the genetic effects of the flanking QTLs, indicating a need for 

caution in interpreting data about the detected QTLs. 

Implications from genealogy 

Analysis of geographic patterns of variation in crops is essential 

to understand their evolving populations and to make efficient 

use of the available variability in plant breeding, even 

though traditional landraces are rapidly replaced by modern 

high-yielding varieties worldwide. The nature of crop gene 

pools suggests that domestication has been achieved largely at 

the intraspecific level although gene flow enables cultivated 

forms to keep pace with abiotic and biotic factors. Genetic 

diversity could result from the diversification of alleles and 

interacting genes (including epistasis) and/or from the diversification 

of structural and regulatory genes. Recent work on 

DNA sequences shows that genetic diversity is affected both 

by current patterns of microevolutionary forces, such as gene 

flow and selection, and by phylogenetic history (Schaal et al 

2003). Alleles at different loci even on different chromosomes 

were not randomly associated with each other, possibly attributable 

to selection, drift, and nonrandom mating (Morishima 

et al 1992). Genealogy provides insight into these evolutionary 

processes. 

Divergent phenotypes are often detected in domesticated 

plants. An example in rice is apiculus coloration associated 

with anthocyanin pigmentation. It has been reported in cultivars 

from the northern area of Japan that varying degrees of 

coloration are caused by a series of alleles at the C locus 

(Takahashi 1982). Based on the synteny map between rice and 

maize, the C gene was expected to be the rice homologue 

(OsC1) of maize C1, which belongs to the group of R2R3- 

Myb regulatory factors. Two different types of deletions causing 

a frameshift were detected in the 3rd exon, and both of the 

deleted nucleotides corresponded to the positions of putative 

base-contacting residues, suggesting that the indica and 

japonica types carry mutations with independent origins (Saitoh 

et al 2004). In addition, replacement substitutions were frequently 

detected in the OsC1 gene of strains carrying the previously 

defined C alleles. Molecular evolutionary analysis revealed 

that haplotype diversity was higher in wild forms than 

in cultivated forms; however, cultivated forms had their specific 

haplotypes showing few shared haplotypes with their wild 

form (Fig. 2). The genealogy of the OsC1 gene suggested that 

allelic diversification causing phenotypic changes might have 

resulted from mutations in the coding region of landraces rather 

than incorporation from others through hybridization. The neutrality 

test also revealed that changes in amino acids might be 

associated with selective forces acting on the lineage representing 

most Asian cultivated forms. This suggests that 

landraces might maintain more agronomically valuable genes 

accumulated since the origin of rice than previously believed, 

even if a reduction in nucleotide diversity was caused by the 

population bottleneck due to domestication. Thus, the allelic 

diversity of the OsC1 gene shows the importance of farmers’ 

efforts; however, it remains to be studied why haplotype diversity 

is high in the northern areas of rice cultivation. 

It is well known that the semidwarf gene from variety 

DGWG greatly contributed to the Green Revolution in rice. 

Recent molecular work revealed that the mutation is due to a 

deletion at the OsGA20ox2 gene (Ashikari et al 2002), which 

made it possible to survey its geographical distribution. It has 

been believed that the semidwarf gene arose and was maintained 

in a local variety until the development of high-yielding 

varieties (Jennings 1966). Recently, we found that wild strains 

and landraces from China carry the same allele as that of 

DGWG, although it was not distributed in tropical areas, except 

for modern high-yielding varieties. The phenotypic expression 

of the semidwarf gene was suppressed in the wild 

strain, showing that the presence of interacting genes enabled 

its wild carriers to maintain it within the population, although 

slightly deleterious mutations would be rapidly selected against 

because of competition in rice fields. This leads us to consider 

that sd1 preexisted in wild forms and was incorporated into 

landraces through gene flow. Therefore, farmers’ efforts seem 

to have greatly contributed to mine the semidwarf gene, excluding 

interacting genes under less-competitive fields apart 

from tropical areas. These findings confirm that genetic resources 

of wild forms and landraces have been equally important, 

giving us suggestions for improving rice in the future. 

References 

Ashikari MA, Sasaki A, Ueguchi-Tanaka M, Itoh H, Nishimura A, 

Datta S, Ishiyama K, Saito T, Kobayashi M, Khush GS, Kitano 

H, Matsuoka M. 2002. Mutation in a gibberellin biosynthetic 

gene, GA20 oxidase, contributed to the rice ‘Green Revolution’. 

Breed. Sci. 52:143-150. 

Jennings PR. 1966. The evolution of plant type in Oryza sativa. 

Econ. Bot. 20:396-402. 

Lauter N, Doebley J. 2002. Genetic variation for phenotypically invariant 

traits detected in teosinte: implications for the evolution 

of novel forms. Genetics 160:333-342. 

Mather K, Jinks JL. 1982. Biometrical genetics. 3rd edition. New 

York, N.Y. (USA): Chapman and Hall. 396 p. 

Matsubara K, Khin-Thidar, Sano Y. 2003. A gene block causing crossincompatibility 

hidden in wild and cultivated rice. Genetics 

165:343-352. 

Morishima H, Sano Y, Oka HI. 1992. Evolutionary studies in cultivated 

rice and its wild relatives. Oxford Surv. Evol. Biol. 

8:135-184. 

Rieseberg LH, Widmer A, Arntz AM, Burke JM. 2003. The genetic 

architecture necessary for transgressive segregation is common 

in both natural and domesticated populations. Philos. 

Trans. R. Soc. Lond. B Biol. Sci. 358:1141-1147. 

Saitoh K, Ohnishi K, Mikami I, Khin-Tidar, Sano Y. 2004. Allelic 

diversification at the C (OsC1) locus of wild and cultivated 

rice: nucleotide changes associated with phenotypes. Genetics 

168:(in press). 

Schaal BA, Gaskin JF, Caicedo AL. 2003. Phylogeography, haplotype 

trees, and invasive plant species. J. Hered. 94:197-204. 

Takahashi M-E. 1982. Gene analysis and its related problems. J. 

Fac. Agric. Hokkaido Univ. 61:91-142. 


Haplotype Strain Group 

SA1 

Jp (T65, A58) 

SA2 

Jp (A18, A56) 

SA3 

Jp (A38, A83, A136, 544) 

Jv (226, 647) In (133) 

SA4 

Jp (A5, A55, A108) 

87 

SA5 

Jp (734) 

60 

SA6 

In (108, IR36, 868, 414, 

719, Acc6622, Acc35618) 

A 

63 SA8 

In (I47) 

65 SA9 

In (I32, I45) Ra (W2012) 

RU1 

Rp (W1944) 

80 

77 

75 

78 

85 

56 

SA7 

RU4 

RU3 

RU5 

RU2 

Rp (W1943) 

In (160, 706, N303, C6172) 

RP (W120) 

Ra (W107) Rin (W1819) 

Ra (W1865) 

B 

95 

RU6 

96 64 

66 

RU7 

RU8 

Rp (W1294) 

Rp (W593) 

Ra (W2002) 

C 

AF 

GLU 

W025, W1468, W1647 

W1187 

Outgroup 

0.001 

Substitution/site 

Fig. 2. Neighbor-joining (NJ) reconstruction of the genealogical relationships among OsC1 haplotypes 

of 43 strains in wild and cultivated rice strains. Oryza glaberrima, O. barthii (AF), and O. glumaepatula 

(GLU) were used as outgroups. The maximum-parsimony (MP) method gave the same topology. Bootstrap 

values for nodes supported in >50% of 1,000 bootstrap replicates are shown above (for NJ) and 

below (for MP) the branches. Jp, Jv, and In represent japonica, javanica, and indica types of O. sativa, 

respectively. Ra, Rin, and Rp represent annual, intermediate, and perennial types of O. rufipogon, 

respectively. AF includes O. glaberrima (W025) and O. barthii (W1468, W1647). 

Notes 

Author’s address: Graduate School of Agriculture, Hokkaido University, 

Sapporo, Japan, e-mail: rysano@abs.agr.hokudai.ac.jp. 

Who was the mother of cultivated rice Differentiation 

of chloroplast genome structure among cultivated rice 

and ancestral wild species 

Koh-ichi Kadowaki 

Phylogenetic analysis of Oryza species based 

on the mitochondrial and chloroplast genomes 

Analysis of variation in the cytoplasmic genome is valuable 

for taxonomic analysis (e.g., Ge et al 1999). Simple sequence 

repeats (SSR) and their flanking regions in the mitochondrial 

and chloroplast genomes were sequenced in order to reveal 

DNA sequence variation (Fig. 1). The information was used to 

gain new insights into phylogenetic relationships among 22 

species in the genus Oryza (Vaughan et al 2003). Five chloroplast 

SSR loci (Ishii and McCouch 2000) and seven mitochondrial 

SSR loci equal to or longer than ten mononucleotide repeats 

were chosen from known rice mitochondrial (Notsu et al 


O. granulata complex 

O. sativa complex 

Related genus 

GG 

AA 

BB 

EE 

O. officinalis complex 

Wheat, 

barley, 

maize, 

etc. 

HHJJ 

FF 

CC 

BBCC 

CCDD 

Genus Oryza 

O. ridleyi complex 

Fig. 1. A model for evolution of the cytoplasmic genome in Oryza species based on 

SSRs and their flanking regions. 

2002) and chloroplast genome sequences (Hiratsuka et al 

1989). A total of 50 accessions of Oryza that represented six 

different genomes of diploid Oryza species and three different 

allopolyploid genomes were analyzed. Many base substitutions 

and deletions/insertions were identified in the SSR loci and 

their flanking regions. Of mononucleotide SSR, G (or C)-repeats 

were more variable than A (or T)-repeats. Results obtained 

by chloroplast and mitochondrial SSR analyses showed 

similar phylogenetic relationships among species, although 

chloroplast SSR were more informative because of their higher 

sequence diversity. The CC genome is suggested to be the 

maternal parent for the two BBCC-genome species (O. 

punctata and O. minuta) and the CCDD species (O. latifolia) 

based on the high level of sequence conservation between diploid 

CC-genome species and these allotetraploid species. 

Polyphyletic domestication of indica, temperate japonica, 

and tropical japonica from O. rufipogon: evaluation 

of genetic diversity by the chloroplast genome 

It is believed that Oryza sativa ecospecies indica and japonica 

were domesticated in Asia from O. rufipogon and/or O. nivara, 

but that domestication process is still not well understood. 

Complete chloroplast genome information is available for one 

japonica variety (Hiratsuka et al 1989). Recently, we have determined 

a whole chloroplast genome of O. nivara, an annual 

type of wild rice and possible progenitor to O. sativa (Masood 

et al 2004). Several insertion/deletion events were identified 

between the two species. To investigate the domestication process 

of rice, three polymorphic loci, 8K, 57K, and 67K, were 

examined using 303 accessions of Asian cultivated rice and 

32 accessions of Oryza species carrying the AA genome. The 

8K was originally found as ORF100 by Chen et al (1993) and 

Kanno et al (1993). The three regions showed a different extent 

of sequence variation. In the 8K region, most indica 

(76.1%) had a deletion type of polymorphism, whereas most 

japonica (98.5%) had a nondeletion type. In contrast, in the 

76K region, no length polymorphism was observed at all within 

the cultivated rice examined. In the 57K region, various length 

polymorphisms, so far six types, were identified by DNA sequencing. 

These results showed that the extent of polymorphism 

is variable depending on the chloroplast locus and that 

occurrence of an insertion/deletion event is independent regarding 

the three regions. Sequence variations at the 8K and 

57K regions were also present in some of the O. nivara and O. 

rufipogon but not in other wild species, suggesting that mutations 

at the two regions happened relatively recently and prior 

to the domestication of Asian cultivated rice. These results 

suggest that indica of type IV and VII were domesticated either 

from a local population of O. rufipogon via O. nivara or 

directly from a local population of O. rufipogon separately. 

Regarding japonica, japonica of type VI and IV (tropical type) 

were derived from a different local population of O. rufipogon 

without O. nivara, polyphyletically and independently. A previous 

study using nuclear element SINEs also showed that O. 

sativa has been derived polyphyletically from O. rufipogon 

(Ohtsubo et al 2004). Our study strongly suggests that there 

are at least four major lineages for the domestication process 

of Asian cultivated rice (Fig. 2). 


0 

Type IV-AA 

1 

Type IV-TT 

16 

2 

Hypothetical 

ancestor 

Type VII-TT 

15 

17 

Type VI 

unidentified 

O. rufipogon 

107 

indica 

japonica 

81 

Fig. 2. A model for the domestication process of cultivated rice from wild species. 

Numerals in the figure show the number of accessions. 

References 

Chen WB, Nakamura I, Sato Y, Nakai H. 1993. Distribution of deletion 

type in cpDNA of cultivated and wild rice. Jpn. J. Genet. 

68:597-603. 

Ge S, Sang T, Lu B-R, Hong D-Y. 1999. Phylogeny of rice genomes 

with emphasis on origins of allotetraploid species. Proc. Natl. 

Acad. Sci. USA 96:14400-14405. 

Hiratsuka J, Shimada H, Whittier R, Ishibashi T, Sakamoto M, Mori 

M, Kondo C, Honji Y, Sun CR, Meng BY. 1989. The complete 

sequence of the rice (Oryza sativa) chloroplast genome: 

intermolecular recombination between distinct tRNA genes 

accounts for a major plastid DNA inversion during the evolution 

of the cereals. Mol. Gen. Genet. 217:185-194. 

Ishii T, McCouch SR. 2000. Microsatellites and microsynteny in the 

chloroplast genomes of Oryza and eight other Gramineae species. 

Theor. Appl. Genet. 100:1257-1266. 

Kanno A, Watanabe N, Nakamura I, Hirai A. 1993. Variations in 

chloroplast DNA from rice (Oryza sativa): differences between 

deletions mediated by short direct-repeat sequences within a 

single species. Theor. Appl. Genet. 86:579-584. 

Masood S, Kadowaki K, Nishikawa T, Fukuoka S, Njenga PK, 

Tsudzuki T, Kadowaki K. 2004. The complete nucleotide sequence 

of wild rice (Oryza nivara) chloroplast genome: first 

genome-wide comparative sequence analysis of wild and cultivated 

rice. Gene. (In press.) 

Notsu Y, Masood S, Nishikawa T, Kubo N, Akiduki G, Nakazono M, 

Hirai A, Kadowaki K. 2002. The complete sequence of the 

rice (Oryza sativa L.) mitochondrial genome: frequent DNA 

sequence acquisition and loss during the evolution of flowering 

plants. Mol. Genet. Genomics 268:434-445. 

Ohtsubo H, Cheng C, Ohsawa I, Tsuchimoto S, Ohtsubo E. 2004. 

Rice retroposon p-SINE1 and origin of cultivated rice. Breed. 

Sci. 54:1-11. 

Vaughan DA, Morishima H, Kadowaki K. 2003. Diversity in the 

Oryza genus. Curr. Opin. Plant Biol. 6:139-146. 

Notes 

Author’s address: National Institute of Agrobiological Sciences, 

Tsukuba, Ibaraki 305-8602, Japan, e-mail: 

kadowaki@affrc.go.jp. 

Diverse mechanisms of low-temperature stress 

response in rice 

Ryozo Imai, Jiangqi Wen, Kentaro Sasaki, and Kiyoharu Oono 

Rice is widely cultivated in a large number of different natural 

environments. Compared with other cereal crops such as wheat 

and barley, rice is much more sensitive to low temperature, 

probably as a result of its subtropical origin. In terms of impact 

on rice production, male sterility is the most damaging 

impact of chilling. The developmental stages from pollen formation 

to fertilization are the most vulnerable to low temperature 

throughout the life cycle of rice plants (Nishiyama 1984). 

It has been reported that the young microspore stage in pollen 

development was the most sensitive to low temperature (Satake 

and Hayase 1970). Exposure of rice plants at the tetrad stage 

to a moderate low temperature (12 ºC) for 4 days resulted in 

male sterility in 80% of the spikelets (Satake and Hayase 1970, 

Nishiyama 1984). Microscopic observation of developing rice 

anthers suggested that one possible reason for male sterility 

after low-temperature treatment was the failure of anther de- 


velopment. Observed abnormalities included the cessation of 

anther development, arrest of pollen development, anthers remaining 

within the flowers after anthesis, and partial or no 

dehiscence (Satake 1976). 

Cytological observation revealed a dilatation of tapetal 

layers in chilling-treated rice anthers (Nishiyama 1976, 1984). 

Dilatation of the tapetal layer was accompanied by a vigorous 

augmentation of cytoplasmic organelles such as mitochondria, 

proplastids, Golgi bodies, and endoplasmic reticula (Nishiyama 

1976, 1984). Chilling temperature treatment also affects the 

physiological status of anthers. Nonreducing sugar content was 

found to increase rapidly, whereas acid phosphatase activity 

decreased in the moderate temperature-treated rice anthers 

(Nishiyama 1984). Possible involvement of phytohormones, 

such as GA and auxin in chilling-induced male sterility, has 

been reported (Nishiyama 1975, Yoshioka and Suge 1996). 

However, it is still largely unknown how chilling temperature 

induces molecular events that result in male sterility in rice 

plants. 

In this paper, we used a large-scale screening of genes 

responsive to a moderate chilling temperature (12 ºC) and functional 

analysis of the gene products. We report on diverse 

molecular responses that occur during chilling stress in rice. 

Materials and methods 

Plant materials, growth conditions, 

and stress treatment 

Seeds of Oryza sativa L. cv. Yukihikari (japonica rice) were 

grown in a growth chamber at 25 ºC under continuous illumination 

(256 mmol m –2 s –1 ). After growing for 7 days, rice seedlings 

underwent environmental stress treatments. To collect 

the flowers and anthers, Yukihikari plants were grown in pots 

with nutrient soil for 2 months in a phytotron room controlled 

at 25/19 ºC (day/night). Anthers and panicles at the tetrad stage 

of microspore development were collected and frozen immediately 

in liquid nitrogen. For the chilling treatment, the pots 

with rice plants were transferred to a phytotron room that was 

precooled to 12 ºC. Anthers and panicles were collected at 48 

h of 12 ºC treatment. 

cDNA subtraction 

Total RNA was isolated using TRIzol reagent (Invitrogen) and 

poly (A) + RNA was purified using Dynabeads Oligo(dT) 25 

(Dynal). The PCR-Select cDNA subtraction kit (CLONTECH) 

was used to isolate chilling-induced clones. Candidate clones 

were arrayed on Hybond-N + membranes (Amersham- 

Pharmacia Biotech) and further selected using differential hybridization. 

Northern blot analysis 

Twenty micrograms of total RNA that was isolated from 

stressed or control samples using TRIzol reagent were separated 

on 1.0% formaldehyde denaturing agarose gels and transferred 

onto Hybond-N + membranes according to standard 

methods. 

Microarray screening 

Total RNA was extracted from seedling roots and used for labeling 

with Cy5. Microarray plates containing about 9,000 

cDNA clones (provided by STAFF) were screened with target 

cDNA from nonstressed (NS) and cold-treated (2 or 24 h) roots 

(CS). Cold-induced clones were selected from spots with a 2 

or higher CS/NS signal ratio. 

Results and discussion 

Subtractive screening of cDNA that accumulated in 12 ºCtreated 

anthers identified a cDNA clone, OsMEK1, encoding 

a protein with features characteristic of an MAP kinase (Wen 

et al 2002). The putative OsMEK1 protein shows 92% identity 

to the maize MEK homolog, ZmMEK1. OsMEK1 transcript 

levels were induced in rice anthers by 12 ºC treatment 

for 48 h. Similar OsMEK1 induction was observed in shoots 

and roots of seedlings that were treated at 12 ºC for up to 24 h. 

Interestingly, no induction of OsMEK1 transcripts was observed 

in 4 ºC-treated seedlings. In contrast, rice lip19, encoding a 

bZIP protein possibly involved in low-temperature signal transduction, 

was not induced by 12 ºC treatment but was induced 

by 4 ºC treatment. Among the three MAP kinase homologs 

cloned, only OsMAP1 displayed a 12 ºC-specific induction 

pattern similar to that of OsMEK1. A yeast two-hybrid system 

revealed that OsMEK1 interacts with OsMAP1, but not with 

OsMAP2 and OsMAP3, suggesting that OsMEK1 and 

OsMAP1 probably function in the same signaling pathway. 

An in-gel assay of protein kinase activity revealed that a protein 

kinase (approx. 43 kDa), which preferentially uses MBP 

as a substrate, was activated by 12 ºC treatment but not by 

4 ºC treatment. These results lead us to conclude that at least 

two signaling pathways for low-temperature stress exist in rice 

and an MAP kinase pathway with OsMEK1 and OsMAP1 components 

is possibly involved in the signaling for the higherrange 

low-temperature stress. 

Microarray slides were screened with target cDNA from 

nonstressed and chilling-stressed (2 or 24 h) roots. Forty-nine 

spots were selected for cDNA clones that are induced by the 

2-h cold treatment. Database searches indicated that 26 clones 

have a putative function and contain lip9 (Aguan et al 1991), 

OsCDPK7 (Saijo et al 2000), and OsMAP1, which are known 

to be cold-inducible. In addition, all of the 12 clones that were 

analyzed by Northern blots showed induction within 2 h of 

12 ºC treatment. Collectively, these data suggest that the rice 

microarray system is highly reliable for identifying stress-related 

genes. The newly identified 12 ºC-induced clones include 

NAC-family transcription factors, a zinc finger protein, and a 

ring finger protein, suggesting that these DNA-binding proteins 

may be involved in chilling signal transduction. Genes 

for biosynthesis of phospholipids, a disaccharide, polyamines, 

and jasmonate were also induced. These data suggest that 

metabolic changes occur as early responses to moderate lowtemperature 

stress. Exposure to moderate low temperature for 

a longer period (24 h) induced different responses in rice roots. 

Screening of spots that gave stronger signals in CS (24 h) than 


in NS identified 94 spots. Functions for 29 clones were assigned 

after database searches. Those clones include genes for 

protein sorting and vesicle transport, translation, and oxidative 

stress protection. 

Our data demonstrated that moderate (12 ºC) and severe 

(4 ºC) chilling stresses could be recognized as distinctive signals 

in rice. We were able to detect a number of responses of 

rice to moderate chilling temperature. Identification of functions 

for such genes will provide us with a clue to understand 

specific molecular events that occur during moderate chilling 

stress in rice. 

References 

Aguan K, Sugawara K, Suzuki N, Kusano T. 1991. Isolation of genes 

for low-temperature-induced proteins in rice by a simple subtractive 

method. Plant Cell Physiol. 32:1285-1289. 

Nishiyama I. 1975. Male sterility caused by cooling treatment at the 

young microspore stage in rice plants. XI. Effects of some 

substances on sterility. Proc. Crop Sci. Soc. Jpn. 44:397-402. 

Nishiyama I. 1976. Male sterility caused by cooling treatment at the 

young microspore stage in rice plants. Proc. Crop Sci. Soc. 

Jpn. 45:254-262. 

Nishiyama I. 1984. Climate influence on pollen formation and fertilization. 

In: Tsunoda S, Takahashi N, editors. Biology of rice. 

Amsterdam (Netherlands): Elsevier. p 153-171. 

Saijo Y, Hata S, Kyozuka J, Shimamoto K, Izui K. 2000. Over-expression 

of a single Ca 2+ -dependent protein kinase confers 

both cold and salt/drought tolerance on rice plants. Plant J. 

23:319-327. 

Satake T, Hayase H. 1970. Male sterility caused by cooling treatment 

at the young microspore stage in rice plants. V. Estimation 

of pollen developmental stage and most sensitive stage 

to coolness. Proc. Crop Sci. Soc. Jpn. 39:468-473. 

Satake T. 1976. Determination of the most sensitive stage to steriletype 

cool injury in rice plants. Res. Bull. Hokkaido Natl. Agric. 

Exp. Stn. 113:1-43. 

Yoshioka T, Suge H. 1996. Damage of seed fertility by cooling treatment 

and endogenous gibberellins in ears of rice plants (Oryza 

sativa L.). Breed. Sci. 46:173-178. 

Wen JQ, Oono K, Imai R. 2002. Two novel mitogen-activated protein 

signaling components, OsMEK1 and OsMAP1, are involved 

in a moderate low temperature signaling pathway in 

rice. Plant Physiol. 129:1880-1891. 

Notes 

Authors’ addresses: Ryozo Imai and Kentaro Sasaki, Winter Stress 

Laboratory, National Agricultural Research Center for 

Hokkaido Region, NARO, Sapporo, Japan. Present address: 

Jiangqi Wen, Division of Biological Sciences, University of 

Missouri-Columbia, MO, USA; Kiyoharu Oono, Research 

Center for Environmental Genomics, Kobe University, Kobe, 

Japan, e-mail: rzi@affrc.go.jp. 

Genetic diversity of Myanmar rice landraces 

Ye Tint Tun, K. Irie, T. Nagamine, Jhon Ba Maw, F. Kikuchi, and H. Fujimaki 

Myanmar is a tropical country in monsoon Asia and is included 

in the center of diversity of cultivated rice along with Yunnan 

Province (China), east Nepal, Bhutan, Assam (India), and northern 

Thailand (Nakagahra and Hayashi 1977, Nakagahra et al 

1984, Glaszmann 1987). Myanmar has varied natural environments, 

various social and cultural customs, different types 

of farming and cropping systems, and tribal diversity. Consequently, 

various ecotypes such as rainfed, irrigated, deepwater, 

floating, and upland rice are cultivated and a number of 

landraces are differentiated in each ecotype. A number of 

landraces are still cultivated locally but are being rapidly replaced 

by improved cultivars. It is thus necessary to collect 

landraces for ex situ conservation and to characterize and evaluate 

them for future rice breeding. 

Our study deals with a thousand or more rice landraces 

collected throughout Myanmar in order to investigate their genetic 

diversity and characterize them. 

Agroecological zoning of rice-growing environments 

The characterization of rice-growing areas was first needed 

for analyzing the diversity of landraces of Myanmar cultivated 

rice. Therefore, agroecological zoning was attempted by using 

monthly meteorological data accumulated at 34 observatory 

sites throughout Myanmar for more than 30 years of rainfall; 

maximum, minimum, and mean temperatures; differences 

between day and night temperatures; evapotranspiration; and 

sunshine, compiled by FAO (1987). Eight agroecological zones 

were elaborated by the results of the principal component analysis: 

Northern Mountainous (NM), Eastern Plateau (EP), Semiarid 

(SA), Western Hilly (WH), Western Coastal (WC), Southern 

Coastal (SC), Southern Plain (SP), and Ayeyarwady Delta 

(AD) (Fig. 1). The NM, EP, and WH zones had the most complex 

topography, with high mountains and deep valleys, but 

the others are rather flat and plain zones. 

In Myanmar, rainfed rice is the most dominant ecotype 

and it is grown extensively in SP and AD zones. In the mountainous 

zones, NM, WH, and EP, upland rice is preferentially 

grown in the slash-and-burn farming system. Irrigated rice is 

predominantly cultivated in the central SA zone and deepwater 

and floating rice are grown exclusively in delta areas in AD 

and WC zones. 


WH 

n = 134 

NM 

n = 232 

NM 

WC 

n = 130 

WH 

EP 

EP 

n = 260 

WC 

SA 

SP 

SP 

n = 144 

AD 

SC 

SA 

n = 125 

AD 

n = 149 

SC 

n = 217 

Low 

Medium 

High 

Myanmar 

n = 1,391 

Fig. 1. Regional variation in photoperiod sensitivity of landraces of cultivated rice in Myanmar. WH = 

Western Hilly, NM = Northern Mountainous, EP = Eastern Plateau, SA = Semiarid, WC = Western Coastal, 

SC = Southern Coastal, SP = Southern Plain, and AD = Ayeyarwady Delta. 


Indica-japonica discrimination 

A core collection of 300 samples randomly chosen from the 

1,391 landrace collection was used to discriminate between 

indica and japonica Myanmar rice landraces. A discriminating 

index (Z) was employed, which was composed of a linear function 

of degree of phenol color reaction (Ph), tolerance of KClO 3 

(K), and apiculus hair length (H). Classification of landraces 

was carried out by using those indices calculated by the following 

formula (Sato 1991): 

Z = Ph + 1.313K – 0.82H – 1.251 

Eighty percent of the landraces were found to belong to 

indica and 20% to japonica. Indica landraces were distributed 

ubiquitously in every zone, whereas japonicas were predominant 

exclusively in the WH zone. Several techniques have been 

proposed by different researchers to discriminate between indica 

and japonica. Their results are not always consistent but 

the presently employed discriminating index seemed to be the 

most useful for identifying subspecific differentiation of 

Myanmar cultivated rice. 

Photoperiod sensitivity 

Some 1,391 landraces were sown at Yezin in Myanmar twice a 

year—on 12 March 2001 under long-day conditions and on 

25 August 2001 under short-day conditions—to observe heading 

responses under different daylength. The photoperiod sensitivity 

(PS) of rice landraces was evaluated based on the differences 

in days to heading under the two different growing 

seasons. The basic vegetative growth (BVG) was estimated 

from days to heading of the landraces when they were sown on 

25 August. 

A great diversity was observed in the BVG and PS of 

landraces. The PS of landraces was categorized into three 

groups as low (less than 59 days), intermediate (60 to 109 d), 

and high (above 110 d). It was disclosed that 15% of Myanmar 

landraces were less sensitive to photoperiodism, 13% were 

intermediate, and 72% were highly sensitive. Figure 1 shows 

the regional variation of PS. A distinctive cline was observed 

from the mountainous to the plain zones. This sort of differentiation 

among zones was considered to be closely associated 

with cultivated ecotypes, the prevailing cropping system, and 

the natural conditions of each rice-growing area. 

Photoperiod sensitivity is the most important factor in 

determining the growth duration of the rice plant and is considered 

to be closely associated with local adaptability and 

with ecological differentiation of rice landraces as well. 

Amylose content of endosperm starch and genotypes 

of esterase isozyme 

The 1,391 landraces were used for analyzing the amylose content 

of endosperm starch and a core collection of 300 samples 

randomly chosen from the landrace collection was used to identify 

esterase isozyme genotypes. Landraces of Myanmar cultivated 

rice were found to have great diversity in amylose content 

of endosperm starch, ranging from 0 to 40%. Glutinous 

landraces distributed ubiquitously in every zone had less varied 

frequencies. However, the amylose content of nonglutinous 

landraces varied from zone to zone. In particular, landraces 

with very low (3–10%) amylose content appeared more frequently 

in the northeastern NM and EP zones, but amylose 

content was less or zero in the middle to southwestern zones 

such as WC, SA, SP, AD, and SC (Fig. 2). Glutinous landraces 

have been grown in every zone and are supposed to be used 

especially for making rice snacks or cakes at seasonal festivals. 

Those who live in the northeastern zones such as NM 

and EP are supposed to have a strong preference for low-amylose 

rice and/or any special way of cooking it. In Japan, lowamylose 

mutants have been artificially induced to improve the 

palatability of cooked rice of japonica cultivars (Okuno et al 

1983). However, low-amylose landraces occurred spontaneously 

in the northeastern zones of Myanmar. The amylose content 

of endosperm starch seems closely associated with the 

palatability of cooked rice and is highly influenced by human 

selection. We need to clarify whether the low amylose content 

of landraces in the northeastern zones is controlled by the du 

genes or any other genetic mechanisms. 

Isozymes have an identical enzymic function and are supposed 

to be less affected by or rather independent of natural 

and/or artificial selection. Consequently, polymorphic diversity 

in isozymes may reflect the phylogenetic differentiation 

of rice landraces. Esterase isozymes have been used repeatedly 

for investigating polymorphic diversity reflecting the 

phylogenic differentiation of various kinds of cultivated plants. 

For rice, three loci of Est1, Est2, and Est3 have been identified 

where two, three, and two alleles are located, respectively. 

Two alleles in the Est1 locus produce the 1A band and null in 

electrophoresis, three alleles in the Est2 do 6A and 7A band 

and null, and two alleles in the Est3 do 12A and 13A band. All 

combinations of those alleles in the three loci produce 12 genotypes 

(Nakagahra and Hayashi 1977). Myanmar landraces are 

found to have 11 genotypes among them. 

In contrast to the amylose content of endosperm, genotypic 

frequencies of esterase isozymes were less varied among 

zones and no significant difference in genotypic frequency was 

observed among agroecological zones. The contrasting frequency 

distributions observed between amylose contents and 

esterase isozyme genotypes are considered greatly influenced 

by differences in effects of natural and/or artificial selection. 

Conclusions 

Myanmar has various social and cultural customs, complex 

topography, and different ethnic groups in its eight 

agroecological zones. The genetic diversity of rice landraces 

is thought to have been brought about by natural factors such 

as various topographic and climatic conditions and by different 

social and agricultural circumstances. 


45 

45 

40 

40 

WH 

35 

35 

n = 134 

30 

30 

25 

25 

20 

20 

15 

15 

10 10 

50 50 

NM 

n = 232 

NM 

45 

40 

35 

30 

25 

20 

15 

10 

5 

0 

WC 

n = 130 

WC 

WH 

SA 

EP 

45 

40 

35 

30 

25 

20 

15 

10 

5 

0 

EP 

n = 260 

45 

40 

35 

30 

25 

20 

15 

10 

5 

0 

SP 

n = 144 

AD 

SP 

SC 

45 

40 

35 

30 

25 

20 

15 

10 

5 

0 

SA 

n = 125 

45 

40 

35 

30 

25 

20 

15 

10 

5 

0 

0 

2 

AD 

n = 149 

4 8 12 16 20 24 28 32 36 

6 10 14 18 22 26 30 34 38 

45 

40 

35 

30 

25 

20 

15 

10 

5 

0 

40 0 

2 

SC 

n = 217 

4 8 12 16 20 24 28 32 36 40 

6 10 14 18 22 26 30 34 38 

No. of landraces 

180 

160 

140 

120 

100 

80 

60 

40 

20 

Myanmar 

n = 1,391 

0 

0 

2 

4 

6 

8 

10 

12 

14 

16 

18 

20 

22 

24 

26 

28 

30 

32 

34 

36 

38 

40 

Amylose content (%) 

Fig. 2. Regional variation in amylose content of endosperm starch in Myanmar rice landraces. WH = Western Hilly, NM = Northern 

Mountainous, EP = Eastern Plateau, SA = Semiarid, WC = Western Coastal, SC = Southern Coastal, SP = Southern Plain, and AD 

= Ayeyarwady Delta. 


Rice landraces in Myanmar show remarkable diversity 

in traits such as grain size and shape, plant type, etc., besides 

the characters investigated here. Their diversity is considered 

to result from adaptation to various natural and agroecological 

environments and to be caused by unconscious selection by 

farmers for several thousand years and by artificial selection 

applied through preference for different qualities and ingredients 

of rice grains. 

The diversity of genetic resources of Myanmar rice is 

expected to be very useful for domestic rice breeding in the 

future and an irreplaceable and valuable property for human 

beings from the global point of view. These resources are becoming 

rapidly eroded because of the extensive diffusion of 

improved high-yielding cultivars. We urgently need to collect 

and characterize rice landraces. Therefore, the Myanmar seed 

bank has collected and stored a number of them. These will be 

very useful genetic resources for future rice breeding but remain 

to be evaluated for useful agronomic traits. 

References 

FAO. 1987. Agro-climatologic data. FAO Plant Production and Protection 

Series, Rome. 

Glaszmann JC. 1987. Isozymes and classification of Asian rice varieties. 


Nakagahra M, Hayashi KI. 1977. Origin of cultivated rice as detected 

by isozyme variations. JARQ 11:1-5. 

Nakagahra M, Akihama T, Hayashi KI. 1984. Geographical distribution 

of esterase genotypes of rice in Asia. Rice Genet. Newsl. 

1:118-120. 

Okuno K, Fuwa H, Yano M. 1983. A new mutant gene lowering 

amylase content in endosperm starch of rice. Jpn. J. Breed. 

33:387-394. 

Sato YI. 1991. Variation in spikelet shape of indica and japonica 

rice cultivars of Asian origin. Jpn. J. Breed. 41:121-134. 

Notes 

Authors’ addresses: Ye Tint Tun, K. Irie, F. Kikuchi, and H. Fujimaki, 

Tokyo University of Agriculture, 1-1-1 Sakuragaoka, Setagaya, 

Tokyo 156-8502, Japan, e-mail: tint@affrc.go.jp; Ye Tint Tun, 

Myanma Agriculture Service, Ministry of Agriculture and Irrigation, 

Yangon, Myanmar; T. Nagamine, National Agricultural 

Research Center for Western Region, Fukuyama, 

Hiroshima 721-8514, Japan; Jhon Ba Maw, Department of 

Agricultural Research, Ministry of Agriculture and Irrigation, 

Yezin, Myanmar. 

QTL analysis for eating quality in japonica rice 

H.G. Hwang, J.P. Suh, Y.C. Cho, S.J. Kwon, I.S. Choi, H.C. Hong, Y.G. Kim, M.K. Kim, H.C. Choi, and Y.T. Lee 

In Korea, the most important goal in japonica rice breeding is 

to improve the eating quality of cooked rice. The eating quality 

of rice is a complex characteristic whose many components 

are difficult to evaluate for selection in a rice breeding program. 

The quality of cooked rice has been examined by a sensory 

test. Recently, we have been using an indirect rapid evaluation 

method by detecting the glossiness of cooked rice for 

eating quality. However, this is difficult to apply for early-generation 

selection in the rice breeding program because of size 

constraints. 

The use of molecular markers has facilitated the understanding 

of quantitative trait loci (QTLs) and marker-assisted 

selection (MAS). Many QTL analyses for grain quality of rice 

have been reported (Bao et al 2002, Tan et al 1999, Wan et al 

2004, Zhou et al 2003). QTLs related to glossiness of cooked 

rice were identified at different loci in Ilpumbyeo—a highquality 

parent (Cho et al 2004, Lee 2003). In this study, we 

report on QTLs associated with physicochemical traits and 

eating quality by glossiness of cooked rice using a japonica × 

japonica cross. Also, we evaluated the allelic effect of QTLs 

associated with glossiness of cooked rice in BC 3 F 1 and BC 4 F 2 

generations developed by MAS. 


Plant materials 

Suweon365 possesses medium eating quality and is highly resistant 

to blast. On the other hand, Chucheongbyeo is known 

for its good eating quality and susceptibility to leaf blast. A 

total of 231 recombinant inbred lines (RILs) derived from a 

cross between Suweon365 and Chucheongbyeo were developed 

using the single-seed descent method. The RIL population 

was field-evaluated by the National Institute of Crop Science, 

Suwon, for three years (1999 to 2001). To develop a 

series of QTL-NILs (near-isogenic lines), four desirable RILs 

carrying QTL segments for glossiness of cooked rice (GCR) 

from a Toyo taste meter value were successively backcrossed 

with Suweon365 for selection using linked markers until the 

BC 4 F 2 . The allele effect of QTLs related to GCR was evaluated 

in BC 3 F 1 and BC 4 F 2 generations. 

Chemical properties 

The amylose content of milled rice was determined by the relative 

absorbency of starch-iodine color in digested solution of 

100-mesh rice flour by the modified method of Juliano (1971). 

Protein content was calculated by total nitrogen multiplied by 


5.95 after determining the nitrogen content of rice material 

using the Micro-Kjeldahl method (Foss:2300 Kjeltec Analyzer). 

Gelatinization characteristics were evaluated using a 

Rapid Visco Analyser (model: RVA-3D) in accordance with 

the operation manual (NewPort Sci. Co., Australia). 

Glossiness of cooked rice (GCR) 

using a Toyo taste meter 

The glossiness of cooked rice was determined in two replications 

using a Toyo taste meter (model: MA-90A and 90B) in 

accordance with the operation manual (TRCM Co.). The whole 

rice grains of milled rice were separated from broken rice grains 

using a length grain separator of a Satake testing rice grader 

(Model: TGR 05A) of mesh size 3.6 mm for 2 min and a 33-g 

sample of whole grain was used. 

Linkage map construction and QTL analysis 

The linkage map was constructed with 236 markers of various 

origins such as SSRs, STS, AFLP, and MITE. The Mapmaker 

program was used to construct a molecular map at an LOD 

value of 3.0 and a maximum recombination fraction q = 0.40 

using the “group” command. QTL analysis was performed by 

one-way analysis of variance using the qGene v.3.06 program. 

An LOD score of 2.0 was used as the threshold for detecting 

QTLs. 


The correlation coefficients among seven chemical properties 

and GCR (Toyo taste meter value) were computed. There was 

a high correlation (r ≥ 0.8) between GCR and the sensory evaluation 

for eating quality of cooked rice by panelists (data not 

shown, Lee 2003). GCR by the Toyo taste meter value showed 

a highly negative correlation with amylose and protein content, 

and a positive correlation with alkali digestion value. This 

result was similar to those of previous reports in japonica × 

japonica (Lee 2003) and japonica × javanica crosses (Cho et 

al 2004). The eating quality of cooked rice can be evaluated 

indirectly based on GCR by the Toyo taste meter value instead 

of palatability score because the sensory test needs much time 

and labor from well-trained panelists. 

The polymorphism between Suweon365 and 

Chucheongbyeo was 16.3%, 4.8%, and 13.0% for SSR, STS, 

and AFLP markers, respectively. This result was similar to that 

of the cross between two Korean japonica cultivars (Lee 2003). 

The linkage map spanning 1,998 cM with an average interval 

size of 8.5 cM was constructed using 236 DNA markers. The 

distal region of the long arm on chromosome 2 showed only 

polymorphism and a map could be constructed of 8.1 cM by 

seven SSR markers. The coverages of chromosomes 1, 5, 6, 

and 10 were relatively poor. There were big gaps in many regions 

of chromosomes 3, 4, 5, 8, 11, and 12. AFLP markers 

were not mapped evenly on all chromosomes, and were combined 

only on chromosomes 4, 7, 8, 10, 11, and 12. 

For chemical properties, gelatinization characteristics, 

and glossiness of cooked rice, 15 putative QTLs were identified. 

Three QTLs for alkali digestion value were located on 

chromosomes 6, 8, and 11. One QTL, qADV-6, was identified 

at the wx locus (RM190). A QTL, qPro-10, for protein content 

was identified near RM333 on chromosome 10. It was different 

from the one identified on chromosomes 1, 4, 5, 6, and 7 

by Hu et al (2004). A QTL, qGIT-11, for gelatinization temperature 

was identified in the same region (em2210) as qADV- 

11. Three QTLs for high viscosity, breakdown, and consistency 

were located in a cluster region (RM213–RM482) at the 

distal end of the long arm on chromosome 2. The QTLs for 

gelatinization characteristics in a DH population from japonica 

× indica were reported on chromosomes 1, 2, 3, 6, and 11 by 

Tian et al (2004). Five QTLs for amylose content, gel consistency, 

water absorption, cooked rice elongation, and volume 

expansion were identified near the wx locus. However, only 

one QTL for alkali digestion value in this mapping population 

was identified at the wx locus. The QTLs for GCR by the Toyo 

taste value were identified on chromosomes 3, 7S, 7L, 8, and 

11 (Fig. 1). These QTLs explained 7.1%, 10.6%, 8.8%, 8.9%, 

and 7.2% of the total phenotypic variation, respectively. A QTL 

for GCR was not detected repeatedly at the same locus in three 

years. In addition, QTLs were different from those from 

japonica × japonica (Lee 2003), japonica × javanica (Cho et 

al 2004), and japonica × indica (Ebitani et al 2002) crosses. 

They were also located at loci different from 13 QTLs affecting 

six palatability properties in japonica × indica (Wan et al 

2004) crosses. Traits related to grain quality appear to be influenced 

by ecotypes as well as by environment and cultural 

method. 

BC 3 F 1 and BC 4 F 2 lines developed by successive backcrossing 

with Suweon365 for four desirable RILs carrying four 

putative QTL segments for GCR, based on MAS, were evaluated 

for allele effect (Table 1). In the BC 3 F 1 generation, the 

lines having DNA markers close to the QTL showed an allelic 

effect of 2.5 for qGCR-7.1, 1.6 for qGCR-7.2, 2.1 for qGCR- 

8, and 4.1 for qGCR-11, contributed by Chucheongbyeo. In 

the BC 4 F 2 generation, the lines homozygous for the 

Chucheongbyeo allele had an allelic effect of 1.1 for qGCR- 

7.1, 2.7 for qGCR-7.2, 1.9 for qGCR-8, and 1.0 for qGCR-11. 

Although the allelic effect was not statistically significant, all 

lines having QTLs from Chucheongbyeo for GCR from the 

Toyo taste meter value increased the value of glossiness of 

cooked rice vis-à-vis the lines with the Suweon365 allele. It is 

difficult to evaluate the dosage effect for QTLs because the 

values of Chucheongbyeo homozygotes were not always as 

high as those of the heterozygous lines (Table 1). The allelic 

effect of QTLs associated with GCR was lower because the 

parents used to develop genetic material had a similar japonica 

genetic background. These results will be helpful for developing 

effective breeding selection for grain quality of japonica 

rice. In the future, we will try to evaluate the effect of QTL 

pyramiding associated with the glossiness of cooked rice. 


Chr. 7 

Chr. 8 

Chr. 11 

qGCR-7-1 

3.2 cM 

RM481 

RM180 

RM214 

RM500 

em623 

em723 

em724 

em127 

em225 

em525 

em524 

em425 

em424 

1031R25 

1031R08 

RM320 

RM560 

OSR22 

qGCR-8 

3.8 cM 

qADV-8 

qHV-8 

em2211 

RM547 

RM72 

em445 

em145 

RM044 

RM404 

em5211 

em429 

em4210 

em7210 

em123 

em522 

RM342A 

RM342B 

qGCR-11 

8.6 cM 

qGIT-11 

qADV-11 

1031R37 

OSR01 

RM20B 

RM332 

RM552 

em1217 

em7215 

em429 

em2210 

em4214 

RM229 

RM021 

RM206 

em426 

qGCR-7-2 

4.2 cM 

RM342D 

em443 

em243 

em143 

RM346 

RM182 

OSR04 

RM336 

RM070 

em529 

em738 

em1210 

em1211 

em428 

em229 

em627 

dasng04 

RM505 

em864 

RM473a 

RM234 

RM515 

1031R28 

RM281 

RM264 

em812 

em427 

em726 

em129 

em526 

em442 

RM144 

em242 

dasng08 

1031R23 

RM47 

RM478 

Fig. 1. Chromosomal location of QTLs associated with glossiness of cooked rice. qGCRs are QTLs for glossiness of cooked rice 

from the Toyo taste meter. 

References 

Bao JS, Wu YR, Hu B, Wu P, Cui HR, Shu QY. 2002. QTL for rice 

grain quality based on a DH population derived from parents 

with similar apparent amylose content. Euphytica 128:317- 

324. 

Cho YC, Hong HC, Suh JP, Jeong YP, Choi IS, Kim MK, Kim YG, 

Choi HC, Hwang HG. 2004. QTL mapping for grain quality 

and shape in japonica × javanica in rice. Korean J. Breed. 

36(Suppl. 1):408-409. 

Ebitani T, Umemoto T, Yano M. 2002. QTL analysis of “Mido” value, 

an inbreed selection index for eating quality of rice, using 

progenies from a cross between japonica and indica variety. 

Breed. Sci. 52(Suppl. 1):371. 

Hu ZL, Li P, Zhou MQ, Zhang ZH, Wang LX, Zhu LH, Zhu YG. 

2004. Mapping of quantitative trait loci (QTLs) for rice protein 

and fat content using doubled haploid lines. Euphytica 

135:47-54. 

Juliano BO. 1971. A simplified assay for milled rice amylose. Cereal 

Sci. Today 16:334-338, 340, 360. 


Table 1. Allelic effect of QTLs for glossiness of cooked rice by Toyo taste value in BC 3 F 1 and BC 4 F 2 

generations. 

Glossiness of cooked rice by Toyo taste value 

QTLs Chr. Markers BC 3 F 1 (2003) BC 4 F 2 (2004) 

Hetero Homo Allelic Homo Hetero Homo Allelic 

of CS a of SS effect b of CC of CS of SS effect 

qGCR7.1 7S RM214-RM500 64.8 62.3 2.5 69.3 69.3 68.2 1.1 

qGCR7.2 7L RM234-RM47 65.9 64.3 1.6 68.8 67.5 66.1 2.7 

qGCR8 8 RM547-RM72 67.6 64.3 2.1 66.1 67.1 64.2 1.9 

qGCR11 11 RM20B-RM332 66.1 63.5 4.1 68.4 67.4 67.4 1.0 

Chucheongbyeo (donor) 72.2 71.4 

Suweon365 (recurrent) 65.4 64.1 

a Homo of CC: one band for Chucheongbyeo; hetero of CS: two bands of Chucheongbyeo and Suweon365; homo of SS: one band 

for Suweon365. b Allelic effects in BC 3 F 1 and BC 4 F 2 are the value of CS-SS and CC-SS, respectively. 

Lee JS. 2003. QTL analysis for grain quality properties in a japonica 

rice combination. Ph.D. thesis. Yeungnam University, Korea. 

64 p. 

Tan YF, Li JX, Yu SB, Xing YZ, Xu CG, Zhang Q. 1999. The three 

important traits for cooking and eating quality of rice grains 

are controlled by a single locus in an elite rice hybrid, Shanyou 

63. Theor. Appl. Genet 99:642-648. 

Tian R, Jiang GH, Shen LH, Wang LQ, He YQ. 2004. Mapping quantitative 

trait loci underlying the cooking and eating quality of 

rice using a DH population. Mol. Breed., on-line, accessed 

on 30 Sept. 2004. 

Wan XY, Wan JM, Su CC, Wang CM, Shen WB, Li JM, Wang HL, 

Jiang L, Liu SJ, Chen LM, Yasui H, Yoshimura A. 2004. QTL 

detection for eating quality of cooked rice in a population of 

chromosome segment substitution lines. Theor. Appl. Genet., 

on-line, 10.1007/s00122-004-1744-3. 

Zhou PH, Tan YF, He YQ, Xu CG, Zhang Q. 2003. Simultaneous 

improvement for four quality traits of Zhenshan 97, an elite 

parent of hybrid rice, by molecular marker-assisted selection. 


Notes 

Authors’ addresses: H.G. Hwang, J.P. Suh, Y.C. Cho, I.S. Choi, H.C. 

Hong, Y.G. Kim, M.K. Kim, H.C. Choi, Y.T. Lee, National 

Institute of Crop Science, RDA, Suwon 441-857, Republic of 

Korea; S.J. Kwon, National Institute of Agricultural Biotechnology, 

RDA, Suwon 441-857, Korea, e-mail: 

yccho@rda.go.kr. 


SESSION 2 

Structure and function of the rice genome 

CONVENER: T. Sasaki (NIAS) 

CO-CONVENERS: H. Hirochika (NIAS) and H. Leung (IRRI)

The complete rice genome sequence 

and its application to breeding and genetics 

Takuji Sasaki 

Rice is one of the main staples in the world and is cultivated 

mainly in Asia, Africa, and Latin America using irrigated and 

upland ground for paddy fields. The major limiting factors for 

rice cultivation are climate, soil, and water. Except for the 

upland rice in Africa, rice plants grow mainly in tropical and 

semitropical areas characterized by a rainy season. The wild 

relatives of cultivated rice, Oryza sativa, are also widely distributed 

within nearly 20° north and south latitude (IRRI 2004). 

Archaeological evidence suggested that rice cultivation began 

about 10,000 years ago in southern China, with wild rice as 

the main resource. Although the exact origin of both O. sativa 

subsp. indica and subsp. japonica is still unknown, rice has 

been well domesticated even in such a “short” period and rice 

cultivation has become widespread, extending to 40° north and 

south latitude. The domestication of rice is undoubtedly one 

of the most important developments in human history and, at 

present, about half of the world population relies on rice for 

its daily main food source. 

Despite the long history of breeding and improvement 

of rice varieties adapted to a wide range of environmental conditions 

as well as varieties highly preferred by both farmers 

and consumers, novel rice varieties are still necessary to meet 

the demands of increasing population pressure coupled with 

decreasing resources for rice production. In addition, varieties 

with tolerance of biotic and abiotic stress, and high yield as 

well as good nutritional and cooking quality, will always be 

targeted by breeders and farmers. The breeding strategy used 

so far for developing new varieties is based on the selection of 

desirable phenotypes among progenies generated after crossing 

elite varieties. Since most agronomic traits are composed 

of more than two genes and, in extreme cases, ten to 15 genes, 

this strategy is similar to a “hit or miss” approach, making it 

extremely difficult to select the most preferable allele for each 

of the genes corresponding to a phenotype. We must be innovative 

in improving our rice breeding strategy for such complex 

traits mainly contributed to farming and agribusiness by 

introducing up-to-date scientific knowledge. 

Recent advances in molecular biology made it possible 

to acquire extensive knowledge of rice genetics such as the 

fundamental factors of heredity and the nucleotide sequence 

of the genome. In Japan, the Ministry of Agriculture, Forestry, 

and Fisheries (MAFF) started the rice genome project in 1991 

with the aim of analyzing the structure of the rice genome, 

understanding the rice plant based on its genetic composition, 

and using the information for innovative rice breeding strategies. 

From 1998, the project has served as a launching pad for 

sequencing the entire genome through international coordination 

and elucidating its function based on this information both 

genetically and reverse-genetically (NIAS 2004a, Sasaki 1998). 

This paper focuses on the rice genome sequencing efforts, with 

emphasis on the analysis and utility of the genome sequence in 

the future. 

How to sequence and decode the information of heredity 

The term “genome” refers to the total genetic information that 

defines a species. The size of the genome corresponds to the 

amount of nucleotide that comprises half of the chromosome 

pairs or haploid. This can be experimentally measured by flow 

cytometry and expressed as a nuclear C-value. The genome 

size of rice with 12 chromosomes has been estimated to be 

430 Mb (Royal Botanic Gardens 2004). Rice has the smallest 

genome size when compared with other major cereal crops 

such as maize, with 2,500 Mb, and wheat, with 16,000 Mb. 

However, even with the most advanced method of nucleotide 

sequencing, the size of rice is not an easy target for complete 

sequencing with high accuracy. In the case of rice, it is also 

important to assign the location of the sequence to its genomic 

position so that the information can be directly useful to breeders. 

This approach in sequencing analysis facilitates the identification 

of gene function by forward and reverse genetics and 

comparison of specific phenotypes among rice varieties based 

on the difference in nucleotide sequence. This is extremely 

important considering that there are about 120,000 rice varieties 

worldwide and each variety is characteristically defined by 

the nucleotide sequences. Thus, an accurate and exact standard 

genome sequence is indispensable for subsequent goals 

in rice genomics. 

The rice genome sequencing effort has been an international 

collaboration since 1998. The International Rice Genome 

Sequencing Project (IRGSP) is currently composed of 

10 participating countries, led by Japan (Sasaki and Burr 2000, 

NIAS 2004b). A clone-by-clone sequencing strategy using a 

standard japonica rice variety, Nipponbare, has been adopted. 

Although each country’s contribution differs, each group has 

made significant progress in decoding the rice genome sequence. 

Japan has contributed in sequencing 55% of the entire 

genome, or an equivalent of 220 Mb genome sequence. In 

December 2002, the IRGSP finished the high-quality draft sequence 

of the entire genome, which is now available in public 

databases. It is expected that by December 2004 the entire 

Nipponbare genome, covered by nearly 3,400 BAC (bacterial 

artificial chromosome) and PAC (P1-derived artificial chromosome) 

clones, will be sequenced with 99.99% accuracy 

without any sequence gap. The genomic positional assignment 

of tiled BAC and PAC clones has been established based on 


genetically mapped DNA markers prior to the launching of 

the genome sequencing project so that every genome sequence 

can be associated with a specific position on the genetic map. 

Several genomic regions with highly repetitive sequence or 

unstable structure in host bacteria have not yet been completely 

sequenced. So far, only the centromeres of three chromosomes 

have been completely covered by BAC or PAC clones, with 

the centromeres of chromosomes 8 and 4, respectively, being 

completely sequenced (Wu et al 2004, Zhang et al 2004). 

Characteristics of the rice genome 

The most frequently asked and reasonable question refers to 

the total number of genes within the rice genome. To precisely 

predict the gene composition of rice, we need sufficient statistical 

data to fit the parameters used by prediction programs 

such as those constructed based on a hidden Markov model. 

The most useful evidence for gene prediction is the full-length 

cDNA sequence with information on the start and stop codons 

without interruption, and which, in correspondence to the genome 

sequence, can provide statistical information on nucleotide 

sequence at splice sites. Several prediction programs have 

been developed using various plant genome sequences. For 

rice, the FGENESH prediction program has been used for the 

analysis of the total number of genes and modeling of gene 

structure. Excluding transcripts of transposable elements, a total 

of 45,000 protein-coding sequences have been predicted at an 

average density of one gene per 8,200 bp. Chromosomes 1 

and 3 appear to be the most gene-rich chromosomes, whereas 

chromosome 12 is relatively gene-poor. 

The entire genome sequence has revealed many structural 

features of the rice genome. Tandem gene duplications 

have been widely observed among the 12 chromosomes. A dot 

matrix plot of all the predicted genes with significant homology 

to a given gene showed a clear diagonal line, indicating 

that a significant number of genes are duplicated and arrayed 

in tandem (Sasaki et al 2002). Although there is not yet enough 

evidence of actual expression of each member of these genes, 

several duplicated genes have been captured as cDNAs or ESTs 

in many cases. Such a high degree of gene multiplication as 

observed in the rice genome has not been observed in any species 

so far sequenced, including the model dicot plant 

Arabidopsis thaliana. For the human genome, the number of 

genes is lower than the total number of transcripts because of 

frequent alternative splicing for one gene. This mechanism may 

be an adaptive mechanism to increase the repertoire of gene 

product corresponding to a different signal for transcription to 

the same gene. In contrast, a high degree of gene multiplication 

in the rice genome may be essential to prepare various 

gene sets for a similar transcript, as shown by the low frequency 

of alternative splicing in the rice genome using the 

information from rice full-length cDNAs. Further study on 

promoter analysis must elucidate how and when each of the 

multiplied similar genes is expressed to function and in which 

tissue of the rice plant. 

How to use sequence information in the future 

Currently, forward and reverse genetics methods have been 

widely used to identify specific genes corresponding to phenotypes. 

The reverse genetics method facilitates the mutation 

of genes to analyze the resulting phenotype. It is systematically 

applicable to a genome-wide survey of disrupted position 

by referring to the genome sequence. Some of the most 

commonly used biological factors for gene disruption include 

T-DNA from Agrobacterium tumefaciens and Ac/Ds from 

maize. In Arabidopsis, many mutants have been generated and 

the corresponding genes have been identified using this method. 

Both factors are being widely used for rice, mainly in countries 

such as Korea, China, France, and the United States. 

However, mutated rice plants using these factors are considered 

as recombinant plants. Because of strict regulations in 

Japan concerning transgenic plants, an alternative biological 

factor based on endogenous retrotransposon Tos17 is widely 

used. Tos17 is activated only by a stress such as tissue culture 

conditions for 3–5 months and the Tos17 transposed into 5–10 

new loci is stably harbored there (Hirochika 2001). If this transposed 

Tos17 occasionally disrupts the open reading frame, regenerated 

rice plants will show mutation in the phenotype. Rice 

plants with transposed Tos17 are not recombinant plants and 

can be cultivated as normal plants. This makes Tos17 a very 

efficient tool for gene-disruption studies in Japan. For screening 

of the disrupted sequence by Tos17, sequences flanking 

inserted Tos17 are collected in a database. Currently, nearly 

42,000 flanking sequences of 4,300 mutant lines have been 

analyzed (Miyao et al 2003). These sequences have been assigned 

to the genome sequence and it has been shown that 

Tos17 prefers disruption of the exon region instead of intron 

and intergenic regions. However, the distribution of disrupted 

regions is not even within the genome. This result suggests the 

necessity of using other complementary factors such as T-DNA 

and Ac/Ds for complete disruption of all rice genes. 

Forward genetics provides a procedure to identify and 

select individuals that possess a phenotype of interest. It normally 

depends on mutants of phenotype, which are generated 

intentionally by physical factors such as irradiation by gamma 

ray or fast neutrons; chemical agents such as N-methyl-Nnitrosourea; 

biological factors such as Tos17, T-DNA, and Ac/ 

Ds; or unintentionally by unknown spontaneous factors. The 

major principle that led to gene identification using these mutants 

is based on precise genetic analysis by DNA markers, 

which have been developed from the genome sequence information. 

As a result, the map-based cloning strategy has advanced 

significantly because of the rapid progress of genome 

sequencing, thereby facilitating accurate detection of polymorphism 

even at a single nucleotide level. This strategy has led 

to the identification of a particular gene corresponding to mutation 

caused by a single gene. Many important genes such as 

genes with resistance to rice bacterial blight, Xa1 and Xa21, 

genes with resistance to rice blast disease, Pib and Pita, and 

genes corresponding to grass height, d1, gid2, slr1, and sd1, 

have been identified by map-based cloning. However, almost 

Session 2: Structure and function of the rice genome 67

all of the traits that are important for rice breeding are not 

controlled by a single gene. Instead, they are controlled by 

many genes known as quantitative trait loci or QTLs, and the 

phenotypes result from a complex mechanism of interaction 

of the gene products of these genes. QTL mapping is performed 

by the interval mapping method and the location of the QTL is 

shown as a score of probability, LOD. Recent progress in genetic 

analysis of QTLs using chromosomal substitution lines 

(CSL) evaluated by DNA markers has made it possible to isolate 

each QTL as a single Mendelian factor and to further identify 

the corresponding gene by map-based cloning. This strategy 

has been applied for the identification of genes of QTLs 

for rice heading date or flowering time. Among 14 QTLs of 

heading date in rice, four have been successfully identified 

(Yano et al 2001). 

Many biological processes that have not yet been characterized 

at the molecular level can now be reanalyzed using 

the genome sequence. It has been widely presumed that continuous 

variation in phenotypes can be attributed to the variation 

of corresponding alleles. However, no molecular tools are 

available to clarify the differences that could be useful for developing 

novel strategies in improving specific phenotypes. 

At present, the identification of genes involved in QTLs involves 

a thorough understanding of the difference of nucleotide 

sequence conferring the difference in degree of function 

as well as obtaining the fundamental information on the gene 

network expressed as a complex trait. The former can be attributed 

to the contribution of natural variation of nucleotide 

sequences in the corresponding member genes involved in 

QTLs. Natural variations derived from wild relatives of cultivated 

rice, which have remained in the genome after subsequent 

selection during the breeding process, are indispensable 

resources to identify more favorable traits for rice improvement. 

The gene network associated with the expression of the 

QTL must be elucidated to clarify the physiological and biochemical 

processes regarding the target QTL. If we could find 

out a meaningful nucleotide difference for phenotype, this information 

would be indispensable for marker-assisted selection 

and screening of germplasm, including elite breeding lines. 

The standard rice genome sequence will provide the major 

source of information to attain this goal. 

Future use of the genome sequence will also rely on an 

efficient informatics infrastructure. Rice genomics data are currently 

increasing exponentially day by day because of the increase 

in the number of rice genomics researchers encouraged 

by the availability of an accurate rice genome sequence. This 

means that rice genome databases should evolve from a storehouse 

for thousands of bases or amino acids to a “workhouse” 

attaching substantial information to the sequence. These databases 

should therefore provide the framework to allow 

postsequencing analysis such as identifying genes and predicting 

the proteins they encode, determining when and where the 

genes are expressed, and how they interact under each specified 

condition. To satisfy these demands, it is necessary to link 

the resources of various types of information. In addition, this 

information must be closely linked to available genetic resources 

for further research. 

Comparative genomics 

The complete genome sequence will be useful for comparing 

genomes at both the intraspecific and interspecific level. Colinearity 

of genes has been widely studied between the genomes 

of japonica rice and indica rice, among the different species of 

Oryza, and also among different cereals representing different 

genera. To understand the cultivated rice plant, O. sativa, comparison 

with its wild relatives such as O. rufipogon or O. barthii 

is indispensable. Comparative analysis of the genome sequence 

such as with single nucleotide polymorphism and insertion/ 

deletion, the number and types of inserted transposable elements, 

and the degree of duplication of specified genes will 

provide important insights into why and how currently cultivated 

rice species arose from their wild relatives. Among the 

wild relatives, tetraploid Oryza species exist. The cultivated 

rice O. sativa is a nearly pure diploid, although several genomic 

regions have been revealed to be duplicated. Polyploidy 

is often observed in cultivated plant species because it is 

advantageous in getting more yield than the diploid. Bread 

wheat is the most famous case and sugarcane is also known 

for its high polyploidy. Until now, no molecular information 

on the factor inducing polyploidy is available. Comparative 

analysis based on bioinformatic analysis of genome sequence 

data on several Oryza species will provide fundamental information 

on how polyploidy of a genome could be introduced 

under cultivation pressure. If we can control the ploidy of the 

rice genome, an increase in yield might be attained much more 

easily than pinpoint modification of target sequences. 

Conclusions 

Rice genome analysis is a relatively young science similar to 

the genome analysis of other species such as humans. However, 

because of rapid progress and technological innovations, 

a lot of new molecular resources and tools such as molecular 

markers for precise genetic mapping and a high-quality genome 

sequence for comprehensive molecular analysis of genome 

structure and function have already become available 

for rice. What is needed next is a collection of genetic resources 

for applying the molecular tools for further identification of 

useful alleles. With the completion of a high-quality genome 

sequence of rice, it would be possible to explore the diversity 

among different rice varieties. Diversity within the genomic 

component is very important for finding new alleles for breeding 

novel varieties and for understanding the rice plant based 

on differences in nucleotide sequences among Oryza species 

and among the members of the Gramineae. After the divergence 

from the common ancestral species, about 60 million 

years ago, Oryza species as well as Zea and Hordeum species 

have evolved as unique cereal crops. A thorough understanding 

of the similarities and differences in genome structure, gene 


structure, and their functional organization could provide the 

key to maintaining a close balance between cereal production 

and world food security. 

References 

Hirochika H. 2001. Contribution of the Tos17 retrotransposon to 

rice functional genomics. Curr. Opin. Plant Biol. 4:118-122. 

IRRI (International Rice Research Institute). 2004. 

www.knowledgebank.irri.org/wildRiceTaxonomy/default.htm. 

Miyao A, Tanaka K, Murata K, et al. 2003. Target site specificity of 

the Tos17 retrotransposon shows a preference for insertion 

within genes and against insertion in retrotransposon-rich regions 

of the genome. Plant Cell 15:1771-1780. 

NIAS (National Institute of Agrobiological Sciences). 2004a. 

www.nias.affrc.go.jp/project/inegenome_e/zenenki/ 

zenenki_outlook_e.htm. 

NIAS (National Institute of Agrobiological Sciences). 2004b. http:/ 

/rgp.dna.affrc.go.jp/IRGSP/. 

Royal Botanic Gardens. 2004. www.rbgkew.org.uk/cval/ 

introduction.html. 

Sasaki T. 1998. The rice genome project in Japan. Proc. Natl. Acad. 

Sci. USA 95:2027-2028. 

Sasaki T, Burr B. 2000. International Rice Genome Sequencing 

Project: the effort to completely sequence the rice genome. 

Curr. Opin. Plant Biol. 3:138-141. 

Sasaki T, Matsumoto T, Yamamoto K, et al. 2002. The genome sequence 

and structure of rice chromosome 1. Nature 420:312- 

316. 

Wu J, Yamagata H, Hayashi-Tsugane M, et al. 2004. Composition 

and structure of the centromeric region of rice chromosome 

8. Plant Cell 16:967-976. 

Yano M, Kojima S, Takahashi Y, Lin H-X, Sasaki T. 2001. Genetic 

control of flowering time in rice, a short-day plant. Plant 

Physiol. 127:1425-1429. 

Zhang Y, Huang Y, Zhang L, et al. 2004. Structural features of the 

rice chromosome 4 centromere. Nucleic Acids Res. 32:2023- 

2030. 

Notes 

Author’s address: Rice Genome Research Program, NIAS/STAFF, 

1-2, Kannondai 2-chome, Tsukuba, Ibaraki 305-8602, Japan, 

e-mail: tsasaki@nias.affrc.go.jp. 

Acknowledgment: The program for rice genome research (Rice Genome 

Project GS-1101, 1102, 1103, 1201, 1301, and 1302) is 

supported by the Ministry of Agriculture, Forestry, and Fisheries 

(MAFF) of Japan. The Rice Genome Research Program 

(RGP) is a joint collaboration of the National Institute of 

Agrobiological Sciences (NIAS) and the Institute of the Society 

for Techno-innovation of Agriculture, Forestry, and Fisheries 

(STAFF). 

Exploitation and use of naturally occurring allelic variations 

in rice 

Masahiro Yano, Yasunori Nonoue, Tsuyu Ando, Ayahiko Shomura, Takehiko Shimizu, Izumi Kono, Saeko Konishi, 

Utako Yamanouchi, Tadamasa Ueda, Shin-ichi Yamamoto, and Takeshi Izawa 

Over the last decade, a tremendous amount of progress has 

been made in rice genome analysis, including in whole-genome 

sequencing (Sasaki 2003). This progress has provided powerful 

tools—DNA markers—for plant genetics and breeding. 

DNA markers have been used in the genetic analysis of agronomically 

important traits, such as disease and pest resistance, 

plant height, and grain quality. Many phenotypic traits of economic 

interest are controlled by naturally occurring allelic 

variations and environmental conditions. These traits often 

show complex and quantitative inheritance. Recent progress 

in rice genomics has had a large impact on our understanding 

of such complex traits. DNA markers are used in the genetic 

analysis of quantitative traits, such as heading date, environmental 

stress tolerance, and yield-related traits. This markerassisted 

approach has also made it possible to exploit hidden 

genetic factors contributing to the improvement of traits with 

agronomic importance and to clone genes involved in such 

complex traits at the molecular level (Yano 2001). Tightly 

linked DNA markers or causal genes can be used in markerassisted 

selection (MAS) in breeding programs. The advent of 

DNA markers has allowed us to practice MAS in rice breed- 

ing. This paper describes our recent progress in the markerassisted 

genetic and molecular dissection of several complex 

traits and our efforts to develop plant materials that are, or will 

be, valuable in the analysis of complex traits in rice. 

Genetic and molecular dissection of complex traits in rice 

Flowering time (heading date) 

Flowering time is a major determinant of the regional and seasonal 

adaptation of rice cultivars and is one of the important 

traits in breeding programs. Analysis of natural allelic variation 

has contributed significantly to the exploration of genes 

involved in flowering time in rice. In quantitative trait locus 

(QTL) analysis, 15 QTLs for heading date have been detected 

genetically by using different types of progeny derived from a 

cross between Nipponbare and Kasalath (Yano et al 2001). 

Near-isogenic lines (NILs) for each QTL have also been developed 

by successive backcrossing and MAS. Consequently, 

eight QTLs have been mapped as single Mendelian factors by 

using advanced backcross progeny (Yano et al 2001). Furthermore, 

four QTLs (Hd1, Hd3a, Hd5, and Hd6) have been cloned 


y map-based strategies (Yano et al 2000, Takahashi et al 2001, 

Kojima et al 2002, Yamanouchi et al, unpublished data). Recently, 

two additional QTLs, Ehd1 (Doi et al 2004) and Lhd4 

(N. Kitazawa et al, unpublished data), have been identified by 

using different cross combinations. The series of studies on 

heading date clearly demonstrated that complex traits could 

be dissected not only genetically, but also at a molecular level 

(Izawa et al 2003). 

Seed dormancy and longevity 

Seed dormancy is an important trait because it affects resistance 

to preharvest sprouting. Seed dormancy has also been 

analyzed by marker-assisted genetic approaches. Five putative 

QTLs affecting seed dormancy have been detected on 

chromosomes 3, 5, 7 (2 regions), and 8 by using backcross 

inbred lines (BILs) derived from a backcross of Nipponbare/ 

Kasalath//Nipponbare (Lin et al 1998). QTLs for seed dormancy 

(Sdr1) and heading date (Hd8) have been mapped to 

approximately the same region on chromosome 3 (Lin et al 

1998). To clarify whether Sdr1 and Hd8 could be dissected 

genetically, we carried out fine mapping by using advanced 

backcross progeny. As a result, Sdr1 and Hd8 were mapped to 

different loci. We thus successfully dissected two tightly linked 

QTLs by a marker-based approach (Takeuchi et al 2003). In 

addition, QTLs controlling seed longevity were identified by 

using 98 BILs derived from a cross between japonica cultivar 

Nipponbare and indica cultivar Kasalath. Three putative QTLs 

(qLG-2, qLG-4, and qLG-9) were detected on chromosomes 

2, 4, and 9, respectively (Miura et al 2002). Kasalath alleles 

increased seed longevity in these QTLs. Among the QTLs detected, 

qLG-9 has exhibited large effects on longevity. 

Cool-temperature tolerance 

Cool temperatures at the booting stage often cause an increase 

in the number of aborted microspores, eventually resulting in 

decreased rice yield. To identify the chromosomal regions controlling 

cool-temperature tolerance at the booting stage in rice, 

we performed QTL analysis with doubled-haploid lines derived 

from crosses between two temperate japonica cultivars, 

Akihikari (moderately cool-temperature sensitive) and 

Koshihikari (cool-temperature tolerant). Three QTLs were 

mapped on each of chromosomes 1, 7, and 11. For each QTL, 

alleles from Koshihikari increased the degree of cool-temperature 

tolerance (Takeuchi et al 2001). One major QTL, qCT-7, 

has been mapped as a single Mendelian factor, and positional 

cloning of qCT-7 has been progressing (Ueda et al, unpublished 

data). 

Exploitation of natural allelic variations 

Plant materials for the exploitation of genes 

with agricultural importance 

Sequencing of the entire rice genome has progressed greatly 

(Sasaki 2003). This achievement has contributed to the molecular 

analysis of complex traits, as mentioned above. Although 

sequence information and molecular tools have already 

been rapidly accumulated, plant materials for genetic analysis 

are still being developed because material development usually 

requires a long time and much labor. Thus, a lack of genetic 

materials would limit our comprehensive understanding 

of quantitative traits. 

In general, primary mapping populations, such as F 2 and 

recombinant inbred lines (RILs), are used for the detection of 

QTLs. QTLs with relatively large effects can be detected in 

the genetic analysis of primary mapping populations. However, 

some QTLs with minor effects and those with epistatic 

interactions with other loci might not be detected in QTL analysis 

(Yano and Sasaki 1997). These QTLs can be detected by 

using advanced backcross progeny. Even though QTLs of interest 

can be identified by using existing primary mapping 

populations, further development of NILs is required for fine 

mapping and cloning of QTLs. Development of these materials 

is laborious and time-consuming. This problem has prevented 

many researchers from performing map-based cloning 

of QTLs as a more general strategy in plant molecular genetics. 

To solve these problems, novel mapping populations— 

chromosome segment substitution lines (CSSLs) or introgression 

lines (ILs)—have been developed in rice (Fig. 1). In the 

CSSLs, a particular chromosome segment from a donor line is 

substituted in the genetic background of the recurrent line. The 

substituted segments cover all chromosomes in a whole set of 

lines (Fig. 1). These materials allow detailed and reliable QTL 

analyses. CSSLs have been developed on the basis of several 

cross-combinations and will be developed from crosses within 

the cultivated species, Oryza sativa (Table 1). 

The use of wild relatives as donor parents would be a 

more powerful way to exploit a wide range of allelic variations 

because they are adapted to specific environmental conditions, 

and variations in genes give advantages for adaptation. 

Many wild relatives exist in the O. sativa complex: seven 

species have been differentiated and have unique geographic 

distributions under special environmental conditions. Accessions 

of these species can be crossed with each other, and progeny 

can be produced by backcrossing. Thus, it is feasible to 

develop ILs covering whole rice chromosomes. These kinds 

of material are already being developed. Some accessions of 

wild relatives from O. glumaepatula, O. glaberrima, and O. 

meridionalis have been used as donor parental lines to develop 

ILs (Sobrizal et al 1999, Yoshimura A, personal communication). 

Potential advantages of CSSLs in QTL analysis 

The potential power of CSSLs and ILs in genetic analysis has 

been demonstrated (Kubo et al 2002, Ebitani et al 2005). For 

example, CSSLs or ILs can be used in genetic analysis to associate 

QTLs with particular chromosome regions and to 

quickly develop NILs of target regions containing QTLs of 

interest. In general, when an association is detected between a 

chromosomal region and a trait, it is often difficult to validate 

the QTLs, especially those representing very small genetic effects. 

In such a case, NILs are required to analyze genetic ef- 


A 

Koshihikari × Kasalath 

F 

BC 1 F 3 

SBC 2 F 1 

SBC F 

CSSL 

no. chr. 1 

CSSLs 

SBC 4 F 2 

× Koshihikari 

SL-201 

SL-202 

SL-203 

1 

SL-204 

SL-205 

SL-206 

BC 1 F 1 

SL-207 

SL-208 

SL-209 

SL-210 

SL-211 


SL-212 

SL-213 

SL-214 

SL-215 

SL-216 

SL-217 


SL-218 

SL-219 

SL-220 


SL-221 

SL-222 

SL-223 

3 1 

SL-224 

SL-225 

SBC 3 F 2 

SL-226 

SBC 4 F SL-227 

1 

SL-228 

SL-229 

SL-230 

CSSLs 

B 

SL-232 

SL-233 

SL-234 

SL-235 

SL-236 

SL-237 

SL-238 

SL-239 

chr. 2 

chr. 3 

chr. 4 chr. 6 

chr. 5 

chr. 7 

chr. 8 

chr. 10 chr. 12 

chr. 9 chr. 11 

Fig. 1. Chromosome segment substitution lines (CSSLs) derived from a cross between Koshihikari and Kasalath. (A): Flowchart of 

the process of selection of CSSLs. (B): Graphical representation of genotypes of the CSSLs. The black segments indicate regions 

homozygous for Kasalath alleles; the white segments indicate regions homozygous for Koshihikari alleles; the hatched segments 

indicate heterozygous regions. Genotype classes of the 129 RFLP markers for the CSSLs can be obtained at the Web site of the 

Rice Genome Resource Center (www.rgrc.dna.affrc.go.jp/ine39.html). 

Table 1. Chromosome segment substitution lines (CSSLs) derived from crosses within the cultivated species 

Oryza sativa. 

Cross combination 

Population/ Number of Genotype data Reference 

Recurrent Donor generation lines 

Nipponbare (J) a Kasalath (I) a CSSL 54 125 RFLPs RGRC (2003) 

Asominori (J) IR24 (I) CSSL 91 85 RFLPs Kubo et al (2002) 

IR24 (I) Asominori (J) CSSL 70 87 RFLPs Kubo et al (2002) 

Sasanishiki (J) Habataki (I) CSSL 45 – – 

Koshihikari (J) Kasalath (I) CSSL 39 129 RFLPs Ebitani et al (2005) 

Koshihikari (J) Nona Bokra BC 3 F 3 /BC 4 F 2 – – – 

Koshihikari (J) Nipponbare (J) BC 3 F 1 – – – 

Nipponbare (J) Koshihikari (J) BC 3 F 1 – – – 

Koshihikari (J) IR64 (I) BC 1 F 1 – – – 

IR64 (I) Koshihikari (J) BC 1 F 1 – – – 

a J and I indicate japonica and indica cultivars, respectively. 

fects in detail. Because CSSLs normally have one chromosomal 

region substituted, they can be used as NILs themselves or as 

starting material to develop NILs. 

Development of NILs for QTLs is necessary for verification 

and characterization of the QTLs detected (Tanksley 

1993, Yano and Sasaki 1997). Once a QTL for a particular trait 

is detected, a CSSL can be used as a NIL itself or as a base 

material for NIL development. As only one chromosome fragment 

has been substituted, only one additional backcross will 

be required for NIL development. Such NILs allow us to combine 

two or three QTLs in one genetic background to clarify 

the epistatic interaction among the QTLs. 

Use of exploited natural allelic variations in rice 

breeding 

The advent of DNA marker technology has made it possible to 

establish new breeding strategies, such as marker-assisted 

breeding (Peleman and van der Voort 2003). That strategy requires 

comprehensive dissection and understanding of complex 

traits of interest. To that end, even if a large number of 

DNA markers are available, novel plant materials (mapping 

populations) such as ILs or CSSLs are necessary for markerassisted 

strategies to be used in practical breeding (Zamir 2001, 

Peleman and van der Voort 2003). As mentioned above, novel 

mapping populations and plant materials have already been 

developed. To enhance the exploitation and use of novel alleles 

for rice breeding, these CSSLs should be grown under 


different environmental conditions, and comprehensive 

phenotyping should be performed on traits of economic importance, 

such as yield, cooking and eating quality, disease 

and pest resistance, and environmental-stress tolerance. 

Once we have found favorable alleles in donor varieties, 

those genes can be introduced systematically and rapidly 

into elite varieties in a given region, by the aid of MAS. Furthermore, 

by using a variety of NILs and DNA markers, we 

can introduce genes derived from different donor cultivars into 

elite rice cultivars (trait pyramiding). This strategy has not yet 

been used in phenotype-based selection in rice breeding. 

References 

Doi K, Izawa T, Fuse T, Yamanouchi U, Kubo T, Shimatani Z, Yano 

M, Yoshimura A. 2004. Ehd1, a B-type response regulator in 

rice, confers short-day promotion of flowering and controls 

FT-like gene expression independently of Hd1. Genes Dev. 

18:926-936. 

Ebitani T, Takeuchi Y, Nonoue Y, Yamamoto T, Takeuchi K, Yano 

M. 2005. Chromosome segment substitution lines carrying 

overlapping chromosome segments of indica rice cultivar 

‘Kasalath’ in a genetic background of japonica elite cultivar 

‘Koshihikari’. Breed. Sci. (In press.) 

Izawa T, Takahashi Y, Yano M. 2003. Comparative biology comes to 

bloom: genomic and genetic comparison of flowering pathways 

in rice and Arabidopsis. Curr. Opin. Plant Biol. 6:113- 

120. 

Kojima S, Takahashi Y, Kobayashi Y, Monna L, Sasaki T, Araki T, 

Yano M. 2002. Hd3a, a rice ortholog of the Arabidopsis FT 

gene, promotes transition to flowering downstream of Hd1 

under short-day condition. Plant Cell Physiol. 43:1096-1105. 

Kubo T., Aida Y, Nakamura K, Tsunematsu H, Doi K, Yoshimura A. 

2002. Reciprocal chromosome segment substitution series 

derived from Japonica and Indica cross of rice (Oryza sativa 

L.). Breed. Sci. 52:319-325. 

Lin SY, Sasaki T, Yano M. 1998. Mapping quantitative trait loci 

controlling seed dormancy and heading date in rice, Oryza 

sativa L., using backcross inbred lines. Theor. Appl. Genet. 

96:997-1003. 

Miura K, Lin SY, Yano M, Nagamine T. 2002. Mapping quantitative 

trait loci controlling seed longevity in rice (Oryza sativa L.). 


Peleman JD, van der Voort JR. 2003 Breeding by design. Trends 

Plant Sci. 8:330-334. 

RGRC (Rice Genome Resource Center). 2003. Materials for genetic 

analysis: Nipponbare/Kasalath chromosome segment substitution 

lines (CSSL), 54 lines. www.rgrc.dna.affrc.go.jp/ 

ine54.html. 

Sasaki T. 2003. Rice genome analysis: understanding the genetic 

secrets of the rice plant. Breed. Sci. 53:281-289. 

Sobrizal, Ikeda K, Sanchez PL, Doi K, Angeles ER, Khush GS, 

Yoshimura A. 1999. Development of Oryza glumaepatula 

introgression lines in rice, O. sativa L. Rice Genet. Newsl. 

16:107-108. 

Takahashi Y, Shomura A, Sasaki T, Yano M. 2001. Hd6, a rice quantitative 

trait locus involved in photoperiod sensitivity, encodes 

the alpha subunit of protein kinase CK2. Proc. Natl. Acad. 

Sci. USA 98:7922-7927. 

Takeuchi Y, Hayasaka H, Chiba B, Tanaka I, Shimano T, Yamagishi 

M, Nagano K, Sasaki T, Yano M. 2001. Mapping quantitative 

trait loci controlling cool-temperature tolerance at booting 

stage in temperate japonica rice. Breed. Sci. 51:191-197. 

Takeuchi Y, Lin SY, Sasaki T, Yano M. 2003. Fine-scale linkage 

mapping enables dissection of closely linked quantitative trait 

loci for seed dormancy and heading in rice. Theor. Appl. Genet. 

107:1174-1180. 

Tanksley SD. 1993. Mapping polygenes. Annu. Rev. Genet. 27:205- 

233. 

Yano M. 2001. Genetic and molecular dissection of naturally occurring 

variation. Curr. Opin. Plant Biol. 4:130-135. 

Yano M, Katayose Y, Ashikari M, Yamanouchi U, Monna L, Fuse T, 

Baba T, Yamamoto K, Umehara Y, Nagamura Y, Sasaki T. 

2000. Hd1, a major photoperiod sensitivity quantitative trait 

locus in rice, is closely related to the Arabidopsis flowering 

time gene CONSTANS. Plant Cell 12:2473-2483. 

Yano M, Kojima S, Takahashi Y, Lin HX, Sasaki T. 2001. Genetic 

control of flowering time in rice, a short-day plant. Plant 

Physiol. 127:1425-1429. 

Yano M, Sasaki T. 1997. Genetic and molecular dissection of quantitative 

traits in rice. Plant Mol. Biol. 35:145-153. 

Zamir D. 2001. Improving plant breeding with exotic genetic libraries. 

Nature Rev. Genet. 2:983-989. 

Notes 

Authors’ addresses: Masahiro Yano, Utako Yamanouchi, Tadamasa 

Ueda, Shin-ichi Yamamoto, and Takeshi Izawa, National Institute 

of Agrobiological Sciences, Tsukuba, Ibaraki 305-8602, 

Japan; Yasunori Nonoue, Tsuyu Ando, Ayahiko Shomura, 

Takehiko Shimizu, Izumi Kono, and Saeko Konishi, Institute 

of the Society for Techno-Innovation of Agriculture, Forestry, 

and Fisheries, Tsukuba, Ibaraki 305-0854, Japan, e-mail: 

myano@nias.affrc.go.jp. 


Allelic and functional diversity of stress-tolerance genes 

in rice 

Hei Leung, Hehe Wang, Jianli Wu, Ma. Elizabeth Naredo, Marietta Baraoidan, Alicia Bordeos, Suzette Madamba, Gay Carrillo, 

Jatinder Sangha, Zenna Negussie, Jill Cairns, Bin Liu, Yolanda Chen, Darshan Brar, Il Ryong Choi, Cassiana Vera Cruz, 

Renee Lafitte, Luca Comai, and Kenneth L. McNally 

Most natural genetic variation in organisms is represented by 

simple base changes or small insertions/deletions (single nucleotide 

polymorphisms, SNPs, or single feature polymorphisms, 

SFPs). It is reasonable to assume that SNPs in both structural 

and regulatory genes are key drivers of phenotypic variation, 

and that they collectively account for many of the varietal differences 

in crop germplasm. SNP analysis has been used successfully 

in dissecting complex traits in other model organisms 

with practical implications (see recent examples in understanding 

the genetic basis of immunity in Drosophila, 

Lazzaro et al 2004, tracking asthma-related traits and cholesterol 

levels in humans, Laitinen et al 2004, Cohen et al 2004). 

Thus, a whole-genome SNP data set derived from diverse 

germplasm (i.e., SNP haplotypes encompassing the natural 

variation at any particular segment of the chromosome) has 

considerable potential for crop improvement. 

SNP discovery is important for breeding in two fundamental 

ways. First, it reveals DNA variation among varieties, 

thus providing the tools for selection in breeding programs. 

Second, SNPs provide the “anchor” to relate other forms of 

polymorphisms—biochemical, metabolic, physiological, and 

phenotypic performance (Fig. 1). We have taken both forward 

and reverse genetics approaches to identify induced and natural 

variants exhibiting a differential response to biotic and abiotic 

stresses. Three sources of genetic materials are used: mutants, 

germplasm from a gene bank collection, and advanced 

breeding lines. We select a set of candidate genes involved in 

stress tolerance based on convergent evidence of cross-species 

sequence inference, QTL position, and expression phenotypes. 

As a test case, a good level of blast resistance in elite 

breeding lines has been achieved using this strategy. We envision 

a systematic approach of associating phenotypes with 

genotypes through a broad sampling of induced and natural 

variation, which can reveal the functional relationships among 

regulatory and effector genes, and provide the needed information 

for maximizing the use of allelic variation in breeding. 

Phenotypic screen of induced variation 

A large collection of induced mutations has been generated by 

institutions and laboratories around the world (Hirochika et al 

2004). We have focused on the use of chemical and irradiation-induced 

mutations in an indica rice variety, IR64. From 

screening the IR64 mutant collection (primarily deletions and 

point mutations), we have genetically defined mutants with an 

altered response to multiple diseases (blast, bacterial blight, 

Genetic variation 

SNP and indels 

Connecting genomic variation 

with phenotypes and performance 

Profiles of 

l Transcripts 

l Proteins 

l Metabolites 

Interactions 

Phenotypes 

Performance 

Fig. 1. Gene function discovery involves relating genetic variation 

to phenotypes. Available whole-genome sequence provides the 

anchor to relate gene expression polymorphism to genomic variation. 

and tungro viruses), brown planthopper, and drought. 

From these screening experiments, we observe that gainof-function 

mutations can be found at considerable frequencies. 

We identified gain-of-resistance to brown planthopper, 

which is apparently controlled by a partially dominant mutation 

(J. Sangha, IRRI, unpublished data). Another mutant showing 

gain-of-resistance to rice tungro spherical virus (RTSV) 

was conditioned by a recessive mutation (N. Zenna and I. Choi, 

IRRI, unpublished data). A putative mutant with enhanced tolerance 

of water stress under field conditions has been isolated 

(Botwright and Lafitte 2002, J. Cairns and R. Laffitte, IRRI, 

unpublished). Zeng et al (2004) found that the lesion mimic 

gene SPL11 encodes a putative E3 ligase that functions as a 

negative regulator of cell death and broad-spectrum host resistance. 

These results suggest a significant role for negatively 

regulated networks in stress tolerance. Thus, a systematic 

screening for an agronomically important phenotype could 

yield better than expected results. 

Reverse genetics to detect induced and natural variation 

Based on results from expression and mapping studies as well 

as published reports (e.g., Cooper et al 2003), we developed a 

panel of candidate genes potentially involved in stress tolerance. 

Currently, a set of 32 candidate genes is used as targets 

for TILLING (Till et al 2003). These genes are selected based 

on several criteria: (1) high-quality sequence information from 


indica and japonica rice, (2) implicated function based on sequence 

annotation, (3) functional role inferred from other species, 

(4) evidence that they are within QTL regions based on 

mapping studies, and/or (5) the region in which they occur 

undergoes allelic shifts under selection. For each gene, we apply 

TILLING to survey haplotypic and SNP variation in the 5′ 

regulatory and coding regions in mutants, O. sativa germplasm, 

and wild relatives. Simultaneously, we evaluate the phenotypes 

associated with these alleles in natural germplasm and derived 

breeding lines. 

Choice of the genetic material is a key factor for the 

success of the approach. We have chosen a set of 200–400 

accessions representative of diverse genetic origins from different 

selection cycles in our rice breeding programs. These 

are taken from the accessions genotyped with neutral simple 

sequence repeat (SSR) markers, randomly distributed along 

the genome, such that the genetic structure can be assessed 

and considered in statistical models to test association. This 

collection will also be used to optimize the choice of subsets 

of inbreds for the SNP discovery analyses, and identification 

of haplotype patterns. To achieve cost-effective genotyping, 

multiples of 96 accessions are chosen to fit a high-throughput 

format. Moreover, to have sufficient power for statistical tests 

of association between polymorphism and phenotypes, at least 

200 inbred lines should be evaluated for phenotypes under 

homogeneous conditions. 

Initial results have been obtained from using primers 

designed from the coding sequence of protein phosphatase 2a- 

4 (PP2a4) on cultivated germplasm and from DREB1 and trehalose 

6-phosphatase (TPS) using cultivated and wild 

germplasm. We have detected mismatches (putative SNPs) for 

these genes across a range of germplasm. Three putative SNP 

haplotypes were detected for PP2a4 in 22 Korean japonica 

and Tongil (indica/japonica hybrid) varieties contrasted to 

Nipponbare. One SNP was detected in Milyang30, Taebaek, 

Samgang, Sindongjin, and Milyang23, a second was found in 

Milyang30 and Milyang23, and a third occurred in Suwon345. 

Furthermore, a putative SNP was also detected between SHZ2 

and LTH, the two parental lines of a breeding population developed 

for disease-resistance studies. 

For TPS on 48 germplasm accessions, the most distinct 

haplotype occurred in the contrast between Kun Min Tsieh 

Hunan and IR64. A mismatch at the same position was also 

detected in the contrast between Kun Min Tsieh Hunan and 

Nipponbare. The TPS primers were also tested on the wild 

germplasm panel, but amplicons were obtained only for the 

AA-genome species. 

In tests of DREB1 on cultivated rice, three putative SNP 

haplotypes have been detected in 96 accessions. The SNP detected 

between IR64 and Nipponbare is common to indica/ 

japonica contrasts. A second haplotype occurred in the 

Mimidam and IR64 contrast, and a third was detected between 

Taichung Native 1 and Nipponbare. Interestingly, primers designed 

for the coding region of DREB1 produced amplicons 

from wild relatives as distant as the GG genome, but none for 

the distant HHJJ and HHKK genomes (E. Naredo and K. 

McNally, IRRI, unpublished data). Many putative SNPs can 

be detected in these distant contrasts, including an SNP that 

appears to be absent from most wild species and indica but 

present in japonica (Fig. 2). 

We have also tested the ability to detect mismatches in 

pools of germplasm. Pooling is an approach that is routinely 

used in TILLING and may be advantageous for EcoTILLING 

as a first-pass screen among lines that are considered closely 

related. Our results indicate that pools of up to 8-fold are possible, 

which will allow 768 accessions to be screened in a single 

run for preliminary analysis. By improving the throughput of 

the system, we will be in a position to identify variants in the 

large collection of germplasm (>100,000) accessions in the 

International Rice Genebank Collection held at IRRI. 

Disease resistance as a case study 

The recent work of Cohen et al (2004) showed that rare alleles 

in a few candidate genes can account for the complex phenotype 

of plasma cholesterol level in humans. In this case, strong 

association was established between SNP in the coding regions 

of three candidate genes and extreme phenotypes in two human 

populations. The study suggests that it is possible to associate 

SNP variants with phenotypes provided that opposing 

phenotypes are adequately described in one or more populations. 

We use disease resistance as a case study to test whether 

allelic variants can be selected that achieve the desired phenotypic 

performance in the field. In this case, we applied candidate 

defense-related genes as selection markers for blast resistance. 

Advanced lines with chromosomal segments carrying 

the positive alleles showed effective quantitative resistance in 

multiple locations over three years (Liu et al 2004). However, 

because each chromosomal segment carries many potential 

defense genes, confirmative evidence is needed. To gather more 

evidence to support that a candidate gene is responsible for 

phenotypic gain, we followed the introgression of specific alleles 

of members of the oxalate oxidase (germin-like-protein) 

gene family. Preliminary data suggest association of indels in 

the 5′ untranslated region of two members of the gene family 

with blast resistance observed in advanced breeding lines (G. 

Carrillo and C. Vera Cruz, IRRI, unpublished data). Work is in 

progress to isolate mutations in these candidate genes through 

a reverse genetics screen. We hope that an integration of genotypic 

and phenotypic data obtained from advanced germplasm 

and mutants could provide the means to put knowledge of gene 

function to use in practical breeding. 

References 

Botwright T, Lafitte R. 2002. Deletion mutants for screening of 

drought tolerance in rice. In: Progress toward developing resilient 

crops for drought-prone areas. Proceedings of an International 

Workshop. Los Baños (Philippines): International 

Rice Research Institute. p 114-115. www.irri.org/publications/ 

limited/pdfs/ResilientCrops.pdf. 


A 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

AA 

CCDC 

EE 

CC 

CC 

CCDC 

GG 

GG 

BBCC 

CC 

CC 

CC 

BB 

BBCC 

CC 

CC 

IIIJJ 

HHKK 

IR64 

NIP 

1 5 10 15 20 25 30 35 40 45 

B 

1 5 10 15 20 25 30 35 40 45 I/N 1 5 10 15 20 25 30 35 40 45 I/N 

C 

1 5 10 15 20 25 30 35 40 45 I/N 1 5 10 15 20 25 30 35 40 45 I/N 

Fig. 2. EcoTILLING of DREB1 on wild Oryza. DREB1 amplicons of a similar size were produced from 

47 wild Oryza species using primers designed from Nipponbare (A). Mixtures of these products 

were combined with either IR64 (left set of lanes) or Nipponbare (right set of lanes). The IRD700 

channel is shown in panel B and the IRD800 in C. Circles indicate clear differences present in 

both channels. Arrows indicate a mismatch present in japonica but absent from most wild and 

indica rice. 

Cohen JC, Kiss RS, Pertsemlidis A, Marcel YL, McPherson R, Hobbs 

HH. 2004. Multiple rare alleles contribute to low plasma levels 

of HDL cholesterol. Science 305:869-872. 

Cooper B, Clarke JD, Budworth P, Kreps J, Hutchison D, Park S, 

Guimil S, Dunn M, Luginbuhl P, Ellero C, Goff SA, 

Glazebrook J. 2003. A network of rice genes associated with 

stress response and seed development. Proc. Natl. Acad. Sci. 

USA 100:4945-4950. 

Hirochika H, Guiderdoni E, An G, Hsing Y, Eun MY, Han CD, 

Upadhyaya N, Ramachandran S, Zhang Q, Pereira A, 

Sundaresan V, Leung H. 2004. Rice mutant resources for gene 

discovery. Plant Mol. Biol. 54:325-334. 

Laitinen T, Polvi A, Rydman P, Vendelin J, Pulkkinen V, Salmikangas 

P, Mäkelä S, Rehn M, Pirskanen A, Rautanen A, Zucchelli M, 

Gullstén H, Leino M, Alenius H, Petäys T, Haahtela T, Laitinen 

A, Laprise C, Hudson TJ, Laitinen LA, Kere J. 2004. Characterization 

of a common susceptibility locus for asthma-related 

traits. Science 304:300-304. 

Lazzaro BP, Sceurman BK, Clark AG. 2004. Genetic basis of natural 

variation in D. melanogaster antibacterial immunity. Science 

303:1873-1876. 

Liu B, Zhang S, Zhu X, Yang Q, Wu S, Mei M, Mauleon R, Leach J, 

Mew T, Leung H. 2004. Candidate defense genes as predictors 

of quantitative blast resistance in rice. Mol. Plant-Microbe 

Interact. 17:1146-1152. 

Till BJ, Reynolds SH, Greene EA, Codomo CA, Enns LC, Johnson 

JE, Burtner AR, Young K, Taylor NE, Henikoff JG, Comai L, 

Henikoff S. 2003. Large-scale discovery of induced pointmutations 

with high-throughput TILLING. Genome Res. 

13:524-530. 

Zeng L, Qu S, Bordeos A, Yang C, Baraoidan M, Yan H, Xie Q, 

Nahm BH, Leung H, Wang GL. 2004. Spl11, a negative regulator 

of plant cell death and defense, encodes a U-Box/ARM 

repeat protein endowed with E3 ubiquitin ligase activity. Plant 

Cell 16:2795-2808. 

Notes 

Authors’ addresses: Division of Entomology and Plant Pathology, 

Genetic Resources Center, Division of Plant Breeding, Genetics, 

and Biochemistry, Division of Crop, Soil, and Water 

Sciences, International Rice Research Institute, DAPO Box 

7777, Metro Manila, Philippines; Department of Biology, University 

of Washington, Seattle, Washington, USA. 

Acknowledgments: This work is supported in part by the Swiss 

Agency for Development and Cooperation (SDC), Rockefeller 

Foundation, and Generation Challenge Program. 


Functional genomics by reverse genetics 

Gynheung An 

Sequencing of the rice genome is nearly complete. Therefore, 

one of the most challenging goals now is to examine the functioning 

of its large number of genes. To facilitate such an evaluation, 

several reverse genetics approaches have been developed. 

Random insertional mutagenesis by transposons or T- 

DNA has been the most widely used for large-scale analyses. 

This technique is not only efficient for identifying knockout 

mutants, but can also be employed for promoter trapping and 

activation tagging. The recent establishment of a large number 

of insertional mutants will accelerate these reverse genetics 

approaches for studying gene function in this model monocot 

species. 

Insertional mutagenesis by T-DNA or transposon 

An efficient means for rice transformation is via the 

Agrobacterium-mediated cocultivation method. This technique 

allows researchers to obtain a suitable T-DNA tagging population. 

For example, An’s group has generated approximately 

100,000 fertile rice lines tagged by T-DNA (Jeon et al 2000, 

Jeong et al 2002), while work by Zhang and colleagues has 

resulted in more than 30,000 T-DNA insertional lines (Wu et 

al 2003). Several other groups have independently produced 

T-DNA insertional mutant lines. Because the average copy 

number of T-DNA inserts per line is 1.4 to 2.0, more than 

300,000 T-DNA tags have now been generated in rice. These 

populations are large enough to find a knockout in a given 

gene at more than 90% probability, assuming that T-DNA is 

randomly inserted into a chromosome. 

The maize Ac/Ds system has been used for large-scale 

generation of gene tagging via the Agrobacterium-mediated 

gene delivery method (Chin et al 1999, Kolesnik et al 2004). 

The germinal transposition frequency of Ds is high. Study of 

the transposition pattern in siblings has revealed that 79% contain 

at least two different insertions, suggesting late transposition 

during rice development (Kolesnik et al 2004). Moreover, 

repetitive ratoon culturing causes new transposition, at about 

30% frequency, demonstrating that this strategy can produce a 

large mutant population (Chin et al 1999). 

Retroelements are a major component of plant genomes. 

At least 17% of the rice genome consists of retrotransposons. 

One such element in rice is Tos17. Although its copy number 

is low, this can increase to up to 30 copies during tissue culture. 

More than 47,000 Tos17 insertion lines have been produced 

to generate insertional mutants in rice (Hirochika 2001, 

Miyao et al 2003). 

Activation tagging 

Insertional mutagenesis usually generates recessive loss-offunction 

mutations, making them unsuitable for functional 

analysis of redundant genes. Jeong et al (2002) have demonstrated 

that the CaMV 35S enhancer element efficiently increases 

the expression of nearby genes from either their 5′ or 

3′ ends. Based on this, they have devised a binary vector that 

carries the tetramerized 35S enhancers next to the left T-DNA 

border. This vector has now been used to generate activationtagging 

pools for more than 50,000 individual transformants 

(Jeong et al 2002). Moreover, examination of randomly selected 

tags in the intergenic regions has shown that approximately 

40% of the enhancer insertions increase nearby gene 

expression, proving that activation-tagging efficiency at the 

transcriptional level is quite high. 

Forward screening 

Insertional mutant lines are useful resources for studying gene 

function. However, most knockout mutations are recessive. 

Therefore, their phenotypes are not visible in the primary 

transgenic lines and can be detected only in the progeny generations. 

In their analyses of T-DNA tagging lines, Jeon et al 

(2000) have observed that the most common mutants are dwarfing 

(7.0%) and leaf-pigment alterations (9.5%). Developmental 

mutants have also been observed in their screening. One 

must determine whether these phenotypes are due to the insertion. 

Cosegregation of phenotypes with the insert DNA is a 

good indication that the phenotype is caused by the insert. 

Ultimately, the mutant phenotype must be confirmed by another 

allele or complemented by the wild-type gene. Tissue 

culturing is known to generate mutations not associated with 

insertions. Retrotransposons, such as Tos17 or Karma, are activated 

during tissue culture. Miniature inverted-repeat transposable 

elements (MITEs), such as mPing or Pong, can also 

act as mutagens. In addition, small insertions, deletions, and 

base substitutions may be induced in cultured cells. Therefore, 

a large portion of mutants are generated in these populations 

through a mechanism not associated with the introduced 

inserts. Nonetheless, insertional mutant populations are good 

resources for studying gene function with forward genetics. 

By analyzing insertional mutant pools, functions of a large 

number of genes have been determined. 

DNA pool screening 

Insertional mutant collections are more beneficial when used 

with reverse genetics approaches. With the near completion of 

rice genome sequencing, researchers are now identifying a large 


number of genes based on their sequence homology. Generally, 

groupings of related genes are formed, in which the functions 

of only one or a few have been elucidated. Reverse genetics 

approaches are excellent methods for studying a group 

of genes relatively quickly. In these approaches, mutants in a 

target gene are first collected, then their phenotypes are investigated. 

Double or multiple mutants may be generated as necessary 

for further assessment. 

Insertional mutants in a given gene can be isolated via 

PCR-based screening. Using a gene-specific primer and a 

primer located near the end of the insert, a DNA fragment flanking 

the insert element can be amplified and its sequence then 

determined. This strategy has already been successfully applied 

with Arabidopsis, petunia, and maize. Enoki et al (1999) 

and Hirochika (2001) have reported PCR-based reverse genetics 

screenings of the Ac and Tos17 insertion lines in rice. 

The DNA pools have been used for isolation of knockout mutations 

in several genes. DNA pools of the T-DNA insertional 

lines have also been generated and used for isolation of knockout 

and activation-tagging mutants (Lee et al 2003). 

Tag end sequence databases 

Tag end sequences (TES) that flank insert elements can be 

obtained by TAIL PCR, iPCR, or adaptor-ligation PCR. When 

numerous flanking sequences are generated, they are catalogued 

into databases. Although large-scale application of this 

alternative strategy requires considerable effort, once established, 

these databases can be easily shared with other scientists, 

thus facilitating the distribution of mutant materials and 

analysis of gene function. Because sequencing of the rice genome 

is nearly complete, flanking-sequence databases will 

become powerful tools for systematically analyzing the functions 

of a large number of genes in those species. The TES 

database has been derived from the Tos17 insertional mutant 

lines (Miyao et al 2003). Hirochika’s group has analyzed more 

than 20,000 TES from 4,316 mutant lines 

(www.postgresql.org). TES databases for T-DNA insertional 

lines have also been created by several laboratories. An et al 

(2003) have reported the establishment of a database of 3,793 

TES. Entries are now 24,299 (www.postech.ac.kr/life/pfg/risd). 

TES analyses reveal the distribution of insert elements in plant 

chromosomes. For example, most transposition events occur 

at a tightly linked site, as reported with Arabidopsis. Therefore, 

a short-range and highly preferential transposition system 

can be effectively used for targeted mutagenesis of closely 

linked genes. 

For any given gene, Ds transposes equally into exons 

and introns, indicating no preference toward the 5′ end for 

genes in Arabidopsis, maize, or rice. Analysis of 2,057 Ds flanking 

sequences in rice has demonstrated that Ds insertions are 

distributed randomly throughout that genome. 

Miyao et al (2003) have investigated more than 20,000 

unique insertion points at Tos17 tagged sites and report that 

insertion events are three times more frequent in genic regions 

than in intergenic regions. Consistent with this, Tos17 prefers 

gene-dense regions over centromeric heterochromatin regions. 

Although insertion target sites are distributed throughout the 

rice chromosomes, they tend to cluster. This phenomenon may 

reduce the efficiency of genome-wide gene disruption by Tos17. 

Analysis of TES has demonstrated that the frequency of T- 

DNA insertion into genic regions is greater than the expected 

probability given random insertion (An et al 2003). Frequencies 

are also higher at the beginning and end of the coding 

regions, as well as upstream near the start ATG codon. The 

overall GC content at the insertion sites is close to that measured 

for the entire rice genome. Functional classification of 

1,846 tagged genes has shown a distribution similar to that 

observed for all genes in the rice chromosomes. This indicates 

that T-DNA insertion is not biased toward a particular functional 

group (An et al 2003). At the chromosomal level, insertions 

are not evenly distributed; frequencies are higher at the 

ends and lower near the centromere. At certain sites, the frequency 

is greater than in the surrounding regions, proving that 

T-DNA insertion is not a random event. 

Prospects 

The number of insertional mutants accumulated thus far by 

the scientific community is large enough to saturate almost all 

the genes in japonica rice. It will be valuable to develop methods 

for tagging indica varieties. The rapid generation of TES 

databases is urgently needed for efficient use of this insertional 

mutant resource. DNA pools are useful for identifying the alleles 

of a given gene. Such reverse genetics materials are valuable 

to functional genomic analyses of a large number of genes 

in rice and for developing new varieties. 

References 

An S, Park S, Jeong DH, Lee DY, Kang HG, Yu JH, Hur J, Kim SR, 

Kim YH, Lee M, Han S, Kim SJ, Yang J, Kim E, Wi SJ, Chung 

HS, Hong JP, Choe V, Lee HK, Choi JH, Nam J, Kim SR, 

Park PB, Park KY, Kim WT, Choe S, Lee CB, An G. 2003. 

Generation and analysis of end sequence database for T-DNA 

tagging lines in rice. Plant Physiol. 133:2040-2047. 

Chin HG, Choe MS, Lee SH, Park SH, Koo JC, Kim NY, Lee JJ, Oh 

BG, Yi GH, Kim SC, Choi HC, Cho MJ, Han CD. 1999. Molecular 

analysis of rice plants harboring an Ac/Ds transposable 

element-mediated gene trapping system. Plant J. 19:615- 

623. 

Enoki H, Izawa T, Kawahara M, Komatsu M, Koh S, Kyozuka J, 

Shimamoto K. 1999. Ac as a tool for the functional genomics 

of rice. Plant J. 19:605-613. 



Jeon JS, Lee S, Jung KH, Jun SH, Jeong DH, Lee J, Kim C, Jang S, 

Yang K, Nam J, An K, Han MJ, Sung RJ, Choi HS, Yu JH, 

Choi JH, Cho SY, Cha SS, Kim SI, An G. 2000. T-DNA insertional 

mutagenesis for functional genomics in rice. Plant J. 

22:561-570. 

Jeong DH, An S, Kang HG, Moon S, Han JJ, Park S, Lee HS, An K, 

An G. 2002. T-DNA insertional mutagenesis for activation 

tagging in rice. Plant Physiol. 130:1636-1644. 


Kolesnik T, Szeverenyi I, Bachmann D, Kumar CS, Jiang S, 

Ramamoorthy R, Cai M, Ma ZG, Sundaresan V, Ramachandran 

S. 2004. Establishing an efficient Ac/Ds tagging system in 

rice: large-scale analysis of Ds flanking sequences. Plant J. 

37:301-314. 

Lee S, Kim J, Son JS, Nam J, Jeong DH, Lee K, Jang S, Yoo J, Lee 

J, Lee DY, Kang HG, An G. 2003. Systematic reverse genetic 

screening of T-DNA tagged genes in rice for functional genomic 

analyses: MADS-box genes as a test case. Plant Cell 

Physiol. 44:1403-1411. 

Miyao A, Tanaka K, Murata K, Sawaki H, Takeda S, Abe K, 

Shinozuka Y, Onosato K, Hirochika H. 2003. Target site specificity 

of the Tos17 retrotransposon shows a preference for insertion 

within genes and against insertion in retrotransposonrich 

regions of the genome. Plant Cell 15:1771-1780. 

Wu C, Li X, Yuan W, Chen G, Kilian A, Li J, Xu C, Li X, Zhou DX, 

Wang S, Zhang Q. 2003. Development of enhancer trap lines 

for functional analysis of the rice genome. Plant J. 35:418- 

427. 

Notes 

Author’s address: Pohang University of Science and Technology, 

Pohang 790-784, Republic of Korea, e-mail: 

genean@postech.ac.kr. 

Tissue culture-induced mutations and a new type 

of activation tagging as tools for functional analysis 

of rice genes 

Hirohiko Hirochika 

With the completion of genomic sequencing of rice, rice has 

been firmly established as a model organism for both basic 

and applied research. The next challenge is to uncover the functions 

of genes predicted by sequence analysis. Knockout of 

genes by insertional mutagenesis is a straightforward method 

for identifying gene functions. In Arabidopsis, whose entire 

genomic sequencing has been completed, mutant populations 

covering almost all the genes have been produced by using 

insertion elements, such as T-DNA, Ac/Ds, and En/Spm. While 

these mutant lines are important resources for forward genetics 

studies of gene function, their applications in reverse genetics 

are even more important with the available Arabidopsis 

genome sequence. Two primary approaches for reverse genetics 

analysis have been established. One is PCR screening of 

DNA pools from mutants using primers corresponding to a 

target sequence. The other is flanking sequence tag (FST) analysis, 

a strategy in which sequences flanking the insertion elements 

are determined in each mutant line to develop a database 

of knockout genes that can be searched electronically. 

Because of the labor- and time-demanding nature of PCR 

screening, FST database development is now the preferred 

strategy. Although investment in the FST database is high, 

screening of mutants is easy once the database is established. 

In several Arabidopsis laboratories (http://signal.salk.edu/ 

tabout.html, http://atidb.cshl.org, and http://genoplanteinfo.infobiogen.fr), 

a large collection of FST data, totaling more 

than 100,000, has been established and opened to the public. 

A large-scale insertional mutagenesis has begun also in rice 

using T-DNA, Ac/Ds, and the endogenous retrotransposon 

Tos17. The current status of rice mutant resources has been 

recently summarized by Hirochika et al (2004). Here, I summarize 

recent progress in insertional mutagenesis in rice using 

Tos17 and its application to forward and reverse genetics studies. 

In addition, I also mention our new challenge, the development 

of activation tagged lines using full-length cDNA. 

Forward and reverse genetics using Tos17 

The unique features of Tos17 that make it a powerful genetic 

tool for forward and reverse genetics studies are summarized 

as follows. (1) Transposition can be regulated since Tos17 is 

activated by tissue culture and becomes silent in regenerated 

plants. (2) Highly mutagenic during tissue culture, Tos17 transposes 

preferentially into gene-rich, low-copy regions and about 

ten loci on average are disrupted in each plant regenerated 

from 5-month-old culture. (3) Integration target loci were 

widely distributed over the chromosomes, so that random insertion 

for saturation mutagenesis is feasible. (4) Induced 

mutations are stable and germinally transmitted in the next 

generation. (5) The original copy number is quite low, one to 

five depending on varieties, so that it is easy to identify the 

transposed copy responsible for the specific mutation. (6) Rearrangements 

are very rare at junctions between Tos17 ends 

and flanking host sequences, so that screening of mutants by 

PCR and analysis of disrupted genes can be carried out with 

high efficiency. (7) The transposon is endogenous, so that 

screening and characterization of mutants in the field are possible 

without any environmental concern. However, some features 

might also be disadvantageous. For example, transposition 

via a copy-and-paste mode means that no revertants can 


e obtained, although revertants are useful for confirming that 

a gene is tagged and can be obtained from the mutants induced 

by class II elements such as Ac/Ds. 

Forward genetics studies 

Traditional transposon tagging is still an important method for 

cloning important genes for functional analysis. The first direct 

evidence for the feasibility of tagging using Tos17 was 

shown by cloning genes involved in abscisic acid biosynthesis 

(Agrawal et al 2001) and causative genes for more than 20 

mutations have been cloned by using this strategy. R 1 (M 2 ) 

generations of regenerated rice were screened for mutants based 

on the phenotypes in the field. About 50% of the regenerated 

lines examined showed many kinds of visible mutant phenotypes, 

such as dwarf, sterile, yellow, albino, virescent, viviparous, 

brittle, and spotted leaf. In addition to screening in the 

field, screening in vitro was also conducted to isolate genes 

involved in salt-stress tolerance, and root growth and development. 

Some of the mutants were picked out and further subjected 

to cosegregation analysis to determine whether the mutations 

were caused by Tos17 insertions. Genetic analysis 

showed that all these mutations are recessive, although some 

did not segregate in a 1:3 ratio. Because the copy number of 

Tos17 is low enough to visualize each transposed copy, the 

Tos17 copy causing the specific mutation can be identified by 

a simple DNA gel-blot analysis. Finally, the causative gene 

can be isolated by using IPCR (inverse PCR) or TAIL-PCR 

(thermal asymmetric interlaced-PCR). This situation is quite 

distinct from the tagging using endogenous transposable elements 

in other plants, such as maize and petunia, in which the 

copy number of transposable elements is high (approx. 100 

copies). 

Although gene-tagging with Tos17 is a powerful strategy 

for cloning genes, one fundamental problem should be 

noted. That is the relatively low tagging efficiency (5–10%) 

(Hirochika 2001, Agrawal et al 2001). This indicates that tissue 

culture-induced mechanisms other than Tos17 insertions 

cause untagged mutations with high frequency. Several lines 

of evidence indicated that untagged mutations involve deletions 

and point mutations induced by unknown mechanisms 

(G.K. Agrawal et al, unpublished results). Recent studies 

showed that these deletion mutations, providing complete lossof-function 

mutations, are also important resources for functional 

analysis of genes. These mutations can also be used as a 

resource for reverse genetics by using TILLING (Till et al 

2003) and PCR-based screening of deletions. 

Reverse genetics studies 

For reverse genetics, DNA pools for PCR screening or the FST 

database are needed to identify mutations in genes of interest. 

The first direct evidence for the feasibility of the use of Tos17 

for the PCR-screening strategy was shown by screening for a 

mutant of the homeobox gene (OSH15) (Sato et al 1999). DNA 

pools from 50,000 Tos17 insertion lines have been constructed 

and, so far, results from screening these DNA pools suggest a 

success rate of 50% for a given target sequence (H. Hirochika, 

unpublished results). Because of the finite nature of the DNA 

pool, it is not practical to distribute DNA pools publicly. In 

addition, PCR screening is labor- and time-intensive. Thus, 

generating an FST database as a public resource is a more 

desirable and practical strategy. As of June 2004, about 20,000 

independent flanking sequences from 6,000 lines have been 

determined (Miyao et al 2003; http://tos.nias.affrc.go.jp). Some 

other research groups constructing insertional mutant lines have 

also started to construct FST databases. T-DNA FST databases 

at Pohang University (www.postech.ac.kr/life/pfg/risd) and at 

Génoplante (http://genoplante-info.infobiogen.fr/oryzatagline) 

are open to the public. FST data are expected to be anchored 

to the public IRGSP genome sequence and positions of both 

predicted genes and FST data will be shown graphically on 

the physical map. 

In addition to FST databases, it is also important to make 

databases on phenotypes. To maximize the utility of phenotype 

databases, a common vocabulary for describing mutants 

should be adopted among the mutant producers in different 

projects. The Plant Ontology Consortium vocabulary provides 

such a common framework and integrates different fields of 

expertise specific to plants. If the phenotype database is linked 

to the FST database, functional assignment of genes will be 

greatly accelerated. 

Limitations and future prospects 

Several problems or limitations associated with features inherent 

in insertional mutagenesis have been noted when functional 

analysis was carried out. One obvious problem is the 

lack of a mutant phenotype. In Arabidopsis, less than 2% of 

the insertional mutant lines showed a mutant phenotype. This 

problem may be largely due to gene redundancy. One possible 

solution is a combination of mutations by crossing, as has been 

successfully demonstrated in Arabidopsis (Liljegren et al 

2000). Considering the higher gene redundancy in rice, this 

problem may become more serious. Another problem is background 

mutations induced during the production of mutant 

lines, which make it unreliable to directly correlate phenotype 

with insertion mutations. Most of the background mutations 

are recessive and probably introduced by tissue culture. The 

molecular natures of the mutations are not known, but they 

may be deletions and point mutations as described in the previous 

section. 

To solve those problems, activation tagging inducing 

dominant mutations will be a complementary strategy. More 

than 50,000 activation-tagged lines have been produced already 

by using T-DNA with multimerized 35S enhancers (Jeong 

et al 2002). Considering the random insertion of T-DNA, a 

large number of tagged lines must be needed to overexpress 

all the genes. Recently, a more efficient activation tagging system 

has been developed in Arabidopsis (T. Ichikawa et al, unpublished 

results). This system, called Fox Hunting (full-length 

cDNA over-expressor gene hunting), uses transformation with 

T-DNA carrying an Arabidopsis full-length cDNA library under 

the control of the modified 35S promoter. By using this 


system, random overexpression of the Arabidopsis full-length 

cDNA is possible. Considering gene redundancy in rice and 

the availability of a large number of full-length cDNAs (Kikuchi 

et al 2003), we have begun producing FOX lines of rice in 

which 15,000 independent rice full-length cDNA are 

overexpressed under the control of the ubiquitin promoter (H. 

Ichikawa et al, unpublished results). 

References 

Agrawal GK, Yamazaki M, Kobayashi M, Hirochika R, Miyao A, 

Hirochika H. 2001. Screening of the rice viviparous mutants 

generated by endogenous retrotransposon Tos17 insertion: 

tagging of a zeaxanthin epoxidase gene and a novel OsTATC 

gene. Plant Physiol. 125:1248-1257. 



Hirochika H, Guiderdoni E, An G, Hsing Y, Eun MY, Han CD, 

Upadhyaya N, Ramachandran S, Zhang Q, Pereira A, 

Sundaresan V, Leung H. 2004. Rice mutant resources for gene 

discovery. Plant Mol. Biol. 54:325-334. 

Jeong DH, An S, Kang HG, Moon S, Han JJ, Park S, Lee HS, An K, 

An G. 2002. T-DNA insertional mutagenesis for activation 

tagging in rice. Plant Physiol. 130:1636-1644. 

Kikuchi S et al. 2003. Collection, mapping, and annotation of over 

28,000 cDNA clones from japonica rice. Science 301:376- 

379. 

Liljegren SJ, Ditta GS, Eshed Y, Savidge B, Bowman JL, Yanofsky 

MF. 2000. SHATTERPROOF MADS-box genes control seed 

dispersal in Arabidopsis. Nature 404:766-770. 

Miyao A, Tanaka K, Murata K, Sawaki H, Takeda S, Abe K, 

Shinozuka Y, Onosato K, Hirochika H. 2003. Target site specificity 

of the Tos17 retrotransposon shows a preference for insertion 

within genes and against insertion in retrotransposonrich 

regions of the genome. Plant Cell 15:1771-1780. 

Sato Y, Sentoku N, Miura Y, Hirochika H, Kitano H, Matsuoka M. 

1999. Loss-of-function mutations in the rice homeobox gene 

OSH15 affect the architecture of internodes resulting in dwarf 

plants. EMBO J. 18:992-1002. 

Till BJ, Reynolds SH, Greene EA, Codomo CA, Enns LC, Johnson 

JE, Burtner AR, Young K, Taylor NE, Henikoff JG, Comai L, 

Henikoff S. 2003. Large-scale discovery of induced pointmutations 

with high-throughput TILLING. Genome Res. 

13:524-530. 

Notes 

Author’s address: Molecular Genetics Department, National Institute 

of Agrobiological Sciences, Kannondai 2-1-2, Tsukuba, 

Ibaraki 305-8602, Japan, e-mail: hirohiko@nias.affrc.go.jp. 

Toward genome-wide transcriptional analysis 

in rice using MAS oligonucleotide tiling-path microarrays 

Lei Li, Xiangfeng Wang, Xueyong Li, Ning Su, Viktor Stolc, Bin Han, Jiayang Li, Yongbiao Xue, Jun Wang, and Xing Wang Deng 

As the international efforts to completely sequence the rice 

genome are nearly completed, an immediate challenge and 

opportunity for the plant community is to comprehensively and 

accurately annotate the rice genome. This will lay a solid foundation 

for rice functional genomics and proteomics. We report 

here our current strategy to use integrative approaches centered 

on whole-genome nucleotide tiling-path microarray and 

array-guided full-length cDNA analysis for conducting genome-wide 

transcriptional analysis. Preliminary results suggest 

that these experimental approaches are useful in verifying 

predicted rice gene models and in identifying a significant 

number of new transcriptional units that were simply missed 

from current annotation. The pilot results further indicate that 

array data-guided RT-PCR cloning of full-length cDNA is 

emerging as an effective protocol to provide the much-needed 

experimental verification for those rice genes that lacked previous 

cDNA support. Thus, success of our effort described 

herein is likely to result in a more accurate annotation of all 

rice genes in the genome. 

Current status of rice genome annotation 

Estimation of the total gene number in rice based upon the 

initial draft sequences of Oryza. sativa L. subsp. japonica and 

indica ranged widely from 30,000 to 60,000 (Goff et al 2002, 

Yu et al 2002). Finished sequences of chromosomes 1, 4, and 

10 allowed a much fine-tuned estimation that placed the total 

gene number of rice between 57,000 and 62,500 (Feng et al 

2002, Sasaki et al 2002, The Rice Chromosome 10 Sequencing 

Consortium 2003). Although the fast-accumulating finished 

sequences will ultimately come to a conclusion for the total 

gene number estimation, it should be noted that about half of 

those annotated gene models lack experimental support. Further, 

between one-third and one-half of the predicted genes 

appear to have no recognizable homology in Arabidopsis and 

features that exhibit striking deviations from experimentally 

verified genes in other species. Therefore, experimental efforts 

complementary to computer-based genome annotation are 

needed to verify predicted genes and to discover new genes in 

rice. 


Tiling array/cDNA analysis 

Gene prediction 

Whole-genome 

tiling array 

Array data 

analysis 

Rice genome 

annotation 

Gene-finding programs 

Comparative genomics 

cDNA cloning/sequencing 

New bioinformatic tools 

Fig. 1. A proposed workflow for using tiling microarray to improve rice genome annotation. Current rice 

genome annotation efforts focus on ab initio gene prediction, comparative genomics, and other 

nonexperimental methods (block arrow on the right). Experiment-oriented approaches such as wholegenome 

tiling microarray and full-length cDNA analysis are emerging as complements to the computation-based 

methods (block arrow on the left). Arrowheads indicate the direction of information flow. See 

text for more details. 

Rice whole-genome tiling-path microarray 

One powerful and proven annotation approach complementary 

to ab initio gene prediction, comparative genomics, and 

other nonexperimental methods is whole-genome tiling-path 

microarray analysis coupled with array-guided cDNA cloning 

(Shoemaker et al 2001, Kapranov et al 2002, Yamada et al 

2003). Figure 1 illustrates the strategy of our effort to enhance 

the rice genome annotation using an integrative approach centered 

on tiling microarrays. Essential to our approach is the 

design and development of custom tiling microarrays that represent 

the whole rice genome. Given the size of the rice genome 

(approx. 430 Mb) and the repetitive nature of our experiments, 

a suitable tiling microarray should offer high feature 

density, versatility of modification, and compatibility with 

our existing conventional microarray facility. These considerations 

led us to choose the Maskless Array Synthesizer (MAS) 

platform developed by NimbleGen (www.nimblegen.com, 

Stolc et al 2004). 

In the initial phase, a minimum tiling strategy was designed 

to efficiently represent the genome with a minimum 

number of interrogating oligos employed that were strategically 

allocated to each potential gene such that expression of 

these genes can be unambiguously assayed. Implementation 

of this design required approximately 6.5 million pairs of interrogating 

oligos to tile essentially all the nonrepetitive sequences 

of the genome (Table 1), resulting in a resolution of a 

pair of 36-mer interrogating oligos every 60-bp genome sequence 

on average. Table 1 also shows the number of tiling 

oligo pairs for each of the 12 chromosomes. When synthesized 

at a density of 389,287 oligos per slide (Stolc et al 2004), 

the interrogating oligos that tile the japonica and indica genomes 

can be accommodated onto a set of 32 and 34 MAS 

arrays, respectively. It is expected that once the workflow of 

processing the minimum tiling arrays is streamlined, higherresolution 

arrays will be designed and used to further our analysis. 

Table 1. Estimation of 36-mer oligo pairs 

for a minimum tiling of the rice genomes. 

Chr. 

japonica 

Oligo pairs 

indica 

1 758,477 843,677 

2 638,666 678,379 

3 618,937 747,025 

4 563,693 597,322 

5 456,837 536,721 

6 526,611 551,022 

7 487,411 478,130 

8 469,369 514,189 

9 367,182 378,664 

10 375,141 419,408 

11 424,379 404,151 

12 440,214 390,756 

Total 6,126,917 6,539,444 

A pilot tiling array hybridization experiment 

To test the rice tiling array hybridization conditions, data acquisition, 

and analysis procedures, a pilot experiment was 

conducted in which two tiling arrays representing a portion of 

chromosome 10 were hybridized with mixed cDNA targets 

derived from four tissues: seedling root, seedling shoot, panicle, 

and suspension-cultured cells. Analysis of the array data detected 

expression of 77% of the reference gene models. Specifically, 

expression of 80.2% of the full-length cDNA matched 

genes (198 out of 247) and 70.9% (175 out of 247) of the 

genes without previous experimental support were detected. 

Thus, the cDNA-confirmed genes served as positive controls 

and demonstrated the sensitivity and feasibility of this approach. 

The relatively lower detection rate for the unsupported 

genes suggests that they are expressed at lower levels or restricted 

to specific cell types/developmental stages. Alterna- 


tively, some of these predicted genes might be false models 

that do not exist in vivo. 

Integration of tiling array data to improve rice genome 

annotation 

The hybridization data of the tiling microarrays can be used to 

further rice genome annotation (Fig. 1). For example, the hybridization 

data can provide support to or verification of predicted 

genes without prior experimental support, as detection 

of hybridization signals is strongly indicative of RNA synthesis 

directed by the genome segment represented by the interrogating 

oligos. Thus, the information generated from tiling 

arrays can be used to improve genome annotation by means of 

array-guided RT-PCR cloning and analysis of corresponding 

cDNAs. Likewise, the array data are also expected to reveal 

novel transcriptional units (i.e., those that were not included 

in any annotation but were detected by tiling arrays). In this 

regard, the array data could serve as a guidepost for cloning 

the corresponding cDNA by means of RT-PCR (Fig. 1). Compared 

with sequencing library-based full-length cDNA clones, 

which is considered the standard of gene annotation (Kikuchi 

et al 2003), array-assisted cDNA cloning and analysis will be 

targeted, and thus be more cost-effective and inclusive to cover 

the remaining portion of the rice genome that lacks corresponding 

expressed sequences. 

Compilation of reference genes for tiling array analysis 

For the sake of simplicity, a set of BGI (Beijing Genome Institute) 

japonica gene models was used as the reference genes in 

the pilot trial (Stolc et al 2004). To effectively decode the tiling 

microarray hybridization data, however, reference genes 

are critically important, for which a comprehensive set of rice 

gene models is precisely anchored in the genomic sequences. 

A unique advantage in rice is the availability of multiple annotations 

such as the TIGR (The Institute for Genomic Research) 

japonica annotation (www.tigr.org/tdb/e2k1/osa1/) and 

the BGI indica and japonica annotations (http:// 

rise.genomics.org.cn). While each annotation has its own forte 

and weakness, comparison and analysis of these annotations 

should yield a more comprehensive inventory of gene models 

to be evaluated by tiling arrays than any single annotation can 

offer. Reference gene compilation also takes advantage of the 

EST and full-length cDNA collections in the TIGR and the 

Kikuchi full-length cDNA data sets (http:// 

cdna01.dna.affrc.go.jp/cDNA/). For example, the full-length 

cDNA-confirmed genes could naturally serve as positive controls 

to test array data and be used as training data to improve 

rice gene-finding algorithms (Fig. 1). Moreover, rice sequences 

have been subjected to extensive comparative genomic analysis 

with other cereals and other plant species. Closer and further 

inspection of those data should help to identify common 

and unique cereal genes and to integrate multiple annotations 

to provide a more comprehensive representation of the rice 

genome content (Fig. 1). 

Development of new bioinformatic tools to facilitate rice 

genome annotation 

It has been realized from the pilot experiments that new computational 

tools need to be developed and validated to interpret 

tiling data and facilitate incorporation of tiling data to 

improve rice genome annotation. For example, algorithms to 

ascertain transcription need to be improved by incorporating 

more statistical parameters. Reliable tiling data in turn will 

provide better training data to improve gene prediction and 

other rice bioinformatic tools to enhance genome annotation 

(Fig. 1). The tiling data can be used in conjunction with several 

other public-funded rice genomics projects to maximize 

the detection of rice genes expressed at different developmental 

stages or under diverse environmental conditions. Consequently, 

improved genome annotation will also permit better 

array design to probe subtle transcriptional events such as alternative 

splicing, differential initiation and termination, etc. 

Finally, to present the tiling data in an accessible and informative 

form, an interactive database is required where the tiling 

data will be correlated with the complete genome sequence 

with all the annotated rice genes and other genomic features 

linked to cDNA/EST sequences and proteomic and structural 

information. Applying these approaches in rice should aid in 

the current genomic efforts to provide a complete and accurate 

rice genome annotation that holds the key to unravel the 

biology of the agriculturally important cereal crops. 

The current research on japonica rice genome tiling array 

analysis in the laboratory of Xing Wang Deng at Yale University 

is supported by a grant from the National Science Foundation 

plant genome program (DBI-0421675). The collaborative 

effort in China was supported by the 863 rice functional 

genomics program from the Ministry of Science and Technology 

of China. 

References 

Feng Q, Zhang YJ, Hao P, Wang SY, Fu G, Huang YC, Li Y, Zhu JJ, 

Liu YL, Hu X, et al. 2002. Sequence and analysis of rice chromosome 

4. Nature 420:316-320. 

Goff SA, Ricke D, Lan TH, Presting G, Wang RL, Dunn M, 

Glazebrook J, Sessions A, Oeller P, Varma H, et al. 2002. A 

draft sequence of the rice genome (Oryza sativa L. ssp. 

japonica). Science 296:92-100. 

Kapranov P, Cawley SE, Drenkow J, Bekiranov S, Strausberg RL, 

Fodor SPA, Gingeras TR. 2002. Large-scale transcriptional 

activity in chromosomes 21 and 22. Science 296:916-919. 

Kikuchi S, Satoh K, Nagata T, Kawagashira N, Doi K, Kishimoto N, 

Yazaki J, Ishikawa M, Yamada H, Ooka H, et al. 2003. Collection, 

mapping, and annotation of over 28,000 cDNA clones 

from japonica rice. Science 300:1566-1569. 

Sasaki T, Matsumoto T, Yamamoto K, Sakata K, Baba T, Katayose 

Y, Wu JZ, Niimura Y, Cheng ZK, Nagamura Y, et al. 2002. 

The genome sequence and structure of rice chromosome 1. 

Nature 420:312-316. 


Shoemaker DD, Schadt EE, Armour CD, He YD, Garrett-Engele P, 

McDonagh PD, Loerch PM, Leonardson A, Lum PY, Cavet 

G, et al. 2001. Experimental annotation of the human genome 

using microarray technology. Nature 409:922-927. 

Stolc V, Li L, Wang X, Li X, Su N, Tongprasit W, Han B, Xue Y, Li 

J, Snyder M, Gerstein M, Wang J, Deng X-W. 2004. Towards 

genome wide transcriptional unit analysis in rice using oligonucleotide 

tiling-path microarrays. Plant Mol. Biol. (In press.) 

The Rice Chromosome 10 Sequencing Consortium. 2003. In-depth 

view of structure, activity, and evolution of rice chromosome 

10. Science 300:1566-1569. 

Yamada K, Lim J, Dale JM, Chen H, Shinn P, Palm CJ, Southwick 

AM, Wu HC, Kim C, Nguyen M, et al. 2003. Empirical analysis 

of transcriptional activity in the Arabidopsis genome. Science 

302:842-846. 

Yu J, Hu S, Wang J, Shu WG, Li S, Liu B, Deng Y, Dai L, Zhou Y, 

Zhang X, et al. 2002. A draft sequence of the rice genome 

(Oryza sativa L. ssp. indica). Science 296:79-92. 

Notes 

Authors’ addresses: Lei Li, Xueyong Li, Ning Su, and Xing Wang 

Deng, Department of Molecular, Cellular, and Developmental 

Biology, Yale University, New Haven, CT 06520; 

Xiangfeng Wang, National Institute of Biological Sciences, 

Zhongguacun Life Science Park, Beijing 102206, China; 

Xiangfeng Wang and Jun Wang, Beijing Institute of Genomics, 

Chinese Academy of Sciences, Beijing 101300, China; 

Xueyong Li and Yongbiao Xue, National Center of Crop Design, 

China Bioway Biotech Group Co., Ltd., Beijing 100085, 

China; Xiangfeng Wang and Ning Su, Peking-Yale Joint Center 

for Plant Molecular Genetics and Agrobiotechnology, College 

of Life Sciences, Peking University, Beijing 100871, 

China; Viktor Stolc, Genome Research Facility, NASA Ames 

Research Center, MS 239-11, Moffett Field, CA 94035; Bin 

Han, National Center for Gene Research, Shanghai Institutes 

for Biological Sciences, Chinese Academy of Sciences, 500 

Caobao Road, Shanghai 200233, China; Jiayang Li, Institute 

of Genetics and Developmental Biology, Chinese Academy 

of Sciences, Datun Road, Beijing 100101, China, e-mail: 

xingwang.deng@yale.edu. 

Candidate gene characterization at the Pup1 locus: a major 

QTL increasing tolerance of phosphorus deficiency 

Matthias Wissuwa, Kristy Gatdula, and Abdelbagi Ismail 

Phosphorus (P) deficiency is a major yield-limiting factor for 

rice, particularly under upland or rainfed lowland conditions 

(Kirk et al 1998). P deficiency is frequently not due to low 

soil-P content in absolute terms but due to tight binding of 

soil-P in forms that are not readily available to the plant. The 

development of rice cultivars capable of using a higher portion 

of this fixed P already present in soils could be an attractive 

and cost-effective approach to increasing rice yields on P- 

deficient soils. 

One promising step toward the development of more P- 

efficient cultivars was the identification of the Pup1 locus in a 

QTL mapping population that was derived from rice cultivars 

Nipponbare (low P uptake) and Kasalath (high P uptake) 

(Wissuwa et al 2002). The effect of Pup1 was confirmed using 

a near-isogenic line (NIL) carrying the positive Pup1 allele 

from Kasalath in the Nipponbare genetic background. NIL- 

Pup1 had three to four times the P uptake from a highly P- 

fixing volcanic ash soil than Nipponbare with equally high 

effects on biomass accumulation and grain yield (see Okada 

and Wissuwa, this volume). Further analysis suggested that 

roots of NIL-Pup1 were slightly more efficient in extracting 

soil-bound P and that this advantage helped to maintain higher 

root growth rates, with additional benefits for P uptake 

(Wissuwa 2004). 

Candidate gene identification 

Efforts at IRRI are now directed toward identifying the gene(s) 

at the Pup1 locus. Further fine mapping of Pup1 has advanced 

considerably and Pup1 has now been mapped to a 195-kb interval 

spanning three BAC clones on chromosome 12 (Fig. 1). 

Gene annotation in the Pup1 region identified 31 putative 

genes. Only one of those is a known gene, whereas some sequence 

similarities to known genes exist for an additional four 

genes. The remaining annotations were for hypothetical proteins 

(13) and transposable elements (13). None of the sequence 

similarities suggested an association with processes involved 

in P uptake or metabolism. This would suggest that Pup1 is 

most likely a novel gene. It is also possible, however, that the 

gene is simply absent or highly distorted in Nipponbare, which 

would make annotations based on Nipponbare sequence data 

impossible. 

Gene-specific primers have been developed for the putative 

genes located in the Pup1 interval with the aim to investigate 

P-deficiency-induced expression patterns. Nipponbare 

and NIL-Pup1 were grown in nutrient solution with three levels 

of P supply: excess P (50 µM), low P (1 µM), and zero P. 

RNA was isolated from roots and shoots of 5-week-old plants. 

RT-PCR performed on transcribed RNA samples identified 

two candidate genes that showed differential expression depending 

on genotype and P supply (Fig. 2). Candidate gene #1 


kb Line A B C 

M18 

50 

B 

AC 

1 

HG28 

134 

HG29 

141 

M50 

227 

B 

AC 

2 

Pup interval 

M59 

273 

T20 

329 

B 

AC 

3 

M69 

354 

Pup 1 region 

pseudomolecule 

Phenotype: N K K 

Genotype: 

Nipponbare 

Kasalath 

Fig. 1. Fine mapping of Pup1 to a 195-kb interval on chromosome 12, using selected 

recombinant lines. The phenotype of lines A and B places Pup1 upward of 

marker T20, while line C places Pup1 downward of HG28. The Pup1 interval spans 

three BAC clones and gene annotation identified 31 hypothetical genes in that 

region. 

Candidate gene #1 

Candidate gene #2 

N 

Pup1 

N 

Pup1 

N 

Pup1 

N 

Pup1 

N 

Pup1 

N 

Pup1 

High P Low P Zero P High P Low P Zero P 

Fig. 2. Expression patterns (RT-PCR) for two Pup1 candidate genes that are 

up-regulated by P deficiency in NIL-Pup1. In Nipponbare (N), gene #1 (at 

110 bp) was not expressed at all, whereas gene #2 (at 540 bp) was constitutively 

expressed. Expression analysis was conducted using root RNA 

sampled from 5-week-old plants that were grown in nutrient solution under 

three levels of P supply (50 µM P, 1 µM P, zero P). 


was not expressed in Nipponbare but expression was induced 

by P deficiency in NIL-Pup1. Further analysis suggests that 

this sequence codes for a small soluble protein (52 aa) most 

likely targeted for the nucleus. The second candidate gene follows 

a similar P-deficiency-induced expression pattern in NIL- 

Pup1, but, in this case, it is strongly and constitutively expressed 

in Nipponbare, regardless of P supply. The protein 

encoded by this gene contains one transmembrane helix. Further 

expression profiling using additional time points and a 

shoot-tissue control is currently being done and should provide 

a more detailed picture of these two candidate genes. 

That none of the annotated genes in the Pup1 region can 

be associated with genes known to be involved in P metabolism 

or uptake is unfortunate in terms of facilitating gene identification. 

However, this may not be very surprising considering 

that most genes related to P uptake that have been identified 

up to now, such as P transporters, phytases, or phosphatases, 

have been identified using low-P nutrient solutions 

as the screening medium. These genes do possess important 

functions in P metabolism and uptake but so far it has not been 

shown that they can be used to improve tolerance of P deficiency 

in the field. In contrast, Pup1 was mapped and its effect 

was confirmed in a P-deficient soil. This may be an indication 

that alleles most useful for improvements of abiotic stress 

tolerance of plants will have to be identified in natural environments. 

The positive effect of Pup1 on tolerance of P deficiency 

has repeatedly been confirmed in the field using NIL-Pup1. 

Recently, it has also been validated that this effect was not 

limited to a Nipponbare genetic background or to the specific 

soil conditions at the field site used to map and confirm the 

QTL. Following a cross to the donor parent Kasalath, Pup1 

was transferred into the background of two tropical rice cultivars, 

IR36 and IAC47. The introgression of Pup1 was monitored 

using flanking markers. Several introgression lines were 

evaluated together with a well-adapted local check variety at a 

P-deficient field site in the Philippines that had soil properties 

very different from those of the original field. The variability 

among introgression lines was high but a majority outperformed 

both recurrent parents. The best line exceeded the tolerant local 

check in grain yield by 26%. This represented a quite remarkable 

achievement considering that none of the lines had 

been selected at the test site. Given that variability among a 

small set of test lines was high, it appeared feasible to make 

additional gains through selection. This success has convinced 

breeders at IRRI to use the most promising introgression line 

as a donor to transfer the Pup1 locus to elite breeding material 

lacking tolerance of P deficiency. 

References 

Kirk GJD, George T, Courtois B, Senadhira D. 1998. Opportunities 

to improve phosphorus efficiency and soil fertility in rainfed 

lowland and upland rice ecosystems. Field Crops Res. 56:73- 

92. 

Wissuwa M. 2004. Combining a modelling with a genetic approach 

in establishing associations between genetic and physiological 

effects in relation to phosphorus uptake. Plant Soil. (In 

press.) 

Wissuwa M, Wegner J, Ae N, Yano M. 2002. Substitution mapping 

of Pup1: a major QTL increasing phosphorus uptake of rice 

from a phosphorus-deficient soil. Theor. Appl. Genet. 105:890- 

897. 

Notes 

Authors’ address: International Rice Research Institute, Los Baños, 

Philippines, e-mail: m.wissuwa@cgiar.org. 

Wrap-up of Session 2 

Extensive analysis of the rice genome in the last decade has led 

to the development of valuable tools that have contributed a great 

deal toward significant progress in rice molecular genetics. With 

the completion of a high-quality map-based sequence of the rice 

genome, it is expected that rice research will take on a new dimension. 

As reviewed by T. Sasaki of NIAS, the sequencing effort 

was organized in 1998 as a collaboration of 10 countries or regions, 

which made up the International Rice Genome Sequencing 

Project (IRGSP). By adopting a common strategy and sharing 

resources, the IRGSP successfully completed the sequencing of 

japonica rice variety Nipponbare at the end of 2004. This session 

highlighted research based on information derived from the 

genome sequence. 

M. Yano of NIAS described studies on map-based cloning 

of QTLs for rice heading date or flowering time. Heading date is a 

crucial characteristic for adaptation to different cultivation areas 

since rice is widely cultivated from the equatorial region to about 

40º N and S. This suggests the existence of several QTLs for rice 

heading date. M. Yano identified at least 14 QTLs for heading 

date and found out that about half of them were photoperiodsensitive 

genes. Using a series of chromosomal substitution lines 

as a resource for the cloning of each targeted gene of the photoperiod-sensitive 

genes, several, such as Hd1, Hd3a, Hd5, Ehd1, 

and Lhd4, have been isolated. The difference in sequences among 

these genes and their combination in rice varieties could explain 

the phenotype of heading date at each cultivation area. This strategy 

relies on genetic analysis and requires accurate genome sequence 

information to facilitate designing DNA markers that can 

be used to narrow down the candidate region of a targeted gene. 

The identification of genes or allelic diversity by reverse genetics 

methods also requires access to genome sequence information. 


H. Leung of IRRI reviewed the importance of allelic diversity 

in evaluating a germplasm collection, which is a key factor in 

sustaining agricultural productivity. The ability to assess genetic 

variation in rice is enhanced by information on single nucleotide 

polymorphism (SNP). Since several germplasm collections exist 

for rice, information on SNPs of important phenotypes and the 

construction of a database are desirable for crop improvement. 

G. An of Pohang University of Science and Technology 

(POSTECH) in Korea and H. Hirochika of NIAS focused on artificial 

knockout of rice genes by insertional mutagenesis. G. An 

used a T-DNA mediated by Agrobacterium and the Ac/Ds system 

of maize for generating insertion mutants in rice. H. Hirochika 

used a rice endogenous retrotransposon, Tos17, as a tool. These 

insertion elements are preferentially inserted in the genome with 

T-DNA normally inserted into any genomic region and Tos17 mainly 

into genic regions. The utility of gene disruption lines depends on 

the degree of saturation of insertion. So far, 50,000 Tos17 mutant 

lines have been generated, but it is still necessary to increase 

the insertion points so that all the genes could be covered. 

However, this may not be feasible considering a relatively 

high redundancy of family genes in the rice genome. If one of the 

family genes is disrupted, the remaining genes might complement 

this disruption and the phenotype will not be altered. In 

addition, since activation of Tos17 is induced during tissue culture, 

the occurrence of somaclonal variation may obstruct the 

identification of affected phenotypes. When a novel phenotype is 

accompanied by this genomic rearrangement, the phenotype must 

be analyzed genetically and not by a reverse genetics method. In 

such cases, it is necessary to develop a novel method of in planta 

gene disruption or silencing to induce stable mutations. 

In the last five years, the correspondence between various 

phenotypes and genes has been widely explored using genetic 

and reverse genetics methods. The phenotypes associated with 

increased yield such as plant height and architecture, which are 

controlled by many genes, are well-known targets of breeding. 

The Green Revolution, which brought dramatic increases in cereal 

grain yields in the late 1960s, was due largely to reduced 

plant height, increased photosynthetic activity, and increased 

efficiency of fertilizer use for high yields. For rice, a mutant of 

semidwarfism (sd1) was used as a donor of reduced height of 

many rice varieties such as IR8 of IRRI and Reimei in Japan. M. 

Matsuoka of Nagoya University has identified many genes affecting 

rice plant height, including sd1, and revealed the outline of a 

signal transduction pathway of gibberellic acid, which is the key 

plant hormone controlling plant height. For wheat, another gene, 

Rht1, has been used for breeding a semidwarf plant. It has been 

reported that this gene is also involved in the signal transduction 

pathway of gibberellin. The corresponding locus of Rht1 in rice is 

known as slr1, which gives a mutation of slender phenotype. 

Both genes carry a mutation in molecular domains, which affect 

the perception of gibberellin and control the signal of this event 

to the next step. As documented by M. Matsuoka, our knowledge 

on the physiological interaction of gibberellin and rice genes 

is still limited and further studies on the molecular events in line 

with signal transduction of gibberellin are necessary. This must 

be implemented using the detailed genomic information on rice. 

Xing-Wang Deng of Yale University reported on a comparative 

tiling-path microarray analysis of the transcription units of 

both japonica and indica subspecies using 60-mer DNA chips. 

The genome-wide transcription analysis facilitated the identification 

of additional expression-supported gene models. Ideally, such 

an array is expected to detect a change in gene expression under 

various environmental or biotic stresses. In addition, a large-scale 

microarray covering the entire genome can be used to validate 

the expression of predicted genes in the genome sequence. 

The topics covered at this conference represent only some 

of the current focus in rice genetics and genomics. The completion 

of the rice genome sequence will pave the way for novel 

approaches in clarifying many biological phenomena. The molecular 

mechanisms for biotic and abiotic stresses can be elucidated 

by identifying the genes involved in these phenotypes. Also, 

linkage disequilibrium can be evaluated practically using the genome 

sequence to provide a novel tool for mapping important 

agronomic traits. It is expected that further advances in rice molecular 

genetics will promote ground-breaking research that will 

facilitate a complete understanding of the biology of rice, the 

inheritance of agronomically important traits, and the control of 

various biological pathways that are necessary to continually improve 

the rice plant. 


SESSION 3 

Opportunities and challenges 

of transgenic rice 

CONVENER: F. Takaiwa (NIAS) 

CO-CONVENERS: N. Oka (NIAS) and S. Datta (IRRI)

Overproduction of C 4 

enzymes in transgenic rice: 

an approach for improved photosynthesis and crop yield 

Mitsue Miyao-Tokutomi and Hiroshi Fukayama 

Most terrestrial plants, including many important crops, such 

as rice, wheat, soybean, and potato, assimilate CO 2 through 

the C 3 photosynthetic pathway and are classified as C 3 plants. 

However, some plants, such as maize and sugarcane, possess 

the C 4 photosynthetic pathway in addition to the C 3 pathway, 

and these are classified as C 4 plants. The C 4 pathway acts to 

concentrate CO 2 at the site of the reactions of the C 3 pathway, 

and thus inhibits photorespiration. This CO 2 -concentrating 

mechanism, together with modifications of leaf anatomy, enables 

C 4 plants to achieve high photosynthetic capacity and 

high water- and nitrogen-use efficiencies. As a consequence, 

the transfer of C 4 traits to C 3 plants is one strategy being adopted 

for improving the photosynthetic performance of C 3 plants. 

The C 4 pathway consists of three key steps: (1) the initial 

fixation of CO 2 in the mesophyll cell cytosol by phosphoenolpyruvate 

carboxylase (PEPC) to form a C 4 acid, (2) 

decarboxylation of a C 4 acid in the bundle-sheath cells to release 

CO 2 , and (3) regeneration of the primary CO 2 acceptor 

phosphoenolpyruvate in the mesophyll cell chloroplasts by 

pyruvate, orthophosphate dikinase (PPDK). The decarboxylation 

reaction is catalyzed by one or more of the three enzymes, 

namely, NADP-malic enzyme (NADP-ME), NADmalic 

enzyme, and phosphoenolpyruvate carboxykinase (PEP- 

CK). The enzymes involved in the C 4 pathway are also present 

in C 3 plants, probably mostly in the photosynthetic mesophyll 

cells, although their activities in C 3 plants are very low. 

In the leaves of C 3 plants, photosynthesis and subsequent 

carbon and nitrogen metabolism proceed mainly in the mesophyll 

cells. To have a significant effect on metabolism, any 

genes for C 4 enzymes (C 4 photosynthetic genes) introduced 

into C 3 plants will need to be expressed at high levels in these 

cells. A strategy based on the evolutionary scenario of C 4 genes 

has enabled C 4 enzymes to be expressed at high levels and in 

desired locations in the leaves of C 3 plants (Matsuoka et al 

2001, Miyao 2003). 

How to overproduce C 4 

enzymes 

in the mesophyll cells of C 3 

plants 

Recent comparative studies have revealed that C 3 plants have 

at least two different types of genes homologous to C 4 genes, 

one encoding enzymes of “housekeeping” function (C 3 -specific 

genes) and the other similar to the C 4 genes of C 4 plants 

(C 4 -like genes), though expression of the latter is very low or 

even undetectable in C 3 plants (Miyao 2003). It is now postulated 

that C 4 genes evolved from a set of preexisting counterpart 

genes in ancestral C 3 plants, with modifications at the site 

and in the level of expression in the leaves and kinetic proper- 

C 4 plant 

Ancestral 

C 4 plant 

C 4 -specific gene 

Enhancer elements 

Cell-specific elements 

Modification of kinetics 

C 4 -like gene 

Light-responsive elements 

Ancestral gene 



Fig. 1. A schematic representation of the evolution of C 4 -specific 

genes. 

ties of enzymes (Fig. 1; Miyao 2003). This hypothesis is based 

on previous observations that the promoters of the maize C 4 - 

specific PEPC and PPDK genes can drive high-level expression 

of a reporter gene in transgenic rice plants in an organspecific 

and mesophyll-cell-specific manner as in maize 

(Matsuoka et al 1993, 1994). It was found recently that this 

could also be the case for genes of C 4 enzymes located in the 

bundle-sheath cells of C 4 plants (see Miyao 2003). These findings 

have important implications that C 3 plants possess the 

regulatory factors necessary for high-level expression of C 4 

genes, and, more importantly, that the introduction of the intact 

C 4 -specific genes would lead to high-level expression of 

the C 4 enzymes in the leaves of C 3 plants. 

Enzymes located in the mesophyll cells of C 4 

plants 

The introduction of the intact maize C 4 -specific gene, which 

contained all exons and introns and its own promoter and terminator 

sequences, was effective in overproducing PEPC and 

PPDK in rice leaves (Ku et al 1999, Fukayama et al 2001). 

The activities of PEPC and PPDK in transgenic rice leaves 

increased up to 110-fold and 40-fold, respectively, that of 

nontransformants (Table 1). The level of the maize PEPC and 

PPDK proteins accounted for 12% and 35%, respectively, of 

total leaf soluble protein at most. In these transgenic rice plants, 

the levels of transcripts and protein and the activity of the introduced 

C 4 enzyme in the leaves all correlated well with the 

copy number of the introduced gene. In addition, the levels of 

transcripts per copy of the maize C 4 -specific gene were comparable 

in both maize and transgenic rice plants and the maize 

gene was expressed in a similar organ-specific manner. These 

observations suggest that the maize C 4 -specific genes behave 

in a qualitatively and quantitatively similar way in both maize 

and transgenic rice plants. 


Table 1. Increase in activities of C 4 enzymes in transgenic rice leaves. 

C 4 enzyme 

Highest enzyme activity (increase in fold) a 

(location in C 4 plants) Introduced construct References 

Over rice activity Over maize activity 

PEPC (MC) Intact maize gene 110 3–4 Ku et al (1999) 

PPDK (MC) Rice Cab prom::maize FL C 4 cDNA 5

An approach for introducing the C 4 

pathway into rice leaves 

Overproduction of the maize C 4 -specific PEPC or PPDK did 

not affect the growth of rice plants, while that of the maize C 4 - 

specific NADP-ME led to stunting and bleaching of leaf color 

(Tsuchida et al 2001). This stunting was not suppressed by cooverproduction 

of any of the maize PEPC, the maize PPDK, 

and the sorghum NADP-malate dehydrogenase (NADP- 

MDH), or a combination of the three. Thus, the rice C 3 -specific 

NADP-ME instead of the maize C 4 -specific isoform has 

been adopted for the introduction of the C 4 pathway into rice. 

Transgenic rice plants overproducing three (PEPC, PPDK, and 

NADP-ME) or four enzymes (PEPC, PPDK, NADP-ME, and 

NADP-MDH) are now being produced. 

References 

Fukayama H, Hatch MD, Tamai T, Tsuchida H, Sudoh S, Furbank 

RT, Miyao M. 2003. Activity regulation and physiological 

impacts of the maize C 4 -specific phosphoenolpyruvate carboxylase 

overproduced in transgenic rice plants. Photosynth. 

Res. 77:227-239. 

Fukayama H, Tsuchida H, Agarie S, Nomura M, Onodera H, Ono K, 

Lee B-H, Hirose S, Toki S, Ku MSB, Makino A, Matsuoka 

M, Miyao M. 2001. Significant accumulation of C 4 -specific 

pyruvate, orthophosphate dikinase in a C 3 plant, rice. Plant 

Physiol. 127:1136-1146. 

Ku MSB, Agarie S, Nomura M, Fukayama H, Tsuchida H, Ono K, 

Hirose S, Toki S, Miyao M, Matsuoka M. 1999. High-level 

expression of maize phosphoenolpyruvate carboxylase in 

transgenic rice plants. Nat. Biotech. 17:76-80. 

Matsuoka M, Furbank RT, Fukayama H, Miyao M. 2001. Molecular 

engineering of C 4 photosynthesis. Annu. Rev. Plant Physiol. 

Plant Mol. Biol. 52:297-314. 

Matsuoka M, Kyozuka J, Shimamoto K, Kano-Murakami Y. 1994. 

The promoters of two carboxylases in a C 4 plant (maize) direct 

cell-specific, light-regulated expression in a C 3 plant (rice). 

Plant J. 6:311-319. 

Matsuoka M, Tada Y, Fujimura T, Kano-Murakami Y. 1993. Tissuespecific 

light-regulated expression directed by the promoter 

of a C 4 gene, maize pyruvate, orthophosphate dikinase, in a 

C 3 plant, rice. Proc. Natl. Acad. Sci. USA 90:9586-9590. 

Miyao M. 2003. Molecular evolution and genetic engineering of C 4 

photosynthetic enzymes. J. Exp. Bot. 54:179-189. 

Miyao-Tokutomi M, Fukayama H, Tamai T, Matsuoka M. 2001. Highlevel 

expression of C 4 photosynthesis enzymes in transgenic 

rice. In: PS2001 Proceedings, 12th International Congress on 

Photosynthesis, S39-001. Canberra (Australia): CSIRO Publishing. 

Tsuchida H, Tamai T, Fukayama H, Agarie S, Nomura M, Onodera 

H, Ono K, Nishizawa Y, Lee B-H, Hirose S, Toki S, Ku MSB, 

Matsuoka M, Miyao M. 2001. High-level expression of C 4 - 

specific NADP-malic enzyme in leaves and impairment of 

photoautotrophic growth of a C 3 plant, rice. Plant Cell Physiol. 

42:138-145. 

Notes 

Authors’ address: Photosynthesis Laboratory, National Institute of 

Agrobiological Sciences, Kannondai, Tsukuba 305-8602, Japan. 

The uptake and translocation of minerals in rice plants 

Naoko K. Nishizawa 

The uptake and translocation of metal nutrients in plants are 

essential for plant growth and, since plants are the primary 

source of food for humans, the nutritional value of plants is of 

central importance to human health. The most widespread human 

nutritional problem in the world is iron (Fe) deficiency 

(WHO 2003). Increasing the ability of plants to provide higher 

levels of minerals, such as Fe, will have a dramatic impact on 

human health. Increasing the Fe uptake from soils is a prerequisite 

to increase the amount of Fe in the edible parts of plants. 

This can be aided by identifying the transporters involved in 

the translocation of Fe within plants. 

Despite its abundance in soils, much Fe is present in the 

insoluble Fe(III) form and not readily available to plants. Therefore, 

plants have evolved two distinct uptake strategies 

(Marschner et al 1986). Graminaceous plants use a chelation 

strategy and release low-molecular-weight compounds, 

phytosiderophores, from their roots to chelate Fe(III) in the 

soil (Takagi 1976). Three molecules of S-adenosyl-methionine 

are combined with nicotianamine synthase (NAS) to form 

nicotianamine (NA) in both graminaceous and 

nongraminaceous plants (Mori and Nishizawa 1987, Shojima 

et al 1990). In graminaceous plants, the amino group of NA is 

transferred by nicotianamine aminotransferase (NAAT), and 

the resultant keto form is reduced to deoxymugineic acid 

(DMA) and other mugineic acid family phytosiderophores 

(MAs). Rice secretes DMA, the first member of MAs. A recent 

study showed that transgenic rice containing NAAT genes 

from barley is more tolerant than nontransgenic individuals of 

low-Fe availability in calcareous soil since transgenic rice 

plants secrete higher amounts of MAs than do wild-type plants 

(Takahashi et al 2001). 

All other plant groups use a reduction strategy composed 

of Fe(III) reductase and Fe(II) transporter. They do not produce 

or release phytosiderophores. However, nongraminaceous 

plants synthesize NA, which is the biosynthetic precursor of 

MAs and structurally similar. NA chelates metal cations, including 

Fe(II) and Fe(III). Unlike MAs, NA is not secreted, 

and is thought to have a role in metal homeostasis in 

nongraminaceous plants. Since Higuchi et al (1999) isolated 

the first NAS gene from Hordeum vulgare, NAS genes have 


A 

ZmYS1 

OsYSL16 

OsYSL2 

OsYSL9 

OsYSL15 

B 

Root 

+Fe 

–Fe 

+Fe 

Leaf 

–Fe 

OsYSL2 

OsYSL10 

OsYSL5 

OsYSL11 

OsYSL6 

OsYSL6 

OsYSL13 

OsYSL14 

OsYSL13 

OsYSL12 

OsYSL14 

OsYSL7 

OsYSL15 

OsYSL17 

OsYSL18 

OsYSL8 

OsYSL1 

OsYSL16 

rRNA 

0.1 

OsYSL3 

OsYSL4 

Fig. 1. (A) The unrooted phylogenic tree for the 18 OsYSL amino acid sequences. ZmYS1 is the Fe(III)-phytosiderophore 

transporter from maize. (B) RNA gel-blot analysis of OsYSL transcripts in Fe-sufficient and Fe-deficient rice plants. Only the 

expression of OsYSL2, OsYSL6, OsYSL13, OsYSL14, OsYSL15, and OsYSL16 was detected in these plants. The bottom panel 

shows ethidium bromide-stained rRNA as a loading control. +Fe: Fe-sufficient plants, –Fe: Fe-deficient plants. 

been isolated from barley again, and from tomato, Arabidopsis 

thaliana, rice, and maize. In graminaceous plants, the Fe(III)- 

phytosiderophore complex is taken up by a specific transporter 

at the root surface. Maize yellow stripe 1 (YS1) is the gene 

encoding the Fe(III)-phytosiderophore transporter (Curie et al 

2001). 

Our search for YS1 homologues in the Oryza sativa L. 

subsp. japonica (cv. Nipponbare) rice genomic database identified 

18 putative YS1-like genes (OsYSL) that exhibited 36% 

to 76% sequence similarity to YS1 (Koike et al 2004) (Fig. 

1A). Further, we examined whether Fe regulates the expression 

of these genes in rice (Fig. 1B). Using Northern blot analysis, 

the transcripts of six genes were detected in roots or leaves 

of Fe-sufficient or Fe-deficient rice plants. OsYSL6 was constitutively 

expressed in both the roots and leaves of either Fesufficient 

or Fe-deficient plants, with Fe deficiency slightly 

reducing its expression. OsYSL13 was preferentially expressed 

in the shoots, and Fe deficiency reduced its expression. 

OsYSL14 was expressed in both roots and shoots, and Fe status 

had no effect on its expression. OsYSL15 was expressed 

only in roots and the expression was highly induced by Fe 

deficiency. OsYSL16 was expressed in both roots and shoots, 

and the expression in roots was slightly increased by Fe deficiency. 

We could not detect the transcripts of other members 

of OsYSLs in these plants by Northern blot analysis. 

The high expression of OsYSL2 in the leaves but not in 

the roots of Fe-deficient plants led us to hypothesize that 

OsYSL2 is a transporter involved in the long-distance transport 

of Fe in rice, but is not a rice Fe(III)-phytosiderophore 

for iron uptake from the soil. Based on the nucleotide sequence, 

OsYSL2 was predicted to encode a polypeptide of 674 

amino acids containing 14 putative transmembrane domains. 

Subcellular localization of OsYSL2 and green fluorescent protein 

(GFP), together with the presence of the putative transmembrane 

domains, showed that OsYSL2 is a transporter that 

is localized in the plasma membrane. 

The cell-type specificity of OsYSL2 expression was investigated 

using promoter (1.5 kb):β-glucuronidase (GUS) 

analysis. Although Northern blot analysis did not detect OsYSL2 

transcripts in the roots, GUS staining resulting from OsYSL2 

promoter activity showed a dotted pattern in phloem companion 

cells in the central cylinder of Fe-sufficient roots. The activity 

in companion cells increased in Fe-deficient roots, but 

no GUS staining was observed in epidermal or cortical cells, 

even in Fe-deficient roots. This absence of OsYSL2 promoter 

activity in the root epidermis or cortex supports the contention 

that OsYSL2 is not an Fe(III)-phytosiderophore transporter 

involved in the uptake of Fe from the soil. GUS staining was 

observed in phloem cells of the vascular bundles of leaves and 

leaf sheaths of Fe-sufficient rice. The phloem-specific expression 

of the OsYSL2 promoter suggested that OsYSL2 is involved 

in the phloem transport of Fe. In Fe-deficient leaves, the 

OsYSL2 promoter was active in all tissues, with particularly 

strong GUS activity evident in companion cells. Since the re- 

Session 3: Opportunities and challenges of transgenic rice 91

productive organs are major sinks for Fe and are important in 

supplying Fe for human needs, we examined OsYSL2 expression 

in flowers and developing seeds. The OsYSL2 promoter 

was active during reproductive development in rice. 

We examined the transport activity and substrate specificity 

of OsYSL2 using Xenopus laevis oocytes. OsYSL2 was 

heterologously expressed in the oocytes and the substrate-induced 

inward currents at –60 mV were measured in response 

to several related compounds. Surprisingly, OsYSL2 transported 

Fe(II)-NA and Mn(II)-NA, but did not transport Fe(III)-DMA 

or Mn(II)-DMA. This suggested, therefore, that OsYSL2 is a 

metal-NA transporter in rice but not an Fe(III)-DMA transporter. 

These studies with rice have established that OsYSL2 is 

an Fe-regulated metal-NA transporter that is involved in 

phloem transport and the translocation of mineral nutrients in 

grains. The presence of a metal-NA transporter in rice and its 

expression in the phloem companion cells clearly demonstrated 

that metals, at least Fe and Mn, are transported in the phloem 

as NA complexes. Together with the expression of OsNAS3 in 

the leaves of Fe-sufficient plants, these results suggest that NA 

functions in intracellular delivery of Fe as well as in its 

phloem transport. The expression of OsYSL2 in reproductive 

organs also suggests that translocation of Fe into grains is mediated 

by NA and that manipulating OsYSL2 and OsNAS3 

changes the Fe content in rice grains. 

This is the first identification of a metal-NA transporter, 

OsYSL2, involved in long-distance transport of Fe and accumulation 

of Fe in grains. Our results with rice, along with those 

of previous studies with tobacco (Takahashi et al 2003), provide 

molecular evidence that NA is a chelator that mediates 

metal transport in the phloem and regulates intracellular metal 

distribution in graminaceous and dicotyledonous plants. Furthermore, 

these results indicate that higher plants have a single 

strategy for the phloem transport of Fe, although graminaceous 

plants and other plant groups have different strategies for Fe 

uptake from the soil. 

An approach to produce rice plants that can tolerate Fe 

deficiency would be the alteration of the expression patterns 

of genes that regulate Fe homeostasis. Therefore, we aimed to 

make superpromoters highly responsive to Fe deficiency by 

combining the recently identified IDE1 and IDE2 elements 

with other enhancer-like sequences to enhance gene expression 

in Fe-deficient roots of rice plants (Kobayashi et al 2004). 

Fe deficiency-responsive element 1 (IDE1) and IDE2 are cisacting 

elements that are responsible for Fe-deficiency-inducible 

and root-specific expression of the barley (H. vulgare L.) 

gene IDS2 (Fe-deficiency-specific clone no. 2) (Kobayashi et 

al 2003). Modules containing IDE1 and IDE2 of the IDS2 promoter 

were used as repeats or were linked to the Fe-deficiencyresponsive 

promoter of barley IDS3 (Fe-deficiency-specific 

clone no. 3), and were connected to known enhancer-like sequences. 

Five artificial promoters, as well as the native promoters 

of barley IDS2 or IDS3, were connected individually 

upstream of GUS and were introduced into rice (Fig. 2). 

Transgenic rice plants were grown under control or Fe-deficient 

conditions, and GUS expression was analyzed. The artificial 

promoter that contained one module of IDE1 and IDE2 

conferred strong Fe-deficiency-inducible GUS expression to 

the roots of rice plants. The artificial promoters induced prominent 

gene expression in Fe-deficient rice plant roots, although 

the repetition of the elements had no obvious effect. Histochemical 

observations revealed that GUS expression driven 

by artificial and native promoters was spatially similar, and 

expression was dominant within vascular bundles and root exodermis. 

These findings suggest that there is coordinated expression 

of the genes that are involved in Fe-deficiency-induced 

Fe uptake in rice. This supports the possibility that a 

common mechanism regulates the expression of graminaceous 

genes that are involved in Fe acquisition. We propose that it 

would be advantageous to use the artificial and native promoters 

of IDS2 and IDS3 to produce crops that are tolerant of Fe 

deficiency. 

References 

Curie C, Panaviene Z, Loulergue C, Dellaporta SL, Briat JF, Walker 

EL. 2001. Maize yellow stripe 1 encodes a membrane protein 

directly involved in Fe(III) uptake. Nature 409:346-349. 

Higuchi K, Suzuki K, Nakanishi H, Yamaguchi H, Nishizawa NK, 

Mori S. 1999. Cloning of nicotianamine synthase genes, novel 

genes involved in the biosynthesis of phytosiderophores. Plant 

Physiol. 119:471-479. 

Kobayashi T, Nakayama Y, Itai RN, Nakanishi H, Yoshihara T, Mori 

S, Nishizawa NK. 2003. Identification of novel cis-acting elements, 

IDE1 and IDE2, of the barley IDS2 gene promoter 

conferring iron-deficiency-inducible, root-specific expression 

in heterogeneous tobacco plants. Plant J. 36:780-793. 

Kobayashi T, Nakayama Y, Takahashi M, Inoue H, Nakanishi H, 

Yoshihara T, Mori S, Nishizawa NK. 2004. Construction of 

artificial promoters highly responsive to iron deficiency. Soil 

Sci. Plant Nutr. (In press.) 

Koike S, Inoue H, Mizuno D, Takahashi M, Nakanishi H, Mori S, 

Nishizawa NK. 2004. OsYSL2 is a rice metal-nicotianamine 

transporter that is regulated by iron and expressed in the 

phloem. Plant J. 39:415-424. 

Marschner H, Römheld V, Kissel M. 1986. Different strategies in 

higher plants in mobilization and uptake of iron. J. Plant Nutr. 

9:695-713. 

Mori S, Nishizawa N. 1987. Methionine as a dominant precursor of 

phytosiderophores in Graminaceae plants. Plant Cell Physiol. 

28:1081-1092. 

Shojima S, Nishizawa NK, Fushiya S, Nozoe S, Irifune T, Mori S. 

1990. Biosynthesis of phytosiderophores: in-vitro biosynthesis 

of 2′-deoxymugineic acid from L-methionine and 

nicotianamine. Plant Physiol. 93:1497-1503. 

Takagi S. 1976. Naturally occurring iron-chelating compounds in 

oat- and rice-root washing. I. Activity measurement and preliminary 

characterization. Soil Sci. Plant Nutr. 22:423-433. 

Takahashi M, Nakanishi H, Kawasaki S, Nishizawa NK, Mori S. 

2001. Enhanced tolerance of rice to low iron availability in 

alkaline soils using barley nicotianamine aminotransferase 

genes. Nature Biotech. 19:466-469. 


Fig. 2. Sequence comparison of IDE1 with other Fe-deficiency-inducible promoters 

and IDE2. IDS2 IDE1+ represents IDE1 and its 3′-flanking 4 bp, which is strikingly 

conserved in the HvNAAT promoters. The palindromic sequence in IDE1 is indicated 

with a red box and arrowhead. Brackets in IDE1 and IDE2 indicate the sites of mutations 

in linker-scanning analysis. Orange letters indicate sequences identical to 

IDE1+. The distances (in bp) from the translation start sites are shown in parentheses. 

Sequences highly homologous to IDE1 were found in many other Fe-deficiencyinducible 

promoters. A sequence highly homologous to IDE1 was also found at –772/ 

–789 of the IDS2 promoter itself. IDE2 was also homologous to IDE1 and its 3′- 

flanking 4 bp. Neither IDE1 nor IDE2 showed substantial homology to other known 

cis-acting elements involved in Fe metabolism or root-specific expression. HvNAAT- 

A and HvNAAT-B, barley nicotianamine aminotransferase genes; HvNAS1, barley 

nicotianamine synthase gene; HvIDS3, barley Fe-deficiency-specific clone no. 3 

(mugineic acid synthase gene); OsNAS1 and OsNAS2, rice nicotianamine synthase 

genes; OsIRT1, rice ferrous Fe transporter gene; AtFRO2, Arabidopsis thaliana ferric-chelate 

reductase gene; AtIRT1, A. thaliana ferrous Fe transporter gene; AtNAS1, 

AtNAS2, and AtNAS4, A. thaliana nicotianamine synthase genes. IDE1 and IDE2 are 

novel cis-acting elements. Conserved cis-acting elements homologous to IDE1 and 

IDE2, along with cognate trans-acting factors, may function in various genes and 

plant species. 

Takahashi M, Terada Y, Nakai I, Nakanishi H, Yoshimura E, Mori S, 

Nishizawa NK. 2003. Role of nicotianamine in the intracellular 

delivery of metals and plant reproductive development. 

Plant Cell. 15:1263-1280. 

WHO (World Health Organization). 2003. www.who.int/nut/ida.htm. 

Notes 

Author’s addresses: Graduate School of Agricultural and Life Sciences, 

The University of Tokyo, 1-1 Yayoi, Bunkyo-ku, Tokyo 

113-8657, Japan, and Core Research for Evolutional Science 

and Technology (CREST), e-mail: annaoko@mail.ecc.utokyo.ac.jp. 


Improving drought and cold-stress tolerance 

in transgenic rice 

Kazuko Yamaguchi-Shinozaki and Kazuo Shinozaki 

Drought, high salinity, and freezing are environmental conditions 

that have adverse effects on plant growth and crop productivity. 

Plants respond and adapt to these stresses to survive 

under stress conditions at the molecular and cellular levels as 

well as at the physiological and biochemical levels. Expression 

of a variety of genes is induced by these stresses in diverse 

plants. The products of these genes are thought to function 

not only in stress tolerance but also in the regulation of 

gene expression and signal transduction in stress responses 

(Shinozaki et al 2003). 

Genetic engineering is useful for improving the stress 

tolerance of plants. Several different approaches have been 

attempted to improve the stress tolerance of plants by gene 

transfer (Shinozaki et al 2003). The genes selected for transformation 

were those involved in encoding enzymes required 

for the biosynthesis of various osmoprotectants. Other genes 

that have been selected for transformation include those that 

encoded enzymes for modifying membrane lipids, LEA protein, 

and detoxification enzymes. In all these experiments, a 

single gene for a protective protein or an enzyme was 

overexpressed under the control of the constitutive 35S cauliflower 

mosaic virus (CaMV) promoter in transgenic plants, 

although several genes have been shown to function in environmental 

stress tolerance and response. The genes encoding 

protein factors that regulate gene expression and signal transduction 

and that function in stress responses may be useful for 

improving the tolerance of plants of stresses by gene transfer 

as they can regulate many stress-inducible genes involved in 

stress tolerance. 

DREB/DRE regulon in Arabidopsis 

To understand the molecular process of signal transduction 

from the perception of the stresses to gene expression, we isolated 

many environmental stress-inducible genes in Arabidopsis 

thaliana. The products of these genes are thought to function 

not only in stress tolerance but also in the regulation of gene 

expression and in signal transduction in the stress response. 

These gene products can be classified into two groups. The 

first group includes proteins that probably function in protecting 

cells from dehydration. The second group of gene products 

contains protein factors that are involved in further regulation 

of gene expression and signal transduction and that function 

in stress response (Shinozaki et al 2003). 

We analyzed the expression of the drought-inducible 

genes and identified at least four independent regulatory systems 

in drought-responsive gene expression: two are ABAdependent 

and two are ABA-independent. In one of the ABA- 

independent pathways, a cis-acting element with nine base pairs 

(TACCGACAT), named the dehydration-responsive element 

(DRE), is involved in dehydration, and high-salt and low-temperature–induced 

gene expression (Yamaguchi-Shinozaki and 

Shinozaki 1994). 

Two cDNA clones that encode DRE-binding proteins, 

DREB1A/CBF3 and DREB2A, were isolated by using yeast 

one-hybrid screening (Liu et al 1998). cDNA clones encoding 

two DREB1A homologs (named DREB1B/CBF1 and 

DREB1C/CBF2) and one DREB2A homolog (DREB2B) were 

also isolated (Liu et al 1998). Expression of the DREB1A gene 

and its two homologs was induced by low-temperature stress, 

whereas expression of the DREB2A gene and its homolog was 

induced by dehydration. These results indicate that two independent 

families of DREB proteins, DREB1 and DREB2, function 

as trans-acting factors in two separate signal transduction 

pathways under low-temperature and dehydration conditions, 

respectively. 

Overexpression of the cDNA encoding DREB1A in 

transgenic Arabidopsis plants activated the expression of many 

of these stress-tolerance genes under normal growing conditions 

and resulted in improved tolerance of drought, salt loading, 

and freezing (Liu et al 1998, Kasuga et al 1999). To identify 

target stress-inducible genes of DREB1A, we performed 

microarray analysis. We searched for downstream genes in 

transgenic plants overexpressing DREB1A using the full-length 

cDNA microarray and GeneChip array. We identified more 

than 40 genes as the DREB1A downstream genes (Seki et al 

2001, Maruyama et al 2004). The products of these genes were 

not only proteins known to function against stress but also 

protein factors involved in further regulation of signal transduction 

and gene expression in response to stress. 

However, use of the strong constitutive CaMV 35S promoter 

to drive expression of DREB1A also resulted in severe 

growth retardation under normal growing conditions. In contrast, 

expression of DREB1A from the stress-inducible rd29A 

promoter gave rise to minimal effects on plant growth while 

providing an even greater tolerance of stress conditions than 

did expression of the gene from the CaMV 35S promoter 

(Kasuga et al 1999). We also analyzed the effect of 

overexpression of DREB1A on stress tolerance and growth 

retardation in transgenic tobacco plants using the constitutive 

CaMV 35S promoter and the stress-inducible rd29A promoter 

(Kasuga et al 2004). Overexpression of DREB1A improved 

drought- and low-temperature stress tolerance in tobacco. The 

stress-inducible rd29A promoter minimized the negative effects 

on plant growth in tobacco. These results indicate that 

the stress-inducible rd29A promoter and the DREB1A gene 


Table 1. List of characterized and predicted OsDREB genes from Oryza sativa var. japonica. 

Gene name Amino acid Genomic sequence Chromosome mRNA 

residues acc. no. no. 

OsDREB1A 238 – – AF300970, AY345233, 

AF494422, AK071519, 

AK105599 

OsDREB1B 218 – – AF300972, AY319971 

OsDREB1C 214 AP002838, AP001168 6 AY327040 

OsDREB1D 253 AP002536, AB023482 6 AF243384, AY345235 

OsDREB1E 236 AP004163, a AP004632 8 

OsDREB1F 219 AP003448 1 AY345234 

OsDREB1G 219 AL606640 4 AY114110 

OsDREB1H 224 AP005775, AP006060 2 AK106041, AK060550 

OsDREB1I 251 AP004163, b AP004632 8 

OsDREB2A 274 AP003301 1 AF300971, AK067313 

OsDREB2B 298 AC134927, AC134928 5 AK099221, AK071850 

OsDREB2C 230 AP004623 8 AY339375, AK108143 

OsDREB2D 261 AC135924 5 AK103822, AK102559 

a 

AP004162 15673 base–14966 base. b AP004162 19564 base–18812 base. 

are quite useful for improving drought, salt, and freezing stress 

tolerance not only in Arabidopsis but also in other kinds of 

dicot plants. 

DREB/DRE regulon in rice 

Rice, an important crop plant, has now emerged as an ideal 

model species for the study of crop genes because of its commercial 

value, relatively small genome size (approx. 430 Mb), 

diploid origin (2x=24), and close relationship with other important 

cereal crops. It is important to analyze the DRE/DREB 

(CRT/CBF) regulon in rice to understand the molecular mechanisms 

of stress tolerance and produce monocots with higher 

stress tolerance by gene transfer. We isolated five cDNAs for 

DREB homologs, OsDREB1A, 1B, 1C, 1D, and 2A, and analyzed 

their functions in rice (Dubouzet et al 2003). The 

OsDREB proteins were specifically bound to the DRE sequence 

and activated the transcription of a GUS reporter gene 

driven by the DRE sequence in rice protoplasts. Expression of 

the OsDREB1A and 1B genes was induced by low-temperature 

stress, whereas the expression of the OsDREB2A gene 

was induced by dehydration. Recently, we searched the rice 

genome database for amino acid sequences of the ERF/AP2 

domains of DREB and ERF, and found an additional five 

DREB1A homologs, OsDREB1E, OsDREB1F, OsDREB1G, 

OsDREB1H, and OsDREB1I, and three DREB2A homologs, 

OsDREB2B, OsDREB2C, and OsDREB2D, in rice (Table 1). 

We generated transgenic Arabidopsis overexpressing the 

OsDREB1A cDNA. The rice OsDREB1A protein exhibits a 

function similar to that of the Arabidopsis DREB1A protein in 

transgenic Arabidopsis plants (Dubouzet et al 2003). We also 

generated transgenic rice plants overexpressing the rice 

OsDREB or Arabidopsis DREB cDNAs and analyzed their 

tolerance of drought, high-salt, and cold stress. These transgenic 

plants showed high tolerance of these stresses. 

On the other hand, we carried out microarray analysis 

using rice cDNAs to identify stress-inducible genes in response 

to dehydration, high salt, and low temperature in rice (Rabbani 

et al 2003). We confirmed stress-inducible expression of the 

candidate genes selected by microarray analysis using RNA 

gel-blot analysis, and finally identified a total of 73 genes as 

stress-inducible genes in rice (Fig. 1). Among them, 36, 62, 

57, and 43 genes were induced by cold, drought, high salinity, 

and ABA, respectively. The products of the identified rice genes 

are classified into functional proteins and regulatory proteins 

like those of Arabidopsis (Rabbani et al 2003). Comparative 

analysis of stress-inducible genes in Arabidopsis and rice revealed 

a considerable level of similarity in stress responses 

between the two genomes at a molecular level. Among 73 identified 

stress-inducible genes in rice, 51 have already been reported 

in Arabidopsis with a similar function or gene name. 

These results indicate that rice has many stress-inducible genes 

in common with Arabidopsis, even though these two plants 

evolved separately a million years ago. Common stress-inducible 

genes include LEA proteins, antifreeze proteins, detoxification 

enzymes, and so on. All these genes are up-regulated in 

response to at least one of the abiotic stresses in rice and reported 

as stress-inducible genes in Arabidopsis. Transcriptome 

analysis of rice also revealed novel stress-inducible genes, 

suggesting some differences between Arabidopsis and rice in 

their response to stress. 

We analyzed expression of the stress-inducible genes 

identified by the cDNA microarray and Northern blot analyses 

in the transgenic rice plants overexpressing the OsDREB1A 

or DREB1C cDNAs and detected overexpression of some 

stress-inducible genes. The promoter regions of these genes 

contain the DRE sequences. These results indicate that there 

is a DRE/DREB regulon in the stress-responsible signal transduction 

in rice similar to that in Arabidopsis. The DREB/DRE 

regulon should be useful for producing transgenic dicots and 


A. Genes induced by cold, drought, and high-salinity 

stresses and by ABA application: 

D. Genes induced by drought and high-salinity stresses and 

by ABA application: 

G. Genes induced by drought and high-salinity stresses: 

H. Genes induced by drought stress and by ABA application: 

B. Genes induced by cold, drought, and high-salinity stresses: 

I. Genes induced by cold stresses: 

C. Genes induced by cold and drought stresses and by ABA 

application: 

E. Genes induced by cold and drought stresses: 

J. Gene expressed constitutively: 

F. Gene induced by cold and high-salinity stresses: 

Fig. 1. RNA gel-blot analysis of stress-inducible genes. Each lane was loaded with 10 µg of total RNA isolated from 2-wk-old rice seedlings that were exposed to H 2 O, dehydration, 

250 mM NaCl, 100 µM ABA, and 4 °C cold treatment for 1, 2, 5, 10, and 24 h. RNA was analyzed by gel-blot hybridization with gene-specific probes of selected stress-inducible 

clones by rice cDNA microarray. Stress-inducible clones were classified into various groups on the basis of their expression patterns in RNA gel-blot analysis under each stress 

treatment. 


monocots with higher tolerance of drought, high salinity, and/ 

or cold stress. 

References 

Dubouzet JG, Sakuma Y, Ito Y, Kasuga M, Dubouzet EG, Miura S, 

Seki M, Shinozaki K, Yamaguchi-Shinozaki K. 2003. 

OsDREB genes in rice, Oryza sativa L., encode transcription 

activators that function in drought-, high-salt- and cold-responsive 

gene expression. Plant J. 33:751-763. 

Kasuga M, Liu Q, Miura S, Yamaguchi-Shinozaki K, Shinozaki K. 

1999. Improving plant drought, salt, and freezing tolerance 

by gene transfer of a single stress-inducible transcription factor. 

Nature Biotechnol. 17:287-291. 

Kasuga M, Miura S, Shinozaki K, Yamaguchi-Shinozaki K. 2004. A 

combination of the Arabidopsis DREB1A gene and stress-inducible 

rd29A promoter improved drought- and low-temperature 

stress tolerance in tobacco by gene transfer. Plant Cell 

Physiol. 45:346-350. 

Liu Q, Kasuga M, Sakuma Y, Abe H, Miura S, Yamaguchi-Shinozaki 

K, Shinozaki K. 1998. Two transcription factors, DREB1 and 

DREB2, with an EREBP/AP2 DNA binding domain separate 

two cellular signal transduction pathways in drought- and lowtemperature-responsive 

gene expression, respectively, in 

Arabidopsis. Plant Cell 10:1391-1406. 

Maruyama K, Sakuma Y, Kasuga M, Ito Y, Seki M, Goda H, Shimada 

Y, Yoshida S, Shinozaki K, Yamaguchi-Shinozaki K. 2004. 

Identification of cold-inducible downstream genes of the 

Arabidopsis DREB/CBF3 transcriptional factor using two 

microarray systems. Plant. J 38:982-993. 

Rabbani MA, Maruyama K, Abe H, Khan MA, Katsura K, Ito Y, 

Yoshiwara K, Seki M, Shinozaki K, Yamaguchi-Shinozaki K. 

2003. Monitoring expression profiles of rice (Oryza sativa 

L.) genes under cold, drought and high-salinity stresses, and 

ABA application using both cDNA microarray and RNA gel 

blot analyses. Plant Physiol. 133:1755-1767. 

Seki M, Narusaka M, Abe H, Kasuga M, Yamaguchi-Shinozaki K, 

Carninci P, Hayashizaki Y, Shinozaki K. 2001. Monitoring 

the expression pattern of 1300 Arabidopsis genes under 

drought and cold stresses using a full-length cDNA microarray. 

Plant Cell 13:61-72. 

Shinozaki K, Yamaguchi-Shinozaki K, Seki M. 2003. Regulatory 

network of gene expression in the drought and cold stress 

responses. Curr. Opin. Plant Biol. 6:410-417. 

Yamaguchi-Shinozaki K, Shinozaki K. 1994. A novel cis-acting element 

in an Arabidopsis gene is involved in responsiveness 

to drought, low-temperature, or high-salt stress. Plant Cell 

6:251-264. 

Notes 

Authors’ addresses: Kazuko Yamaguchi-Shinozaki, Biological Resources 

Division, Japan International Research Center for 

Agricultural Sciences (JIRCAS), Tsukuba, Ibaraki; Laboratory 

of Plant Molecular Physiology, Graduate School of Agricultural 

and Life Sciences, The University of Tokyo, Tokyo, 

CREST, JST, Japan; Kazuo Shinozaki, Laboratory of Plant 

Molecular Biology, RIKEN Tsukuba Institute, Tsukuba, 

Ibaraki, CREST, JST, Japan, e-mail: 

kazukoys@jircas.affrc.go.jp. 

Broad-spectrum disease resistance in transgenic rice 

Motoshige Kawata 

Environmental stresses exert a critical influence on rice yields 

and pathogen attacks are sometimes the most devastating biotic 

stress. In a humid temperate climate, fungal rice blast 

caused by Magnaporthe grisea is the most serious problem. 

Bacterial leaf blight caused by Xanthomonas oryzae is a serious 

disease of rice in subtropical and tropical areas. The enhancement 

of disease resistance in rice has made a great contribution 

toward increasing the productivity of rice and decreasing 

the application of pesticides, which can negatively 

affect human health and the environment. Genetic engineering 

has given us opportunities and challenges for transgenic rice 

for improving disease resistance, based on recent advances in 

molecular biology. Cloning of several disease-resistance genes 

is providing insights into their function and how they protect 

plants against pathogen attacks. 

Developing domestic gene technologies 

Plant gene engineering meant to bring transgenic rice into practical 

use must address environmental safety and consumer concerns, 

and thus the development of new technologies is impor- 

tant. To develop new domestic gene technologies, we hit upon 

the idea of incorporating disease-resistance genes from vegetables, 

employing a selection technique that uses rice genes, 

and keeping the transgenes from functioning in the edible portions 

of transgenic rice. Based on this idea, we successfully 

developed new transgenic rice strains with multiple disease 

resistance by integrating new domestic gene technologies, including 

new selection marker genes from rice, defensin genes 

from vegetables, which are highly effective in conferring disease 

resistance, and new callus-specific and green-tissue-specific 

promoters from rice. We confirmed that the transgenes 

are specifically expressed in the callus or green tissue, whereas 

expression is suppressed in rice grains. 

Defensin genes 

Defensins are antimicrobial proteins found in a wide variety 

of organisms and that exhibit powerful antimicrobial activity 

against fungi and bacteria. In recent years, progress has been 

achieved in analyzing the characteristics of their antimicrobial 

activity against plant pathogens. Studies of the general 

mechanism on antimicrobial activity have been reported and 


ALS Trp Ser 

Cells that expressed ALS gene suffer growth inhibition in the 

existence of bispyribac sodium (BS) 

BS 

mutated ALS 

(mALS) 

Leu 

Ile 

Cells that expressed mALS gene do not suffer growth 

inhibition even in the existence of BS 

BS 

Fig. 1. Effect of the mutated acetolactate synthase gene (mALS). 

the detailed mechanisms of antimicrobial activity on defensin 

peptide have been studied (Thevissen et al 2004). Plant 

defensins have four disulfide bonds, distinguishing them structurally 

from the defensins of insects and mammals, which have 

three disulfide bonds (Fant et al 1998). 

Mature plant defensin proteins comprise 45 to 54 amino 

acid residues, and are small, basic proteins rich in cysteine. 

Their expression is observed in Cruciferae plants, as well as 

in sorghum, dahlias, and many other plants. Although their 

amino acid structures are diverse, cysteine residues are well 

preserved, and the alpha helix motif stabilized by cysteine is 

common to these defensins. Cruciferae defensin proteins, 

which belong to the group exhibiting the strongest antimicrobial 

activity, are expressed in places such as seed surfaces. 

They exist as gene clusters, and their expression characteristics 

and structures differ. One characteristic is that slight amino 

acid sequential differences change the strength of antibacterial 

activity and effectiveness against pathogens (De Samblanx 

et al 1997). Defensin genes exist in commonly consumed 

Cruciferae vegetables. They exhibit antibiotic activity against 

a wide variety of pathogens such as fungi and bacteria, but 

rice plants do not have them. We isolated and determined the 

structures of defensin gene clusters from eight kinds of vegetables, 

and found that adding defensin proteins suppressed 

the growth of rice blast hyphae in vitro (Kawata et al 2004). 

New promoters 

Developing transgenic plants endowed with practical traits such 

as disease resistance requires taking steps so that transgenes 

are expressed only in the necessary places, and not in the rice 

grains. We isolated the green-leaf-specific promoter, which was 

expressed in stems and leaves, and the callus-specific promoter. 

These promoters were not expressed in the rice grains, including 

the germ. Moreover, we modified the isolated promoters, 

such as adding restriction enzyme recognition sites, thereby 

making it possible to easily construct the vectors used for transformation. 

These new technologies make it possible to induce 

the functioning of “disease-resistance genes” and “marker genes 

that can select transformed cells” at the necessary time for each, 

and these techniques suppress the accumulation of transgene 

proteins in rice grains. 

New selection markers 

Acetolactate synthase (ALS) is one of the enzymes that is related 

to metabolism, and it is widely found in plants. It is one 

of the target enzymes for herbicides. Cells that expressed the 

ALS gene suffer growth inhibition in the existence of bispyribac 

sodium (BS), but a natural mutation with resistance to this 

herbicidal effect has been found among cultured rice plant cells. 

An investigation into the reason for this resistance revealed 

that inhibition of enzyme activity changed because of the replacement 

of amino acid sequences at two locations in the ALS 

protein (mALS) (Fig.1). Formerly, the marker genes used in 

the development of transgenic plants involved mainly antibiotic 

resistance genes obtained from microorganisms or other 

organisms. This new selection marker gene (mALS) can be used 

in place of antibiotic resistance selection marker genes. 

Transgenic rice that integrates domestic gene technologies 

Transgenic rice strains developed with multiple disease-resistance 

genes are characterized by the integration of original 

new technologies, incorporating genes with high disease resistance 

from vegetables. In addition to mALS ligated downstream 

from the rice callus-specific promoter, and a defensin 

gene ligated downstream from the rice green-leaf-specific promoter, 

we constructed a binary vector incorporating an expression 

cassette consisting of the rice P10 terminator ligated downstream 

from each gene (Fig. 2). Transformation was accomplished 

by using seeds of rice variety Dontokoi. Estimation of 

the rice blast resistance and bacterial leaf blight allowed us to 

obtain transgenic plants that had been endowed with resistance 

to these pathogens. We confirmed that this resistance is stably 

inherited (Kawata et al 2003). This positively confirmed that 

the defensin gene makes it possible to develop transgenic rice 

strains with high resistance to fungi (the rice blast fungus) and 

bacteria (bacterial leaf blight). We also found that the resistance 

conferred by this gene is not specific to a particular race 


pCsp mALS P10 pLsp Def P10 

Fig. 2. Vector construction for the development of transgenic rice by making use of 

domestic gene technologies. pCsp = callus-specific promoter from rice, mALS = mutated 

acetolactate synthase gene as selection marker, P10 = rice 10K prolamine 

terminator, pLsp = green-tissue-specific promoter from rice, Def = defensin gene as 

disease-resistance gene. 

of rice blast fungus. The transgenic plants did not have any 

significant difference with a nontransgenic variety in traits such 

as leaf shape or fertility, and we were able to obtain progeny 

seeds. We analyzed the expression specificity of the callusspecific 

promoter and the green-leaf-specific promoter, and 

confirmed that expression was suppressed in the rice grains. 

We investigated whether defensin has any significant homology 

between known allergens and toxic proteins. With reference 

to the Report of a Joint FAO/WHO Expert Consultation 

(2001), homology was assessed as significant when either or 

both of the following were true: (1) overall homology is 35% 

or more and (2) eight or more neighboring amino acid sequences 

match. We confirmed that defensin peptides have no 

significant amino acid sequence homology either with known 

allergen proteins or with toxic proteins. 

Creating new demand 

There is potential consumer demand for natural antimicrobial 

agents that offer eco-friendliness and excellent safety (Zasloff 

2002). By focusing on the defensins, which exhibit broad antimicrobial 

activity, as new natural antimicrobial agents, we 

expect that new antimicrobial agents will be developed, leading 

to the creation of new demand among consumers. It has 

been shown that animal and insect defensins inhibit the growth 

of human disease-related bacteria (Saido-Sakanaka et al 1999). 

This indicates the possibility that these defensins can be used 

for the prevention and treatment of human diseases and food 

poisoning. Analytical research on plant defensins has also 

broadened in scope from only targeting plant pathogens, and 

there are hopes for the elucidation of pharmacological characteristics 

through antimicrobial activity spectrum analyses of 

diverse pathogenic organisms. 

References 

De Samblanx GW, Goderis IJ, Thevissen K, Raemaekers R, Fant F, 

Borremans F, Acland DP, Osborn RW, Patel S, Broekaert WF. 

1997. Mutational analysis of a plant defensin from radish 

(Raphanus sativus L.) reveals two adjacent sites important 

for antifungal activity. J. Biol. Chem. 272(2):1171-1179. 

Fant F, Vranken W, Broekaert W, Borremans F. 1998. Determination 

of the three dimensional solution structure of Raphanus sativus 

antifungal protein 1 by 1H NMR. J. Mol. Biol. 279:257-270. 

Kawata M, Nakajima T, Yamamoto T, Mori K, Oikawa T, Fukumoto 

F, Kuroda S. 2003. Genetic engineering for disease resistance 

in rice (Oryza sativa L.) using antimicrobial peptides. JARQ 

37(2):71-76. 

Kawata M, Nakajima T, Mori K, Oikawa T, Kuroda S. 2004. Genetic 

engineering for blast disease resistance in rice, using a 

plant defensin gene from Brassica species. Proceedings of 

the 3rd IRBC Congress. (In press.) 

Saido-Sakanaka H, Ishibashi J, Sagisaka A, Momotani E, Yamakawa 

M. 1999. Synthesis and characterization of bactericidal 

oligopeptides designed on the basis of an insect anti-bacterial 

peptide. Biochem. J. 338:29-33. 

Thevissen K, Warnecke DC, Francois IE, Leipelt M, Heinz E, Ott C, 

Zahringer U, Thomma BP, Ferket KK, Cammue BP. 2004. 

Defensins from insects and plants interact with fungal 

glucosylceramides. J. Biol. Chem. 279(6):3900-3905. 

Zasloff M. 2002. Antimicrobial peptides of multicellular organisms. 

Nature 415:389-395. 

Notes 

Author’s address: Hokuriku Department of Rice Research, National 

Agriculture Research Center, Inada, Joetsu, Niigata 943-0193, 

Japan, and Graduate School of Natural Science and Technology, 

Niigata University, Ikarashi, Niigata 950-2181, Japan. 

Golden Rice and improvement of human nutrition 

Swapan Datta, Vilas Parkhi, Mayank Rai, Jing Tan, Niranjan Baisakh, Lina Torrizo, Editha Abrigo, Norman Oliva, 

Md. Alamgir Hossain, Russel Julian, Anindya Bandyopadhyay, and Karabi Datta 

Deficiencies of vitamin A, iron, and zinc are widespread in 

Asia and in developing countries, where the main diet is plantbased 

and the staple food is rice. Rice provides 40–70% of the 

total calories consumed in developing Asian countries. In 

Southeast Asia, an estimated 250,000 people go blind each 

year because of vitamin-A deficiency (VAD). Diet is the only 

source of vitamin A in mammals since they cannot manufacture 

it on their own. Most of the dietary vitamin A is of plantfood 

origin in the form of provitamin A that is converted to 

vitamin A in the body (Sivakumar 1998). So far, none of the 


screened exotic rice germplasm has shown the presence of β- 

carotene in polished seeds. Nevertheless, there is tremendous 

potential for using the genetic variability of carotenoid levels 

in unpolished rice germplasm (Tan et al 2004). 

More than half of all women and children in South and 

Southeast Asian countries are anemic. Anemia limits growth 

and cognitive development in children and increases the incidence 

of death among severely anemic women during childbirth. 

Biofortified rice with enhanced minerals and provitamin 

A could complement existing nutrition interventions and 

provide a sustainable and low-cost way of reaching people who 

do not have access to the normal food market because of their 

poor economic conditions. Food consumption studies suggest 

that doubling the iron in rice can increase the iron intake of the 

poor by 50%. Rice germplasm screening shows that a doubling 

of iron and zinc in unmilled rice is feasible (Gregorio et 

al 2000). 


Several indica cultivars (IR64, BR29, Mot Bui, NHCD, BR28, 

IR68144) were used for transformation in conferring β-carotene 

biosynthesis in rice seeds. BR29, IR68144, and IR64 were 

used for iron improvement in rice endosperm. Transformation, 

regeneration, molecular analysis (PCR, Southern, RT-PCR), 

event selection, and biochemical analysis (carotenoid profiling 

by HPLC analysis) were conducted as described earlier 

(Datta et al 2003, Vasconcelos et al 2003, Tan et al 2004). The 

iron content of both polished and brown rice was estimated by 

using either atomic absorbance mass spectrometry (AA-MS) 

or inductively coupled plasma-optical emission spectrophotometry 

(ICP-OES). Transformation (biolistic and 

Agrobacterium) protocols were as described before (Datta et 

al 2003, Vasconcelos et al 2003). The pGPTV-bar/Fer vector 

for iron (Vasconcelos et al 2003) and pBaal3 and pCaCar 

(based on mannose selection) were used for β-carotene engineering 

in rice seeds (Datta et al 2003; unpublished data). 

Lysine improvement is in progress using the dapA gene from 

Corynobacterium, introduced in two rice cultivars by biolistic 

transformation. 


Iron deficiency is the leading human nutritional disorder in 

the world and very limited work has been done to reduce iron 

deficiency compared to progress in enhancing vitamin A and 

iodine in plant-based diets. Goto et al (1999) reported enhancement 

of iron in japonica rice seeds. We showed that the expression 

of the soybean ferritin gene, driven by the endospermspecific 

glutelin promoter, led to higher iron and zinc levels in 

transgenic rice grains. This is the first report showing enhanced 

zinc and iron in rice grains after polishing. Brown rice is rarely 

consumed and polishing of the rice grain brings about a considerable 

loss of micronutrients by removing the outer layers. 

Expression of the soybean ferritin gene under the control of 

the glutelin promoter in rice has proven to be effective in enhancing 

grain nutritional levels, not only in brown grain but 

also in polished grain (Vasconcelos et al 2003). 

Zinc deficiency has been associated with complications 

in pregnancy and delivery, as well as growth retardation, congenital 

abnormalities, and retarded neuro-behavioral and immunological 

development in the fetus (Vasconcelos et al 2003). 

Interestingly, we found that, along with enhanced iron, zinc 

levels were likewise increased in our transgenic ferritin rice. 

We obtained 27.9 µ g –1 iron and 55.5 µ g –1 zinc in unpolished 

transgenic rice seeds. This accounts for a 2–3-fold increase in 

iron compared with that of the control. The high levels of concentration 

of iron and zinc remained in rice grains after polishing 

(12–15 µ g –1 iron in transgenic seed in the T 3 generation 

versus 5–7 µ g –1 in control seed). Iron and zinc levels in 

rice grain may vary because of soil properties, weather, and 

some unknown factors (Gregorio et al 2000). 

Vitamin A plays an important role in a wide variety of 

physiological functions in human beings. Biosynthesis of carotenoids 

in plants takes place within the plastids, chloroplasts 

of photosynthetic tissues, and chromoplasts of fruits and flowers. 

Chlorophyll, tocopherols, plastoquinone, phylloquinone, 

gibberellins, and carotenoids all share a common biosynthetic 

precursor, geranylgeranyl/diphosphate (GGPP), which is derived 

from a plastidic isoprenoid metabolism. The genes necessary 

for the four enzymes (phytoene synthase, phytoene 

desaturase, ζ-carotene desaturase, and lycopene cyclase) catalyze 

to complete the pathway toward β-carotene (provitamin 

A) biosynthesis from GGPP (Misawa et al 1990). Conventional 

interventions (distribution, fortification, diet diversification) 

have been helpful in defeating VAD but are not effective 

enough. Conventional plant breeding to alter, modify, or introduce 

this biosynthetic machinery into the target endosperm 

tissues in rice has been impossible as of now, as no endospermactive 

carotenoid biosynthetic genes have been found thus far 

in the available rice gene pool (Tan et al 2004, Fig. 1). Thus, 

the transgenic approach offers the most realistic way to improve 

rice by incorporating genes coming from sources other 

than rice into rice and showing β-carotene expression in rice 

seeds (Figs. 1 and 2). Based on the proof of concept of β- 

carotene rice in japonica rice published by Ye et al (2000), we 

carried out this work in indica rice and several tropical indica 

cultivars (BR29, IR64, Mot Bui, NHCD, BR28, IR68144) were 

transformed and showed expression of β-carotene genes (Datta 

et al 2003, Parkhi et al, unpublished data, Fig. 2). 

Agrobacterium-mediated transformation along with the 

POSITECH TM (mannose selection) system works well with 

several cultivars. Expression of β-carotene varies among cultivars, 

as it was found that IR64 did not show much expression 

compared with BR29. Introgressed IR64 from original Golden 

Rice (T309) also showed lower carotenoid levels (Baisakh et 

al, unpublished data). Efforts have been made to select a 

transgenic event with a simple and low-copy transgene for a 

Mendelian 3:1 segregation pattern and eventually to develop 

homozygous lines with acceptable agronomic performance. 

The bioavailability of micronutrients (iron and zinc) and 

β-carotene needs to be studied. Since ferritin is used as a natu- 


AU KDML 105 

AU 

Ase Pindjau 

0.040 

0.035 

0.030 

0.025 

0.020 

0.015 

0.010 

0.005 

0.000 

–0.005 

–0.010 

10 20 30 40 50 60 70 

0.040 

0.035 

0.030 

0.025 

0.020 

0.015 

0.010 

0.005 

0.000 

lut 

b-crt 

5 10 15 20 25 30 

AU 

0.040 

0.035 

0.030 

0.025 

0.020 

0.015 

0.010 

0.005 

0.000 

IR64 (transgenic) 

lut 

5 10 15 20 25 30 5 10 15 20 25 30 

Minutes 

b-crt 

Fig. 1. Carotenoid profile for selected rice cultivars under both polished and unpolished conditions. 

Curves in red = polished seeds. Curves in blue = unpolished seeds. Curves in green = control seeds. 

KDML 105 and Ase Pindjau (from Thailand and Indonesia) are nontransgenic seeds, whereas IR64 and 

BR29 are genetically engineered Golden indica rice (lut = lutein, β-crt = β-carotene). 

AU 

0.040 

0.035 

0.030 

0.025 

0.020 

0.015 

0.010 

0.005 

0.000 

BR29 (transgenic) 

lut 

Minutes 

b-crt 

Fig. 2. Golden indica BR29 (control at left and transgenic segregating 

line at right) showing expression of beta-carotene in 

rice seeds after polishing (carotenoid levels ranged from 0.5 

to 3.14 µg g –1 . Transformation was carried out with mannose 

selection (PMI) and the Agrobacterium system. 

ral source of iron in the early development of animals and 

plants, ferritin-iron bioavailability should not be a problem. In 

fact, iron in transgenic rice with ferritin has already been tested 

in iron-deficient rats, and rice diets were as effective as the 

FeSO 4 diet in replenishing hematocrit (Murray-Kolb et al 

2002). Further studies suggest that a relatively small supplement 

of vitamin A or β-carotene can double the absorption of 

endogenous nonhaem iron from cereal, resulting in a significant 

increase in carotenoid in cereal diets (Graham and Rosser 

2000). Thus, β-carotene-enriched rice helps reduce vitamin-A 

deficiency and protect against iron-deficiency anemia. 

Biofortified rice with enhanced nutrition (β-carotene, iron, zinc, 

and lysine) along with other superior agronomic traits and yield 

could support the food-health and livelihood of the millions of 

people who need them most. 

References 

Datta K, Baisakh N, Oliva N, Torrizo L, Abrigo E, Tan J, Rai M, 

Rehana S, Al-Babili S, Beyer P, Potrykus I, Datta SK. 2003. 

Bioengineered ‘golden’ indica rice cultivars with beta-carotene 

metabolism in the endosperm with hygromycin and mannose 

selection systems. Plant Biotechnol. J. 1:81-90. 

Goto F, Yoshihara T, Shigemoto N, Toki S, Takaiwa F. 1999. Iron 

fortification of rice seed by the soybean ferritin gene. Nature 

Biotechnol. 17:282-286. 


Graham RD, Rosser JM. 2000. Carotenoids in staple foods: their 

potential to improve human nutrition. Food Nutr. Bull. 

21(4):404-409. 

Gregorio G, Senadhira D, Htut H, Graham R. 2000. Breeding for 

trace mineral density in rice. Food Nutr. Bull. 21(4):382-387. 

Misawa N, Nakagawa M, Kobayashi K, Yamano S, Izawa Y, 

Nakamura K, Harashima K. 1990. Elucidation of the Erwinia 

uredovora carotenoid biosynthetic pathway by functional 

analysis of gene products expressed in Escherichia coli. J. 

Bacteriol. 172:6704-6712. 

Murray-Kolb L, Takaiawa F, Goto F, Yoshihara T, Theil E, Beard J. 

2002. Transgenic rice is a source of iron for iron-depleted 

rats. J. Nutr. 132(5):957-960. 

Sivakumar B. 1998. Current controversies in carotene nutrition. Ind. 

J. Med. Res. 108:157-166. 

Tan J, Baisakh N, Oliva N, Torrizo L, Abrigo E, Datta K, Datta SK. 

2004. HPLC carotenoid profiles of rice germplasm. Int. J. Food 

Sci. Technol. (Submitted.) 


Krishnan S, Oliveira M, Goto F, Datta SK. 2003. Enhanced 


gene. Plant Sci. 64(3):371-378. 

Ye X, Al-Babili S, Klöti A, Zhang J, Lucca P, Beyer P, Potrykus I. 

2000. Engineering the provitamin A (β-carotene) biosynthetic 

pathway into (carotenoid-free) rice endosperm. Science 

287:303-305. 

Notes 

Authors’ address: Plant Breeding, Genetics, and Biochemistry Division, 

International Rice Research Institute, DAPO Box 7777, 

Metro Manila, Philippines. 

Acknowledgments: We thank the USAID and HarvestPlus Challenge 

Program (“Breeding Crops for Better Nutrition”) for financial 

and research support and Dr. Peter Beyer for providing 

us with the pCaCar and pBaal3, Dr. S.C. Falco (DuPont) for 

dapA, Syngenta for pmi, and Dr. F. Goto for the ferritin gene. 

Health-promoting transgenic rice suppressing 

life-related disease and type-I allergy 

Fumio Takaiwa 

Higher plants are an attractive host for the production of recombinant 

proteins for pharmaceutical and nutraceutical applications. 

Plant production offers many advantages, the most 

prominent of which are cost savings relative to animal 

transgenic systems, easy control of the production scale, and 

freedom from animal-derived pathogens. When recombinant 

proteins are expressed in seed under the control of a seedspecific 

promoter, they are highly and stably accumulated in a 

specific organelle designated as a protein body. Taking advantage 

of these properties, the seed has been used as an ideal 

bio-reactor for the production of recombinant proteins. 

Rice is one of the most important crops and food resources 

in the world. Rice seed has several advantages over 

the other cereal crops, such as easier storage and processing, 

greater biomass, and lower production costs. In addition, a 

transformation system of rice has been established and the 

whole genome sequence, including mitochondria and chloroplasts, 

is now available. Therefore, rice seed can be used as a 

good platform for producing recombinant proteins. 

The rice seed production system 

For use as a bio-reactor for the production of recombinant proteins 

or a direct delivery system of bioactive peptides or vaccines, 

establishment of a high expression system is required. 

Final accumulation levels are regulated at several steps of tran- 

scription and posttranscription, including translation, mRNA 

stability, transport, protein stability, and folding, among others. 

It has been found that the transcription level primarily 

determines the expression level. Therefore, screening for strong 

promoters from genes expressed in maturing seed has been 

carried out first. Promoter activities of more than 20 genes 

were assayed in stable transgenic rice seed by fusing their promoters 

to the GUS reporter gene, and then introducing their 

chimeric genes into the rice genome by Agrobacterium-mediated 

transformation. From the results of this promoter analysis, 

rice seed promoters are classified into several types based 

on their expression patterns in seed. It was shown that glutelin, 

26-kDa globulin, and 10-kDa prolamin promoters exhibited 

strong promoter activities in this order and different expression 

patterns in the endosperm (Qu and Takaiwa 2004). 

Prolamin, glutelin, and globulin promoters directed expression 

in the outer layer, subaleurone layer, and inner starchy 

endosperm, respectively. Therefore, we have selected and used 

the glutelin, 26-kDa globulin, and 10-kDa prolamin promoters 

for the expression of foreign proteins in transgenic rice 

seed dependent upon that purpose. 

Irrespective of high transcripts, it has been occasionally 

observed that final products were at a very low level. In this 

case, the space for accumulation of recombinant proteins or 

subcellular localization as well as codon usage or mRNA stability 

is critical. Low-storage-protein mutants such as LGC-1 


pmol (4 MU min –1 mg –1 protein) 

60 

50 

40 

30 

20 

10 

1.3 kB GluB-1 

2.3 kB GluB-1 

GluB-2 

GluB-4 

10 kDa Pro 

13 kDa Pro 

16 kDa Pro 

Glb-1 

REG-2 

Ole18 

b-conglycinin 

AlaAT 

GOGAT 

AGPase 

PPDK 

SBE1 

Ubiquitin 

Fig. 1. GUS activities expressed by various promoters in maturing seeds at 17 days after flowering. 

GUS activity is expressed as pmol (4 MU min –1 µg –1 protein). 

will become good hosts, giving rise to a high accumulation of 

foreign proteins, because they can provide ample space for 

accumulation. It has been confirmed by crossing between 

transgenic rice and a mutant that the transgene product was 

enhanced more than twofold in a low storage background (Tada 

et al 2003). 

Transgenic rice accumulating soybean glycinin 

Soybean protein has a function to lower the cholesterol level 

in human serum and is one of the best plant food proteins in 

terms of nutritional quality. The major storage proteins of soybean 

are glycinin and β-conglycinin. Glycinin is superior to 

β-conglycinin in nutritional value as well as in functional properties. 

Soybean proteins are deficient in sulfur-containing amino 

acids and rich in lysine, and rice proteins are the opposite. 

Therefore, if soybean glycinin were highly accumulated in rice 

seeds, an improvement of nutritional value and endowment of 

functional properties such as gel-formation and emulsifying 

abilities would be expected. We previously generated 

transgenic rice accumulating soybean glycinin A1aB1b 

(Katsube et al 1999). In this study, two representative types of 

glycinin (A1aB1b and A3B4) belonging to different subclasses 

were expressed in the endosperm tissue under the control of 

glutelin GluB-1 and 26-kDa globulin promoters, respectively. 

Furthermore, to enhance their accumulation levels, they were 

transferred into the low-storage-protein background of LGC- 

1 by crossing. It was shown that A1aB1b was much more highly 

accumulated than A3B4 in transgenic rice. When both glycinins 

were expressed together, the A3B4 level increased about threefold 

more than the level obtained by the expression of A3B4 

alone, thus indicating a requirement of assembly between different 

subunits for the stable accumulation of A3B4. In the 

LGC background, soybean glycinin was accumulated as a major 

storage protein, accounting for 30% to 40% of total rice seed 

protein. The glycinin level also increased more than twice compared 

with that of the normal background. Glycinin was extracted 

as salt-soluble globulin and acid-soluble glutelin fractions, 

thus suggesting that hybrid glycinin-glutelin oligomers 

were also formed in transgenic rice seed. When the lysine and 

protein levels in the seed were examined and compared between 

a transgenic line expressing both glycinins and the lowglutelin 

mutant background line, they were remarkably enhanced 

from 0.23 to 0.41 g per 100 g of seed and from 6.9 to 

10.9 g per 100 g of seed, respectively. 

Transgenic rice accumulating bioactive peptides 

It has been observed that some bioactive peptides derived from 

digests of food proteins are active after oral administration 

even though specific activities are not high compared with the 

endogenous bioactive peptides. Such peptides are expected to 

be effective in preventing lifestyle-related diseases such as 

hypertension, high cholesterol, and diabetes, among others. 

Biological activities of peptides could be strengthened by replacement 

of amino acid residues. Taking advantage of these 


modified peptides with high specific activity, transgenic crops 

to prevent lifestyle-related diseases can be generated. 

An antihypertensive peptide ovokinin-like sequence derived 

from egg white alubumin (Matoba et al 2001) was inserted 

into highly variable regions of seed storage proteins to 

confer upon them a potent anti-hypertensive function. Bioactive 

peptides were usually introduced into hypervariable regions 

of rice’s major storage protein glutelins, which are located in 

the C terminal regions of acidic and basic subunits. In this 

study, tandem modified ovokinin was inserted into at least two 

sites of the C terminal region of the acidic subunit of glutelin. 

To release bioactive peptide from the seed storage protein, the 

bioactive peptide was flanked by Asn or Lyn amino acids, the 

target site of the typsin small-intestine digestive enzyme. It is 

anticipated that bioactive peptide could be released from storage 

protein in the small intestine after digestion with digestive 

enzymes after oral feeding. This functional glutelin gene containing 

antihypertensive peptides was transcriptionally fused 

to the rice glutelin promoter and then introduced into the rice 

genome. The modified glutelin was highly accumulated in rice 

seed. On the other hand, when a small size of bioactive peptide 

of less than 30 amino acids was directly synthesized in 

transgenic plants, transgenic plants accumulating it have not 

yet been obtained. This is due to rapid clearance in the plant 

cell after synthesis. Therefore, it has been demonstrated that 

the best way for production in transgenic plants is for bioactive 

peptide to be synthesized as a part of the seed storage protein. 

Transgenic rice alleviating Japanese cedar pollinosis 

Peptide immunotherapy using dominant T-cell epitopes is safer 

and more effective than the conventional treatment of type-I 

allergic diseases due to no binding to the specific IgE antibody 

and high dose administration. It is possible to use rice 

seed as a direct delivery system for peptide vaccine. Japanese 

cedar pollinosis is an important allergic rhinoconjunctivitis in 

Japan in spring. We have generated transgenic rice accumulating 

epitope peptide of Japanese cedar pollen allergens (Cryj1 

and Cryj2) in the endosperm. The codon-optimized gene coding 

for hybrid epitope peptide consisting of seven major human 

T-cell epitopes (96 amino acids) derived from the cedar 

pollen allergens was synthesized (Hirahara et al 2001). To enhance 

the accumulation level of the 7 crp peptide, glutelin 

GluB-1 signal peptide and the KDEL ER retention signal were 

attached to the N and C termini of the 7 crp peptide, respectively. 

When this gene was introduced and expressed in the 

endosperm tissue of rice under the control of the glutelin GluB- 

1 promoter, the 7 crp peptide was detected as a visible band in 

CBB-stained gel. About 60 µg of 7 crp peptide was accumulated 

in one grain of the highest accumulating line, accounting 

for about 3% to 4% of the total seed protein (Fig. 2). The 7 crp 

peptide in seed was not glycosylated and was resistant to boiling 

for 20 min. Oral administration of these transgenic rice 

seeds to specific mice before priming with pollen proteins suppressed 

the IgE level to about one-third of that obtained by 

feeding nontrasgenic rice seeds, thus indicating that immunological 

tolerance was induced by this transgenic rice. Taken 

together, these results provide insight into the efficacy of using 

transgenic rice for oral immunotherapy against Japanese 

cedar pollinosis. 

References 

Hirahara K, Tatsuta T, Takatori T, et al. 2001. Preclinical evaluation 

of an immunotherapeutic peptide comprising 7 T-cell determinants 

of Cry j1 and Cry j2 , the major Japanese cedar pollen 

allergens. J. Allergy Clin. Immunol. 108:94-100. 

Katsube T, Kurisaka N, Ogawa M, Maruyama M, Ohtsuka R, Utsumi 

S, Takaiwa F. 1999. Accumulation of soybean glycinin and its 

assembly with the glutelins in rice. Plant Physiol. 120:1063- 

1073. 

Matoba N, Doyama N, Yamada Y, Maruyama N, Utsumi S, Yoshikawa 

M. 2001. Design and production of genetically modified soybean 

protein with anti-hypertensive activity by incorporating 

potent analogue of ovokinin(2-7). FEBS Lett. 497:50-54. 

Qu LQ, Takaiwa F. 2004. Evaluation of tissue specificity and expression 

strength of rice seed component gene promoters in 

transgenic rice. Plant Biotechnol. J. 2:113-125. 

Tada Y, Utsumi S, Takaiwa F. 2003. Foreign gene products can be 

enhanced by introduction into low storage protein mutants. 

Plant Biotechnol. J. 1:411-422. 

Notes 

Author’s address: Laboratory of Gene Engineering, Department of 

Plant Biotechnology, National Institute of Agrobiological 

Sciences, Kannondai 2-1-2, Tsukuba, Ibaraki 305-8602, Japan. 


A 

2.3 k GluB-1 pro GluB-1 sig 7 crp KDEL 0.6 k GluB-1 3′ 

B 

Western 

analysis 

CBB staining 

NT TF 

NT 

TF 

kDa 

kDa 

75 

75 

50 

50 

37 

Glutelin 

(acidic) 

37 

25 

Glutelin 

(basic) 

25 

15 

Prolamin 

7 crp 

15 

10 

10 

7 Crp 

Fig. 2. (A) Vector construct used to generate transgenic rice expressing the 7 crp gene in 

endosperm tissue. (B) Accumulation of the 7 Crp peptide in seeds of transgenic rice plants. 

Total protein was extracted from a seed of nontransformant (NT) or transgenic rice (TF: 

pGPTV-HPT-GluB-1pro sig/7Crp+KDEL #10), separated by 15% SDS-PAGE, and then stained 

with CBB or immunoblotted with anti-7 Crp antibody. 

Producing rice plants with a site-specific base change 

in the acetolactate synthase gene by chimeraplast-directed 

gene targeting 

A. Okuzaki and K. Toriyama 

Chimeraplasts, which are also called chimeric RNA/DNA oligonucleotides, 

have been reported to cause site-specific base 

changes in chromosomal targets in mammalian cells (see Hohn 

and Puchta 1999, Oh and May 2001, Graham and Dickson 

2002 for reviews). Chimeraplasts consist of 68 synthesized 

oligonucleotides that have a DNA “mutator” region of five 

nucleotides complementary to the target site flanked by 2′-Omethyl 

RNA bridges of 8–12 nucleotides each. This DNA mutator 

region can cause a site-specific base change in the endogenous 

gene. The first uses of chimeraplasts causing sitespecific 

base changes in plant cells were reported in maize 

(Zhu et al 1999) and tobacco (Beetham et al 1999). In both 

cases, gene targeting has been directed toward generating base 

changes that result in a chemically selectable phenotype. The 

targeted gene has been an acetohydroxyacid synthase (AHAS) 

gene in maize (Zhu et al 1999, 2000) and an acetolactate synthase 

(ALS) gene in tobacco (Beetham et al 1999). Both ALS 

and AHAS are the first common enzymes in the biosynthetic 

pathway of the branched-chain amino acids leucine, isoleusine, 

and valine. Mutations of certain amino acids in these proteins 

have been known to confer resistance to ALS-inhibiting herbicides 

(Shimizu et al 2002). The efficiency of gene conversion 

mediated by chimeraplasts in maize was estimated to be 10 –4 . 

In the same manner, we demonstrated that chimeraplastdirected 

gene targeting is feasible in rice (Okuzaki and 

Toriyama 2004). In our current study, we further confirmed 

the feasibility in rice using anther culture–derived haploid cells 

as well as scutellum-derived diploid cells. We also investigated 

whether the haploid state promoted the efficiency of 

chimeraplast-directed gene targeting. 


′ ′ 

Fig. 1. Chimeraplast Ch-W548L, which is designed to substitute TTG (a leucine codon) at codon 548 for 

TGG (a tryptophan codon). Uppercase letters, DNA residues; lowercase letters, 2′-0-methyl-RNA residues; 

bold letters, target codon. 


A japonica rice variety, Taichung 65 (Oryza sativa L.), was 

used in our study. A 6–8-day-old callus derived from scutellum 

of mature seeds and a 5-week-old callus derived from 

anther culture were used as target cells. The calli were placed 

on a plate in a circle 2.5 cm in diameter for delivery of 

chimeraplasts. 

Chimeraplast design and delivery 

Three chimeraplasts were designed based on the sequence of 

the rice ALS gene (accession numbers AB049822 and 

AB049823 in the DDBJ databases). A Ch-W548L (Fig. 1) was 

designed to substitute TTG (a leucine codon) at codon 548 for 

TGG (a tryptophan codon). In chimeraplast S627I, AGT for 

serine was replaced by ATT for isoleucine. Ch-P171A was 

designed with the Pro-171 codon CCC altered to contain a 

mismatch codon, GCC, which encodes alanine. Chimeraplasts 

were delivered to callus cells by particle bombardment 

(Okuzaki and Toriyama 2004). Calli were selected on N6 medium 

supplemented with 0.25–0.5 µM Bispyribac-sodium (BS, 

Kumiai Chemical Industry Co., Ltd., Shizuoka, Japan). A plasmid 

containing cauliflower mosaic virus 35S promoter and 

green fluorescent protein (35S-GFP) was also delivered for 

the estimation of the number of cells receiving gold particles. 

Molecular characterization of herbicide-resistant 

plants 

Genomic DNA was extracted from the leaves of regenerated 

plants. Target sequences were amplified from the DNA using 

Ex Taq polymerase and the ALS primers (Okuzaki and 

Toriyama 2004). Amplified fragments were gel-purified and 

were either directly sequenced or sequenced after cloning into 

a pCR 2.1 TOPO vector. 


Feasibility of chimeraplast-directed gene targeting 

At first, we delivered a mixture of Ch-W548L and Ch-S627I 

to scutellum-derived calli by particle bombardment. Three independent 

calli resistant to BS were obtained out of 20 plates 

of calli (Table 1). Direct sequencing of the PCR products of 

the regenerated plants showed two overlapping peaks of T and 

G, and consequently the N nucleotide designation in the chromatograms 

at the target nucleotide in codon 548, in the same 

manner as we reported previously (Okuzaki and Toriyama 

2004). However, no base changes were observed in the whole 

coding region of the ALS gene, including another target codon, 

627. Sequencing of the cloned PCR products demonstrated 

that both unconverted wild-type (TGG) and converted mutant 

alleles (TTG) were present in codon 548 in all three plants. 

The base change in ALS was transmitted to T 1 seedlings. The 

frequency of calli having predicted conversion at the target 

site of W548L after delivery of Ch-W548L was 0.12 per plate. 

Out of 44 control plates, which included the delivery of Ch- 

S627I, delivery of gold particles only, and no bombardment 

treatment, 4 plants showed resistance to BS (Table 1). These 

plants also contained the same base change in codon 548, which 

seemed to be caused by spontaneous mutation. The frequency 

of spontaneous mutation of W548L was calculated to be 0.09 

per plate. 

Next, we delivered chimeraplasts to anther culture–derived 

calli. Bombardment of a mixture of three chimeraplasts 

(Ch-W548L + Ch-S627I + Ch-P171A) and a mixture of two, 

including Ch-W548L, yielded 22 BS-resistant calli, all of which 

contained the predicted conversion sequence, TTG, at Trp-548 

codon (Table 1). Out of 23 control plates, in which Ch-W548L 

was not delivered, 11 calli showed resistance to BS, having 

the same base change, TTG, in codon 548, which was likely 

caused by spontaneous mutation. The frequency of calli having 

a predicted conversion at the target site of Ch-W548L after 

delivery of Ch-W548L was 0.96 per plate, and was twice 

as high as the frequency of spontaneous mutation, which was 

0.47 per plate. These results confirmed that the chimeraplastdirected 

gene targeting was feasible in rice. 

Estimation of gene conversion frequency per cell 

We estimated gene conversion efficiency using the following 

calculation: Gene conversion frequency = (Number of calli 

having predicted gene conversion per plate – Number of calli 

having spontaneous mutation per plate) / Total cells receiving 

chimeraplasts per plate. The total cells receiving chimeraplasts 

were estimated by transient expression of 35S-GFP. Gene conversion 

frequency mediated by chimeraplasts was estimated 

to be 3.3 × 10 –5 and 3.2 × 10 –4 for scutellum-derived calli and 

anther culture–derived calli, respectively. The frequency of 

chimeraplast-directed ALS W548L change in anther-culture 

calli was ten times higher than that of scutellum calli. 


Table 1. Number of calli having the predicted base change when chimeraplast was 

delivered to scutellum calli or anther-culture calli. In parentheses is the frequency of 

gene conversion (TTG) at codon 548 per plate. 

Number of Number of calli having predicted 

Delivered chimeraplast plates bombarded base change 

W548L S627I P171A 

Scutellum calli 

Ch-W548L + Ch-S627I 20 3 0 0 

Ch-W548L 5 0 0 0 

Total with Ch-W548L 25 3 (0.12) 

Ch-S627I 6 2 0 0 

Gold particles only 13 1 0 0 

No bombardment 25 1 0 0 

Total without Ch-W548L 44 4 (0.09) 

Anther-culture calli 

Ch-W548L + Ch-S627I + ChP171A 8 13 2 0 

Ch-W548L + Ch-S627I 7 5 0 0 

Ch-W548L + Ch-P171A 8 4 1 0 

Total with Ch-W548L 23 22 (0.96) 

Ch-S627I + Ch-P171A 8 5 0 0 

Gold particles only 7 1 0 0 

No bombardment 8 5 0 0 

Total without Ch-W548L 23 11 (0.48) 

Does the haploid state promote targeting efficiency 

In this study, we used anther culture–derived calli to investigate 

whether the haploid state promotes the efficiency of 

chimeraplast-directed gene targeting. Ploidy analysis of anther-derived 

calli after 5 wk of culture, which was carried out 

using flow cytometry after staining nuclei with DAPI, showed 

that these calli contained 60% haploid-state cells. The ploidy 

of calli having a predicted gene conversion was estimated from 

the sequence pattern of direct sequencing analysis. If there is a 

single peak of the converted nucleotide at a target site, gene 

conversion is considered to occur in a haploid-state cell. In 

contrast, double peaks of the converted and unconverted nucleotide 

indicate that gene conversion occurred in a doubled-haploid 

cell. Direct sequencing at codon 548 in 22 calli having 

W548L change after delivery of Ch-W548L demonstrated that 

8 calli had a single peak of T, whereas 13 calli had double 

peaks of G and T. For calli obtained by spontaneous mutation, 

six calli had a single peak of T and five had double peaks of T 

and G. There was no significant difference in the gene conversion 

frequency between haploid cells and doubled-haploid 

cells. This result indicated that the haploid state did not promote 

targeting efficiency in rice. A similar result has also been 

reported in tobacco (Kochevenko and Willmitzer 2003). 

Co-conversion at separate sites is the next subject 

In our study, calli containing co-conversion at two sites were 

not obtained in the co-delivery of two or three distinct 

chimeraplasts. Co-conversions at separate sites is the next subject 

to be examined in order to direct base changes for 

nonselectable genes. 

References 

Beetham PR, Kipp PB, Sawycky XL, Arntzen CJ, May GD. 1999. A 

tool for functional plant genomics: chimeric RNA/DNA oligonucleotides 

cause in vivo gene-specific mutations. Proc. 

Natl. Acad. Sci. USA 96:8774-8778. 

Graham IR, Dickson G. 2002. Gene repair and mutagenesis mediated 

by chimeric RNA-DNA oligonucleotides: chimeraplasty 

for gene therapy and conversion of single nucleotide polymorphisms 

(SNPs). Biochim. Biophys. Acta 1587:1-6. 

Hohn B, Puchta H. 1999. Gene therapy in plants. Proc. Natl. Acad. 

Sci. USA 96:8321-8323. 

Kochevenko A, Willmitzer L. 2003. Chimera RNA/DNA oligonucleotide-based 

site-specific modification of the tobacco 

acetolactate synthase gene. Plant Physiol. 132:174-184. 

Oh TJ, May GD. 2001. Oligonucleotide-directed plant gene targeting. 

Curr. Opin. Biotech. 12:169-172. 

Okuzaki A, Toriyama K. 2004. Chimeric RNA/DNA oligonucleotidedirected 

gene targeting in rice. Plant Cell Rep. 22:509-513. 

Shimizu T, Nakayama I, Nagayama K, Miyazawa T, Nezu Y. 2002. 

Acetolactate synthase inhibitors. In: Boger P, Wakabayashi 

K, Hirai K, editors. Herbicide classes in development. Berlin: 

Springer-Verlag. p 1-41. 

Zhu T, Peterson DJ, Tagliani L, St. Clair G, Baszczynski C, Bowen 

B. 1999. Targeted manipulation of maize genes in vivo using 

chimeric RNA/DNA oligonucleotides. Proc. Natl. Acad. Sci. 

USA 96:8768-8773. 

Zhu T, Mettenburg K, Peterson DJ, Tagliani L, Baszczynski CL. 

2000. Engineering herbicide-resistant maize using chimeric 

RNA/DNA oligonucleotides. Nat. Biotechnol. 18:555-558. 


Notes 

Authors’ address: Laboratory of Plant Breeding and Genetics, Graduate 

School of Agricultural Science, Tohoku University, Aobaku, 

Sendai 981-8555, Japan, e-mail: 

torikin@bios.tohoku.ac.jp, okku@bios.tohoku.ac.jp. 

Transgenic rice plants expressing wheat catalase 

show improved tolerance for chilling-induced damage 

in membranes 

Haruo Saruyama, Hidenori Onodera, and Matsuo Uemura 

Active oxygen species (AOS) such as superoxide radical 

(O 2 – ), hydrogen peroxide (H 2 O 2 ), and hydroxyl radical (OH • ) 

are generally produced during the biological process. The mitochondrial 

respiratory chain and the photosynthetic process 

are the major sites where AOS are produced. Indeed, approximately 

2–3% of the oxygen used by the mitochondria can be 

converted into superoxide and H 2 O 2 (Puntarulo et al 1988). 

Since plant growth depends on the balance between the production 

and detoxification of AOS, functional stability of AOSscavenging 

enzymes, such as superoxide dismutase, catalase, 

and ascorbate peroxidase, is critical for plant survival under 

stress environments. In fact, many transgenic plants expressing 

such antioxidant enzymes were reported. We previously 

confirmed that the expression of wheat catalase in transgenic 

rice plants shows tolerance for chilling injury (Matsumura et 

al 2002). Although improved tolerance for stress circumstances 

was reported, the functional mechanism of the tolerance has 

not been revealed. To make clear the reasons for improved 

tolerance in transgenic rice plants for low-temperature injury, 

we examined the effect of chilling (5 o C up to 8 days) on electrolyte 

leakage, mitochondrial membrane integrity, fatty acid 

peroxidation, and photosynthetic and respiratory activities. 


Transgenic rice plants (CT 2-6-4) in which wheat catalase 

cDNA was expressed and nontransgenic control rice plants 

(cultivar Yuukara) (Matsumura et al 2002) were used in this 

study. All plants were grown in a growth chamber for 20 days 

under conditions of 12 h light (262 µmol m –2 s –1 ) at 25 o C and 

12 h dark at 20 o C. 

The effects of chilling (12 h light, 60 µmol m –2 s –1 at 5 

o C and 12 h dark at 5 o C) on these plants were examined with 

the method previously reported (Matsumura et al 2002). 

Catalase (CAT; EC 1.11.1.6) activity in leaves was measured 

according to Matsumura et al (2002). H 2 O 2 concentration 

in leaves was measured according to Patterson et al (1984). 

Electrolyte leakage was measured as follows. Leaf cuttings 

(3 cm) from plants after various treatments were added 

to 3.0 mL of water in a test tube and shaken for 2.5 h. The 

conductivity of the solution (value 1) was measured with a 

conductivity meter (Twin Cond B-173, Horiba, Japan). Next, 

the samples were boiled for 30 min and shaken again for 2.5 h, 

and the conductivity (value 2) was measured. The electrolyte 

leakage value (%) was expressed as (value 1)/(value 2) times 

100. 

Mitochondrial activity was measured by TTC reduction 

according to Steponkus and Lanphear (1967). Fatty acid 

peroxidation was measured by the TBA reaction according to 

Heath and Packer (1968). Photosynthetic activity was examined 

by measuring the value of Fv/Fm using a Plant Efficiency 

Analyzer (Hansatech) according to Clarke and Campbell 

(1996). Respiratory activity was analyzed by measuring the 

decrease in oxygen using Oxytherm (Hansatech). 


Visible damage by chilling 

After plants were exposed at 5 o C for 8 d, visible damage by 

chilling was compared. As Figure 1 shows, leaves of 

nontransgenic rice after 8 d of chilling were as severely damaged 

as withered. However, such severe damage did not occur 

to the leaves of transgenic rice. These data clearly demonstrate 

that transgenic rice plants are more cold-tolerant than 

nontransgenic rice. 

Chilling effect on catalase activity and H 2 

O 2 

content 

While the catalase activity in leaves of nontransgenic rice at 

25 o C was around 100 (µmol mg –1 protein min –1 ), the transgenic 

rice showed values of approximately 700. After 4 d of chilling, 

the activities in both nontransgenic and transgenic rice 

were reduced to approximately 60 and 390, respectively (Fig. 

2). After 8 d, the transgenic rice showed 4 times higher residual 

activity than the control (Matsumura et al 2002). 

In the transgenic plants, the content of H 2 O 2 (about 0.36 

µmol g –1 FW) was kept low during chilling. In contrast, the 

nontransgenic plants showed an increased content of H 2 O 2 by 

chilling, and, at the end of the 8-d chilling period, the content 

reached a level two times that of the control (about 0.68 µmol 

g –1 FW). Thus, high catalase activities remaining in the 


Catalase activity (mmol min –1 mg –1 protein) 

800 

Nontransgenic 

Transgenic 

700 

600 

500 

400 

300 

Fig. 1. Visible damage from chilling. Transgenic and nontransgenic 

plants were grown in a growth chamber for 1 mo and then chilled 

at 5 °C for 8 d. (A) Nontransgenic rice plant (control), (B) transgenic 

rice plant. 

200 

100 

transgenic plants seem to prevent an increase in H 2 O 2 content 

after chilling. 

0 

0 1 2 3 4 

Chilling period (day) 

Electrolyte leakage induced by chilling 

Both the transgenic and nontransgenic plants showed the same 

leakage value (10%) after 2 d of chilling. However, the leakage 

of the control plants increased dramatically to approximately 

60% at 4 d and reached 69% at 8 d. In contrast, the 

transgenic plants indicated a gradual increase in the leakage, 

and the value was only 36% after 8 d. This lower electrolyte 

leakage, which is essential for membrane integrity (Uemura et 

al 1995), suggests that effective detoxification of H 2 O 2 by the 

elevated catalase activity contributes to the maintenance of 

membrane integrity in the transgenic plants at low temperatures. 

Effect of chilling on photosynthetic activity 

Photosynthetic activities in both transgenic and nontransgenic 

plants were decreased gradually by chilling. However, we could 

not find a difference in Fv/Fm values between them during the 

chilling up to 8 d. Because the Fv/Fm value indicated only the 

activity of photosystem II, other systems or components in the 

photosynthesis machinery may be considered as the injury sites. 

Fig. 2. Effect of chilling on catalase activity. Rice plants grown for 

1 mo were chilled at 5 °C for the indicated period and catalase 

activity in leaves was then measured at 25 °C. 

Effect of chilling on respiratory 

and mitochondrial activity 

Although respiratory activity was decreased by the chilling, 

approximately 80% of the activity remained after 8 d of chilling 

in both the transgenic and nontransgenic plants, suggesting 

that there were no differences in the chilling sensitivity of 

the respiratory chain between them. Next, we examined the 

effect of chilling on mitochondrial activity, with TTC reduction 

as an index of membrane integrity as reported by Steponkus 

(1971). TTC reduction did not change until 2 d of chilling in 

both the transgenic and nontransgenic plants. A prolonged chilling 

treatment resulted in a decrease in value in nontransgenic 

rice to 68% and 46% at 6 and 8 d, respectively. In contrast, the 

transgenic plants showed no decrease in value during chilling 

of up to 6 d and only decreased to 57% at 8 d. Therefore, we 

conclude that the oxidative damage from chilling on the membrane-associated 

mitochondrial function in transgenic plants 

was minimized but not in nontransgenic rice plants. The reason 

for the discrepancy between the results of the respiratory 

and TTC reduction activities, however, is not clear at this 

moment. 

Membrane fatty acid peroxidation 

Because membrane fatty acids are the major target of 

peroxidation by H 2 O 2 (Iturbe-Ormaetxe et al 1998), we examined 

the formation of peroxidation products of fatty acids, TBA, 


y chilling. When the value at the beginning of chilling was 

taken as 100%, the value of nontransgenic rice increased dramatically 

to 301%, 339%, and 343% at 4, 6, and 8 d of chilling, 

respectively. In the transgenic rice plants, in contrast, the 

value was lower than that of the nontransgenic plants (183% 

and 289% at 4 and 6 d, respectively). However, at 8 d, the 

value in the transgenic rice showed almost the same value as 

in the nontransgenic rice (344%). Thus, membrane fatty acids 

in the nontransgenic plants, but not in the transgenic plants, 

are considered to be readily peroxidized by H 2 O 2 by chilling. 

From these results, we concluded that the enhanced tolerance 

of chilling in transgenic rice plants is partly due to the 

effective removal of H 2 O 2 by the overexpressed catalase and 

increased stability of the membrane during chilling. 

References 

Clarke AK, Campbell D. 1996. Inactivation of the petE gene for 

plastocyanin lowers photosynthetic capacity and exacerbates 

chilling-induced photoinhibition in the Cyanobacterium 

Synechococcus. Plant Physiol. 112:1551-1561. 

Heath RL, Packer L. 1968. Photoperoxidation in isolated chloroplasts. 

I. Kinetics and stoichimetry of fatty acid peroxidation. 

Arch. Biochem. Biophys. 125:189-198. 

Iturbe-Ormaetxe I, Escuredo PR, Arrese-Igor C, Becana M. 1998. 

Oxidative damage in pea plants exposed to water deficit or 

paraquat. Plant Physiol. 116:173-181. 


Matsumura T, Tabayashi N, Kamagata Y, Souma C, Saruyama H. 

2002. Wheat catalase expressed in transgenic rice can improve 

tolerance against low temperature stress. Physiol. Plant. 

116:317-327. 

Patterson BD, MacRae EA, Ferguson IB. 1984. Estimation of hydrogen 

peroxide in plant extract using titanium (IV). Anal. 

Biochem. 139:487-494. 

Puntarulo S, Galleano M, Sanchez RA, Boveris A. 1988. Hydrogen 

peroxide metabolism in soybean embryonic axes at the onset 

of germination. Plant Physiol. 86:626-630. 

Steponkus PL, Lanphear FO. 1967. Refinement of the triphenyl tetrazolium 

chloride method of determining cold injury. Plant 

Physiol. 42:1423-1426. 

Steponkus PL. 1971. Effect of freezing on dehydrogenase activity 

and reduction of triphenyl tetrazolium chloride. Cryobiology 

8:570-573. 

Uemura M, Joseph RA, Steponkus PL. 1995. Cold acclimation of 

Arabidopsis thaliana: effect on plasma membrane lipid composition 

and freeze-induced lesions. Plant Physiol. 109:15- 

30. 

Notes 

Authors’ addresses: H. Saruyama, Hokkaido Green-Bio Institute, 

Naganuma, Hokkaido 069-1317, Japan, e-mail: 

66379635@people.or.jp; H. Onodera and M. Uemura, 

Cryobiosystem Research Center, Faculty of Agriculture, Iwate 

University, Morioka, Iwate 020-8550, Japan. 

Acknowledgments: This work was supported in part by a Grant-in- 

Aid for the 21st Century COE Program to Iwate University 

from the MEXT, Japan, to MU. 

Rice is the most important food crop for half the world’s population. 

Because of the growing population in Asian countries, more 

rice will have to be produced from less land for cultivation. Genetic 

engineering of rice could solve this problem by permitting 

higher productivity, including more resistance to biotic and abiotic 

stresses as an alternative technology of conventional breeding. 

Biotechnology using genetic engineering makes it possible 

to introduce valuable genes from all organisms into the rice genome. 

The application of genetic engineering is not limited to 

agriculture but is relevant to industry as molecular farming. Rice 

can be used as a bio-reactor for producing high added value 

such as biopharmaceuticals (antibodies, vaccines). 

In Session 3, six leading researchers reviewed the present 

status of studies on transgenic rice. Transgenic rice for improved 

input traits showing resistance to biotic and abiotic stress and 

output traits with improved quality of rice grain has been developed. 

To increase rice productivity, M. Tokutomi’s research group 

has introduced genes coding for maize C 4 -specific photosynthetic 

enzymes such as PEPC, PPDK, and NADP malic enzyme. Although 

transgenic rice overexpressing C 4 enzymes exhibited a higher 

photosynthetic rate, these plants showed little increase in yield. 

Tolerance of unfavorable circumstances such as not using Fedeficient 

soil is important for expanding cultivation land to enhance 

rice production worldwide. N. Nishizawa showed that 

transgenic rice plants were tolerant of low Fe availability in calcareous 

soil when barley genes encoding the enzymes involved 

in the biosynthesis of mugineic acid family phyotosiderophores 

(MAs) were introduced into the rice genome. This is because 

transgenic rice plants secrete higher amounts of deoxymugineic 

acid. Her group has characterized the transporter genes involved 

in the uptake and translocation of iron and zinc from the soil. 

Plant productivity is affected severely by environmental 

stresses such as cold, drought, and high salinity. K. Shinozaki’s 

group has identified a cis-regulatory element designated as the 

dehydration response element box (DREB), which is conserved 

in promoters of many genes implicated in gene regulation in response 

to the above stress. They demonstrated that the DREB 

transcription factor gene is quite useful for improving tolerance 

of environmental stresses in transgenic rice, when the DREB recognizing 

the conserved sequence was overexpressed under the 

control of a stress-inducible promoter. 

Yield is severely decreased by pathogen attacks such as 

fungal rice blast. Disease resistance can be enhanced by intro- 


ducing disease-resistance genes. M. Kawata suggested that an 

antimicrobial peptide such as defensin expressed in transgenic 

rice exhibited effective resistance against rice blast and bacterial 

leaf blight in field tests. 

To engineer transgenic rice acceptable to many consumers, 

the introduction of output traits with consumer benefits into 

rice grain is necessary. Output traits are those for improving not 

only macronutrients such as proteins, carbohydrates, and oils 

but also micronutrients such as vitamin A, iron, and zinc for food 

and feed. Furthermore, both these nutrients and biofunctional 

materials contributing to human health are targets of transgenic 

rice. Target characters are quite different between developing 

and developed countries. 

Many people in South Asia suffer from deficiency in iron 

and vitamin A, which causes blindness, anemia, and poor growth. 

Rice grain offers an excellent delivery system for these missing 

micronutrients because of the ease of storage and high daily 

consumption. S. Datta reported that iron-fortified rice and “golden 

rice,” rich in provitamin A, have been developed by expression of 

the iron storage protein ferritin and several enzyme genes involved 

in beta-carotene biosynthesis, respectively, in the endosperm 

tissue of several tropical indica rice varieties. This 

transgenic rice will help solve health problems in Asia. 

In developed countries, life-style-related diseases such as 

hypertension, diabetes, and high cholesterol are social problems 

and these have been gradually increasing. In addition, patients 

suffering from allergenic diseases such as pollen and food allergy 

are rapidly increasing. If rice were transformed to alleviate or 

prevent these diseases with an appropriate daily diet, transgenic 

rice grain would be of direct benefit to many consumers in developed 

countries, thus indicating a high potential for acceptance 

by consumers. F. Takaiwa discussed transgenic rice grain with a 

function to decrease serum cholesterol level by accumulating 

the soybean storage protein glycinin in the endosperm tissue. 

The soy-rice was rich in the essential amino acid lysine deficient 

in rice grain. 

It was demonstrated that oral feeding of transgenic rice 

grain to mice before sensitization with allergen suppressed the 

IgE level in spleen cells, indicating that oral feeding induces immunological 

tolerance. 


SESSION 4 

Improving rice yield potential 

CONVENER: M. Kondo (NARO) 

CO-CONVENERS: H. Ikehashi (Nihon Univ.) and S. Peng (IRRI)

Photosynthesis improvement in rice: 

Rubisco as a target for enhancing N-use efficiency 

Amane Makino 

Ribulose-1,5-bisphosphate (RuBP) carboxylase/oxygenase 

(Rubisco) catalyzes two competing reactions, photosynthetic 

CO 2 fixation and photorespiratory carbon oxidation. Quantitative 

analyses of the photosynthetic gas exchange and Rubisco 

activity in C 3 plants have revealed that Rubisco is a rate-limiting 

factor for potential photosynthesis under the present atmospheric 

air conditions. In addition, Rubisco is the most abundant 

leaf protein and large amounts of nitrogen are invested in 

this protein, accounting for 15% to 30% of leaf-N content. 

Therefore, genetic modification to enhance Rubisco efficiency 

in crops would have great agronomic importance as a prospect 

to enhance N-use efficiency for potential photosynthesis. 

This paper discusses the possible approaches for photosynthetic 

improvement by enhancing Rubisco efficiency in rice. 

Is enhancement of Rubisco efficiency feasible 

Large variations in the kinetic properties of Rubisco exist 

among enzymes from different groups of photosynthetic species 

(for a review, see Bainbridge et al 1995). Such natural 

variations suggested that it is possible to alter the enzyme to 

favor the specificity for CO 2 relative to O 2 and to enhance the 

specific activity (catalytic turnover rate per unit of enzyme). 

However, a negative correlation has also been found between 

the CO 2 /O 2 specificity and the catalytic turnover rate. In addition, 

Rubiscos from higher plants have the highest specificity 

for CO 2 to O 2 among all photosynthetic organisms, and their 

variation is very small. Although a higher CO 2 /O 2 specificity 

than higher-plant Rubiscos was previously discovered for a 

few red algae, the catalytic rate of algal Rubiscos was too low. 

Thus, it may be difficult to find a super Rubisco in nature. 

In spite of no variation in CO 2 /O 2 specificity among 

higher-plant Rubiscos, comparative studies of Rubisco activity 

among higher plants revealed significant differences in specific 

activity (Table 1). Rubisco from rice had the lowest specific 

activity, which was significantly lower than that from 

wheat. In addition, the difference in the specific activity between 

rice and wheat reflected the observed difference in the 

light-saturated rate of photosynthesis at atmospheric CO 2 levels 

(Fig. 1). These results indicate that Rubisco limits the potential 

photosynthesis in C 3 plants and that the lower specific 

activity from rice would make it possible to improve photosynthesis 

by introducing a more efficient Rubisco into rice 

cultivars. Unfortunately, no variability in the kinetic properties 

of Rubisco has been found among rice varieties, including 

old and modern cultivars (Makino et al 1987). Significant variations 

existed only on the basis of different genotypes in the 

Oryza genus (Makino et al 1987), suggesting that Rubisco 

Table 1. Rubisco-specific activity from several C 3 

higher plants. a 

Species 

Specific activity 

(mol CO 2 m –2 s –1 at 25 °C) 

Oryza sativa 15.5 ± 0.8 

Triticum aestivum 23.0 ± 0.4 

Triticum monococcum 16.2 ± 0.9 

Aegilops squarrosa 16.4 ± 0.3 

Spinacia oleracea 24.1 ± 0.8 

Phaseolus vulgaris 16.6 ± 0.2 

Pisum sativum 20.4 ± 0.4 

Nicotiana tobaccum 18.3 ± 1.1 

a Data are taken from Makino (2003) and Evans and Austin 

(1986). 

Photosynthesis at 36 Pa CO 2 (mmol m –2 s –1 ) 

50 

40 

30 

20 

10 

Triticum aestivum 

Oryza sativa 

0 0 1.0 2.0 3.0 4.0 5.0 

Rubisco (g m –2 ) 

Fig. 1. Light-saturated photosynthesis in air versus Rubisco content 

in leaves of rice (open circles) and wheat (closed circles) 

(Makino 2003). Measurements were made on leaves at different 

ages at a temperature of 25 °C, an irradiance of 1,800 µmol 

quanta m –2 s –1 , and an external CO 2 partial pressure of 36 Pa. 

cannot be a target for classical crossbreeding among cultivars. 

Similarly, in the Triticum genus, variation in the specific activity 

for a main cultivar of wheat, T. aestivum, was not significant, 

whereas there were some differences correlated with 

the genome constitution in Triticum and Aegilops genotypes 


(Evans and Austin 1986, see Table 1). In addition, Evans and 

Austin (1986) found that Rubisco from T. aestivum has the 

highest specific activity and that the occurrence of the highest 

specific activity is associated with possession of the B-type 

chloroplast genome, which encodes a large subunit of Rubisco. 

Although the divergence in the small subunit had been considered 

to be a more important causal factor for the variations in 

Rubisco activity because the large subunit is more conserved, 

the results of Evans and Austin suggested that the rbcL gene 

of a highly efficient Rubisco can be targeted for improving 

photosynthesis. 

Recently, technology for plastid transformation, at least 

with tobacco, has enabled a more precise manipulation and 

replacement of the chloroplast-coded large subunit of Rubisco 

(for a review, see Andrews and Whitney 2003). Such plastid 

transformation provides a transgenic tobacco with a hybrid of 

Rubisco with large and small subunits derived from different 

higher plants. For example, replacing tobacco rbcL with sunflower 

rbcL showed that, although the hybrid Rubisco is enzymatically 

active, its activity is extremely low. These results 

indicate that hybrid Rubisco with a large foreign subunit, once 

introduced to a host plant, does not function to its full potential. 

Thus, the replacement of Rubisco with an existing more 

efficient Rubisco requires the complete replacement of both 

large and small subunits. Of course, another problem with the 

manipulation of Rubisco is to develop plastid transformation 

techniques for the major crop species, including rice. 

Rubisco activase is essential for the full function of 

Rubisco activity, but Rubisco activase from tobacco is not effective 

for the activation of Rubisco from spinach, barley, and 

wheat. Therefore, it is possible that the replacement of Rubisco 

activase is also required for the full function of foreign Rubisco. 

Genetic manipulation of the amount of Rubisco 

An increase in Rubisco content theoretically leads to an improvement 

of photosynthesis under CO 2 -limited conditions. 

An attempt to increase Rubisco content by supplementation 

with an additional sense rbcS gene showed increases in Rubisco 

content for a given leaf-N content (Makino et al, unpublished). 

However, gas exchange analyses on transgenic rice with 

overexpressed Rubisco indicated that photosynthesis is CO 2 - 

saturated under normal CO 2 levels (37 Pa CO 2 ). These greater 

increases in Rubisco are assumed to lead to an imbalance 

among the in vivo capacities of Rubisco and other photosynthetic-limiting 

factors. This may have been caused by the nonspecific 

reallocation of N to Rubisco from other components 

limiting photosynthesis, which resulted in a decrease in the 

capacity for RuBP regeneration. 

When a leaf is senescent, the Rubisco content decreases 

faster than that of other photosynthetic components. Therefore, 

the maintenance of Rubisco content in a leaf during senescence 

also has agronomic importance. However, although 

the degradation of Rubisco is the major determinant of Rubisco 

content in a senescent leaf (Makino et al 1984), its degradation 

mechanism is largely unknown. Although the degradation 

of Rubisco by reactive oxygen was observed for isolated chloroplasts 

in the presence of inhibitors for antioxidant enzymes, 

the degradation process in a leaf may be more strictly regulated 

than expected. For example, a selective suppression of 

the degradation of Rubisco was suggested to occur in the senescent 

leaf of rbcS antisense rice with a decreased Rubisco 

content (Ishizuka et al 2004). 

What is a target for improving RuBP regeneration 

To maintain higher Rubisco efficiency, higher RuBP regeneration 

capacity is required. Comparative analyses of the photosynthetic 

characteristics between rice and wheat indicated 

that the capacity for RuBP regeneration is also greater in wheat 

than in rice (Sudo et al 2003). This higher capacity for RuBP 

regeneration in wheat was most closely related to a greater 

cytochrome f content. As RuBP regeneration capacity is limited 

by the electron transport capacity and/or Pi regeneration 

capacity during sucrose and starch syntheses, the cytochrome 

b6/f complex may conceivably be a factor in the RuBP regeneration 

capacity. Thus, the enhancement of the cytochrome b6/ 

f complex content may be another important target for improving 

photosynthesis to maintain the balance between the capacities 

of Rubisco and RuBP regeneration. 

Introduction of C 4 

characteristics into rice 

Generally, C 4 plants show a higher rate for photosynthetic CO 2 

assimilation than C 3 plants because C 4 plants have a mechanism 

for eliminating photorespiration by increasing the concentration 

of CO 2 around Rubisco. Therefore, another approach 

to improving photosynthesis is to introduce C 4 characteristics 

into rice (for a review, see Miyao 2003). However, even when 

photosynthesis was measured under saturating CO 2 levels (100 

Pa CO 2 ), the rate of photosynthesis for a given leaf-N content 

was higher in maize than in rice (Makino et al 2003). Since the 

amount of Rubisco in C 3 plants is excessive for CO 2 -saturated 

photosynthesis, it is possible that an excess investment of N in 

Rubisco in C 3 plants led to lower photosynthesis in rice than 

in maize under conditions of saturating CO 2 . However, the 

photosynthetic rate in maize leaves was still higher than that 

in leaves of the rbcS antisense rice with optimal Rubisco for 

CO 2 -saturated photosynthesis. These results indicate that, even 

if photorespiration is eliminated and Rubisco content is optimized 

for saturating CO 2 conditions, photosynthetic efficiency 

is still higher in maize than in rice. This higher efficiency of 

photosynthesis in maize resulted from the saving of N due to 

the lower amount of Rubisco of a higher V max /Km(CO 2 ) type 

and allowed a greater N investment in thylakoid components 

(Makino et al 2003). The N cost for the C 4 cycle enzymes was 

not large in maize, and the greater content of the thylakoid 

components may have led to an enhancement of RuBP regeneration. 

Recent genetic manipulation to express high levels of 

C 4 cycle enzymes from maize at target locations in rice is becoming 

well established (for a review, see Miyao 2003). Such 

Session 4: Improving rice yield potential 115

technology may lead to the eventual successful driving of the 

complete C 4 cycle in rice in the near future. To improve photosynthetic 

efficiency in rice, however, optimal protein allocation 

with a lower amount of Rubisco of a type different from 

that in C 3 plants, as well as the Kranz leaf anatomy to obtain 

an efficient CO 2 -concentrating mechanism, is essential. 

References 

Andrews TJ, Whitney SM. 2003. Manipulating ribulose bisphosphate 

carboxylase/oxygenase in the chloroplasts of higher plants. 

Arch. Biochem. Biophys. 414:159-169. 

Bainbridge G, Madgwick P, Parmar S, Mitchell R, Paul M, Pitts J, 

Key AJ, Parry MAJ. 1995. Engineering Rubisco to change its 

catalytic properties. J. Exp. Bot. 46:1269-1276. 

Evans JR, Austin RB. 1986. Ribulose-1,5-bisphosphate carboxylase 

specific activity in relation to wheat genotype. Planta 167:344- 

350. 

Ishizuka M, Makino A, Suzuki Y, Mae T. 2004. The amount of ribulose-1,5-bisphosphate 

carboxylase/oxygenase (Rubisco) protein 

and levels of mRNAs of rbcS and rbcL in the leaves at 

different positions in transgenic rice plants with decreased 

content of Rubisco. Soil Sci. Plant Nutr. 50:233-239. 

Makino A. 2003. Rubisco and nitrogen relationships in rice: leaf 

photosynthesis and plant growth. Soil Sci. Plant Nutr. 49:319- 

327. 

Makino A, Mae T, Ohira K. 1984. Relation between nitrogen and 

ribulose-1,5-bisphosphate carboxylase in rice leaves from 

emergence through senescence. Plant Cell Physiol. 25:429- 

437. 

Makino A, Mae T, Ohira K. 1987. Variations in the contents and 

kinetic properties of ribulose-1,5-bisphosphate carboxylases 

among rice species. Plant Cell Physiol. 28:799-804. 

Makino A, Sakuma H, Sudo E, Mae T. 2003. Differences between 

maize and rice in N-use efficiency for photosynthesis and protein 

allocation. Plant Cell Physiol. 44:952-956. 

Miyao M. 2003. Molecular evolution and genetic engineering of C 4 

photosynthetic enzymes. J. Exp. Bot. 54:179-189. 

Sudo E, Makino A, Mae T. 2003. Differences between rice and wheat 

in ribulose-1,5-bisphosphate regeneration capacity per unit 

of leaf-N content. Plant Cell Environ. 26:255-263. 

Notes 

Author’s address: Department of Applied Plant Science, Graduate 

School of Agricultural Science, Tohoku University, 1-1 

Tsutsumidori-Amamiyamachi, Aoba-ku, Sendai 981-8555, 

Japan. 

Carbon metabolism to improve sink and source function 

R. Ohsugi 

Sink and source function and their relationship are important 

aspects to understand the regulatory mechanism of crop production 

such as rice. In terms of carbon metabolism, the source 

function in rice leaves includes photosynthetic carbon fixation, 

sucrose formation, and its loading into sieve elements. 

The sink function in rice panicles includes sucrose unloading, 

its breakdown, and starch biosynthesis. In addition, leaf sheaths 

and culms show sink and source functions before and after 

heading, respectively. In the latter stage, starch breakdown is 

a function of major carbon metabolism. 

Rice genome research has been developing widely, providing 

a huge amount of gene information, DNA markers, and 

high-throughput technologies such as DNA microarray, allowing 

us to analyze these complex networks of carbon metabolism. 

At the same time, this makes it possible to elucidate the 

genetic basis of physiological traits such as sink and source 

function using a genetic map and DNA markers. 

I introduce some examples of these new approaches on 

sink and source functions. 

Source function 

Recent progress in the generation of a molecular genetic map 

and markers has made it possible to map individual genes as- 

sociated with complex traits, quantitative trait loci (QTLs). 

Ishimaru et al (2001a) applied QTL analysis to physiological 

traits, including sink and source functions. Using backcross 

inbred lines of Nipponbare (japonica) × Kasalath (indica), a 

rice function map was constructed by collating the results on 

QTLs for 23 important agronomic, physiological, and morphological 

traits. The source function-related traits such as 

chlorophyll content in leaves, photosynthetic ability, and δ 13 C 

value were mapped. These traits were first mapped in the QTL 

analysis of rice in their study. One of three QTLs controlling 

chlorophyll content partly overlapped with that controlling 

yield, although most of these source function-related traits did 

not overlap with that controlling yield. 

Among these source function-related QTLs, Ishimaru et 

al (2001b) further analyzed Rubisco content, soluble protein 

content, and nitrogen content in flag leaves of rice. The kinetics 

of Rubisco and the ratio of Rubisco to total leaf nitrogen 

(N) are the two main factors determining N-use efficiency for 

photosynthetic carbon fixation. Since the kinetics of Rubisco 

is very difficult to change, the ratio of Rubisco to total leaf N 

content is the main target for the improvement of N-use efficiency. 

In their function map, QTLs controlling Rubisco content 

were not detected near QTLs for total leaf N content, indicating 

that contents of Rubisco and total leaf N are controlled 


y different genetics. QTLs that controlled the ratio of Rubisco 

to total leaf N were detected, suggesting that some genetic 

factor(s) may be involved in determining this ratio. 

Temporary sink function (sink to source transition) 

Grain filling and yield are affected by the amount of carbohydrates 

supplied from two organs: mature leaves and leaf sheaths 

plus culms. Nonstructural carbohydrates (NSCs) in leaf sheaths 

and culms are mainly translocated to panicles, and play an 

important role in compensating for the lack of a source supply 

by photosynthesis after heading. NSC accumulation may be 

particularly important for grain filling in high-yielding cultivars 

that bear a large number of spikelets, which require a large 

supply of carbohydrates. 

Nagata et al (2002) applied QTL analysis to elucidate 

the genetic basis of NSC accumulation in leaf sheaths and culms 

of rice, for the first time in the main cereal crops, using backcross 

inbred lines from the cross Sasanishiki(japonica)/ 

Habataki (indica)//Sasanishiki///Sasanishiki. They found QTLs 

affecting NSC accumulation on chromosomes 1, 4, 5, 7, 11, 

and 12. Of these, the indica allele of the QTLs found on chromosomes 

5 and 11 increased NSC accumulation and exerted 

some effects on the reduction in the percentage of incompletely 

filled spikelets, suggesting that the expression of these QTLs 

could improve grain filling of rice. 

Recently, Ebitani et al (2004) developed introgression 

lines (ILs) with a chromosome segment of Kasalath (indica) in 

the genetic background of Koshihikari (japonica). Once a QTL 

of interest is detected on a particular chromosome using ILs, 

near-isogenic lines for QTLs (QTL-NILs) can be developed 

in one or two additional generations from advanced backcross 

populations. Using these 39 ILs grown in the field in 2003 

when it was a cold summer, we examined starch accumulation 

in the second leaf sheath below the flag leaf. Measurements 

were done twice: at the heading stage and 14 days after heading 

(DAH). Two ILs whose starch contents were similar to 

that of Koshihikari at the heading stage but significantly decreased 

(P

efficiently with the large-scale measurement of gene expression 

for NILs and/or ILs that include target loci. The combination 

of QTL analysis and high-throughput technologies such 

as microarray could be very useful for this purpose. 

References 

Ebitani T, Yano M, Takarada T, Omoteno M, Takeuchi Y, Nonoue S, 

Yamamoto Y. 2004. Novel breeding approach through development 

of introgression lines in rice. Proceedings of the 4th 

International Crop Science Congress, 26 September-1 October 

2004, Brisbane, Australia. CD-ROM. 

Ishimaru K, Yano M, Aoki N, Ono K, Hirose T, Lin S Y, Monna L, 

Sasaki T, Ohsugi R. 2001a. Toward mapping of physiological 

and agronomic characters on a rice function map: QTL analysis 

and comparison between QTLs and expressed sequence 

tags. Theor. Appl. Genet. 102:793-800. 

Ishimaru K, Kobayashi N, Ono K, Yano M, Ohsugi R. 2001b. Are 

contents of Rubisco, soluble protein and nitrogen in flag leaves 

of rice controlled by the same genetics J. Exp. Bot. 52:1827- 

1833. 

Ishimaru T, Matsuda T, Ohsugi R, Yamagishi T. 2003. Morphological 

development of rice caryopses located at the different positions 

in a panicle from early to middle stage of grain filling. 

Funct. Plant Biol. 30:1139-1149. 

Ishimaru T, Hirose T, Takahashi K, Sasaki H, Ohsugi R, Yamagishi 

T. 2004. Relationship between endosperm development and 

the gene expression of cell wall invertase isoforms: comparison 

between superior and inferior caryopses. Jpn. J. Crop Sci. 

73(Extra issue):170-171. 

Nagata K, Shimizu H, Terao T. 2002. Quantitative trait loci for 

nonstructural carbohydrate accumulation in leaf sheaths and 

culms of rice (Oryza sativa L.) and their effects on grain filling. 

Breed. Sci. 52:275-283. 

Yamagishi J, Miyamoto N, Hirotsu R, Laza RC, Nemoto K. 2004. 

QTLs for branching, floret formation, and pre-flowering floret 

abortion of rice panicle in a temperate japonica × tropical 

japonica cross. Theor. Appl. Genet. 10.1007/s00122-004- 

1795-5. 

Zhu T, Budworth P, Chen W, Provart N, Chang H-S, Guimil S, Su 

W, Estes B, Zou G, Wang X. 2003. Transcriptional control of 

nutrient partitioning during rice grain filling. Plant Biotech. 

J. 1:59-70. 

Notes 

Author’s address: Graduate School of Agricultural and Life Sciences, 

The University of Tokyo, e-mail: aohsugi@mail.ecc.utokyo.ac.jp. 

Improving radiation-use efficiency: the acclimation 

of photosynthesis to high irradiance in rice leaves 

E.H. Murchie, S. Hubbart, S. Peng, and P. Horton 

Sunlight is the principal energy source for crops. It is also highly 

variable; differences in light levels caused by season, climate, 

and position can affect the photosynthetic rate and productivity 

of crops. High yields are generally attained under highirradiance 

conditions, in which photosynthesis is often saturated 

and in full sunlight the number of photons may exceed 

photosynthetic requirements. The converse may be true in lowirradiance 

conditions such as the tropical wet season, in which 

light interception is lower (Weerakoon et al 2000). Here, photosynthesis 

may often be light-limited and growing periods are 

longer and low-yielding (IRRI 1998). These long-term patterns 

are punctuated by short-term variation in light. Ideally, a 

crop would rapidly maximize photosynthesis according to prevailing 

local conditions. 

Understanding the processes that affect the efficiency of 

transduction of light energy by crops is central to the improvement 

of yield potential: this was highlighted by one recent study 

that showed that the radiation-use efficiency (biomass produced 

per unit of radiation energy intercepted) of rice is low, even 

among C 3 crops, and this is a major limitation to yield potential 

(Mitchell et al 1998). An increase in leaf-level photosynthetic 

capacity at ambient CO 2 levels is likely to lead to an 

increased radiation conversion factor (RCF) (Sheehy et al 

2000). It is essential to understand how such variation is accommodated 

efficiently within the plant such that biomass production 

rates are not affected, or can even be improved. This 

is largely carried out by processes termed acclimation. Acclimation 

is defined as a change in composition or morphology 

of the plant that is a specific response to an environmental 

change. There is ample evidence in plant science for significant 

variation in the capacity for acclimation to a number of 

environmental factors. 

Additionally, it is vital to consider that the requirements 

for light and N are strongly linked: the efficient use of highirradiance 

levels (high photosynthetic rates) requires the assimilation 

of large quantities of N. Leaf N content stored in 

the lower leaves of the canopy is essential for high grain yield 

(Sheehy et al 2000) and any attempt to improve photosynthesis 

must take this into account. 

This paper summarizes the responses of rice plants to 

excess irradiance. Genotype variation in light responses is 

described. Emphasis is placed on photosynthesis and the identification 

of limitations to biomass productivity. 


Photosynthesis in saturating light 

This section identifies the limitations to photosynthesis in highlight 

environments at the leaf level. However, it is recognized 

that a crop canopy consists of a population of leaves with different 

photosynthetic capacities and operating at varying degrees 

of saturation. In response to an incrementing light level, 

starting from darkness, photosynthesis in a given leaf will in 

the first instance show light-limiting rates that give way to light 

saturation at higher light intensities. The light intensity at which 

saturation is reached is dependent on the capacity for assimilation 

by the leaf, which is in turn dependent on growth and 

measurement parameters. However, the intensity at which saturation 

is reached is particularly important since it defines the 

efficiency of photosynthesis at high-light levels. For rice growing 

in the field in the tropical dry season, irradiance can exceed 

2,000 µmol m –2 s –1 (Murchie et al 1999). However, it has 

been shown that photosynthesis in rice leaves generally saturates 

at just over 1,000 µmol m –2 s –1 . Therefore, a large proportion 

of incoming irradiance is in excess of assimilatory requirements. 

This causes photo-protective responses to be induced 

even under otherwise ideal conditions (Murchie et al 

1999). If photosynthesis is limited by additional factors such 

as water availability, this can exacerbate stress responses. 

The saturation of photosynthesis at the leaf level is a 

recognized contributor to RCF values (Mitchell et al 1998). 

Additional events can be observed at high irradiance, which, 

under certain conditions, may limit photosynthesis. Mid-day 

depression of photosynthesis is a temporary reduction in photosynthetic 

rate often associated with the high leaf temperatures 

and vapor pressure deficits that occur in bright conditions. 

It has been observed in rice (Black et al 1995, Murchie 

et al 1999) though mechanistically it remains somewhat undefined. 

Photoinhibition is a reduction in quantum yield caused 

by prolonged exposure (hours or days) to high irradiance. It is 

caused by damage to, or down-regulation of, primary photosynthetic 

reactions. It may also contribute to a reduction in 

photosynthetic rate. It has been observed in rice plants in the 

dry season in the Philippines even under well-watered, optimal 

conditions (Murchie et al 1999). Under saturating irradiance, 

photoinhibition will not necessarily affect the rate of assimilation. 

However, since it affects primary photosynthetic 

events, efficiency at low light may be affected: should a transition 

to low light follow a prolonged period of high light, the 

assimilation rate will be reduced (Horton 2000, Zhu et al 2004). 

Acclimation to high irradiance: genotype variation 

Plants respond to irradiance levels in the long term by adjusting 

the components of photosynthesis. This can occur on a 

morphological level, in terms of leaf thickness, as well as a 

chloroplast level, in which the stoichiometry of individual components 

can vary. Generally, these responses maintain the efficiency 

of photosynthesis in low light while enhancing the capacity 

of photosynthesis in high light. 

Difference in photosynthetic capacity 

between high and low growth irradiance (%) 

25 

20 

15 

10 

5 

0 

IR8 IR22 IR40 IR42 IR45 IR54 IR68 IR72 NPT 

Cultivar 

Fig. 1. Photosynthetic responses to growth under low light (LL: 

250 µmol m –2 s –1 ) and high light (HL: 1,500 µmol m –2 s –1 ) in historical 

IRRI cultivars. Values shown are differences between HL 

and LL for light-saturated photosynthesis measured at ambient 

CO 2 (2,000 µmol m –2 s –1 , 360 µL L –1 CO 2 . 

Through a series of field and laboratory experiments, 

we have demonstrated large changes in the response of rice 

leaf photosynthesis to irradiance during and after leaf expansion 

(Murchie et al 2002, Murchie, Hubbart, Horton, unpublished 

data). One important observation to arise from these 

studies was that acclimation of photosynthesis seemed to be 

restricted to relatively low light intensities: above around 1,000 

µmol m –2 s –1 , acclimation of photosynthetic capacity was limited. 

Additionally, variation among genotypes was noted: the 

new plant type, for example, had very high photosynthetic capacities 

at very low growth irradiance. 

Such data may present a novel angle in future research 

for understanding limitations to photosynthesis in dynamic systems. 

Measurements of photosynthetic capacity and leaf composition 

at growth irradiances that are (1) limiting and (2) saturating 

for photosynthesis should give information on the range 

of conditions over which a particular plant or genotype is able 

to respond. Therefore, we grew a range of historical and modern 

IRRI cultivars to the ninth-leaf stage in controlled conditions 

at two light levels (LL: 250 µmol m –2 s –1 and HL: 1,500 

µmol m –2 s –1 ). Photosynthesis, Rubisco, protein, and chlorophyll 

content were measured along with morphological parameters 

such as leaf area and dry weight. 

The difference in light-saturated photosynthesis between 

HL and LL plants is shown in Figure 1. The variation is striking: 

in particular, the older rice varieties possessed a much 

greater range of variation, which was largely caused by a higher 

photosynthetic capacity at higher irradiance. The mechanism 

of this is unknown since correlation with Rubisco amount or 

chlorophyll content was not high. Stomatal conductance but 

not stomatal frequency correlated with P max . Additionally, 

changes in leaf area were not correlated with changes in either 


P max or Rubisco. Currently, it is difficult to ascribe this variation 

to any consistent feature of biochemistry or morphology. 

It is notable that the highest rate of photosynthesis and the 

highest capacity for acclimation were as recorded by the oldest 

cultivar, IR8. Such variation in leaf photosynthesis in old 

and new rice cultivars has not been extensively recorded. It 

has been suggested that essential factors such as leaf area index 

and N responses may have acted as a surrogate for lightsaturated 

rates of leaf photosynthesis during breeding programs 

over the last 30 years (Richards 2000). If so, research into the 

causes of this loss, and of the genotypic variation in leaf photosynthesis 

in general, may generate extremely useful data for 

the improvement of yield potential. 

References 

Black CC, Tu Z-P, Counce PA, Yao P-F, Angelov MN. 1995. An 

integration of photosynthetic traits and mechanisms that can 

increase crop photosynthesis and grain production. Photosynthesis 

Res. 46:169-175. 

Horton P. 2000. Prospects for crop improvement through the genetic 

manipulation of photosynthesis: morphological and biochemical 

aspects of light capture. J. Exp. Bot. 51:475-485. 

IRRI (International Rice Research Institute). 1998. IRRI research 

report for 1998: irrigated rice ecosystems. Manila (Philippines): 

IRRI. 

Mitchell PL, Sheehy JE, Woodward FI. 1998. Potential yields and 

the efficiency of radiation use in rice. IRRI Discussion Paper 

Series No. 32. Manila (Philippines): International Rice Research 

Institute. 62 p. 

Murchie EH, Hubbart S, Chen YZ, Peng SB, Horton P. 2002. Acclimation 

of rice photosynthesis to irradiance under field conditions. 

Plant Physiol. 130:1999-2010. 

Murchie EH, Chen Y-Z, Hubbart S, Peng S, Horton P. 1999. Interactions 

between senescence and leaf orientation determine in 

situ patterns of photosynthesis and photoinhibition in fieldgrown 

rice. Plant Physiol. 119:553-563. 

Richards RA. 2000. Selectable traits to increase crop photosynthesis 

and yield of grain crops. J. Exp. Bot. 51:447-458. 

Sheehy JE, Mitchell PL, Dionara J, Ferrer A. 2000. What governs 

ceiling yield and its rate of attainment In: Peng S, Hardy B, 

editors. 2001. Rice research for food security and poverty alleviation. 

Proceedings of the International Rice Research 

Conference, March-April 2000. Los Baños (Philippines): International 


Weerakoon WMW, Ingram KT, Moss DN. 2000. Atmospheric carbon 

dioxide and fertiliser effects on radiation interception by 

rice. Plant Soil 220:99-106. 

Zhu X-G, Ort DR, Whitmarsh J, Long SP. 2004. The slow reversibility 

of photosystem II thermal energy dissipation on transfer from 

high to low light may cause large losses in carbon gain by 

crop canopies: a theoretical analysis. J. Exp. Bot. 55:1167- 

1175. 

Notes 

Authors’ addresses: E.H. Murchie, S. Hubbart, and P. Horton, University 

of Sheffield, Department of Molecular Biology and 

Biotechnology, Firth Court, Western Bank, S10 2TN, UK; S. 

Peng, Crop, Soil, and Water Sciences Division, International 

Rice Research Institute, DAPO Box 7777, Metro Manila, Philippines, 

e-mail: e.h.murchie@sheffield.ac.uk. 

Sink-source relationship and yield potential of rice: effect of 

ethylene on grain filling of late-flowering spikelets 

P.K. Mohapatra and Rashmi Mohapatra 

Poor filling of spikelets on the basal primary branches of the 

rice panicle limits genetic potential for yield and produces 

grains unsuitable for human consumption (Mohapatra et al 

1993). The problem is more acute in the newly developed super 

rice with heavy panicles. Spikelet development is very slow 

in the proximal part of the panicle; it results in poor partitioning 

of dry matter for grain filling (Mohapatra et al 1993). Assimilates 

for grain filling are received from the flag leaf and 

photo-assimilatory tissues located within the panicle structure 

(Yoshida 1981). Senescence of these organs is programmed 

genetically, and has a fixed time schedule. Therefore, it will 

be of interest to know if the late-developed basal spikelets of 

the rice panicle become source-limited for assimilates during 

their peak period of growth because of the coincidence of senescence 

of the photosynthetic tissues. Similarly, the formation 

of ABA or ethylene during senescence of these organs 

may make the spikelets sink-limited for growth. Ethylene has 

been identified as a regulator of senescence in detached rice 

leaves (Lin et al 2002), but nothing is known about its role in 

the regulation of senescence in intact plants and grain filling. 

The application of triazolic compounds reduces ethylene production 

in plants, and thereby delays senescence (Fletcher et 

al 2000). In contrast, 2-chloroethylphosphonic acid (CEPA) 

application enhances ethylene production and may expedite 

senescence. We need to ascertain the role of ethylene in regulating 

the filling of rice spikelets. It is also desirable to distinguish 

its action from that of senescence-controlled reduction 

in assimilate supply through depletion of source activity. 


A high-yielding rice plant, Oryza sativa cv. Mahanadi 

(IR19661-131-1-3 × Savitri), was cultivated in irrigated field 

conditions of the Regional Research Station, Chiplima, during 


Table 1. Variations in yield parameters of rice (cv. Mahanadi) panicle at maturity following different types of chemical treatments 

during the wet season of 2000. The ± values indicate standard deviation. 

Wt. of Wt. of No. of No. of Av wt. of Percentage of 

Type of panicle total No. of filled barren spikelets 

treatment (g) spikelets spikelets spikelets spikelets (mg) Filled Barren 

(g) spikelets spikelets 

H 2 O 6.12 ± 0.95 3.76 ± 0.48 190.7 ± 31.7 151.7 ± 34.1 40.0 ± 8.2 19.70 ± 0.80 79.5 ± 5.3 21.0 ± 2.4 

Uniconazole 6.85 ± 0.03 3.94 ± 0.15 201.0 ± 1.4 167.5 ± 6.4 33.5 ± 7.7 19.62 ± 0.85 83.3 ± 3.3 16.7 ± 2.8 

Paclobutrazol 7.14 ± 0.79 4.10 ± 0.67 202.7 ± 4.6 176.7 ± 14.5 26.0 ± 1.1 20.24 ± 2.88 87.3 ± 5.0 12.8 ± 5.0 

CEPA 5.54 ±0 .56 2.78 ± 0.40 204.3 ± 36.5 126.7 ± 13.8 77.7 ± 5.0 13.58 ± 0.46 62.0 ± 4.8 38.0 ± 4.8 

the wet season of 2000. The experimental area was divided 

into three parts. In each part, uniformly developed plants were 

divided into four groups for growth regulator treatments. On 

each day, 0.5 mL of the chemicals paclobutrazol (100 µL L –1 ) 

and uniconazole (0.5 mg L –1 ), 2-CEPA (10 –5 M), and H 2 O in 

aqueous solution were injected into the boot (the leaf sheath 

enclosure confining the young panicle before emergence) of 

the flag leaf. The treatments began 5 d before the day of anthesis 

and continued for 5 d. 

Plants were harvested from the day of treatment to maturity 

at 5–7-d intervals. Six spikelets from the uppermost primary 

branch (apical) and eight spikelets from the lowermost 

primary branch (basal) were collected from the panicle on the 

main shoot of four different plants in each treatment. The apical 

spikelets emerged 5–6 d earlier than the basal spikelets. In 

each set, the lemma and palea (glumes) were dissected out, 

weighed fresh, and used for biochemical analyses. The four 

sets of spikelets were used for estimation of pigments, soluble 

carbohydrates, lipid peroxidation, and activity of peroxidase 

enzyme, respectively. At maturity, grain yield of the panicle 

was estimated. Separate samples were taken for measuring ethylene 

concentration of the panicle, basal and apical primary 

branches, flag leaf, and boot of the flag-leaf sheath. The cut 

end of the plant material was dipped in 1 mL of water inside a 

test tube. The test tube was sealed and incubated for 2 h in 

darkness before 1 mL of head space gas was used for ethylene 

estimation in a gas chromatograph (model 6890, Hewlett- 

Packard Company, USA, equipped with a flame ionization 

detector) (Naik and Mohapatra 2000). 

Results 

Grain yield and yield components 

Uniconazole and paclobutrazol treatments improved panicle 

weight, spikelet weight, and number of filled spikelets, and 

reduced the number of barren spikelets significantly (Table 

1). In contrast, CEPA application had the opposite effect. The 

triazolic compounds did not improve the average weight of 

spikelets, but CEPA application depressed it significantly. 

Ethylene concentration of panicle, 

primary branches, flag leaf, boot, and spikelets 

Ethylene concentration of the plant organs was low at 5 d before 

anthesis; it reached a peak at anthesis and declined thereafter 

(Fig. 1). It was minimum at maturity. The concentration 

of ethylene inside the boot of the flag leaf also increased temporally 

during the pre-anthesis period. The ethylene concentration 

of the apical primary branch was lower than that of the 

basal branch of the panicle. The treatment with triazolic compounds 

decreased the concentration of ethylene of the boot 

and the plant organs significantly, but CEPA increased the concentration. 

The treatment with the chemicals was more effective 

on the basal branch than on the apical branch. 

Pigment concentration of glumes 

The concentrations of chlorophylls, carotenoids, soluble sugars, 

proteins, and free amino acids of the glumes increased 

temporally from 5 d before to 1 wk after anthesis. It declined 

thereafter till maturity in the apical branch (data not included). 

In the basal branch, the concentration of these materials increased 

for a longer time before declining. During the declining 

phase, the glumes of basal spikelets possessed a higher 

concentration of the materials than those of the apical spikelets. 

The treatment with the triazolic compounds improved the 

concentration significantly in the glumes of basal spikelets, 

but CEPA application had the opposite effect. The chemical 

treatments did not change the concentration of the materials in 

the apical spikelets. 

Peroxidase activity and peroxidation 

of lipids in glumes 

The activity of the enzyme peroxidase and the amount of lipid 

peroxidation in the glumes of apical spikelets increased from 

5 d before to 16 d past anthesis and declined thereafter till 

maturity (data not presented). The activity was higher in the 

glumes of the basal spikelets than in the apical spikelets. Peroxidase 

activity increased up to 23 d past anthesis in the basal 

spikelets and declined thereafter. The treatment with triazolic 

compounds increased activity of the enzyme, whereas CEPA 

decreased it significantly. The treatment with chemicals was 

not effective on the apical spikelets. 


Ethylene concentration (nmol g –1 fresh wt. h –1 ) 

6 

(Apical primary branch, 2000) 

5 

4 

H 2 O 

UCZ 

PBZ 

CEPA 

6 

5 

4 

(Basal primary branch, 2000) 

3 

3 

2 

2 

1 

1 

0 

10 

8 

6 

0 

12 

(Flag leaf, 2000) (Panicle, 2000) 

10 

8 

6 

4 

4 

2 

2 

0 

–5 0 5 

0 

–5 0 5 

Ethylene concentration (nmol mL –1 of air) 

5.0 

4.5 

4.0 

H 2 O 

UCZ 

PBZ 

CEPA 

(Boot, 2000) 

3.5 

3.0 

2.5 

2.0 

1.5 

1.0 

0.5 

0 

–5 –4 –3 –2 –1 

Days from anthesis 

Fig. 1. The effect of uniconazole, paclobutrazol, and 2-chloroethylphosphonic acid on the ethylene 

concentration of apical and basal primary branches, panicle, flag leaf, and boot of the flag 

leaf of rice at anthesis during the wet season of 2000. Anthesis date is recorded when the apical 

spikelets reach that stage. This time scale is used for all organs except the basal primary branch. 


Discussion 

A comparison of the concentration of proteins, photosynthetic 

pigments, peroxidase activity, and lipid peroxidation of glumes 

of apical and basal spikelets quantified the magnitude of difference 

existing between the senescence patterns of the two 

contrasting spikelets. The juvenile phase of development was 

slower and the senescence phase was faster in the basal spikelets 

than in the apical spikelets. The concentration of soluble 

carbohydrates and amino acids of the glumes exhibited a similar 

temporal fluctuation. This indicated the close correlation 

between assimilate concentration and developmental stages of 

the organs. Rice seed is a caryopsis. The real function of the 

seed coat is carried out by the lemma and palea, and their development 

is related to that of the seed. The lemma and palea 

unite to enclose hull space and this space determines grain 

size (Yoshida 1981). Hence, a close association between endosperm 

filling and the senescence pattern of the glumes is 

expected. In our study, the glumes of the inferior spikelets remained 

green for a longer time than those of the superior spikelets, 

but endosperm filling did not improve. Therefore, the senescence 

pattern of the glumes may not be necessarily coincidental 

to or causative of grain filling. 

The estimation of ethylene from different parts of the 

panicle and flag leaf indicated its involvement in the regulation 

of senescence and grain maturation. Ethylene concentration 

increased and decreased in the pre- and postanthesis stages, 

respectively. This is similar to the observation from another 

high-yielding rice cv., Lalat (Naik and Mohapatra 2000). Ethylene 

produced by the dominant spikelets of the panicle and 

flag leaf at anthesis accumulated inside the boot of the flag 

leaf sheath and increased the concentration temporarily. The 

apical spikelets emerged from the boot at anthesis, but the basal 

spikelets remained inside it for about 5–6 d more. Exposure to 

ethylene inside the boot could have retarded the development 

of basal spikelets. In contrast, the apical spikelets developed 

faster and enjoyed freedom from the constraints of the flagleaf 

enclosure. When the basal spikelets reached anthesis, the 

inferior spikelets produced more ethylene (Fig. 1) and this ethylene 

might be responsible for the faster degradation of lipids 

of the glumes during the senescence phase. Ethylene is easily 

mobile within flower organs and works as a signal for interorgan 

communication during senescence; in maize, ethylene 

released by the dominant spikelets at pollination retards the 

growth of inferior spikelets (Cheng and Lur 1996). Our study 

and our previous work (Mohapatra et al 2000, Naik and 

Mohapatra 2000) lead to the conclusion that ethylene produced 

by the apical spikelets and flag leaf at anthesis accumulated in 

the boot. It enhanced the senescence of the basal spikelets of 

the panicle and resulted in poor filling (the r value between 

boot ethylene concentration and % of filled spikelets was 

–0.77). 

In our study, grain filling of the basal spikelets occurred 

during the senescence phase of the flag leaf and panicle. This 

apparent coincidence suggests the possibility of a source limitation 

for assimilates. However, deficiency of assimilates restricting 

grain filling of irrigated rice was ruled out (Mohapatra 

et al 1993, Yang et al 2002). Instead, it was opined that poor 

filling is affected by the low cytokinin level of the inferior 

endosperm (Yang et al 2002). A compromise between our study 

and that of Yang et al (2002) suggests an assumption that ethylene 

may be triggering a deficiency of cytokinin in the inferior 

endosperm and causing poor cell division and grain filling. 

Simultaneously, ethylene is also responsible for the promotion 

of senescence of the source organs. But poor filling 

cannot be consequential to a shortage of assimilates through 

depletion of source activity. Murchie et al (2002) found no 

consistent relationship between the rate of grain filling and 

photosynthesis of rice leaves. The gradient of metabolic potential 

between superior and inferior spikelets is established at 

the early stage of seed development (Qu-Lee and Setter 1985). 

High metabolic potential precludes adverse ethylene action on 

apical spikelets, but not on basal spikelets. At this stage, competition 

for limiting assimilates is less likely to occur because 

of the low demand for sink activity (Mohapatra et al 1993, 

Yang et al 2002). 

References 

Cheng CY, Lur HS. 1996. Ethylene may be involved in abortion of 

maize caryopses. Physiol. Plant. 98:245-252. 

Fletcher RA, Gilley A, Sankhla N, Davis TD. 2000. Triazoles as 

plant growth regulators and stress protectants. Hort. Rev. 

24:55-138. 

Lin CC, Hsu YT, Kao CH. 2002. Ammonium ion, ethylene and NaClinduced 

senescence of detached rice leaves. Plant Growth 

Regul. 37:85-92. 

Mohapatra PK, Naik PK, Patel R. 2000. Ethylene inhibitors improve 

dry partitioning and development of late flowering spikelets 

on rice panicles. Austr. J. Plant Physiol. 22:311-323. 

Mohapatra PK, Patel R, Sahu SK. 1993. Time of flowering affects 

grain quality and spikelet partitioning within the rice panicle. 

Austr. J. Plant Physiol. 20:231-241. 

Murchie EH, Yang J, Hubbart S, Horton P, Peng S. 2002. Are there 

associations between grain-filling rate and photosynthesis in 

the flag leaves of field-grown rice J. Exp. Bot. 53:2217-2224. 

Naik PK, Mohapatra PK. 2000. Ethylene inhibitors enhanced sucrose 

synthase activity and promoted grain filling of basal 

rice kernels. Austr. J. Plant Physiol. 27:997-1008. 

Qu-Lee T-M, Setter TL. 1985. Enzyme activities of starch and sucrose 

pathways and growth of apical and basal maize kernels. 


Yang J, Zhang J, Huang Z, Wang Z, Zhu Q, Liu L. 2002. Correlation 

of cytokinin levels in the endosperm and roots with cell number 

and cell division activity during endosperm development 

in rice. Ann. Bot. 90:369-377. 

Yoshida S. 1981. Fundamentals of rice crop science. Manila (Philippines): 



Notes 

Authors’ address: School of Life Science, Sambalpur University, Jyoti 

Vihar, Sambalpur 768019, India, e-mail: 

pravat48@hotmail.com. 

Acknowledgments: This work was supported by a research grant to 

P.K. Mohapatra by the Indian Council of Agricultural Research, 

New Delhi. 

Impact of increased source capacity on rice yield: 

a case study with CO 2 

enrichment 

Toshihiro Hasegawa, Kazuhiko Kobayashi, Mark Lieffering, Han Yong Kim, Hidemitsu Sakai, Hiroyuki Shimono, 

Yasuhiro Yamakawa, Mayumi Yoshimoto, and Masumi Okada 

Grain yields of rice (Oryza sativa L.) as well as other major 

cereals have increased markedly in the last century. Genetic 

improvements have contributed largely to these increases. Yield 

potential can be increased by increasing biomass production 

and/or increasing allocation to harvested organs. Genetic improvements 

in the yield potential of major cereals have largely 

been through the latter. Photosynthesis plays a central role in 

biomass production and recent varieties tend to have longer 

photosynthetic duration than older varieties, but only limited 

evidence exists that the maximum photosynthetic capacity has 

been improved. Even where improved photosynthetic capacity 

was observed, no direct evidence of biomass increase was 

shown (review: Evans 1993). Efforts are under way to improve 

the genetic potential of photosynthesis, but the question “Does 

improving photosynthesis benefit crop production” needs to 

be answered experimentally. 

Photosynthetic capacity under current atmospheric CO 2 

concentrations ([CO 2 ]) is generally limited by the carboxylation 

process in the Calvin cycle. Increases in [CO 2 ] accelerate 

carboxylation velocity by increasing [CO 2 ] supply from stomata 

and reducing photorespiration (review: Long et al 2004). 

In general, CO 2 enrichment studies aim to identify the magnitude 

of changes in crops with increased [CO 2 ], but these results 

can also be used to infer the impact of increased photosynthesis 

on crop yield. 

Among various experimental facilities that have been 

used in elevated CO 2 studies, free-air CO 2 enrichment (FACE) 

systems, which do not have enclosures and therefore a minimum 

of artifacts, are usually regarded as providing the best 

estimates of yield responses to elevated [CO 2 ] in the open field. 

Across all species, fewer FACE than chamber experiments have 

been conducted, though the rice FACE system has been running 

at two sites (Japan and China) for more than three years. 

Many data have been accumulated, leading to reviews of the 

yield responses of a range of plant species to FACE (Kimball 

et al 2002, Long et al 2004). To expand on these reviews, this 

paper summarizes the FACE effects on photosynthesis, biomass, 

and yield of rice in order to examine the effect of increased 

source capacity on grain yield. The rice results are 

compared with those of other C 3 species responses reviewed 

by Kimball et al (2002) and Long et al (2004). 

Data 

Most of the rice responses to elevated [CO 2 ] presented in this 

paper are from the FACE experiment conducted in Shizukuishi, 

Iwate, Japan (39 o 38′N, 140 o 57′E) from 1998 to 2000. The 

design of the FACE system was reported by Okada et al (2001) 

and growth conditions were given by Kim et al (2003a). Briefly, 

rice was grown in farmers’ paddies equipped with 12-m-diameter 

FACE systems. Pure CO 2 gas was released from peripheral 

emission tubes 0.5 m above the canopy, the sides of the 

emission depending on the wind direction. The target [CO 2 ] at 

the center of the FACE plots was 200 µmol mol –1 above the 

ambient [CO 2 ]. Each plot had some subplot treatments such as 

N rates and varieties, but here we present only the results from 

cultivar Akitakomachi under standard N conditions (8–9 g 

m –2 of N). 

For this paper, growth and yield data in Shizukuishi were 

from Kim et al (2003a,b). Leaf photosynthesis data were from 

Seneweera et al (2002) and Anten et al (2003). In the former 

study, [CO 2 ] and [H 2 O] exchange rates were measured on the 

top fully expanded leaves on two occasions during the season, 

whereas in the latter study light-saturated CO 2 assimilation 

(Asat) and maximum apparent quantum yield of CO 2 uptake 

(QY) were measured at different positions of the canopy at 

around flowering. By also measuring canopy light attenuation 

and leaf area distribution, Anten et al (2003) estimated canopy 

photosynthesis by integrating the photosynthesis of sunlit and 

shaded leaves. 

Other rice data presented in this paper come from the 

rice-wheat FACE experiment conducted in Wuxi (31 o 37′N, 

120 o 28′E), Jiangsu, China, in 2001, 2002, and 2003. A FACE 

system similar to the one in Shizukuishi was used. Final biomass 

and grain yield data of cultivar Wuxiangjing 14 (Huang 

et al 2004) are presented in this study. 

To compare rice growth and yield responses to FACE 

with those of other C 3 species, we present data from a review 

by Kimball et al (2002) and a meta-analysis conducted by Long 

et al (2004) as a reference; the latter reviewed results from 

large-scale (>8 m) FACE experiments only. 



Biomass production is a result of canopy light interception 

and conversion of the intercepted light to biomass; leaf area 

influences the former and photosynthesis the latter. The effect 

of FACE on peak leaf area was not significant for all years; 

this agreed with the average C 3 responses in the meta-analysis 

of Long et al (2004). Although this implies that carbon supply 

is not limiting to leaf area growth, it also indicates that the 

FACE effects on biomass should appear largely through increased 

source capacity rather than increased light capture, so 

the FACE studies can provide us with evidence as to how much 

increased grain yield depends on increased photosynthetic rates. 

Light-saturated photosynthesis (Asat) is generally limited 

by carboxylation or by ribulose 1,5-bisphosphate regeneration. 

These processes are well represented by two parameters 

obtained from leaf gas exchange measurements: maximum 

carboxylation rate (Vc,max) and maximum rate of electron 

transport (Jmax). For rice, the FACE treatment decreased 

both Vc,max and Jmax, with the larger reduction observed in 

the former; this was commonly observed in other C 3 species 

(Fig. 1). In rice, the FACE effects were larger in the flag leaf 

than in the 8th leaf at the active tillering stage. Nitrogen content 

per unit leaf area (Narea) was also reduced with the FACE 

treatment by 6% in rice, and a similar reduction was reported 

in the meta-analysis (4%). 

The data reported by Seneweera et al (2002) in the rice 

FACE experiment indicated that both Vc,max and Jmax are 

highly sensitive to the decrease in Narea; these relationships 

largely explain the variation in both parameters with elevated 

[CO 2 ] and leaf age. According to the relationships among 

Vc,max, Jmax, and Narea, enhancement of Asat by elevated 

[CO 2 ] is greater where Narea is large, and becomes progressively 

smaller as Narea declines. In fact, Asat measured on the 

8th leaf at the active tillering stage showed a large enhancement 

of about 40% (Seneweera et al 2002), and measurements 

thereafter by Anten et al (2003) and those on the flag leaf by 

Seneweera et al (2002) resulted in lower rates of enhancement, 

about 20% and 4%, respectively. These changes related well 

to the changes in Narea. 

Elevated CO 2 also has a positive influence on the maximum 

apparent quantum yield of CO 2 uptake (QY, Fig. 1). In 

rice, Anten et al (2003) reported about a 20% increase in QY 

at around the heading stage with FACE; this was slightly larger 

than the enhancement in the meta-analysis of C 3 species (Long 

et al 2004). Consequently, both Asat and QY contribute to the 

carbon gain enhancement of the canopy. 

The enhancement of canopy photosynthesis (Acanopy) 

estimated from the leaf photosynthetic parameters was 24% 

for rice (Anten et al 2003), which was slightly lower but comparable 

to the daily integral of leaf CO 2 uptake (A’) in the 

meta-analysis. The crop growth rate (CGR) around flowering, 

however, was enhanced by elevated [CO 2 ] to a lesser extent of 

4–15%. Canopy gas exchange was not measured in the FACE 

experiments, but, according to Sakai et al (2001), who measured 

canopy gas exchange in controlled chambers, CO 2 enrichment 

by 300 µmol mol –1 increased net canopy CO 2 uptake 

and the CGR around flowering by similar ratios (11% and 9%, 

respectively). The relationship between leaf- and canopy-level 

gas exchange and crop growth needs to be better understood 

to account for why leaf photosynthetic enhancement did not 

translate well into increased biomass production. 

The enhancement of final biomass by FACE ranged from 

9% to 15% in Shizukuishi. This was similar to that obtained in 

Wuxi, China (7–14%), despite the large difference in climatic 

conditions. However, the enhancement ratios of rice biomass 

are lower than the average response of C 3 species. 

Seed yield enhancement was similar to but slightly larger 

than biomass enhancement: 13–16% in Shizukuishi and 8–18% 

in Wuxi. These values lie at the lower edge of the confidence 

intervals of the average yield responses of C 3 species (mean: 

24%). A large variation in the grain yield response to elevated 

[CO 2 ] exists among C 3 crops. Wheat yield responses to FACE 

under ample N and water were similar to those of rice, but 

substantially larger harvestable yield enhancements were observed 

in potato and cotton (Kimball et al 2002). It is worth 

noting that no enhancement of the aboveground biomass was 

recorded for potato. This may suggest that many important 

traits can regulate the effect of increased source capacity on 

yield. 

Conclusions 

In rice, a 20–40% increase in Asat occurred with FACE before 

anthesis. Thereafter, photosynthetic enhancement was rather 

limited because both Jmax and Vc,max decreased with decreasing 

leaf N content: this was low with elevated [CO 2 ] and decreased 

as the crop aged. The rice FACE experiments in Japan 

and China revealed that grain yield enhancement by elevated 

CO 2 under standard N supplies ranged from 8% to 18% (mean: 

13.5%), with no marked differences between locations. The 

magnitude of the yield enhancement was smaller than that of 

other C 3 species (the meta-analysis mean: 24%). Plant traits 

that moderate the effects of elevated [CO 2 ] on grain yield need 

to be identified under season-long examinations to fully take 

advantage of increased source capacity for yield improvement. 

References 

Anten NPR, Hirose T, Onoda Y, Kinugasa T, Kim HY, Okada M, 

Kobayashi K. 2003. Elevated CO 2 and nitrogen availability 

have interactive effects on canopy carbon gain in rice. New 

Phytol. 161:459-471. 

Evans LT. 1993. Crop evolution, adaptation and yield. Cambridge 

(UK): Cambridge University Press. 500 p. 

Huang JY, Yang HJ, Yang LX, Liu HJ, Dong GC, Zhu JG, Wang YL. 

2004. Effects of free-air CO 2 enrichment (FACE) on yield 

formation of rice (Oryza sativa L.) and its causes. Sci. Agric. 

Sin. (In press.) 

Kim HY, Lieffering M, Kobayashi K, Okada M, Mitchell MW, 

Gumpertz M. 2003a. Effects of free-air CO 2 enrichment and 

nitrogen supply on the yield of temperate paddy rice crops. 

Field Crops Res. 83:261-270. 


SF 

S8 

VC,max 

Meta (C 3 species) 

Rice 

Other C 3 crops 

SF 

S8 

Jmax 

SF 

S8 

V,max/Jmax 

A 

Narea 

Asat 

SF A S8 

QY 

A’ 

Acanopy 

CGR 

Final biomass 

Rice (Shizukuishi) 

Rice (Wuxi) 

Agricultural yield 

Rice (Shizukuishi) 

Rice (Wuxi) 

Wheat (Maricopa) 

Potato (Rapolano) 

Cotton (Maricopa) 

–40 –20 0 20 40 60 

% increase with elevated [CO 2 ] 

Fig. 1. The percent change under elevated [CO 2 ] of (1) maximum carboxylation 

rate (Vc,max), (2) maximum rate of electron transport (Jmax), 

(3) the ratio of Vc,max:Jmax, (4) N content per unit leaf area (Narea), (5) 

light-saturated CO 2 assimilation (Asat), (6) maximum apparent quantum 

yield of CO 2 uptake (QY), (7) daily integral of leaf CO 2 uptake (A’), (8) 

estimated daily net canopy CO 2 assimilation at the heading stage 

(Acanopy), (9) crop growth rate (CGR) around the heading stage, (10) 

final biomass, and (11) seed yield. Meta-analysis results of large-scale 

free-air CO 2 enrichment experiments for C 3 species along with 95% confidence 

intervals (Long et al 2004) are presented as a reference. Rice 

leaf photosynthetic properties are from Seneweera et al (2002) and Anten 

et al (2003); S8 and SF beside symbols stand for the 8th-leaf and flagleaf 

measurements, respectively, of Seneweera et al (2002), and A for 

Anten et al’s measurements. Rice growth and yield data in Shizukuishi, 

Japan, are from Kim et al (2003a,b) and in Wuxi, China, from Huang et al 

(2004). Agricultural yield data of other C 3 crops (seed for wheat, tuber 

for potato, and boll for cotton) were from Kimball et al’s (2002) review; 

only results from crops grown under ample water and nitrogen conditions 

are presented. 


Kim HY, Lieffering M, Kobayashi K, Okada M, Miura S. 2003b. 

Seasonal changes in the effects of elevated CO 2 on rice at 

three levels of nitrogen supply: a free-air CO 2 enrichment 

(FACE) experiment. Global Change Biol. 9:826-837. 

Kimball BA, Kobayashi K, Bindi M. 2002. Responses of agricultural 

crops to free-air CO 2 enrichment. Adv. Agron. 77:293- 

368. 

Long SP, Ainsworth EA, Rogers A, Ort DR. 2004. Rising atmospheric 

carbon dioxide: plants FACE the future. Annu. Rev. Plant Biol. 

55:557-594. 

Okada M, Lieffering M, Nakamura H, Yoshimoto M, Kim HY, 

Kobayashi K. 2001. Free-air CO 2 enrichment (FACE) using 

pure CO 2 injection: system description. New Phytol. 150:251- 

260. 

Sakai H, Yagi K, Kobayashi K, Kawashima S. 2001. Rice carbon 

balance under elevated CO 2 . New Phytol. 150:241-249. 

Seneweera SP, Conroy JP, Ishimaru K, Ghannoum O, Okada M, 

Lieffering M, Kim HY, Kobayashi K. 2002. Changes in sourcesink 

relations during development influence photosynthetic 

acclimation of rice to free-air CO 2 enrichment (FACE). Funct. 

Plant Biol. 29:945-953. 

Notes 

Possibilities for and constraints to 

improving canopy photosynthesis 

J.E. Sheehy, A.V. Elmido, P. Pablico, M.J.A. Dionora, and A.B. Ferrer 

Authors’ addresses: Toshihiro Hasegawa, Hidemitsu Sakai, and 

Mayumi Yoshimoto, National Institute for Agro-Environmental 

Sciences, Tsukuba, Ibaraki 305-8604, Japan, e-mail: 

thase@niaes.affrc.go.jp; Kazuhiko Kobayashi, University of 

Tokyo, Japan; Mark Lieffering, AgResearch Grassland, New 

Zealand; Han Yong Kim, Chonnam University, South Korea; 

Hiroyuki Shimono, Yasuhiro Yamakawa, and Masumi Okada, 

National Agricultural Research Center for Tohoku Region, 

Japan. 

We have a unique opportunity to solve all of the rice production 

problems of Asia by redesigning one plant mechanism: 

photosynthesis. A lack of scientific understanding of the yieldshaping 

mechanisms, combined with a triumph of correlation 

over common sense, have led to the commonly held belief 

that leaf photosynthesis and yield are unrelated. To counter 

that misconception, we argue the link between photosynthesis 

and yield and explore what it would take to be a C 4 rice plant. 

Recent discoveries of single-cell C 4 photosynthesis could provide 

new impetus to convert rice from a C 3 to a C 4 photosynthetic 

system without having to incorporate the Kranz anatomy 

(Sage 2002). 

Solar energy captured in photosynthesis gives plants the 

capacity to grow. Two aspects of energy capture set the limits 

for crop growth in a given thermal environment. The first is 

the quantity of energy captured, and that depends on the available 

irradiance and the fraction absorbed by the crop canopy. 

The second is the efficiency with which the absorbed energy 

is used for the chain of synthetic processes that culminate in 

harvestable yield. Modern cultivars with erect leaves absorb 

all the available photosynthetically active radiation, leaving 

little room for further improvements in energy capture. In highyielding 

rice, sink size is unlikely to limit yield; the potential 

sink size is almost double that actually realized. 

In plants using the C 3 photosynthetic pathway (rice, 

wheat, etc.), oxygen competes for the active sites on the key 

photosynthetic enzyme, Rubisco, leading to as much as a 30% 

loss of captured CO 2 in a process known as photorespiration. 

The C 4 syndrome evolved from the C 3 syndrome and C 4 plants 

have a photosynthetic system that concentrates CO 2 around 

Rubisco, and so nearly saturates the carboxylation reaction 

and eliminates photorespiration. Consequently, C 4 plants use 

water more effectively. In addition, C 4 crops require less nitrogen 

fertilizer because they make better use of nitrogen in 

photosynthesis. The maximum rate of leaf photosynthesis for 

both C 3 and C 4 plants appears to vary linearly with nitrogen 

content per unit leaf area over a wide range of nitrogen concentrations 

(Evans and von Caemmerer 2000). Assuming the 

data shown by Evans and von Caemmerer (2000) could be 

described by linear relationships, then maize had a gradient of 

approximately 0.74 µmol CO 2 s –1 mmol –1 N compared with 

that for C 3 plants (rice and wheat) of 0.26 µmol CO 2 s –1 

mmol –1 N: a threefold difference in leaf photosynthetic nitrogen-use 

efficiency (Table 1). 

Photosynthesis in C 3 plants should benefit more from 

rising concentrations of atmospheric carbon dioxide than the 

photosynthesis in C 4 plants. However, work of Baker et al 

(1990) showed that canopy photosynthesis in rice did not respond 

to increases in atmospheric CO 2 beyond 50 Pa. Average 

atmospheric concentration of carbon dioxide increased by approximately 

4.5 Pa from 1968 to 1998. However, over that 

period, yield for the same cultivar (IR8) grown on the IRRI 

farm at nitrogen inputs of approximately 150 kg N ha –1 fell by 

2.6 t ha –1 (Peng et al 1999). 


Table 1. A comparison of three attributes of maize (C 4 ) and rice (C 3 ) 

and the advantage of the C 4 over the C 3 system under current CO 2 

conditions. WUE (transpiration) is water-use efficiency, RUE (PAR) 

is radiation-use efficiency, PNUE is photosynthetic nitrogen-use 

effectiveness. Source: (a) from various sources, (b) Kiniry et al 

(1989), and (c) Evans and von Caemerrer (2000). 

WUE RUE (PAR) PNUE 

(g DW kg –1 H 2 O) (g DW MJ –1 ) (µm CO 2 s –1 mmol –1 N) 

Zea mays (C 4 ) 2.9 a 3.3 b 0.74 c 

Oryza sativa (C 3 ) 1.6 a 2.2 b 0.26 c 

C 4 /C 3 advantage 1.8 1.5 2.8 

Dry weight (t ha –1 ) 

12 

10 

8 

6 

4 

2 

The link between yield and photosynthesis 

There are two nearly equal phases of growth in rice (Fig. 1), 

with about half of the first phase of vegetative growth preceding 

the second phase of reproductive growth (largely the 

panicle). The growth of the panicle is largely heterotrophic 

and in high-yielding rice has a maximum growth rate about 

the same as that achieved by the autotrophic vegetative component. 

How could the growth rates of both components be 

increased by 50% There is a linear relationship between accumulated 

intercepted photosynthetically active solar radiation 

and accumulated biomass; the slope of that linear relationship 

is known as the radiation-use efficiency (ε). The relationship 

between the growth rate of shoots (dW s /dt) and ε is 

dW s /dt = ε I int (1) 

where I int is the total amount of photosynthetically active radiation 

(PAR) intercepted by the crop for the same day. To 

increase growth rates by 50%, the value of ε would have to 

increase by 50%, from the average value of ε for rice of 2.2 g 

dry weight (DW) MJ –1 PAR to 3.3 g DW MJ –1 PAR, the average 

value for maize (Table 1, Kiniry et al 1989). The link between 

ε and photosynthesis on any given day can be written 

as 

ε = (1 - β) [(P g (t) – R(t) – D(t))] /I int (t) (2) 

where β is the root weight ratio, P g is the rate of canopy gross 

photosynthesis, R is the rate of shoot and root respiration, D is 

the loss of dry matter through detachment, and t is time in 

days. On a given day, to increase ε by 50%, photosynthesis 

would have to increase by a little more than 50%, that is, to 

values comparable with those of C 4 plants. 

Upper limits to canopy photosynthesis 

The simple model of Monteith (1965) shows that the maximum 

photosynthetic leaf area index of an erect canopy is approximately 

6.6. The sunlit leaf area index (L o ) reached a maximum 

of 3.3 and photosynthetically active shaded leaf area index 

(L s ) reached its maximum at approximately 3.0. Sinclair 

and Sheehy (1999) proposed that the third class of leaves in 

0 

0 20 40 60 80 100 120 

DAT 

Fig. 1. The two phases of rice growth for IR72 growing in the dry 

season of 2001 at IRRI with nonlimiting fertilizer applications. Open 

squares denote vegetative parts and filled squares denote panicles. 

DAT = days after transplanting. 

the canopy (L n ) were those that were close to the light compensation 

point and largely useful for the N they contributed 

to the grain (L n = L – L o – L s ; where L is the total LAI). Sheehy 

(2000) went on to show that the loss through senescence of L s 

and L n would reduce canopy photosynthesis by about 22–27%. 

The maximum rate of canopy photosynthesis (P gmax ) has an 

asymptotic value given by 

P gmax = L o A mo + L s A ms as I → ∞ (3) 

where I is the incident PAR, A mo is the maximum rate of sunlit 

leaf photosynthesis, and A ms is the maximum rate of shaded 

leaf photosynthesis. The maximum quantum yield of the canopy 

at full light interception is 

dP g /dI = α(1 + τ) as I → 0 (4) 

where α is the leaf quantum yield in low irradiance and τ is 

leaf transmittance. It can be seen that the limits for canopy 

photosynthesis are set by the properties of the individual leaves. 

At LAIs greater than approximately 6.0, improvements in 

canopy photosynthesis can only result from improvements in 

leaf quantum yield and/or leaf photosynthesis. 

At daytime temperatures typical of the tropics (>26 °C), 

the quantum yield of C 4 plants is greater than that of C 3 plants 

(Ehleringer and Björkman 1977) and the optimum temperature 

for rice plants is close to 30 °C, whereas that for C 4 plants 

is closer to 36 °C. Climate change models predict that by 2100 

mean planet-wide surface temperatures will rise by 1.4 to 5.8 

°C, making C 4 -ness even more desirable for tropical and semitropical 

environments. 


Redesigning photosynthesis 

There are four strategies for producing a C 4 rice plant: (1) 

achieve single-cell C 4 photosynthesis by combining the C 4 

pathway and the Calvin cycle in one cell, (2) produce a C 3 -C 4 

intermediate with reduced photorespiration, (3) produce a C 4 

rice with Kranz anatomy and specialization of chloroplast function 

in mesophyll and bundle-sheath cells, and (4) produce 

more Rubisco, or better Rubisco, or find ways to make Rubisco 

work harder. 

All of the genes for the C 4 syndrome are already present 

in C 3 plants and some plants switch between C 3 and C 4 modes 

of photosynthesis. In the sedge Eleocharis vivipara, the submerged 

culms are C 3 but the emergent culms are C 4 with the 

Kranz anatomy (Ueno et al 1988). Orcuttia has been shown to 

have two types of anatomy: the terrestrial leaves are C 4 with 

Kranz anatomy but submerged leaves have C 4 biochemistry 

without Kranz anatomy (Keeley 1998). Hibberd and Quick 

(2002) showed that cells in stems and petioles of a typical C 3 

plant, tobacco, displayed C 4 characteristics, opening up the 

possibility that a form of C 4 -ness may already exist in some 

rice plants. 

Recently, another plant has been reported to operate the 

C 4 mechanism within single cells (Voznesenskaya et al 2002). 

In the chenopod Borszczowia aralocaspica, the elongated 

chlorenchymatous cells are closely packed, without intercellular 

air spaces, at the inner end where they have abundant, 

large chloroplasts. The outer two-thirds of the cells have the 

appearance of mesophyll cells, with normal air spaces. 

Work in progress 

At IRRI, we have been screening the wild rice collection for 

evidence of C 4 traits. The δ 13 C values range from –31% to 

–25% and the interveinal spacing (at the middle of the blade) 

ranges from 167 to 313 µm (the value for IR72 is 227 µm). 

The number of vascular bundles per unit width was not related 

to the overall width of the leaf. Currently, we are counting the 

number of chloroplasts in the bundle-sheath cells and measuring 

the mesophyll cell size. Work is under way to transfer maize 

genes for PEP carboxylase (PEPC) and pyruvate, orthophosphate 

dikinase (PPDK), individually and in combination, into 

elite IRRI indica cultivars. Table 1 shows the advantage of 

maize over rice in water-use efficiency, radiation-use efficiency, 

and photosynthetic nitrogen-use efficiency. It is hard to imagine 

any trait other than C 4 photosynthesis that could increase 

the value of rice by US$44 billion annually through increased 

yield and reduced water and nitrogen costs. 

References 

Baker JT, Allen LH Jr, Boote KJ, Jones P, Jones JW. 1990. Rice 

photosynthesis and evapotranspiration in subambient, ambient 

and superambient carbon dioxide concentrations. Agron. 

J. 82:834-840. 

Ehleringer J, Björkman O. 1977. Quantum yields for CO 2 uptake in 

C 3 and C 4 plants: dependence on temperature, [CO 2 ], and [O 2 ] 

concentration. Plant Physiol. 59:86-90. 

Evans JR, von Caemmerer S. 2000. Would C 4 rice produce more 

biomass than C 3 rice In: Sheehy JE, Mitchell PL, Hardy B, 

editors. Redesigning rice photosynthesis to increase yield. 

Proceedings of the Workshop on The Quest to Reduce Hunger: 

Redesigning Rice Photosynthesis, 30 Nov.-3 Dec. 1999, 

Los Baños, Philippines. Makati City (Philippines): International 

Rice Research Institute and Amsterdam (The Netherlands): 

Elsevier Science B.V. p 53-71. 

Hibberd JM, Quick WP. 2002. Characteristics of C 4 photosynthesis 

in stems and petioles of C 3 flowering plants. Nature 415:451- 

454. 

Keeley JE. 1998. C 4 photosynthetic modifications in the evolutionary 

transition from land to water in aquatic grasses. Oecologia 

116:85-97. 

Kiniry JR, Jones CA, O’Toole JC, Blanchet R, Cabelguenne M, 

Spanel DA. 1989. Radiation use efficiency in biomass accumulation 

prior to grain-filling for five crop species. Field Crops 

Res. 20:51-64. 

Monteith JL. 1965. Light distribution and photosynthesis in field 

crops. Ann. Bot. 29:17-37. 

Peng S, Cassman KG, Virmani SS, Sheehy JE, Khush GS. 1999. 

Yield potential trends of tropical rice since the release of IR8 

and the challenge of increasing rice yield potential. Crop Sci. 

39:1552-1559. 

Sage RF. 2002. C 4 photosynthesis in terrestrial plants does not require 

Kranz anatomy. Trends Plant Sci. 7:283-285. 

Sheehy JE. 2000. Limits to yield for C 3 and C 4 rice: an agronomist’s 

view. In: Sheehy JE, Mitchell PL, Hardy B, editors. Redesigning 

rice photosynthesis to increase yield. Proceedings of 

the Workshop on The Quest to Reduce Hunger: Redesigning 

Rice Photosynthesis, 30 Nov.-3 Dec. 1999, Los Baños, Philippines. 

Makati City (Philippines): International Rice Research 

Institute and Amsterdam (The Netherlands): Elsevier Science 

B.V. p 39-52. 

Sinclair TR, Sheehy JE. 1999. Erect leaves and photosynthesis in 

rice. Science 283:1456-1457. 

Ueno O, Samejima M, Muto S, Miyachi S. 1988. Photosynthetic 

characteristics of an amphibious plant, Eleocharis vivipara: 

expression of C 4 and C 3 modes in contrasting environments. 

Proc. Natl. Acad. Sci. USA 85:6733-6737. 

Voznesenskaya EV, Franceschi VR, Kiirats O, Artyusheva EG, Freitag 

H, Edwards GE. 2002. Proof of C 4 photosynthesis without 

Kranz anatomy in Binertia cycloptera (Chenopodiaceae). Plant 

J. 31:649-662. 

Notes 

Authors’ address: Crop, Soil, and Water Sciences Division, International 

Rice Research Institute (IRRI), DAPO Box 7777, Metro 

Manila, Philippines, e-mail: j.sheehy@cgiar.org. 


Progress in breeding the new plant type 

for yield improvement: a physiological view 

Shaobing Peng, Rebecca C. Laza, Romeo M. Visperas, Gurdev S. Khush, and Parminder Virk 

Increasing yield is still the most important objective of rice 

breeding programs in developing countries because of the 

growing demand for food resulting from population growth 

and a reduction in area devoted to rice production. The International 

Rice Research Institute (IRRI) began developing the 

new plant type (NPT) rice through ideotype breeding approaches 

in 1989 (Khush 1995). The goal was to develop an 

NPT with a yield potential 20–25% higher than that of existing 

semidwarf rice varieties under a tropical environment during 

the dry season. The NPT was designed based on the results 

of simulation modeling and the new traits were mostly morphological 

since these are easier to select than physiological 

traits in breeding programs. The proposed NPT has a low 

tillering capacity (3 to 4 tillers when direct-seeded), few unproductive 

tillers, 200 to 250 grains per panicle, a plant height 

of 90 to 100 cm, thick and sturdy stems, leaves that are thick, 

dark green, and erect, a vigorous root system, 100 to 130 days’ 

growth duration, and increased harvest index (Peng et al 1994). 

In 1989, about 2,000 entries from the IRRI germplasm 

bank were grown to identify donors for the desired traits (Khush 

1995). Donors for the low-tillering trait, large panicles, thick 

stems, a vigorous root system, and short stature were identified 

in the “bulu” or javanica germplasm mainly from Indonesia. 

This germplasm is now referred to as the tropical japonica 

(Khush 1995). Hybridization began in the dry season of 1990. 

The first-generation NPT lines based on tropical japonicas were 

developed in less than 5 years. They were grown in a replicated 

observational trial for the first time in late 1993. As intended, 

the NPT lines had large panicles, few unproductive 

tillers, and lodging resistance. However, they did not yield well 

because of limited biomass production and poor grain filling 

(Peng and Khush 2003). Low biomass production was caused 

by an excessive reduction in tillering capacity and low photosynthetic 

rate. Poor grain filling was associated with panicle 

morphology and source limitation. The poor grain filling of 

NPT lines was probably caused by the lack of apical dominance 

within a panicle (Yamagishi et al 1996), the compact 

arrangement of spikelets on the panicle (Khush and Peng 1996), 

a limited number of large vascular bundles for assimilate transport 

(S. Akita, personal communication), and source limitation 

because of early leaf senescence (Ladha et al 1998). The 

first-generation NPT lines were also susceptible to diseases 

and insects and had poor grain quality. Therefore, they could 

not be released for planting in farmers’ fields. However, they 

are valuable genetic materials used in rice breeding programs 

worldwide. 

Breeding of the second-generation new plant type 

In 1995, development of second-generation NPT lines began 

by crossing first-generation tropical japonica NPT lines with 

elite indica parents. Multiple site-year comparisons of firstgeneration 

NPT lines with the highest-yielding indica varieties 

have shown that the original NPT design did not have sufficient 

tillering capacity. An increase in tillering capacity is 

needed to increase biomass production and to improve compensation 

when tillers are lost because of insect damage or 

other causes during the vegetative stage. A slightly smaller 

panicle size without a change in panicle length also appeared 

to be advantageous to reduce the compact arrangement of spikelets. 

Genes from indica parents have effectively reduced panicle 

size and increased tillering capacity in the second-generation 

NPT lines. Indica germplasm also helped improve other NPT 

attributes such as grain quality and disease and insect resistance. 

Some second-generation NPT lines (F 5 generation) with 

the above refinements were then selected and were planted in 

a replicated observational trial for the first time in the 1998 

wet season. These second-generation NPT lines have been 

tested in breeders’ replicated yield trials since the 2001 dry 

season and in replicated agronomic trials since the 2002 dry 

season. However, only data from the replicated agronomic trials 

in the dry seasons of 2003 and 2004 will be presented in 

this paper. 

Performance of the second-generation new plant type 

Field experiments were conducted under flooded irrigation at 

the IRRI farm in the dry season of 2003 and 2004. Five second-generation 

NPT lines and five indica inbred check varieties 

were grown. Seedlings were transplanted at a hill spacing 

of 20 × 20 cm. The plants received a basal fertilizer supply of 

30 kg P ha –1 , 40 kg K ha –1 , and 5 kg Zn ha –1 incorporated 1 d 

before transplanting. Total N fertilizer applied was 200 kg 

ha –1 in four splits: basal (60 kg ha –1 ), midtillering (40 kg ha –1 ), 

panicle initiation (60 kg ha –1 ), and heading (40 kg ha –1 ). Standard 

cultural management practices were followed. To avoid 

yield loss, pests were intensively controlled using recommended 

pesticides. 

In the 2003 dry season, PSBRc52 produced the highest 

yield among indica inbred check varieties and IR72967-12-2- 

3 was the top yielder among the second-generation NPT lines 

(Table 1). IR72967-12-2-3 produced 10.16 t ha –1 , which was 

significantly higher than the yield of PSBRc52. The higher 

yield was associated with the higher aboveground total biomass 

production and greater grain weight in IR72967-12-2-3. 


Table 1. Crop growth duration, grain yield, and yield attributes of second-generation new plant type lines and 

check varieties grown at the IRRI farm in the dry season of 2003. 

Crop Grain Biomass Harvest Panicles Spikelets Grain 1,000-grain 

Genotype duration yield production index m –2 panicle –1 filling weight 

(d) (t ha –1 ) (g m –2 ) (%) (no.) (no.) (%) (g) 

Indica checks 

PSBRc52 117 9.51 1,762 49.8 483 94.4 86.9 22.2 

IR72 117 9.31 1,731 49.2 468 84.5 89.4 24.1 

PSBRc54 119 9.06 1,766 47.2 412 106.0 84.1 22.7 

IR8 133 8.72 1,858 41.8 379 94.8 76.0 28.4 

PSBRc82 111 8.54 1,515 50.8 452 82.2 87.6 23.7 

Mean 119 9.03 1,726 47.8 439 92.4 84.8 24.2 

Second-generation new plant type lines 

IR72967-12-2-3 125 10.16 1,877 47.1 385 112.3 73.6 27.8 

IR73930-41-5-3-1 117 9.00 1,729 47.0 395 100.4 82.8 24.8 

IR71700-247-1-1-2 111 8.97 1,480 54.1 493 106.6 77.0 19.8 

IR73935-51-1-3-1 117 8.59 1,566 46.3 366 97.7 84.8 24.0 

IR72158-16-3-3-1 121 6.27 1,712 32.5 343 113.4 56.9 25.1 

Mean 118 8.60 1,673 45.4 396 106.1 75.0 24.3 

LSD (0.05) 0.49 126 2.3 28 6.4 3.6 1.2 

Table 2. Crop growth duration, grain yield, and yield attributes of second-generation new plant type lines and 

check varieties grown at the IRRI farm in the dry season of 2004. 

Crop Grain Biomass Harvest Panicles Spikelets Grain 1,000-grain 

Genotype duration yield production index m –2 panicle –1 filling weight 

(d) (t ha –1 ) (g m –2 ) (%) (no.) (no.) (%) (g) 

Indica checks 

IR72903-121-2-1-2 124 9.32 1,907 44.7 376 118.6 79.1 24.2 

PSBRc18 116 8.37 1,775 42.6 418 103.0 78.6 22.4 

PSBRc52 111 8.04 1,612 43.3 466 95.3 78.8 20.0 

IR68440-36-2-2-3 116 7.96 1,758 42.1 447 107.9 79.1 19.4 

IR72 116 7.25 1,720 41.4 481 81.8 82.1 22.0 

Mean 117 8.19 1,754 42.8 438 101.3 79.5 21.6 

Second-generation new plant type lines 

IR72158-16-3-3-1 118 8.76 1,795 42.0 350 135.6 61.8 25.7 

IR71700-247-1-1-2 110 8.65 1,493 47.5 512 99.5 74.7 18.6 

IR72967-12-2-3 118 8.28 1,786 42.4 352 120.0 66.0 27.2 

IR73459-120-2-2-3 124 7.78 1,850 39.7 339 108.3 81.9 24.4 

IR73930-41-5-3-1 116 7.39 1,617 41.6 364 98.3 80.1 23.5 

Mean 117 8.17 1,708 42.6 383 112.3 72.9 23.9 

LSD (0.05) 0.65 112 2.9 26 9.3 4.0 0.4 

Overall, the second-generation NPT lines did not outyield the 

indica inbred check varieties. There was a small difference in 

biomass production and harvest index between the two groups. 

Grain filling of the second-generation NPT lines was poorer 

than that of the indica inbred check varieties. The spikelet number 

per panicle of the second-generation NPT lines was less 

than 115, and was on average 15% greater than that of the 

indica inbred check varieties. 

In the 2004 dry season, an indica inbred line (IR72903- 

121-2-1-2) produced the highest yield (Table 2). IR71700-247- 

1-1-2 recorded the highest harvest index in both seasons. The 

grain yield in the 2004 dry season was lower than that in the 

2003 dry season because of lower solar radiation and higher 

minimum temperature in the 2004 dry season. Overall, there 

was no significant difference in grain yield, biomass production, 

and harvest index between the second-generation NPT 

lines and indica inbred check varieties. As observed in the 2003 

dry season, grain filling of the second-generation NPT lines 

was poorer than that of the indica inbred check varieties. On 

average, spikelet number per panicle of the second-generation 

NPT lines was 11% greater than that of the indica inbred check 

varieties, which was just enough to compensate for the reduction 

in panicle number per m 2 . 


Directions for further improvement 

Progress has been made in the second-generation NPT lines 

developed by crossing elite indica with improved tropical 

japonica and by modifying the original plant type design. Indica 

germplasm also helped improve other NPT attributes such 

as grain quality and disease and insect resistance. Many second-generation 

NPT lines outyielded the first-generation NPT 

lines in both the dry and wet seasons because of their improved 

crop biomass production and grain-filling percentage. Occasionally, 

the second-generation NPT lines outyielded the indica 

inbred check varieties. Overall, the second-generation NPT 

lines have not shown yield superiority over the indica inbred 

check varieties because of the following reasons: 

1. The second-generation NPT lines did not show higher 

biomass production or harvest index than the indica 

inbred check varieties. 

2. The panicle size of the second-generation NPT lines 

was not substantially greater than that of the indica 

inbred check varieties. 

3. The grain-filling percentage of the second-generation 

NPT lines has not reached the same level as that of 

the indica inbred check varieties. 

To further improve the grain yield of the second-generation 

NPT lines, the source-sink relation should be balanced by 

improving photosynthesis and delaying leaf senescence of the 

top three leaves during the ripening phase. The morphology of 

the top three leaves and the panicle position within the canopy 

are also critical for maintaining the balance between source 

and sink. Panicles with an average size of 150 spikelets per 

panicle are necessary to assure the increase in sink capacity. 

These morpho-physiological traits can be expressed as the yield 

attributes that are relatively easy to measure in breeding programs. 

These yield attributes are 330 panicles per m 2 , 150 

spikelets per panicle, 80% grain filling, 25 mg of grain weight 

(based on oven-dry weight), 20 t ha –1 of aboveground total 

biomass (based on oven-dry weight), and 50% harvest index. 

These attributes are required to produce a yield of 11 t ha –1 at 

14% moisture content. As the effort of breeding for secondgeneration 

NPT continues, we expect that more elite secondgeneration 

NPT lines with improved yield potential, disease 

and insect resistance, and grain quality will be developed at 

IRRI. These elite second-generation NPT lines should increase 

the yield potential of irrigated lowland rice by about 10% in 

the tropics. Concurrent modification of crop management practices 

such as transplanting with different seedling age, planting 

geometry, fertilization, irrigation scheme, and weed control 

is required for the NPT lines to fully express their yield 

potential. 

References 

Khush GS. 1995. Breaking the yield frontier of rice. GeoJournal 

35:329-332. 

Khush GS, Peng S. 1996. Breaking the yield frontier of rice. In: 

Reynolds MP, Rajaram S, McNab A, editors. Increasing yield 

potential in wheat: breaking the barriers. El Batán (Mexico): 

International Maize and Wheat Improvement Center. p 36- 

51. 

Ladha JK, Kirk GJD, Bennett J, Peng S, Reddy CK, Reddy PM, 

Singh U. 1998. Opportunities for increased nitrogen-use efficiency 

from improved lowland rice germplasm. Field Crops 

Res. 56:41-71. 

Peng S, Khush GS. 2003. Four decades of breeding for varietal improvement 

of irrigated lowland rice in the International Rice 

Research Institute. Plant Prod. Sci. 6:157-164. 

Peng S, Khush GS, Cassman KG. 1994. Evaluation of a new plant 

ideotype for increased yield potential. In: Cassman KG, editor. 

Breaking the yield barrier. Proceedings of a Workshop on 

Rice Yield Potential in Favorable Environments. Los Baños 


Yamagishi T, Peng S, Cassman KG, Ishii R. 1996. Studies on grain 

filling characteristics in “New Plant Type” rice lines developed 

in IRRI. Jpn. J. Crop Sci. 65(Extra issue No. 2):169- 

170. 

Notes 

Authors’ addresses: Shaobing Peng, Rebecca C. Laza, and Romeo 

M. Visperas, Crop, Soil, and Water Sciences Division, International 

Rice Research Institute (IRRI), DAPO Box 7777, 

Metro Manila, Philippines; Gurdev S. Khush, 416 Cabrillo 

Avenue, Davis, CA 95616, USA; Parminder Virk, Plant Breeding, 

Genetics, and Biochemistry Division, IRRI. 

Current status and prospects of rice breeding 

for increased yield in China 

Wan Jianmin 

Rice is the main staple food in China, being planted on 23% of 

the world’s rice area and contributing 37% of global rice production. 

Rice production in China plays an important role for 

the Chinese people and also affects the world rice market. Although 

China, where one-fifth of the world’s population de- 

pends on 7% of the world’s arable land, will not encounter any 

food security problems in the next few years, it has hidden 

risks of food shortage as its rapid economic growth eats away 

at arable land. 


Area (million ha) 

40 

35 

30 

25 

Area of rough rice 

Area of hybrid rice 

20 

15 

10 

5 

0 

1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 

Year 

Fig. 1. Area of hybrid rice (million ha). 

Current situation of breeding to increase rice yield in China 

A look back at crop breeding for super high yield 

A series of high-yielding semidwarf varieties has made dramatic 

progress in rice cultivation since the 1950s. Hybrid rice 

at the end of the 1970s marked another breakthrough in China’s 

rice production (Yang et al 1996, Zhou et al 1997, Huang et al 

2003). For the first time in 1977, hybrid rice was produced in 

farmers’ fields. Subsequently, the area of hybrid rice expanded 

rapidly to 54% of the total rice area in 1997, and has remained 

at more than 60% since then (Fig. 1). 

Despite this progress, China’s grain output dipped from 

a record high of 512 million tons in 1998 to 435 million tons 

in 2003. In view of the rising population and falling number of 

hectares of land across the world, the most effective and economical 

way to increase production is considered to be growing 

hybrid rice. Hybrid rice, which has been grown in China 

for a long period, will continue to play a significant role in 

guaranteeing China’s food security. 

Advancement of the program 

for super high-yielding rice 

The current technology of hybrid rice can increase the yield 

ceiling by 15–20% over that of the best commercial varieties. 

Several breeding programs for super high-yielding rice have 

been proposed since the 1980s. In 1996, the Chinese Ministry 

of Agriculture set up a super high-yielding rice program with 

the targets listed in Table 1. 

In recent years, a cooperative research program in hybrid 

rice breeding has been conducted by the Jiangsu Academy 

of Agricultural Sciences and China National Hybrid Rice 

Research Center. In this program, TGMS (thermosensitive 

genetic male sterility) line Pei’ai64S was used as the female 

parent and test-crossed with a number of breeding lines. Then, 

several combinations with a super high yield potential were 

screened, including Pei’ai64S/E32, which attained an average 

yield as high as 13.26 t ha –1 , with growth duration of 130 days 

on a total area of 0.24 ha at three locations in 1997. This hy- 

Table 1. Yield standards (t ha –1 ) set for super rice in China. a 

Phase 

Hybrid rice 

Yield 

Early-season Late-season Single-season increase 

indica indica rice 

Before 1996 7.50 7.50 8.25 0 

1996-2000 9.75 9.75 10.50 >20% 

2001-2005 11.25 11.25 12.00 >40% 

a It is required that grain yield be up to standards at two locations of an ecological 

area with a planting scale of 6.7 ha at each location for two consecutive years. 

brid has met the standard of super high-yielding rice, though 

only in a yield trial in a small area. Pei’ai64S/E32 is characterized 

by a set of yield components that lead to a theoretical 

yield of 13.95 t ha –1 and actual yield of 12.87 t ha –1 . Several 

hybrids, such as Chuanxiangyou2, P088S/0293, and 

IIyouming86, increased their yield by more than 20%. 

Theory of breeding for super high-yielding rice 

Two-line and three-line systems 

Currently, the three-line system of hybrid rice production is 

being followed. But it is known that the two-line system, based 

on the photosensitive or thermosensitive genetic male sterility 

(P or T-GMS) system, is more efficient and cost-effective. 

Combination of ideotype and heterosis 

Since Johan and Jeroen (2003) proposed the ideotype concept, 

several models have been proposed for super high-yielding 

rice: the low-tillering and large panicle model by Khush 

(1996), the bushy type and rapid-growing model by Huang 

(2003), the ideal plant type and huge rice model by Yang et al 

(1996), and the heavy panicle model by Zhou et al (1997). 

The new plant type (NPT) being developed by IRRI might 

raise current yield by 20–25%. These models, yet to be proven 

in practice, provide the leading concepts for super high-yield- 


ing programs since they are based on certain theories and practical 

experiences. 

Based on the characteristics of Pei’ai64S/E32 in hybrid 

rice breeding, Yuan (1997) proposed a morphological model 

of super high-yielding rice in terms of (1) plant height, (2) the 

uppermost three leaves, (3) plant type, (4) panicle weight and 

number, (5) leaf area index (LAI) and ratio of leaf area to grains, 

and (6) harvest index of above 0.55. 

Use of intersubspecific heterosis 

As the heterosis of intersubspecific hybrids is much stronger 

than that of intervarietal hybrids, its use is one of the most 

feasible approaches for realizing super high yield. To exploit 

intersubspecific super high-yielding hybrid rice, the development 

of various lines with wide compatibility (WC), especially 

lines with a broad spectrum of compatibility, is important (Wan 

and Ikehashi 1996). By incorporating WC genes into restorer 

lines and male sterile lines of indica, japonica, or intermediate 

type, each with a different growth duration, various super highyielding 

hybrids will be developed for different ecological 

environments. 

Genotype × environment interactions 

in breeding super high-yielding rice 

The present superior-yielding varieties exhibit variable performance 

because of a high proportion of G × E interaction. 

There is a need to identify and release stably yielding varieties 

even on a specific-area basis, instead of relatively less stable 

varieties on a wide-area basis. There are strong genotypic differences 

among varieties for this interaction as well as methods 

for selecting varieties that are more stable across environments. 

Prior to releasing varieties, it is possible to select varieties 

with a stable performance even in unfavorable environments 

or management regimes. 

Future prospects of breeding rice for higher yield in China 

The extension of already released varieties is an immediate 

possibility for breaking the yield barrier. Further, the new plant 

type or super rice, hybrid rice, and genetically engineered 

transgenic rice are on the horizon. 

Germplasm enhancement 

The historic discovery of the semidwarfing gene (sd1) in 

Taichung (Taiwan, China) revolutionized rice production in 

the world, extending varieties carrying this gene to almost all 

the rice-growing countries. Following this model, favorable 

QTLs have been identified from a set of wild rice based on 

molecular analyses and field experiments. Near-isogenic lines 

carrying the QTLs are created by marker-facilitated backcrossing 

and selection. 

× indica hybrids, the next breakthrough in yield could be set in 

motion by the use of indica × tropical japonica and indica × 

NPT rice. 

Breeding by design 

Breeding by design is a concept that aims to control all allelic 

variation at all loci of agronomic importance (Steven 1998, 

Ashikari and Matsuoka 2002, Johan and Jeroen 2003), which 

involves the integrative, complementary application of the technological 

tools and materials currently available to develop 

superior varieties. This concept can be achieved by a combination 

of precise genetic mapping, high-resolution genotyping, 

and extensive phenotyping. This approach to all traits of agronomic 

importance will require great organizational skills and 

will be a major effort. The systematic application of markerassisted 

selection could lead to the production of superior varieties 

within five to ten years. 

References 

Ashikari M, Matsuoka M. 2002. Application of rice genomics to 

plant biology and breeding. Bot. Bull. Acad. Sin. 43:1-11. 

Huang Y. 2003. Construction and advancement of rice ecological 

breeding system. World Sci. Tech. Res. 4:1-8. (In Chinese.) 

Johan DP, Jeroen RV. 2003. Breeding by design. Trends Plant Sci. 

8:330-334. 

Khush GS. 1996. Prospects of and approaches to increasing the genetic 

yield potential of rice. In: Evenson RE et al, editors. 

Rice research in Asia: progress and priorities. Wallingford 

(UK): CAB International and IRRI. p 59-71. 

Steven JK. 1998. Marker-assisted selection as a strategy for increasing 

the probability of selecting superior genotypes. Crop Sci. 

38:1164-1174. 

Wan J, Ikehashi H. 1996. Two new loci for hybrid sterility in cultivated 

rice. Theor. Appl. Genet. 92:183-190. 

Yang S, Zhang B, Chen W, Xu Z, Wang J. 1996. Theories and methods 

of rice breeding for maximum yield. Acta Agron. Sin. 

22(3):295-304. 

Yuan L. 1997. Hybrid rice breeding for super high yield. Hybrid 

Rice 12(6):1-6. 

Zhou K, Wang X, Li S, Li P, Li H, Huang G, Liu T, Shen M, 1997. 

The study on heavy panicle type of inter-subspecific hybrid 

rice (Oryza sativa L.). Sci. Agric. Sin. 30(5):91-93. 

Notes 

Author’s address: Institute of Crop Sciences, Chinese Academy of 

Agricultural Sciences, Beijing 100081, China, e-mail: 

wanjm@caas.net.cn. 

Hybrid rice 

All the rice hybrids grown so far are indica hybrids, except for 

japonica hybrids in the northern part of China. Since the indica 

× tropical japonica hybrids are shown to outyield indica 


A large-grain rice cultivar, Akita 63, exhibits high yield 

and N-use efficiency for grain production 

Tadahiko Mae, Ayako Inaba, Yoshihiro Kaneta, Satoshi Masaki, Mizuo Sasaki, and Amane Makino 

It is estimated that the world’s population will increase 1.4- to 

1.5-fold by 2025, and this projected increase will be mostly in 

Asia (IRRI 1995). More than 90% of the world’s production 

of rice is in Asia. It is therefore crucial to increase rice production 

within a relatively short period. With little scope for expanding 

the area of land available for rice cultivation, this increase 

in rice production must be achieved by an increase in 

yield from the land currently used for such cultivation. On the 

other hand, interest is increasing in the environmental impact 

of nitrogen (N) management practices (Cassman et al 1998). 

Therefore, it is desirable to have high-yielding rice cultivars 

with high N-use efficiency for grain production. However, selection 

or breeding of such rice cultivars has not been seriously 

attempted (Ladha et al 1998, Ying et al 1998a,b). 

A new large-grain japonica-type cultivar of rice, Akita 

63, has recently been released by the Akita Agricultural Experimental 

Station in northern Japan and its grain size is 30– 

31 mg, which is 25–50% larger than that of current japonica 

cultivars. In test cultivation with standard levels of fertilization 

at different locations in Akita Prefecture and in different 

years, the average yield of Akita 63 was 18% higher than that 

of Akitakomachi, which is the current leading cultivar of Akita 

Prefecture (Masaki et al, unpublished data), indicating that 

Akita 63 might be a new type of high-yielding cultivar with 

high N-use efficiency. Although high productivity with large 

grain cultivars of rice has been previously reported (Kobayashi 

et al 1990, Wang et al 1995), their physiological (internal) N- 

use efficiency (NUE) defined as grain yield per unit of N ac- 

cumulated in the aboveground part at harvest was not examined. 

Herein, we report on the achievement of a high yield of 

over 12 t ha –1 of rough rice with Akita 63 and analyze the 

factors responsible for its high yield compared with the yield 

of common reference cultivars. Finally, the yield-limiting factors 

of recent japonica cultivars are discussed in terms of physiological 

NUE. 


Akita 63 was grown with different levels of N supply (0–160 

kg N ha –1 ) in an experimental field of the Agricultural Experimental 

Station of Akita Prefecture, Japan (39°34′N, 140°11′E, 

16 m altitude), for three years (2000-02). As reference cultivars 

of japonica rice, a high-yielding local cultivar 

(Yukigesyou), a common cultivar (Toyonishiki), and a modern 

cultivar (Akitakomachi) were also grown with the same N 

treatments. The soil type was a Gley soil (Eutric Gleysols; FAO) 

with pH 5.3, 31.6 g total C kg –1 , 2.4 g total N kg –1 , and 25.2 

cmol kg –1 cation exchange capacity. Phosphorus (43 kg P 

ha –1 as fused phosphate) was applied to all plots 30 days before 

transplanting. Potassium was not supplied because of its 

sufficient level in the soil. Two to four levels of N fertilization 

were applied by using ammonium sulfate and controlled-release 

fertilizers (urea LP100 type and LPS100 type polyolefincoated, 

Chisso Co., Japan) (Table 1). A basal application was 

conducted a week before transplanting. Weeds, insects, and 

Table 1. Rates of nitrogen application for rice cultivation. 

Year 

N treatment 

Nitrogen application (g N m -2 ) 

Basal dressing Topdressing Total 

2000 High level 2 a + 6 b 2 a × 4 times 16 

Standard level 4 a 2 a × 1 time 6 

No application 0 0 0 

Single application 10 b 0 10 


Standard level 4 a 2 a × 1 time 6 

No application 0 0 0 

Single application 10 b + 4 c 0 14 


Standard level 4 a 2 a × 2 times 8 

a Ammonium sulfate. b Controlled-release fertilizer (urea LP100 type polyolefin-coated, Chisso 

Co., Japan). c Controlled-release fertilizer (urea LSP100 type polyolefin-coated, Chisso Co., 

Japan). 


Panicle dry weight (g m –2 ) 

1,500 

1,200 

900 

600 

300 

Akita 63 in 2000 



Yukigeshou in 2000 

Toyonishiki in 2001 

Akitakomachi in 2000 

Akitakomachi in 2002 

Akita 63: y = 48.32x + 276.7 

R 2 = 0.853 

Reference y = 52.89x + 91.1 

cultivars: R 2 = 0.961 

0 

0 5 10 15 20 25 

Plant nitrogen content (g m –2 ) 

Fig. 1. Relationship between panicle dry weight and plant nitrogen content 

per unit land area at harvest in Akita 63 and the reference cultivars. 

diseases were controlled as required to avoid yield loss. All 

the experimental plots were arranged in a randomized design 

with three replicates, except in the case of Toyonishiki, for 

which there were only two replicates at the standard level and 

no application treatments in 2001. The area of each plot was 

26.25 m 2 (3.5 m wide and 7.5 m long). About 30 five-day-old 

seedlings were transplanted with a rice transplanter on a day 

in the middle of May at a hill spacing of 0.3 × 0.14 m (24 hills 

m –2 ) with 4–5 seedlings per hill. 

The sample plants were separated into leaf blades, culms 

plus leaf sheaths, panicles, and other parts (dead parts and nonreproductive 

tillers). Leaf area was measured with a leaf area 

meter (Type-AMM, Hayashi-Denko, Tokyo, Japan). All 

samples were oven-dried at 105 °C for several days, weighed, 

and powdered. At the time of harvest, plants from three additional 

hills with the mean panicle number from each plot were 

collected and hand-threshed for measurement of the number 

of filled and unfilled spikelets, and grain weight. A survey of 

yield was carried out as follows: plants from 80 hills were collected 

from the center part of each plot. Winnowed unhulled 

rice weight was measured after reaping, threshing, and wind 

selection. Unhulled rice of 80 hills was husked to obtain whole 

brown rice, and then put through a 1.8-mm sieve to remove 

any immature kernels (brown rice). The weight of winnowed 

unhulled rice and that of brown rice was adjusted to a moisture 

content of 0.15 g H 2 O g –1 fresh weight. The harvest index 

was defined as the ratio of panicle dry weight to the total dry 

weight of the aboveground part. 

N content was determined with Nesler’s reagent after 

Kjeldahl digestion of powdered samples with H 2 SO 4 and H 2 O 2 . 


The grain yield of Akita 63 was 22–58% higher than that of 

the reference cultivars for all N treatments throughout the threeyear 

period. The yields were high in 2000 and 2001 and low in 

2002, mainly because of the differences in climatic conditions 

during the culture period among the three years. The highest 

yield was 12.8 t ha –1 of rough rice (9.83 t ha –1 of brown rice), 

close to the highest yield of japonica rice (10.52 t ha –1 of brown 

rice) previously recorded in Japan (Honya 1989). The dry 

weight of the aboveground part, the number of total spikelets, 

and the amount of N accumulated in the aboveground part 

(plant N) per unit land area at harvest were almost the same 

between Akita 63 and the reference cultivars in 2000 and larger 

in Akita 63 than in the reference cultivars in 2001 and 2002 

for all the N treatments. However, the dry weight of the 

aboveground part and the number of total spikelets for a given 

amount of plant N per unit land area at harvest did not differ 

between Akita 63 and the reference cultivars. Clear differences 

in sink capacity (total spikelet number per unit land area × the 

weight of 1,000 kernels) and panicle dry weight for a given 

amount of plant N per unit land area were found between Akita 

63 and the reference cultivars. The sink capacity was 30–40% 

greater in Akita 63 than in the reference cultivars because the 

weight of 1,000 kernels was 29–36% greater in Akita 63. Figure 

1 shows the relationship between panicle dry weight and 

plant N content per unit land area at harvest. The panicle dry 

weight for a given amount of plant N per unit land area was 

12–43% greater in Akita 63 than in the reference cultivars. 

This was because the ratio of dry matter partitioning to the 


panicle (harvest index) was higher in Akita 63 than in the reference 

cultivars. The difference was more pronounced with 

less plant N content. No difference in the relationship between 

leaf area index and plant N content per unit land area was found 

between Akita 63 and the reference cultivars, but panicle dry 

weight for a given unit of leaf area index was 13–48% greater 

in Akita 63 than in the reference cultivars. 

Reaccumulation of dry matter or starch in the culms and 

leaf sheaths was remarkable in the later stage of grain filling in 

the reference cultivars, but not in Akita 63. The reaccumulation 

in the reference cultivars could be attributed to their insufficient 

sink capacity because the proportion of the number of 

ripened grains to the total number of grains was more than 

90% even for the high N treatment in the reference cultivars, 

whereas the proportion was lower than 79% for the same N 

treatment in Akita 63. Those results indicate that Akita 63, 

which has a greater sink capacity for a given amount of plant 

N per unit land area, still had the capacity for further accumulation 

of dry matter in its panicles even in the later stage of 

grain filling, while the reference cultivars, which have relatively 

smaller sink capacities, had already nearly reached the 

upper limit of accumulation and had almost no space for further 

accumulation in their panicles in the later stage of grain 

filling. Thus, dry matter was mostly translocated into the culms 

and leaf sheaths in the late stage of grain filling in the reference 

cultivars under given growth conditions. As a consequence, 

the physiological (internal) N-use efficiency was higher 

in Akita 63 than in the reference cultivars. When the partitioning 

of 13 C into panicles, culms, leaf sheaths, leaf blades, and 

roots was examined at harvest after the photosynthetic assimilation 

of 13 CO 2 at different stages of grain filling, the proportion 

of 13 C partitioned into panicles was higher at all stages of 

grain filling in Akita 63 than in reference cultivar Toyonishiki 

(Mae et al, unpublished data). Reaccumulation of nonstructural 

carbohydrates in the culms and leaf sheaths at the later stages 

of grain filling seems to be a characteristic of the common 

cultivars of japonica rice. Similar trends have been previously 

reported between a high-yielding indica cultivar, Milyang 23, 

and a common japonica cultivar, Nipponbare (Saito et al 1991, 

Tsukaguchi et al 1996). Our study indicates that the enlargement 

of sink capacity per unit of plant N might increase the 

yield potential of common japonica cultivars of rice and improve 

their physiological N-use efficiency. 

Honya K. 1989. Analysis of high-yield cultivation of rice. Tokyo 

(Japan): Hakuyosha. 36 p. (In Japanese.) 

IRRI (International Rice Research Institute). 1995. Rice facts. Manila 

(Philippines): IRRI. p 2. 

Kobayashi A, Koga Y, Uchiyamada H, Samoto S, Horiuchi H, Miura 

K, Okuno K, Fujita Y, Uehara Y, Ishizaka S, Nakagahara M, 

Yamada T, Maruyama K. 1990. Breeding a new rice variety, 

Oochikara. Bull. Hokuriku National Agric. Exp. Stn. 32:85- 

104. (In Japanese.) 

Ladha JK, Kirk GJD, Bennett J, Peng S, Reddy CK, Reddy PM, 

Singh U. 1998. Opportunities for increasing nitrogen-use efficiency 

from improved lowland rice germplasm. Field Crops 

Res. 56:41-71. 

Saito K, Kashiwagi N, Kinoshita T, Ishihara H. 1991. Characteristics 

of dry matter production process in high yielding rice 

varieties. IV. Dry matter accumulation in the panicle. Jpn. J. 

Crop Sci. 60:255-263. 

Tsukaguchi T, Horie T, Ohnishi M. 1996. Filling percentage of rice 

spikelets as affected by availability of non-structural carbohydrates 

at the initial phase of grain filling. Jpn. J. Crop Sci. 

65:445-452. 

Wang W, Yamamoto Y, Nitta Y. 1995. Analysis of the factors of high 

yielding ability for a japonica type rice line, 9004, bred in 

China. 1. Comparison of yield ability with a Japanese rice 

variety under the same level of spikelet number per area. Jpn. 

J. Crop Sci. 64:545-555. 

Ying J, Peng S, He Q, Yang H, Yang C, Visperas RM, Cassman KG. 

1998a. Comparison of high-yield rice in tropical and subtropical 

environments. I. Determinants of grain and dry matter 

yields. Field Crops Res. 57:71-84. 

Ying J, Peng S, Yang C, Neng Zhou, Visperas RM, Cassman KG. 

1998b. Comparison of high-yield rice in tropical and subtropical 

environments. II. Nitrogen accumulation and utilization 

efficiency. Field Crops Res. 57:85-93. 

Notes 

Authors’ addresses: Tadahiko Mae, Ayako Inaba, Mizuo Sasaki, and 

Amane Makino, 

Department of Applied Plant Science, School of Agricultural Science, 

Tohoku University, 1-1 Tsutsumidori-Amamiyamachi, 

Aoba-ku, Sendai 981-8555, Japan; Yoshihiro Kaneta and 

Satoshi Masaki, the Agricultural Experimental Station of Akita 

Prefecture, Aikawa Yuuwamachi Kawabe-gun, Akita Prefecture 

010-1231, Japan, e-mail: 

hikomae@biochem.tohoku.ac.jp. 

References 

Cassman KG, Peng S, Olk DC, Ladha JK, Reichardt W, Dobermann 

A, Singh U. 1998. Opportunities for increased nitrogen-use 

efficiency from improved resource management in irrigated 

rice systems. Field Crops Res. 56:7-39. 


Using heterosis and hybrid rice 

to increase yield potential in China 

Xuhua Zhong, Shaobing Peng, Feng Wang, and Nongrong Huang 

The use of heterosis and hybrid rice has been one of the most 

important achievements in China. Hybrid rice occupies about 

15 million ha of fields, or half of China’s rice-planting area. 

The yield potential of hybrid rice is about 20% greater than 

that of inbreds. Since its adoption in 1976, hybrid rice had 

been planted on 271 million ha of rice land until 2001 in China. 

The increased grain from hybrid rice can feed 60 million people 

annually. 

China’s rice yield has experienced two quantum leaps 

since the 1950s (Peng et al 2004). The first one resulted from 

the extension of semidwarf rice varieties in the 1960s. Rice 

yield increased from 2.08 t ha –1 in 1961 to 3.14 t ha –1 in 1966, 

with an annual increase of 0.21 t ha –1 . The second leap was 

brought about by the development of hybrid rice in 1976. Grain 

yield increased from 3.50 t ha –1 in 1976 to 5.36 t ha –1 in 1984, 

with an annual increase of 0.23 t ha –1 . To feed the increasing 

population, China’s rice production must increase from 177 

million tons in 2003 to 240 million t in 2030. Yield must increase 

from 6.27 to 7.66 t ha –1 during the same period. Hybrid 

rice is expected to play an important role in the new increase 

in rice yield potential. 

Three generations of hybrid rice 

Three-line hybrid rice 

The first-generation hybrids used intervarietal heterosis through 

the three-line approach. Cytoplasmic genetic male sterility 

(CMS) was used to facilitate hybrid seed production. Yuan 

started his pioneering work on hybrid rice breeding in 1964. 

However, no significant progress was achieved until 1970, 

when the male sterile wild rice plant, “wild abortive,” was 

found. In 1973, all three lines—male sterile lines (A lines), 

maintainer lines (B lines), and restorer lines (R lines)—were 

made available. The heterosis of hybrid rice was shown by 

faster tillering, greater leaf area index, higher dry matter production, 

and a stronger root system. The three-line hybrids 

yielded 20% more than the best inbred check varieties. Planting 

area of hybrids was 0.135 million ha in 1976, which expanded 

rapidly to 8.84 million ha in 1984 and 16.6 million ha 

in 1990. 

Two-line hybrid rice 

The yield of hybrid rice remained stagnant until the mid-1980s. 

To break the yield barrier, the two-line approach using 

intersubspecific heterosis was proposed (Yuan 1997). Environment-sensitive 

genic male sterility, including photoperiodand 

thermosensitive sterility (P/TGMS), was employed in this 

approach. In contrast with the three-line system, normal vari- 

eties can be used as restorers to produce hybrid F 1 . The possibility 

to find good combinations is much greater than with the 

three-line approach. Second-generation hybrids yielded 5–10% 

more than three-line hybrids although their grain filling was 

poor. 

Super hybrid rice 

The Ministry of Agriculture of China established a mega project 

on the development of super rice in 1996 (Min et al 2002). 

The objectives were 

1. To develop “super” rice varieties with a maximum 

yield of 9–10.5 t ha –1 by 2000, 12 t ha –1 by 2005, and 

13.5 t ha –1 by 2015 measured from a planting area of 

at least 6.7 ha. 

2. To develop “super” rice varieties with a yield potential 

of 12 t ha –1 by 2000, 13.5 t ha –1 by 2005, and 15 

t ha –1 by 2015. These yields will be achieved in experimental 

and demonstration plots. 

3. To raise the national average rice yield to 6.9 t ha –1 

by 2010 and 7.5 t ha –1 by 2030. 

In addition, the super rice varieties should outyield the 

locally wide-grown check varieties by 10%, or daily yield 

achieve 100 kg ha –1 , with acceptable grain quality and pest 

resistance. Significant progress has been achieved in China’s 

super hybrid rice breeding in recent years. 

Strategy for super hybrid rice breeding 

The strategy for super hybrid rice breeding was to combine 

the ideotype approach with the use of intersubspecific heterosis 

(Yuan 1997). This was different from the counterpart 

projects of other countries and/or organizations. According to 

the principles of hybrid rice breeding, there are two ways to 

achieve super high yield. One is to make full use of the dominant 

complementary effects of the two parents to improve the 

morphological characteristics of the hybrid. Another is to extend 

the genetic diversity of the parents to increase the heterosis 

level. The heterosis of intersubspecific hybrids is much 

stronger than that of intervarietal ones. The use of 

intersubspecific hybrids is thus the most feasible approach for 

super high yield. The discovery of P/TGMS and wide compatibility 

(WC) genes made it possible to directly use 

intersubspecific heterosis. Several WC lines were developed 

as “bridges” to overcome the too great genetic difference in 

typical intersubspecific hybrids, which resulted in physiological 

barriers. The WC lines could freely cross with both indica 

and japonica varieties without problems in grain filling (Yang 

et al 1996). 


Table 1. Selected super hybrid rice released by the provincial or national seed board. 

Variety Breeder Planting area Year of Province of 

(000 ha) release release 

Lianyoupeijiu Jiangsu Academy of Agricultural Sciences; 3,700 1999 Jiangsu 

China National Hybrid Rice R&D Center 

Xieyou 9308 China National Rice Research Institute 673 1999 Zhejiang 

D you 527 Sichuan Agricultural University 667 2001 Sichuan 

Ganyou 527 Sichuan Agricultural University 533 2002 Sichuan 

Ganyou 827 Sichuan Agricultural University 133 2003 Sichuan 

Eryouhang 1 Fujian Academy of Agricultural Sciences 167 2004 Fujian 

Teyouhang 1 Fujian Academy of Agricultural Sciences 67 2002 Fujian 

Yueza 122 Guangdong Academy of Agricultural Sciences 80 2001 Guangdong 

Yueza 889 Guangdong Academy of Agricultural Sciences 20 2004 Guangdong 

Lianyou 932 Hubei Academy of Agricultural Sciences 13 2002 Hubei 

Liaoyou 1518 North China Japonica Hybrid Rice R&D Center n.a. a 2002 Liaoning 

P88S/0293 China National Hybrid Rice R&D Center n.a. n.a. n.a. 

a n.a. = not available. 

Morphological design of super hybrid rice 

The poor grain filling of second-generation hybrids and 

progress in IRRI’s new plant type breeding led Chinese scientists 

to pay more attention to morphology. Different versions 

of morphological design were developed according to regional 

ecological conditions. Well adopted are the “panicle-underleaves” 

type for the Yangtze River region (Yuan 2001), the 

“heavy-panicle” type for western China (Zhou et al 1997), and 

the “rapid-early-growth-and-deep-root” type for South China 

(Huang 2001). 

Yuan (2001) suggested the panicle-under-leaves type 

with the following morphological traits: 

Plant height about 100 cm, with culm length of 70 

cm. 

Top three leaves: 

1. Long: flag leaf is 50 cm and -2nd and -3rd leaves 

55 cm in length. The top two leaves are higher 

than the top of the panicle. 

2. Erect: leaf angles of the flag, -2nd, and -3rd leaves 

are around 5°, 10°, and 20°, respectively. 

3. Narrow and V-shaped: the leaves look narrow but 

they are 2 cm wide when flattened. 

4. Thick: 55 g m –2 of specific leaf weight for the 

top three leaves. 

Plant type: moderately compact type with moderate 

tillering capacity; droopy panicles, the panicle top 

about 60 cm from the ground after filled; erect leaf 

canopy without appearance of panicles. 

Panicle weight and number: grain weight is 5 g per 

panicle; 270–300 panicles m –2 . 

Leaf area index (LAI): the LAI of the top three leaves 

is about 6. 

Harvest index: about 0.55. 

Performance of selected super hybrid rice 

More than 10 hybrid varieties that meet the super rice criteria 

have been released by the provincial or national seed board 

(Table 1). The cumulated planting area of super hybrid rice is 

estimated at 6 million ha. 

Lianyoupeijiu is a two-line intersubspecific hybrid (Lu 

and Zou 2003). The female parent Pei’ai64S is an intermediate 

type between indica and japonica, whereas the restorer 9311 

is a typical indica line. In the 1997-98 yield trial, Lianyoupeijiu 

achieved 9.56 t ha –1 and outyielded check variety Shanyou 63 

by 7.75%. An average yield of 12.15 t ha –1 was achieved in a 

67.8-ha area in Jiangsu Province in 2001. 

Xieyou 9308 is a three-line intersubspecific hybrid 

(Cheng and Zhai 2000). The female parent Xieqingzao-A is a 

typical indica sterile line. The restorer line Zhonghui 9308 is 

an intermediate type. Grain yield ranged from 10.53 to 11.95 t 

ha –1 in farmers’ fields of 6.7 ha in Zhejiang Province during 

1999-2001 and a yield of 12.23 t ha –1 was recorded in a 697- 

m –2 field. 

Liaoyou 1518 is a japonica hybrid (Hua et al 2003). The 

female parent 151A is a japonica line. The male parent C418 

is an indica line with good compatibility with the japonica variety. 

In field trials in 2001, Liaoyou 1518 yielded from 8.37 

to 12.08 t ha –1 and outyielded check variety Liaojing 454 by 

6.37–30.2%. 

Common characteristics of released super hybrid rice 

The super hybrids have several characteristics in common: 

Many super hybrids, such as Lianyoupeijiu, Xieyou 

9308, Yueza 122, Liaoyou 1518, and P88S/0293, use 

indica-japonica intersubspecific heterosis. 

Large sink. The panicle size is significantly enlarged, 

with a small reduction in panicle number m –2 . Spikelets 

reach 35,000–45,000 m –2 . 


Conclusions 

Large source. The top three leaves of Lianyoupeijiu 

are longer and wider than those of the check varieties 

and age more slowly (Cheng and Zhai 2000). 

Good plant type. Plant height is 110–120 cm and taller 

than that of check varieties. The top three leaves are 

more erect during grain filling. For example, the leaf 

angle was 5.1 o , 8.1 o , and 15.1 o for the flag, 2nd, and 

3rd leaves of Lianyoupeijiu, and 7.3 o , 16.3 o , and 24.0 o 

for those of Shanyou 63, respectively. 

Super hybrids have more root biomass and a greater 

ratio of roots is located in deep soil than the check 

variety Shanyou 63. This is related to the slower senescence 

of leaves during grain filling. 

The increase in grain yield is mainly due to the increase 

in biomass while the harvest index remains unchanged. 

Hybrid rice has contributed significantly to increasing China’s 

rice production. The use of rice heterosis has experienced three 

stages, three-line, two-line, and super hybrid rice, and yield 

potential has increased continuously. China’s super hybrid rice 

breeding program is successful. The combination of the 

ideotype approach and the use of intersubspecific heterosis is 

a feasible strategy to raise rice yield potential. This is verified 

by the release of more than 10 super hybrids in recent years. 

The common characteristics of super hybrid rice are a large 

sink, large source, taller plant, good plant type, and strong root 

system with prolonged activity. Greater biomass production is 

the main reason for the higher grain yield of super hybrid rice. 

References 

Cheng SH, Zhai HQ. 2000. Comparison of some plant type components 

in super high-yielding hybrids of intersubspecies rice. 

Acta Agron. Sin. 26(6):713-718. 

Hua ZT, Wang YR, Wang Y, Dai GJ, Cai W, Zhang ZX, Hao XB, Su 

YA, Li QY. 2003. Breeding and application of japonica super 

hybrid rice Liaoyou 1518. Liaoning Agric. Sci. 2003(5):51- 

52. 

Huang YX. 2001. Rice ideotype breeding of Guangdong Academy 

of Agricultural Sciences in retrospect. Guangdong Agric. Sci. 

3:2-6. 

Lu CG, Zou JS. 2003. Comparative analysis on plant type of two 

super hybrid rice and Shanyou 63. Sci. Agric. Sin. 36(6):633- 

639. 

Min SK, Cheng SH, Zhu DF. 2002. China’s super rice breeding and 

demonstration in rice production fields: an overview. China 

Rice 2:5-7. 

Peng S, Laza RC, Visperas RM, Khush GS, Virk P. 2004. Rice: 

progress in breaking the yield ceiling. Proceedings of the 4th 

International Crop Science Congress, 26 Sept-1 Oct 2004, 

Brisbane, Australia. Published on CD-ROM. Web site 

www.regional.org.au/au/cs. 

Yang SR, Zhang LB, Chen WF, Xu ZJ, Wang JM. 1996. Theories 

and methods of rice breeding for maximum yield. Acta Agron. 

Sin. 22(3):295-304. 

Yuan LP. 1997. Hybrid rice breeding for super high yield. Hybrid 

Rice 12(6):1-6. 

Yuan LP. 2001. Breeding of super hybrid rice. In: Peng S, Hardy B, 

editors. Rice research for food security and poverty alleviation. 

Los Baños (Philippines): International Rice Research 


Zou KD, Wang XD, Li SG, Li P, Li HY, Huang GT, Liu TQ, Shen 

MS. 1997. The study on heavy panicle type of intersubspecific 

hybrid rice (Oryza sativa L.). Sci. Agric. Sin. 30(5):91-93. 

Notes 

Authors’ addresses: Xuhua Zhong, Feng Wang, and Nongrong 

Huang, The Rice Research Institute, Guangdong Academy of 

Agricultural Sciences, Guangzhou 510640, China, e-mail: 

xzhong1@pub.guangzhou.gd.cn; Shaobing Peng, Crop, Soil, 

and Water Sciences Division, International Rice Research Institute 

(IRRI), DAPO Box 7777, Metro Manila, Philippines, 

e-mail: s.peng@cgiar.org. 

Breeding and prevalence of japonica hybrid rice 

variety Mitsuhikari 

Atsushi Nakamura 

In most rice-growing countries in Asia, achieving self-sufficiency 

in rice production and maintaining price stability are 

important. Since the 1980s, many Asian countries, and especially 

China, have developed breeding of hybrid rice to increase 

yield. 

Since 1986, Mitsui Chemicals, Inc. (MCI) has developed 

a breeding program for a japonica hybrid rice variety 

acceptable to Japanese markets. To breed a new variety possessing 

both high yield and good eating quality, we aimed at 

traits such as plant type (erect leaf), ear type (panicle weight), 

strength against lodging (stiffness), and grain quality (amy- 

lose content, etc.). In 2000, MCI registered the first commercial 

japonica hybrid rice varieties, Mitsuhikari 2003 and 

Mitsuhikari 2005, which were bred by combining a conventional 

breeding method and modern biotechnology in Japan. 

MCI started selling these seeds in Japan in 2002. 

Mitsuhikari 2003 and 2005 possess high yield and good 

eating quality. To gain wide acceptance in Japan, it is necessary 

to detect suitable districts to make full use of these varieties. 

On the other hand, Mitsuhikari varieties have some undesirable 

traits such as long culms and a long growing period. To 

improve these traits, we have continuously developed a hy- 


Table 1. Yield performance and grain quality of Mitsuhikari varieties and MH 3001. a 

Culm Panicle Brown rice Amylose Protein 

Variety Yield Heading length length Panicles content content Eating 

(t ha –1 ) (d) (cm) (cm) m –2 100-grain Length Width (%) (%) quality b 

wt (g) (mm) (mm) 

Mitsuhikari 

2003 7.1 140 91 26.3 283 2.11 5.5 3.0 19.5 8.2 70 

2005 6.7 138 90 26.2 262 2.06 5.6 3.0 18.8 7.4 77 

MH 3001 7.1 126 82 26.0 305 2.25 5.7 3.0 19.2 7.5 74 

Check 5.5 120 91 17.7 400 2.14 5.2 3.0 19.0 7.9 75 

(Koshihikari) 

a Values mean average of 2001-02. Input dose of fertilizer was 100 kg N ha -1 . b Eating quality means the analyzed value of the grain quality inspector 

(Satake RCTA11A). 

brid rice breeding program. The fruit of our efforts, a new 

japonica hybrid rice line, MH 3001, appeared in 2002. 

This study presents the productivity and prevalence of 

both Mitsuhikari varieties and MH 3001. 


Mitsuhikari varieties 

Mitsuhikari 2003 and 2005 were selected after the evaluation 

of 800 different F 1 s crossed between Japanese varieties as females 

and japonica-like restorer lines from China as males. 

They have erect leaves, panicle weight type, long growing 

periods, long culms, and strength to resist lodging. For the 

easy production of F 1 seeds, the female lines were cytoplasmic 

male sterile (CMS) lines produced through asymmetric 

fusion (Akagi et al 1989). We have also developed DNA markers 

such as PCR, RAPD, and microsatellites that can identify 

CMS lines, restorer lines, and F 1 s (Akagi et al 1995, Ichikawa 

et al 1997). These markers are powerful tools for breeding 

parental lines and certifying F 1 seed purity. 

MH 3001 

Since 1993, we have developed some dwarf restorer lines. From 

1998 to 2002, performance tests of F 1 s produced between 

japonica CMS lines and these restorer lines were practiced. 

One combination, whose productivity and eating quality were 

the same as those of Mitsuhikari 2005, but whose growing 

period and plant height were shorter than those of Mitsuhikari 

2005, was selected. The F 1 of this combination was named 

MH 3001. 

Yield performance and grain quality 

We assessed the productivity and grain quality of Mitsuhikari 

varieties and MH 3001, as compared with those of Koshihikari 

(check variety) at MCI’s research paddy field in Ibaraki Prefecture 

from 2001 to 2002. Their productivity was expressed 

by brown rice yields after screening with a net of 1.75-mm 

mesh. Also, agronomic traits such as culm length, panicle 

length, panicles m –2 , and both the weight and size of brown 

rice were measured. We analyzed grain quality such as amylose 

content, protein content, and eating quality with a rice 

quality inspector (Satake RCTA11A). 

Performance test of Mitsuhikari 

Since 1995, performance tests of Mitsuhikari varieties have 

been carried out at the prefectural agricultural research center 

in seven districts, Kanto, Tokai, Kinki, Chugoku, Shikoku, 

Kyusyu, and Okinawa. Cropping practices followed those of 

each district. The input doses of fertilizer were 70–120 kg N 

ha –1 . Productivity was expressed by the yield of brown rice 

after screening. 


Table 1 shows the yield performance and grain quality of 

Mitsuhikari 2003 and 2005 and MH 3001. They yielded nearly 

30% more than Koshihikari. Their growing period to heading 

date was 30–40 days longer than that of Koshihikari. The 

panicle length of these hybrids was also about 10 cm longer 

than that of Koshihikari, and their period from heading to 

maturation took more than twice as long as that of Koshihikari. 

The culm length of Mitsuhikari varieties and Koshihikari was 

about 90 cm, and that of MH 3001 was about 80 cm. This trait 

of MH 3001 will be an advantage for tolerance of lodging. 

The grain weight of MH 3001 was the heaviest among 

the four varieties. The length of brown rice of all the hybrids 

was longer than that of Koshihikari. For grain quality, both the 

amylose and protein contents of these hybrids were much the 

same as those of Koshihikari. Also, eating quality of the hybrids 

was similar to that of Koshihikari. 

Table 2 shows the results of the performance test. In all 

of the districts, the productivity of Mitsuhikari 2003 was higher 

than that of Mitsuhikari 2005. Mitsuhikari varieties cropped 

in the southern districts of Japan have outyielded the checks 

by more than 20%. This result means that these southern districts 

are suitable for growing Mitsuhikari varieties to exploit 

their real productivity. We estimated that Mitsuhikari 2003 and 

2005 were cropped on about 1,000 ha last year, mostly in southern 

districts. On the other hand, the new hybrid line, MH 3001, 

is earlier in heading than Mitsuhikari varieties. We expect that 


Table 2. Productivity (yield in t ha -1 ) of Mitsuhikari varieties in different districts. a 

Variety 

District 

Kanto Tokai Kinki Chugoku Shikoku Kyusyu Okinawa 

Mitsuhikari 

2003 6.9 (117) 7.3 (124) 7.4 (123) 7.5 (132) 7.2 (126) 6.9 (128) 4.3 (130) 

2005 6.7 (114) 6.6 (112) 7.3 (122) 6.5 (114) 6.1 (107) 6.9 (128) 4.1 (124) 

Check 5.9 (100) 5.9 (100) 6.0 (100) 5.7 (100) 5.7 (100) 5.4 (100) 3.3 (100) 

a Values mean average of each district (1995-2003). Input doses of fertilizer were 70-120 kg N ha -1 . Numbers in parentheses 

mean increase rate vis-à-vis the check. 

this line will be accepted where Mitsuhikari 2003 and 2005 

have not prevailed. MH 3001 is now under registration. 

References 

Akagi H, Sakamoto M, Negishi T, Fujimura T. 1989. Construction 

of rice cybrid plants. Mol. Gen. Genet. 215:501-506. 

Akagi H, Nakamura A, Sawada R, Oka M, Fujimura T. 1995. Genetic 

diagnosis of cytoplasmic male sterile cybrid plants of 

rice. Theor. Appl. Genet. 90:948-951. 

Ichikawa N, Kishimoto N, Inagaki A, Nakamura A, Koshino Y, 

Yokozeki Y, Oka M, Samoto S, Akagi H, Higo H, Shinjyo C, 

Fujimura T, Shimada H. 1997. A rapid PCR-aided selection 

of a rice line containing the Rf-1 gene which is involved in 

restoration of the cytoplasmic male sterility. Mol. Breed. 

3:195-202. 

Notes 

Author’s address: Mitsui Chemicals Incorporation, Agrochemicals 

Division, Bio-product Department, 2893, Kaminoshima, 

Azumamachi, Inashiki-gun, Ibaraki-ken 300-0732, Japan, e- 

mail: atsushi.nakamura@mitsui-chem.co.jp. 

Dry-matter production and nitrogen distribution 

in a female-sterile line of rice 

Morio Kato, Sachio Maruyama, and Masao Yokoo 

The functional relations between the source capacity of leaves 

and the sink capacity of panicles affect dry-matter production 

and determine rice yield. The source-sink relation has so far 

been studied on rice plants from which their panicles were 

artificially removed. In rice, however, the role of panicle photosynthesis 

is not neglected and the role of hulls as a nutrient 

pool is important. We planned to analyze the source-sink relation 

and the accumulation of dry matter in intact plants. Under 

the assumption that the sink capacity of panicles is negligible, 

we examined the source-sink relation in a female-sterile line 

of rice (FS1), and characterized the effects of losing the sink 

function of panicles on dry-matter production and nitrogen 

distribution among plant organs. 


A female-sterile line (FS1) and a normal fertile line (Fujisaka 

5) of rice (Oryza sativa L.) were used. FS1 is a progeny from 

the cross between Japanese variety Fujisaka 5 and Indonesian 

variety Tjina four times backcrossed by Fujisaka 5 as a recur- 

rent parent. FS1 resembles Fujisaka 5 for most plant types, but 

has panicles with few ripened grains because of incomplete 

embryo sac formation (Yokoo 1984). 

Germinated seeds were sown on 21 April in a plastic 

box containing rice nursery soil. Two seedlings at the 6-leaf 

stage were transplanted to a 1/5,000 area of a Wagner pot. 

Plants were placed under natural conditions in a vinyl house 

from April to September 2000 at the Agricultural and Forestry 

Research Center, University of Tsukuba. 

At 10 days before heading and during the period from 

heading to maturity (42 d after heading), 6 plants were harvested 

every 7 d and divided into 6 parts: roots, culms plus 

leaf sheaths (stems), leaf blades, dead leaves, panicles, and 

late tillers that appeared after heading. All the samples were 

oven-dried at 80 °C for 72 h and their dry weights were determined. 

The dried plant materials were ground in a mill, and 

their carbon and nitrogen contents were determined by a C-N 

Corder (MT600, Yanaco, Japan). The amounts of carbon and 

nitrogen of each organ were determined by multiplying dry 

weights by each content. 


Dry weight (g plant –1 ) 

80 

70 

Root 

Culm + leaf sheath 

Leaf blade 

Dead leaf 

Panicle 

Late tiller 

FS1 

Fujisaka 5 

60 

50 

40 

30 

20 

10 

0 

–10 0 7 14 21 28 35 42 

Days after heading 

–10 0 7 14 21 28 35 42 


Fig. 1. Changes in dry weights of each organ in FS1 and Fujisaka 5 during the ripening stage. 

Results 

Growth characteristics 

FS1 and Fujisaka 5 headed 88 and 89 d after sowing, respectively. 

Culm and panicle lengths of FS1 were much longer than 

those of Fujisaka 5. The number of panicles was almost equal 

in both lines. In FS1, late tillers developed rapidly from basal 

and upper lateral buds in the last half of the ripening stage. 

Changes in dry-matter production 

and distribution among organs 

Total dry weights were 48.5 and 40.4 g plant –1 at heading time 

and 75.3 and 58.8 g plant -1 at maturity in FS1 and Fujisaka 5, 

respectively. The distribution of dry weight to each organ differed 

during the period from heading to maturity (Fig. 1). The 

dry weight of panicles increased rapidly 10 to 20 d after heading 

in Fujisaka 5, but remained unchanged in FS1 during maturity. 

The dry-matter partitioning ratios of panicles at maturity 

were 9% and 40% in FS1 and Fujisaka 5, respectively. 

Though the dry weight of leaf blades decreased after heading 

in both lines, the decrease in FS1 was slower than that in 

Fujisaka 5. The dry weight of stems in Fujisaka 5 decreased at 

14 d after heading with the rapid increase in panicle weight, 

and remained steady during the late ripening stage. On the other 

hand, the stem dry weight in FS1 continued to increase until 

28 d after heading and then decreased rapidly with the 

appearence of late tillers. Stems occupied more than half of 

the total dry matter in both lines at heading. During the ripening 

stage, the dry-matter partitioning ratio of stems decreased 

to 34% at maturity with the increase in panicle weight in 

Fujisaka 5, but that in FS1 increased to 62% until 4 wk after 

heading and still remained at 43% at maturity. The dry weight 

of late tillers increased after 28 d of heading and occupied 

24% of the total dry weight at maturity in FS1. Root dry weight 

of Fujisaka 5 decreased during the ripening stage, but that of 

FS1 remained steady or increased slightly. 

Changes in carbon and nitrogen contents 

of each organ 

Average carbon contents in a whole plant of both lines were 

42–43% during the ripening stage. While there were no significant 

differences in carbon contents of each organ between 

FS1 and Fujisaka 5 at heading, the differences in leaf blades, 

stems, and panicles were significant at maturity. However, the 

changes in carbon accumulation of each organ showed a trend 

similar to that of the dry weights in both lines because the 

difference in the content between lines and the changes during 

ripening were small. 

Average nitrogen contents in a whole plant were 1.0% 

and 1.2% at heading and those decreased to 0.6% and 0.9% at 

maturity in FS1 and Fujisaka 5, respectively. The nitrogen contents 

in leaf blades of both lines were 2.5–2.7% at heading, 

and decreased relatively slower in FS1 than in Fujisaka 5, so 

those at maturity were 1.8% and 1.4% in FS1 and Fujisaka 5, 

respectively. Though there were no significant differences in 

the nitrogen content of each organ between FS1 and Fujisaka 

5 at heading, the differences were significant in every organ at 

maturity. Particularly, nitrogen contents of panicles and stems 

were significantly lower in FS1 than in Fujisaka 5. 

Changes in nitrogen accumulation 

and distribution of each organ 

Total nitrogen accumulation in a whole plant did not increase 

during the ripening stage and the total amounts per plant were 

439 mg in FS1 and 504 mg in Fujisaka 5 at maturity (Fig. 2). 

After heading, the amount of nitrogen in leaf blades decreased 

rapidly and that in panicles increased, which occupied above 

a half of the whole-plant nitrogen 21 d after heading in Fujisaka 


Nitrogen (g plant –1 ) 

0.6 

0.5 

Root 

Culm + leaf sheath 

Leaf blade 

Dead leaf 

Panicle 

Late tiller 

FS1 

Fujisaka 5 

0.4 

0.3 

0.2 

0.1 

0.0 

–10 0 7 14 21 28 35 42 –10 0 7 14 21 28 35 42 



Fig. 2. Changes in nitrogen accumulation of each organ in FS1 and Fujisaka 5 during the ripening stage. 

5. The partitioning ratio to stems also declined rapidly during 

the period from 7 to 21 d after heading. On the contrary, in 

FS1, the decline in nitrogen partitioning ratio to leaf blades 

was slower, and the partitioning ratio to panicles was almost 

steady. The nitrogen partitioning ratio to stems increased gradually 

until 21 d after heading, and then decreased rapidly to 

21% at maturity with the apperance of late tillers. The partitioning 

ratio to roots decreased with ripening in Fujisaka 5, 

but that of FS1 remained comparatively higher. While panicles 

accumulated 55% of the total nitrogen in Fujisaka 5 at maturity, 

late tillers were the largest sink organ for nitrogen in FS1, 

occupying 32% of the total nitrogen of a whole plant. 

Conclusions 

FS1 did not decrease dry-matter production irrespective of the 

loss of sink function of panicles. The slower decrease in nitrogen 

concentration in leaf blades caused retarded leaf senescence 

and the maintenance of relatively higher photosynthetic 

activity in FS1 (Kato et al 2004). Its vegetative organs such as 

culms, leaf sheaths, late tillers, and roots substituted for panicles 

as photosynthate and nitrogen sinks. The results of the changes 

in dry-matter and nitrogen distribution indicated that, in the 

final ripening stage, the carbohydrate and nitrogen accumulated 

in stems rapidly translocated to late tillers. This unique 

pattern of dry-matter and nitrogen distribution retarded the 

senescence of leaves and the decline in leaf photosynthesis, 

and then contributed to the increase in dry-matter production 

during the ripening stage. FS1 is considered to have larger 

amounts of carbon and nitrogen in stems than Fujisaka 5. This 

suggests the possibility of using FS1 for biological production 

as a forage crop. However, the sink capacity for nitrogen in 

panicles was so strong that vegetative organs such as stems 

and late tillers could not fulfill it. The difference between the 

dry-weight and nitrogen distribution patterns in organs caused 

the change in carbon-nitrogen balance during the ripening stage. 

The present results on dry-matter production and nitrogen accumulation 

were obtained from the limited conditions for root 

growth and nitrogen nutrition in pots. Further field experiments 

under higher nitrogen nutrition levels are needed to evaluate 

the potential biomass productivity and nitrogen accumulation 

of this female-sterile line of rice. 

References 

Kato M, Kobayashi K, Ogiso E, Yokoo M. 2004. Photosynthesis 

and dry-matter production during ripening stage in a femalesterile 

line of rice. Plant Prod. Sci. 7:184-188. 

Yokoo M. 1984. Female sterility in an Indica-Japonica cross of rice. 

Jpn. J. Breed. 34:219-227. 

Notes 

Authors’ address: Graduate School of Life and Environmental Sciences, 

University of Tsukuba, e-mail: 

katomo@sakura.cc.tsukuba.ac.jp. 


Cytokinin as a causal factor of varietal differences 

in the reduction in leaf level of ribulose-1,5-bisphosphate 

carboxylase/oxygenase during senescence in rice plants 

Taiichiro Ookawa, Yukiko Naruoka, Ayumi Sayama, and Tadashi Hirasawa 

Dry matter production and yield are higher in rice cultivar 

Akenohoshi than in cultivar Nipponbare, primarily because of 

a smaller decrease in the rate of photosynthesis during the ripening 

stage of Akenohoshi (Jiang et al 1988). The rate of 

photosynthesis and nitrogen content during senescence are both 

closely correlated with leaf levels of ribulose-1,5-bisphosphate 

carboxylase/oxygenase (Rubisco). During leaf senescence, 

Akenohoshi shows a smaller decrease in levels of Rubisco than 

Nipponbare, and we have previously demonstrated that 

Akenohoshi maintains larger amounts of nitrogen in leaves 

during ripening. This may account for the differences in the 

ability of these two cultivars to maintain high leaf levels of 

Rubisco (Ookawa et al 2003). The high leaf nitrogen content 

in Akenohoshi results from both greater total accumulation of 

nitrogen and greater partitioning of nitrogen to leaves. 

It was also observed that larger amounts of cytokinin 

were transported from the roots to the aboveground parts of 

the plant during the ripening stage by Akenohoshi and 

Nipponbare (Soejima et al 1995). Cytokinin can delay senescence 

and recent studies have illustrated the importance of 

cytokinin in the control of senescence (Gan and Amasino 1995). 

Cytokinins might also contribute to the maintenance of high 

Rubisco content. However, it remains to be determined whether 

cytokinin suppresses a decline in Rubisco content. 

Several studies have reported that cytokinin can induce 

the expression of photosynthetic genes and promote protein 

synthesis. It can be assumed that suppressing the decrease in 

accumulation of Rubisco gene transcripts by cytokinin results 

in the maintenance of high Rubisco levels during leaf senescence. 

Cytokinin also affects nitrogen partitioning in the whole 

plant (Jordi et al 2000). High nitrogen partitioning to leaves 

by cytokinin probably results in the maintenance of high 

Rubisco levels during senescence. 

In the following experiments, our main aims were therefore 

(1) to compare the levels of Rubisco, nitrogen, and rbcL 

and rbcS transcripts between Nipponbare and Akenohoshi, (2) 

to analyze the effects of exogenous cytokinin on levels of 

Rubisco, nitrogen, and rbcL and rbcS transcripts in leaves of 

rice during the ripening stage, and (3) to analyze the effects of 

exogenous cytokinin on the nitrogen content of leaves by determining 

nitrogen absorption and partitioning to various organs 

in the whole plant. 

Methods 

Rice seedlings (Oryza sativa L., cvs. Nipponbare and 

Akenohoshi) were transplanted to Wagner pots (1/2,000 a) 

filled with soil on 27 May. In this study, Nipponbare plants 

were treated with additional nitrogen fertilizer and cytokinin. 

Ammonium sulfate was applied to some pots on 1 September 

at a dose of 10 g per pot as additional nitrogen fertilizer (NF). 

For the treatment with cytokinin, 30 mL of a 10 -4 M solution 

of 6-benzylaminopurine (BA), containing 0.05% Tween 20 as 

a surfactant, was sprayed on the aboveground parts of each 

hill at 2-day intervals from 5 September. Levels of Rubisco 

and nitrogen were determined in the same flag leaf of the main 

stem. Levels of Rubisco were determined by the single radial 

immunodiffusion method. Nitrogen was quantitated using a 

CN analyzer (MT-600, Yanaco Inc., Kyoto, Japan). For quantification 

of mRNA, the flag leaves on the main culms were 

collected between 1000 and 1100 on a clear day and were immediately 

frozen in liquid nitrogen. Levels of rbcL and rbcS 

mRNA were determined by Northern blotting analysis. 

Comparison of Rubisco, nitrogen, and rbcL 

and rbcS mRNA levels in leaves 

Changes in Rubisco and rbcL and rbcS mRNA levels in leaves 

during ripening were compared between the two cultivars. At 

the heading stage, there were no differences between the two 

cultivars, but, after heading, levels of Rubisco and rbcL and 

rbcS mRNA remained higher in Akenohoshi than in Nipponbare 

(Fig. 1A, C, and D). 

The nitrogen content in flag leaves was higher in 

Nipponbare and Akenohoshi at the heading stage. After heading, 

the nitrogen content decreased more slowly in Akenohoshi 

than in Nipponbare (Fig. 1B). 

Effects of treatment with 6-benzylaminopurine 

and nitrogen fertilizer on Rubisco, nitrogen, and rbcL 

and rbcS mRNA levels in leaves 

In Nipponbare controls, Rubisco content in flag leaves decreased 

with time after heading. Rubisco content remained 

substantially higher in plants treated with NF or BA in 

Nipponbare than in the controls (Fig. 1A). Levels of rbcL and 

rbcS mRNA decreased with time in the controls. However, 

levels remained high in plants treated with NF or BA during 

ripening (Fig. 1C, D). 

In the controls, nitrogen content declined with time in 

flag leaves. In contrast, nitrogen content decreased but remained 

relatively high during ripening when compared with 

the controls (Fig. 1B). 


Rubisco content (g m –2 ) 

Nitrogen content (g m –2 ) 

3 

A 1.5 

c 

B 

2 

c 

c 

b 

a 

b 

b 

a 

a 

1.0 

a 

c 

b 

c 

b 

ab 

a 

1 

0 

Control (NI) 

BA (NI) 

NF (NI) 

Akenohoshi 

0.5 

0 

Relative level of rbcL mRNA (%) 

Relative level of rbcS mRNA (%) 

120 

100 

80 

60 

40 

20 

a 

b 

b 

b 

a 

a 

a 

C 

120 

100 

80 

60 

40 

20 

c 

c 

b 

a 

c 

bc 

b 

a 

D 

0 

31 Aug 10 Sep 20 Sep 30 Sep 

Date 

0 

31 Aug 10 Sep 20 Sep 30 Sep 

Fig. 1. Effects of 6-benzylaminopurine (BA) and additional nitrogen fertilizer (NF) application on the Rubisco 

content (A), nitrogen content (B), and relative levels of rbcL mRNA (C) and rbcS mRNA (D) in flag leaves of 

Nipponbare (NI). The plants of Akenohoshi were grown in the same conditions as the control for Nipponbare. 

Vertical bars represent standard deviations (n = 3). Symbols with different letters are significantly different 

at the 5% level (LSD). 

Date 

Relationships between Rubisco content, levels of rbcL 

and rbcS mRNA, and nitrogen content 

The relationships between Rubisco content, levels of rbcL and 

rbcS mRNA, and nitrogen content were compared among the 

controls, BA- and NF-treated Nipponbare plants, and 

Akenohoshi plants. There were close relationships between 

Rubisco content and levels of rbcL and rbcS mRNA, irrespective 

of treatment and cultivar (rbcL: r = 0.90, rbcS: r = 0.98). 

In previous studies, cytokinin promoted the expression of genes 

for photosynthetic proteins by increasing transcription (Suzuki 

et al 1994). Therefore, cytokinin likely also contributed to the 

maintenance of high leaf levels of Rubisco in NF- and BAtreated 

plants. 

Over a wide range of nitrogen contents, there was also a 

close positive correlation between Rubisco content and nitrogen 

content (r = 0.80). There was also a close correlation between 

nitrogen content and levels of both rbcL and rbcS mRNA 

(rbcL: r = 0.74, rbcS: r = 0.85). 

Effects of BA and NF on accumulation 

and partitioning of nitrogen 

The high leaf nitrogen content was maintained not only in NFtreated 

Nipponbare plants but also in BA-treated Nipponbare 

plants. The total amount of accumulated nitrogen during the 

late-ripening stage was much larger after NF treatment when 

compared with the controls (Table 1), and the partitioning of 

nitrogen to leaves increased significantly as a result of NF treatment. 

Therefore, the maintenance of high leaf nitrogen content 

in NF-treated plants resulted not only from an increase in 

total nitrogen accumulation by the entire plant but also from 

an increase in nitrogen partitioning to leaves. There was no 

difference in terms of the increase in nitrogen whole-plant content 

between BA-treated plants and controls from the heading 

stage to the late-ripening stage. The partitioning of nitrogen to 

leaves at the late-ripening stage was considerably higher in 

BA-treated plants than in the controls, but less nitrogen was 

partitioned to panicles in the former than in the latter (Table 

1). These results showed that maintenance of high leaf nitrogen 

content in BA-treated plants was caused by an increase in 

nitrogen partitioning to leaves and a decrease in partitioning 


Table 1. Comparisons of nitrogen content, changes in nitrogen content, and nitrogen partitioning to various 

organs in entire plants. a 

Nitrogen content (mg hill –1 ) Nitrogen partitioning (%) 

Changes in nitrogen 

Organ Heading Late ripening content (mg hill –1 ) Heading Late ripening 

(31 Aug) (25 Sep) (31 Aug) (25 Sep) 

Whole plant 

Control 592.4 ± 35.1 635.7 ± 19.3 a 43.3 100 100 

BA – 653.9 ± 15.8 a 61.5 – 100 

NF – 1,024.6 ± 6.2 b 432.2 – 100 

Leaves 

Control 292.4 ± 12.1 142.9 ± 4.9 a –149.5 49.4 ± 1.0 22.6 ± 0.3 a 

BA – 238.7 ± 3.2 b –53.7 – 36.5 ± 1.3 b 

NF – 307.2 ± 3.5 c 14.8 – 30.0 ± 0.6 c 

Culms + leaf sheaths 

Control 202.0 ± 18.2 114.1 ± 3.1 a –87.9 34.0 ± 1.0 17.9 ± 0.5 a 

BA – 115.2 ± 0.3 a –86.8 – 17.6 ± 0.4 a 

NF – 202.3 ± 12.2 b 0.3 – 19.7 ± 1.0 b 

Panicles 

Control 58.2 ± 1.6 328.7 ± 8.9 a 270.5 9.8 ± 0.4 51.5 ± 0.7 a 

BA – 259.7 ± 18.8 b 201.5 – 39.8 ± 2.0 b 

NF – 459.4 ± 3.1 c 401.2 – 44.8 ± 0.5 c 

Roots 

Control 39.8 ± 4.2 50.1 ± 7.4 a 10.3 6.8 ± 0.4 8.0 ± 1.0 a 

BA – 40.3 ± 1.6 b 0.5 – 6.2 ± 0.4 b 

NF – 55.7 ± 0.6 a 15.9 – 5.5 ± 0.0 b 

a Data represent means ± standard deviations of results from three replicates. Nitrogen partitioning is expressed as the nitrogen content 

of the indicated organs as a percentage of the total nitrogen content of the whole plant. Means followed by different letters are 

significantly different at the 5% level of probability (LSD). 

to panicles, rather than an increase in whole-plant accumulation 

of nitrogen. 

Jordi et al (2000) used a 15 N tracer to investigate patterns 

of nitrogen partitioning in transgenic tobacco plants, in 

which the gene for cytokinin-synthetic enzyme, 

isopentenyltransferase, was specifically expressed in senescent 

leaves. They found that high levels of Rubisco were maintained, 

reflecting the partitioning of a large amount of 15 N to leaves. A 

large amount of nitrogen was partitioned to leaves in 

Akenohoshi, with a high cytokinin activity in xylem sap compared 

with that in Nipponbare (Ookawa et al 2003). These 

results indicate that cytokinin has an effect on nitrogen partitioning. 

Conclusions 

Cytokinin affects both the synthesis of Rubisco in leaves and 

the partitioning of nitrogen to leaves in rice plants during ripening. 

This suggests that cytokinin causes the differences in 

Rubisco reduction during senescence, thus resulting in differences 

in the reduction of photosynthesis between Akenohoshi 

and Nipponbare. Therefore, capacity of the roots for cytokinin 

synthesis is an important determinant of the ability of rice 

plants to maintain photosynthetic rate in leaves during ripening 

in high-yielding rice cultivars. 

References 

Gan S, Amasino RM. 1995. Inhibition of leaf senescence by 

autoregulated production of cytokinin. Science 270:1986- 

1987. 

Jiang CZ, Hirasawa T, Ishihara K. 1988. Physiological and ecological 

characteristics of high yielding varieties in rice plants: leaf 

photosynthetic rates. Jpn. J. Crop Sci. 57:139-145. 

Jordi W, Schapendonk A, Davelaar E, Stoopen GM, Pot CS, De Visser 

R, Rhijn JA, Gan S, Amasino RM. 2000. Increased cytokinin 

levels in transgenic P SAG12 -IPT tobacco plants have large direct 

and indirect effects on leaf senescence, photosynthesis 

and N partitioning. Plant Cell Environ. 23:279-289. 

Ookawa T, Naruoka U, Yamazaki T, Suga J, Hirasawa T. 2003. A 

comparison of the accumulation and partitioning of nitrogen 

in plants between two rice cultivars, Akenohoshi and 

Nipponbare, at the ripening stage. Plant Prod. Sci. 6:172-178. 

Soejima H, Sugiyama T, Ishihara K. 1995. Changes in the chlorophyll 

contents of leaves and in levels of cytokinins in root 

exudates during ripening of rice cultivars Nipponbare and 

Akenohoshi. Plant Cell Physiol. 36:1105-1114. 

Suzuki I, Cretin C, Omata T, Sugiyama T. 1994. Transcriptional and 

posttranscriptional regulation of nitrogen-responding expression 

of phosphoenolpyruvate carboxylase gene in maize. Plant 

Physiol. 105:1223-1229. 

Notes 

Authors’ address: Tokyo University of Agriculture and Technology, 

e-mail: ookawa@cc.tuat.ac.jp. 


Measurement and evaluation of rice plant type 

by means of an image analysis method and a 3-D digitizer 

Masaaki Oka and Takashi Ogawa 

In the breeding of high-yielding rice varieties, plant type is 

very important. Plant type is closely related to sunlight-receiving 

efficiency. The plant type of rice has changed over time 

and many varieties with varied plant type have been developed 

recently to achieve higher yield. To advance the research 

on high-yielding rice breeding further, new methods for measuring 

plant type are indispensable. A method for evaluating 

what kind of plant type is ideal is also required. 

We developed a method for measuring rice plant type 

by means of image analysis and a 3-D digitizer. The system 

using a 3-D digitizer can also calculate the light-receiving leaf 

area of a rice plant. By using this method, we compared the 

characteristics of plant type and the light-receiving efficiency 

of some rice varieties. (All computer software in this experiment 

was written in Microsoft VisualBasic.net.) 

Calculation 

Koshihikari-ah01 Threshold value 126 

Image analysis of plant type 


Fifteen rice varieties (three popular varieties, five high-yielding 

varieties, two hybrid rice varieties bred in Japan, four new 

plant type—NPT—lines bred by IRRI, and a Korean variety) 

were cultivated in a paddy field. The side-view photographs 

of a single rice plant were taken before heading, at heading 

time, and at the ripening stage. After changing these image 

data to a binary image (Fig. 1, upper right), the two-dimensional 

distribution of aerial part density was calculated (lower 

left). To compare the characteristics of plant type, the following 

numerical factors were obtained (lower right): 

Results 

Factor 1: Plant height 

Factor 2: Two-dimensional area of the plant on a binary 

image 

Factor 3: (Height at which the plant area is divided into 

two equally)/(Plant height) 

Factor 4: (Horizontal distance between the center axis 

and the portion in which 90% of the plant area is contained)/(Plant 

height) 

At the ripening stage, most NPT lines (IR65600-127-6-2-3, 

IR71218-5-2-1, etc.) showed lower plant height and higher 

values of Factors 3 and 4. New high-yielding varieties bred in 

Japan (Fukuhibiki, Habataki, etc.) showed a similar tendency. 

In contrast, high-yielding hybrid rice bred in Japan (Mitsuhikari 

2003 and 2005) with a longer culm showed a higher value of 

Factor 3 and lower value of Factor 4. Both Factor 3 and 4 

Factor 1 124 

Factor 2 2108.617 

Factor 3 0.5 

Factor 4 0.1290323 

Fig. 1. Image analysis of the plant type of Koshihikari at the ripening 

stage. Binary image (upper right), the distribution of plant density 

(lower left), and calculated factors 1–4 (lower right). 

values of popular varieties in Japan (Hitomebore, etc.) were 

lower. Milyang 23, a Korean variety, bred from a cross between 

japonica and indica rice, showed a value near the middle 

of the new high-yielding varieties and popular varieties (Oka 

and Ogawa 2004). 

Measuring plant type by using a 3-D digitizer 


We used an electromagnetic 3-D digitizer to measure a detailed 

solid structure of rice plant type. Using a 3-D digitizer, 

three-dimensional coordinates of points on the edge of all 

leaves of the rice plant (nine points per leaf) were obtained. 

One plant had about 100–150 leaves and it took 1.0 to 1.5 

hours to measure all the leaves of one plant. After a single leaf 

was divided into 13 triangles by using the coordinate data, the 

area, the angle to the level surface, and the coordinate of the 

center of gravity of each triangle were calculated. Total leaf 

area and averaged leaf angle were calculated in each space 

divided at intervals of 10 cm horizontally and vertically, and 

the spatial distribution of leaf area and leaf angle was drawn 


Calculation 

Total leaf area (a): 2,570 cm 2 

Leaf area receiving direct sunlight (b) 475 cm 2 

Percentage of (b) to (a): 18% 

South 

Fig. 2. The calculation of leaf area receiving direct sunlight of Hitomebore about 20 days before 

heading. The canopy consists of six plants and the target plant is in the center of the back row. The 

light-receiving triangles and the shaded triangles are drawn by thick solid lines and thin dotted lines, 

respectively. 

by computer graphics. The characteristics of leaf distribution 

of three rice varieties (Hitomebore, Milyang 23, and 

Mitsuhikari 2005) were compared by this method. 

Results 

The differences of leaf distribution among three varieties were 

not clear before heading, but, as the developmental stage advanced, 

the characteristics of leaf distribution became more 

prominent. At the ripening stage, Hitomebore had many slanting 

leaves that protruded outward in the middle height. In 

Milyang 23, leaves were more erect at the upper height than in 

Hitomebore. Mitsuhikari 2005 was taller than the others, and 

its leaves stood straight at about every height and were densely 

distributed near the vertical center line of the plant (Oka et al 

2003). 

Calculation of light-receiving efficiency 

It is thought that the rice plant with ideal plant type has a larger 

area of leaves that receive direct sunlight and shows a higher 

percentage of sunny leaf area to total leaf area. To evaluate the 

light-receiving efficiency of the rice plant, we made computer 

software to calculate the area of light-receiving leaves when 

the sun rose at any angle. 

Light-receiving leaf area of an isolated single plant 

The three-dimensional structure of a rice plant was rotated in 

the direction that was the viewpoint of the sun, and we inves- 

tigated whether each triangle into which a leaf was divided 

received direct sunlight or not. When the center of gravity of 

one triangle was in the inside of the other triangle, we judged 

that they overlapped each other. In addition, the triangle near 

the sun was recorded as sunny and the other was recorded as 

shade. After this calculation was performed for the combination 

of all triangles, the percentage of leaf area receiving direct 

sunlight was determined. 

Light-receiving leaf area of the rice canopy 

In our experimental paddy field, rice plants were planted at 24 

by 24 cm. The three-dimensional coordinates of the leaves of 

five plants that were placed at 24-cm intervals (east, southeast, 

south, southwest, and west of the target plant) were calculated. 

By examining the mutual shading among all triangles 

of leaves of these five plants, the light-receiving efficiency of 

the target plant was evaluated by the same method as for an 

isolated single plant. 

The light-receiving efficiencies of three rice varieties 

were compared. The result for Hitomebore before heading is 

shown in Figure 2. When the angle of elevation of the sun was 

70 degrees and the azimuth angle was 15 degrees from the 

south to the east, the leaf area receiving direct sunlight was 

475 cm 2 and the percentage to the total leaf area was 18%. At 

the same angle of the sun, the percentages of Milyang 23 and 

Mitsuhikari 2005 were 19% and 9%, respectively. 


Conclusions 

By using the image analysis method, the rough characteristics 

of plant type of many rice varieties can be recognized in a 

short time. As the calculation speed of the computer became 

faster, it became possible to do image analysis easily using a 

personal computer. We are making the database of a rice profile 

image now. By combining this database and the image 

analysis method in this experiment, we will be able to develop 

an epochal analysis system for the rice plant type. 

The detailed solid structure of the rice plant can be measured 

by a 3-D digitizer. In addition, the area of light-receiving 

leaves can be calculated. The light-receiving efficiency of 

a rice plant can be evaluated by performing calculations of 

sunny leaf area of a plant irradiated from various directions. 

This system can also calculate the light-receiving efficiency 

of a virtual rice plant type, for example, a plant whose leaf 

angle, leaf shape, or internode length of an actual plant was 

changed. The ideal rice plant type will be found by repeating 


this simulation experiment. The target of our research is an 

automatic evaluation system on rice plant type by using new 

measurement techniques, information technology, and simulation 

experiments. 

References 

Oka M, Yamauchi M, Imaizumi Y. 2003. Measurement of spatial 

leaf distribution of three rice varieties by means of 3-D digitizer. 

Breed. Sci. 5(supple. 1):204. (In Japanese.) 

Oka M, Ogawa T. 2004. Characteristics of plant type of some new 

high-yielding rice varieties measured by computer image 

analysis. Breed. Sci. 6(supple. 1):83. (In Japanese.) 

Notes 

Authors’ address: Miyagi University of Education, e- 

mail:maoka@staff.miyakyo-u.ac.jp. 

Improving rice yield potential is a persistent research target for 

food security. Its significance is magnifying with the increasing 

need for improving the efficiency in the use of land, water, and 

other natural resources, and labor for rice culture. With the 

completion of the rice genome sequence, more tools are becoming 

available for understanding the genetic control of plant 

functions. Considering this background, this session had two major 

aims: 

1. To elucidate the key physiological and morphological 

traits to be targeted to improve yield potential by summarizing 

research carried out after the Green Revolution. 

2. To characterize recently developed high-yielding genetic 

resources, including hybrid rice, and to evaluate the 

possible contribution of molecular breeding to the improvement 

of rice yield potential. 

Six papers were presented on physiological processes that 

govern rice yield potential and five papers on breeding for increased 

rice yield potential. The following key points and future 

prospects can be extracted from the papers presented: 

1. A balanced increase in source and sink capacity is the 

key to enhancing rice yield potential since it is obvious 

that source and sink capacity interact in determining 

yield (P.K. Mahapatra, S. Peng). Understanding and 

quantifying sink strength are a major challenge in explaining 

the differences among genotypes in 

remobilization efficiency, grain filling, and harvest index. 

High remobilization ability was shown to improve yield, 

harvest index, and N-use efficiency in new japonica variety 

Akita-63, with large grain size (T. Mae). It is expected 

that recent progress in understanding carbon 

metabolism related to grain filling at the molecular level 

will help to manipulate sink strength (R. Ohsugi). 

2. It is clear that enhancing biomass production without 

reducing harvest index would be an effective way to improve 

rice yield potential. The increase in biomass production 

will largely depend on improving radiation-use 

efficiency and canopy photosynthetic rate. Improving 

Rubisco efficiency would have great importance as a 

prospect for enhancing N-use efficiency (A. Makino). 

The introduction of C 4 photosynthesis is a challenging 

option (J. Sheehy). The results of experiments under 

elevated CO 2 conditions suggested that increased 

canopy photosynthesis capacity increased biomass and 

yield, but to a lesser extent than single-leaf photosynthesis 

(T. Hasegawa). The relationship among leaf- and 

canopy-level gas exchange and crop growth needs to 

be studied. During the entire crop-growing season, enhancement 

of biomass production during the ripening 

phase would be most effective in improving rice yield 

potential, followed by high biomass production during 

the reproductive growth stage. Significant genotypic 

variation has been found in biomass production at these 

stages. Further understanding is needed on the mechanisms 

through which sink strength affects source-related 

ability, for example, photosynthesis and stomatal 

conductivity, and biomass production after the reproductive 

stage. 

3. The interaction of yield determination with environment 

needs further attention. The effectiveness of traits should 

be evaluated in adaptability to different environments. 

For example, breeding for large panicle size with more 


spikelets per panicle has proven to be effective in increasing 

rice yield potential in temperate and subtropical 

regions. However, the following questions remain 

unanswered: (1) Does a panicle-weight-type rice variety 

perform well in the tropics (2) Can a panicle-weighttype 

rice variety achieve good grain filling without using 

heterosis (S. Peng). High radiation in the dry season 

under tropical conditions causes photoinhibition and 

lower radiation-use efficiency (E. Murchie). Adaptation 

to tropical/temperate conditions or dry-/wet-season conditions 

must be considered. Exhaustive results from QTL 

analyses for yield traits indicated a general tendency 

that each of the QTLs for yield traits made a rather small 

contribution, with a large interaction of environmental 

and genetic factors (Z. Li). A statistical tool for QTL × 

environment analysis should be useful for figuring out 

the adaptability of traits and solving complexity in markerassisted 

selection for yield under various conditions. 

4. The secondary improvement of new plant type (NPT) 

tropical japonica by hybridizing indica and tropical 

japonica types steadily showed a higher yield level than 

indica check cultivars but a slightly lower yield than hybrids. 

Further potential was suggested by the concurrent 

improvement of several traits of the NPT. 

5. There is no question that hybrid rice technology has 

contributed to the increase in rice yield potential. Recent 

progress in China for hybrid rice breeding was reviewed 

(Zhong et al). Several new hybrids, called “super” 

hybrids, released starting in the late 1990s attained 

a yield increase of 10–15% over that of previous 

hybrids, indicating further potential for genetic improvement. 

A commercial hybrid in Japan, Mituhikari, achieved 

an acceptable level of eating quality in cooked rice as 

well as a yield increase of 20–30%. This attempt indicated 

a way toward high-productivity rice farming with 

an assured marketability in the gray area of the rice 

industry in industrialized countries (A. Nakamura). 

6. Further improvement in rice yield potential will continue 

to rely on conventional cross-breeding approaches. 

Ideotype breeding, hybrid rice breeding, and the modern 

biotechnology approach are powerful supplements 

to the empirical breeding approach in crop improvement 

for yield potential. An efficient collaboration mechanism 

between physiology and breeding is essential in these 

efforts. 


SESSION 5 

Broadening the gene pool 

and exploiting heterosis in cultivated rice 

CONVENER: Y. Fukuta (JIRCAS/IRRI) 

CO-CONVENERS: D. Mackill (IRRI) and R. Ikeda (JIRCAS)

Developing aerobic rice in Brazil 

B. da S. Pinheiro, E. da M. de Castro, O.P. de Moraes, and F. Breseghello 

In the mid-1980s, upland rice varieties of tropical japonica 

extraction attained an area bigger than 4.5 million ha under 

Brazilian savannas. Afterward, crop area decreased gradually 

and markedly, attaining only 1.8 million ha in 2003. However, 

production has not declined at the same level because average 

yield doubled in the period 1986-2003, rising from 1.1 to 1.9 

t ha –1 . This was due to both crop migration toward more favored 

areas in rain distribution, and the adoption of modernplant-type 

varieties, tropical japonica × indica derivatives. The 

impact of improved-plant-type varieties and associated technologies 

in the favored localities of the savanna region put 

forward a new concept of upland rice, leading to a new denomination—aerobic 

rice. The shift in breeding strategies over 

time and the genetic gain attained in the latter part of the 1990s 

are presented and discussed in this paper. 

The National Upland Rice Breeding Program 

Since its inception in 1974, the former National Research Center 

for Rice and Beans (CNPAF) of Embrapa, located in Goiás, 

has been developing and coordinating the National Upland Rice 

Breeding Program. The objectives and priorities of the program 

have changed from time to time to cope with the alterations 

in geographic distribution of the crop, cropping system, 

and consumers’ taste for the product. According to Pinheiro 

(2003), it can be divided into three distinct periods: (1) phase 

1, 1975-85, higher emphasis on drought tolerance, blast resistance, 

and yield stability, targeting exclusively the unfavored 

savanna conditions; (2) phase 2, 1985-90, breeding strategy 

expanded to include selection for high yield potential, targeting 

favorable savanna conditions as well; (3) phase 3, 1990 to 

the present, higher emphasis on blast resistance, grain appearance, 

and yield potential, targeting mainly favored conditions. 

Drought was a strong priority during phase 1 of the breeding 

program, which relied on progenitors of the tropical 

japonica group, from both national and African origin, but will 

not be covered in this paper. After the studies of Steinmetz et 

al (1988a,b), the breeding strategy was expanded to obtain 

genotypes to be grown under supplementary irrigation and in 

favored microregions for rainfall distribution during phase 2. 

Progenitors of the indica group were involved in crosses, segregating 

generations were evaluated under supplementary irrigation, 

and advanced lines tested exclusively in favorable 

locations. A noticeable move from the frontier land to more 

favored areas for water distribution led to a decrease in the 

priority for drought tolerance in phase 3. Currently, drought 

evaluation results do not restrict the release of aerobic rice 

varieties. 

Even with alterations in the program objectives and plant 

ideotype, some basic strategies have been maintained for a 

long time. To create plant populations to be used in the subsequent 

steps of the program, recurrent selection schemes are 

widely used. Progenitor combination capacity is evaluated to 

obtain more successful crosses. Pure lines are developed 

through population advances by mass selection within families. 

Some characteristics, such as plant type, grain appearance, 

growth cycle, and resistance to major diseases, are selected 

for starting from early generations. Yield trials are carried 

out at three levels, encompassing observational trials, preliminary 

trials, and advanced yield trials under a multi-institutional 

network composed of breeders and connected specialists, 

located in various localities of the savanna region. 

Table 1 presents relevant data on aerobic rice lines, 

pooled from the best advanced yield trials, 86 out of 198, of 

the National Upland Breeding Network, conducted for three 

years at representative sites of the savanna region. 

Main program strategies 

Breeding for blast resistance 

In spite of the difficulties in breeding for blast resistance, breeders 

and pathologists were able to cooperate strongly since the 

initial steps of the program (Prabhu et al 1999). Donors possessing 

ample spectra of resistance were selected for the National 

Blast Nurseries conducted in nine locations throughout 

the country, and widely used in crosses. Results of subsidiary 

tests for leaf blast performed by pathologists in field nurseries 

were provided to breeders before field selection, in which segregating 

lines are evaluated under high pressure for leaf and 

panicle blast. Lines showing high scores in these environments 

are eliminated. Recurrent selection schemes have been used 

to generate resistant lines (Filippi et al 1994). The strong priority 

for blast resistance still holds true in the aerobic rice breeding 

program. But, although great care has been taken to produce 

and release blast-resistant varieties, their average life is 

very short because of the high variability of the pathogen under 

the upland ecosystem. The severity of the disease decreases 

the sustainability of the exploitation. 

Two strategies are currently being used to cope with this 

problem: increasing the number of releases possessing contrasting 

blast-tolerance genes and submitting the most promising 

advanced lines to a side backcross scheme, to transfer different 

genes of blast resistance to a given line. When the resistance 

of the original line breaks down, various isolines would 

be available to replace it. This procedure is accompanied by 

monitoring of the prevailing blast races in the region. 

Breeding for grain quality 

Grain quality was not a strong priority during phase 1 of the 

breeding program because the long and bold grains of tradi- 


Table 1. Mean values for yield and agronomic characteristics of aerobic rice varieties and 

lines, compared with those of traditional upland checks, 1 measured in the best advanced 

yield trials of the National Upland Breeding Network. 

Yield Days to Plant Lodging Leaf Panicle Grain 

Cultivar or line (kg ha –1 ) 50% height score blast blast discoloration 

flowering (cm) score score score 

CNAs 8812 4,419 87 100 1.3 1.8 1.9 2.7 

CNAs 8817 4,089 79 104 2.2 2.0 2.6 3.0 

CNAs 8989 4,739 80 99 1.9 2.3 2.5 3.2 

CNAs 8983 4,747 80 99 1.9 2.2 2.4 3.0 

CNAs 9019 4,757 77 104 3.3 2.0 2.1 2.1 

CNAs 9023 4,299 77 104 2.8 1.6 2.0 2.2 

CNAs 9025 3,964 77 107 2.5 1.7 2.4 2.5 

CNAs 9026 4,148 74 98 3.0 2.0 3.2 2.8 

CNAs 9045 4,162 85 109 1.9 2.0 2.1 2.7 

CNAs 10217 4,029 81 99 0.7 2.1 2.3 2.3 

CNAs 10222 4,272 79 101 1.0 2.0 2.1 2.3 

CNAs 10260 4,288 78 96 0.9 2.0 2.2 2.7 

BRSMG Conai 4,216 71 96 2.0 2.6 4.0 3.3 

BRS Bonança 4,315 82 98 1.2 1.9 1.8 2.3 

BRS Primavera 4,124 76 110 3.6 2.4 3.2 2.6 

BRS Talento 4,440 86 98 1.9 2.0 2.3 2.3 

Caiapó 1 4,054 90 122 3.7 2.2 2.1 3.0 

Canastra 3,433 87 107 2.3 2.4 2.5 2.8 

Carajás 1 4,422 77 103 2.3 2.1 2.3 2.5 

Maravilha 4,110 94 105 1.8 2.1 2.3 3.0 

Progresso 3,904 97 102 2.3 2.2 1.9 3.0 

BRS Soberana 3,920 76 114 3.2 2.2 2.9 2.7 

tional upland varieties were considered the standard for quality. 

During the 1980s, a successful marketing strategy was proposed 

by the irrigated rice producers’ associations and industry 

in southern Brazil to change consumer preferences and increase 

the national demand for long and slender grains. Visual 

selection for grain dimensions (length, width, thickness) and 

endosperm appearance is routinely made starting with the first 

segregating generations. To assist in the process, a chart was 

developed, based on standard varieties, which allows for quick 

field and laboratory scoring of grain dimensions. Starting with 

the F 5 generation, individual plants are selected for intermediate 

values of both amylose content and gelatinization temperature. 

This routine methodology has been adapted to use 

dehulled grains, thus reducing the sample size and decreasing 

the factor of variation induced by the milling process. In the 

F 6 generation, when grain availability is increased, grain translucency 

and milling recovery are also evaluated. The result of 

this concentrated effort is a high proportion of lines possessing 

long, slender, and translucent grains, with high to intermediate 

amylose content, low to intermediate gelatinization temperature, 

and milling recovery higher than 55%. The complete 

set of grain quality criteria is also applied during the yield 

testing program. 

Although there have been releases of improved-planttype 

varieties for the favored savanna conditions since the early 

1990s, the biggest impact was attained with varieties Maravilha 

and BRS Primavera, released in 1996. They were the first upland 

varieties that combine high grain quality and desirable 

aerobic rice ideotype, that is, 90–110 cm of plant stature, 200– 

250 tillers m –2 , erect leaves, resistance to lodging, and yield 

potential around 6 t ha –1 (Pinheiro 1999). 

A preference survey among rice millers indicated that 

BRS Primavera received preference ratings very close to those 

of BR IRGA 409, currently considered the most commercially 

competitive irrigated rice variety (Guimarães et al 2001). Most 

recent releases, such as BRS Talento and BRS Soberana, have 

grains that warrant ample acceptance by industry and consumers 

(Table 2). These varieties have cooking characteristics and 

commercial qualities similar to those of American long-grain 

varieties. 

Measuring advances of the breeding program 

The estimation of genetic gain, considering the period 1995- 

2000, indicates a consistent decrease in growth cycle duration 

of 0.52 day year –1 , which is desirable for the new profile of 

aerobic rice as a component of cropping systems. Besides the 

reduced time of exposition to biotic and abiotic stresses, the 

short growth cycle opens up an opportunity to insert it as a 

succession crop after soybean, to benefit from the remaining 

part of the rainy season in the savanna. 

Plant height decreased by 0.85 cm year –1 . Till 1998, there 

was a consistent decrease of 2.6 cm year –1 , when the program 

was reoriented to include lines with an intermediate height, 

which are more competitive with weeds. Lodging score has 

decreased at 4.16% per year because of both direct selection 

and plant size reduction. Positive gains were also attained for 

Session 5: Broadening the gene pool and exploiting heterosis in cultivated rice 155

Table 2. Grain quality data of representative aerobic rice varieties, compared 

with those of irrigated 1 and upland 2 quality checks (int. = intermediate). 

Varieties Milling Amylose Gelatinization Grain Grain 

yield content temperature length width 

(cm) (cm) 

BR IRGA 409 1 High 27.0 (int.) 7.0 (low) 6.81 2.06 

Caiapó 2 High 26.2 (int.) 3.9 (int.) 6.29 2.43 

Maravilha High 22.1 (low) 3.1 (high) 6.91 2.02 

BRS Primavera Low 26.3 (int.) 4.9 (int.) 7.29 1.96 

Canastra High 21.3 (low) 3.0 (high) 6.99 2.14 

BRS Bonança High 27.3 (int.) 3.4 (high) 6.22 2.29 

Carisma Medium 26.9 (int.) 3.9 (int.) 6.65 1.84 

BRS Talento Medium 26.8 (int.) 3.8 (int.) 6.93 1.92 

BRS Soberana Medium 25.8 (int.) 4.7 (int.) 7.08 1.95 

panicle blast and grain discoloration, whose scores have decreased 

at 1.59% and 1.81% per year, respectively. Although 

the overall yield gain was only 0.37% per year, it was very 

significant when only the favored region of the Brazilian savanna 

was considered, attaining a value as high as 1.85% per 

year in the state of Mato Grosso. 

The introduction of indica germplasm to the genetic base 

of the traditional upland tropical japonica population, associated 

with strong selection pressure for grain characteristics, 

probably imposed a restriction on possible gains for drought 

tolerance. The first aerobic rice releases do not possess the 

same level of drought tolerance as the traditional upland varieties, 

derived from crosses within only the japonica group 

(Pinheiro 2003). However, the program now holds a representative 

number of elite lines, japonica versus indica derivatives, 

providing a sound basis for further improving yield, blast resistance, 

and drought tolerance. 

Concluding remarks 

The alteration in plant type and grain quality of upland rice, 

toward its evolution to aerobic rice, led to an increase in yield 

potential and national market acceptance. Its increased participation 

in grain cropping systems, under no-tillage or minimum 

tillage, is foreseen. Moreover, aerobic rice could greatly 

contribute to environmental sustainability, avoiding new savanna 

deforestation, when used in the rice-pasture association, 

to renew the extensive areas of degraded pasture in the Brazilian 

savannas. 

References 

Filippi MC, Prabhu, AS, Neves PCF, Notteghen JL. 1994. Eficiência 

da seleção recorrente sobre a resistência parcial à brusone em 

arroz de sequeiro. Fitopatol. Bras. (Brasília): 19:279. 

Guimarães EP, Vieira NRA, Pinheiro BS. 2001. Breeding for specialty 

rice in Latin America: status and perspectives. In: 

Chaudhary RC, Tran VD, editors. Specialty rices of the world: 

breeding, production and marketing. Rome: FAO. p 317-322. 

Pinheiro BS. 1999. Características morfológicas da planta 

relacionadas à produtividade. In: Vieira NRA, Santos AB, 

Sant’ana EP, editors. A cultura do arroz no Brasil. Santo 

Antônio de Goiás (Brazil): Embrapa Arroz e Feijão. p 116- 

147. 

Pinheiro BS. 2003. Integrating selection for drought tolerance into a 

breeding program: the Brazilian experience. In: Fischer KS, 

Lafitte R, Fukai S, Atlin G, Hardy B, editors. Breeding rice 

for drought-prone environments. Los Baños (Philippines): 

International Rice Research Institute. p 75-83. 

Prabhu AS, Filippi MC, Ribeiro AS. 1999. Doenças e seu controle. 

In: Vieira NRA, Santos AB, Sant’ana EP, editors. A cultura 

do arroz no Brasil. Santo Antônio de Goiás (Brazil): Embrapa 

Arroz e Feijão. p 262-307. 

Steinmetz S, Reyniers FN, Forest F. 1988a. Caracterização do regime 

pluviométrico e do balanço hídrico do arroz de sequeiro 

em distintas regiões produtoras do Brasil: síntese e 

interpretação dos resultados. EMBRAPA-CNPAF Documentos 

23. Goiânia (Brazil): EMBRAPA-CNPAF. v. 1. 66 p. 

Steinmetz S, Reyniers FN, Forest F. 1988b. Caracterização do regime 

pluviométrico e do balanço hídrico do arroz de sequeiro 

em distintas regiões produtoras do Brasil: catálogo básico de 

dados. EMBRAPA-CNPAF Documentos 23. Goiânia (Brazil): 

EMBRAPA-CNPAF. v. 2. 278 p. 

Notes 

Author’s address: Embrapa Arroz e Feijão, Caixa Postal 179, 75375- 

000 Santo Antônio de Goiás, GO, Brazil, e-mail: 

beatriz@cnpaf.embrapa.br. 


Broadening the gene pool of rice 

through introgression from wild species 

D.S. Brar 

World rice production has more than doubled, from 257 million 

tons in 1966 to 600 million t in 2000. To meet the growing 

needs of the human population, rice production must increase 

by 25% during the next 20 years. Further, several biotic 

and abiotic stresses adversely affect rice productivity. Some 

of the major diseases and pests affecting rice production are 

bacterial blight (BB), blast, sheath blight, tungro virus disease, 

and rice yellow mottle virus (RYMV), and insects such as 

brown planthopper (BPH), stem borer, and Asian and African 

gall midge. Similarly, abiotic stresses such as drought, cold, 

salinity, acidity, iron toxicity, and submergence reduce rice 

production. Changes in insect biotypes and disease races are 

becoming a continuing threat to rice production. There is thus 

an urgent need to broaden the rice gene pool through introgression 

of genes from new and diverse sources. Wild species 

are an important reservoir of useful genes for tolerance of biotic 

and abiotic stresses. 

Useful traits of Oryza species 

The genus Oryza has two cultivated and 22 wild species. Of 

the two cultivated species, O. sativa (2n=24 AA), commonly 

referred to as Asian rice, is grown worldwide, whereas O. 

glaberrima (2n=24 AA), “African rice,” is cultivated in a limited 

area in West Africa. The wild species have 2n=24 or 48 

chromosomes representing 10 genomic types (AA, BB, BBCC, 

CC, CCDD, EE, FF, GG, HHJJ, and HHKK). The wild species 

germplasm maintained in the IRRI genebank contains 2,500 

accessions of AA-genome types, 500 accessions of O. 

officinalis complex, 35 accessions of O. meyeriana complex, 

21 accessions of O. ridleyi complex, 19 accessions of O. 

brachyantha, and 1 accession of O. schlechteri. 

Several incompatibility barriers, such as low crossability, 

increased sterility, and limited recombination between chromosomes 

of wild and cultivated species, limit the transfer of 

useful genes (Brar and Khush 1986, 2002). Recent advances 

in tissue culture and genomics have enabled the production of 

wide hybrids among distantly related species and allowed researchers 

to precisely monitor the introgression of chromosome 

segments from wild into cultivated species. More recently, 

BAC libraries of wild species representing 10 different genomes 

have been developed by the Arizona Genomics Institute 

in Tucson (Rod Wing, personal communication, 2004), 

which offer new opportunities for map-based cloning of useful 

genes/QTLs. 

Alien gene introgression 

The main objectives of our wide hybridization program are 

(1) to broaden the gene pool of rice by transferring useful genes 

for resistance to major diseases and insects and tolerance of 

abiotic stresses and to enhance the grain yield of rice through 

the introgression of QTLs/yield-enhancing loci from wild species, 

(2) to tag alien genes/QTLs introgressed from wild species 

with molecular markers for use in marker-assisted selection 

(MAS), (3) to characterize alien introgression using molecular 

cytogenetic techniques, and (4) to isolate agronomically 

important genes/QTLs using BAC libraries of Oryza. 

Some accomplishments made in collaboration with national 

agricultural research and extension systems (NARES) 

partners and advanced research institutes (ARI) are discussed 

below. 

Production of interspecific hybrids, 

alien introgression lines, and mapping populations 

Direct crosses and embryo rescue techniques have been used 

to successfully produce hybrids between rice (AA) and all other 

wild species (except O. schlechteri). Backcrossing with the 

recurrent rice parent is used to produce fertile progenies. Introgression 

lines (2n=24) have been produced from crosses of 

AA with various wild species representing AA, BBCC, CC, 

CCDD, EE, FF, GG and HHJJ genomes except for O. sativa 

(AA) × O. coarctata (HHKK) (Brar and Khush 2002). 

Doubled haploids (DH) have been produced from O. 

sativa × O. glaberrima. Similarly, recombinant inbred lines 

(RILs) from crosses of rice with A-genome wild species and 

near-isogenic lines (NILs) derived through backcrossing carrying 

small segments from distant genomes serve as mapping 

populations. Segregating populations derived from crosses of 

alien introgression lines × recurrent rice parents are also used 

in mapping genes/QTLs. 

Production of monosomic alien addition lines 

(MAALs) and chromosome segmental substitution 

lines (CSSL) 

MAALs (2n=25) have been produced representing 6 to 10 

chromosomes from 7 wild species (CC, BBCC, CCDD, EE, 

FF, GG, and HHJJ genomes). CSSL are being developed from 

O. rufipogon, O. longistaminata, and O. glaberrima in the 

background of O. sativa using molecular markers. These CSSL 

are important resources in fine mapping of genes and functional 

genomics. 


Table 1. Progress in the transfer of agronomically important genes 

from wild Oryza species into cultivated rice at IRRI. a 

Donor Oryza species 

Trait Wild species Genome Accession 

number 

Transferred to O. sativa 

Grassy stunt resistance O. nivara AA 101508 

Bacterial blight resistance O. longistaminata AA – 

O. officinalis CC 100896 

O. minuta BBCC 101141 

O. latifolia CCDD 100914 

O. australiensis EE 100882 

O. brachyantha FF 101232 

Blast resistance O. minuta BBCC 101141 

Brown planthopper O. officinalis CC 100896 

resistance O. minuta BBCC 101141 

O. latifolia CCDD 100914 

O. australiensis EE 100882 

Whitebacked planthopper O. officinalis CC 100896 

resistance 

Cytoplasmic male sterility O. perennis AA 104823 

O. glumaepatula AA 100969 

Tungro resistance O. rufipogon AA 105908 

O. rufipogon AA 105909 

O. rufipogon AA 106423 

Introgression lines under evaluation 

Yellow stem borer O. longistaminata AA 110404 

O. rufipogon AA – 

Sheath blight resistance O. minuta BBCC 101141 

O. rufipogon AA – 

Increased elongation ability O. rufipogon AA CB751 

Tolerance of acidity and iron O. glaberrima AA Many 

and aluminum toxicity O. rufipogon AA 106412 

O. rufipogon 106423 

Resistance to nematodes O. glaberrima AA Many 

a 

Modified from Brar and Khush (2002). 

Alien gene transfer and varietal development 

Genes for resistance to BPH, BB, blast, and grassy stunt and 

tungro virus, tolerance of acidity, and cytoplasmic male sterility 

have been introgressed from different A-genome wild species, 

including distantly related genomes (CC, BBCC, CCDD, 

EE, FF), into elite breeding lines of rice (see Brar and Khush 

2002, Multani et al 2003, Table 1). Some of these lines carrying 

genes from wild species for resistance to BPH and tungro 

and tolerance of acid sulfate soil conditions have been released 

as commercial varieties. 

During 2002, three varieties were released from wide 

crosses. One of the breeding lines, IR73678-6-9-B from IR64 

× O. rufipogon, was released as variety AS996 for commercial 

cultivation in the Mekong Delta of Vietnam. This variety is 

grown on more than 100,000 ha. Another variety, Matatag 9 

(IR73385-1-4-3-2-1-6), has been released for cultivation in 

tungro-prone areas of the Philippines. Similarly, another highyielding 

elite line (IR72102-4-159-1-3) from O. sativa × O. 

longistaminata has been released as variety NSICRC112 in 

the Philippines. Earlier, five IRRI breeding lines from O. sativa 

× O. officinalis were released as BPH-resistant varieties 

(MTL98, MTL103, MTL105, MTL110, MTL114) in Vietnam. 

Recently, variety Dhanrasi (IE15358) from the cross of B32- 

sel-4 × O. rufipogon was released by the Directorate of Rice 

Research in Hyderabad, India (T. Ram, personal communication, 

2004). 

Some of the indica alien introgression lines developed 

at IRRI have shown a wide spectrum of resistance to BPH and 

blast in Korea. These lines are being used by the Rural Development 

Administration (RDA) to broaden the gene pool of 

japonica rice. 

We are evaluating advanced progenies derived from different 

wide crosses for resistance to stem borer and tolerance 

of aluminum toxicity, P deficiency, and iron toxicity. Some O. 

rufipogon accessions resistant to sheath blight have been identified 

and are being used to produce advanced introgression 

lines. Progenies derived from O. sativa × O. glaberrima have 

shown increased tolerance of iron toxicity under field conditions. 

Advanced introgression lines are under evaluation for 

the transfer of weed competitive ability from O. glaberrima 

into O. sativa. 

Enhancing the yield potential of rice 

Wild species are phenotypically inferior but are valuable genetic 

resources for enhancing the yield potential of rice (Xiao 

et al 1998). Preliminary results at IRRI from the crosses of 

new plant type (NPT) × O. longistaminata and IR64 × O. 

rufipogon indicate possibilities to improve rice grain yield by 

introgressing yield-enhancing loci/QTLs. 

Molecular mapping of introgressed genes/QTLs 

and characterization of introgression 

The introgressed genes Bph10, Pi9(t), and Xa21 have been 

mapped. One of the genes, Xa21 for BB resistance, has been 

used via MAS in gene pyramiding at IRRI and by NARES in 

the Philippines, India, China, and Thailand. Genes introgressed 

for BPH and tungro tolerance are being mapped for use in 

MAS. One of the major QTLs for tolerance of aluminum toxicity 

introgressed from O. rufipogon has been mapped on chromosome 

3, which is conserved across other cereal species 

(Nguyen et al 2003). 

Molecular marker analysis has revealed limited introgression 

of small chromosome segments from distant genomes 

(FF, GG) of Oryza into rice. However, introgression among 

cultivated rice and A-genome wild species is frequent for all 

12 chromosomes. We are developing genome-specific clones 

using representational difference analysis (RDA) to characterize 

alien introgression. 

Molecular characterization using GISH 

GISH techniques have been used successfully to characterize 

parental genomes, extra alien chromosomes, and 

homoeologous pairing between rice and genomes of several 

wild species, which were difficult to identify through conventional 

cytogenetic techniques. 


Two new genomes (GG and HHJJ) have been assigned 

for O. granulata and O. ridleyi based on molecular divergence 

involving total genomic DNA hybridization (Aggarwal et al 

1997). RDA was used to identify transposable elements in rice 

(Panaud et al 2002). 

Development of novel genetic resources 

for functional genomics 

Near-isogenic alien introgression lines, particularly those having 

short introgressed segments derived from crosses of O. 

sativa with species having distant genomes of Oryza, have 

become novel genetic resources for functional genomics. 

Research priorities and future outlook 

With the advances in tissue culture and molecular markers, 

genomics, and molecular cytogenetics, the outlook for broadening 

the gene pool of rice for tolerance of major biotic and 

abiotic stresses through the introgression of genes from wild 

species seems more promising than before. Some of the priorities 

for future research in wide hybridization are listed below. 

Broadening the gene pool for tolerance 

of major biotic and abiotic stresses 

Emphasis should be given to transferring genes from wild species 

for traits for which genetic variability is limited in the 

cultivated germplasm. Priority traits include resistance to sheath 

blight and stem borer and tolerance of salt and drought. There 

is a need to search for micronutrients such as high iron and 

zinc in the polished grains of wild species. Genes with a broad 

spectrum of tolerance of biotic and abiotic stresses introgressed 

from wild species should be transferred into elite indica and 

japonica cultivars. 

Pyramiding of genes/QTLs with a broad spectrum 

of resistance 

A number of genes for tolerance of biotic stresses have been 

introgressed from wild species into rice and also tagged with 

molecular markers. MAS should be practiced to pyramid genes 

to develop durable resistance to pests. Emphasis should be 

given to enhancing tolerance of abiotic stresses by combining 

genes/QTLs governing tolerance at different stages of development 

and for different component traits/mechanisms of tolerance. 

Identifying and introgressing QTLs/yield-enhancing 

loci from wild species 

Molecular analysis has revealed novel QTLs/yield-enhancing 

loci in wild species. There is a need to identify and introgress 

such alleles/QTLs, particularly from A-genome wild species, 

into modern rice varieties for enhancing yield potential. 

Enhancing homoeologous pairing/recombination 

One of the major bottlenecks in alien gene transfer is the limited 

homoeologous pairing between rice and distantly related 

species of Oryza. One approach would be to isolate the gene(s) 

controlling homoeologous pairing such as the Ph 

(homoeologous pairing) gene in wheat and use such genes to 

promote homoeologous recombination in rice. This would be 

particularly important in species such as O. brachyantha (FF), 

O. granulata (GG), and O. ridleyi (HHJJ), which otherwise 

show restricted recombination with the cultivated rice genome. 

Isolation of agronomically important genes/QTLs 

through positional cloning 

Wild species carry genes with a wide spectrum of tolerance of 

biotic and abiotic stresses. BAC libraries of different wild species 

representing each genomic type of Oryza should be used 

for fine mapping and isolation of agronomically important 

genes/QTLs. 

GISH-assisted selection for alien introgression 

Emphasis should be given to developing efficient GISH protocols 

to detect and monitor alien introgression in the backcross 

progenies. Molecular cytogenetic techniques involving 

molecular markers, GISH, FISH, and fiber-FISH should be 

used to precisely determine the mechanism of alien gene transfer 

in the face of limited homoeologous pairing. 

Producing haploids from species crosses 

Chromosome elimination through species crosses has proven 

to be a valuable mechanism to produce haploids in wheat, barley, 

and oat. Such a haploid-inducing system needs to be identified 

in rice, which could be complementary, to produce haploids 

in indica rice, in which anther culture response is low 

and limited to only specific genotypes. 

References 

Aggarwal RK, Brar DS, Khush GS. 1997. Two new genomes in the 

Oryza complex identified on the basis of molecular divergence 

analysis using total genomic DNA hybridization. Mol. Gen. 

Genet. 254:1-12. 

Brar DS, Khush GS. 1986. Wide hybridization and chromosome 

manipulation in cereals. In: Evans DH, Sharp WR, Ammirato 

PV, editors. Handbook of plant cell culture. Vol. 4. Techniques 

and applications. New York (USA): MacMillan Publishers. 

p 221-263. 

Brar DS, Khush GS. 2002. Transferring genes from wild species 

into rice. In: Kang MS, editor. Quantitative genetics, genomics 

and plant breeding. Wallingford (UK): CAB International. 

p 197-217. 

Multani DS, Khush GS, delos Reyes BG, Brar DS. 2003. Alien gene 

introgression and development of monosomic alien addition 

lines from Oryza latifolia Desv. to rice, Oryza sativa L. Theor. 

Appl. Genet. 107:395-405. 

Nguyen BD, Brar DS, Bui BC, Nguyen TV, Pham LN, Nguyen HT. 

2003. Identification and mapping of the QTLs for aluminum 

tolerance introgressed from the new source, Oryza rufipogon 

Griff., into indica rice (Oryza sativa L.). Theor. Appl. Genet. 

106:583-593. 


Panaud O, Vitte C, Hivert J, Muzlak S, Talag J, Brar DS, Sarr A. 

2002. Characterization of transposable elements in the genome 

of rice (Oryza sativa L.) using representational difference 

analysis (RDA). Mol. Genet. Genom. 268:113-121. 

Xiao J, Li J, Grandillo S, Ahn SN, Yuan L, Tanksley SD, McCouch 

SR. 1998. Identification of trait-improving quantitative trait 

loci alleles from a wild rice relative, Oryza rufipogon. Genetics 

150:899-909. 

Notes 

Author’s address: International Rice Research Institute, DAPO Box 

7777, Metro Manila Philippines. 

Acknowledgment: The help provided by NARES partners, IRRI colleagues, 

and the wide hybridization team is duly acknowledged. 

Mutation in seed reserves and its use 

for improving grain quality in rice 

Hikaru Satoh, Ken-ichi Ohtsubo, and Yasunori Nakamura 

The improvement of grain quality is one of the most important 

subjects in rice breeding. For this, genetic resources of endosperm 

properties must be collected, characterized, and evaluated. 

In maize, various kinds of mutants for polysaccharides, 

lipids, or proteins are known. These mutants greatly contributed 

to the improvement of the grain quality of maize, and 

thus expanded the use of maize not only as a food but also as 

an important industrial material in food chemistry. In addition, 

they offered valuable information on the study of gene 

action in the biological processes of metabolic regulation in 

higher plants and led us to develop the novel genetic resources 

improving these properties. 

Starch plays an important role in the eating quality of 

cooked rice and the processing quality of industrial uses. In 

rice, the mutants of starch properties had been little known 

except for the waxy endosperm. We treated the fertilized egg 

cells of rice with MNU (Satoh and Omura 1979) and obtained 

several thousand rice mutants for embryo or endosperm properties 

(Satoh 1985), in which various kinds of mutants modifying 

starch properties were found, as many as in maize (Satoh 

and Omura 1981, Satoh et al 2003). 

Structure of starch and its biosynthesis 

Rice starch consists of two types of glucose polymers, 20% 

amylose and 80% amylopectin. Amylose is an essentially linear 

molecule composed of α-(1,4)-linked glucosidic chains, 

although recently another type of amylose possessing some 

very short-branched chains was reported. Amylopectin is a 

highly branched glucan with α-(1,6) glucosidic bonds that connect 

linear chains. Hizukuri (1986) proposed a cluster model 

for amylopectin. In this model, A and B 1 chains form a single 

cluster, whereas B 2 and B 3 chains extend to two and three clusters, 

respectively. Amylopectin is much larger in molecular size 

than amylose, tenfold or more. The amylose to amylopectin 

ratio and their structures greatly influence the rheological properties 

of starch. 

Amylose is synthesized by ADP glucose 

pyrophosphorylase (AGPase) and granule-bound starch synthase 

I (GBSSI), which is encoded by the Waxy gene. Amy- 

lopectin is synthesized by concerted reactions catalyzed by 

soluble starch synthase (SS), starch-branching enzyme (BE), 

and starch-debranching enzyme (DBE), using ADP-glucose as 

a substrate. In addition, multiple isoforms were found in each 

enzyme, that is, three isoforms of SS and BE, and two isoforms 

of DBE in developing endosperm of rice, respectively. Therefore, 

at least these eight genes are responsible for amylopectin 

biosynthesis in rice endosperm. 

Mutants modifying amylose content 

To improve starch quality, novel genetic resources must be 

developed and evaluated. For this purpose, we need to elucidate 

the genetic regulation mechanism of starch biosynthesis 

and the interrelationships among the gene, enzyme, starch structure, 

and rheological properties. A mutant is one of the most 

helpful materials for this purpose. 

After the MNU treatment, many mutants with modified 

amylose content were obtained in rice (Yano et al 1988). They 

were classified into waxy type and dull type with lower amylose 

content. Amylose content in endosperm starch varied from 

0.3% to 12.8%, depending on the mutant line. Phenotypes of 

dull mutants were intermediate between those of nonwaxy rice 

and waxy rice, and there was good agreement between the phenotype 

and amylose content. Thermal gelatinization of rice 

starch is influenced by amylose content. There is little difference 

in the initiation of gelatinization of endosperm starch 

among the waxy, dull, and wild type. However, the termination 

temperature of gelatinization is remarkably different among 

them. When the amylose content increases, the termination temperature 

of gelatinization of endosperm starch shifts to a high 

temperature. When starch is gelatinized in 4 M urea solution, 

granules with lower amylose content swell more. However, 

these mutations did not alter amylopectin fine structure. These 

facts indicate that amylose content affects the termination temperature 

of gelatinization of starch and the swelling power of 

gel, but it has little influence on the initiation of gelatinization. 


Mutants involved in amylopectin biosynthesis 

The branching enzyme is the only enzyme that can introduce 

α-1,6 glucosidic linkages into α-polyglucans and therefore it 

plays an essential role in the biosynthesis of amylopectin. Three 

isoforms of branching enzyme (BE I, BE IIa, and BE IIb) are 

observed in rice endosperm. Recently, we isolated mutants for 

the respective isoforms of BE, that is, mutants lacking in BE I 

(Satoh et al 2003) or BE IIb activity (Nishi et al 2001) and a 

mutant decreasing remarkably in BE IIa activity (Satoh et al 

2003). All of the BE mutants were controlled by respective 

single genes independent from each other. The grains of the 

BE IIb mutant are markedly smaller than those of the wild 

type and are usually floury in appearance, indicating that this 

enzyme affects starch accumulation. 

In contrast, the endosperm of BE I and BE IIa mutants 

exhibited the normal phenotype and contained the same amount 

of starch as the wild type. BE IIb mutation specifically decreased 

the short chains with less than DP 17, suggesting that 

BE IIb contributes to the synthesis of A chains of amylopectin. 

In contrast, the amylopectin of the BE I mutant was characterized 

by a significant decrease in long chains with more than 

DP 38 and short chains with DP 12 to 23, and a marked increase 

in short chains with less than DP 12, suggesting that BE 

I specifically synthesizes both B 1 chains and B 3 chains. However, 

a notable alteration in the chain length distribution profile 

was not observed in the BE IIa mutation. 

It is therefore reasonable to assume that the three forms 

of BE play distinct roles in the formation of amylopectin 

branches in rice. The endosperm starch from the BE I mutant 

had the lowest onset temperature for thermo-gelatinization. In 

contrast, the BE IIb mutant showed the highest onset temperature 

for thermo-gelatinization. The BE IIa mutation had little 

influence on the gelatinization of starch but exhibited the highest 

enthalpy. These findings indicate that the genetic modification 

of amylopectin fine structure is responsible for changes 

in physicochemical and rheological properties of the starch. 

The results support the view that alterations in amylopectin 

structure, in particular in short chains within clusters, might 

play a critical role in the rheological properties of starch. It is 

therefore likely that genetic modification of the gene for BEs 

will lead to the synthesis of novel types of starch as new materials 

for the food and starch industries. 

Native-PAGE/activity staining analysis for the 

debranching enzyme showed that sugary1 mutation is lacking 

in isoamylase activity (ISA) and is followed by a marked reduction 

in expression levels of pullulanase (Nakamura et al 

1997). Two kinds of nonallelic sugary mutants have been identified 

in rice. Sugary mutations increased short chains and decreased 

long chains dramatically. Rheological properties of 

the sugary1 mutant are characterized by extremely lower viscosity 

in contrast to that of the wild type (Wong et al 2003). 

These results also imply that amylopectin fine structure is important 

in determining rheological properties of starch and the 

genetic modification of DBE also led to the development of 

novel starch in rice. 

Recently, we isolated some of the mutants involved in 

soluble starch synthase (SS). A preliminary examination showed 

that these SS mutations altered the fine structure of amylopectin 

and its rheological properties as well as BE and DBE mutations. 

The details are under investigation. In addition to these 

mutations, we isolated various kinds of mutants by MNU treatment. 

These mutations also modified the amylopectin fine structure 

and rheological properties. 

Spontaneous mutation for starch biosynthesis 

A wide variation in apparent amylose content was detected in 

local rice germplasm collected from Bangladesh. Genetical 

and biochemical analyses showed that this wide variation is 

caused by spontaneous mutations in the GBSS gene, especially 

the indica-type GBSS gene, Wx-a. At least three spontaneous 

mutations lowering the gene expression of Wx-a have been 

found in Bangladeshi local rice germplasm (Jahan et al 2002a). 

Cooking quality is markedly different between japonica 

and indica rice. Alkali digestibility is one of the most important 

indicators for the difference in cooking quality between 

them as well as the amylose content because there is a tight 

correlation between alkali digestibility and gelatinization temperature 

in rice starch. Indica rice is resistant to alkali digestibility, 

whereas japonica rice is easily digested by alkali solution. 

Umemoto et al (2002) made clear that the difference in 

alkali digestibility and urea gelatinization is caused by the difference 

in the chain length distribution of amylopectin, and 

these characters are controlled by the same gene. RFLP analysis 

showed clearly that a gene encoding SSIIa is responsible 

for this character. 

However, wide variation was found in amylopectin chain 

length distribution as well as in alkali digestibility in BGD 

culivars (Jahan et al 2002b). Similar variations were observed 

in local rice germplasm collected from other countries such as 

Myanmar, Pakistan, African countries, and so on. These variations 

might be caused partly by the other amylopectin biosynthetic 

enzymes in addition to SSIIa because wide variation was 

found in the activity level not only of SSIIa but also of other 

SSs, BEs, and DBEs. RFLPs for the genes involved in starch 

biosynthesis supported this assumption. 

Many kinds of mutant genes involved in starch biosynthesis 

have been identified in both rice and maize. The analysis 

of these mutants affecting the production of starch-synthesizing 

enzymes will provide the in vivo function of the corresponding 

enzymes and lead to breeding rice varieties having 

novel properties of starch. 

References 

Hizukuri S. 1986. Polymodal distribution of the chain lengths of 

amylopectins, and its significance. Carbohydr. Res. 147:342- 

347. 

Jahan MS, Kumamaru T, Hamid A, Satoh H. 2002a. Diversity of 

granule-bound starch synthase (GBSS) level in Bangladesh 

rice cultivars. Rice Genet. Newsl. 19:69-71. 


Jahan MS, Nishi A, Hamid A, Satoh H. 2002b. Variation in amylopectin 

fine structure of Bangladesh rice cultivars. Rice Genet. 

Newsl. 19:72-74. 

Nakamura Y, Kubo A, Shimamune T, Mastuda T, Harada K, Satoh 

H. 1997. Correlation between activities of starch debranching 

enzyme and α-polyglucan structure in endosperms of sugary- 

1 mutants of rice. Plant J. 12:143-153. 

Nishi A, Nakamura Y, Tanaka N, Satoh H. 2001. Biochemical and 

genetic analyses of the effects of amylose-extender mutation 

in rice endosperm. Plant Physiol. 127:1-14. 

Satoh H. 1985. Genic mutations affecting endosperm properties in 

rice. Gamma Field Symp. 24:17-37. 

Satoh H, Omura T. 1979. Induction of mutation by the treatment of 

fertilized egg cell with N-methyl-N-nitrosourea in rice. J. Fac. 

Agric. Kyushu Univ. 24:165-174. 

Satoh H, Omura T. 1981. New endosperm mutations induced by 

chemical mutagens in rice, Oryza sativa L. Jpn. J. Breed. 

31:316-325. 

Satoh H, Nishi A, Fujita N, Kubo A, Nakamura Y, Kawasaki T, Okita 

WT. 2003. Isolation and characterization of starch mutants in 

rice. J. Appl. Glycosci. 50:225-230. 

Satoh H, Nishi A, Yamashita K, Takemoto Y, Tanaka Y, Hosaka Y, 

Sakurai A, Fujita N, Nakamura Y. 2003. Starch branching 

enzyme I-deficient mutation specifically affects the structure 

and properties of starch in rice endosperm. Plant Physiol. 

133:1111-1121. 

Umemoto T, Yano M, Satoh H, Shomura A, Nakamura Y. 2002. 

Mapping of a gene responsible for the difference in amylopectin 

structure between japonica-type and indica-type rice varieties. 


Wong KS, Kubo A, Jane JL, Harada K, Satoh H, Nakamura Y. 2003. 

Structures and properties of amylopectin and phytoglycogen 

in the endosperm of sugary-1 mutants of rice. J. Cereal Sci. 

37:139-149. 

Yano M, Okuno K, Satoh H, Omura T. 1988. Chromosomal location 

of genes conditioning low amylose content of endosperm 

starches in rice, Oryza sativa L. Theor. Appl. Genet. 76:183- 

189. 

Notes 

Authors’ addresses: Hikaru Satoh, Faculty of Agriculture, Kyushu 

University (Hakozaki, Fukuoka 812-8581, Japan); Ken-ichi 

Ohtsubo, National Food Research Institute (Kan-nondai, 

Tsukuba, Ibaraki 305-8642; Yasunori Nakamura, Faculty of 

Bioresource Science, Akita Prefectural University 

(Shimoshinjo-Nakano, Akita-City 010-0195, Japan), e-mail: 

hdatoh@agr.kyushu-u.ac.jp. 

Heterosis in rice for increasing yield, production efficiency, 


Sant S. Virmani 

Heterosis is known to be a major factor for increased production 

in several crops, including rice. It has become the basis of 

multibillion-dollar agribusiness worldwide. Hybrid varieties 

in maize, sorghum, sunflower, and rice added 90 million tons 

annually to global food production, which spared about 34 

million ha of land from cultivation of these crops to meet global 

demand (Duvick 1999). Hybrid rice technology, commercialized 

in China, has spread to about 50% of the 30 million 

ha of rice area of that country, contributing 60% of its national 

paddy production. Inbred rice covers the other 50% of the area, 

contributing only 40% of the national paddy production. In 

China, rice hybrids (6.9 t ha –1 ) on average have outyielded 

inbred rice (5.4 t ha –1 ) by 1.5 t ha –1 (Ma and Yuan 2003). Since 

1979, IRRI has been exploring the prospects of exploiting 

heterosis in the tropics to increase rice yield potential in the 

light of increasing rice demand, decreasing land area and water 

resources for rice production, and the need for sparing rice 

land for crop diversification to increase farmers’ income. This 

paper summarizes the progress made so far and discusses the 

future prospects of hybrid rice in the tropics. 

Extent of yield heterosis 

Data gathered at IRRI and in India, the Philippines, Bangladesh, 

Indonesia, Vietnam, Myanmar, and Sri Lanka clearly provide 

evidence for a 15–20% yield advantage of rice hybrids over 

inbred high-yielding varieties (HYVs) (Virmani et al 2003). 

Even in Korea and Egypt, where yield levels are considerably 

higher, a yield advantage (10–15%, 1–1.5 t ha –1 ) of rice hybrids 

over inbred HYVs has been reported (Virmani et al 2003). 

Heterosis for yield in rice hybrids has been attributed to their 

increased dry matter production caused by higher leaf area 

index, higher crop growth rate, and increased harvest index 

resulting from increased spikelet number and increased grain 

weight. Parental lines developed at IRRI and shared freely with 

national programs during the past 15–20 years have contributed 

significantly toward achieving these results. More than 

40 heterotic rice hybrids, developed by the public and private 

sector, have been commercialized during the past ten years in 

some countries outside China. Those known to be derived from 

IRRI-bred parental lines are listed in Table 1. The yield advantage 

of rice hybrids over inbred HYVs was even higher 

(20–30%) in farmers’ fields in the Philippines and India, perhaps 

because of their homeostatic effects. During 2004, al- 


Table 1. List of some hybrids (derived from IRRI-bred germplasm) released for commercial cultivation in some tropical 

countries during 1994-2004. 

Nature of Country Year 

Name of hybrid Parentage breeding released released 

institution 

APHR-1 IR58025A/Vajram Public India 1994 

APHR-2 IR62829A/MTU 9992 Public India 1994 

MGR-1 (IR64610H) IR62829A/IR9761-19-1R Public India 1994 

KRH-1 (IR64611H) IR58025A/IR9761-19-1R Public India 1994 

Magat (PSB Rc26H or IR64616H) IR62829A/IR29723-143-3-2-1R Public Philippines 1994 

CNRH-3 IR62829A/Ajaya Public India 1995 

DRRH-1 (IR65489H) IR58025A/IR40750-82-2-2-3R Public India 1996 

KRH-2 IR58025A/KMR-3 Public India 1996 

Pant Sankar Dhan-1 IR58025A/UPRI 93-133 Public India 1997 

PHB-71 IRRI-bred A line Private India 1997 

ADTRH-1 IR58025A/IR66 Public India 1998 

CORH-2 IR58025A/C20R Public India 1998 

Narendra Sankar Dhan-2 IR58025A/NDR 3026 Public India 1998 

Sahyadri (IR69690H) IR58025A/BR827-35-2-1-1-1R Public India 1998 

Mestizo (PSB Rc72H or IR68284H) IR58025A/IR34686-179-1-2-1R Public Philippines 1997 

HYT-57 (IR69690H) IR58025A/BR827-35-2-1-1-1R Public Vietnam 1999 

Proagro 6201 IRRI-bred A line Private India 2000 

BRRI Dhan Hybrid 1 (IR69690H) IR58025A/BR827-35-2-1-1-1R Public Bangladesh 2001 

HRI 120 (6444) IRRI-bred A line Private India 2001 

Intani 1 IRRI-bred A and R line Private Indonesia 2001 

Intani 2 IRRI-bred A and R line Private Indonesia 2001 

Rokan (IR69690H) IR58025A/BR827-35-2-1-1-1R Public Indonesia 2002 

Maro IR58025A and IRRI-developed R line Public Indonesia 2002 

Mestizo 2 (NSIC Rc114H or IR75207H) IR68888A/IR62161-184-3-1-3-2R Public Philippines 2002 

Mestizo 3 (NSIC Rc116H or IR75217H) IR68897A/IR60819-34-2R Public Philippines 2002 

Bigante (NSIC Rc124H or Mestizo 4) Developed using IRRI germplasm Private Philippines 2004 

IR78386H a IR68897A/IR71604-4-1-4-4-4-2-2-2R Public Philippines 2004 

a Recommended for release by the Rice Technical Working Group, Philippines, 25 November 2004. 

most 1.4 million ha were covered with rice hybrids in India 

(560,000 ha), the Philippines (200,000 ha), Vietnam (600,000 

ha), Bangladesh (40,000 ha), Indonesia (5,000 ha), and 

Myanmar (5,000 ha). Rice hybrids have also been commercialized 

in the United States (Walton 2003) and Japan (Takita 

2003). Several other countries (Sri Lanka, Thailand, Egypt, 

Brazil, Iran, Malaysia, and Pakistan) should be commercializing 

this technology in the next 3–5 years. 

Rice hybrids have also shown significant heterosis for 

earliness. The higher yield coupled with slightly shorter duration 

resulted in higher per-day productivity (68–78 kg d –1 

ha –1 ) than inbred HYVs (62–72 kg d –1 ha –1 ) (Virmani and 

Kumar 2004). The shorter duration of rice hybrids can result 

in a savings of water and creation of niches where an additional 

crop can be grown to allow crop diversification. Peng et 

al (2003) observed higher N-use efficiency in an IRRI-rice 

hybrid—IR68284H (named as Mestizo 1 in the Philippines)— 

than in a widely grown inbred HYV, IR72. Earlier studies at 

IRRI had shown a significantly higher response of an IRRI 

hybrid (IR64616H, named as Magat in the Philippines) to the 

application of N at booting than IR72. These results clearly 

show the increased production efficiency of rice hybrids compared 

with inbred rice and illustrate how hybrid rice becomes 

relevant in the current competitive global economic scenario. 

Breeding of parental lines to exploit heterosis in rice 

Cytoplasmic male sterility and the fertility restoration system 

have been primarily used to develop heterotic rice hybrids in 

and outside China (Virmani 1996). During the past decade, 

thermo- (TGMS) and photoperiod-sensitive genic male sterility 

(PGMS) systems have also been developed in China (Ma 

and Yuan 2003) and outside China (Virmani and Ilyas-Ahmed 

2001). Commercially usable parental lines to develop tropical 

rice hybrids must have adaptability to tropical conditions. In 

addition, the male sterile lines used for this purpose must have 

complete and stable male sterility, and good outcrossing potential 

essential for economically viable hybrid seed production. 

Likewise, commercially usable restorer lines should restore 

normal fertility in the derived hybrids and provide adequate 

pollen for a prolonged period for economically viable 

seed production. The TGMS system is the most effective for 

developing tropical rice hybrids; the PGMS system cannot be 

used for lack of sufficient variation in daylength during the 

year (Virmani and Ilyas-Ahmed 2001). Tropical TGMS lines 

possessing a low critical sterility point (CSP), complete and 

stable male sterility in the wet and dry seasons, and good fertility 

reversion in available high-altitude locations are now 

available at IRRI and in national programs. 


Among the first set of IRRI-bred CMS lines, IR58025A, 

IR62829A, IR68888A, and IR68897A have been used to develop 

several tropical commercial rice hybrids in India, the 

Philippines, Vietnam, Indonesia, and Bangladesh (Table 1). 

The most widely used CMS line, IR58025A, resulted in hybrids 

possessing aroma and somewhat sticky grains on cooking. 

More recently, CMS lines possessing better grain quality 

and higher outcrossing than IR58025A have been developed 

(Virmani and Kumar 2004). Some of these also possess a different 

CMS system. These are being used extensively at IRRI 

and are shared with public and private research institutions in 

national agricultural research and extension systems (NARES) 

to enable them to breed improved hybrids. 

Elite indica and indica/tropical japonica derivative lines 

possess 30–40% restorer lines, which are usable as such to 

develop heterotic rice hybrids in tropical countries. Restorer 

frequency in temperate and tropical japonica rice is negligible 

and is very low in basmati rice. To develop rice hybrids in 

these genetic backgrounds, restorer lines have to be bred by 

transferring Rf genes from indica rice. To increase the frequency 

of restorer and maintainer lines in the tropics, R × R and B × B 

crosses are made routinely in IRRI’s hybrid rice breeding program. 

The nuclear male sterility–facilitated recurrent selection 

approach is also being used to develop maintainer and 

restorer composite populations from which B and R lines are 

extracted. These populations are also shared with national programs 

to enable them to extract locally adapted B and R lines. 

Marker-aided selection, using STS markers found linked with 

Rf3 and Rf4 genes, is being used at IRRI to increase the efficiency 

of selecting fertility restorer lines for wild abortive 

(WA), Dissi, and Gambiaca cytoplasm inducing male sterility 

in rice. 

Six IRRI-bred TGMS lines—IR68301S, IR73727-23S, 

IR73834S, IR75589-31S, IR75589-41-13-17-15-3S, and 

IR75589-41-13-17-15-22S—have been found to have a low 

CSP and are being used to develop tropical two-line rice hybrids. 

Other tropical rice-growing countries (India, Vietnam, 

and the Philippines) are also developing TGMS lines. Markeraided 

selection has been used at IRRI to pyramid tms genes to 

develop more stable TGMS lines. 

Grain quality and biotic resistance of rice hybrids 

Acceptable grain quality of the commercial rice hybrids is essential 

to ensure profitability to hybrid rice farmers. Hybridity 

does not impair grain quality if the parents chosen to develop 

heterotic hybrids possess acceptable grain quality (Khush et 

al 1988). Therefore, a critical evaluation of parents for grain 

quality is necessary before using them in hybrid breeding. Similarly, 

hybrids showing strong heterosis for yield should be 

evaluated critically for grain quality before their release and 

commercialization. 

Biotic stress resistance of hybrids was determined by 

the resistance of their parental lines and evaluating whether 

this resistance was dominant or recessive. Hybrid vigor does 

not make rice hybrids more or less resistant than the parental 

lines (Cohen et al 2003). The most widely used WA-CMS 

system was not found associated with susceptibility to blast, 

bacterial blight, brown planthopper, and whitebacked 

planthopper. Nevertheless, the use of genetically diverse CMS 

systems is advocated to avoid potential genetic vulnerability 

of rice hybrids in the tropics where disease/insect pressure is 

high. 

Hybrid rice seed production technology 


Extensive research and seed production experiences in China, 

at IRRI, and in other countries have helped in identifying guidelines 

and practices for hybrid rice seed production, which are 

packaged in manuals (Virmani and Sharma 1993) and research 

papers (Mao 1988, Virmani 1996, Virmani et al 2002, Virmani 

and Kumar 2004). A seed production manual is also available 

in CD form and can be accessed from the IRRI Knowledge 

Bank at www.knowledgebank.irri.org/hybridRiceSeed/ 

hybridRiceSeed.htm. Using these practices, hybrid seed yields 

ranging from 0.7 to 4 t ha –1 (average 1–1.5 t ha –1 ) have been 

obtained in the tropics. Seed yields are getting higher as seed 

growers and their supervisors gain experience. 

The production of hybrid rice seed is labor-intensive but 

economically viable enough to attract seed companies. More 

than 60 seed companies in the public, private, and NGO sectors 

are working in Asia to produce and market hybrid rice 

seeds. Seed production operations (such as differential planting 

of male and female rows, roguing, flag-leaf clipping, GA 3 

application, separate harvesting and processing of pollen and 

seed parents, etc.) require at least 50 person-days of additional 

labor compared with normal nonmechanized inbred rice cultivation. 

Increasing area under hybrid rice creates increasing 

demand for hybrid rice seeds, resulting in large areas of hybrid 

seed production, which in turn creates additional rural 

employment opportunities in labor-surplus countries through 

the seed industry. In labor-scarce countries, such as Japan and 

the U.S., hybrid rice seed production technology is mechanized 

(Walton 2003, Takita 2003). 

Future opportunities 

Opportunities exist to enhance heterosis through indica/tropical 

japonica crosses since the level of heterosis of intervarietal 

groups is higher than in intravarietal groups (Virk et al 2003). 

Recent results at IRRI have shown stronger heterosis for yield 

in indica/new plant type (NPT) crosses in which the NPT parent 

was derived from indica/tropical japonica crosses (Table 

2). The identification and use of heterotic groups and gene 

blocks may also help to further enhance heterosis. Hybrid 

breeding efficiency is increased by using the TGMS system 

and nuclear male sterility–facilitated recurrent selection to 

breed parental lines. Also, molecular marker-aided selection 

helps in identifying fertility restorers, TGMS lines, and heterotic 

gene blocks. Transgenic parental lines possessing resistance 

to bacterial blight and stem borers can be developed by 


Table 2. Percent heterosis for yield in indica/NPT hybrids versus indica/indica hybrids 

evaluated in retestcross nursery at IRRI, 2004 wet season. 

Total no. % Percent yield advantage over 

Hybrid of hybrids heterotic best inbred check 

tested hybrids 

Range 

Mean 

Indica/indica 85 40.0 1.33– 81.40 30.57 ± 23.68 

Indica/NPT 40 47.5 5.81–131.40 42.7 ± 37.66 

using Xa21 and Bt genes to develop transgenic rice hybrids 

possessing resistance to these stresses. 

Early vegetative vigor and the better developed root system 

of hybrid rice make it adaptable to low-temperature-prone 

(during the boro season) and salinity-prone irrigated rice areas. 

Recent results at IRRI showed better adaptation of certain 

hybrids under alternate wetting-and-drying conditions and indicated 

prospects for increasing water-use efficiency through 

hybrids. Hybrids have also shown better adaptation to aerobic 

conditions (George et al 2002) and certain rainfed lowland 

conditions. Hence, the prospects for hybrid rice in these ecosystems 

should be explored. Brown rice of some rice hybrids 

was found to show a higher iron content in their grain than 

popular inbred HYVs and there was evidence of heterosis for 

this trait (Gregorio, pers. comm.); hence, commercial rice hybrids 

in and outside China should be critically evaluated for 

iron and zinc content in their brown and polished rice to find 

out whether any of these happen to combine high yield with 

high iron and zinc content. Continuous diversification of CMS 

systems of commercial rice hybrids would be a good strategy 

to avoid the risk of potential genetic vulnerability associated 

with some disease/insect problems. Apomixis is the ultimate 

genetic tool to fix heterosis in rice, which would make it widely 

used even by resource-poor farmers. Attempts should therefore 

be intensified to develop this tool using modern genetic 

tools. 

Heterosis in rice is useful for increasing farmers’ income, 

contributing to national food security, and increasing production 

efficiency. The associated labor-intensive hybrid seed production 

helps to create additional rural employment opportunities 

through the seed industry. 

References 

Cohen MB, Bernal CC, Virmani SS. 2003. Do rice hybrids have 

heterosis for insect resistance A study with Nilaparvata lugens 

(Hemiptera: Delphacidae) and Marasmia patnalis (Lepidoptera: 

Pyralidae). J. Econ. Entomol. 96(6):1935-1941. 

Duvick DN. 1999. Heterosis: feeding people and protecting natural 

resources. In: Coors JG, Pandey S, editors. The genetics and 

exploitation of heterosis in crops. Madison, Wis. (USA): 

American Society of Agronomy, Crop Science Society of 

America, Soil Science Society of America. p 19-29. 

George, T, Magbanua R, Laza M, Atlin G, Virmani S. 2002. Magat, 

a wetland semidwarf hybrid rice for high-yielding production 

on irrigated dryland. Int. Rice Res. Notes 27(1):26-28. 

Khush GS, Kumar I, Virmani SS. 1988. Grain quality of hybrid rice. 

In: Hybrid rice. Proceedings of the International Symposium 

on Hybrid Rice, Changsha, Hunan, China, 6-10 Oct. 1986. 

p 201-215. 

Ma GH, Yuan LP. 2003. Hybrid rice achievements and development 

in China. In: Virmani SS, Mao CX, Hardy B, editors. Hybrid 

rice for food security, poverty alleviation, and environmental 

protection. Proceedings of the 4th International Symposium 

on Hybrid Rice, Hanoi, Vietnam, 14-17 May 2002. Los Baños 

(Philippines): International Rice Research Institute. p 247- 

256. 

Mao CX. 1988. Hybrid rice seed production in China. In: Rice seed 

health. Manila (Philippines): International Rice Research Institute. 

p 277-282. 

Peng S, Yang J, Laza RC, Sanico AL, Visperas RM, Son TT. 2003. 

Physiological bases of heterosis and crop management strategies 

for hybrid rice in the tropics. In: Virmani SS, Mao CX, 

Hardy B, editors. Hybrid rice for food security, poverty alleviation, 

and environmental protection. Proceedings of the 4th 

International Symposium on Hybrid Rice, Hanoi, Vietnam, 

14-17 May 2002. Los Baños (Philippines): International Rice 

Research Institute. p 153-172. 

Takita T. 2003. Hybrid rice research and development in Japan. In: 

Virmani SS, Mao CX, Hardy B, editors. Hybrid rice for food 

security, poverty alleviation, and environmental protection. 

Proceedings of the 4th International Symposium on Hybrid 

Rice, Hanoi, Vietnam, 14-17 May 2002. Los Baños (Philippines): 


Virk PS, Khush GS, Virmani SS. 2003. Breeding strategies for enhancing 

heterosis in rice. In: Virmani, SS, Mao CX, Hardy B, 

editors. Hybrid rice for food security, poverty alleviation, and 

environmental protection. Proceedings of the 4th International 

Symposium on Hybrid Rice, Hanoi, Vietnam, 14-17 May 2002. 

Los Baños (Philippines): International Rice Research Institute. 

p 21-30. 

Virmani SS. 1996. Hybrid rice. Adv. Agron. 57:377-462. 

Virmani SS, Ilyas-Ahmed M. 2001. Environment-sensitive genic 

male sterility (EGMS) in crops. Adv. Agron. 72:139-195. 

Virmani SS, Kumar I. 2004. Development and use of hybrid rice 

technology to increase rice productivity in the tropics. Int. 

Rice Res. Notes 29(1):10-20. 

Virmani SS, Mao CX, Hardy B, editors. 2003. Hybrid rice for food 

security, poverty alleviation, and environmental protection. 

Proceedings of the 4th International Symposium on Hybrid 

Rice, Hanoi, Vietnam, 14-17 May 2002. Los Baños (Philippines): 

International Rice Research Institute. 407 p. 


Virmani SS, Mao CX, Toledo RS, Hossain M, Janaiah A. 2002. 

Hybrid rice seed production technology and its impact on seed 

industries and rural employment opportunities in Asia. Technical 

Bulletin 156. Taiwan (China): Food & Fertilizer Technology 

Center. 

Virmani SS, Sharma HL. 1993. Manual for hybrid rice seed production. 

Manila (Philippines): International Rice Research Institute. 

Walton M. 2003. Hybrid rice for mechanized agriculture. In: Virmani 

SS, Mao CX, Hardy B, editors. Hybrid rice for food security, 

poverty alleviation, and environmental protection. Proceedings 

of the 4th International Symposium on Hybrid Rice, 

Hanoi, Vietnam, 14-17 May 2002. Los Baños (Philippines): 


Notes 

Author’s address: Principal scientist (plant breeding), International 

Rice Research Institute, DAPO Box 7777, Metro Manila, 

Philippines, e-mail: s.virmani@cgiar.org. 

Harnessing molecular markers in hybrid rice 

commercialization in the Philippines 

E.D. Redoña, L.M. Perez, L.R. Hipolito, V.E. Elec, I.A. Pacada, L.M. Borines, R.O. Solis, S.A. Ordoñez, and J. Agarcio 

In 2002, commercialization of hybrid rice technology became 

Philippine agriculture’s banner program for attaining self-sufficiency 

and increasing productivity and profitability in rice, 

and generating rural employment. Four public hybrids (Magat, 

Mestizo, Mestiso 2, and Mestiso 3) and four proprietary hybrids—Magilas 

(Monsanto), SL8 (SL Agritech), Bigante 

(Bayer CropScience), and Rizalina 28 (HyRice)—were made 

available to farmers. With intensified training efforts, 40 seed 

growers’ cooperatives were formed that produce 60–70% of 

the program’s seed requirements. From 2001 to 2003, hybrids 

yielded 6.03 t ha –1 versus the 4.44 t ha –1 average for certified 

inbred seeds. Average seed production yields increased from 

A 

PR1A 

PR2A 

IR62829A 

JinanteA 

913A 

PMS8A 

PragathiA 

IR58025A 

IR68901A 

IR68886A 

IR68901A 

IR68896A 

IR68902A 

IR78371A 

IR70960A 

IR69626A 

IR68280A 

IR69627A 

IR69622A 

28A 

IR70963A 

LianA 

PhilRice 


IRRI 

YAU 

YAU 















YAU 


YAU 

1 

2 

3 

4 

5 

6 

7 

8 

9 

Cluster 1 

Cluster 2 

B 

0.00 0.25 0.50 0.75 1.00 

PR24067A a 

PR35731 a 

GinA c 

28 A c 

913 A c 

PMS8 A a 

IR58025A b 

IR62829A b 

IR66901A b 

IR68902A b 

IR69627A b 

IR70371A b 

IR16365-5 A b 

Lian A c 

IR16549A b 

IR17076A b 

BO A d 

IR68886A b 

IR69625A b 

IR69625A b 

IR70369A b 

IR72079A b 

PR35732A a 

IR72083A b 

IR70368A b 

IR72795A b 

PR35733A a 

PR35734A a 

PR35735A a 

PR35736A a 

PR35737A a 

PMS10 A c 

PR35738A a 

Pragathi A c 

MH841-21 A c 

IR68896A b 

IR69626A b 

IR68280A b 

IR16727-7A b 

IR69617A b 

IR16331A b 

IR16375A b 

IR69622A b 

PR35739A a 

IR68888A b 

IR70370A b 

1 

2 

3 

4 

5 

6 

7 

8 

C 

IR58025B 

85.1 

100 

IR78378B 

IR78375B 

IR78376B 

IR79123B 

IR78369B 

(59.2%) 

94.9 

IR68888B 

IR68897B 

IR70369B 

(69%) 

IR73328B 

PR3B 

97.2 

IR78367B 

(~58%) 

PR2B 

PR4B 

PR9B 

0.57 0.68 0.78 0.89 1.00 

Coefficient 



98.1 

(76.5%) 

100.0 

0.52 0.63 0.74 0.85 0.96 

IR68301s 

IR73827-23s 

IR73834s 

Fig. 1. Diversity analysis of hybrid rice germplasm. (A) CMS line dendrogram using 20 microsatellite markers, 25 RAPDs, and 10 +3/+3 

AFLP primer combinations (Redoña et al 1998). (B) Clustering of 46 diverse CMS lines based on 65 microsatellite loci. (C) Dendrogram of 

15 maintainer (B) lines based on 64 microsatellite loci and 255 allele types (Perez 2004). (D) Six TGMS lines clustered using 64 microsatellite 

loci and 129 allele types (Perez 2004). 

D 

(52.1%) 

(63.5%) 

Coefficient 

100.0 

(90.1%) 

(96.4%) 

TGMS1 

TGMS4 

TGMS6 



performance of the hybrids was highest for grain yield and 

percent spikelet fertility at 62% and 25.6%, respectively. However, 

the relationship between heterosis and genetic diversity 

was generally weak for most of the traits. 

Marker-aided pyramiding of bacterial blight (BB) 

resistance genes 

Some commercial hybrids in the Philippines, including the 

high-yielding and popular Mestizo, are susceptible to the 

Xanthomonas oryzae pv. oryzae (Xoo)-caused BB disease that 

is prevalent in the wet season. The disease can cause tremendous 

yield losses if severe infection occurs before flowering. 

To address this potential problem, single and/or pyramids of 

Xa4, Xa7, and Xa21 BB resistance genes were introgressed 

into five B lines of commercialized and promising hybrids— 

IR58025B, IR62829B, 913B, LianB, and BoB (Borines et al 

2003). B lines with 2–3 gene pyramids showed much shorter 

lesions in response to diagnostic Xoo races, indicating an increased 

and wider resistance spectrum against BB than B lines 

containing single genes. Xa4 and Xa7 genes together in different 

maintainer backgrounds imparted a complementary gene 


action in conferring a higher level of resistance to race 3- 

PXO340, and in LianB background to race 10-PXO341. 

In two-line hybrid breeding, three potential TGMS materials 

(F 2 plants no. 450, 473, and 700) homozygous for Xa4, 

Xa7, and Xa21 genes were identified in a segregating population 

developed from the cross TGMS1(–Xa)/AR32-19-3- 

3(+Xa21)/DiR32(+Xa4/+Xa7) (Perez 2004). STS PCR markers 

M5F/R and Xa21F/R were used to detect the presence of 

Xa7 and Xa21, respectively (Fig. 2A). These lines exhibited 

high degrees of resistance to Xoo race 1 (PXO61), race 2 

(PXO86), and race 6 (PXO99) in the greenhouse using an induced 

screening method and they are potential genetic stocks 

for the development of BB-resistant hybrids. 

Molecular tagging of TGMS genes in rice 

In hybrid rice breeding, screening for the TGMS trait is an 

intensive process influenced by environment. Tagging TGMS 

genes from new donors with DNA markers would increase the 

speed and efficiency of developing two-line hybrids. Based 

on a 301 microsatellite marker survey, four possible chromosomal 

locations (i.e., chromosomes 1, 2, 6, and 12) of the tgms 

gene from a new donor were identified by single-marker analysis 

and interval mapping procedures using One-Way ANOVA 

and Qgene software, respectively. The highest LOD score was 

6.48 between RM24 and RM113 on chromosome 1 (Fig. 2B). 

Most of the 11 microsatellite markers highly associated with 

sterility were located on chromosome 6. Three peaks were identified 

in this region with LOD scores >3.0, indicating the presence 

of a possible gene/QTL for the tgms gene in chromosome 

6 (Fig. 2B). Saturating this region with more markers and identifying 

tightly linked markers for use in TGMS line selection 

would expedite two-line hybrid development, thus further widening 

the genetic base of hybrids for the Philippine commercialization 

program. 

Hybrid seed purity analysis based on mitochondrial 

genome analysis 

Genetic purity in hybrid seeds is essential for full heterosis 

expression and is therefore a prerequisite for a successful commercialization 

program. Based on analysis of mitochondrial 

DNA sequences of CMS and fertile rice, the orf61-atp6-orf79 

region of the rice mitochondrial genome revealed two regions 

polymorphic among IR58025A (CMS), its B line (IR58025B), 

and hybrid (Mestizo). The a6P5 and MdF1 primers designed 

to amplify these regions produced 300-bp and 500-bp bands, 

respectively, that discriminated the B line from the CMS line 

and the hybrid. Successful detection of B line off-types in mixtures 

was demonstrated using the bulked DNA assay. This technique 

is starting to be used by the Philippine National Seed 

Quality Control Services for seed purity testing in the hybrid 

commercialization program. 

Conclusions 

DNA marker applications are finding relevant uses in addressing 

problems and challenges arising from commercializing 

hybrid rice in the Philippines. Genetic diversity analyses using 

markers have been useful for ensuring a diverse germplasm 

pool for hybrid breeding and avoiding the redundant use of 

similar genetic materials. While the utility of randomly selected 

markers for predicting heterosis cannot as yet be ascertained, 

the usefulness of markers in the tagging and pyramiding of 

useful genes in parental materials, and in seed purity analysis, 

has been demonstrated. As intellectual property protection 

becomes emphasized, with the recent passage of the Philippine 

Plant Variety Protection Law, DNA profiling and identification 

of diagnostic alleles for specific hybrid germplasm 

will find useful applications. An exciting but yet unsuccessful 

application of molecular markers is in the development of heterotic 

pools to enhance and sustain higher heterosis levels. As 

the Philippines needs to produce 65% more rice relative to 

current production by 2025, promising biotechnological tools 

such as molecular marker technology will continue to be used 

for increasing the speed, efficiency, and accuracy of hybrid 

rice breeding procedures while finding other relevant applications 

in other aspects of hybrid rice technology commercialization. 

References 

Borines LM, Redoña ED, Porter B, White F, Natural MP, Vera Cruz 

CM, Leung H. 2003. Marker-assisted pyramiding of bacterial 

blight resistance in parental lines of hybrid rice. In: Proceedings 

of the Arnel R. Hallauer International Symposium on Plant 

Breeding, Mexico City, Mexico. p 168-169. 

Ordoñez SA. 2003. Genetic variation of restorer lines in rice (Oryza 

sativa L.). Unpublished MS thesis. University of the Philippines 

Los Baños, Laguna, Philippines. 83 p. 

Perez LM. 2004. Development of bacterial blight-resistant 

thermosensitive genetic male sterile (TGMS) lines for hybrid 

rice (Oryza sativa L.) production. Unpublished MS thesis. 

University of the Philippines Los Baños, Laguna, Philippines. 

92 p. 

Redoña ED, Hipolito LR, Ocampo TD, Sebastian LS. 1998. Molecular 

polymorphism of rice cytoplasmic-genetic male sterile 

(CMS) lines based on AFLP, RAPD and microsatellite 

markers. Philipp. J. Crop Sci. 23(1):12-20. 

Notes 

Authors’ addresses: E.D. Redoña, L.M. Perez, L.R. Hipolito, V.E. 

Elec, I.A. Pacada, R.O. Solis, S.A. Ordoñez, and J. Agarcio, 

Philippine Rice Research Institute (PhilRice), Muñoz, Nueva 

Ecija, Philippines; L.M. Borines, Leyte State University 

(LSU), Baybay, Leyte, Philippines, e-mail: 

edredona@philrice.gov.ph. 


A 

M 

TGMS1 

IRBB4/7 

AR32-19-3-3 

IRBB7 

IRBB4 

IR24 

450 

473 

700 

M 

1,170 bp 

294 bp 

1,000 bp 

Chr 1 

Chr 6 

B 

RM128 

RM170 

RM259 

RM220 

RM275 

RM237 

RM1 

RM276 

RM5 

RM111 

RM140 

RM113 

RM3 

RM248 

RM24 

RM162 

RM102 

6.48 3.0 0.0 

LOD 

RM243 

3.0 0.0 

LOD 

RM50 

Fig. 2. Use of DNA markers in two-line hybrid breeding. (A) Genotyping of potential TGMS 

materials (F 2 plants no. 450, 473, and 700) showing the 294-bp allele for Xa7 (M5F/R primer) 

and 1,000-bp allele for Xa21 (Xa21F/R primer) for bacterial blight resistance in rice. TGMS1 is 

a susceptible parent, whereas IRBB4/7 (DiR32) and AR32-19-3-3 are donor parents for Xa4/ 

Xa7 and Xa21, respectively. IRBB7, IRBB4, and IR24 are check varieties (Perez 2004). (B) 

Possible location of tgms gene (LOD = 6.48) on chromosome 1. Three peaks with LOD>3.0 

found on chromosome 6 suggest the presence of a QTL. 


Genetic evolution of Rf1 locus for the fertility restorer gene 

of BT-type CMS rice 

Tomohiko Kazama and Kinya Toriyama 

Cytoplasmic male sterility (CMS) is a maternally inherited trait 

that results in the inability to produce fertile pollen. It is usually 

attributed to mitochondrial defects. In some cases, pollen 

fertility is recovered by a nuclear-encoded gene, the fertility 

restorer gene (Rf). Recently, Rf genes have been cloned in 

Petunia (Bentolila et al 2002) and kosena radish (Koizuka et 

al 2003). Both of the Rf genes have been shown to encode the 

PPR motif. PPR proteins contain a characteristic tandem array 

of 35 amino acid motifs, and therefore are termed a 

pentatricopeptide repeat (Small and Peeters 2000). The identification 

of Rf genes in Petunia and radish as PPR-containing 

genes suggests that searching for PPR motif genes near known 

restorer loci should be a useful strategy to identify Rf genes in 

other species. 

In rice, one particular CMS system has been obtained 

by combining the cytoplasm of Chinsurah Boro II with the 

nuclear genome of Taichung 65 (Shinjyo 1975). It is called 

BT-type. In this CMS line, pollen fertility can be restored 

gametophytically by the presence of a single dominant nuclear 

gene, Rf1. It has been reported that there are two copies of 

atp6 in BT-type mitochondria (Akagi et al 1994, Iwabuchi et 

al 1993, Kadowaki et al 1990). One is a normal atp6 (N-atp6), 

which is usually present in rice mitochondria. The other is an 

additional atp6 (B-atp6), which contains a unique sequence 

(orf79) located downstream of the normal atp6 sequence. 

We followed a positional cloning strategy to identify the 

Rf1 gene. After mapping of the Rf1 locus, we searched for 

ORFs containing a mitochondrial targeting signal and PPR 

motif in the genome of Nipponbare. Three such genes, named 

PPR8-1, PPR8-2, and PPR8-3, respectively, existed in a tandem 

array. The genomic fragments of each gene were obtained 

from cultivar Milyang23, whose genotype is Rf1/Rf1, and were 

introduced into a CMS line, BTA (rf1/rf1) (Kazama and 

Toriyama 2003). 

Here we report on the pollen and seed fertility in the 

complementation test. Restoration of fertility by the PPR8-1 

gene demonstrated that PPR8-1 was the Rf1 gene. Processing 

of the B-atp6 and N-atp6 RNA was examined in transgenic 

plants. In view of the sequence similarity of Rf1 to PPR8-2 

and PPR8-3, the genomic evolution of the Rf1 gene was discussed. 

Complementation tests and analysis of transgenic plants 

Each DNA fragment containing Rf1, PPR8-2, and PPR8-3 was 

introduced into a CMS line, BTA (rf1/rf1), by Agrobacteriummediated 

gene transfer. Pollen fertility was examined in 35 

A B C 

Fig. 1. Restoration of pollen and seed fertility in the complementation 

test. (A) Sterile pollen of a CMS line (BTA [ms-bo]rf1/rf1). 

(B) Pollen of the transgenic BTA line with the PPR8-1 gene. Half of 

the pollen grains were stained with I 2 -KI solution, indicating that 

the fertility of pollen grains with the PPR8-1 gene was restored. 

(C) Normal seed set of the transgenic BTA line with the PPR8-1 

gene. 

transgenic plants with the PPR8-1 gene. Twenty plants showed 

near 50% fertility, indicating that the transgene was segregated 

in the pollen, and pollen with PPR8-1 restored the fertility 

(Fig. 1). These transgenic plants with the PPR8-1 showed normal 

seed set (Fig. 1), demonstrating that the PPR8-1 gene restored 

the fertility. Fertility was not restored by PPR8-2 or 

PPR8-3 genes. These results demonstrated that PPR8-1 was 

the Rf1 gene. 

To reveal the possible interaction between Rf1 protein 

and the CMS-associated gene in mitochondria, we focused on 

the transcripts of orf79 that had been reported to be specific 

for Rf1 genotype and derived from the processing of the B- 

atp6 RNA. A portion of the 3′ end of B-atp6 containing orf79 

was used as a probe for Northern blot analysis of mature anthers. 

The transcripts of orf79 were detected in BTR (Rf1/Rf1) 

and F 1 (Rf1/rf1), but not in BTA (rf1/rf1), as previously reported. 

In the transgenic BTA with Rf1, the signal was detected. 

In contrast, the signal was not detected in the transgenic plants 

with PPR8-2 and PPR8-3. This result indicates that protein 

encoded by Rf1 is involved in processing of B-atp6 RNA. 

The N-atp6 RNA contains the 3′ noncoding region that 

may be cleaved by Rf1 protein. To investigate whether the N- 

atp6 RNA is processed by Rf1 protein or not, we performed 

RT-PCR analysis to detect specifically unprocessed N-atp6 

RNA. The unprocessed RNA was detected in BTA, transgenic 

BTA with PPR8-2, and transgenic BTA with PPR8-3. On the 

other hand, it was not detected in BTR (Rf1/Rf1) and transgenic 

BTA with Rf1, indicating that the primer site of the 3′ noncoding 

region in the N-atp6 was cleaved off by the action of Rf1 protein. 

This result indicates that Rf1 protein is also involved in 

N-atp6 RNA processing. 


Sequencing of each gene 

To compare each gene with Rf1, we sequenced each fragment 

of Milyang23. The ORFs of PPR8-2 and PPR8-3 are 93.2% 

and 93.7% identical to Rf1 at the nucleotide level, respectively 

(Table 1). The ORFs of Rf1, PPR8-2, and PPR8-3 encode 791 

aa, 683 aa, and 794 aa, respectively. The predicted proteins of 

PPR8-2 and PPR8-3 are 92.3% and 89.0% similar to that of 

Rf1, respectively. The number of PPR motifs is 18 for Rf1, 16 

for PPR8-2, and 18 for PPR8-3 (Table 1). These three genes 

are present in a tandem array in Milyang23 and appear to be 

originated from triplication in the coding region, although there 

is a deletion in the PPR8-2 coding region. PPR8-2 and PPR8- 

3 were shown to be expressed in anthers. 

Discussion 

Table 1. Similarity of Rf1 to PPR8-2 and PPR8-3. 

Gene No. of No. of Identity 

amino acids PPR motifs to Rf1(%) 

Rf1 791 18 – 

PPR8-2 683 16 93.2 

PPR8-3 794 18 93.7 

Rf1 protein is likely to directly regulate the processing event 

of B-atp6 RNA. The structure of the 3′ flanking region of N- 

atp6 is quite different from that of B-atp6, yet Rf1 protein 

promotes the processing of N-atp6 RNA. This indicates that 

Rf1 protein recognizes atp6 RNA structure rather than orf79 

RNA or the junction between atp6 and orf79. 

It is interesting that there are two highly homologous 

PPR genes adjacent to the Rf1 gene, and these genes do not 

contribute male fertility restoration and B-atp6 RNA processing. 

The high homology to Rf1 suggests that PPR8-2 and PPR8- 

3 protein may recognize a certain mitochondrial housekeeping 

gene RNA, in the same manner as Rf1 protein recognizes 

housekeeping atp6 RNA. It will be worthwhile to determine 

whether PPR8-2 and PPR8-3 protein affect the profile of mitochondrial 

housekeeping gene transcripts in transgenic plants. 

We propose the hypothesis that the Rf1 gene is evolved from 

highly homologous genes that regulate expression of certain 

housekeeping genes in mitochondria in order to reduce the 

expression of the CMS-associated gene. 



We used near-isogenic lines that differ at the Rf1 gene locus. 

These lines have been derived from a backcross of Chinsurah 

Boro II × Taichung 65. They are [normal] rf1/rf1 Taichung 65 

as a maintainer line (BTB), B 21 F 1 [ms-bo] rf1/rf1 as a CMS 

line (BTA), and B 12 F 8 [ms-bo] Rf1/Rf1 as a restorer line (BTR). 

The genotype of Nipponbare is rf1/rf1 and that of Milyang23 

is Rf1/Rf1. 

Northern blot analysis 

Total RNA was isolated from 100 mg of mature anthers and 

calli of BTA, BTR, F 1 (BTA × BTR), and transformants using 

the RNeasy Plant mini kit (Qiagen, Hilden, Germany), and 10 

µg of each RNA underwent Northern blot analysis. A part of 

the B-atp6 gene of BTR was amplified and labeled with 

digoxigenin using the PCR DIG probe synthesis kit. 

Reverse transcription (RT)-PCR analysis 

First-strand cDNA was synthesized with pd(N)6 Random 

Hexamer primer using the First-strand cDNA synthesis kit 

(Amersham Biosciences Corp., Piscataway, NJ, USA). The 

forward primer, primer-a, lies in the 5′ noncoding region of 

the N-atp6 gene. The reverse primer, primer-b, lies in the 3′ 

untranslated region of the N-atp6 gene. 

References 

Akagi H, Sakamoto M, Shinjo C, Shimada H, Fujimura T. 1994. A 

unique sequence located downstream from the rice mitochondrial 

atp6 may cause male sterility. Curr. Genet. 25:52-58. 

Bentolila S, Alfonso AA, Hanson MR. 2002. A pentatricopeptide 

repeat-containing gene restores fertility to cytoplasmic malesterile 

plants. Proc. Natl. Acad. Sci. USA 99:10887-10892. 

Iwabuchi M, Kyozuka J, Shimamoto K. 1993. Processing followed 

by complete editing of an altered mitochondrial atp6 RNA 

restores fertility of cytoplasmic male sterile rice. EMBO J. 

12:1437-1446. 

Kadowaki K, Suzuki K, Kazama S. 1990. A chimeric gene containing 

the 5′ portion of atp6 is associated with cytoplasmic malesterility 

of rice. Mol. Gen. Genet. 224:10-16. 

Kazama T, Toriyama K. 2003. A pentatricopeptide repeat-containing 

gene that promotes the processing of aberrant atp6 RNA 

of cytoplasmic male sterile rice. FEBS Lett. 544:99-102. 

Koizuka N, Imai R, Fujimoto H, Hayakawa T, Kimura Y, Kohno- 

Murase J, Sakai T, Kawasaki S, Imamura J. 2003. Genetic 

characterization of a pentatricopeptide repeat protein gene, 

orf687, that restores fertility in the cytoplasmic male-sterile 

Kosena radish. Plant J. 34:407- 415. 

Shinjyo C. 1975. Genetical studies of cytoplasmic male sterility and 

fertility restoration in rice, Oryza sativa L. Sci. Bull. Coll. 

Agric. Univ. Ryukyu 22:1-57. 

Small ID, Peeters N. 2000. The PPR motif—a TPR-related motif 

prevalent in plant organellar proteins. Trends Biochem. Sci. 

25:46-47. 

Notes 

Authors’ address: Graduate School of Agricultural Science, Tohoku 

University, e-mail: tomo-k@bios.tohoku.ac.jp. 

Acknowledgments: We thank Kazue Imataka for her helpful technical 

assistance for the transformation of rice plants. This study 

was partially supported by a grant-in-aid from the Ministry of 

Education, Science, and Culture, Japan, and by a grant for the 

Rice Genome Project from the Ministry of Agriculture, Forestry, 

and Fisheries, Japan. 


A polygenic balance model in yield components as revealed 

by QTL analysis in rice 

Wilhelm E. Hagiwara, Kazumitsu Onishi, Itsuro Takamure, and Yoshio Sano 

Under stabilizing selection, typical or intermediate phenotypes 

are favored, encouraging the development of hidden variation 

within and between populations (Mather and Jinks 1982). 

Based on the polygenic balance theory, repulsion phase linkage 

could be accumulated along a chromosome as demonstrated 

in Drosophila (Thoday 1960). Such a linked group of polygenes 

might be reshuffled through recombination, resulting in 

the formation of newly derived extreme phenotypes or transgression. 

Our study was conducted to gain insight into the genetic 

architecture of quantitative traits in grain characteristics 

between the two rice subspecies, the so-called indica and 

japonica types. The grain dimension is well differentiated between 

these two rice subspecies in a quantitative manner, affecting 

grain quality (Redoña and Mackill 1998). To enhance 

the power to detect linked QTLs with small effects, a large 

chromosome segment from an indica type was first introgressed 

by backcrosses into a japonica type of rice. Then, the 

introgressed segment was dissected by repeated self-pollinations 

after hybridization with the recurrent parent. The resulting 

recombinant inbred lines (RILs) are expected to enhance 

the efficiency to detect minor QTLs because of the reduction 

in the residual variation. We report here that a cluster of QTLs 

affecting grain characteristics is present on a chromosome 6 

segment showing repulsion and coupling linkages, and that 

transgressive segregation appears through recombination within 

the chromosome segment. 



T65wx was a near-isogenic line (NIL) of T65 (a japonica type 

of Oryza sativa from Taiwan) carrying wx (waxy) from 

Kinoshitamochi (japonica). T65Wx-pat was a NIL of T65 carrying 

a chromosome 6 segment of Patpaku, an indica type from 

Taiwan (Dung et al 1998). An F 1 plant between T65Wx-pat 

and T65wx was self-pollinated to give the F 2 generation, which 

was advanced to the F 7 generation using the single-seed descent 

(SSD) method. The plants were expected to carry different 

lengths of introgressed segments from Patpaku in the T65wx 

background and they were regarded as near-isogenic RILs. In 

total, 163 RILs were established and used in our experiment. 

Grain characteristics were measured in seeds (unhulled grains) 

and kernels (hulled grains) of the RILs (F 7 generation) and the 

parents. For each of the 163 RILs, two plants were examined. 

The average length and breadth (expressed in mm) from 6 seeds 

per plant and the weight (in grams) of 10 grains per plant were 

calculated. 

Genotyping with molecular markers 

A total of 17 PCR-based markers were used to determine the 

extent of the introgressed chromosome segments in the RILs. 

Nine of these markers (RM508, RM589, RM510, RM204, 

RM314, RM253, RM136, RM527, and RM3) were 

microsatellite markers (McCouch et al 2002), and one marker 

(C214) was a cleaved amplified polymorphic sequence (CAPS) 

marker (Rice Genome Research Program, http:// 

rgp.dna.affrc.go.jp). In addition, three PCR-based markers were 

designed from two RFLP (restriction fragment length polymorphism) 

markers, R2291 and S1520, and one STS (sequenced 

tagged site) marker, R2349. In addition, gene-specific 

markers were designed from the Wx, Se1, RFT1, and alk 

genes. 

Data analysis 

A linkage map was constructed using Map Manager QTX software 

(Manly et al 2001). To detect QTLs, we employed interval 

mapping and MQM mapping methods using MapQTL version 

4 (Van Ooijen et al 2002). The empirical LOD thresholds 

corresponding to the genome-wide significance at the 1% level 

were estimated by permutation tests, for each trait. The additive 

genetic effect, the percentage of variance explained by 

each QTL (PVE), and the total variance explained by all the 

QTLs affecting a trait were obtained with MapQTL in the final 

multiple-QTL model in which one cofactor marker was 

fixed per QTL. Two-way ANOVA was used to detect digenic 

interactions between QTLs. The markers with the highest LOD 

score found by MQM mapping for each QTL were assumed to 

be at the possible position of the QTL. 


Segregation for seed dimension in the RILs 

The seed and kernel were shorter and wider in T65wx than in 

T65Wx-pat, suggesting that genes for the traits might be located 

on the introgressed segment of chromosome 6. T65wx 

showed higher values in the weights of seeds and kernels (SW 

and KW) than did T65Wx-pat. Each of the six traits showed a 

continuous variation among the RILs. 

Detection of putative QTLs for grain characteristics 

The linkage map was constructed based on the segregation of 

17 markers in all the RILs, and had a length of 72.5 cM. In 

total, 10 QTLs responsible for the six traits were detected 

(Table 1). The percentage of the phenotypic variance explained 

(PVE) ranged from 6.8% for a QTL controlling seed breadth 

(SB) to 43.7% for a QTL controlling kernel length (KL). 


Table 1. QTLs for the six grain characteristics detected in the RILs based on the 

MQM mapping method. a Map position Effects on phenotype 

Trait NML (cM) 

LOD PVE a 

Seed length (SL) RFT1 12.6 8.82 18.4 –0.22 

S1520 19.3 4.77 9.4 0.17 

Hd1 51.8 7.04 14.3 0.14 

Seed breadth (SB) RFT1 12.6 6.77 14.4 –0.07 

S1520 19.3 3.38 6.8 0.06 

Hd1 51.8 5.43 11.3 –0.05 

Kernel length (KL) Hd1 51.8 21.30 43.7 0.16 

Kernel breadth (KB) RM3 62.2 10.41 24.4 –0.06 

Seed weight (SW) RM3 62.2 9.51 22.6 –0.01 

Kernel weight (KW) RM3 62.2 9.63 22.8 –0.01 

a 

NML indicates the marker locus nearest to the QTL or the marker with the highest LOD score peak. 

PVE is the percentage of the phenotypic variance explained. a is the additive effect of the Patpaku 

allele and the positive value indicates that the effect of the genotype of the indica parent (Patpaku) is 

in the direction of increasing the measured value of the trait. 

Three QTLs were detected for seed length (SL) and SB, 

at approximately the same locations, near the RFT1, S1520, 

and Hd1 markers (Table 1). Regarding the Patpaku-derived 

QTLs near RFT1, S1520, and Hd1, it should be noted that the 

QTL near RFT1 decreased both SL and SB, while that near 

S1520 increased both of them and the QTL near Hd1 increased 

SL but decreased SB (Table 1). The detection of the effect of 

each QTL on SL depended on the genotypes of the other two 

QTLs (Table 2); however, it could be due to a small number of 

some genotypes because two-way ANOVA failed to detect any 

significant interactions between QTLs for SL. For SB, twoway 

ANOVA detected distinct epistatic effects (interactions) 

between the RFT1 and Hd1 markers (F = 14.64, P = 0.0002) 

as well as between the Hd1 and S1520 markers (F = 7.92, P = 

0.0055), suggesting that the effect of each QTL also depended 

highly on the genotypes of the other linked QTLs for SB (Table 

2). For seed weight, only one QTL was detected near the RM3 

marker. 

In contrast to the characteristics of seeds, the length and 

breadth of kernels were each controlled by a single QTL with 

high LOD scores and PVE (Table 1). For KL, one QTL was 

detected near the Hd1 marker, and that for KB was detected 

near RM3, showing their slightly different position (Table 1). 

The Patpaku allele increased KL, whereas the T65wx allele 

increased KB. For KW, a single QTL was also detected near 

RM3. 

The detection of multiple QTLs of small effects in our 

experiments is possibly related to the elimination of effects of 

segregating genes on other chromosomes. The variance caused 

by segregation on other chromosome regions increases the residual 

variance for the interval under consideration, reducing 

the power for QTL detection (Lynch and Walsh 1998). Consequently, 

multiple segregating QTLs tend to reduce the efficiency 

to detect significant associations between phenotype 

and genotype in the whole-genome QTL mapping. 

Hidden genetic variation 

Actually, the RILs with the highest or lowest values in SL and 

SB resulted from an accumulation of QTLs with positive or 

negative effects (Table 2). Thus, the transgressive segregation 

revealed that a hidden genetic variation was concealed in the 

introgressed segment even though the phenotypic difference 

in grain characteristics was slight between the parental lines. 

Linked QTLs with different signs in effects are suggested to 

occur alternately along a chromosome under the polygenic 

balance model (Mather 1943). Such linked QTLs on a chromosome 

behave as an effective factor in segregating populations, 

although the polygenic balance model is an extreme 

metaphor (Lynch and Walsh 1998). This leads us to consider 

that adaptive traits might be created in part through the reconstruction 

of gene combinations in the flanking region using 

preexisting variation rather than newly derived mutations. Some 

agriculturally valuable traits might also have been reconstructed 

in this way, as predicted from frequent transgressions. 

References 

Dung LV, Inukai T, Sano Y. 1998. Dissection of a major QTL for 

photoperiod sensitivity in rice: its association with a gene 

expressed in an age-dependent manner. Theor. Appl. Genet. 

97:714-720. 

Lynch M, Walsh B. 1998. Genetics and analysis of quantitative traits. 

Sinauer Associates, Inc., Sunderland. 

Manly KF, Cudmore Jr RH, Meer JM. 2001. Map Manager QTX, 

cross-platform software for genetic mapping. Mammalian 

Genome 12:930-932. 

Mather K, Jinks JL. 1982. Biometrical genetics. 3rd ed. New York, 

NY (USA): Chapman and Hall. 

Mather K. 1943. Polygenic inheritance and natural selection. Biol. 

Rev. 18:32-65. 


Table 2. Phenotypic effects of each marker allele depending on the genotypes of the 

other two markers (background) for seed length (SL) and breadth (SB). a 

Genotype 

Trait (in mm) 

Number 

Background Marker SL t-test SB t-test of plants 

allele 

Hd1(T); RFT1(T) S1520(T) 6.57 

ns 

3.48 * 76 

S1520(P) 6.80 3.63 3 

Hd1(T); RFT1(P) S1520(T) 5.95 ** 3.30 

ns 

7 

S1520(P) 6.43 3.37 20 

Hd1(P); RFT1(T) S1520(T) 6.80 

ns 

3.34 

ns 

17 

S1520(P) 6.93 3.38 4 

Hd1(P); RFT1(P) S1520(T) 6.38 * 3.11 ** 2 

S1520(P) 6.74 3.37 34 

S1520(T); RFT1(T) Hd1(T) 6.57 ** 3.48 ** 76 

Hd1(P) 6.80 3.34 17 

S1520(T); RFT1(P) Hd1(T) 5.95 

ns 

3.30 

ns 

7 

Hd1(P) 6.38 3.11 2 

S1520(P); RFT1(T) Hd1(T) 6.80 

ns 

3.63 * 3 

Hd1(P) 6.93 3.38 4 

S1520(P); RFT1(P) Hd1(T) 6.43 ** 3.37 

ns 

20 

Hd1(P) 6.74 3.37 34 

Hd1(T); S1520(T) RFT1(T) 6.57 ** 3.48 ** 76 

RFT1(P) 5.95 3.30 7 

Hd1(T); S1520(P) RFT1(T) 6.80 

ns 

3.63 ** 3 

RFT1(P) 6.43 3.37 20 

Hd1(P); S1520(T) RFT1(T) 6.80 ** 3.34 ** 17 

RFT1(P) 6.38 3.11 2 

Hd1(P); S1520(P) RFT1(T) 6.93 

ns 

3.38 

ns 

4 

RFT1(P) 6.74 3.37 34 

a (T) and (P) refer to the T65wx- and Patpaku-derived alleles, respectively. SL (seed length) and SB (seed 

breadth) were measured in mm. The statistical difference was based on the t-test. ns = not significant. * 

and ** show significance at 5% and 1% probability, respectively. 

McCouch SR, Teytelman L, Xu Y, Lobos KB, Clare K, Walton M, 

Fu BY, Maghirang R, Li ZK, Xing YZ, Zhang QF, Kono I, 

Yano M, Fjellstrom R, DeClerck G, Schneider D, Cartinhour 

S, Ware D, Stein L. 2002. Development and mapping of 2240 

new SSR markers for rice (Oryza sativa L.). DNA Res. 9:199- 

207. 

Redoña ED, Mackill DJ. 1998. Quantitative trait locus analysis for 

rice panicle and grain characteristics. Theor. Appl. Genet. 

96:957-963. 

Thoday JM. 1960. Effects of disruptive selection. III. Coupling and 

repulsion. Heredity 14:35-49. 

Van Ooijen JW, Boer MP, Jansen RC, Maliepaard C. 2002. 

MapQTL® 4.0, software for the calculation of QTL position 

on genetic maps. Wageningen (Netherlands): Plant Research 

International. 

Notes 

Authors’ address: Plant Breeding Laboratory, Graduate School of 

Agriculture, Hokkaido University, Sapporo 060-8589, Japan, 

e-mail: rysano@abs.agr.hokudai.ac.jp. 


Three conveners, Y. Fukuta (Japan International Institute for Agricultural 

Sciences, JIRCAS), D. Mackill (International Rice Research 

Institute, IRRI), and R. Ikeda (JIRCAS), organized Session 5, 

chaired by B. Mishra, Directorate of Rice Research, Indian Council 

of Agricultural Research. Three scientists presented papers on 

broadening the gene pool and two on exploiting heterosis in cultivated 

rice. 

B.S. Pinheiro, of Embrapa Arroz e Feijão, discussed “Aerobic 

rice development in Brazil.” She summarized upland breeding 

research under Brazilian savannas from the mid-1970s to 2003. 

In the mid-1980s, upland rice varieties of a tropical japonica 

background were grown in the Brazilian savannas. Afterward, crop 

area gradually and then markedly decreased. However, production 

did not decline at the same level because average yield 

doubled from 1986 to 2003. This was due to both crop migration 

toward areas with more favorable rainfall distribution and 

the adoption of modern plant-type varieties, tropical japonica × 

indica derivatives. The impact of improved plant-type varieties 

and associated technologies in favored localities of the savanna 

region gave rise to a new concept of upland rice, leading to a 


new designation—aerobic rice. Drought was a strong priority in 

the first breeding program, which relied on tropical japonica progenitors. 

In the next step, the genotypes to be grown under supplementary 

irrigation and in favored microregions for rainfall distribution 

were developed using indica varieties in crosses. A noticeable 

move from the frontier land to more favored areas for 

water distribution led to a decrease in priority for drought tolerance. 

Breeders and pathologists were instead able to cooperate 

strongly for breeding for blast resistance and grain quality. The 

introduction of indica germplasm to the genetic base of the traditional 

upland tropical japonica population was combined with 

strong selection pressure for grain characteristics. The first aerobic 

rice releases did not possess the same level of drought tolerance 

as the traditional upland varieties derived from crosses only 

within the japonica group. However, the program now possesses 

a representative number of elite lines, japonica versus indica 

derivatives, providing a sound basis to further improve yield, blast 

resistance, and drought tolerance. The alteration of plant type 

and grain quality of upland rice, toward its evolution to aerobic 

rice, led to an increase in yield potential and national market 

acceptance. Increased involvement in grain cropping systems, 

under no tillage or minimum tillage, is foreseen. Moreover, aerobic 

rice could greatly contribute to environmental sustainability, 

thus avoiding new savanna deforestation, when used in the ricepasture 

association, to renew the extensive areas of degraded 

pasture in the Brazilian savannas. 

H. Satoh of Kyushu University talked about “Mutation in 

seed reserves and its use for improving grain quality in rice,” 

improvement of grain quality being one of the most important 

subjects in rice breeding. For this purpose, genetic sources with 

different endosperm properties must be collected, characterized, 

and evaluated. Although a wide variation in seed storage starch 

and seed storage protein properties is found among cultivated 

rice and has been used widely in rice varieties, for example, variation 

in cooking quality between japonica and indica rice, only a 

few genes involved in this variation have been identified so far. 

By using a treatment of egg cells of rice fertilized with MNU, 

several thousand rice mutants were found for embryo or endosperm 

properties. In addition, a wide variation originated spontaneously 

in seed reserves was detected in local rice germplasm. 

The genetics and effects of specific genes on qualitative and 

quantitative changes in carbohydrates, proteins, or lipids in the 

rice endosperm have also been observed. The possibility of using 

mutants and germplasm was discussed for endosperm breeding. 

D.S. Brar of IRRI analyzed “Broadening the gene pool of 

rice through introgression from wild species.” Wild species are 

an important reservoir of genes for tolerance of biotic and abiotic 

stresses. However, several genetic barriers, including low crossability, 

sterility, and reduced homoeologous pairing/recombination, 

limit the transfer of genes from wild species into cultivated 

rice (Oryza sativa L.). Applications of embryo rescue, anther culture, 

molecular markers, and genomic in situ hybridization (GISH) 

facilitate alien introgression. Genes for resistance to brown 

planthopper, bacterial blight, blast, tungro, and grassy stunt; tolerance 

of acid sulfate conditions; and cytoplasmic male sterility 

have been introgressed from different wild species (AA, BBCC, 

CC, CCDD, EE, FF genomes). Breeding lines carrying alien genes 

for resistance to and tolerance of these traits have been released 

as varieties. Examples are AS996 from IR64 × O. rufipogon in 

Vietnam, NSICRC112 with high yielding ability from O. sativa × 

O. longistaminata in the Philippines, and Matatag 9 from IR64 × 

O. rufipogon for tungro-prone areas of the Philippines. Advanced 

progenies are under evaluation for tolerance of stem borer, nematodes, 

and aluminum and iron toxicity. O. rufipogon accessions 

resistant to sheath blight have been identified. O. sativa × O. 

glaberrima progenies are being evaluated for weed competitive 

ability. Some of the introgressed alien genes have been mapped. 

A major QTL for tolerance of aluminum toxicity introgressed from 

O. rufipogon has been mapped on chromosome 3, and is conserved 

across other cereals. 

S.S. Virmani of IRRI covered “Heterosis in rice for increasing 

yield, production efficiency, and rural employment opportunities.” 

IRRI started hybrid rice research in 1979 for the tropics 

collaboratively with interested national agricultural research and 

extension systems. With an appropriate choice of parents, a 15– 

20% yield advantage (1–1.5 t ha –1 ) from hybrids was also obtained 

in the tropics. Heterosis for early maturity and yield resulted 

in increased per-day productivity of hybrids, which also 

showed increased nitrogen-use efficiency and better adaptability 

than inbreds to saline soils, the aerobic rice environment, the 

alternate wetting-and-drying irrigation system, and more favorable 

rainfed lowland ecosystems. Increased leaf area index and 

increased grain number per panicle and grain weight contributed 

to the higher yield of hybrids. IRRI has facilitated their commercialization 

in the tropics by developing and freely sharing genetically 

diverse CMS, maintainer, and restorer lines (for 3-line hybrids) 

and temperature-sensitive genetic male sterile (TGMS) lines 

(for 2-line hybrids). IRRI-bred hybrids and/or parental lines resulted 

in coverage of about 1 million ha under hybrid rice in Vietnam, 

India, the Philippines, Bangladesh, Indonesia, and Myanmar. 

Labor-intensive seed production and processing helped generate 

additional rural employment opportunities through the seed industry. 

Through hybrid rice research at IRRI, progress in and contributions 

of heterosis breeding were shown. Future work should 

focus on increasing yield and stability, grain quality, disease/insect 

resistance, and seed production efficiency. Genetic analysis 

for each character under heterosis was also indicated. 

E.D. Redoña, Philippine Rice Research Institute (PhilRice), 

analyzed “The commercialization of hybrid rice in the Philippines.” 

Commercialization of this technology has become the Philippine 

government’s banner program for agriculture. Hybrid rice is being 

pursued to attain rice self-sufficiency, increase rice productivity 

and profitability, and generate additional employment. Progress 

has been made in areas such as breeding, seed production, seed 

industry development, and capacity building since 1998. Four 

public hybrids and four proprietary hybrids were provided to farmers. 

With increased efforts in enhancing seed production capacity, 

with the assistance of various training courses, 21 hybrid 

seed growers’ cooperatives have been established that now produce 

about 70% of the program’s seed requirements. From 2001 

to 2003, hybrids attained an average yield of 6.03 t ha –1 across 

provinces and seasons versus the 4.44 t ha –1 average of certi- 


fied inbred seeds, for a 34% advantage. Among the key factors 

related to the progress made are strong support at the policy 

level, engagement of the private sector, increased capacity enhancement 

activities, front-line demonstrations, development of 

production support, rewards and marketing assistance systems, 

involvement of local government units, and a massive information 

campaign. A successful communication and extension system 

for hybrid rice breeding was also demonstrated. 

In summary, the potential of wide hybridization (between 

indica and japonica, or between cultivated and wild rice), and 

mutants and germplasm for different endosperm traits, was discussed 

for broadening of the gene pool and application to rice 

breeding. In particular, the exploitation of heterosis through hybrid 

rice production was considered a good example of broadening 

the gene pool to achieve higher yield. 


SESSION 6 

Trends in crop establishment 

and management in Asia 

CONVENER: M. Yamauchi (NARO) 

CO-CONVENER: D. Johnson (IRRI)

Trends in crop establishment methods 

in Asia and research issues 

Sushil Pandey and Lourdes Velasco 

Economic factors and recent changes in rice production technology 

are the major drivers that have led to shifts from transplanting 

to the direct-seeding method of rice establishment in 

Asia. The rising cost of labor, the need to intensify rice production 

through double and triple cropping, the development 

of high-yielding short-duration modern varieties, and the availability 

of chemical weed control methods have jointly led to 

this switchover. As the rice production systems of Asia undergo 

further adjustments in response to the rising scarcity of 

land, water, and labor, pressure will increase for a shift toward 

direct-seeding methods. This paper provides a brief overview 

of the trends in crop establishment methods in Asia, their impact, 

and implications for research and technology development. 

The direct-seeded area in Asia is estimated to be about 

29 million ha, which is approximately 21% of the total rice 

area in Asia (Table 1). This includes upland and submergenceprone 

ecosystems where direct seeding is the traditional 

method. If these ecosystems are excluded, the direct-seeded 

area is approximately 15 million ha. The shift toward direct 

seeding occurred during the late 1980s to mid-1990s, mainly 

in rapidly growing economies such as Malaysia and Thailand 

as well as in countries where rapid intensification of the rice 

production system took place (for example, Vietnam). However, 

the growth rate in direct-seeded rice area appears to have 

slowed down considerably in recent years. 

Determinants of adoption of alternative 

crop establishment methods 

At a general level, the availability of water and the opportunity 

cost of labor can be considered the major determinants of 

Table 1. Direct-seeded rice area by ecosystem in various Asian countries. 

Irrigated + Irrigated + Area direct- 

Flood-prone rainfed lowland rainfed lowland seeded as a % Total rice Total area % of total 

Country + upland area a area direct- of irrigated + area a direct-seeded area directrice 

area a seeded rainfed seeded 

(million ha) lowland area (million ha) 

South Asia 8.4 47.8 6.3 13 56.2 14.9 26 

Bangladesh 1.9 8.8 0.1 1 10.7 2.0 19 

India 6.5 36.0 5.5 15 42.5 12.0 28 

Pakistan 2.1 2.1 

Sri Lanka 0.9 0.7 78 0.9 0.7 77 

Southeast Asia 4.0 68.2 8.3–10.3 11–14 72.2 12.3–14.3 17–20 

Cambodia 0.2 1.7 1.9 0.2 10 

China 0.5 31.6 1–2.5 3–8 32.1 1.5–3 5–9 

Indonesia 1.2 9.8 0.8 8 11.0 2.0 18 

Laos 0.2 0.4 0.6 0.2 33 

Malaysia 0.1 0.6 0.4 67 0.7 0.5 71 

Myanmar 0.6 5.7 6.3 0.6 9 

Philippines 0.2 3.4 1.3 38 3.6 1.5 42 

Thailand 0.5 9.1 2.8 31 9.6 3.3 34 

Vietnam 0.5 5.9 2–2.5 34–42 6.4 2.5–3 39–47 

East Asia 3.2 0.1 3 3.2 0.1 3 

Japan 2.1 2.1 

Korea 1.1 0.1 9 1.1 0.1 9 

Total 12.4 119.2 14.7–16.7 12–14 131.6 27.3–29.3 21–22 

a 

Huke and Huke (1997). Sources of direct-seeded area: Bangladesh—Huke and Huke (1997) and personal communication with S. Bhuiyan. India— 

Palaniappan and Purushothaman (1991). Sri Lanka—Pathinayake et al (1991). Cambodia—Helmers (1997). China—personal communication with Lu 

Ping. Indonesia—Huke and Huke (1997) and personal communication with Dr. Hamdane Pane. Malaysia—Huke and Huke (1997) and own estimate. 

Myanmar—Huke and Huke (1997) and own estimate. Philippines—PhilRice-BAS (1995) and own estimate. Thailand—personal communication with 

Dr. Booribon Somrith and data from Agricultural Extension Office, Khon Kaen. Japan—personal communication with Yujiro Hayami. Vietnam—personal 

communication with T.P. Tuong and Government Statistical Office (1997). Korea—Kim (1995). 


Water 

availability 

High 

Low 

TPR 

DSR/TPR 

Low 

Fig. 1. Wage rate and water availability as determinants of crop 

establishment methods. DRS = direct-seeded rice, TPR = transplanted 

rice. 

crop establishment methods (Fig. 1). A low wage rate and assured 

supply of adequate water are conditions favorable for 

transplanting. Incentives for direct seeding increase when water 

availability is low (or uncertain) and wage rates are high. 

Much of the recent spread of direct seeding in Southeast Asian 

countries has been in response to rising wage rates. When water 

availability is low (or uncertain) and the wage rate is low, 

either the direct-seeding or transplanting method can be used, 

depending on field hydrological conditions. 

Two major types of adjustments in crop establishment 

methods have been made in response to the rising cost of labor. 

In temperate Asian countries such as Japan, Korea, and 

Taiwan (China), the shortage of farm labor led to a change 

from manual to mechanical transplanting. On the other hand, 

in tropical countries such as Malaysia and Thailand, the labor 

shortage induced a shift to direct seeding. Small farm size, 

intensive cultivation of rice, a long history of transplanting 

culture, and the relatively high price of rice in East Asian countries 

partly explain their adherence to transplanting. 

Impact of the shift to direct seeding 

Wage rate 

High 

DSR/TPR 

DSR 

Farm-level studies have shown that direct-seeding methods 

produce higher income than transplanting (Table 2). Despite a 

slightly lower average yield than that of transplanted rice under 

farmers’ field conditions, direct-seeded rice has been found 

to result in higher net profit as the savings in labor cost outweighs 

the value of lost output. This has occurred especially 

in areas where the cost of labor has risen rapidly in relation to 

the price of rice. In addition, total farm income has increased 

where direct seeding facilitated double cropping of rice in areas 

that had only one crop of transplanted rice previously (such 

as the Mekong Delta, Vietnam, and Iloilo, Philippines). 

The likely future trend 

Despite the rapid spread of direct seeding in several Southeast 

Asian countries, transplanting is still the dominant method of 

crop establishment. Because of differences in rice production 

systems and economic conditions, it is more convenient to 

examine the likely scenario for East, Southeast, and South Asia 

separately. 

In East Asia, where rice production systems are inputintensive, 

a major shift to direct seeding in response to a further 

escalation in the wage rate is unlikely to occur. In comparison 

with other Asian countries, these countries are more 

industrialized and have higher per capita incomes. Farm income 

is maintained at a higher level through policies that keep 

the price of rice high in comparison with the international 

market price. Under this situation, a potential threat to farmer 

income because of a wage increase is likely to be addressed by 

policy changes that compensate farmers for a loss in profits 

rather than by changes in the method of crop establishment. 

Direct seeding is likely to expand further in Southeast 

Asian countries with low population densities, especially in 

areas where the cost of labor is escalating. However, this type 

of wage-induced shift is a function of the rate of growth of the 

economy. A slowdown in economic growth rate because of 

macroeconomic constraints may result in some shift back to 

transplanting, such as what happened during the Asian economic 

crisis of 1997. In very densely populated areas such as 

Java, western China, and the Red River Delta of Vietnam, transplanting 

is likely to continue to remain the dominant culture in 

the near to medium term. 

In South Asia, where population density is high and overall 

economic growth is slow, economic incentives for a shift to 

direct seeding are likely to remain weak. The pattern of adoption 

of direct seeding in Southeast Asia shows that direct seeding 

is first adopted in the dry season, because of better water 

control, rather than in the wet season. In South Asia, dry-season 

rice accounts for only about 12% of the total rice area 

versus 22% in Southeast Asia. In addition, the overall proportion 

of rainfed rice area in South Asia is higher. These features 

of rice systems may result in a slower adoption of direct seeding 

in South Asia. 

Research implications 

The primary economic motives for a shift to direct seeding are 

the savings in labor cost and the possibility of crop intensification. 

The priority research issue depends on which of the 

two motives is likely to play the dominant role in a particular 

ecoregion. If the main driving force for the transition to direct 

seeding is the rapidly rising wage rate, research to generate 

labor-saving technological innovations would have a high priority, 

such as mechanical tillage and labor-saving weed control 

methods. Where drought and early submergence impede 

the adoption of direct seeding, research to develop varieties 

Session 6: Trends in crop establishment and management in Asia 179

Table 2. Economic returns to different crop establishment methods in the wet 

season. 

Place Dry-seeded Wet-seeded Transplanted % Difference 

($ ha –1 ) 

Suphanburi, Thailand a 

Cash cost b 152 148 3 

Gross returns c 505 476 6 

Gross margin d 353 328 8 

Net returns e 168 132 27 

Pangasinan, Philippines f 

Cash cost 230 273 –16 

Gross returns 608 666 –9 

Gross margin 378 393 –4 

Net returns 288 247 17 

Iloilo (double cropped), Philippines g 

Cash cost 736 441 67 

Gross returns 1,627 904 80 

Gross margin 891 464 92 

Net returns 695 382 82 

a Source: Isvilanonda (1990). All values converted to $ using the exchange rate US$1 = B20. b Cost of 

all purchased inputs. c Gross value of output. d Gross value of output minus the cost of purchased inputs. 

e Gross value of output minus the cost of purchased and family-owned inputs. f Pandey et al (1995). 

g Pandey and Velasco (1998). The comparison is between two crops of wet-seeded rice versus only one 

crop of transplanted rice. 

and crop management practices to relax these constraints is 

needed. 

However, if crop intensification is the major reason for 

direct seeding, research to facilitate early establishment and 

early harvest of the direct-seeded crop would have a higher 

priority, as this will permit timely planting of the subsequent 

crop. The development of short-duration varieties would be 

important in this case. Even though the cost of labor may be 

initially low in these areas, intensification of land use may lead 

to labor shortages because of peak labor demand during the 

harvesting of the previous crop and establishment of the succeeding 

crop within a short period. Suitable mechanical devices 

for land preparation that can reduce the turnaround time 

between crops could help achieve a higher and more stable 

yield of the second crop. 

The high costs of weed control could be a major factor 

constraining the widespread adoption of direct seeding. The 

key to the success of direct-seeded rice is the availability of 

efficient weed control techniques. Varieties with early seedling 

vigor and crop management technologies that help reduce 

the competitive effects of weeds on crops are needed. It is 

essential, however, to evaluate the environmental and health 

consequences of potential technologies that are based on chemical 

means of weed control. 

Empirical analyses have indicated that the technical efficiency 

of rice production is lower and more variable for direct-seeded 

rice than for transplanted rice (Pandey and Velasco 

1999). This suggests the existence of a higher “yield gap” between 

the “best-practice” farmer and the average farmer when 

rice is direct-seeded. A greater variability of technical efficiency 

of direct-seeded rice could be partly due to the use of varieties 

that were originally developed for transplanted culture. Varieties 

that are specifically targeted for direct-seeded methods 

could help reduce such yield gaps. Better crop management 

practices, especially those that facilitate early and more uniform 

establishment, can be similarly helpful. 

Precise water management is critical for high productivity 

of wet-seeded rice (De Datta and Nantasomsaran 1991). A 

high level of control of water flow on irrigated fields is hence 

desirable. However, most irrigation systems in Asia have been 

designed to supply water to transplanted rice for which precision 

in water management is not as critical. Suitable modifications 

of irrigation infrastructure could ensure a high yield of 

direct-seeded rice and also improve water-use efficiency. In 

addition, appropriate mechanical systems of field leveling that 

ensure uniformity in field water level are needed. 

References 

De Datta SK, Nantasomsaran P. 1991. Status and prospects of direct 

seeded flooded rice in tropical Asia. In: Direct seeded flooded 

rice in the tropics. Selected papers from the International Rice 

Research Conference. Manila (Philippines): International Rice 


Helmers K. 1997. Rice in the Cambodian economy: past and present. 

In: Nesbitt HJ, editor. Rice production in Cambodia. Manila 



Huke RE, Huke EH. 1997. Rice area by type of culture: South, Southeast, 

and East Asia. Manila (Philippines): International Rice 

Research Institute. 59 p. 

Isvilanonda S. 1990. Effects of pregerminated direct seeding technique 

on factor use and the economic performance of rice 

farming: a case study in an irrigated area of Suphan Buri. In: 

Fujimoto A, editor. Thai rice farming in transition. Tokyo (Japan): 

World Planning Commission. 

Kim SC. 1995. Weed control technology of direct seeded rice in 

Korea. Paper presented at the International Symposium on 

Weed Control under Direct Seeded Rice, 31 July 1995, 

Omagari, Akita, Japan. 

Palaniappan SP, Purushothaman S. 1991. Rainfed lowland rice farming 

system in Tamil Nadu (India): status and future thrust. In: 

Proceedings of the Rainfed Lowland Rice Farming Systems 

Research Planning Meeting, Myanmar, August 1991. Manila 

(Philippines): International Rice Research Institute. 

Pandey S, Velasco LE. 1998. Economics of direct-seeded rice in 

Iloilo: lessons from nearly two decades of adoption. Social 

Sciences Division Discussion Paper. Manila (Philippines): 


Pandey S, Velasco LE. 1999. Economics of alternative rice establishment 

methods in Asia: a strategic analysis. Social Sciences 

Division Discussion Paper. Manila (Philippines): International 

Rice Research Institute. 

Pandey S, Velasco L, Masicat P, Gagalac F. 1995. An economic analysis 

of rice establishment methods in Pangasinan, Central 

Luzon. Social Sciences Division Discussion Paper. Manila 


Pathinayake BD, Nugaliyadde L, Sandanayake CA. 1991. Direct 

seeding practices for rice. In: Sri Lanka in direct-seeded 

flooded rice in the tropics. Manila (Philippines): International 


PhilRice-BAS (Bureau of Agricultural Statistics). 1995. Provincial 

rice statistics. Muñoz, Nueva Ecija (Philippines): Philippine 


Notes 

Authors’ address: International Rice Research Institute, e-mail: 

sushil.pandey@cgiar.org. 

Direct seeding and weed management in Korea 

Soon-Chul Kim and Woon-Goo Ha 

In 1991, the Rural Development Administration (RDA) introduced 

dry drill seeding and water seeding of rice with water 

seeding being either wet-drill-sown or water broadcast-seeded. 

Despite problems, the direct-seeding area has increased rapidly; 

however, only about 8–9% of the total rice area was under 

direct seeding in 2003 (RDA 2003). The main constraints 

to direct seeding are weed problems and rainfall fluctuation 

during seeding time. 

Weed ecology in direct-seeded paddy 

Weed growth is greatly influenced by cultivation method (Kim 

et al 1992) and thus the adoption of direct-seeding technology 

results in shifts in weed growth in terms of dry matter production 

and composition of dominant species. The greatest weed 

growth is recorded in dry-seeded rice, followed by waterseeded 

rice, mechanically transplanted rice, and manually transplanted 

rice (Table 1). 

Early rice growth and water management are considered 

as the main factors accounting for the above differences. Yield 

losses from weed competition are related closely to weed 

growth, with complete crop loss occurring where weeds are 

not managed. Changes in crop establishment method from 

transplanting to direct seeding changed not only weed dry 

matter production but also the composition of the weed species. 

In transplanted rice fields, the most troublesome weeds 

were Eleocharis kuroguwai (20%), Sagittaria trifolia (16%), 

S. pygmaea (13%), Echinochloa crus-galli (12%), and 

Monochoria vaginalis (11%) (Kim et al 1992). Approximately 

50% of weed biomass was accounted for by perennial weeds. 

In dry-seeded rice, on the other hand, annual grasses were the 

most predominant weeds, such as E. crus-galli (54%), Digitaria 

adscendens (9%), Leptochloa chinensis (8%), Setaria viridis 

(6%), and weedy rice (Oryza sativa subsp. spontanea, 5%). In 

water seeding, E. crus-galli (37%) was also the most important 

weed species even though the degree of dominance was 

lower than in dry-seeded rice. The second most important weed 

was Aneilema keisak (15%), followed by Ludwigia prostrata 

(11%) and Leersia japonica (7%). Weed growth in directseeded 

rice was dominated by E. crus-galli, which became the 

most important weed and had a slightly lower value in waterseeded 

rice than in dry-seeded rice. Most of the annual grass 

weeds occurring in direct-seeded rice fields belonged to the 

C 4 photosynthetic pathway (Table 2). 

Recently, weedy rice (including red rice) has become 

widely distributed in farmers’ fields. The possible origins of 

weedy rice were summarized by Kim (1995b) as follows: 

Shattered grains of the previous year 

Outcross between cultivated rice species 

Outcross between cultivated rice and red rice (dormant 

seeds) 

Outcross between wild-type rice and weedy rice (red 

rice) 

Outcross between red rice species 

Dormant red rice or wild-type rice itself 

The degree of contamination by weedy rice observed in 

fields was highly variable, ranging from 0.5% to 35.2% for 

the southern area. However, it was difficult to estimate the 

correct degree of contamination and distribution pattern for 

all farmers’ fields because field observations were made and 


Table 1. Weed growth and yield losses caused by weed and rice 

competition in various cultivation methods. 

Cultivation method Weed weight Yield loss 

(g m –2 ) a (%) 

Transplanting 

Manual (40–45 days old) 741 10–20 

Mechanical 

Old seedlings (30–35 days old) 843 25–30 

Young seedlings (8–10 days old) 1,020 30–35 

Direct seeding 

Water seeding 1,643 40–60 

Dry seeding 2,300 70–100 

a 

Weeds were harvested in the nonweeded plot. 

Table 2. Dominant weed species associated with establishment 

method. a 

Transplanting Dry seeding Water seeding 

El. kuroguwai (20) b Ec. crus-galli (54) Ec. crus-galli (37) 

Sa. trifolia (16) D. adscendens (9) An. keisak (15) 

Sa. pygmaea (13) Lp. chinensis (8) Lu. prostrata (11) 

Ec. crus-galli (12) Se. viridis (6) Lr. japonica (7) 

M. vaginalis (11) O. sativa subsp. spontanea (5) El. kuroguwa i(6) 

Sc. juncoides (6) Ae. indica (4) Sc. juncoides (6) 

Lu. prostrata (4) An. keisak (3) M. vaginalis (5) 

An. keisak (4) Pl. hydropiper (3) Pt. distinctus (4) 

Cy. serotinus (3) Al. aequalis var. amurensis (3) Sa. pygmaea (3) 

Pt. distinctus (3) Ca. bursa-pastoris (3) Cy. serotinus (3) 

Diversity 0.118 0.317 0.190 

Index 

a 

El = Eleocharis, Sa = Sagittaria, Ec = Echinochloa, M = Monochoria, Sc = Scirpus, 

Lu = Ludwigia, An = Aneilema, Cy = Cyperus, Pt = Potamogeton, D = Digitaria, Lp 

= Leptochloa, Se = Setaria, Ae = Aeschynomene, Pl = Polygonum, Al = Alopecurus, 

Ca = Capsella, Lr = Leersia, O = Oryza. b Values in parentheses are % relative biomass. 

Source: Kim et al (1992). 

sampling done only in fields that had conspicuous rice plants 

that were markedly different from cultivated rice in heading 

date and plant height. Sometimes, hybrid swarms of many intergrades 

existed in the same place and thus it was impossible 

to collect all the variations. 

The taxonomic traits of weedy rice were highly variable 

in grain shape, awn length, awn color, apiculus color, pericarp 

color, and endosperm type. Similarly, the important agronomic 

traits of weedy rice were also highly variable, including maturity, 

plant height, spikelet fertility, grain shattering, and dormancy 

(Kim 1995a). Because of introgression, most of these 

traits exhibited discrete characters. In dry-seeded rice, early 

seeding enhanced the growth of weedy rice. Double-cropped 

dry-seeded rice with barley, rye, or Italian ryegrass significantly 

reduced the growth of weedy rice by 33–66%. 

Sulfonylurea herbicide was first introduced in 1989 in 

rice paddy fields, starting with bensulfuron-methyl, followed 

by pyrazosulfuron-ethyl (Kim 1994a). In 2004, nine sulfonylurea 

herbicides, including these two, were widely used and 

about 75% of rice herbicides (117 sulfonylurea mixtures /155 

in total herbicides) were mixtures of sulfonylurea herbicides 

(KCPA 2004). This was equivalent to about 63.3% of the total 

market value. Since 1999, six weed species were reported as 

sulfonylurea-resistant, exhibiting cross resistance among sulfonylurea 

herbicides. Monochoria korsakowii was the first, 

reported in 1999, followed by M. vaginalis, Lindernia dubia, 

Rotala indica, Scirpus juncoides, and Cyperus difformis (Im 

et al 2004). 

Several herbicides effective for controlling sulfonylurearesistant 

weeds were screened. Before seeding or rice transplanting, 

oxadiazon, thiobencarb, fentrazamide, oxadiargyl, and 

butachlor were highly effective, whereas those for application 

at the 2–3-leaf stage of weeds were simetryn, carfentrazonethyl, 

pyriminobac-methyl, and benzobicyclon. 

Effective weed control 

Dry-seeded rice 

Good seedling establishment, proper weed management, and 

freedom from lodging are important for good yield. However, 

weed control is the key factor in dry-seeded rice and thus the 

success of direct seeding is almost totally reliant on effective 

weed control. Dry seeding has a dry period (up to about 30 

days from seeding) and subsequent permanent flood period. 

Most weeds occur during the dry period and Echinochloa species 

are of prime importance in this period. Echinochloa species 

grow faster than rice because of the different temperature 

regime for seed germination and different physiological and 

morphological traits (Kim and Moody 1989). Differences between 

rice and Echinochloa species in the period of seedling 

emergence were greater at early seeding than at late seeding 

and this difference can be important in deciding optimum herbicide 

application time. For example, seedling emergence at 

early seeding (1 April) required 36 days for rice emergence 

and 20 days for Echinochloa, whereas for late seeding of June 

20 these were 9 days and 5 days, respectively. 

Weed control should be practiced before 20 days after 

seeding and, among five herbicide applications, herbicidal efficacy 

was the greatest at the 12–15 DAS application (rice 

emergence stage). 

Seedling emergence of rice and Echinochloa species 

varies by seeding time and moisture regime and, thus, deciding 

on herbicide application time should be based on seedling 

growth status, not on days from seeding. 

In the permanent flood period, the application of oneshot 

herbicide is recommended either as early or intermediate 

application after flooding. Early applications (0–10 days after 

flooding) consist mostly of soil-applied herbicides for grass 

weeds such as butachlor, thiobencarb, benfuresate, dithiopyr, 

anilofos, molinate, esprocarb, pyrazolate, pentoxazone, and 

several mixtures of these with sulfonylurea, bensulfuron-methyl, 

pyrazosulfuron-ethyl, and imazosulfuron. At intermediate 

application (15–25 days after flooding), the majority are 


one-shot sulfonylurea mixtures for controlling annual grasses 

and perennial weeds. For late application (30–40 days after 

flooding), mostly, foliar-applied herbicide mixtures are recommended 

using bispyribac-sodium, cyhalofop-butyl, 

penoxaprop-P-ethyl, pyribenzoxim with propanil, bentazon, 

azimsulfuron, ethoxysulfuron, etc. 

For dry-seeded rice, in general, two herbicide applications 

are recommended: one at the dry period either just before 

rice emergence or just after rice emergence and the other 

at the flood period. 

Wet-seeded rice 

The newly developed “press-type” wet seed drill is an improved 

technology that avoids the problem of lodging for the soilsurface 

seeding method or poor seedling establishment for the 

subsoil seeding method (Kim 1994b). However, this technology 

has some problems for seedling establishment because of 

the difficulty in uniform harrowing operation, water management 

after seeding, herbicide application, bird damage, and 

algae infestation. 

The field should not be flooded for at least 6–8 days 

after seeding to provide enough oxygen for seed germination 

and enable good root anchorage. Furthermore, soil drying is 

necessary to have good root anchorage and good seedling establishment 

and to minimize the infestation of fresh-water algae. 

The improved technology of seed pressing and no irrigation 

during 6–8 days after seeding allowed all types of seed 

regimes, whether pregerminated or not. Good root anchorage 

significantly reduces the incidence of herbicidal phytotoxicity, 

particularly to sulfonylurea mixtures such as dimepiperate/ 

bensulfuron-methyl or molinate/pyrazosulfuronethyl that are 

currently being recommended in the wet/water-seeding method. 

One-shot herbicides of sulfonylurea mixtures are currently 

recommended after good root anchorage. Effective weed 

control has mostly relied on good tillage operations (particularly 

on harrowing operations) and the degree of root anchorage 

when the first herbicide is applied. In general, herbicide 

selection for wet seeding where there is flooding is relatively 

narrow compared with dry seeding because of the better rice 

root anchor regime at herbicide application time. Therefore, 

herbicide selection should be undertaken with care in wet seeding. 

For early applications (10–20 days after seeding), single 

or sulfonylurea mixtures of molinate, dimepiperate, dymron, 

fenclorim, pyrazolate, mefenacet, cyhalofop-butyl, 

pyriminobac-methyl are recommended. For late applications 

(30–40 days after seeding), herbicide recommendations are 

the same as for dry seeding. 

In general, pregerminated seeds are much safer than 

ungerminated seeds to most herbicides. This suggests that 

pregerminated seeds should be used and roots should be well 

anchored with shoot emergence through the soil when the first 

herbicide is applied. Based on these results, two herbicide recommendations, 

either a one-shot single application or two systematic 

applications, are suggested. The choice of herbicide 

application is mainly determined by the weed flora in the target 

field. 

Water-seeded rice 

Water seeding requires good management at the farmers’ level 

in terms of irrigation management to ensure a good stand and 

have no lodging problems. Farmers usually make a small irrigation/drainage 

canal in the field at regular intervals using tractor 

passing or a small machine or manually pulled stone. The 

surface water after seeding is drained out through this canal to 

allow good anchorage by the root system, thus ensuring a good 

seedling stand, good lodging tolerance, and less herbicidal 

phytotoxicity. The general recommendations for herbicide 

application are similar to those of wet seeding, except that 

thiobencarb can be used before rice seeding. Besides this, early 

and intermediate applications (10–20 days after seeding) and 

late applications (30–40 days after seeding) are the same as in 

wet seeding. 

The relative importance of weed control in the total rice 

labor requirement is greatest in direct-seeded rice, and is 8– 

10% for transplanted rice versus 13–15% for direct-seeded 

rice. This requirement is likely to surpass 25% when technologies 

such as minimum or no-tillage technology and large-scale 

mechanization are adopted. Currently, weed control practices 

are too reliant on herbicide. Therefore, an integrated weed 

management system is needed in the future. 

The basic concepts for integrated weed control combine 

the possible control methods of agrochemical, physical, ecological, 

and biological methods. These can be applied repeatedly 

without risking harm to the environment, and they may 

be more economical and require less input. 

In direct seeding, the weedy rice problem could have 

prime importance. Therefore, thorough and systematic research 

is needed. Weeds resistant to sulfonylurea are also an important 

consideration, in both direct seeding and rice cultivation 

as a whole. 

Reference 

Im IB, Kim JK, Kim SK, Kuk YI, 2004. Control of sulfonylurea 

resistant Lindernia dubia (L.) Pennell in the rice paddy field. 

Korean J. Weed Sci. 24(1):7-13. 

KCPA (Korea Crop Protection Association). 2004. Manual of agrochemicals. 

Seoul (Korea): KCPA. 989 p 

Kim SC.1995a. Latest advancement in cultivation methods and weed 

management of rice production in Korea. Weeds of Kyushu 

25:29-36 

Kim SC.1995b. Weed control in the rice crop. In: Labor saving culture 

of rice crop. Text for ‘95 training of extension workers. 

Seoul (Korea): Rural Development Administration. p 63-86. 

Kim SC. 1994a. The success of dry seeding of rice relies on weed 

management. Agric. Tech. RDA 335:1-4. 

Kim SC.1994b. Improvement of seedling establishment in wet drillseeded 

rice. Res. Ext. RDA 35(3):44-47. 

Kim SC, Oh YJ, Kwon YW. 1992. Weed flora of agricultural area in 

Korea. Korean J. Weed Sci. 12(4):317-334. 


Kim SC, Moody K. 1989. Germination and seedling development 

of rice and Echinochloa species. Korean J. Weed Sci. 9(2):108- 

115. 

RDA (Rural Development Administration). 2003. Reference for rice 

crop. Korea: RDA. 216 p. 

Notes 

Authors’ address: National Yeongnam Agricultural Research Institute, 

RDA, Milyang 627-130, Korea, e-mail: 

kim0sc@rda.go.kr. 

Direct-seeding cultivation of rice in Japan: stabilization 

of seedling establishment and improvement 

of lodging resistance 

Satoshi Yoshinaga 

The aging of farmers and falling rice prices are creating severe 

problems for farmers in Japan, and this has led to demand 

for a reduction in the costs and labor required for rice (Oryza 

sativa L.) production in recent years. Direct seeding, which 

does not require seedlings to be raised or transplanted, is regarded 

as the most effective method of reducing costs and labor. 

However, direct-seeded fields currently represent less than 

1% of the total rice area, although dry-seeded fields increased 

during the 1970s before the use of transplanting machines became 

widespread. One disadvantage of direct seeding, hindering 

its widespread use, is the instability of rice yields of direct-seeded 

fields compared with those of transplanted fields. 

The main factors responsible for the lower yields in wet-seeded 

fields are poor seedling establishment and frequent plant lodging. 

Recently, improved methods that stabilize seedling establishment 

and promote lodging resistance have been developed 

and the area of wet-seeded rice cultivation has doubled within 

the past 5 years. This paper discusses recent improvements 

that stabilize seedling establishment and lodging resistance in 

wet-seeded rice. 

Stabilization of seedling establishment 

In Japan, most popular rice cultivars with high taste quality do 

not show sufficient lodging resistance under wet-seeded conditions, 

as they are bred for transplanting. Seeding such cultivars 

on the soil surface often causes severe lodging during the 

ripening period. This disadvantage can be minimized by seeding 

in the soil, but lower and slower rates of seedling emergence 

result from the anaerobic conditions around the seed. 

Furthermore, direct seeding in Japan is usually conducted in 

periods of low temperature of less than 20 °C daily mean temperature, 

which may result in lower and slower rates of seedling 

emergence. Solutions to these problems are important to 

further develop wet-seeding cultivation and recent research 

has focused on stabilizing seedling establishment in the soil. 

Development of oxygen-generating products 

Ota and Nakayama (1970) reported that coating rice seeds with 

calcium peroxide (CaO 2 ), acting as a source of oxygen, pro- 

motes the seedling emergence of direct-seeded rice in flooded 

soil. Experiments on the practical use of CaO 2 in the 1970s 

contributed to the development of an oxygen-generating product 

containing CaO 2 and its commercial release (Calper Fine 

Granule). Thereafter, pregerminated seeds coated with the 

oxygen-generating product (simply called “coated seeds” hereafter) 

could be sown in soil at depths of 5–20 mm in submerged 

conditions. 

Water management after seeding 

Until the mid-1990s, fields were flooded after direct seeding 

because it was believed that flooding was necessary to maintain 

a more constant and higher soil temperature during seedling 

establishment. At that time, to control weeds, herbicide 

was usually applied under flooded conditions. However, seedling 

establishment was unstable with this type of water management, 

even when coated seeds were used. Oba (1997) suggested 

that drainage after seeding enhanced seedling establishment 

compared with flooding for rice coated and directseeded 

in the soil. Several reports (Yoshinaga et al 2000, Sato 

and Maruyama 2002, Tsuchiya et al 2004) on the effects of 

drainage after seeding show that oxidized soil conditions surrounding 

the seeds improved seedling establishment and stimulated 

initial growth (Table 1). It has also been shown that the 

time taken for seedling emergence in drained conditions is similar 

to that in flooded conditions, although the minimum soil 

temperature during emergence may be lower in drained conditions. 

Furthermore, improved herbicides that can control weeds 

at a later growth stage were developed and their use has supported 

the widespread drainage after seeding because improved 

herbicides can control weeds even if they are applied after 

rice seedling emergence subsequent to drainage. For these reasons, 

drainage subsequent to the seeding of coated seeds had 

become widespread by the end of the 1990s and it has contributed 

to the recent increases in the area cultivated by wet seeding. 

Further experiments are required to decrease the dose of 

CaO 2 or replace the use of CaO 2 and reduce costs by developing 

cultivars with high emergence ability at low temperature. 


Table 1. Relationship between water management and seedling emergence 

in field conditions. a 

Water Soil Seedling Floating Dry weight 

management temperature emergence seedlings (mg plant –1 ) 

after seeding (°C) (%) (%) 

Flooding 25.8 a 67.1 b 6.0 a 43.2 b 

Drainage 25.2 a 76.5 a 0.2 b 63.3 a 

a Seeds were coated with calcium peroxide and seeded in the soil at 10-mm depth. Data are 

means of 2–3 years. Soil temperature is the average of 2 weeks after seeding at 10-mm depth. 

Dry weight was measured at 2–3 weeks after seeding. Values followed by the same letter are 

not significantly different at the 5% level according to LSD. (Yoshinaga et al 2000.) 

Improvement of lodging resistance 

Lodging is problematic for rice production as it makes harvesting 

by machine difficult. Furthermore, the lodging of rice 

plants during the ripening period results in a yield reduction 

owing to decreased canopy photosynthesis from self-shading 

and also decreases grain quality because of the increased coloring 

of brown rice and/or decreased taste (Matsue et al 1991, 

Setter et al 1997). Wet-seeded rice is more susceptible to lodging 

during the ripening period than is transplanted rice. Popular 

cultivars are used in direct seeding to achieve better producer 

prices, although they do not generally show sufficient 

lodging resistance. Therefore, alternative means of improving 

the lodging resistance of direct-seeded rice are required such 

as amending nitrogen application, seeding depth, and seeding 

method. A decrease in nitrogen fertilization limits culm elongation 

and improves lodging resistance. The smaller amount 

of nitrogen application, however, might reduce yield in directseeded 

rice. Deeper seeding depths have been tested in conjunction 

with the use of coated seeds, and this has reduced the 

root lodging of direct-seeded rice. The base of the rice hills, 

however, is still shallow in these conditions and lodging resistance 

is insufficient. When comparing different seeding methods, 

lodging resistance is considered to be the highest in hillseeded 

rice and higher in row-seeded rice than in broadcastseeded 

rice (Ogata and Matsue 1998). However, practical hillseeders 

did not exist until the late 1990s, and previously direct 

seeding had been conducted by broadcast or row seeding. 

A practical hill-seeder was developed in the late 1990s. 

The hill-seeder, combined with a harrow, enabled seeding and 

puddling to be performed simultaneously (Shimotsubo and 

Togashi 1996). The seeder effectively drives seeds into the 

puddled soil intermittently by using a rotating disk, and the 

established seedlings form hills. The greater seeding depth 

effectively prevents the floating and lodging of seedlings, while 

hill seeding effectively increases lodging resistance. Hillseeded 

rice has a greater resistance to lodging than broadcastseeded 

rice across a range of plant densities (Fig. 1) or seeding 

depths (Yoshinaga et al 2001). The higher lodging resistance 

of hill-seeded rice is due to its larger number of panicles 

per hill, which is composed of several plants (Yoshinaga et al 

2001). 

Hill seeding has enabled popular cultivars with low lodging 

resistance to be grown in wet-seeding areas. The area cultivated 

by the hill-seeder has been increasing since 1998 and 

overtook the area cultivated by broadcasting in 2002. Hill seeding 

occupied 24% of the total area of wet-seeded rice in 2003 

and it can be said that the hill-seeder has contributed to the 

recent widespread application of wet seeding. 

To further improve lodging resistance, the introduction 

of cultivars with improved lodging resistance is important together 

with studies on the interaction of nitrogen application 

with seeding methods. 

Conclusions 

The recent improvements in the stability of establishment under 

direct seeding have contributed to the recent increase in 

direct-seeded area. For the further stabilization and widespread 

adoption of direct seeding, developments in plant breeding are 

required to allow the introduction of cultivars with desirable 

traits for direct-seeding cultivation. Improved crop management 

that corresponds to seeding methods and cultivars is also 

required. 

References 

Matsue Y, Mizuta K, Furuno K, Yoshida T. 1991. Studies on palatability 

of rice grown in northern Kyushu. Jpn. J. Crop Sci. 

60:490-496. 

Oba S. 1997. Seedling emergence improving method by draining 

(just after seeding) in wet seeding cultivation of rice. Nogyo 

Gijutsu 52:33-34. 

Ogata T, Matsue Y. 1998. Studies on direct sowing culture of rice in 

northern Kyushu. Jpn. J. Crop Sci. 67:485-491. 

Ota Y, Nakayama M. 1970. Effects of seed coating with calcium 

peroxide on germination under submerged conditions in rice 

plant. Proc. Crop Soc. Jpn. 39:535-536. 

Sato T, Maruyama S. 2002. Seedling emergence and establishment 

under drained conditions in rice direct-sown into puddled and 

leveled soil. Plant Prod. Sci. 5:71-76. 


Pushing resistance (g culm –1 ) 

100 

80 

60 

40 

20 

1998 BS 

1999 BS 

1998 HS 

1999 HS 

0 

0 50 100 150 200 

Plant density (no. m –2 ) 

Fig. 1. Relationship between plant density and pushing resistance in 

broadcast-seeded rice (BS) and hill-seeded rice (HS). Pushing resistance 

was measured at 20 days after heading by bending a hill 45 o 

from the vertical at 15-cm height. Vertical bars indicate standard 

error (n = 3). (Yoshinaga et al 2001.) 

Setter TL, Laureles EV, Marazedo AM. 1997. Lodging reduces yield 

of rice by self-shading and reductions in canopy photosynthesis. 


Shimotsubo K, Togashi T. 1996. Study on submerged direct seeding 

combined puddling with seeding operation. Jpn. J. Crop Sci. 

65(Extra issue 1):12-13. 

Tsuchiya M, Sato T, Maruyama S. 2004. Growth enhancement by 

drainage during seedling establishment in rice direct-sown into 

puddled and leveled soil. Plant Prod. Sci. 7:324-328. 

Yoshinaga S, Nishida M, Wakimoto K, Tasaka K, Matsushima K, 

Togashi T, Shimotsubo K. 2000. Effects of drainage after submerged 

direct-seeding on the effectiveness of fertilized nitrogen 

and the growth and yield of rice plants. Jpn. J. Crop Sci. 

69:481-486. 

Yoshinaga S, Wakimoto K, Tasaka K, Matsushima K, Togashi T, 

Shimotsubo K. 2001. Improvement of lodging resistance in 

submerged direct seeding rice cultivation using a newly developed 

’Shooting hill-seeder’. Jpn. J. Crop Sci. 70:186-193. 

Notes 

Author’s address: National Agricultural Research Center for Tohoku 

Region, NARO, Japan, e-mail: yosinaga@affrc.go.jp. 

Direct seeding of aerobic rice in China 

Guang Hui Xie, Jun Yu, Jing Yan, Huaqi Wang, and Xiurong Zhu 

The food supply in China must increase by 50–60% over the 

next 30 years for a population that is estimated to reach 1.6 

billion by 2030. Rice is vital to more than half of the population 

and it is the most important field crop in China. However, 

lowland rice cultivation area dropped from 33.76 million ha 

in 1981 to 29.96 million ha in 2000 mainly because of water 

shortages. Current projections suggest that by 2030 there will 

be a shortfall of 12.9 billion m 3 of water to meet the demand 

of the country (Li and Duan 2002). Since rice is the most water-consuming 

crop, alternative rice cultivation strategies that 

require less water and increase water productivity are urgent. 

Variety improvement and water saving of aerobic rice 

Many traditional upland rice types have been grown in hilly 

rainfed regions in China for hundreds of years with very low 


inputs. However, these varieties produce a low yield even under 

relatively high-input conditions. During the past two decades, 

rice breeders have worked at variety improvement and 

more and more high-yielding and good-quality rice varieties 

adapted to aerobic soil conditions have been emerging and 

released officially. Most of these varieties were claimed to have 

a high yield potential, from 6.0 to 8.0 t ha –1 . Since the mid- 

1980s, a new way of rice cultivation took place in northern 

China: high-yielding rice seeded directly in aerobic nonpuddled 

and nonflooded soils (Bouman 2001). This high-input system 

is replacing lowland rice in areas where water scarcity makes 

lowland rice impossible. 

Like wheat or maize, aerobic rice is grown on land prepared 

without water: a fine and firm seedbed is established 

with a wide range of soil water content. It is seeded either in 

holes or in strips. A shallow irrigation is applied immediately 

after sowing if the soil is too dry for seed to germinate and 

emerge. During the whole growth duration, no standing water 

is required. Consequently, irrigation water for aerobic rice is 

less than 50–70% of that for lowland rice in northern China 

(Wang et al 2002). 

Distribution of aerobic rice in China 

According to incomplete statistics, during a few years, the cultivation 

area of aerobic rice reached around 190,000 ha in 2000, 

constituting only 0.6% of China’s total rice cultivation area 

(Wang et al 2002). Because water shortages are more severe 

in the northern part of the country than in the southern, aerobic 

rice is mainly distributed in regions of northern and northeastern 

China, with 80,000 and 60,000 ha, respectively. Lowland 

rice area in northern China, which includes Beijing and 

Tianjin municipalities, and Hebei, Shandong, and Henan provinces, 

decreased dramatically by 17.5% from 1980 to 2001. 

Northeastern China, comprising Liaoning, Jilin, and 

Heilongjiang provinces, is the leading region for producing 

good-quality japonica rice, for which cultivation area dropped 

by 5.1% from 1990 to 2001. However, Liaoning is the only 

province suited to growing aerobic rice climatically. In Jilin 

and Heilongjiang, located to the north of Liaoning, the frostfree 

period is too short to grow aerobic rice, which needs direct 

seeding instead of raising seedlings in a greenhouse. In 

regions of central and southwestern China, where rainfall is 

more than in the north regions, about 50,000 ha of aerobic rice 

are grown. 

Direct seeding of aerobic rice in northern China 

Aerobic rice is grown soon after winter wheat or the vegetable 

harvest in the region, with cumulative temperature ranging from 

4,000 to 5,000 °C and precipitation from 580 to 1,000 mm. 

Varieties Han Dao 277, Han Dao 297, Han Dao 65, Wushi 

Han Dao, Xiahan 51, and Han 946 are popularly or potentially 

cultivated. Optimal seeding dates have been documented from 

1 to 10 June. It is not recommended to sow aerobic rice after 

15 June, after which yield decreases by 324 kg ha –1 for every 

day late. For most cases, seeding rates in the region vary from 

120 to 150 kg ha –1 , and rice is sown in strips 25–30 cm apart 

and 2–3 cm deep, as breeders recommended (Wang et al 2001). 

An amount of 60–80 kg N ha –1 is applied as basal, with a total 

of 115–150 kg N ha –1 for most farmers. Generally, two other 

nitrogen splits (40–60 and 20–40 kg N ha –1 ) are placed at 

tillering and elongation stages, respectively. All phosphorus 

and potassium fertilizers are applied as basal at rates of 75– 

120 kg P 2 O 5 ha –1 and 20–30 kg K 2 O ha –1 ; however, in many 

cases, farmers do not use any potassium fertilizer. Irrigation is 

essential for relatively high yield. In the region, a shallow irrigation 

is conducted soon after sowing because of no irrigation 

during a very short period between harvest and seeding. Irrigation 

at the tillering stage is not required in most cases since 

it coincides with the rainy season owing to summer monsoon 

activities, until it lasts 15–20 days without available rain for 

sandy soils and 20–30 days for clay soils (Fan et al 2004). 

However, it was found that aerobic rice showed more tolerance 

of water stress at an early stage than at a late stage. Enough 

soil moisture during panicle forming and heading stages must 

be maintained, and irrigation is recommended for every 10 

days without rain (Fan et al 2004). Weed control is more important 

for direct seeding aerobic rice than for transplanting 

rice. Oxadiazon, butachlor, thiobencarb, and quinclorac are 

used as preemergence herbicides, and a mixture of butachlor 

and oxadiazon has been recommended and accepted widely. 

Propanil, butachlor, bentazone, and oxadiazon or their combinations 

are frequently applied for postemergence herbicides. 

Direct seeding of aerobic rice in northeastern China 

In northeastern China, aerobic rice is mainly grown as a single 

crop in Liaoning Province, where cumulative temperature fluctuates 

from 2,731 to 3,674.3 °C, and precipitation is 350–1,100 

mm. Han Dao 297, Han 946, Dangeng 8, and Danhandao 1 

are the predominant varieties and they are seeded in strips at 

112.5 kg ha –1 directly at the end of April. Amounts of 50–75 

kg N ha –1 , 45 kg P 2 O 5 ha –1 , and 5 kg K 2 O ha –1 are recommended 

to be applied as basal, and 50–75 kg N ha –1 is applied 

as topdressing. A mixture of 4.5 kg ha –1 oxadiazon and 4.5 kg 

ha –1 butachlor is spread as a preemergence herbicide; a mixture 

of 3 kg ha –1 butachlor and 375 g ha –1 quinclorac is applied 

after emergence if needed. 

Direct seeding of aerobic rice in central China 

Multiyear average rainfall is 700–1,600 mm and cumulative 

temperature is 4,500–6,500 °C in the region. Aerobic rice is 

mainly grown on hilly rainfed land or intercropped with perennial 

crops such as trees of tea, fruit, and mulberry in their 

young period. It is also cultivated following winter wheat or 

rapeseed. Predominant varieties are Han Dao 277, Han Dao 

502, Wushi Handao, Zhonghan 3, and Han Dao 86-76. Aerobic 

rice, being rich in frost-free days and diverse in cropping 

systems, is directly seeded from the end of March to the beginning 

of June in central China. In most cases, it is seeded in 


Table 1. Recommended target yield components of aerobic rice and 

the effect of high seed rates. 

Panicle Grain Filled Grain Yield 

Variety number number grain % weight (kg ha –1 ) 

m –2 panicle –1 (g 10 –3 grains) 

Suggestions of breeders for high yield 

Han Dao 65 270 100 90 26 6,318 

Han Dao 277 255 98 95 28 6,647 

Han Dao 297 320 85 90 25 6,120 

Han Dao 502 280 110 90 29 8,039 

Farmers’ practice with 120–150 kg ha –1 seeding rate 

Han Dao 65 292 88 68 28.1 5,079 

Han Dao 277 344 57 85 26.5 3,870 

Han Dao 297 340 64 86 24.9 4,883 

Han Dao 502 334 83 70 28.4 5,438 

holes with 25–28 cm between rows and 15–18 cm in rows 

since the seeding rate is as low as 40–60 kg ha –1 . Farmers also 

used direct seeding in strips at 75–90 kg ha –1 . Fertilizer application 

rates are 120–150 kg N ha –1 , 40–90 kg P 2 O 5 ha –1 , and 

80–100 kg K 2 O ha –1 . 

Challenges and future emphasis 

Although a potential yield of 6.0–8.0 t ha –1 is claimed for aerobic 

rice, actual yield is estimated at 4.0–5.0 t ha –1 across all 

regions in China (Table 1). Aerobic rice yields in Anhui and 

Hubei provinces averaged 3.4–4.5 t ha –1 in 2003 according to 

field estimates. This yield has not satisfied farmers and for 

this reason the area of aerobic rice has not increased as in previous 

years. Moreover, compared with traditional lowland rice, 

improved aerobic rice exhibits a yield potential that is usually 

70% of that of paddy conditions. 

Agronomists have paid little attention to aerobic rice cultivation 

for the last two decades, during which improved varieties 

increasingly emerged, and no data are available for managers 

and technicians to make proper suggestions to farmers. 

Recommendations tend to follow those used for upland rice 

cultivation. For example, a seeding rate of 120 kg ha –1 or more 

is used widely in northern and northeastern China for aerobic 

rice. This rate is used for upland rice cultivation because the 

percentage of seedling survival and tillering ability is very low 

in poor hilly soils with little input. For high-yielding aerobic 

rice, assuming 80% emergence, a 120 kg ha –1 seeding rate 

would give 330–370 plants m –2 , which would lead to much 

more than the 250–300 panicles m –2 , which is the recommended 

optimal panicle number for yield of around 6–8 t ha –1 . This 

could lead to weak plant growth, very small panicles, and a 

low yield of 3.8–5.4 t ha –1 (Table 1). Field experiments with 

seeding-rate treatments showed that a seeding rate of 60–80 

kg ha –1 produced a significantly higher yield (more than 6.0 t 

ha –1 ) because of the higher proportion of productive tillers 

(Feng et al 2003). It also suggested that nitrogen fertilizer of 

no more than 120 kg N ha –1 could meet the need for aerobic 

rice growth with a yield of 5.0–6.0 t ha –1 . 

Therefore, it is crucial to promote and stabilize the yield 

of aerobic rice with a target level of 6.0 t ha –1 , and we believe 

that this is possible. Greater attention should be paid to investigating 

crop establishment according to the following questions: 

1. Is the high yield of aerobic rice dependent on main 

stems or tillers for different varieties, or what percentage 

of tillers contributes to yield Is this directly 

related to seeding rate Experience in lowland rice in 

China across the north to south suggests that it is essential 

to promote as many tillers as possible and to 

promote a productive tiller percentage as high as 90% 

or more for high yield (Lin 2000). 

2. How do we manage fertilization according to nutrient 

uptake by aerobic rice to obtain high yields Improved 

varieties differ from lowland or traditional 

upland rice with low productivity. However, no data 

on mechanisms for nutrient uptake by the new type 

rice in China are available on which to base fertilizer 

recommendations (Xie et al 2003). 

Additionally, as rice establishment changes from transplanting 

to direct seeding, we face challenges for weed control, 

irrigation regimes, and sowing mechanisms. 

References 

Bouman BAM. 2001 Water-efficient management strategies in rice 

production. Int. Rice Res. Notes 16(2):17-22. 

Fan JH, Cai CH, Zhan Y, Li XC, Sun PZ. 2004. Water-saving cultivation 

technology of aerobic rice. Shandong Agric. Sci. 1:45- 

46. 

Feng SZ, Yang HG, Liu WX, Peng MX, Liu FM, Ma DL. 2003. 

Effects of sowing rates on the yield formation of aerobic rice 

after wheat. Crop Res. 17(2):73-77. 

Li YN, Duan AW. 2002. Introduction to water resources and watersaving 

agriculture. In: Qian YB, Li YN, Yang G, editors. Studies 

on new technology of water-saving agriculture. Zhengzhou 

(China): Yellow River Water Conservancy Press. p 8-12. 

Lin QH. 2000. Rice population quality and regulation. In: Lin QH, 

editor. Crop population quality. Shanghai (China): Shanghai 

Scientific & Technical Publishers. p 42-217. 

Wang H, Bouman BAM, Zhao D, Wang C, Moya PF. 2002. Aerobic 

rice in northern China: opportunities and challenges. In: 

Bouman BAM, Hengsdijk H, Hardy B, Bindraban PS, Tuong 

TP, Ladha JK, editors. Water-wise rice production. Manila 

(Philippines): International Rice Research Institute. p 143- 

154. 

Wang SX, Yin HQ, Tang BJ, Wang YT, Leng GX. 2001. High-yielding 

cultivation technology of aerobic rice in Huanghuai Plain. 

Henan Agric. Sci. 11:11-12. 

Xie GH, Wang SY, Wang HQ, Zhao M. 2003. Mineral nutrition uptake 

and fertilization effects of aerobic rice. Sci. Agric. Sin. 

36:1171-1176. 

Notes 

Authors’ address: College of Agronomy and Biotechnology, China 

Agricultural University, Beijing 100094, China, e-mail: 

xiegh@cau.edu.cn. 


An overview on direct seeding 

for rice crop establishment in the Philippines 

Jovino L. de Dios, Evelyn F. Javier, Myrna D. Malabayabas, Madonna C. Casimero, Alex J. Espiritu 

The Philippines is situated in Southeast Asia, with tropical climate 

and wet and dry seasons. Rice is the staple food and has 

been cultivated since time immemorial. Before the introduction 

of wetland rice, direct seeding in the “caiñgin” or slash 

and burn on mountain slopes (Fig. 1A) and secano or “dry 

planting” on undulating lands to flatlands (Fig. 1B) were practiced. 

Shifting cultivation 

Migrants from mainland Asia carried the traditions of lowland 

rice cultivation to the Philippines during the second millen- 

nium B.C. (www.riceweb.org/History.htm#Origin), later called 

“tubigan” or transplanting (Camus 1921). When irrigation systems 

were introduced in 1840, only 30% of the total area 

planted to rice was irrigated until 1913 and only a few farms 

had sufficient irrigation water for two rice crops a year (Alfonso 

and Catambay 1948). 

Transplanting became common when irrigation pumps 

and canals were constructed in the 1950s (Sta Iglesia and Lawas 

1959) and the National Irrigation Administration was created 

in 1963. Almost all flatlands (approx. 3.126 million ha) were 

planted to rice during the wet season but the majority are still 

rainfed. When irrigated area increased to 1,387,855 ha (NIA 

Fig. 1. Different methods of rice planting in the Philippines (A-D) and adoption of 

direct seeding in average percentage of the regional area (E). A: caiñgin on mountain 

slopes, B: secano or dry seeding on flatland, C: lipat-tanim or transplanted 

rice, D: dalatan or sabog-tanim in wet paddies. 


Table 1. Relative distribution (%) of farms reported practicing direct seeding or 

transplanting as crop establishment in the Philippines. 

Year 

Direct seeding 

Transplanting 

Jan to Jun cropping Jul to Dec cropping Jan to Jun cropping Jul to Dec cropping 

(dry season) (wet season) (dry season) (wet season) 

1998 62.77 37.23 66.05 33.95 

1999 67.85 32.15 64.27 35.73 

2000 65.88 34.12 68.82 31.18 

2001 69.23 30.77 68.33 31.67 

2002 67.02 32.98 71.27 28.73 

Source: PhilRice-BAS (2004). 

2004), there were more dry-season rice crops in areas with 

enough water. The introduction of short-duration 

nonphotosensitive rice varieties and effective fertilizer management 

by IRRI led to improved yields. 

The shift from transplanting (Fig. 1C) to wet direct seeding 

(WDS) or “sabog-tanim” (Fig. 1D) with improved management 

practices for weeds, pests, water, and fertilizer could 

be due to high production costs (De Datta 1986, De Datta and 

Flinn 1986), labor scarcity for transplanting during the peak 

season, the opportunity for early establishment for more cropping 

and maximizing available soil moisture, avoidance of bad 

weather conditions during harvesting, and conserving more 

water. In Central Luzon, Philippines, for example, the adoption 

rate of WDS was from

t ha –1 season –1 

t ha –1 season –1 

5 

4 

Regional average yield of rice in the wet season 

Transplanted 

Direct seeded 

A 

5 

4 

Regional average yield of rice in the dry season 

Transplanted 

Direct seeded 

B 

3 

3 

2 

2 

1 

1 

0 

1 2 3 4 5 6 7 8 9 10 11 12 13 

Region 

ARMM 

0 

1 2 3 4 5 6 7 8 9 10 11 12 13 

Region 

ARMM 

% 

% 

Dry season 

Wet season 

80 

C 

60 

D 

60 

40 

40 

20 

20 

0 

Broadleaves Sedges Grasses 

0 

Broadleaves Sedges Grasses 

Weed type 

Weed type 

mm % 

60 

800 

Water used (mm) 

E 

Water saved (%) 

50 

600 

40 

400 

200 

30 

20 

10 

0 

Continuously 

flooded 

Irrigation when 

no floodwater 

Shallow irrigation 

(0–2 cm) 

Irrigation every 

14 d 

0 

Water management 

Fig. 2. Regional average yields of direct-seeded and transplanted rice in the Philippines (A and B), weed-type distribution for wet- and 

dry-season wet direct-seeded rice (C and D), and water used and saved by using controlled irrigation (E). 

whereas pregerminating means soaking the seeds for 12–24 h, 

followed by incubating for 24–35 h. Dry and primed seeds are 

usually sown in dry to moist soil, whereas pregerminated seeds 

are usually sown in saturated soil. 

Sowing methods. Broadcast seeding is the most common 

method because of simplicity. Drilling or distributing seeds 

evenly in rows is done manually or by a seeder. Dibbling or 

localized planting of seeds is done in small holes created by a 

pointed object thrust into the soil. This is common in caiñgin 

(minimal or zero-tillage). 

Seeding rate. Most farmers used seeding rates of up to 

200 kg of seed ha -1 although the recommended rate is 60–80 

kg ha –1 (Table 2). The wide variation among regions could be 

attributed to their culture, skills, technical knowledge, and the 

presence of adverse conditions for DS such as uncontrolled 

water and the presence of pests that affect crop establishment. 

Higher seeding rates would be beneficial if no weed control is 

planned (Guyer and Quadranti 1985), but it is not necessary to 

use high seeding rates to suppress weeds if an effective herbicide 

is used (Castin and Moody 1989). 


Table 2. Average seeding rates (in kg ha -1 ) in 

the Philippines. 

Year Direct seeding Transplanted 

1998 155 101 

1999 151 92 

2000 150 91 

2001 143 89 

2002 146 94 

Mean 151 94 

Recommended 60–80 20–40 

Source: PhilRice Statistics 1970-2002. 

Crop care 

Weed problems. Early control of weeds efficiently alleviates 

weed problems because it lessens competition between rice 

and weeds that emerge simultaneously in direct seeding. 

Farmers normally practice hand weeding. Nevertheless, 

in wet broadcast seeding, hand weeding is more difficult and, 

before weeds are removed, damage has been done to rice. Hand 

weeding might be too slow and it might not coincide with the 

best time to minimize competition. If weeding is delayed beyond 

20 d after emergence, irreparable damage is done. During 

early establishment, weeds have 20–30% of their growth, 

while the crop has 2–3% of its growth (Moody 1990). 

Major weeds include grasses such as Echinochloa spp. 

and Leptochloa chinensis, sedges such as Cyperus spp. and 

Fimbristylis miliacea, and broadleaves such as Monochoria 

vaginalis, Sphenoclea zeylanica, and Ludwigia octovalvis. 

During the wet season, sedges and broadleaves dominate, while 

only broadleaves dominate during the dry season (Fig. 2C,D). 

However, some cultural management practices help mitigate 

this problem. Properly prepared land, managed water, and selected 

herbicides based on types of weeds present, water status, 

and growth stage of the crop are very important. Some 

herbicides such as safened butachlor at 0.75 kg ai ha –1 and 

safened pretilachlor at 0.3 kg ai ha –1 applied at 3 DAS work 

well in WDS with controlled irrigation. 

Weed control strategies must be well integrated for better 

long-term control and environmental protection. Timing of 

herbicide application is very important. 

Fertilizer management. Basal N is not recommended. 

Instead, NPK fertilizers are applied 10–20 d after seeding 

(DAS). More nitrogen fertilizer during panicle initiation is 

recommended. The most efficient time for fertilizer N application 

is during panicle initiation, while basal is less efficient 

because (1) the surge of available-N from inorganic fertilizer 

is a mismatch for the low N requirement of young plants, (2) 

there is still high N from the soil during the early part of the 

season, (3) the practice of suspending irrigation 10–20 DAS 

to hasten seedling survival and to stabilize soil may trigger 

high N loss, and (4) an early high-N supply may give more 

chance for weeds to compete with rice. Around panicle initiation, 

the N uptake rate and requirement are high, whereas N 

application during tillering is condition-specific. The use of 

real-time N management such as the leaf color chart strategy 

was also introduced in direct seeding. 

Water management. Pond water during land preparation 

aids in land leveling. Shallow pond water 10–14 days after the 

crop had established until the time when weed germination is 

still critical helps prevent weeds from germinating or slows 

the growth of others. Mabbayad (1967) reported that weed 

weight decreased markedly even at a shallow water depth of 

2.5 cm. 

Water is also very important in establishing good WDS. 

The field is drained within 2 wk after seeding by constructing 

shallow drainage ditches or “canaletes” inside the paddy to 

prevent floods when rains come just after seeding. 

Dry seeding and semidryland preparation save a significant 

amount of water during land preparation. Controlled irrigation 

can save as much as 50% of water and the corresponding 

labor cost for irrigation (Fig. 2E). Good land preparation 

also helps in water savings. Because of water scarcity, the localized 

testing of aerobic rice technology to save more irrigation 

water is under way. Aerobic rice technology means growing 

rice in aerated soil with characteristic high yields and fertilizer 

response comparable with those of irrigated lowland 

rice culture. 

References 

Alfonso DJ, Catambay AB. 1948. A study of the net duty of irrigation 

water for lowland rice in Calamba, Laguna. Philipp. Agric. 

32(2):87-113. 

Camus JS. 1921. Rice in the Philippines. Philipp. Agric. Rev. 14(1):7- 

86. 

Castin EM, Moody K 1989. Effect of different seeding rates, moisture 

regimes, and weed control treatments on weed growth 

and yield of wet-seeded rice. In: Proceedings of the 12th Asian- 

Pacific Weed Science Society Conference, Seoul, Korea. 

p 337-343. 

De Datta SK. 1986. Technology development and the spread of direct-seeded 

flooded rice in Southeast Asia. Exp. Agric. 22:417- 

426. 

De Datta SK, Flinn JC. 1986. Technology and economics of weed 

control in broadcast-seeded flooded tropical rice. In: Noda K, 

Mercado BL, editors. Weeds and the environment in the tropics. 

Bangkok (Thailand): Asian Pacific Weed Agency. 

Guyer R, Quadranti M. 1985. Effect of seed rate and nitrogen level 

on the yield of direct wet-seeded rice. In: Proceedings of the 

10th Asian-Pacific Weed Science Society Conference, 

Chiangmai, Thailand. p 304-311. 

Khan MAK, Bhuiyan SI, Undan RC. 1992. Assessment of directseeded 

rice in an irrigated system in the Philippines. 

Bangladesh Rice J. 3(1&2):14-20. 

Mabbayad BB. 1967. Tillage techniques and planting methods for 

lowland rice. M.Sc. thesis. University of the Philippines College 

of Agriculture, College, Laguna, Philippines. 

Moody K. 1990. Post-planting weed control in direct-seeded rice. 

Paper presented at a Rice Symposium, 25-27 September 1990, 

Malaysian Agricultural Research and Development Institute, 

Penang, Malaysia. 


Moody K, Cordova VG. 1985. Wet-seeded rice. In: Women in rice 

farming. Manila (Philippines): International Rice Research 


NIA (National Irrigation Administration). 2004. Irrigation development. 

www.nia.da.gov.ph/Systems/Irrigation_Development.htm, 

19 August 2004. 

PhilRice, BAS. 2004. Philippine rice statistics 1970-2002. Volume 

2. Muñoz, Nueva Ecija (Philippines): PhilRice, BAS. 

Sta Iglesia JC, Lawas JM. 1959. Effects of pump irrigation on farmers’ 

organization, operation and income of rice farms in Sto. 

Domingo and La Purisima, Nabua, Camarines Sur. In: A halfcentury 

of Philippine agriculture. BA Golden Jubilee Committee. 

Manila (Philippines): Graphics House. 463 p. 

Notes 

Authors’ address: Philippine Rice Research Institute, Science City 

of Muñoz, Nueva Ecija 3119, Philippines, e-mail: 

jldeDios@philrice.gov.ph. 

Rice establishment in drought-prone areas of Bangladesh 

M.A. Mazid, M.A. Mortimer, C.R. Riches, A. Orr, B. Karmaker, A. Ali, M.A. Jabbar, and L.J. Wade 

The High Barind Tract of northwest Bangladesh is droughtprone, 

with the majority of the 1,200–1,400-mm mean annual 

rainfall occurring in June to October. Limited irrigation potential 

restricts cropping intensity to 175%, considerably less 

than in districts where irrigation allows two or three rice crops 

each year (Nur-E-Elahi et al 1999). The majority of farmers 

produce a single crop of transplanted rainfed rice, grown in 

this monsoon season. Some 80% of the area then lies fallow in 

the postrice rabi season. The challenge in the Barind is to simultaneously 

improve the reliability and yield of rice while 

increasing total system productivity by increasing the area 

planted to postrice rabi crops, including chickpea, linseed, and 

mustard (Mazid et al 2003). Mazid et al (2002, 2003) have 

proposed that the productivity of Barind soils can be increased 

by switching from transplanted rice (TPR) to direct-seeded 

rice (DSR) to allow more reliable establishment of rabi crops 

on residual moisture immediately after the rice harvest. 

Chickpea, a drought-tolerant and high-value crop, can be grown 

successfully when seeded after rice in late October to mid- 

November. This can make significant contributions to higher 

productivity and improved farm income. A late onset of the 

monsoon delays transplanting as a minimum of 600 mm of 

cumulative rainfall is needed to complete ploughing, puddling, 

and transplanting. Direct seeding can be completed after 

ploughing, however, following only 150 mm of cumulative 

rainfall (Saleh and Bhuiyan 1995). Earlier planted DSR matures 

1–2 weeks before transplanted rice, thus reducing the 

risk of terminal drought and allowing earlier planting of a following 

nonrice crop (Saleh et al 2000). An earlier rice harvest 

can also be achieved by planting early-maturing rice varieties. 

Swarna, the most widely grown cultivar, matures after 140 to 

145 days and, when transplanted, may not be harvested until 

early to mid-November. In many years, soil dries rapidly at 

this time, reducing the likelihood of successful chickpea establishment. 

DSR reduces labor and draft power requirements for rice 

establishment by 16% and 30%, respectively, compared with 

TPR. However, weeds are a major constraint to the adoption 

of DSR as the inherent advantage of weed control afforded by 

transplanted rice in standing water is lost (Mazid et al 2002). 

Labor shortages for many households prevent timely first weeding 

of transplanted rice so that with current practices 34% of 

farmers lose over 0.5 t ha –1 of the attainable rice yield because 

of weed competition (Mazid et al 2001). The additional weed 

problems in DSR may be overcome, however, by applying a 

preemergence herbicide (Mazid et al 2001). We report on the 

productivity of direct seeding of an early-maturing rice cultivar 

with herbicide application, followed by chickpea in the 

postrice season. 

Methods 

Systems trial 

The productivity of two rice cultivars when direct-seeded or 

transplanted was evaluated in the Barind from 2001 to 2003 

under differing nutrient regimes. Modern cultivar BRRI dhan 

39 (BR39) (maturity 120–125 days) was compared with the 

widely grown Swarna (maturity 145–150 days). The experiment 

was conducted with a split-split plot design with main 

plots (3) as crop establishment and associated weed management, 

subplots (4) as nutrient management, and sub-subplots 

(2) as cultivars. Establishment treatments were (1) transplanted 

rice (TPR)—soil puddled prior to transplanting and plots handweeded 

twice at 30 and 45 days after transplanting (DAT); (2) 

direct-seeded rice (DSR)—soil ploughed prior to dry seeding 

(2001) or ploughed and puddled before direct seeding of 

pregerminated seed (2002 and 2003) in rows by hand, with 

hand weeding at 21, 33, and 45 days after sowing (DAS); (3) 

direct-seeded rice with chemical weed control (DSRH)—as 

with DSR but with oxadiazon (375 g a.i. ha –1 ) applied 2–4 

days after seeding, with one hand weeding at 33 DAS. Nutrient 

regimes were (kg ha –1 ) (1) single superphosphate, 40 P + 

40 K; (2) compound 60 N + 40 P + 40 K; (3) farmyard manure 

(FYM) + inorganic fertilizer totaling 60 N + 50 P + 50 K; and 


Table 1. Effect of rice establishment method and rice cultivar on grain yield of rice 

and a postrice chickpea crop (t ha –1 ± S.E.), 2002-04, Rajabari, northwest 

Bangladesh. 

Crop Transplanted rice Direct-seeded rice 

BR39 Swarna BR39 Swarna 


2001 1.92 ± 0.10 2.79 ± 0.19 2.91 ± 0.12 2.85 ± 0.11 

2002 2.80 ± 0.13 2.59 ± 0.11 2.96 ± 0.08 3.75 ± 0.15 

2003 0.61 ± 0.08 0.51 ± 0.05 1.62 ± 0.21 2.67 ± 0.16 

Chickpea 

2001-02 – a – 1.01 ± 0.06 0.91 ± 0.05 

2002-03 – – 0.76 ± 0.05 0.49 ± 0.04 

2003-04 – – 0.38 ± 0.04 0.16 ± 0.02 

a 

= chickpea not sown. 

(4) diammonium phosphate (DAP) (18% N) + Guti slow-release 

urea (45%N) totaling 43 N + 40 P + 40 K. Rice was 

harvested in a 5-m 2 area. Biomass of individual weed species 

was recorded in two unweeded quadrats per plot at 28 DAS/ 

DAT and total weed biomass at 45 DAS/DAT. 

Trials began at 16 sites during the 2003 monsoon season 

to verify the profitability of a DSR rice-chickpea system. 

Chickpea (cv. Barisola 2) was sown after the harvest of Swarna 

or three shorter-duration BRRI dhan cultivars established by 

either transplanting or direct seeding. Prior to dry direct seeding 

in June, the land was ploughed (at least 3 times) with an 

animal-drawn country plough and leveled with a ladder. Seed 

was sown in lines by hand into furrows opened by a handpulled 

lithao (plough). Seedbeds were established at the same 

time and seedlings were transplanted at approximately 30 DAS 

following conventional ploughing and puddling operations. In 

direct-seeded rice, a single application of oxadiazon (375 g 

a.i. ha –1 ) was made to control weeds, whereas, in transplanted 

rice, pretilachlor (450 g a.i. ha –1 ) was applied. 

Seasonal variation in rainfall was considerable. The annual 

rainfall in 2001, 2002, and 2003 was 1,475, 1,464, and 

932 mm, respectively. In July 2003, rainfall was 2.5 times less 

than in the same month in previous years and in 2001 it was 

highest in October. 

Results 

Systems trial 

There were significant effects of cropping system on the phenology 

of rice. Flowering was always later with cv. Swarna 

and under transplanting (P broadleaf 

weeds. 

The yields of chickpea after direct-seeded rice were significantly 

higher following BR39 than after Swarna in 2002 

and 2003 (P

Grain yield (t ha –1 ) 

8 

6 

4 

0.65 ± 0.07 

0.67 ± 0.06 

0.70 ± 0.07 

0.61 ± 0.06 

the relatively low chickpea yields obtained on-farm in 2003, 

after either transplanted or direct-seeded rice, provided worthwhile 

additional income given the crop’s low input costs and 

high market value. Further studies are continuing to evaluate 

the profitability and sustainability of direct-seeding and ricechickpea 

systems under farmer management. 

2 

0 

Conclusions 

BR31 BR32 BR39 Swarna 

Cultivar 

Fig. 1. Productivity (t ha –1 ) of rice and chickpea 

grown in transplanted rice (open columns) and 

direct-seeded rice (solid columns) systems. 

Data are means of 16 on-farm sites in 2003 

(monsoon and rabi seasons). Weed control in 

rice is by preemergence herbicide. Data above 

each pair of bars are chickpea yields (t ha –1 ± 

S.E.). 

The systems trial confirmed that replacing transplanted rice 

with direct-seeded rice could improve farm productivity in the 

Barind by allowing a greater opportunity to raise a high-value 

rabi crop. The early-season weed flush associated with direct 

seeding can be successfully controlled by oxadiazon applied 

preemergence. However, one subsequent manual weeding will 

be essential for yield protection from weed competition and to 

prevent a buildup of Alternanthera sessilis, Cyperus iria, and 

Paspalum distichum in particular. While extensive cultivar 

evaluation is under way, cvs. BR31 and BR32 represent promising 

lines for direct seeding. BR39, on the other hand, is not 

suitable for direct seeding as sterility is a major problem when 

it is planted early because it tends to flower in rains. Successful 

chickpea cropping after rice is contingent upon the presence 

of residual soil moisture and the time-window for successful 

chickpea establishment may be difficult to exploit. 

Chickpea yields in the systems trial reported above were always 

higher because crops were established immediately after 

the rice harvest. In on-farm trials in 2003, the potential advantage 

from direct seeding was not evident because widespread 

rain showers during the first two weeks of October favored 

establishment regardless of the time of the rice harvest. High 

yields were not achieved, however, because of late-season 

drought. Our associated socioeconomic studies indicate that, 

although farmers are keen to gain additional income from growing 

chickpea, many, particularly resource-poor share-croppers 

who pass 50% of their production to the landlord, need practices 

that maximize rice yield. To be widely adopted for direct 

seeding in place of Swarna, an early-maturing rice cultivar will 

need to be high-yielding and sheath-blight-resistant. The reduction 

in input costs (no nursery and substitution of labor 

with a herbicide) that can be achieved with direct seeding and 

early planting of chickpea was evaluated favorably by an onfarm 

trial with farmers in 2003. Farmers considered that even 

References 

Mazid MA, Bhuiyan SI, Mannan MA, Wade LJ. 2002. Dry-seeded 

rice for enhanced productivity of rainfed drought-prone lands: 

lessons from Bangladesh and the Philippines. In: Pandey S, 

Mortimer M, Wade LJ, Tuong TP, Hardy B, editors. Direct 

seeding in Asian rice systems: strategic research issues and 

opportunities. Manila (Philippines): International Rice Research 


Mazid MA, Jabber MA, Riches CR, Robinson EJZ, Mortimer M, 

Wade LJ. 2001. Weed management implications of introducing 

dry-seeded rice in the Barind Tract of Bangladesh. Proceedings 

of the BCPC Conference–Weeds 2001. 1:211-216. 

Mazid MA, Jabber MA, Mortimer M, Wade LJ, Riches CR, Orr AW. 

2003. Improving rice-based cropping systems in north-west 

Bangladesh: diversification and weed management. The BCPC 

International Congress, Crop Production and Protection. 

p 1029-1034. 

Nur-E-Elahi AH, Khan MR, Siddique A, Saha M, Nasim M, 

Shahidullah SM. 1999. Existing cropping patterns of 

Bangladesh: potential techniques and strategies for improving 

system productivity. In: Mandal MR, editor. Proceedings 

of the Workshop on Modern Rice Cultivation in Bangladesh. 

Gazipur (Bangladesh): Bangladesh Rice Research Institute. 

p 107-169. 

Saleh AFM, Bhuiyan SI. 1995. Crop and rainwater management strategies 

for increasing productivity of rainfed lowland rice systems. 

Agric. Syst. 49:259-276. 

Saleh AFM, Mazid MA, Bhuiyan SI. 2000. Agrohydrologic and 

drought-risk analyses of rainfed cultivation in northwest 

Bangladesh. In: Tuong TP, Kam SP, Wade LJ, Pandey S, 

Bouman BAM, Hardy B, editors. Characterizing and understanding 

rainfed environments. Manila (Philippines): International 


Notes 

Authors’ addresses: M.A. Mazid, Bangladesh Rice Research Institute 

(BRRI), Post Rangpur Cadet College, Rangpur 5404, 

Bangladesh, e-mail: brrirang@bdonline.com; M.A. Mortimer, 

University of Liverpool, L69 3BX, UK; C.R. Riches and A. 

Orr, Natural Resources Institute, University of Greenwich, 

ME4 4TB, UK; B. Karmaker and A. Ali, BRRI, Post 

Shyampur, Rajshahi, Bangladesh; M.A. Jabbar, BRRI, Post 

BRRI, Gazipur, Bangladesh; L.J. Wade, The University of 

Western Australia, Crawley, WA 6009, Australia. 

Acknowledgments: This work was partially funded by the UK Department 

for International Development (Crop Protection 

Programme Project R8234). However, the views expressed 

are not necessarily those of DFID. Support has also been received 

from the Consortium for Unfavorable Rice Environments 

(ADB/IRRI). 


Emerging issues in weed management of direct-seeded rice 

in Malaysia, Vietnam, and Thailand 

M. Azmi, D.V. Chin, P. Vongsaroj, and D.E. Johnson 

Malaysia’s rice area was 692,600 ha for the combined dryand 

wet-season crops in 2000. In Peninsular Malaysia, rice 

double cropping in the eight granary areas accounted for a 

third of this total. Direct seeding (DS) as an alternative to the 

traditional transplanting method in the main granaries in Malaysia 

began in the late 1970s and thereafter expanded rapidly 

(Azmi and Abdullah 1998). By 2000, this method accounted 

for more than 90% of the total rice area planted. Wet seeding 

rather than dry is favored but is only feasible where good water 

management is possible. DS is undertaken by hand broadcasting 

or using a motorized blower. 

Rice is the most important crop in Vietnam and Thailand. 

The crop is grown in Vietnam on 4.2 million ha or 

74.1% of the arable land and produces 29 million tons every 

year. The Mekong River Delta in the south and Red River Delta 

in the north are the two main regions for rice production, occupying 

52% and 18% of the national rice area. In the Red 

River Delta, rice is largely transplanted and double-cropped, 

whereas, in the Mekong Delta, about half the rice area is irrigated, 

there are three growing seasons, and the crop is mainly 

direct-seeded. Wet seeding is the most common practice (about 

60% of the area), with water seeding, zero tillage, and dry 

seeding making up the balance. Rice cultivation in Thailand 

covers about 10 million ha. At the end of the 1990s, the majority 

of the rainfed rice area was dry direct-seeded and the remainder 

transplanted, whereas irrigated areas in the Central 

Plain were largely wet-seeded. Ten years earlier, however, most 

of the rainfed areas had been transplanted, as were 30% of the 

irrigated areas of the Central Plain, with the remaining areas 

being dry-seeded. This rapid change in establishment methods 

was the result of labor shortages in rural areas, particularly 

during the peak season. 

Weed succession in rice ecosystems 

The widespread introduction of DS, the repeated use of herbicides, 

and limited irrigation supplies in the 1990s are the factors 

responsible for the shift in weed species populations in 

rice ecosystems. Grasses, such as Echinochloa crus-galli, along 

with other Echinochloa spp. (E. oryzicola, E. colona, E. 

stagnina, and E. picta), Leptochloa chinensis, and Ischaemum 

rugosum, which were not prevalent and dominant in Malaysian 

rice fields in the 1970s (Azmi et al 1993), became widespread 

in the 1990s. Weed surveys were carried out in areas 

where direct seeding was being introduced to replace transplanting 

from 1989 to 1993. In 1989, while transplanting was 

the principal system, Monochoria vaginalis, Sphenoclea 

guyanensis, Fimbristylis miliacea, and Leptochloa flava were 

dominant weeds. In 1993, these were replaced by L. chinensis, 

E. crus-galli, M. vaginalis, and E. colona in areas that were 

by then mostly direct-seeded. From 1981 to 1994, the areas 

infested by L. chinensis grew by an estimated 4% per annum 

and all areas in Muda were infested by 1992. A similar trend 

was observed in Vietnam, where E. crus-galli, L. chinensis, 

and weedy rice were dominant weeds in the DS systems (Chin 

2001). In Thailand, Sphenoclea zeylanica, Monochoria 

vaginalis, and Marsilea minuta are dominant in the transplanted 

systems, whereas E. crus-galli, L. chinensis, Cyperus iria, and 

C. difformis are dominant in the wet-seeded areas (Vongsaroj 

1997). Some of the shifts in the importance of weed species as 

a result of changes in crop establishment method from traditional 

transplanting (1970s) to DS culture are shown in Table 

1. Weedy rice is difficult to control because of its similarity to 

the rice plant. It was detected in Malaysia in 1988 (Azmi and 

Abdullah 1998) and Vietnam in 1994 (Chin 2001), and there 

has also been a rapid ingression of weedy rice in the DS area 

of the Central Region of Thailand in recent years. Recent weed 

surveys conducted in Malaysia found that weedy rice (Oryza 

sativa) had become a problem, and the weed is thought to be a 

serious threat to rice production (Azmi et al 2003). 

Yield loss from weeds 

Yield losses largely depend on season, weed species, weed 

density, rice cultivar, and growth rate and density of weeds 

and rice. Weedy rice at 35% infestation can cause about a 60% 

yield loss (Watanabe et al 1997) and, under a serious infestation, 

a yield loss of 74% was recorded in DS rice (Azmi and 

Abdullah 1998). 

Impact of prolonged herbicide usage in direct-seeded rice 

Labor shortages have led to an increase in direct seeding and a 

rapid increase in the use of and reliance on herbicides. This 

reliance has resulted in an undesirable shift in weed species 

and concerns about environmental contamination. Phenoxy and 

sulfonylurea compounds are widely used herbicides in Malaysia, 

Vietnam, and Thailand to control broadleaf weeds and 

sedges in DS rice. Weeds resistant to these herbicides have 

evolved and there is evidence that weed species such as S. 

zeylanica, Marsilea minuta, and F. miliacea have developed 

resistance to phenoxy herbicides (Watanabe et al 1997). Lately, 

acetolactate-synthase (ALS) inhibitor-resistant biotypes have 

been reported in several weed species in many countries, including 

Malaysia (Azmi and Baki 2003). ALS-inhibitor herbicides, 

including sulfonylurea, are widely used in the world 


Table 1. Weed shift from transplanting to the direct-seeding method. 

Irrigated transplanting Extensive direct Intensive direct 

seeding seeding 

Grasses 

Isachne globosa Echinochloa crus-galli complex E. crus-galli 

Leersia hexandra Leptochloa chinensis L. chinensis 

Ischaemum rugosum 

I. rugosum 

Oryza sativa (weedy rice) 

O. sativa (weedy rice) 

Broadleaf weeds 

Limnocharis flava L. flava L. flava a 

Monochoria vaginalis M. vaginalis S. guyanensis b 

M. minuta L. hyssopifolia Sphenoclea zeylanica b 

Sphenoclea zeylanica M. minuta M. crenata b 

Limnophila erecta a 

Sedges 

Scirpus grossus Cyperus iria C. iria 

Fimbristylis miliacea 

F. miliacea b 

C. difformis 

a 

Biotypes with herbicide resistance against 2,4-D and ALS-inhibitor herbicides. b Species/biotypes with herbicide 

resistance against 2,4-D. 

because of their low dosage, mammalian toxicity, and herbicide 

injury in many crops. 

Weedy rice 

The term “weedy rice” refers to populations of annual Oryza 

species that diminish farmer income both quantitatively through 

yield reduction and qualitatively through lowered commodity 

value at harvest (Mortimer et al 2000). Weedy rice populations 

are easy-shattering weedy types of rice, and are unwanted 

plants among cultivated rice. These weedy rice populations 

are found in Malaysia, Vietnam, and Thailand. In Malaysia, a 

close relationship between weedy rice and cultivated varieties 

has been shown, giving a strong indication that evolutionary 

forces are still present in the rice ecosystems (Abdullah et al 

1997). According to Mortimer et al (2000), three factors that 

determine the population of weedy rice are (1) seed remaining 

dormant in the soil over crop seasons, (2) dissemination through 

crop seed contamination, and (3) seeds returning from plants 

in the previous crop. No single control measure will effectively 

control weedy rice. An integrated approach involving 

cultural, physical, and chemical interventions is expected to 

be effective in managing the weedy rice problem in a sustainable 

manner. Unless the problem is addressed, weedy rice in 

many areas poses a major threat to sustainable DS rice production. 

Water seeding 

Besides tillage practices, good water management is an important 

prerequisite for controlling weedy rice and other weed 

species. Further, water seeding (WS) can be practiced in leveled 

fields with levees where the water depth at seeding time 

is 5–10 cm, sufficient to suppress weedy rice (Chin et al 2000). 

Water seeding as opposed to wet seeding is effective in weedy 

rice suppression and can reduce populations by 70–76%. In 

Vietnam, WS can be practiced 10–20 days earlier than wet 

seeding, allowing farmers to save water. Further, energy inputs 

can be reduced by nearly half because of the lack of drainage 

at the beginning of the cropping period and reduced irrigation 

requirements from tillering to ripening. Weed control 

input for WS is also reduced by one-half to a third that of wet 

and dry seeding, and a 69% reduction in the weedy rice population 

was recorded after WS practices in Selangor, Malaysia 

(Azmi 2004, unpublished data). WS practices for two consecutive 

seasons reduced weedy rice infestation to below 10%. 

Conclusions 

Effective weed management practices are an important prerequisite 

in DS rice culture with herbicide application seemingly 

indispensable. With the advent of herbicide-resistant weed 

biotypes and unwarranted environmental contamination following 

the continuous and indiscriminate use of herbicides in 

Malaysia, Vietnam, and Thailand, a more sustainable integrated 

weed management (IWM) technology needs to be developed. 

This might be through the thorough understanding of the biology, 

ecology, and socioeconomics of the major weeds in the 

rice ecosystems. IWM, the promotion of biological and cultural 

control options, and minimizing the use of herbicides are 

seen as keys to sustainable rice-farming systems in the region. 


References 

Abdullah MZ, Vaughan DA, Watanabe H, Okuno K. 1997. The origin 

of weedy rice in Muda and Tg. Karang areas in Peninsular 

Malaysia. MARDI Res. J. 24(2):169-174. 

Azmi M, Abdullah MZ. 1998. A manual for the identification and 

control of padi angin (weedy rice) in Malaysia. Serdang (Malaysia): 

MARDI Publication. 18 p. 

Azmi M, Baki BB, Mashhor M. 1993. Weed communities in principal 

rice-growing areas in Peninsular Malaysia. MARDI Report 

No. 165. 15 p. 

Azmi M, Baki BB. 2003. Weed species diversity and management 

practices in the Sekinchan Farm Block, Selangor’s South West 

Project – a highly productive rice area in Malaysia. Proceedings 

1, 19th Asian-Pacific Weed Science Society Conference, 

17-21 March 2003, Philippines. p 174-184. 

Azmi M, Sivapragasam A, Abdullah MZ, Muhamad H. 2003. Weedy 

rice management through integration of cultural, physical and 

chemical interventions in direct-seeded rice. Paper presented 

at the International Rice Conference 2003, 13-16 Oct. 2003, 

Alor Setar, Kedah, Malaysia. 

Chin DV, Hien TV, Hien TV. 2000. Weedy rice in Vietnam. In: Baki 

BB, Chin DV, Mortimer M, editors. Wild and weedy rice in 

rice ecosystems in Asia: a review. Limited Proceedings No. 2. 

Los Baños (Philippines): International Rice Research Institute. 

p 45-50. 

Mortimer M, Pandey S, Piggin C. 2000. Weedy rice: approaches to 

ecological appraisal and implications for research priorities. 

In: Baki BB, Chin DV, Mortimer M, editors. Wild and weedy 

rice in rice ecosystems in Asia: a review. Limited Proceedings 

No. 2. Los Baños (Philippines): International Rice Research 


Vongsaroj P. 1997. Weed management in paddyfields. Botany and 

Weed Science Division, Department of Agriculture, Bangkok. 

Bangkok (Thailand): Amarin Printing Company. 175 p. 

Watanabe H, Azmi M, Md Zuki I.1997. Emergence of major weeds 

and their population change in wet-seeded rice fields in the 

Muda area, Peninsular Malaysia. Proceedings of 16th Asian 

Pacific Weed Science Society. p 246-250. 

Notes 

Authors’ addresses: M. Azmi, MARDI Rice Research Centre, Locked 

Bag 203, 13200 Kepala Batas, Penang, Malaysia, 

azmiman@mardi.my; D.V. Chin, Cuu Long Delta Rice Research 

Institute, Omon, Can Tho, Vietnam; P. Vongsaroj, Department 

of Agriculture, Bangkok, Thailand; D.E. Johnson, 

International Rice Research Institute, DAPO Box 7777, Metro 

Manila, Philippines. 

Changing from transplanted rice to direct seeding 

in the rice-wheat cropping system in India 

Y. Singh, Govindra Singh, David Johnson, and Martin Mortimer 

Rice and wheat are the staple food crops of India, contributing 

nearly 80% of the total food-grain production, and the ricewheat 

rotation is the principal cropping pattern of the Indo- 

Gangetic Plains, where the system occupies some 13.5 million 

ha. Demand for food grains in India is expected to grow and 

the requirement for 2025 is estimated to be a 40% production 

increase compared with 2003-04. There has, however, been 

stagnation in rice productivity in recent years, and long-term 

experiments show a declining rice yield trend (Padre and Ladha 

2004). Further, increasing costs are leading to demand for 

change in rice production methods to improve returns to farmers. 

Direct seeding in place of transplanting is one alternative 

being considered. 

In the Indo-Gangetic Plains, rice is traditionally transplanted 

at the end of the dry season (May/June) and, after the 

rice harvest, wheat is sown (November/December). Direct 

seeding has already replaced transplanting in many parts of 

Southeast Asia. The combined effects of changing water availability 

and the opportunity cost of labor are driving reevaluation 

of crop establishment methods. Pandey and Velasco (2002) 

argued that low wage rates and adequate water supply favor 

transplanting but direct seeding is likely to increase in circumstances 

of labor scarcity and increasing wage rates. These latter 

conditions are now common in parts of India and farmers 

are considering the alternatives offered by direct seeding. 

Balasubramanian and Hill (2002) commented that direct 

seeding offers the advantages of reduced labor requirements 

and drudgery, earlier crop maturity, more efficient water 

use and higher tolerance of water deficit, fewer methane 

emissions as the system is more aerobic, and often higher profit 

in areas with an assured water supply. Weeds, however, are the 

main constraint for farmers practicing direct seeding since the 

inherent weed control from standing water at crop establishment 

is lost. Differing levels of farmers’ resources, land development 

and infrastructure, or weather patterns may dictate different 

direct-seeding practices in the Indo-Gangetic Plains. 

Options available to farmers include wet seeding with sprouted 

seeds (broadcast or drum seeding) on puddled fields, dry drilling 

of seed with or without prior cultivation, or broadcasting 

dry seed after dry-cultivating the field. This paper summarizes 

results from an ongoing study of the options for direct-seeded 

rice in the Indo-Gangetic Plains. 


Table 1. Effects of rice establishment method and weed control on grain yield (t ha –1 ) and 

standard errors (SE) of rice in 2001 and 2002 in on-station and on-farm experiments. See 

text for details. 

2001 weed management 2002 weed management 

Establishment 

method On-station On-farm On-station On-farm 

CW HW TO CW TO CW HW TO CW TO 

TP 7.8 7.2 5.9 4.9 3.4 6.1 5.5 4.6 5.8 3.6 

WS 8.1 5.6 1.0 – – 6.8 5.4 0.9 5.9 2.2 

DS 6.1 1.0 0.0 4.9 2.6 6.7 5.1 1.0 6.3 3.6 

ZT 6.6 1.5 0.0 – – 5.7 5.4 0.2 – – 

SE 0.04 0.24 0.23 0.21 


Treatments compared rice establishment methods as main-plot 

treatments with subplots of weed management treatments in 

rice, as follows: 

Main plots (60 m 2 )—rice establishment: 

TP = conventional transplanting of approximately 21- 

day-old plants after soil puddling, 20 × 20-cm spacing 

(seedling nursery established at the same time 

as direct seeding). 

WS = wet seeding (pregerminated, drum-seeded, 35 kg 

ha –1 ) after soil puddling, 20-cm row spacing. 

DS = dry seeding (50 kg ha –1 ) after conventional tillage, 

20-cm row spacing. 

ZT = dry seeding (50 kg ha –1 ), zero-tillage after flush 

irrigation (7 d before glyphosate application). 

Subplots (20 m 2 )—intensity of weeding, postemergence weed 

control in rice: 

TO = unweeded checks. 

CW = weed-free “best-bet herbicide treatment” followed 

by two manual weedings. In TP plots, butachlor 

(1.5 kg a.i. ha –1 ) was applied 2 DAT; on the WS 

plots, anilofos (0.4 kg a.i. ha –1 ) was used 5 days 

after seeding (DAS); and, for the DS and ZT, 

pendimethalin (1.0 kg a.i. ha –1 ) was applied 1 DAS. 

HW = one manual weeding 30 DAS. 

From each subplot, weed counts and biomass (fresh 

weight) by species were taken from two 0.25 m × 1 m quadrats 

covering 5 crop rows. After the rice harvest, the experimental 

area was sown to wheat after either conventional tillage or 

zero-tillage (as strips). The overall experimental design was a 

strip, split-plot design, with four complete randomized replications 

as follows: wheat (cv. PBW-343 or PBW-154, 100 kg 

ha –1 , 20-cm row spacing) was sown in December into either 

conventionally prepared plots (harrowed, rotavated, and leveled) 

or zero-tilled plots (paraquat, 0.5 kg a.i. ha –1 ), 1 week 

before seeding. Yield estimates were taken from 5 m 2 . Full 

experimental details are given in Singh et al (2003). In farmers’ 

fields, CW and TO treatments were imposed as described 

above for TP, WS, and DS establishment methods. 

Results 

In 2001, grain yield from transplanted rice was similar to that 

of wet-seeded rice when weeds were controlled (CW), whereas 

yield from dry-seeded rice (DS, ZT) was about 20% less (Table 

1). A single hand weeding (HW) was insufficient to prevent 

major yield loss from weeds in either the DS or ZT rice and 

yield loss was complete without weed control (TO). In 2002, 

when weeds were controlled (CW), yields under WS and DS 

were similar or higher than under TP and ZT. Yield losses because 

of weed competition, however, were less severe than in 

2001 and, with a single weeding (HW), yield in dry-seeded 

areas decreased by an average of 16% versus 80% in 2001. In 

farmers’ trials in 2001, equivalent rice yields were achieved 

from DS and TP rice, with potential yield losses from weeds 

(TO) being 30% in TP and almost 50% in DS. In 2002, DS 

rice gave the highest grain yield, while potential losses from 

weeds were the lowest in TP (37%), followed by DS (44%) 

and WS (62%). 

Weed species shifts in response to crop establishment 

are well known. At the start of the experiment in 2000, in the 

absence of weeding, Echinochloa colona, Caesulia axillaries, 

and other broadleaf weeds (other dicots) and Fimbristylis 

miliacea were the most abundant weeds at 56 DAS in both the 

transplanted and dry-seeded plots (Fig. 1). After three seasons 

under transplanting, however, E. colona, Commelina diffusa, 

and Ischaemum rugosum were dominant, whereas, in dryseeded 

rice, C. diffusa dominated. Furthermore, by 2002, 

Leptochloa chinensis, an annual grass previously unrecorded 

at the site, was noted, and Cyperus rotundus, a perennial sedge, 

had also become proportionally more abundant in dry-seeded 

areas over the three years up to 2002. The population dynamics 

of individual weeds responded noticeably to establishment 

method. E. colona, C. rotundus, C. diffusa, and I. rugosum all 

declined in abundance (density at 28 DAT/56 DAS) under trans- 


Biomass (g m –2 ) at 28 DAT 

10 

2000 

Transplanted rice 

Density (m –2 ) at 28 DAT 

Echinochloa colona 

Cyperus rotundus 

1.0 

0.1 

10 

Other 

(dicot) 

2002 

C. axillaris 

E. colona 

F. miliacea 

I. rugosum 

C. diffusa 

C. rotundus 

100 

10 

E. colona 

0.01 

1.0 

C. diffusa 

I. rugosum 

C. axillaris 

100 


Commelina diffusa 

0.1 

C. rotundus 

0.01 

F. miliacea 

10 

Biomass (g m –2 ) at 56 DAS 

10 

2000 

Dry drill-seeded rice 

Density (m –2 ) at 56 DAS 

Echinochloa colona 

Cyperus rotundus 

1.0 

C. axillaris 

E. colona 

Other 

(dicot) F. miliacea 

I. rugosum 

C. diffusa 

100 

0.1 

10 

2002 

C. rotundus 

10 

C. diffusa 

E. colona 

0.01 

1.0 

C. rotundus 

L. chinensis 

C. difformis F. miliacea 

I. rugosum 

100 


Commelina diffusa 

0.1 

C. axillaris 

10 

0.01 

2000 2002 2004 

2001 2003 

Year 

2000 2002 2004 

2001 2003 

Year 

Fig. 1. Changes in abundance of weed species in relation to crop establishment method. Diagrams on the left are rank abundance curves 

based on biomass at 28 DAT or 56 DAS comparing plots in 2000 and 2002. Diagrams on the right illustrate change in density (mean ± 

standard error of mean) of selected species (at 28 DAT or 56 DAS) over five seasons. Log 10 scales. 

Caesulia axillaris- = C. axilliaris Echinochloa colona = E. colona Fimbristylis miliacea = F. miliacea 

Commelina diffusa = C. diffusa Cyperus rotundus = C. rotundus Ischaemum rugosum = I. rugosum 

Cyperus difformis = C. difformis Leptochloa chinensis = L. chinensis 


planting. Conversely, E. colona and C. diffusa increased in 

density over 2000-04 under direct seeding, while seasonal 

variation was evident in C. rotundus and I. rugosum. 

Analyses of the costs and returns of rice production by 

the different establishment systems, based on farm yields, indicate 

that gross returns were broadly similar for each of the 

crop management systems. The costs of nursery beds, land 

preparation, and labor for transplanting, however, resulted in 

a benefit-cost ratio for TP approximately half that of DS rice. 

Relative to TP, DS required approximately 40% of the total 

water by volume and substantially less labor. 

Discussion 

Direct seeding offers farmers in the Indo-Gangetic Plains more 

independence from the hired migrant labor required for transplanting 

and enables the rice crop to be established earlier in 

the monsoon season. In turn, early rice establishment, combined 

with the shorter duration of direct-seeded rice, facilitates 

earlier sowing of the following wheat crop and improves 

wheat yield. Coupled with water savings in rice, the introduction 

of direct seeding offers considerable system-wide benefits. 

However, successful direct seeding of rice is contingent 

upon effective weed management early in the life of the crop. 

This study has shown that the use of a herbicide (pendimethalin 

for drill-seeded rice and anilofos for wet-seeded rice) followed 

by one hand weeding will protect yield from weed competition. 

Although in some seasons supplementary hand weeding 

may be unnecessary, removal of “escapes” through failed application 

or tolerance is essential for competitive grasses 

such as I. rugosum. It may also be necessary to alter crop management 

practices to prevent an undesirable shift in weed species, 

particularly to Cyperus rotundus. This may be achieved 

by rotation of crop establishment methods or the incorporation 

of other cultural measures or stale seedbed treatments. 

Current trends of increasing labor costs in many regions 

of the Indo-Gangetic Plains, and the increasing concerns over 

water availability in the long term, suggest that direct seeding 

will become an increasingly attractive option for farmers. Further 

research is required, however, to determine the 

sustainability of weed management practices and the extent to 

which the crop establishment methods are suited to farmers’ 

circumstances that vary in terms of resources, soil, and infrastructure. 

Delivery of knowledge on the integration of crop 

establishment and weed management practices to farmers remains 

a pressing need. 

References 

Balasubramanian V, Hill JE. 2002. Direct seeding of rice in Asia: 

emerging issues and strategic research needs for the 21st century. 

In: Pandey S et al. Direct seeding: research strategies 

and opportunities. Proceedings of the International Workshop 

on Direct Seeding in Asian Rice Systems, 25-28 January 2000, 

Bangkok, Thailand. Los Baños (Philippines): International 


Padre AT, Ladha JK. 2004. Integrating yield trends of rice-wheat 

systems using the linear mixed effects model and meta-analysis. 

In: Rice-wheat information sheet. Rice-Wheat Consortium, 

New Delhi, India. p 1-3. 

Pandey S, Velasco L. 2002. Economics of direct seeding in Asia: 

patterns of adoption and research priorities. In: Pandey S et 

al, editors. Direct seeding: research strategies and opportunities. 

Proceedings of the International Workshop on Direct 

Seeding in Asian Rice Systems, 25-28 January 2000, Bangkok, 

Thailand. Los Baños (Philippines): International Rice Research 


Singh G, Singh Y, Singh VP, Singh RK, Singh P, Johnson DE, 

Mortimer M, Orr A. 2003. Direct seeding as an alternative to 

transplanted rice for the rice-wheat system of the Indo- 

Gangetic plains: sustainability issues related to weed management. 

In: Proceedings of an International Congress held 

at the SECC, Glasgow, Scotland, UK, 12-14 Nov. 2003. p 7 

F-9-1035. 

Notes 

Authors’ addresses: Y. Singh and Govindra Singh, G.B. Pant University 

of Agriculture & Technology, Pantnagar 263145, India; 

David E. Johnson, International Rice Research Institute, 

DAPO Box 7777, Metro Manila, Philippines; Martin 

Mortimer, School of Biological Sciences, University of 

Liverpool, Liverpool, UK, e-mail: 

pratibhasingh_84@rediffmail.com. 




are not necessarily those of DFID. 


Seedling recruitment in direct-seeded rice: 

weed biology and water management 

A.M. Mortimer, O. Namuco, and D.E. Johnson 

In rice cropping, water constitutes a powerful selective agent 

for weed management (Mortimer and Hill 1999). Permanently 

flooded fields of transplanted rice that are unweeded by 45 

days after transplanting (DAT) exhibit a weed flora different 

from fields where the soil remains saturated. Noticeable differences 

are also evident across toposequences where land 

formation imposes different drainage and flooding regimes in 

rainfed rice (Pane et al 2000). However, it is the timing, duration, 

and depth of flooding events in association with the 

method of crop establishment that govern the precise nature 

of weed suppression by water. Wet seeding onto a saturated 

puddled soil with subsequent flooding 10–15 days afterward 

in comparison with transplanting into standing water alters the 

dominance hierarchy of weeds with the replacement of obligate 

aquatic dicotyledonous species by semiaquatic species, 

including competitive grasses. Weed species shifts may be relatively 

rapid. From observations of farmers’ fields, Azmi and 

Mashor (1995) recorded that seven seasons of wet seeding 

resulted in the appearance of 21 new weed species, particularly 

Echinochloa species and forms of “weedy” rice, and the 

exclusion of 12 others. 

Understanding the processes underlying the selective 

nature of water in governing the recruitment and survival of 

weed species involves comparative autecological studies of 

seed germination and seedling and plant growth in relation to 

flooding regimes. These regimes may vary from permanently 

aerobic but moist soils, through continuously saturated anaerobic 

soils, to saturated soils subject to varying water depths 

arising from periodic flooding and drainage, either as a result 

of rainfall variation or as a policy of water management. Potentially, 

plants may experience all of these flooding regimes 

at different stages in their life cycle and the diversity in rice 

weed floras is in part because, in many semiaquatic species, 

the rooting systems of adult plants are tolerant of, or adapted 

to, submergence in flooded soils. However, controlled studies 

of weed community dynamics in transplanted and wet-seeded 

rice emphasize the fact that weeds are recruited early in the 

life of both crops. In transplanted rice continuously flooded at 

shallow (< 50 mm) or deeper (approx. 100 mm) depth, recruitment 

of plants to the weed community occurred only in 

the first 10–20 DAT and subsequently losses continued until 

the rice harvest. In wet-seeded rice initially sown on saturated 

soil and subsequently flooded to these depths by 10–15 days 

after seeding (DAS), recruitment of sedges and dicotyledonous 

species was restricted to the same early period in contrast 

to graminaceous species that increased in overall density up to 

60 DAS (Hill et al 2001). Cohort recruitment of weeds in dry- 

seeded rice may, however, extend over a longer period than in 

wet-seeded rice, depending on the time of initiation and depth 

of flooding. 

Intrinsically, seed size confers an advantage in seedling 

establishment (autotrophic net gain in biomass after germination). 

From a phyto-centric seedling point of view, the emergence 

of photosynthetic tissue above a water surface is a critical 

life-history event in that there is a relatively abrupt increase 

in available photosynthetically active radiation (PAR) and absence 

of hypoxic conditions. Successful water seeding of rice 

relies upon rapid mobilization of these reserves that is partly 

dependent upon dissolved oxygen in floodwater and, similarly, 

successful direct seeding of rice requires the absence of flooding 

immediately after sowing. Understanding seed and seedling 

biology at this stage in the life cycle is critical in understanding 

the role of water management in governing weed seedling 

recruitment. We briefly review the available knowledge. 

Weed biology: toward a definition of functional weed types 

Germination 

Seed size in rice weeds varies on a logarithmic range from 

approximately 10 µg (Sphenoclea zeylanica, Cyperus 

difformis) to greater than 10 4 µg (Rottboellia cocchinensis) and 

seed size in weed species is well known to be associated with 

suites of dormancy/germination traits. In small-seeded species, 

the absence of oxygen (Monochoria vaginalis) or lowered 

oxygen concentration (Zinzania aquatica), fluctuating diurnal 

temperatures (Najas graminea), and differential responses to 

spectral light ratios (Sphenoclea zeylanica) have all been shown 

to be cues involved in response mechanisms that govern seed 

germination rates underwater. Phytochrome-mediated switches 

in response to red/far-red light ratios provide another probable 

mechanism governing germination although there has been 

little detailed work on the features of photocontrol of germination 

of many small-seeded rice weeds (Sanders 1994). Gap 

detection mechanisms are moreover likely to confer fitness 

advantages in a rice-weed community that will experience 

largely uniform canopy closure, often within 30–40 days. 

Equally evident is the fact that small-seeded species (e.g., 

Cyperus difformis) exhibit polymorphisms in germination response 

to flooding, providing “bet-hedging” tactics against 

unpredictable flooding events. Conversely, some species exhibit 

little or no innate or induced dormancy and germinate 

rapidly on the surface of saturated soils (Fimbristylis miliacea, 

Echinochloa colona, Kim and Moody 1989) but not underwater. 

It is useful, therefore, to attempt to qualitatively group weed 


species by cues for germination response although such a classification 

may be blurred by ecotypic differentiation. The classification 

of species on character states such as obligate requirements 

for germination, physiological (Cyperus difformis) 

or somatic dormancy polymorphisms (Ischaemum rugosum), 

and gap detection mechanisms is a first step in identifying 

groups of species that respond differentially to flooding and 

that may not reflect taxonomic fidelity. 

Transition to autotrophy 

Very early plant growth patterns are characterized by the rate 

of transition from heterotrophy to autotrophy. Once committed 

by germination to growth, seedling ability to photosynthesize 

underwater at lowered light intensities and to mobilize 

resources to develop photosynthetic structures out of water 

are key traits influencing survivorship. Understanding the role 

of water depth in prohibiting weed recruitment is then concerned 

with the growth dynamics of seedlings underwater and 

response to depth-dependent light regimes. Recent research at 

IRRI has measured the growth responses of a range of weed 

species underwater in fully illuminated conditions and total 

darkness. Echinochloa crus-galli and Cyperus difformis illustrate 

dramatic differences in resource mobilization and partitioning 

(Fig. 1). In the larger-seeded grass, carbohydrate mobilization 

and partitioning to root and shoot resulted in the 

onset of autotrophy before leaf emergence from an illuminated 

water column. Under total darkness, in contrast, the maximum 

shoot height achieved before exhaustion of reserves was 80 

mm, with evidence of enhanced height extension rate. In C. 

difformis, net assimilation of biomass occurred below a 45- 

mm water column in full illumination, but biomass was not 

partitioned into photosynthetic tissue that emerged above the 

water surface, by 300 thermal units (TU). Root length (a partial 

surrogate of root biomass) extension was noticeable immediately 

after germination under both light regimes, suggesting 

that carbohydrate assimilate was partitioned to roots in the 

first instance. Maximum shoot height was achieved in the dark 

regime over 300 TU. Emergence of leaf tissue from illuminated 

water columns was not achieved until 700 TU (data not 

shown) but plants returned to lighted conditions from after 300 

TU in darkness did not grow. These two species differ sharply 

in their response to flooding regimes, with C. difformis maintaining 

a longer aquatic stage before emergence from the water 

column because of underwater photosynthesis. In contrast, 

in E. crus-galli, seedling growth depended on emergence from 

the water column before reserve use was complete. 

Water management for direct-seeded rice 

The preceding discussion allows for the ready identification 

of two functional types in response to water at the time of direct 

seeding: those species in which early establishment success 

is critically dependent upon height extension in relation 

to water depth and those in which it is not. It is a likely supposition 

that some species in the latter category may also possess 

ecological correlates in terms of gap detection mechanisms at 

germination. While the extent to which these types simply represent 

ends of a continuum and the possible role of 

heterophyllic tissue are unclear, quantification of growth responses 

as illustrated above enables two questions to be answered: 

(1) What is the minimum depth required to prohibit 

establishment in clear and dark turbid water, thus capturing 

the range of field conditions (2) How long must seedlings be 

submerged for seed reserves to be exhausted While answers 

are species-specific and require thermal time measurement, 

they are pertinent in defining the suppressive effect of water 

and providing a causal explanation of variation in weed recruitment 

patterns. 

Comparative studies on Leptochloa chinensis illustrated 

a growth strategy similar to that of E. crus-galli with respect 

to seed reserve partitioning to shoot and leaf structures, although 

both the specific growth rate and height extension rate 

of aboveground biomass were twofold lower. Germination in 

this species is strongly suppressed by standing water (>15 mm) 

and establishment was found to be prohibited by water depths 

of >50 mm. Experimental dose-response studies with 

pretilachlor for early postemergence control of both E. crusgalli 

and L. chinensis have shown that significantly reduced 

doses of the herbicide can be employed if coupled with precise 

flooding regimes that take into account plant size and 

height extension rate after application. 

Jackson and Ram (2003) have recently emphasized the 

opportunities for selecting for tolerance of complete submergence 

in established plants based on the SUB1 locus. Moreover, 

Caton (2002) has illustrated the importance of quantifying 

seed reserve mobilization in direct seeding. The extent to 

which genes that ensure rapid establishment and submergence 

tolerance can be exploited for germplasm adapted for direct 

seeding remains a pressing research issue (Yamauchi and 

Biswas 1997). However, it is worth remarking that the extent 

to which submergence tolerance can be of use in weed management 

may be constrained by the evolutionary responses 

within the Echinochloa genus, which reflects a wide spectrum 

of flood tolerance (Kennedy et al 1989). 

References 

Azmi M, Mashor M. 1995. Weed succession from transplanting to 

direct seeding methods in Kembubu rice area. Malaysian J. 

Biosci. 6:143-153. 

Caton BP. 2002. Simulating seed reserve mobilization and seedling 

growth of rice in DSRICE1. Field Crops Res. 76:55-69. 

Hill JE, Mortimer AM, Namuco OS, Janiya JD. 2001. Water and 

weed management in direct-seeded rice: are we headed in the 

right direction In: Peng S, Hardy B, editors. Rice research 

for food security and poverty alleviation. Los Baños (Philippines): 


Jackson MB, Ram PC. 2003. Physiological and molecular basis of 

susceptibility and tolerance of rice plants to complete submergence. 

Ann. Bot. 91:227-241. 


Biomass (mg) 

9 

Echinochloa crus-galli 

0.04 

Primary root length (mm) 

Cyperus difformis 

15 

Root and shoot 

biomass (illuminated) 

0.03 

Root length 

(illuminated) 

12 

6 

Total biomass 

(illuminated) 

9 

3 

Structural seed 

biomass 

Root and shoot 

biomass (dark) 

0.02 

0.01 

Total biomass 

(dark) 

Root length (dark) 

6 

3 

0 

0.00 

0 

Shoot height (mm) 

200 

15 

150 

12 

Under full illumination 

9 

100 

In dark 

In dark 

6 

Under full illumination 

50 

3 

0 

0 100 200 300 

Thermal units (°C days) 

0 

0 100 200 300 

Thermal units (°C days) 

Fig. 1. Seedling growth responses in Echinochloa crus-galli and Cyperus difformis in response to flooding. Germinated 

seeds were placed in water columns containing nutrients to a depth of 50 mm and either illuminated (light/dark) on a 12- 

h cycle or maintained in total darkness at a temperature of 30/20 °C day/night. PAR = 450 µE m –2 s –1 . Repeated destructive 

harvests were taken over the time course, enabling fit of the functions shown. 

Kennedy RA, Fox TC, Dybiec LD, Rumpho ME. 1989. Biochemical 

adaptations to anoxia in rice and Echinochloa seeds. In: 

Taylorson RB, editor. Recent advances in the development 

and germination of seeds. New York (USA): Plenum Press. 

p 151-163. 

Kim SC, Moody K. 1989. Germination of two rice cultivars and 

several weed species. Korean J. Weed Sci. 9:116-122. 

Mortimer AM, Hill JE. 1999. Weed species shifts in response to 

broad-spectrum herbicides in sub-tropical and tropical crops. 

Brighton Crop Protection Conference (1999) 2:425-437. 

Pane H, Noor ES, Dizon M, Mortimer AM. 2000. Weed communities 

of gogorancah rice and reflections on management. In: 

Tuong TP, Kam SP, Bouman B, Pandey S, Wade L, Hardy B, 

editors. Characterizing and understanding rainfed rice environments. 

Manila (Philippines): International Rice Research 


Sanders BA. 1994. The need for understanding the life cycle of the 

rice weed Cyperus difformis. Aust. J. Exp. Agric. 34:1031- 

1038. 


Yamauchi M, Biswas JK. 1997. Rice cultivar difference in seedling 

establishment in flooded soil. Plant Soil 189:145-153. 

Notes 

Authors’ addresses: A.M. Mortimer, School of Biological Sciences, 

University of Liverpool, Liverpool L69 7ZB, UK; O. Namuco 

and D.E. Johnson, International Rice Research Institute, 

DAPO Box 7777, Metro Manila, Philippines, e-mail: A.M. 

Mortimer@liverpool.ac.uk. 




are not necessarily those of DFID. 

The crop protection industry’s view on trends 

in rice crop establishment in Asia and their impact 

on weed management techniques 

Jean-Louis Allard, Kee Fui Kon, Yasuo Morishima, and Ruediger Kotzian 

Farmers in South and Southeast Asia will have to increase rice 

production significantly to meet the needs of a growing population. 

Competition between industry and agriculture for water, 

labor, and land will oblige rice farmers to intensify production 

while using less labor and water. The risk of herbicide 

overuse coming from those changes was highlighted in the 

1990s (Naylor 1994). Weed management practices must therefore 

be integrated and adapted to new cropping systems and 

associated weed problems (Labrada 2004). 

In northeast Asia, the change in diet, opening of commodity 

markets, aging farmers, and reduction in the cultivated 

area also compel rice farmers to increase productivity while 

maintaining higher product quality. These changes will have a 

significant impact on crop establishment and water management 

practices. They will also affect weed populations and 

herbicide use. 

Until recently, the crop protection industry was focusing 

on the discovery of new active ingredients. High development 

costs and a stagnating and fragmented rice crop protection 

market are leading to a reduced investment in discovery 

and development of new rice herbicides. With less chemical 

innovation, the crop protection industry in the future will have 

to adapt use recommendations for current active ingredients 

in order to solve new weed problems and ensure reliable performance 

under changing cropping practices and conditions. 

In the future, recommendations for herbicide use will increasingly 

take into account integrated weed management techniques 

(Kudsk and Streibig 2003). 

This paper describes changes in weed problems and their 

management in three major Asian rice-cropping systems. Sustainable 

weed management solutions based on development 

activities from Syngenta in Asia are used to illustrate how the 

crop protection industry responds to changes. 

Trends and impact on weed management techniques 

Direct seeded wet-sown rice 

Direct seeded wet-sown rice is widely practiced in Southeast 

Asia, the southern provinces of China, and South Korea. The 

technique has been widely known since the 1980s and, after a 

rapid initial development, further expansion has slowed down 

because of the high constraints of this cropping system. Chemical 

weed control is essential to avoid yield losses but high crop 

sensitivity at the early stages restricts the number of active 

ingredients that can be used. Accurate leveling and good water 

management are also required to ensure adequate weed 

control. High infestations of grasses such as Echinochloa crusgalli 

(L.) Beauv. and Leptochloa chinensis (L.) become dominant 

after a few cropping seasons. Weedy rice (Oryza sativa 

L. fp. spontanea) is expanding in areas where wet-sown rice is 

grown continuously, such as Vietnam, Malaysia, Sri Lanka, 

Thailand, and Korea (Baki et al 2000). Resistance of grass 

weeds such as Echinochloa or Leptochloa spp. to propanil, 

quinclorac, and ACC-ase inhibitors is starting to appear in several 

countries. Resistance of annual broadleafs and sedges to 

ALS-inhibitors has been frequently detected in Korea (Park 

2004) and is starting to appear in China. 

During recent years, the crop protection industry has introduced 

more flexible and powerful postemergence herbicides 

from the ACC-ase and ALS-inhibitor groups for grass or broadspectrum 

weed control. Increasing reliance on such chemical 

classes will intensify the risk of developing weed resistance 

(Valverde et al 2000). Syngenta is adopting an alternative approach. 

The 25-year-old herbicide/safener combination 

pretilachlor/fenclorim is being rejuvenated by using an earlier 

application timing 3–7 days before seeding. The earlier application 

eliminated the major constraint for farmers, which is 

that it must be sprayed on mud within a few days after seeding 

(Fig. 1). Direct application by dripping the concentrate or 


Fig. 1. Conventional herbicide application in direct-seeded wet-sown rice in Thailand. 

Fig. 2. Application of pretilachlor/fenclorim during leveling. 

slightly diluted product into water at the last leveling is easy 

(Fig. 2). Crop tolerance is excellent. Performance is stable and 

less affected by weather conditions or water management than 

conventionally applied products. The method provides reliable 

broad-spectrum weed control, including Leptochloa sp. 

and weedy rice. It also offers farmers a herbicide with a different 

mode of action as a resistance management tool. This example 

shows how exploring novel use recommendations for 

an old product can provide a solution to new weed problems. 

Transplanted rice in South and Southeast Asia 

In areas with increasing labor costs, farmers will implement 

techniques that reduce production costs and increase yield and 

cropping intensity. Techniques such as the use of younger seedlings 

and reduction in hand weeding and herbicides such as 

2,4-D or sulfonylureas will lead to an increase in Echinochloa 

spp. infestations. In areas with reduced water availability such 

as “tail areas” of irrigated perimeters or during El Niño years, 

or in areas with high cropping intensity, farmers will need techniques 

to reduce water use and time for soil preparation. 

The role of industry will be to register herbicides for 

controlling grasses and train farmers in the safe and effective 

use of such products. New formulations for easier application 

with a shaker bottle may be favored over products applied 

through only foliar application. Syngenta is also promoting 

preplanting applications with nonselective herbicides such as 

paraquat, allowing quick planting without tillage in several 

Chinese provinces (Fig. 3) or reduced tillage in Indonesia and 

eastern India. The technique helps farmers to speed up soil 

preparation and planting by up to 1 month and reduce water 


Fig. 3. Farmers planting rice without tillage in southern China. 

consumption by up to 15% (Kartaatmadja et al 2004). Application 

of paraquat followed by no-tillage is currently the only 

technique allowing sustainable rice cultivation in Indonesian 

tidal rice. The development of reduced- or no-tillage together 

with an adapted rice cropping system remains a possibility in 

the fast-growing no-till wheat area in northwest India. 

Transplanted rice in northeast Asia 

In northeast Asia, stringent pesticide laws will be enforced to 

ensure the safety of users, consumers, and the environment. 

Aging “weekend or part-time” farmers and also professional 

farmers managing a large number of fields will require broadspectrum 

herbicides to control annual and perennial weeds 

(“one-shot herbicides”), which can be easily applied through 

dripping at planting or later from dikes or into water inlets. 

Sulfonylurea herbicide-resistant (SU-R) annual broadleafs or 

sedges such as Monochoria vaginalis (Burm. f.) Presl var. 

plantaginea (Roxb.) Solms-Laub. and Scirpus juncoides Roxb. 

subsp. juncoides Roxb. are already widely distributed and will 

continue to expand (Itoh and Uchino 2002). The control of 

these weeds will require resistance prevention measures or the 

addition of an antiresistance component into herbicide combinations. 

Maintenance and reinforcement of registration dossiers 

for new and old herbicidal active ingredients will continue to 

be a key priority for the crop protection industry. However, 

because of limited financial investment in a stagnating market, 

cost and benefit will need to be balanced, which could 

lead to the divestment or phasing-out of older or economically 

nonviable active ingredients. 

Responding to farmers’ demands, the industry will increasingly 

focus new developments on so-called “easy-to-apply” 

herbicides such as “Jumbos” or SC formulations. Herbicides 

that can be applied during other operations such as leveling 

or transplanting will also gain in importance. 

The crop protection industry is also working intensively 

on new solutions to SU-R problems. Specialized active ingredients 

such as carfentrazone, bromobutide, benzobicyclon, or 

others are being added into the already complex “one-shot” 

mixtures, adding to production costs and restricting formulation 

options. It is pertinent to compare the cost of a resistance 

prevention strategy with the cost of the cure. Calculations based 

on Italian rice production systems have, for example, shown 

that rotation of different herbicide mode of actions can prevent 

or delay the appearance of resistance at a much lower 

cost than using curative treatments when the resistant weed 

population has developed (Orson 1999). In northeast Asia, it 

has been recognized that repeated use of “one-shot” products 

in rice, relying only on sulfonylureas to control annual 

broadleafs and sedges, is one of the main causes of the development 

of ALS-resistance (Itoh 2001). In line with the Herbicide 

Resistance Action Committee (HRAC 2004) guidelines 

for the prevention of resistance, Syngenta recommends an application 

of pretilachlor applied preplanting or preemergence 

in sequence with a pyriftalid-based “one-shot” to delay the 

development of SU-R weeds in farmers’ fields in both Korea 

and Japan (Table 1). 

Recommendations and conclusions 

In the coming years, scarcer land, labor, and water in conjunction 

with the need for intensification will drive significant 

changes in cultivation techniques and weed management in 

Asian rice production. 

The number of farmers switching from transplanting to 

direct seeding is expected to be minor and limited to large 


Table 1. Effect on SU-R SCPJO (Scirpus juncoides Roxb. subsp. juncoides Roxb.) with sequential use of pretilachlor 

followed by a pyriftalid/pretilachlor-based mixture. Results of a field trial in Miyagi Prefecture, Japan, in 2003. 

Chemicals (formulation) Dosage Timing a % Control on SCPJO 

(g a.i. ha –1 ) (evaluated at 40 DAT) 

Bensulfuron-methyl (WP) 51 3 DAT 46 

Pretilachlor (EW) 400 3 DAT 90 

Pyriftalid/pretilachlor/azimsulfuron/bensulfuron-methyl (GR) 180/180/6/30 3 DAT 85 

Pretilachlor (EW) → 400 → 3 DAT→ 

pyriftalid/pretilachlor/azimsulfuron/bensulfuron-methyl (GR) 180/180/6/30 24 DAT b 100 

a 

Chemicals were applied at 3 DAT (3 days after transplanting of rice) at preemergence of SCPJO. b The sequential application followed at 24 

DAT. 

farms where this technique increases productivity significantly. 

In most cases, farmers will improve productivity by adapting 

their current cultivation system. Reduced- or no-tillage is one 

of those opportunities to reduce water consumption and increase 

cropping intensity. Herbicides will replace hand-weeding 

to reduce costs. Younger seedlings will also be used more 

frequently to reduce the workload and cost of nursery maintenance. 

Those changes will lead to a higher use of herbicides, 

particularly in tropical Asia. 

The “thrown seedling” method, currently widely practiced 

in China, may be a further alternative to transplanting for 

smaller farms. Larger farms will aim at improving productivity 

through mechanization. Dry soil cultivation with large tractors 

will be the first option, but reduced- or no-tillage may 

also become popular. Intensive dry-seeded rice as it is widely 

practiced in the United States, Brazil, and now Italy could appear 

in limited areas in Asia. However, this system relies on 

good availability of irrigation water and a high input of preand 

postemergence herbicides (2–4 applications at a cost of 

US$100–250 ha –1 ). The high chemical input can be managed 

in an environmentally acceptable way. Favorable areas could 

be large farms in northwest India, Malaysia, and Thailand. 

In areas with an already high use of herbicides such as 

direct-seeded wet-sown or mechanically transplanted rice, 

farmers will require herbicides, which are easier to apply and 

solve new weed problems such as weedy rice or SU-R weeds. 

The role of the crop protection industry will be to introduce 

new solutions responding to new farmers’ needs through novel 

use recommendations, formulations, and antiresistance strategies 

based mostly on current active ingredients. To ensure safety 

to farmers and the environment, strong regulatory standards 

will be required and will have to be implemented. Training 

and stewardship programs will also be needed to ensure the 

safe and sustainable use of herbicides. As a global leader in 

weed control, Syngenta remains committed to delivering innovative 

solutions to Asia’s rice farmers. 

References 

Baki BB, Chin DV, Mortimer M, editors. 2000. Wild and weedy 

rice in rice ecosystems in Asia: a review. Limited Proceedings 

No. 2. Los Baños (Philippines): International Rice Research 


HRAC (Herbicide Resistance Action Committee). 2004. Guidelines 

to the management of herbicide resistance. 

www.plantprotection.org/HRAC/Guideline.html. 

Itoh K. 2001. An overview of resistance in Japan and Asia. Japan- 

Australia Seminar, Utsunomiya University. p 3-4. 

Itoh K, Uchino A. 2002. Sulfonylurea-resistant weeds in paddy rice 

fields of Japan. In: Agrochemical resistance, extent, mechanism, 

and detection. Washington, D.C. (USA): The American 

Chemical Society. p 168-180. 

Kudsk P, Streibig JC. 2003. Herbicides: a two-edged sword. Weed 

Res. 43(2):90-102. 

Kartaatmadja S, Pane H, Wirajaswadi L, Sembiring H, Simatupang 

S, Bachrein S, Ismadi D, Fagi AM. 2004. Optimizing the use 

of natural resources and increasing rice productivity. Proceedings 

of the 4th ISCO Conference, Brisbane, Australia. Paper 

758. 

Park TJ. 2004. Herbicide resistance in rice in Korea. 

www.weedscience.org/in.asp. 

Labrada R. 2004. The need for improved weed management in rice. 

www.fao.org/DOCREP/006/Y4751E/y4751e01.html. 

Orson JH. 1999. The cost of herbicide resistance. 

www.plantprotection.org/HRAC/cost.html. 

Naylor K. 1994. Economic growth threatens to bring herbicide overuse 

to Asia. www.stanford.edu/dept/news/pr/94/ 

940405Arc4350.html. 

Valverde BE, Riches CR, Caseley JC. 2000. Prevention and management 

of herbicide resistant weeds in rice. Electronic Book, 

p 123. www.weedscience.org/in.asp. 

Notes 

Authors’ addresses: Jean-Louis Allard and Kee Fui Kon, Syngenta 

Asia Pacific Pet. Ltd., 250 Northbridge Road #39-00, Raffles 

City Tower, Singapore 179101; Yasuo Morishima, Syngenta 

Japan K.K. 21F, Office Tower X, 1-8-10, Harumi, Chuo-Ku, 

Tokyo 104-6021, Japan; Ruediger Kotzian, Syngenta Crop 

Protection AG, Scwarzwaldallee 256, CH 4002, Basel, Switzerland, 

e-mail: jean-louis.allard@syngenta.com. 


Improved anchorage and bird protection with iron-coated 

seeds in wet direct seeding of rice crops 

Minoru Yamauchi 

The shift in rice establishment method from transplanting to 

direct seeding has recently become prominent in Asia. Direct 

seeding constitutes either dry or wet seeding. Puddling the field 

is characteristic of rice crop production and is generally considered 

to increase soil nutrient supply, suppress weed infestation, 

and increase water holding, thus contributing to 

sustainability. Puddling is conducted for transplanting and wet 

seeding, but not for dry seeding. Therefore, poor water holding 

and heavy weed infestation are major problems in dryseeded 

rice fields. 

A variable rice plant stand is the major constraint to wet 

seeding. Puddling creates an anaerobic soil environment, thus 

making seed germination and growth unstable, particularly 

when seeds are sown beneath the soil surface. In the tropics, 

farmers broadcast pregerminated seeds onto the puddled soil 

after drainage. It was recently shown that rice cultivars tolerant 

of anaerobic soil could be sown underground (Yamauchi 

et al 2000). In Japan, farmers are recommended to seed “oxygen-supplier” 

chemical (major ingredient CaO 2 )-coated 

pregerminated seeds at a soil depth of 10–20 mm after drainage. 

When the seeding depth is too deep, however, emergence 

and plant stand are poor. In contrast, seeding on the surface or 

in shallow soil may lead to floating seedlings when water is 

subsequently introduced. Drainage at seeding tends to stimulate 

weed growth. 

Water seeding is a form of wet seeding in which 

pregerminated seeds are sown in standing water. In the tropics, 

it is practiced when drainage is difficult. In Japan, it was 

practiced in a limited area before the materialization of “oxygen-supplier” 

coating technology. Although water seeding is 

advantageous in water-saving and weed control, the occurrence 

of floating or lodging of seedlings limits adoption. 

This paper describes the new technology developed to 

overcome the floating and lodging problems of rice seedlings 

in standing water. 

Hypothesis 

Seedlings in water are buoyant and tend to be poorly anchored, 

leading to floating or lodged seedlings. We can improve anchorage 

by increasing the specific gravity of seeds by adhering 

heavy material on the husk. Among the materials commonly 

used in agriculture, iron has a high specific gravity of 7.9. We 

confirmed by use of pregerminated seeds that the occurrence 

of floating seedlings was reduced by iron coating (Yamauchi 

2002a). 

Seed vigor is important to achieve a rapid and uniform 

plant stand in fields (Yamauchi and Tun Winn 1996). The ger- 

mination process could be subdivided into imbibition, activation, 

and postgermination stages. We found that drying seeds 

at the activation stage could increase seed vigor by increasing 

the germination rate (Yamauchi 2002b). Dry seeds with high 

vigor produced in this manner were found to have the potential 

to attain the same level of plant stand as pregerminated 

seeds. 

We hypothesized that coating “high-vigor dry seeds” with 

iron could make wet seeding more adaptable and labor-saving 

(Yamauchi 2003). Because dry seeds are storable, iron-coated 

dry seeds can be prepared in advance of the busy farm season. 

Preparation of iron-coated seeds 

The procedure described here is designed for and being tested 

by rice farmers in Japan. Seeds with more than a 95% germination 

percentage are surface-sterilized by soaking in sterilant 

solution for 24 h, followed by ordinary soaking in tap water 

for 24–48 h at 15–20 °C. The soaking time depends on the 

temperature and cultivar. Soaking is stopped just before coleoptile 

emergence on the husk, then the seeds are dried. The 

seeds thus prepared are vigorous. 

Seeds are coated with a mixture of iron powder and calcined 

gypsum (CaSO 4 0.5 H 2 O, 5–25% of iron powder weight) 

(Yamauchi 2004). The iron is in a reduced form. The amount 

of iron used for coating is determined by the ratio of iron to 

seed weight. When the coating ratio is as high as 1, it is recommended 

that an electric motor-driven granulator be used because 

the weight of the mixture is too heavy to handle manually. 

When the ratio is 0.1, granulation could be achieved in a 

container by mixing seeds and the mixture by hand. Portions 

of the mixture are poured successively onto the seeds while 

water is sprayed occasionally to help the granulation. When 

the granulation is over, calcined gypsum alone is dressed so 

that the surface of coated seeds becomes smooth and not chalky. 

Because wetting the mixture of iron powder and gypsum 

initiates rapid oxidation of iron, the temperature of coated 

wet seeds increases. Therefore, the seeds must be spread in a 

shallow layer and dried. Otherwise, there is a risk that seeds 

will be injured and die because of the heat. 

The well-dried coated seeds can be stored at room temperature 

for several months without a significant decrease in 

germination percentage. The specific gravity and single-seed 

weight of seeds thus produced are increased with the increase 

in iron-coating ratio (Figs. 1 and 2). 


Single-seed weight (mg) 

150 

100 

50 

0 

Specific gravity 

4 

3 

Fig. 2. Intact (top) and iron-coated (bottom) rice seeds. 

2 

1 

0 

0 1 2 3 4 

Iron-coating ratio 

Fig. 1. The relationship between iron-coating 

ratio and specific gravity and singleseed 

weight. Iron-coating ratio = iron powder 

weight/seed weight. 

Characteristics of direct seeding with iron-coated dry seeds 

Performance in wet and water seeding 

We tested the seedling establishment and grain yield of rice 

from iron-coated seeds with coating ratios of 0.5, 1, and 2 in 

comparison with “oxygen-supplier” chemical-coated seeds in 

wet and water seeding. “Oxygen-supplier” chemical-coated 

seeds floated in water seeding, achieving a plant stand of 62%, 

whereas iron-coated seeds did not float, achieving a plant stand 

of 82–94%. In wet seeding, the “oxygen-supplier” chemicalcoated 

seeds were damaged by sparrows, whereas the ironcoated 

seeds were affected only slightly, exhibiting 64% and 

81–94% plant stand, respectively. Thus, the plant stand of ironcoated 

seeds is more consistent than that of “oxygen-supplier” 

chemical-coated seeds under the experimental conditions used 

here. There were little differences in grain yield (5,280–5,660 

kg ha –1 ) among the treatments. Our recent study showed that 

the coating ratio could be reduced as low as 0.1, although this 

depends on soil properties, crop management, and seeding 

practices. 

Seeding at the right interval after puddling is important 

to the success of direct seeding and poor timing may end in a 

poor plant stand. We tested the relationship between the time 

of seeding after puddling and plant stand in water seeding. 

The iron-coated seeds anchored well and established seedlings 

successfully even when the time of seeding was 5–6 d after 

puddling. However, noncoated dry seeds failed to achieve anchorage 

and floated, even when seeded on the same day of 

puddling. 

In wet seeding, the soil surface tends to become hard 

after drainage. When the seeds are broadcast-seeded onto such 

hard soil, the contact between seed and soil is poor and seeds 

may become desiccated and die. Iron-coated seeds are advantageous 

because the seeds are heavy and become buried more 

easily in soil. Thus, iron-coated dry seeds seem to allow greater 

flexibility in terms of seeding time in water and wet seeding. 

Bird damage 

Seeds and seedlings are vulnerable to bird attack in the field. 

When the field is drained as in wet seeding, they were often 

destroyed by sparrows (Passer montanus). When the field is 

flooded for water seeding, the seeds are eaten by ducks (Anas 

poecilorhyncha). Therefore, it is difficult to avoid bird damage, 

particularly when sparrows and ducks are in the vicinity. 

It was noted, however, that iron-coated seeds are more resistant 

to sparrow attack at seeding. Sparrows cannot break the 

husk of iron-coated seeds and thus cannot eat the caryopsis. 

Ducks are still able to eat the iron-coated seeds in flooded 

fields, however, and, therefore, when ducks are expected to 

attack, it is recommended to drain the field. 

Environmental effects 

The use of iron-coated seeds for direct seeding should be discussed 

and evaluated in terms of sustainability. While most 

water used for puddling has to be drained in ordinary wet seeding, 

it is not necessary with iron-coated seeds because these 

seeds can be sown and establish plants successfully in the presence 

of water. The introduction of iron-coated seed technology 

may save water used for puddling. In addition, the flexibility 

of water control with iron-coated seeds would make 

herbicide application easier than in ordinary wet seeding. 

Water seeding is advantageous in the control of weeds, 

including weedy rice. Despite this advantage, water seeding 


has not yet been fully employed in direct seeding because of 

the unstable plant stand. Water seeding with iron-coated seeds 

provides an opportunity for ecologically sound weed control 

practices. 

Continuous application of iron to paddy fields may 

amend the soil and reduce the occurrence of physiological rice 

plant diseases, such as Akagare disease. In Japan, it is recommended 

to apply 340 kg Fe ha –1 annually to maintain the fertility 

of paddy soil. Seeding iron-coated seeds results in the 

application of 5–50 kg Fe ha –1 with each crop when the seed 

rate is 50 kg ha –1 . We confirmed that application of iron does 

not increase the ferrous iron concentration or cause any physiological 

problem. Thus, the introduction of iron-coated seed 

technology would be beneficial in maintaining soil fertility. 

Cost and labor requirement 

A new technology should be low-cost and labor-saving from 

the perspective of farmers. Granulation could be done manually 

if the coating ratio were low. Iron powder costs approximately 

US$1–3 kg –1 in Japan. Assuming a seed rate of 50 kg 

ha –1 and a coating ratio of 0.1, the cost for iron powder would 

be $5–15 ha –1 . 

In ordinary wet seeding, farmers have to follow a strict 

time schedule for seed preparation, land preparation, and seeding, 

watching the weather conditions for successful seeding. 

Because iron-coated seeds can be prepared in advance, avoiding 

the busy farming season, this method contributes to laborsaving. 

In Japan, many farmers are currently purchasing seedlings 

prepared for transplanting machines. We can imagine that 

once iron-coated seeds are produced commercially and sold 

on the market, part-time and weekend farmers could use such 

seeds so that they can save on labor for seed preparation. 

Conclusions and future research needs 

Because the technology of iron-coated seeds is new in Japan, 

we need to continue the evaluation. Furthermore, the feasibility 

of this technology in tropical Asia should be discussed. 

This technology could be advantageous for developing a sustainable 

technology oriented for water-saving, reducing the use 

of chemicals, and promoting ecological weed control. 

References 

Yamauchi M. 2002a. Reducing floating rice seedlings in wet direct 

sowing by increasing specific gravity of seeds with iron powder 

coating. Jpn. J. Crop Sci. 71(extra issue 1):150-151. (In 

Japanese.) 

Yamauchi M. 2002b. Production of rice seeds with increased vigor 

and their application to wet direct sowing. Jpn. J. Crop Sci. 

71(extra issue 1):152-153. (In Japanese.) 

Yamauchi M. 2003. Developing rice direct sowing technology with 

iron-coated dry seeds. Jpn. J. Crop Sci. 72(extra issue 1):4-5. 

(In Japanese.) 

Yamauchi M. 2004. Rice direct sowing with iron-coated dry seeds: 

coating and sowing methods. Jpn. J. Crop Sci. 73(extra issue 

1):10-11. (In Japanese.) 

Yamauchi M, Aragones DV, Casayuran PR, Sta. Cruz PC, Asis CA, 

Cruz RT. 2000. Seedling establishment and grain yield of tropical 

rice sown in puddled soil. Agron. J. 92:275-282. 

Yamauchi M, Tun Winn. 1996. Rice seed vigor and seedling establishment 

in anaerobic soil. Crop Sci. 36:680-686. 

Notes 

Author’s address: National Agricultural Research Center for Western 

Regions, NARO, Nishifukatsu 6-12-1, Fukuyama 721- 

8514, Japan, e-mail: myamauch@affric.go.jp. 

Issues for integrated weed management 

and decision support in direct-seeded rice 

D.E. Johnson and A.M. Mortimer 

Rice production has been transformed over much of Asia, with 

total harvest increasing by 2.4% per annum from 1968 to 1999 

(IRRI 2004) through the introduction of improved germplasm, 

agronomy, pest management, and, in many cases, mechanization. 

At the same time, the declining availability of labor for 

agriculture and increasing labor costs have required farmers 

to seek alternatives to manual weeding, which has long provided 

farmers with the means to limit losses to weeds. Herbicide 

use in rice grew dramatically in Asia from 1980 to 1995, 

with more than a threefold increase in sales, to more than $900 

million per annum (Naylor 1996), allowing a massive release 

of labor from agriculture that is expected to continue. The use 

of herbicides has been accompanied globally by the evolution 

of herbicide resistance in weeds, weed species population shifts, 

and concerns about environmental contamination and human 

health. 

In addition to labor, there are increasing concerns over 

supplies of irrigation water. Farmers in many rice-growing areas 

are likely to have only limited availability of irrigation 

water and, in the future, most of the 22 million ha of dry-season 

rice in South and Southeast Asia will fall into an “economic 

water scarcity zone” (Bouman and Tuong 2003). Water 

scarcity threatens the sustainability of irrigated rice ecosystems 

since it may no longer be feasible for farmers to undertake 

wet cultivation and flood in fields to ensure good crop 

establishment and control weeds. The development and adoption 

of alternative irrigation strategies such as “alternate wetting 

and drying” and “aerobic rice” may enable good crop 


growth but the lack of sustained flooding will greatly increase 

potential losses from weeds. These systems may integrate direct 

seeding and herbicide use, yet, to be sustainable, effective 

weed management strategies are required. 

In many areas of Asia, transplanting of rice is being replaced 

by direct seeding as farmers respond to increased costs 

or decreased availability of labor or water (Pandey and Velasco, 

this volume). Direct-seeded systems tend not to be as robust 

as transplanting as elements of soil moisture, irrigation, drainage, 

and weed control are more critical to successful crop establishment 

and growth. With management playing a more 

decisive role, as a practice, direct seeding could be described 

as “knowledge-intensive.” 

Technical constraints and economic demands are leading 

to change in production systems and are causing farmers 

to be increasingly dependent on information from outside 

sources. We discuss below some issues related to direct seeding, 

its adoption, and weedy rice and herbicide resistance, for 

which farmers could benefit from having a greater availability 

of current knowledge. 

Weed management issues related to direct seeding 

As described above, weed management is one of the major 

constraints to direct seeding. Studies in the rice-wheat system 

in India, in Malaysia, and in Bangladesh (this volume) have 

established a substantial knowledge base on weed management 

in some of the most productive rice systems in Asia. These 

systems have commonality in many of the weed species present 

but also reflect a continuum from transplanting to direct seeding. 

Four pertinent observations from this database can be 

made. 

First, there is considerable variability in soil type, irrigation, 

and drainage, and at the farm level because of resources 

and cropping practices that affect the selection of direct-seeding 

options. 

Associated with direct seeding is an inevitable shift in 

the weed flora toward competitive grasses, including 

Echinochloa species, Leptochloa chinensis, and Ischaemum 

rugosum in wet-seeded rice and in dry-seeded rice the perennial 

sedge Cyperus rotundus. Management of such weeds requires 

farmers to have the ability to anticipate changes in weed 

populations and, to reduce losses, exploit integrated strategies 

comprising tillage, water, and crop management to complement 

herbicide application. 

Third, “weedy” rice (Oryza sativa) has become a relatively 

recent problem in Southeast Asia although “red” rice 

has been known in Latin America for decades. In the late 1980s, 

it was observed in direct-seeded rice of Malaysia and by the 

mid-1990s several of the major rice production schemes had 

infested areas (Azmi et al, this volume). In the Philippines, 

weedy rice was reported in 1990, in Vietnam in 1994, and it is 

now widespread in the central region of Thailand. Its vigorous 

growth results in serious yield losses, and its rapid spread makes 

it a considerable threat to direct-seeded rice production. Control 

strategies combining preventive and cultural measures, 

however, have been shown to be effective. Prevention is one 

of the immediate steps that should be implemented in many 

countries. This may involve sensitizing farmers to risks, closer 

inspection of seeds, and roguing at the initial appearance of 

weedy rice in fields. 

Finally, intensive herbicide use in rice has resulted in 

certain weed species, including Sphenoclea zeylanica and 

Fimbristylis miliacea, developing resistance to 2,4-D herbicide 

(Watanabe et al 1997). Lately, possible ALS (acetolactatesynthase) 

inhibitor-resistant biotypes of Bacopa rotundifolia 

and Limnophila erecta have been reported (Azmi and Baki 

2003). The risks of herbicide resistance evolving are known 

and strategies exist to mitigate the risks of occurrence. What is 

required is a greater awareness within farm communities of 

the resistance problem and suitable management strategies. 

Information needs for improved weed management systems 

As argued elsewhere in this volume by various authors, the 

transition from traditional transplanting to direct seeding is 

knowledge-intensive and renders ineffective the experience of 

traditional rice production systems with their reliance on indigenous 

knowledge and manual inputs (often where opportunity 

costs of labor are discounted). Information must be acquired 

from external agencies and uncertainty is often involved 

over the reliability of information and the implementation of 

actions. Decision rules or heuristics (see Heong and Escalada 

1997 for discussion) may address this difficulty but, to be successful, 

must focus appropriately and in context. 

Table 1 categorizes the issues that must be addressed in 

a decision support framework for weed management in directseeded 

rice. Since decision support frameworks may apply at 

different scales (regional, production system in locality, and 

individual farms) and have multiple target audiences 

(policymakers, researchers and extension personnel, and farmers), 

it is important to delineate the context with respect to 

production systems. Characterization of the existing cropping 

system(s) through constraint analyses enables identification of 

the magnitude and variance of yield gaps caused by the presence 

of weeds. In direct-seeded rice, these may be large (typically 

10–50% of attainable yield) and are often a consequence 

of an inability to implement timely weeding (20–30 DAS), 

which in turn may reflect difficulties in water management (irrigation 

and drainage), land leveling, and appropriate water 

management for selected herbicides. Constraints in credit, 

knowledge, and resources at the farm level contribute to poor 

agronomic practice. Characterization studies in weed management 

typically involve on-farm trials comparing yields from 

plots with intensive weed management as opposed to farmers’ 

practices, coupled with focus-group interviews to describe 

farmer decision-making in weed management and to explore 

perceptions and the availability or lack of information. 

Comparative technology evaluation provides a partial 

budgeting analysis of system components and identifies intervention 

points for change. Typically, and at its simplest, a matrix 

of components versus systems is constructed and subjected to 


Table 1. Issues to be considered in a decision support framework for direct-seeding options and weed management. 

Domains 

Policy adviser 

Focus 

Extension 

Farmer 

Socioeconomic 

Biophysical characteristics of production 

systems 

Weed managementtechnologies 

Implication of changes to establishment 

or weed management 

practices: 

Herbicide use 

Displacement of labor 

Patterns of land and water 

use 

Farm productivity 

Public/private partnerships 

Output: policy brief 

 

 

 

 

 

Yield potential 

Yield gaps because of weeds 

Stability of rice-based cropping 

systems 

Diversification of cropping 

systems 

Crop establishment methods 


Choice of mechanical/chemical 

control 


Cost-benefit analysis of component 

technologies and system 

profitability 

Output: technology options/knowledge 

banks 

Integration of weed management 

technologies, agronomic 

practices, and cultivar 

selection 

Prevention of herbicide resistance 

Preventive methods to prohibit 

weed species shifts 

Environmental protection 


banks 

Information on individual component 

technologies 


banks 

Choice of weed management 

methods in relation to 

labor 

knowledge 

resource availability 

input costs 

Output: leaflets, advice/decision 

trees 

Weed management options in 

relation to cropping systems, 

site location, water management, 

and other agronomic variables 


trees 

Information on 

Target weed species 

Herbicide use 

Timing of application 


trees 

sensitivity analysis by partial budgeting. Pandey and Velasco 

(2002) have pointed to the likelihood of tradeoffs. For instance, 

the introduction of direct seeding may have a negative impact 

on labor demand for the adoption of direct seeding in place of 

transplanting or a positive effect where it allows for cropping 

intensification, both of which may have policy implications 

(Pandey and Velasco 2002). Likewise, improvements in water-use 

efficiency may be achievable only by a switch in herbicide 

product (e.g., butachlor to pretilachlor), which reduces 

the spectrum of control and necessitates more costs in later 

manual weeding. 

As a third framework component, decision trees provide 

farm-level information in the form of structured questions 

that enable answers in the form of options to be chosen. 

These trees specifically focus on technical issues related to 

the adoption of a particular system component and are most 

effective when heuristics are employed. The question “Are 

weeds that emerge early in the life of a rice crop the only ones 

to contribute to yield loss” may initiate a tree that can lead to 

the recognition of the risk of intransigent weeds that in practice 

require identification skills in scouting rice fields. For a 

farmer in the favorable rainfed environment, the question “What 

are my options for rice establishment” might initiate a tree 

involving several steps and covering the range of direct-seeding 

options (see Fig. 1 for an example). 

Conclusions 

Traditional and modern rice systems alike are complex, the 

management often sophisticated, and substantial information 

is required to enable farmers to judge objectively what the 

best technology options are. This is certainly true in the transition 

to direct seeding and the management of intractable weed 

problems. Gaining access to such information may present a 

major obstacle for potential adopters. The challenge for researchers 

is to adequately address the variability of rice-farming 

systems for which they are making recommendations and 

to synthesize the results in ways that will make the conclusions 

available to those who will use them. Only in this way 

will farmers obtain benefit from advances in technology and 

be able to meet the challenges of a changing production environment. 


Crop establishment 

Can field be drained 

Yes 

No 

Can field be dry-cultivated 

Is Cynodon dactylon or Cyperus rotundus absent 

Transplant 

Yes 

No 

Dry seeding into a seedbed 

Wet seeding, sowing onto 

puddled saturated soil 

Are soil conditions suitable for line 

seeding by machinery 

Is there a need for interow cultivation 

or substantial hand weeding 

Yes No Yes No 

Broacast Drill-seed Broacast Drum-seed 

Weed management 

Are annual grasses absent 

Good water management possible 

Yes 

No 

Apply herbicide 

+ limited manual weeding 

Apply herbicide 

+ limited manual weeding or 

interflow cultivation 

Fig. 1. Illustrative decision tree for adoption of direct seeding with respect to favorable rainfed lowland rice. 

References 

Azmi M, Baki BB. 2003. Weed species diversity and management 

practices in the Sekinchan Farm Block, Selangor’s South West 

Project—a highly productive rice area in Malaysia. In: Proceedings 

1, 19th Asian-Pacific Weed Science Society Conference, 

17-21 March 2003, Philippines. p 174-184. 

Bouman BAM, Tuong TP. 2003. Growing rice with less water. Issues 

of water management in agriculture: compilation of essays. 

Comprehensive Assessment Secretariat, IWMI, Colombo, 

Sri Lanka. p 49-54. 

Heong KL, Escalada MM. 1997. Perception change in rice pest 

management: a case study of farmers’ evaluation of conflict 

information. J. Appl. Comm. 81:3-17. 

IRRI (International Rice Research Institute). 2004. World rice statistics. 

Manila (Philippines): IRRI. www.irri.org/science/ 

ricestat. 

Naylor R, editor. 1996. Herbicides in Asian rice: transitions in weed 

management. Palo Alto, Calif. (USA): Institute for International 

Studies, Stanford University, and Manila (Philippines): 


Pandey S, Velasco L. 2002. Economics of direct seeding in Asia: 

patterns of adoption and research priorities. In: Pandey S, 

Mortimer M, Wade L, Tuong TP, Lopez K, Hardy B, editors. 

Direct seeding: research issues and opportunities. Proceedings 

of the International Workshop on Direct Seeding in Asian 

Rice Systems: Strategic Research Issues and Opportunities, 

25-28 January 2000, Bangkok, Thailand. Los Baños (Philippines): 


Watanabe H, Zuki MI, Ho NK. 1997. Response of 2,4-D resistant 

biotypes of Fimbristylis miliacea (L.) Vahl to 2,4-D 

dimethylamine and its distribution in Muda plain, Peninsular 

Malaysia. J. Weed Sci. Technol. 42(3):240-249. 

Notes 

Authors’ addresses: D.E. Johnson, International Rice Research Institute, 

DAPO Box 7777, Metro Manila, Philippines; A.M. 

Mortimer, School of Biological Sciences, University of 

Liverpool, Liverpool, UK. 


Control of Leptochloa chinensis (L.) Nees 

in wet-seeded rice fields in Sri Lanka 

Anuruddhka S.K. Abeysekera and U.B. Wickrama 

Weeds are the major biotic stress in rice cultivation in Sri Lanka, 

causing yield losses of as much as 30–40% (Herath Banda et 

al 1999). Leptochloa chinensis (L.) Nees, Red Sprangletop, is 

becoming an important and serious threat to rice production in 

all rice ecosystems in Sri Lanka. It has the same outbreak level 

as Echinochloa crus-galli (L.) Beauv. (Abeysekera 1999). 

Chemical weed control is commonly practiced by rice farmers 

in Sri Lanka (90%) and propanil 36% EC was the most widely 

used grass-killer herbicide in rice fields before the introduction 

of bispyribac sodium (Abeysekera 1999, Amarasinghe and 

Marambe 1998). In 1998, bispyribac sodium was the leading 

herbicide and L. chinensis was selected as a serious weed in 

wetland rice fields. After it germinates, it grows profusely in 

waterlogged spots in rice fields because of very poor land leveling 

in farmers’ fields. At present, many herbicides have difficulty 

in controlling this weed. Therefore, the objective of 

this study was to evaluate the control efficacy of L. chinensis 

with postemergence herbicides in wet-seeded rice. 


A field experiment was conducted at the Rice Research and 

Development Institute, Sri Lanka, during the 1999 minor and 

1999-2000 major season. The soil type was red-yellow podsolic 

with pH 6.2 and annual rainfall of the region at 2,250–2,285 

mm, with an average temperature of 23–28 ºC. Rice (Oryza 

sativa) variety Bg 300a (3 months old) was used and broadcast 

at 100 kg ha –1 . The experimental design was a randomized 

complete block design with 3 replicates; plot size was 6 × 

3 m 2 , each separated by 30-cm bunds. Six common 

postemergence herbicides were used: (1) clomozone + propanil 

60% EC (Cl + Pro) at 1.2 L ha –1 applied at 7–10 days after 

sowing (DAS), (2) cyhalofop-butyl 100% EC (Cy) at 1 L ha –1 

applied 7–10 DAS, (3) fentrazmide + propanil 44% WP (Fe + 

pro) at 3 kg ha –1 applied at 7–10 DAS, (4) propanil 36% EC 

(Pro) at 8.5 L ha –1 applied at 7–10 DAS, (5) quinclorac 250 g 

L –1 SC (Qui) at 0.6 kg ha –1 applied at 7–10 DAS, and (6) 

bispyribac sodium 10% SC (Bis) at 0.3 kg ha –1 applied at 10– 

12 DAS. These were evaluated with and without removing L. 

chinensis as the control treatment. MCPA was used to control 

broadleaf and sedge weeds at the experimental site and, except 

for L. chinensis, the other grasses were removed manually. 

All the other cultural practices and pest control followed 

Department of Agriculture recommendations. Rice germination 

count 2 weeks after sowing (WAS), visual phytotoxicity 

symptoms 2, 4, 6, and 8 d after herbicide application, the count 

of L. chinensis 4 and 6 WAS, dry weight of L. chinensis at 6 

WAS, yield, and yield components of rice were recorded at 

harvest time. The data were analyzed by ANOVA using the 

SAS computer package and the treatment means were compared 

by Duncan’s multiple range test. 


Grasses were the dominant weed group found at the experimental 

site, where the grass L. chinensis was dominant. 

Echinochloa crus-galli, Isachne globosa, Paspalum distichum, 

and Ischaemum rugosum were also observed. None of the herbicides 

affected the germination and growth of rice seedlings. 

Phytotoxicity 

Cyhalofop-butyl, propanil, fentrazmide + propanil, quinclorac, 

and bispyribac sodium applied to rice plants did not show any 

visual phytotoxicity symptoms in rice plants. Clomozone + 

propanil–treated rice plants showed a slight whitening effect 2 

d after spraying. But these symptoms did not affect the grain 

yield of rice and were not visible after 7 d. 

Weed control efficacy 

The lowest L. chinensis dry weight and count were observed 

from the clomozone + propanil treatment and cyhalofop-butyl 

treatment, followed by fentrazmide + propanil, quinclorac, and 

propanil treatments during both seasons (Fig. 1). The highest 

L. chinensis dry weight was observed from the no-weeding 

treatment, followed by the bispyribac sodium treatment (Fig. 

1). All the herbicides gave excellent control of E. crus-galli 

and I. rugosum but were less effective on I. globosa and P. 

distichum. 

Yield 

The grain yield of rice was highly affected by the biomass 

production of L. chinensis. It showed significant differences 

among herbicide-treated plots and untreated plots. Weed competition 

was more severe in the dry season than in the wet 

season. In both seasons, the highest yield was obtained from 

clomozone + propanil, followed by cyhalofop-butyl, 

fentrazmide + propanil, propanil, quinclorac, and bispyribac 

sodium treatments, respectively (Fig. 2). Infestation of L. 

chinensis reduced rice yield significantly in the respective treatments. 

A significant linear relationship was found between the 

dry weight of L. chinensis and rice grain yield (Fig. 3A, B). 

The highest L. chinensis dry weight gave the lowest rice yield. 

Conclusions 

Cyhalofop-butyl 100% EC at 1 L ha –1 and clomozone + 

propanil 60% EC at 1.2 L ha –1 applied at 7–10 DAS gave excellent 

control efficacy (9% and 94%, respectively) of L. 


Dry weight (g m –2 ) 

40 

35 

30 

Minor 

Major 

a 

a 

a 

25 

20 

b 

15 

b 

bc 

10 

bc 

bc 

5 

c 

c 

d d 

d 

0 

Cl + pro Cy Fe + pro Qui Pro Bis Nw Hw 

Treatment 

Fig. 1. Dry weight of L. chinensis as affected by different herbicides at 6 weeks 

after sowing. 

Yield (t ha –2 ) 

6 

5 

4 

3 

a 

a 

a 

a 

Minor 

Major 

b 

a 

b 

b 

b 

b 

b 

c 

a 

c 

d 

2 

a 

1 

0 

Cl + pro Cy Fe + pro Qui Pro Bis Nw Hw 

Treatment 

Fig. 2. Effect of different herbicides on rice grain yield. 

chinensis. Fentrazmide + propanil 44% WP at 3 kg ha –1 , 

quinclorac 250 g L –1 SC at 0.6 kg ha –1 , and propanil 36% EC 

at 8.5 L ha –1 applied at 7–10 DAS gave moderate control efficacy 

(60%, 58%, and 40%, respectively) of L. chinensis. The 

lowest control efficacy of L. chinensis was observed with 

bispyribac sodium 10% SC at 0.3 kg ha –1 applied at 10–12 

DAS. No adverse effect was observed from all the tested herbicides 

on rice grain yield. 

References 

Abeysekera ASK. 1999. Current status of weed control in rice in Sri 

Lanka. Proceedings of the 17th Asian Pacific Weed Science 

Society Conference, Bangkok, Thailand, 22-27 November. 

p 174-179. 

Amarasinghe L, Marambe B. 1998. Trends in weed control of rice 

cultivation in Sri Lanka. Proceedings of a Multidisciplinary 

International Conference, University of Peradeniya, Sri Lanka. 

p 272-274. 

Bambaradeniya CNB, Gunatilaka CVS. 2002. Ecological aspects of 

weed flora in an irrigated rice field ecosystem in Sri Lanka. 

JNSF, Sri Lanka. 2 p. 

Herath Banda RM, Dhanapala MP, Silva GAC, Hossain M. 1998. 

Constraints to increase rice production in Sri Lanka. Workshop 

on the prioritization of rice research, International Rice 

Research Institute, Los Baños, Philippines. 

Notes 

Authors’ address: Rice Research and Development Institute, 

Batalagoda, Ibbagamuwa, Sri Lanka, e-mail: anuru@sltnet.lk. 



6 

5 

4 

3 

2 

1 


4.5 

4.0 

3.5 

3.0 

2.5 

2.0 

y = –0.0514x + 4.7006 

R 2 = 0.7528 1.5 

1.0 

y = –0.0741x + 3.7667 

R 2 = 0.9622 

0 

0 10 20 30 40 


0.5 

0.0 

0 5 10 15 20 25 30 


Fig. 3. Rice yield as affected by dry weight of Leptochloa chinensis. (A) minor season, (B) major season. 

Village-level modeling of environment-friendly 

and appropriate technologies and practices 

for direct seeding 

A 

Rowena G. Manalili, Bernard D. Tadeo, Emmanuel R. Tiongco, Wilfredo B. Collado, Rodolfo V. Bermudez, Constancio A. Asis, Jovino L. De Dios, Marvin 

F. Adap, Mario dela Cruz, Ulysses G. Duque, Leonardo V. Marquez, Cheryll B. Casiwan, Roy F. Tabalno, Placida C. Lanuza, and Belen C. Tejada 

B 

Direct seeding has increasingly become popular and adopted 

by farmers in many rice-growing areas in the Philippines, especially 

during the dry season. A recent survey of the Philippine 

Rice Research Institute (PhilRice) revealed that more than 

75% of the farmers in the major rice-producing areas are adopting 

direct seeding. Iloilo has the largest area devoted to directseeded 

rice, followed by Occidental Mindoro and Negros Occidental 

(PhilRice-BAS 1997). Wet seeding of pregerminated 

seed in puddled soil is more common than dry seeding and 

more farmers wet-seed during the dry season. 

The economic advantages of this mode of crop establishment 

include reducing the labor requirement as a result of 

the elimination of the seedbed and transplanting operations, 

mitigating methane emissions, advancing harvesting by more 

than 1 week, and other advantages such as using less water to 

mitigate drought and allowing dryland tillage in relation to 

farm management techniques and practices. 

A gamut of mature technologies exists for the directseeding 

rice ecosystem. A strategy identified for more effective 

promotion of technologies is to develop a village into a 

community of farmers for direct seeding, where indigenous 

and modern technologies and practices merge. The village 

model serves as a springboard for developed technologies before 

these are eventually promoted. It is where technologies 

are fine-tuned with the participation of farmers. The community 

is envisioned to use ecologically sound, sustainable, and 

affordable practices and is closely monitored and supervised 

by an interdisciplinary team of research specialists and extension 

experts from the Philippine Rice Research Institute. The 

project is piloted in the Agrarian Reform Community of 

Maragol, Science City of Muñoz, Nueva Ecija, Philippines. 

The objectives of the study were to (1) enhance the capacity 

of rice farmers, (2) increase the yield and income of 

rice farmer-partners by promoting improved rice production 

technologies, (3) link and network farm products to the mainstream 

and local market, and (4) assess the socioeconomic 

impact of technologies and provide insights for researchers. 

Methodology 

Site selection was conducted based on the accessibility of the 

area: large contiguous irrigated rice areas and farmers practicing 

direct seeding. 

A baseline survey and participatory rural appraisal were 

conducted to generate a location-specific database crucially 

important in the determination of rice farm households’ technology 

base, needs, and problems. Site mapping was also conducted 

using the global positioning system (GPS). 

Spatial and temporal pest dynamics in the area were 

monitored. Rat population dynamics were monitored using the 

trap barrier system + trap crop (TBS + TC) for four rice-cropping 

seasons in 2002-04 (PhilRice 2000). The soil test kit 


Site selection 

Support 

services 

Biophysical and 

socioeconomic 

characterization 

Technology 

adaptation, 

demonstration, 

and promotion 

Village-level 

adoption of 

technologies 

Monitoring and 

evaluation 

Fig. 1. Schematic diagram of the methodology in the direct-seeding rice village model. 

(STK), minus one element test (MOET), nutrient omission plot, 

and laboratory analysis were used to assess the nutrient status 

of various rice fields at the study site (PhilRice 2002). 

Researchers and farmers identified technologies and 

adaptable farmers’ practices for direct-seeded rice. Demonstration 

trials on the package of technologies, which include 

the use of 40 kg ha –1 of certified seeds, the drum seeder, and 

carbonized rice hulls and organic fertilizer, were set up in the 

village. The performance of hand-tractor-drawn riding-type 

implements such as the harrow and leveler was also tested. 

The technologies introduced are being monitored through 

the farm record-keeping (FRK) approach. Farm record-keeping 

calendars were distributed to farmer-partners every cropping 

season for them to record their daily farm activities, expenses, 

and revenues. Monitoring, documentation, and copying 

of data were done on a weekly basis by technical staff to 

check whether farmers were religiously recording their daily 

farm activities, and problems encountered by the farmers were 

discussed. The gathered data were coded and encoded using 

Microsoft Access. These were processed and analyzed using 

frequencies, means, averages, and costs and returns analysis. 

The profitability and productivity between direct-seeded and 

transplanted rice were also compared. 

Figure 1 shows the schematic diagram of the methodology 

for the direct-seeding rice village model. 


The project focused on about 200 ha of contiguous rice fields 

in the Agrarian Reform Community of Maragol. It is one of 

the 37 barangays in the Science City of Muñoz, Nueva Ecija. 

It has a population of 2,294 or around 480 households. It has a 

total land area of 520 ha, of which 490 ha is devoted to rice 

production. Rice is planted twice a year with the National Irrigation 

System (NIS) as the main source of irrigation. Direct 

seeding is practiced by 90% of the farmers during the dry season. 

However, 90% of the farmers practiced transplanting dur- 

ing the wet season because of drainage problems as a result of 

the difficulty of water control. 

The farmers’ association and organization are considered 

as an effective conduit in extending assistance to individual 

farmers. A farmers’ organization called Ugnayan ng 

Magsasaka Makabagong Teknolohiya at Yamang Kalikasan 

para sa Kaunlaran ng Maragol, Inc. (UMAYKAMI) was 

formed, and was officially registered at the Securities and Exchange 

Commission (SEC). Identified leaders were trained on 

basic leadership and community empowerment. Its members 

were taught about the technologies in direct seeding. 

The population of insect pests was higher during the late 

growth stages of the rice crop. Occurrence of major pests was 

prevented by the 2-month fallow period and the use of resistant 

varieties. Using the trap barrier system + trap crop (TBS 

+ TC), it was observed that fewer rats were trapped as soon as 

the TBS +TC was established. This increased a week after the 

start of the field operations because of flooding and the destruction 

of rat burrows in the levees and surrounding areas. 

More rats were trapped up to the early maximum tillering period 

than at the later rice growth stages. Community-wide physical 

rat control was recommended at the start of the cropping 

season. It was recommended that the time to use baits was 

during the early rice growth stages, not throughout the cropping 

period as is now being practiced. 

Relative grain yields of the nutrient omission plots ranged 

from 50% to 98% compared with those of the complete plots. 

Results showed that indigenous nitrogen (N) and potassium 

(K) were limited in supply but not phosphorus (P). The recommended 

rate of 90-40-40 kg NPK ha –1 provided an average 

grain yield of 5 t ha –1 . However, by increasing the yield target 

to 7 t ha –1 , the soil nutrient supply would be limited for all 

three nutrients (N>K>P) and a higher P application would be 

needed to attain that yield target. 

The indigenous nitrogen-supplying capacity (INS) and 

the agronomic efficiency of applied N (AEN) of the farmercooperators’ 

fields ranged from 50 to 70 kg N ha –1 and 5 to 25 


Table 1. Comparative costs and returns analysis of direct-seeded (DSR) and transplanted (TPR) rice production in Maragol, Science City of Muñoz, Nueva Ecija, Philippines, 2001-03. 

Indicator 2001 2002 2003 

Dry season Wet season Dry season Wet season Dry season Wet season 

TPR DSR TPR DSR TPR DSR TPR DSR TPR DSR TPR DSR 

Gross returns (pesos ha –1 ) 39,682 44,435 35,032 – 52,773 58,673 25,478 30,353 54,298 56,063 34,861 32,250 

Yield (t ha –1 ) 4.90 5.49 4.30 – 6.21 6.90 3.23 3.84 6.46 6.67 4.73 4.30 

Price (pesos kg –1 ) 8.10 8.10 8.15 – 8.50 8.50 7.90 7.90 8.41 8.41 7.64 7.50 

Production costs (pesos ha –1 ) 

Materials 7,081 6,535 4,216 – 6,795 6,977 5,488 5,425 5,998 6,514 4,832 3,865 

Labor 14,173 11,997 11,940 – 16,950 14,785 10,944 10,639 15,445 13,044 11,538 7,324 

Other costs 2,066 5,407 6,306 – 7,080 7,446 7,908 3,664 4,966 5,879 6,308 4,679 

Total costs (pesos ha –1 ) 23,319 23,939 22,462 – 30,825 29,208 24,340 19,728 26,409 25,438 22,679 15,868 

Net income (pesos ha –1 ) 16,363 20,496 12,570 – 21,948 29,464 11,138 10,625 27,889 30,625 12,182 16,383 

Net profit-cost ratio 0.70 0.86 0.56 – 0.71 1.01 0.05 0.54 1.06 1.20 0.54 1.03 

Cost per kg (pesos ha –1 ) 4.76 4.36 5.23 – 4.96 4.23 7.55 5.13 4.09 3.82 4.79 3.69 

kg grain produced per kg N applied, respectively. These results 

showed that the areas tested have a high N-supplying 

capacity and that a small amount of N is needed as basal fertilizer 

or during the early vegetative stage. 

The minus one element test (MOET) showed that most 

of the fields were deficient in macronutrients but not in micronutrients. 

Farmers were encouraged and trained to optimize the 

use of biomass residue from the farm, for example, carbonization 

of rice hull (CRH) and animal manure as organic fertilizer 

and soil conditioner. The use of inorganic fertilizer was reduced. 

The performance of the ride-on tillage implements for 

hand tractors was tested in farmers’ fields. The technology for 

the 40–60 kg ha –1 of certified seeds was demonstrated with 

the use of the hand-tractor-drawn paddy seeder and the manually 

drawn drum seeder. The use of 40–60 kg ha –1 of certified 

seeds not only reduced the input costs in rice production but 

also helped save seeds for other farmers’ use. This also had a 

yield advantage of 10% over the use of ordinary seeds. 

Most of the farmers belonged to the medium- (3–5 t 

ha –1 ) and high-yield group (>5 t ha –1 ). Direct-seeded rice appeared 

to be more profitable than transplanted rice as indicated 

by its higher gross returns, net income, and net profitcost 

ratios and its lower cost of production. Table 1 shows the 

comparative costs and returns analysis between direct-seeded 

and transplanted rice production in the village. 

Conclusions 

Village-level integration is an effective strategy to verify and 

promote technologies for direct-seeding rice production. The 

strong collaboration among farmers, local government units, 

village officials, and researchers has contributed to the successful 

implementation of the project. 

The upcoming activities of the project include verification 

trials based on the results of the MOET and NOPT; provision 

of a fertilizer domain and delivery of nutrient management 

technologies; development of an integrated rat management 

system; linkage with support services such as inputs and 

product markets, and credit services; and impact evaluation of 

introduced technologies. 

References 

PhilRice-BAS (Philippine Rice Research Institute-Bureau of Agricultural 

Statistics). 1997. Survey, 1997. Maligaya, Science City 

of Muñoz, Nueva Ecija, Philippines. 

PhilRice (Philippine Rice Research Institute). 2000. Management 

of field rats. Rice Technology Bulletin No. 28. DA-PhilRice 

Maligaya, Science City of Muñoz, Nueva Ecija, Philippines. 

12 p. 

PhilRice (Philippine Rice Research Institute). 2002. Integrated nutrient 

management for rice production. Technology Bulletin 

No. 45. DA-PhilRice Maligaya, Science City of Muñoz, Nueva 

Ecija, Philippines. 24 p. 


Notes 

Authors’ addresses: R. Manalili, B. Tadeo, E. Tiongco, W. Collado, 

R. Bermudez, C. Asis, J. De Dios, M. Adap, M. Dela Cruz, U. 

Duque, and L. Marquez, Philippine Rice Research Institute, 

Maligaya, Science City of Muñoz, Nueva Ecija 3119, Philippines; 

P. Lanuza and B. Tejad, Department of Agriculture- 

Local Government Unit, Science City of Muñoz, Nueva Ecija 

3119, Philippines, e-mail: rgmanalili@philrice.gov.ph; 

btadeo@philrice.gov.ph. 


The change in rice crop establishment from transplanting to direct 

seeding has occurred in many Asian countries because of 

rapid economic growth, increasing labor costs, and shortages of 

water. In addition, there is a need to reduce the costs of rice 

production in order to maintain profitability despite the declining 

market price, a trend that is expected to continue in the future. A 

change in cultivation methods could have a substantial influence 

on the environment; therefore, a change from transplanting to 

direct seeding should be carried out with a focus on the sustainable 

use of water and concern regarding agricultural chemical 

use. 

Direct seeding is practiced in three main forms: dry seeding, 

wet seeding, and water seeding. In dry seeding, 

nongerminated seeds are sown beneath the soil surface. The 

plant stand is more consistent than in wet and water seeding as 

the soil is aerobic. Dry seeding is advantageous when the water 

supply is limited at the time of land preparation and crop establishment. 

The cultivation system of “aerobic rice” in China, which 

combines dry seeding with a “nonflooding” water regime, is suggested 

as a water-saving technology. As the land is not flooded 

and puddled, heavy weed infestation and high rates of water infiltration 

are problems in dry seeding. 

Wet seeding is a widely practiced form of direct seeding in 

which the land is flooded and puddled in the same way as in 

transplanting. Pregerminated seeds are sown on the puddled soil 

surface just after drainage. Drainage is important to achieve reliable 

crop establishment. In Malaysia, Vietnam, Thailand, the Philippines, 

and Korea, the repeated use of herbicides and limited 

irrigation water have seriously complicated weed problems with 

the rapid shift in weed populations to annual grasses (e.g., 

Ischaemum rugosum and Leptochloa chinensis) and weedy rice 

(Oryza sativa) that are difficult to control. 

In Japan, wet seeding is practiced differently from the 

method used in other Asian countries. Pregerminated seeds are 

commonly coated with an “oxygen-supplier” based on calcium 

peroxide and seeded under the surface of puddled soil. As the 

coating alone is not sufficient to overcome anaerobic soil conditions, 

the field must be drained at the crop establishment stage. 

Seeding beneath the puddled soil surface has the advantage of 

reducing seedling float and bird damage. Further, “hill seeding” 

of coated seeds under the soil surface using a specialized seeder 

has been shown to reduce crop lodging. 

Sowing pregerminated seeds into standing water (water 

seeding) has advantages over wet seeding in terms of water savings 

and weed control. The water used for puddling is not drained 

and conserved in the field, whereas the presence of water in the 

paddy field, particularly at the time of rice crop establishment, 

efficiently suppresses weed growth. A major limitation of water 

seeding, however, is the tendency of seedlings to float, which 

causes uneven rice populations. 

Direct seeding in place of transplanting can have a positive 

impact on the productivity of cropping systems, particularly 

in rainfed systems where water availability is limited. In the rainfed 

and drought-prone areas of Bangladesh, advancement of the rice 

crop harvest through direct seeding increases the chance of a 

second crop. Direct seeding also has advantages in the ricewheat 

cropping system of northern India, where water availability 

and labor cost are the major determinants of successful farming. 

Weed infestation is a major threat to yield and the further expansion 

of direct seeding in Asia. Where direct seeding has been 

introduced, there is commonly not only an increase in weed biomass 

but also a change in the weed species. Repeated use of 

the same herbicide or those with the same mode of action risks 

the occurrence of herbicide resistance in weeds, and biotypes 

with resistance to acetolactate synthase inhibitor have been reported. 

Furthermore, the control of weedy rice presents major 

challenges. 

With the high development costs for new herbicides, the 

agrochemical industry currently focuses on novel uses and mixtures 

of current active ingredients in order to solve new weed 

problems associated with direct-seeded rice. Water management 

in terms of timing, duration, and depth of flooding at the crop 

establishment stage is critical in determining the scope for the 

use of herbicides. Understanding the biology of weeds in relation 

to water profile is important in developing ecologically sound weed 

control methods and to enable the exploitation of flooding for 

weed control. A new technology that coats rice seeds with iron 

powder to increase the specific gravity would overcome the problem 

of seedling float and make water seeding possible. This may 

provide a means to control weedy rice and water seeding is suggested 

as a potential control measure. Thus, the weed problems 

associated with direct-seeded rice could be alleviated to some 

extent by greater emphasis on combined herbicide use and water 

management. 


Crop and weed management are expected to become increasingly 

knowledge-intensive as direct seeding becomes widespread 

and as technology develops. To meet the challenges of 

the declining availability of labor and water, increasingly difficult 

weed problems, and growing concerns regarding sustainability, 

farmers require greater access to information and guidance to 

enable them to make more objective decisions for crop and weed 

management. 


SESSION 7 

Improving efficiency through innovations 

in mechanization 

CONVENER: N. Itokawa (NARO) 

CO-CONVENER: J. Rickman (IRRI)

Rice as a key resource for saving our planet 

Nobutaka Ito 

Rice has become an extremely important crop being cultivated 

worldwide. In fact, the total production of rice in the world 

has reached almost 500 million tons and 90% of the total production 

comes from Asian countries. Taking into consideration 

the world population of 6.3 billion, food per capita per year 

has been estimated at 400 kg. This rough estimate was based 

on the total world production of rice in addition to that of other 

cereals (such as wheat, sorghum, soybean, and maize), almost 

equal to 2 billion tons, plus 400 million tons from vegetables, 

fruit, and meat. However, this amount is just enough for now 

and the Chinese government would like to increase production 

to satisfy a growing population. In this paper, basic data 

related to rice and its cultivation are discussed prior to developing 

solutions to increase rice production and protect the environment. 

Rice is cultivated basically for food consumption; however, 

it can also be used as energy in the form of an environment-friendly 

fuel. This is possible by using its biomass, for 

which production efficiency is the highest among agricultural 

crops. About 430 L of ethanol can be produced from 1 ton of 

rice biomass. Reliability is based on the fact that almost 17 to 

19 tons of carbon dioxide per hectare are absorbed by the growing 

rice plants, which is nearly equal to the amount of carbon 

dioxide absorbed by some forest trees. In addition, the rice 

maturity stage takes only 4 months, whereas trees take more 

than 10 years to reach maturity. In this case, ethanol from rice 

can easily be obtained as the demand for energy consumption 

increases. This will also help minimize the world’s dependency 

on fossil fuel. The most significant advantage of using ethanol 

produced from biogas is the clean emission of gas compared 

with that of fossil fuel during combustion. 

These facts illustrate the important role of rice in helping 

resolve four global issues (called tetralemma), particularly 

concerning the increased consumption of fossil fuel, which 

jeopardizes the environment. This condition will sooner or later 

result in fossil fuel exhaustion; therefore, we must develop an 

alternative solution to combat this threat. 

New technology in rice mechanization 

Direct sowing by use of coated rice 

Direct sowing of rice is popularly accepted in many developing 

countries because of its low cost and energy-saving technology 

in planting; however, the yield per unit area is almost 

half that obtained in many developed countries. Only the United 

States practiced direct sowing of rice successfully, combined 

with low-cost production and high application of technology. 

Yield is almost 7.7 t ha –1 , which is 1.1 t ha –1 higher than in 

Japan’s case using the transplanting method. A new type of 

direct sowing by using calcium peroxide-coated rice seed has 

good potential because of its low production cost. An automatic 

coating machine has already been developed and commercialized. 

Coating technology will be expanded for the purpose 

of reducing some operations by the simultaneous application 

of herbicide, pesticide, and fertilizer. 

Combine harvester 

A special head-feeding-type self-propelled combine harvester 

was developed and is commercially available in Japan. This 

machine can harvest from two to six rows depending on the 

specified width of cut. I present new concepts of the machine’s 

functions below. 

Combine husker (named by the author) 

Normally, cutting of straw, threshing, winnowing, and cleaning 

of grains are the main functions of a commercial rice combine 

harvester. However, husking can be done simultaneously 

through an additional mechanism I designed. Using the combine 

husker, rice can be directly harvested as brown rice, throwing 

out other foreign materials such as rice chaff, husks, empty 

grains, and weeds after husking and scattering them on the 

paddy field. The following improvements can be made by introducing 

this type of harvester: first, a savings of energy for 

transportation because of the reduction in bulk density through 

removal of rice husks and other foreign materials during harvesting; 

second, a saving of drying space equal to almost half 

of the rice volume. 

The turntable combine harvester 

To improve operating efficiency, the total time for completing 

the harvesting operation and the total distance traveled to complete 

the harvesting operation are normally used for evaluating 

how much this could be improved. The turntable combine 

harvester turns normally using the turntable mounted on the 

main chassis. The shape ratio of the paddy field expressed by 

the ratio of length to width, the pattern of promoting harvesting 

operation, the operating speed, the operating width of the 

cutter bar, and the specifications of the machine must be considered 

as the factors for improvement. For the turntable combine 

harvester, the total distance required to complete harvesting 

could be reduced by 10% compared with that of the conventional 

commercial machine. The total distance under the 

same conditions could be reduced almost 25% at maximum. 

The turntable combine harvester can be recommended for its 

safety, automation, and energy savings. 

The combine harvester equipped 

with a pivot turn mechanism 

The improvement in turnability of the tracked vehicle can be 

achieved by applying the pivot turn mechanism. The ground 


contact length of the braked track has a key for that purpose 

because the width of the track cannot be changed or controlled 

once it has been mounted or equipped. 

Brown rice dryer 

As mentioned above, the removal of rice husks during harvesting 

can save dryer space and allow normal air to be used 

for efficient brown rice drying only. A prototype brown rice 

dryer has already been fabricated for the test. There is no need 

to use hot air for drying, which produces carbon dioxide 

through kerosene burning of rice chaff. 

The rice transplanter using a long seedling mat 

In automated transplanting operations, time consumed for a 

frequent supply of nursery plants is one obstacle for continuous 

machine operation. The capacity for carrying a young nursery 

or rice seedlings should be big enough to eliminate such a 

problem. A long nursery plant mat is one alternative for solving 

this problem. 

Precision farming 

Precision farming can be characterized mainly by (1) hightechnology 

agricultural machinery, (2) large-scale farming, (3) 

combined application of a global positioning system and geographic 

information system, and so on. Precision farming is 

seemingly impossible to introduce in Japanese conditions. 

However, newly reclaimed national area can be used for possible 

validation. This possibility is intended for improving food 

self-sufficiency. Furthermore, this would assure safe food production 

with a satisfactory level of security. Precision farming 

contributes a lot as an environment-friendly technology. 

The world faces many threats. This paper has presented some 

suggestions to help overcome these threats. First, the world 

requires enough food, particularly rice, as the world population 

is increasing dramatically. Second, world dependency on 

fossil fuel must be minimized. Third, emissions of carbon dioxide 

must be controlled. We must improve rice production 

and consider the crop’s potential. Mechanization has a vital 

role in this endeavor as it functions as a force multiplier to 

compensate the human labor shortage for those engaged in 

food production. This includes developing new technologies 

to strengthen rice production and add more value to make the 

crop more competitive in the world market. Biomass derived 

from rice can provide us with ethanol to be used as an environment-friendly 

fuel. Thus, pollution from toxic gas emissions 

with fossil fuel will be lessened and a safe environment restored. 

However, in Japan, any increase in rice production is 

controlled politically to keep the price high and maintain the 

farmers’ standard of living. This condition protects farmers’ 

lives but tends to neglect human welfare. This should not be 

the case because it is more important to produce enough to 

achieve a good distribution of food globally and keep the environment 

safe. Rice has the potential to overcome various 

global threats. I also hope that environmental protection will 

be given much more priority. 

Bibliography 

Ito N. 2002. Global tetralemma and its overcome through agri-techno 

fusion strategy. J. Electrification 55(5):7-12. 


fusion strategy. J. Electrification. 55(6):8-13. 


fusion strategy. J. Electrification 55(7):15-21. 

Ito N. 2000. Global tetralemma and its overcome through rice production. 

Presented at the 5th Regional Science and Technology 

Policy Research Symposium, 5-7 September, Mie, Japan. 

Conclusions 

Notes 

Mechanizing paddy rice cultivation in Korea 

Woo-Pung Park and Sang-Cheol Kim 

Author’s address: Laboratory of Energy Utilization Engineering, 

Department of Environmental Science and Technology, Faculty 

of Bio-resources, Mie University, e-mail: ito-n@bio.mieu.ac.jp. 

The field agricultural production system in Korea has focused 

on paddy rice and plot size is relatively small (less than 1 ha). 

Until the late 1950s, manual and animal power were used in 

rice production. With the execution of the Five-Year Economic 

Development Plans started in the early 1960s, agricultural 

machinery such as water pumps and pesticide applicators were 

disseminated for disaster-prevention purposes. In the 1970s, 

as a large portion of rural labor moved to factories in urban 

areas, full-scale agricultural mechanization projects began. The 

power tiller replaced animal power in the 1970s, and rice trans- 

planters and harvesters were distributed in the 1980s to overcome 

labor peaks and shortages. 

Agricultural mechanization technologies were developed 

and extended in the 1990s to improve the agricultural system 

and productivity. The change from manual and animal power 

to about 98% mechanization in rice production took place 

within only 30 years, which is regarded as a significant achievement 

by other developing countries. 

Session 7: Improving efficiency through innovations in mechanization 225

Table 1. Mechanization rate (%) in rice production. 

Year Tillage Transplanting Spraying Harvesting Drying 

1985 70 23 68 17 2.1 

1990 84 78 93 72 15 

1995 97 97 98 95 32 

2002 100 98 100 99 49 

Review of agricultural mechanization 

Population changes in the rural community 

The farm population was 58% of the national population in 

1960, but it gradually decreased as labor demand for other 

industries increased. The number of farmers decreased, from 

5.0 million in 1975 to 4.4 million in 1980, 3.1 million in 1990, 

and 2.2 million in 2000. The annual reduction rate is above 

3%. The current labor shortage is caused by (1) an increasing 

portion of old and female labor and (2) the higher labor wages 

in the rural community have promoted farm mechanization 

more actively. 

Farmers from 20 to 59 years old gradually decreased 

and farmers more than 60 years old gradually increased. In 

2000, farmers older than 60 made up 33.1% of the farm population. 

This is a big problem in Korea. 

Expanding farm size 

Farm size averaged 0.9 ha in 1965. It increased to 1.37 ha in 

2000. However, the farm size is not big enough to carry out 

farm-restructuring effects. A small farm is defined as a field 

less than 1.0 ha in size. Medium-sized farms from 1 to 2 ha 

have decreased rapidly. 

On the other hand, the number of large farms (more than 

2.0 ha) has been increasing. The ratio of small farms among 

total farms decreased slowly. Farmers gave up the profession 

during the industrialization period and were absorbed by the 

industrial sector. 

Number of agricultural machinery 

Since the 1960s, water pumps, two-wheel tractors, and pesticide 

applicators have been extended to farms. Two-wheel tractors 

were 11,884 in 1970, 289,779 in 1980, and 751,236 in 

1990. In 2003, the number reached 857,809. 

Four-wheel tractors have been propagated on a large scale 

since 1980, and their number reached 2,664 in 1980, 41,203 

in 1990, and 211,576 in 2003. With the rapid propagation of 

agricultural machinery, rice cultivation operations, including 

tillage, land preparation, transplanting, and harvesting, became 

mechanized almost completely (about 98%) as of 1998. This 

was enabled by the government’s strong enthusiasm for farm 

mechanization and concentrated financial support to attain selfsufficiency 

in major crops. 

Current progress in agricultural mechanization 

In 2002, the mechanization rate of tillage and land preparation 

was 100%, rice transplanting 98%, pest control 100%, harvesting 

99%, and paddy rice drying 49% (Table 1). Except for 

drying, all operations seem to be fully mechanized. Average 

tractor power was in the 45-hp (33.8 kW) class. Major attachments 

for tractors are the Irang-Janggi (plow), rotavator, trailer, 

and others, and the width of the rotavator for the 40-hp class 

tractor is 170 cm. 

Four-row walking-type and six-row riding-type transplanters 

are the main machines extended. Three- and four-row 

head-feed paddy rice combine harvesters are now the main 

rice combine harvester. The number of paddy rice dryers increases 

as the paddy rice combine harvester is being propagated, 

but sun drying is still practiced because of good weather 

during the harvest period. Drying in the rice-processing center, 

which reached 340 units in 2003, is another choice for rice 

drying. 

Research on and development of agricultural machinery 

for paddy rice 

The following machinery has been developed: 

Partial-tillage direct rice seeder (1999) 

Attachable to a tractor, 8 rows, operation width 2 m 

(tillage 0.6 m, no-tillage 1.3 m, drain ditch 0.1 m). 

Paddy-field leveling rotavator (1999) 

For preparing paddy fields before transplanting or direct 

seeding of rice, 3.4 m in width and less than ±35 

mm in unevenness. 

Unmanned tractor (2000) 

Autonomous travel technology based on a GPS navigation 

system and remote-control technology with a 

field monitoring system for the unmanned tractor. 

Paper-mulching rice transplanter (2003) 

Developed for sustainable agriculture and no herbicide 

application. 

Partial-tillage rice transplanter (2002) 

Developed after experiments considered important 

factors in untilled paddy fields and the design parameters 

of a rotary-type machine. 

A rice wet-seeding broadcaster (2003) 

To mechanize wet seeding, a broadcaster was developed 

for wet-seeding operations. The prototype is 

composed of a metering device, a blowing part, and a 

seeding part, which can be attached to a tractor. 


Automatic hose winding device for power sprayers 

(2002) 

This study was conducted to develop the device for 

power sprayers, designed with a remote-control system. 

Criteria for securing parts for the rice transplanter and 

combine (2001) 

This study looked into the major repair parts for the 

rice transplanter and combine. It estimated the quantity 

of repair parts needed so that agricultural machinery 

agencies could secure the appropriate volume of 

parts. A computer program was developed for this 

purpose. 

A total information system for agricultural machinery 

(2001) 

A system that users could easily access through the 

Internet. 

Dry-type rice polisher (1999) 

Decreases rice turbidity by 7% versus the wet-type 

rice polisher. 

A grain circulating-type natural air in-bin dryer (1999) 

It took 10 days to dry 8 tons of paddy rice from 21.9% 

(wet basis) to 16.7% (w.b.) moisture content using 

the prototype dryer. 

 

A computer vision system to evaluate rice kernel quality 

(2001) 

The algorithm developed in this study was able to 

correctly classify rice features. For segmentation, an 

overall accuracy of 98.9% was achieved. Overall accuracy 

was 99.6% for milling rate, 99.5% for broken 

rice, 97.6% for cracked rice, and 94.7% for chalky 

rice. 

The processing system for prewashed rice (2002) 

Turbidity decreased significantly in response to increased 

amounts of water, and dropped to the target 

of 50 ppm or even lower when water increased to 

423 mL kg –1 of rice or higher. Turbidity varied from 

47.33 to 48.00 and 48.67 ppm, respectively, for water 

rates of 800, 1,000, and 1,200 kg h –1 , which meant 

successful rice processing free from prewashing. 

The far-infrared ray grain dryer (2002) 

Considering crack ratio, the inspection test of the drying 

simulation program indicated that optimum drying 

temperature should be decreased gradually 

(100→90→80→75→72 °C) for the concurrent-flow 

circulation dryer, whereas the cross-flow dryer performed 

better at 55 °C. 

Future research and development for farm mechanization 

Paddy rice is the major crop in Korea and is the staple food of 

most Korean people, who eat more than 240 g per day. Annual 

production is about 5 million metric tons from around 1 million 

hectares of paddy fields using mechanized farming. But 

Korean farmers have some problems: 

It is difficult to make profits from farming because 

farming is small-scale. 

More than 48% of Korean farmers are more than 50 

years old. 

Korean paddy rice farming is not strong enough for 

international competition because rice production 

costs and the market price are very high vis-à-vis 

imported products. 

To solve these problems, Korean rice farming needs new 

strategic studies of advanced technology development, such 

as reducing production costs through large-scale farming and 

rice production for products of high quality and excellent taste 

and appearance, including verification of food safety. 

The following need to be developed: 

Rice processing technology for keeping head rice, 

which makes an even taste, appearance, glutinousness, 

glossy surface, and color. 

A precision agricultural model for the small field plots 

in Korea, which is based on field information to decide 

the optimum amount of fertilizer and pesticide 

considering the conditions of soil and paddy rice. 

Traceability technology that offers information from 

seed to the table for human well-being (a flow route 

for farm products). 

Notes 

Authors’ address: National Institute of Agricultural Engineering, 

RDA, Suwon, Korea, e-mail: wppark@rda.go.kr. 


The status and prospects 

of rice production mechanization in China 

Li Yaoming 

The planting area of rice occupies nearly 31,000,000 hectares 

on the mainland of China, and field yields account for 39% of 

the total grain production. Rice is the largest grain crop, with 

the largest planting area, highest single yield, and most total 

production. China has 34.5% of the world’s grain yield, with 

only 21.4% of the rice planting area in the world. So China 

has played a major role in global rice production. 

The status of rice production mechanization in China 

Upon entering the 21st century, China had made great progress 

in major areas such as tillage, breeding and planting, field management, 

and harvest. By the end of 2003, China had doubled 

the area of rice under mechanization since 1995. Mechanized 

planting was 5.08% in 1995 and has increased nearly 3% per 

year since then. The mechanized area is 8 times more than that 

in 1995, whose level of rice mechanization reached 24%, with 

an increase of 21.6%. 

The status of rice production mechanization in China is 

such that the total level is relatively low, except for the level of 

tillage mechanization, which is higher. Mechanized planting 

and harvest are lower, and planting is the weakest operation 

and has the most difficulties. There are many reasons why farm 

mechanization does not combine well with agronomy. China 

has a vast territory with complex agricultural situations, and a 

large population with limited arable lands. The weak agricultural 

economy, the large labor force, and expensive farm machinery 

have all contributed to the slow adoption of rice mechanization. 

Technical analysis of rice mechanization in China 

Tillage mechanization 

For the different steps in rice production mechanization, the 

level of tillage mechanization is highest. There are many kinds 

of tillage machines: the plow, furrower, drive-disk plow, paddyfield 

harrow machine, and rotary machines that meet requirements 

for use in the main rice production areas. 

With the economic change, the most imminent technique 

is the straw-chopper. 

Rice planting mechanization 

Mechanized direct sowing of rice. China has two kinds of direct-sowing 

techniques for rice: wet direct sowing and dry direct 

sowing. The wet direct sowing is mainly aperture sowing, 

and some implements can sow the rice seed after it has been 

soaked and pregerminated. Other machines can sow sprouted 

rice seeds with a sprout of 3 mm. In contrast, the dry direct 

sowing is mainly drilling, which usually uses a wheat-drilling 

machine to directly sow seed in the paddy without irrigation, 

to a depth of 2 cm. This method is required for crop uniformity. 

China now increasingly uses direct sowing of rice. 

The rice factory-bred seedling technique. The rice factory-bred 

seedling is one means of mechanizing rice transplanting. 

According to the Chinese situation, in recent years, Chinese 

farmers have welcomed farm machinery and techniques 

such as bred seedlings, precision-sowing equipment, carpettype 

seedlings, and pan-type seedlings using a shallow-transplanting 

machine. 

Mechanized rice transplanting. There are two ways: 

transplanting and throwing seedlings. The transplanter uses 

carpet-type seedlings. Throwing seedlings uses pan-type seedlings 

with soil having less-bruising roots, more rapid greening, 

and higher yield. Moreover, the efficiency of throwing 

seedlings is higher and costs less, so Chinese farmers welcome 

it. 

Transplanters mainly adopt the crank-swing-type capture-seedling 

mechanism and entire supporting slide plates with 

a single wheel. The row distance is generally 30 cm, with manipulation 

by three operators. The thrown seedling is broadcast 

or falls into a row. The broadcasting machine throws seedlings 

using a centrifugal disc mechanism. 

Harvest technology. In the past few years, mechanized 

rice harvesting in China has developed. Implements include 

the general-purpose wheel-type combine harvester for rice and 

wheat, as well as a cutting machine and binder. Recently, some 

new implements have been developed, such as the rubber-track 

crawler, full-fed and head-fed combine harvesters, and stripper 

combine harvester suitable for harvest operations, thus 

basically meeting needs in various areas. 

Dry farming technology. China has been slow in developing 

machines suitable for dry cultivation. Even today, the 

level of mechanization is low. Rice mechanization has now 

become more important, so suitable machines for dry farming 

technology have become important. 

Prospects 

There is still much room for the development of new tillage 

technology, and also perfecting existing technology for sowing 

and applying fertilizer. This would add much value. Implements 

for returning straw to the field are urgently needed but 

are still in the development stage. Laser cultivation implements 

would be useful. 

Under controllable environmental conditions, factorybred 

seedlings are being provided to mechanized, large-scale, 

centralized commercial farms according to a standardized flow- 


chart. The goal of this technology is to upgrade planting techniques. 

Transplanting, throwing seedlings, and direct sowing in 

rice mechanization are now a mature technology for rice plant 

establishment. The technique chosen will depend on local conditions. 

The development of rice direct-sowing technology is 

now increasing. To improve the production rate of the labor 

force, save costs, and promote overall mechanization, we suggest 

that direct sowing is an ideal planting technique. For all 

of China, mechanized transplanting still plays a major role, 

thus developing into a combination of transplanting and manure 

application. Furthermore, technology is maturing with the 

further development of shallow transplanting. Swaying seedlings 

and sowing seedlings and precision-throwing-seedling 

machines are likely to obtain a good market share. 

Rice harvest mechanization has a good future. It will be 

a qualitative leap forward after overcoming problems of field 

transportation and grain cleaning in double cropping in southern 

China. A new-generation stripper combine harvester will 

have improved reliability, high efficiency, and less power consumption. 

Importance is increasingly given to using machinery for 

dry farming after solving problems of lower performance, meeting 

the special needs of rice, and enlarging the production scale. 

China will emphasize developing rice transplanting and machinery 

for dry farming, especially for the hybrid rice highspeed 

transplanter, bred-seedling technology and equipment, 

and hybrid rice precision seeding and small-quantity directsowing 

machines. The assembly, integration, and complete 

setup are speeded up, thus improving overall rice mechanization 

in various areas, and further strengthening international 

exchange and cooperation. This will promote technical renovation 

and lift the level of mechanized rice production in China. 

Bibliography 

Chen Zhi. 2003. Unique approach to develop the farm machinery 

industry by innovation of science and technology. Trans. 

CSAM 34(3):131-134. 

Editorial Committee of China Agricultural Machinery Yearbook. 

2000. China agricultural machinery yearbook 2000. Beijing 

(China): Editorial Committee of China Agricultural Machinery 

Yearbook. p 206-214. 

Liu Min. 2002. The development tactics and countermeasures of 

agricultural mechanization in the 21st century in China. Chinese 

Agric. Mechan. (6):3-5. 

Song Jiannong, Zhuang Naisheng, Wang Lichen. 2000. The development 

tendency of Chinese rice planting mechanization in 

the 21st century. J. China Agric. Univ. 5(2):30-33. 

Notes 

Author’s address: Jiangsu University, China, e-mail: 

ymli@ujs.edu.cn. 

The PhilRice-JICA rotary rice reaper: redesigning 

a technology for Filipino farmers and manufacturers 

Eulito Bautista, Manuel Jose Regalado, Arnold Juliano, Shuji Ishihara, Hiroyuki Monobe, Joel Ramos, and Leo Molinawe 

Harvesting and related operations are among the most laborintensive 

operations in rice production in Asia. In the Philippines, 

harvesting and threshing consume roughly 60% of the 

total labor devoted to rice (Takahashi 1994). Harvesting alone 

costs around 6–8% of total produce, whereas losses (including 

gathering) can reach as high as 5% of total yield (Andales 1998). 

Back-breaking manual reaping alone requires 10–16 persondays 

ha –l (Juarez et al 1988). 

Attempts have been made to modernize rice harvesting 

in Asia, the most successful of which was reported by Ezaki 

(1970) on walking-type reapers in Japan. Later, modified versions 

of this design were locally produced in China, Thailand, 

India, and Pakistan. This design made use of a cutterbar assembly, 

composed of reciprocating triangular blades placed 

above a stationary ledger plate. This cutterbar mechanism reciprocates 

at low speed, requires low power, and is highly suitable 

for windrowing operations since the plants have practically 

no horizontal movement after cutting prior to conveying 

to one side of the machine. 

In 1980, the Los Baños-based International Rice Research 

Institute (IRRI) developed with Chinese engineers a 

local version of this type of reaper for release to Asian countries. 

The IRRI reaper was a simplified version of the Japanese 

and Chinese designs and was released to small manufacturers 

in the Philippines and a few other countries (IRRI 1987). 

This design, however, was not as successful as the Japanese 

model released shortly afterward in the Philippines because of 

its durability problem as a result of poor blade tolerance 

(Stickney et al 1986), less convenient operation, and high walking 

speed. The imported model is more durable, lighter, and 

more convenient to use, with a slower forward speed and reverse 

gear that allows easy headland turning. 

Except for its high price, this imported reaper was found 

to be suitable to the variable field conditions in the Philippines. 

Eventually, it slowly replaced the IRRI models in most 

areas. No cheaper version was attempted locally to compete 

with this imported design because of a lack of alternative components 

such as light diecast components and a small trans- 


mission with forward and reverse options. Thus, we aimed to 

develop a cheaper design suitable to the field conditions in the 

Philippines and that could be locally manufactured with less 

precision. 

Development of a reaping mechanism 

Focus was directed to replacing the reciprocating cutterbar 

assembly with a rotary cutting system borrowed from grass 

cutters (Takahashi 1993). Such a rotary cutter would require 

fewer blades and less manufacturing tolerance. To minimize 

development work, some mechanisms of the IRRI reaper were 

adapted on the new design, such as the header-tiller-engine 

arrangement, windrowing mechanism, crop guides, and 

starwheels. For a simpler, low-cost, and easy-to-maintain power 

transmission, an open-carriage tiller unit was adopted. 

Laboratory studies 

A series of laboratory trials on potential rotary cutters combined 

with a starwheel and vertical flatbelt conveyor were made 

to determine the functionality and optimum operating parameters 

(blade number and tip speed, forward speed) with grain 

shattering, quality of cut, and power requirement as critical 

parameters. 

From these trials, the number of blades per disc was set 

at three, blade tip speed was set at 23–30 m s –1 , and forward 

speed could range from 2.8 to 3.3 km h –1 . 

Prototyping and field testing 

A series of prototypes was prepared to improve technical functionality 

and field performance. The basic design has three 

rotary discs with serrated blades, driven by one hexagonal belt 

through a bevel gear transmission at the drive shaft, a single 

flatbelt with sheetmetal lugs for conveying and windrowing, a 

plastic starwheel above each disc to hold the plants during 

cutting, and three inclined header plates in front for gathering 

prior to cutting (Fig. 1A). Power transmission for the carriage 

unit used an open-case chain-sprocket transmission mechanism 

connected by belt and pulley to the engine. A 3.7-kW gasoline 

engine was placed behind the tiller similar to previous designs 

for balance and operator access. 

Initially, we encountered problems of straw scattering 

and clogging, failure to attain orderly windrows, and belt slippage 

that limited power to the cutting and conveying components 

(Togasi 1994, Maeoka 1997). However, by properly 

positioning the starwheel relative to the cutting device and 

vertical conveyor, a prototype that successfully windrowed cut 

plants was finalized. This prototype, with an 80-cm cutting 

width, harvests at 0.11–0.16 ha h –1 , with grain losses of 1– 

1.5% and 70–90° windrow angles from machine direction. 

Commercialization 

Three commercial models were developed for the Philippine 

market by three cooperating manufacturers (Fig. 1B). Two of 

the designs were similar to our final prototype except for the 

Table 1. Comparison of the rotary reaper and the imported reaper 

models. 

Parameter Rotary reaper Imported reaper 

Overall length (cm) 244 239 

Overall width (cm) 138 147 

Overall height (cm) 118 90 

Weight (kg) 158–185 116 

Engine (kW) 4.5 1.7 

Cutting width (m) 1.1–1.3 1.2 

Traveling speed (km h –1 ) 2.6–3.2 2.6 

Cutting device 

38 cm dia., rotary 120 cm, reciprocating 

Minimum cutting height (cm) 8.0 10.0 

Field capacity (ha h –1 ) 0.22–0.34 0.29 

Shattering losses (%) 0.1–1.3 about 1–2 

cutting widths (1.1 and 1.3 m), while a third model followed 

the conveying system of the Japanese model employing chains 

for crop conveyance and windrowing. 

The commercial models were improvements of the final 

prototype. The hexagonal belt drive for the cutters was replaced 

by two sets of chains for durability. A reverse system employing 

belts and pulleys was also incorporated for easier headland 

turning. A bigger engine (4.5 kW) was installed to allow 

extra power for difficult soil conditions. The steel cagewheels 

replaceable by rubber tires have been changed into a combined 

tire-cagewheel system for operation in any type of soil 

surface. 

Field tests indicate that the commercial reapers have performances 

similar to those of the imported model (Table 1). 

However, in terms of harvesting ability in extreme field conditions 

such as tall crops, weedy fields, and soft soil conditions, 

the rotary reaper has better ability. In addition, repair and maintenance 

at village shops are easier and cheaper. 

Farmers’ feedback 

Central Luzon farmers found the new rotary reaper suitable to 

their conditions but also expected it to work in both transplanted 

and direct-seeded fields in any season. Because rice 

plots are generally small, maneuverability and transfer across 

levees were important and reverse motion was necessary. Adjustment 

for cutting height was also suggested in the commercial 

models since some farmers cut lower while others prefer 

higher cutting height for thresher efficiency. 

In addition, the new design was expected to work in 

weedy fields, with short, low-density crops, and beside levees 

without manual cutting of plant rows. Operators also suggested 

adopting a higher forward speed to increase field capacity, 

which is important for custom service. 

Economics 

Economic analysis indicated that using the new reaper design 

could incur a harvesting cost of US$22 ha –l versus $25 ha –l for 

the imported model. This is mainly due to the lower investment 

cost of the rotary reaper ($1,360 compared with $2,700 

per unit of the imported design). The rotary reaper has to work 


2,200 

1,200 

Gasoline 

engine 

Converter lug 

Cutter blade 

disc 

806 

Starwheel 

Cage wheel 

360 

Fig. 1. The PhilRice-JICA rotary reaper prototype (A) and commercial model (B). (A) Schematic view of 

the prototype, (B) a commercial model during field operation. 


for 33 ha y –l while the imported unit has to complete 58 ha 

y –l at a benefit-cost ratio of 1.0. 

In a separate study, Yasunobo (2001) noted that the rotary 

reaper should be economical to own when harvesting at 

least 4.13 ha. In selected areas, farmers have greater incentives 

to own such a reaper because of the income-generating 

potential for rice harvesting (from custom fees and extra opportunities 

for thresher-owners). 

References 

Andales SC. 1998. Report on postharvest loss assessment. Muñoz, 

Nueva Ecija (Philippines): Bureau of Postharvest Research 

and Extension. 

Ezaki H. 1970. Binders and combines. Tokyo: Agricultural Books 

Co. (In Japanese.) 

IRRI (International Rice Research Institute). 1987. IRRI annual report. 

Los Baños (Philippines): IRRI. p 512-513. 

Juarez F, Te A, Duff B, Crissman L, Stickney RE, Manaligod HT, 

Salazar GC, Fernandez CP. 1988. The development and impact 

of mechanical reapers in the Philippines. Agric. Econ. 

Paper No. 88-23. Los Baños (Philippines): International Rice 

Research Institute. 

Maeoka K. 1997. Report on the development of the Maligaya rice 

reaper. Report to JICA Technical Cooperation Project. Rice 

Engineering and Mechanization Division, PhilRice, Muñoz, 

Nueva Ecija. 

Stickney RE, Salazar GC, Manaligod HT, Juarez FS, Duff B, Bockhop 

CW, Fernandez CPo 1986. CAAMS-IRRI mechanical reaper: 

experiences in the Philippines. In: Small farm equipment for 

developing countries. Proceedings of the International Conference 

on Small Farm Equipment for Developing Countries: 

Past Experiences and Future Priorities. Los Baños (Philippines): 


Takahashi H. 1993. Report on the development of the Maligaya 



Nueva Ecija, Philippines. 

Takahashi H. 1994. Present situation and problems of rice production 

in the Philippines. Paper presented at the Workshop on 

Farming Management of Rice-Based Lowland Farms, 

PhilRice, Muñoz, Nueva Ecija, 5 December 1994. 

Togasi T. 1994. Report on the test and evaluation of the Maligaya 



Nueva Ecija. 

Yasunobo K, Casiwan C, Manalili R, Francisco S. 2000. Factors 

influencing farm mechanization in the Philippines: socioeconomic 

context. Report to JICA Technical Cooperation Project, 

PhilRice, Maligaya, Muñoz, Nueva Ecija. 

Notes 

Authors’ address: Philippine Rice Research Institute, e-mail: 

eubautista@philrice.gov.ph. 

Acknowledgments: We wish to acknowledge the support and suggestions 

of Dr. Santiago R. Obien and Dr. Hitoshi Takahashi 

in this PhilRice-JICA collaborative project. The assistance of 

JICA short-term experts Hiroyuki Takahashi, Tatsusi Togashi, 

Kunihiko Maeoka, and Koji Inooku was crucial during 

prototyping and testing. 

Long-mat seedling culture and the transplanting system: 

an innovative one-person operational technology 

for mechanical rice transplanting 

Hisashi Kitagawa, Akio Ogura, Kouhei Tasaka, Hiroyuki Shiratsuchi, and Mikio Yashiro 

Rice transplanting was mechanized in the 1970s in Japan, and 

now more than 99% of paddy fields are cultivated by mechanized 

transplanting. Even large-scale transplanters with higher 

efficiency and precision are becoming common, and mechanical 

transplanting is assumed to have reached its technical perfection. 

But a problem remains. 

Usually, a plastic box (60 × 28 × 3 cm), called a “nursery 

box,” is used for raising rice seedlings. Soil is packed into 

it, and seeds are sown. Then, nursery boxes are arranged in 

vinyl houses and the seedlings are raised. When the seedlings 

grow up, the nursery boxes are put on a truck and taken to the 

paddy fields. The seedlings are then transplanted by a transplanter. 

This conventional soil-seedbed system (CSS) has major 

problems for the following reasons: (1) a nursery box filled 

with soil weighs about 6 kg, (2) 2,000 nursery boxes are necessary 

for paddy fields of 10 ha, and (3) it is necessary to carry 

nursery boxes many times from sowing to transplanting. The 

larger the farmer is, the harder it is to carry the nursery boxes. 

Our center developed the long-mat seedling culture system 

(long-mat system: LMS) to save labor and facilitate work in 

comparison with the CSS (Tasaka et al 1996, 1997), and has 

improved it further. 

In this paper, we introduce the general characteristics of 

this long-mat system and discuss the recent development of 

rolling-up equipment that enabled a one-person operation, and 

the improvement of the higher-precision transplanter for longmat 

seedlings. 


Characteristics of the long-mat system 

Raising of seedlings by hydroponics 

The nursery device is composed of a nursery tray (60 × 28 × 5 

cm), a tank for liquid fertilizer, and a water-jet pump. A nonwoven 

cloth is put on the nursery tray, and germinated seeds 

are sown on it. The water pump circulates culture solution, 

and the seedlings are ready for transplanting in 2 weeks. The 

mat-shape seedlings about 6 m long grown here are called 

“long-mat seedlings (LS).” 

The length of the conventional nursery box in the CSS is 

60 cm. Therefore, the length of one LS is equivalent to that of 

ten nursery boxes in the CSS. Therefore, the number of LS is 

only one-tenth of that in the CSS. And, for the same amount of 

seedlings, the weight of LS is about one-fifth of that of the 

CSS, and therefore it is easier to carry LS to the field. 

Rolling up seedlings 

Long-mat seedlings are too long to carry. So, the technology 

to roll up LS in a compact form was developed, which enabled 

farmers to set seedlings on the transplanter in a normal position 

(Fig. 1). The most marked aspect of this technology is the 

method of rolling up LS, which is called the “reverse roll 

method.” The details of this technology will be explained below. 

The normal roll method is not as good as the reverse roll 

method. 

Manual rolling up by two people using iron plates. Three 

steps are followed: 

1. By rolling a heavy iron roller on a long-mat seedling 

in one direction, the leaves of the seedlings are bent 

down. 

2. Ten sheets of iron plates (60 cm long) are placed on 

the seedlings, and pressed down. 

3. Two people are needed for rolling up. While one 

slowly shifts an iron plate, the other rolls up the LS 

using PVC pipe as a core. 

In this method, the cost is lower, but the labor requirement 

is higher. Therefore, this technology is suitable for smallscale 

farmers. 

Mechanical one-person rolling up. A machine to support 

the rolling up of raised long-mat seedlings was developed 

for use by large-scale farmers. Instead of pressing down seedlings 

with iron plates, this machine has a structure to draw 

seedlings into the front with a rotating rubber belt and to press 

the seedlings down. This machine is self-propelled with a motor 

driven by a battery. One person follows the machine, and rolls 

up the seedlings. Using this machine, LS can be rolled up easily 

with only one operator. 

The time required for rolling up was reduced to less than 

half of that using iron plates. This machine will soon come 

onto the market. 

Loading of the rolled seedlings. The volume of the rolled 

seedlings is about one-fifth that of the equivalent amount of 

seedlings in the CSS. The seedlings can be transported by a 

pickup truck, which many Japanese farmers have. A pickup 

truck can carry seedlings for a rice field of 1.5 ha at one time. 

Transplanting long-mat seedlings. For the long-mat 

transplanter, the folders for rolled seedlings and seedling press 

controls are attached to the normal transplanter on the market. 

However, because the root-mat of LS is softer and more flexible 

than that of the CSS, LS can be injured at the time of 

machine transplanting, and some seedlings die before rooting. 

The transplanter was improved to compress the seedling mat 

by force and to increase the density of the root-mat (Fig. 2). 

For precision of transplanting, there is no inferiority in the 

transplanting of LS compared with that of the CSS with this 

easy adjustment. 

Labor savings and yield 

At the time of transplanting in the CSS, each nursery box should 

be handed over to the operator in the field. Therefore, both an 

operator and a helper are required. Mainly spouses and elderly 

people help with the loading of the nursery boxes, which 

is heavy labor. 

But, if long-mat technology is introduced, just one operator 

can do transplanting. Moreover, if the machine that supports 

rolling up is adopted, the whole operation from sowing 

to transplanting can be done completely by one person. The 

working hours for transplanting in LMS are less than one-third 

of those in the CSS. Therefore, LMS is an excellent laborsaving 

technology. 

The labor intensity of long-mat technology is less than 

half of that in the CSS. The typical working posture in the 

CSS, bending the waist to lift heavy nursery boxes, almost 

disappeared. The posture used in the LMS is much easier. 

It is estimated that the applicability of this technology is 

high because there is no inferiority in growth and yield in comparison 

with growth and yield in the CSS. And, some largescale 

farmers in Kanto and Tohoku regions have already 

adopted this technology. 

Epilogue 

The National Agricultural Research Center (NARC) developed 

the basic technology of LMS. The basic equipment was developed 

by private companies in the project funded by the Biooriented 

Technology Research Advancement Institution 

(BRAIN). And, operational technologies such as raising and 

transplanting seedlings were developed by the agricultural 

experiment stations of Iwate, Saitama, Nagano, and Ibaraki 

prefectures in cooperation with NARC. NARC is now promoting 

the diffusion of long-mat technology to farmers in cooperation 

with the prefectural diffusion sections, private corporations, 

and the farmers who had introduced LMS. 

Long-mat technology makes the loading and transplanting 

of seedlings easy. So, it is effective not only in large fields 

on a flat plain but also in the small and scattered fields in 

middle-mountainous places. Hydroponic devices are used only 

in early spring. The technology to grow tulips and strawberry 

seedlings was developed to increase the effective use of the 


Normal roll method 

Reverse roll method 

Push down 

roller 

Core for rolling up 

Seedlings are cut 

off because leaves 

come out first 

Transplanting 

finger 

Running 

direction 

Running 

direction 

Iron weight 

plate 

Core for 

rolling up 

Seedlings fall down in this 

direction 

Running 

direction 

Motor 

Belt 

Nursery tray 

Fig. 1. The principle of rolling up (upper right), manual rolling up by two people (middle), and mechanical 

one-person rolling up (below). 

facilities, too. It is expected that the long-mat technology will 

spread to stimulate agriculture in the region. 

References 

Tasaka K, Ogura A, Karahashi M. 1996. Development of hydroponic 

raising and transplanting technology for mat type rice 

seedlings. Part 1. Raising test of seedlings. J. Jpn. Soc. Agric. 

Mach. 58:89-99. 

Tasaka K, Ogura A, Karahashi M, Niiyama H, Namoto H, Kaneko T. 

1997. Development of hydroponic raising and transplanting 

technology for mat type rice seedlings. Part 2. Development 

and field test of rice transplanters for long mat type hydroponic 

rice seedlings. J. Jpn. Soc. Agric. Mach. 59:87-98. 

Notes 

Author’s address: National Agricultural Research Center, e-mail: 

kitaga@narc.affrc.go.jp. 


Seedling press 

controls 

Folders for rolled 

seedlings 

Seedlings 

Planting finger 

Guides to gather 

seedlings 

Fig. 2. The outline of the long-mat transplanter (left) and remodeling for higher precision transplanting 

(right). 

High-precision autonomous operation using 

an unmanned rice transplanter 

Yoshisada Nagasaka, Yutaka Kanetani, Naonobu Umeda, and Takuo Kokuryu 

With farm size increasing and the number of farmers decreasing, 

more efficient agricultural practices are needed. However, 

in Japan, fields are not as yet sufficiently consolidated. They 

are frequently dispersed and it is difficult to use large machines 

in areas that would be unable to bear their weight. This efficiency 

problem would be solved, however, if one operator were 

able to control several lightweight machines simultaneously. 

The goal of our research is to develop an autonomous operating 

system for paddy fields that would permit one operator to 

control multiple machines. The objective of this study is to 

develop an autonomous navigation system for the rice transplanter. 

Several pieces of research on autonomous agricultural 

machinery have been published (Reid et al 2000). Recently, 

an autonomous tractor using an optical surveying device and a 

terrestrial magnetism sensor for plowing was developed 

(Matsuo et al 2002). These researchers obtained good results 

but found it difficult to control multiple machines because each 

vehicle required its own optical surveying device. Some researchers 

used a differential global positioning system (DGPS) 

(Cho and Lee 2002). Kalman filtering of DGPS can effectively 

correct DGPS position error (Han et al 2002), but it does not 

provide enough precision for operation in paddy fields. 

Noguchi et al (2002) developed a robot tractor and Nagasaka 

et al (2000) developed an automated rice transplanter. Both of 

them employed RTKGPS and FOG sensors that established 

precise operations. In this paper, the results of automated rice 

transplanting operation experiments are reported. We first describe 

the sensors, actuators, and controllers used for the automated 

rice transplanter and how data are processed. Then we 

describe the vehicle control methods and how the desired operating 

path is chosen. The experimental results are discussed, 

and the final section gives conclusions. 


Vehicle, sensors, actuators, and controllers 

We modified a commercial 6-row rice transplanter (PH-6, Iseki 

Co., Ehime, Japan). Figure 1 shows the modified rice transplanter 

controlled by a computer with a 486-compatible 66 

MHz central processing unit. Figure 2 is a schematic for an 

automated rice transplanting system. An RTKGPS (MS750, 

Trimble Navigation Ltd., Sunnyvale, California) with 2-cm 

precision at 10 Hz data output was used to locate the position 

of the rice transplanter. For communication between the 

RTKGPS reference station and the rover station, 5-mW output 

wireless radio modems (YRM211, Vertex Standard Co., 

Ltd., Japan) were used with baud rates set at 9,600 bps. A FOG 

sensor (JG-35FD, Japan Aviation Electronics Ind., Ltd., Tokyo) 

was used to measure the yaw angle, and an inclinationmeasuring 

apparatus (JCS-7401, Japan Aviation Electronics 

Ind., Ltd., Tokyo) composed of three FOG sensors and three 

accelerometers was used to measure the roll and pitch angles. 

The GPS position data, yaw, and roll and pitch angle 

data were transferred to a computer through an RS232C interface. 

The sampling rate was 10 Hz for the GPS, 20 Hz for the 

yaw angle, and 25 Hz for the roll and pitch angles. The main 

computer corrected the GPS position data, which were influenced 

by the vehicle inclination, and calculated the control 

parameters, then sent them to an industrial programmable logic 


Fig. 1. Unmanned rice transplanter. 

Data 

Modem 

Modem 

RTKGPS base station 

RTKGPS rover station 

RS232C 

Main computer 

RS232C 

Steering 

HST 

Clutch 

Brake (L, R) 

Engine throttle 

Attachment up-down 

Encoder 

Limit switch 

Motor 

Controller 

PLC 

R5232C 

Yaw 

Roll 

Pitch 

Fig. 2. A schematic for an automated rice transplanting system. 


controller (PLC) (KZ-350, Keyence CO., Osaka) through the 

serial port every 100 milliseconds. The PLC receives control 

parameters from the main computer through the RS232C interface 

link unit to control the actuators. The steering angle is 

sensed by means of an absolute rotary encoder and is controlled 

by a DC motor. Proximity sensors detect the clutch and 

brake positions and electrical linear cylinders control the clutch 

and brake pedals. The positions of the transplanting instrument 

control levers, engine throttle, and hydrostatic transmission 

(HST) lever were measured using absolute rotary encoders 

and operated by electrical linear cylinders. The PLC control 

loop was 2 milliseconds. 

Path planning 

Before starting operation, the computer must create a desired 

path along which the rice transplanter travels. In this study, the 

paddy field was assumed to be rectangular. The four corners 

in the field were measured beforehand. The sowing width of a 

6-row rice transplanter is 1.8 m. Since a paddy field usually 

has only one entrance, the desired path depends on the width 

of the field. 

Vehicle control 

In paddy fields, it is thought that the paddy ground is flimsier 

than upland fields. Several papers discuss steering control in 

upland fields. O’Connor et al (1996) used a steering controller 

based on a set of linear motion equations. Cho and Lee (2000) 

used a fuzzy controller for autonomous operation for an orchard 

speed sprayer. Kise et al (2002) developed an optimal 

steering controller and obtained good curved-path guidance 

results. Inoue et al (1997) developed an adaptive steering controller 

that corrected steering system delay. Most of these researchers 

analyzed vehicle dynamics and made a dynamic 

model, then decided on the control parameters. 

In this study, we used a simple proportional controller 

for steering control before evaluating the effectiveness of advanced 

controllers in loose-ground-like paddy fields. 

At the headland, the rice transplanter moves forward and 

backward during a turn so as to minimize the headland space. 

When the rice transplanter reaches the edge of the field, it first 

stops transplanting and lifts the transplanting attachment, then 

the HST lever is set to the forward position. While the rice 

transplanter is turning and the yaw angle is less than 160 degrees, 

only the yaw angle is obtained. Under these conditions, 

the steering angle is kept at 40 degrees and the brake is applied 

on one side. If the yaw angle exceeds 160 degrees, the 

rice transplanter is moved as close as possible to the next desired 

path. Then it moves backward. 


The experiment was conducted 4 days after puddling. The field 

size is a 20 m × 100 m square. When the operating program 

started, the transplanter moved into the paddy field and reached 

the starting point. Then it started operation and went forward 

and backward in the field 4 times. Finally, it planted the seedlings 

around the headland. After finishing operation, it went 

out from the field and stopped automatically. The deviation 

from the desired straight path was less than 12 cm at a traveling 

speed of 0.6 m s –1 during operation. As the turning radius 

of the vehicle at the headland was around 2 m, it was easy to 

get back as close as possible to the new desired path after turning. 

The operating efficiency was about 0.3 ha h –1 when long 

mat-type hydroponic seedlings (Tasaka 1998) were used. 

When the data communication between the RTKGPS 

base and rover station via radio link was disconnected, the 

clutch was released and operation was interrupted. Then, as 

soon as the radio link was connected, operation started again. 

Conclusions 

An automated rice transplanter was developed and autonomous 

guidance operation was conducted. The maximum deviation 

from the desired straight path is 12 cm and RMS deviation is 

5.5 cm at a 0.6 m s –1 operating speed. This is only accuracy 

enough for rice transplanting operations, and not enough for 

spraying and mechanical weeding operations after rice transplanting. 

Since it is necessary to drive an autonomous vehicle 

between the crop row to do spraying and mechanical weeding 

operations, the maximum deviation must be reduced to less 

than 5 cm. It is also necessary to evaluate the control algorithms. 

References 

Cho SI, Lee JH. 2000. Autonomous speed sprayer using differential 

global positioning system, genetic algorithm and fuzzy control. 

J. Agric. Eng. Res. 76:111-119. 

Han S, Zhang Q, Noh H. 2002. Kalman filtering of DGPS position 

for a parallel tracking application. Trans. ASAE 45(3):553- 

559. 

Inoue K, Otuka K, Sugimoto M, Murakami N. 1997. Estimation of 

place of tractor and adaptive control method of autonomous 

tractor using INS and GPS. Preprints of the International 

Workshop on Robotics and Automated Machinery for Bio- 

Productions, 27-36, EurAgEng, Gandía, Spain. 

Kise M, Noguchi N, Ishii K, Terao H. 2002. The development of the 

autonomous tractor with steering controller applied by optimal 

control. Proceedings of Automation Technology for Off- 

Road Equipment. Chicago, Ill. (USA): ASAE. p 367-373. 

Matsuo Y, Yamamoto S, Yukumoto O. 2002. Development of tilling 

robot and operation software. Proceedings of Automation 

Technology for Off-Road Equipment. Chicago, Ill. (USA): 

ASAE. p 184-189. 

Nagasaka Y, Otani R, Shigeta K, Taniwaki K. 2000. A study about 

an automated rice transplanter with GPS and FOG. ASAE 

Paper No.00-1045. Milwaukee, Wis. (USA): ASAE. 

Noguchi N, Kise M, Ishii K, Terao H. 2002. Field automation using 

robot tractor. Proceedings of Automation Technology for Off- 

Road Equipment. Chicago, Ill. (USA): ASAE. p 239-245. 

O’Connor M, Bell T, Elkaim G, Parkinson B. 1996. Automatic steering 

of farm vehicles using GPS. Paper presented at the 3rd 

International Conference on Precision Agriculture. Madison, 

Wis. (USA): ASA, CSSA, and SSSA. 


Reid JF, Zhang Q, Noguchi N, Dickson M. 2000. Agricultural automatic 

guidance research in North America, Comput. Electron. 

Agric. 25(1/2):155-167. 

Tasaka K. 1998. Outline of raising and transplanting technology for 

long mat type hydroponic rice seedlings. Agric. Eng. Paper 

98-A-061. EurAgEng, Oslo, Norway. 

Notes 

Authors’ address: National Agricultural Research Center, e-mail: 

zentei@affrc.go.jp. 

Precision-drilling methods for the direct sowing 

of rice in flooded paddy fields 

Yoh Nishimura, Kazunobu Hayashi, Takashi Goto, and Mitsuhiro Horio 

Dry seeding is the most common direct-sowing method used 

in Japanese rice paddy fields. In 1996, before this study began, 

only 7,000 ha (< 0.5%) of the rice fields in Japan relied 

on direct-sowing methods, and dry seeding accounted for approximately 

70% of the total. However, considering the climatic 

conditions in Japan and the working system available 

for rice production, we believe that wet seeding could considerably 

improve rice productivity. Because wet seeding relies 

on cultivating methods similar to rice transplanting, however, 

it is generally considered to be more expensive than dry seeding, 

in terms of both labor and machinery requirements. 

The major challenge related to implementing wet seeding 

in Japan has been crop stabilization and minimizing the 

occurrence of lodging. Most rice varieties that are currently 

grown in Japan have been selectively bred for eating quality, 

and, as a result, few varieties are specifically suited for directsowing 

methods and are also resistant to low temperatures and 

lodging. However, several studies have been carried out investigating 

direct-sowing methods using several different rice 

varieties. Japanese rice farmers have expressed a desire to produce 

high-quality rice (Koshihikari, etc.) that can attain the 

highest market prices. As a result, it is not broadcast seeding 

that places seeds on the soil surface but the sowing method 

that buries seeds in the soil absolutely that could be the best 

option. 

Traditional wet seeding in flooded paddy fields 

An improved technology, aimed at stabilizing crop growth and 

improving the efficiency of crop establishment by burying seeds 

in the soil, was developed in the 1980s. A method involving 

the coating of seeds with CaO 3 was found to improve both 

crop emergence and establishment. Furthermore, a riding-type 

drill seeder, which could be used to open shallow furrows on 

the puddled soil surface and re-cover the furrow after seeding, 

was also developed. Although these machines were available 

on the market, their widespread use was limited, with the majority 

of growers continuing to use more conventional transplanting 

methods. 

Commercial success of this machinery was limited because 

of several problems. First, the coating machine required 

a high level of skill for its operation, especially for calculating 

the appropriate doses of chemical powder and water. This process 

required constant observation to control the coating efficiency 

after adding rice seeds to the inclined-rotator drum. 

High concentrations of chemical dust in the working environment 

also caused a significant problem and a potential health 

risk. The drill seeder was found to have poor accuracy for the 

control of seeding depth affected by the hardness of the soil 

surface. Several mechanical problems also occurred related to 

the stability of the furrow-opener and the covering plate. 

Development of a precision-drilling method 

for flooded paddy fields 

Following the Uruguay Round agreement in 1994, it was decided 

that there should be a gradual reduction in the tariff placed 

on imported rice. This agreement encouraged the production 

of low-cost rice that could be produced with minimum labor. 

Research carried out in our laboratory led to the development 

of a direct-sowing machine, with financial assistance from an 

“Urgent Development of Agricultural Machinery and Its Commercial 

Enterprise” project aimed at facilitating improvements 

in agricultural machinery. This research led to the development 

of an automatic coating machine for rice paddy seeds, 

and a precision drill seeder. 

 

Automatic coating machine for paddy seed 

Design. The coating machine has several characteristics that 

lead to improved efficiency. 

The machine uses dewatered, dipped seeds to stabilize 

the moisture content. 

Although the coating machine employs the same 

working principles as the earlier developed models 

(consecutive water input and intermittent chemical 

input after seed input), the rotator drum has a much 

greater capacity. Water and chemical inputs are controlled 

automatically. 

The coating mix can be set to ratios of 1:2 and 2:2 

(seeds to chemical, dry weight). Seed input modes 

can be set to 15 kg or 10 kg for 1:2 mixing or 15 kg or 

20 kg for 1:1 mixing. 


Chemical hopper 

Water nozzle 

Agitator 

Cover 

Inclined-rotator 

drum 

Screw feeder 

Vent for products 

Water pump 

Controller 

Fig. 1. Side view of a developed automatic coating machine. 

 

 

The machine is contained within a shatterproof cover 

and was designed to be resistant to chemical dust 

particles. 

A vent on the side of the rotator drum enables the 

easy removal of the coated seeds. 

Overview of the coating machine. The coating machine, 

as shown in Figure 1, is composed of an inclined-rotator drum, 

a screw-type chemical feeder, a constant-pressure water nozzle 

with associated pump system, and a program-sequence controller. 

Combining a large-diameter agitator and a screw-type 

feeder, a stable chemical feeding performance was achieved. 

Several research experiments were carried out to test different 

input patterns of water and coating chemicals. Based on the 

results of these experiments, a prototype machine was produced. 

This prototype was equipped with an automatic sequencing 

system to control the input of water and chemicals into the 

inclined-rotator drum. Several performance tests were then 

carried out. 

Performance of the coating machine. The efficiency of 

the newly developed coating machine was compared with that 

of other more traditional models. Efficiency was evaluated 

based on the output of coated seeds, the mass of one coated 

seed, the hardness of the coating layer, and germination. The 

results of these tests showed that, in most respects, the newly 

developed coating machine performed similarly to the earlier 

models. Although differences were few between the models 

for coating time, the newly developed coating machine had 

double the capacity and was therefore much faster. Much of 

the preparation work such as dewatering, preparation of chemicals, 

drying of coated seed, and packing of the coated product 

needed to be carried out either before or after coating using 

the old model. However, using the newly developed coating 

machine, all of these tasks could be carried out simultaneously. 

The precision drill seeder in flooded paddy fields 

Design. The depth at which the seeds were sown was known 

to be affected by the hardness of the soil surface in the field. 

Research was therefore carried out to improve the seeding 

characteristics of the traditional drill seeder. 

The drill seeder mounts on a paddy vehicle had a 

working width of eight rows or more. 

A float is used to improve the accuracy of placing 

seeds at depth. A function was also provided to accurately 

control the angle of the covering plate. 

Seeding depth is adjusted depending on soil characteristics 

and the rice variety being planted. 

The machine adds fertilizer to the planted seeds. 

Overview of the seeding machine. The seeder, as shown 

in Figure 2, consists of an 8-row seeder mounted on a paddy 

vehicle. The seeder consists of a seeding part and a fertilizing 

part. This seeding machine was designed to improve seeding 

accuracy. As a high-performance fertilizing part is already in 

commercial use, commercial items were employed. The basic 

components of the furrow-opener and the covering plate are 

also commercially available, as is the traditional drill seeder. 

To overcome some of the problems associated with variations 

in soil-surface hardness, it was decided that the seeder 


Fig. 2. Developed drill seeder in a flooded paddy field. 

should be equipped with a soil-surface hardness sensor. Several 

experiments were carried out to investigate soil-surface 

hardness in relation to covering-plate angle. The results of these 

experiments showed that control of the covering-plate angle 

with a sensor could be used to overcome some of the soilcovering 

problems associated with differences in soil-surface 

hardness. The drill seeder was found to be more accurate for 

seeding depth than other traditional drill seeders. A float was 

also installed to increase the area of ground-surface contact, 

which significantly improved efficiency. The precision drill 

seeder was therefore provided with a mechanism to control 

the covering-plate angle, a sensor for soil-surface hardness, 

and a large float with a two-step adjustable furrow-opener. 

Performance of the seeding machine. Field tests using 

the newly developed seeder were conducted at 14 sites in Japan. 

The results of these tests showed that the seeder was much 

more efficient than traditional seeding machines and led to a 

33% decrease in the number of surface seedlings and a 40% 

decrease in the number of lodged and floating seedlings. In 

addition, the working capacity was improved by 30% as a result 

of the wider working width and the higher operation speed 

compared with those of other traditional seeders. 

Commercialization and diffusion 

Since the development of an improved automatic coating machine, 

its use has increased dramatically. This machine, which 

assists with the direct sowing of rice, was found to improve 

crop emergence and establishment. The development of a precision 

drill seeder also contributes to crop emergence and establishment 

by decreasing the number of surface seedlings and 

floating seedlings, and decreasing lodging. Moreover, these 

newly developed machines have working properties similar to 

those of rice transplanters and will undoubtedly lead to more 

efficient rice fields because of the application of direct-sowing 

techniques in flooded paddy fields. 

The automatic coating machine has been marketed since 

1999 by HATSUTA Industrial Co., Ltd. and Yanmar Agricultural 

Equipment Co., Ltd. A precision drill seeder for use in 

flooded paddy fields has been available and marketed since 

2000 by Iseki & Co., Ltd.; Kubota Corporation; Mitsubishi 

Agricultural Machinery Co., Ltd.; and Yanmar Agricultural 

Equipment Co., Ltd. The use of these machines has increased 

dramatically in recent years. Some 250 coating machines and 

300 drill seeders are now being used. As a result of the increased 

use of these machines, the amount of land that can be 

used for direct sowing in flooded paddy fields will likely increase 

to three times the area that was available in 1998 (3,600 

ha). 

Notes 

Authors’ address: IAM, Bio-oriented Technology Research Advancement 

Institution, e-mail: Y_nishimura@affrc.go.jp. 


The rice direct-seeding system using multiple seed pellets 

in northern Tohoku 

Hiroyuki Sekiya, Hitoshi Ogiwara, Shoichi Kimura, Ryuji Otani, Yukio Yaji, Satoshi Morita, Tatsushi Togashi, and Hiroaki Watanabe 

In Japan, direct-seeding methods of rice are now being introduced 

to lower rice production costs. However, in the northeast 

area, direct seeding did not spread easily. Under chilly 

spring weather conditions, seed establishment in direct seeding 

was unstable in wet-seeded fields. Also, growth of broadcast 

rice is difficult to control. Moreover, good-tasting varieties, 

such as Akitakomachi (Oryza sativa L.), suffer from lodging. 

Recently, to solve these problems, a precise direct-seeding 

system was developed that stabilizes seedling establishment 

and lodging resistance in wet-seeded fields. We developed 

this new rice direct-seeding system with multiple seed 

pellets as one type of hill-seeding cultivation. This paper discusses 

the granulation method of multiple seed pellets, and 

paddy field cultivation tests using multiple seed pellets in wetseeded 

rice. 

Developing the granulation method of multiple seed pellets 

Developing a granulation method 

of a clay-rod cutting system 

We developed a granulation method of a clay-rod cutting system 

as a high-precision new granulation method (Fig. 1, Togashi 

et al 2004b). The clay-rod cutting method follows these steps: 

clay rods are injected from nozzles by a compressing pump, 

rice seeds are spread over the rods, reciprocating movement is 

given to the rods and seeds, and the rods are cut between cutting 

boards after the seeds are attached to the rod surface. The 

pellets are coated in Calper Fine Granule (an oxygen-generating 

product) using a coating machine. The round-shaped pellet 

of about 10 mm in diameter contains 5–7 seeds, made by 

the clay-rod cutting method. The working efficiency of the 

granulation instrument is about 60,000 pellets h –1 (0.3 ha h –1 ). 

This work needs about 1–2 people. The seeds used for granulation 

of multiple seed pellets are pregerminated to about 30% 

in pigeon-breasted conditions. The coated multiple seed pellets 

are put in a sealed plastic bag, and have been stored for 

about 1 month in refrigeration conditions (Ogiwara et al 2003). 

Before seeding work, 48 h of warm processing (25 °C) of 

multiple seed pellets were effective for bud stabilization under 

low-air-temperature conditions. Moreover, we have made 

a small and simple experimental granulation instrument of the 

clay-rod cutting system for general marketing by joint research 

with a private corporation. This granulation instrument was 

sharply miniaturized by simplifying the clay feeding device. 

However, the clay feeding unit needs to be improved for continuous 

operation. At present, no granulation instrument for 

multiple seed pellets is on the market. 

Clay supply device 

(A) Seed adhesion process 

Seed supply device 

Seeds are adhered to the clay rod 

Sorting device 

(B) Granulation process 

Clay-rod cut device 

Ditch board conveyor 

Cloth belt for conveyance 

The clay rod is cut and rolled to the ditch 

board conveyor and seeds are rounded 

Fig. 1. The machine for manufacturing multiple seed pellets by the clay-rod cutting method (Togashi et al 2004a). 


Table 1. Yield components and lodging index for planted multiple seed pellets of cultivar Akitakomachi. a 

Seeding method Fertilizer application Yield Panicle no. Culm length Lodging index Pushing resistance 

(g N m –2 ) (g m –2 ) (m –2 ) (cm) (0–4) (g × cm) 

Seeding multiple Standard 5-3-2-2 610 434 80.3 0.83 1,635 

seed pellets 

Seeding multiple Large quantity 7-3-2-2 638* 451 84.6* 1.33 2,594 

seed pellets 

Drilling Standard 5-3-2-2 578 431 78.7 1.17 525 

Drilling Large quantity 7-3-2-2 638* 465 84.2* 2.17* 0 

Broadcast seeding Standard 5-3-2-2 548 387 80.3 2.08* 31 

Hill-seeding Standard 5-3-2-2 588 436 78.3 0.67 1,760 

Transplanting Standard 5-3-2-2 586 441 73.7* 0* 2,729 

a 

Multiple-seed-pellet seeding, hill-seeding, and transplanting were used with about 15-cm interrow spacing and 30-cm row spacing. Drilling was done with about 30-cm row 

spacing. Hill-seeding was done every five grains within 6 cm of the direction of a rider and 4-cm width. Seeding depth was about 10 mm. The multiple seed pellets of a clayrod 

cutting system were used for the test. Large-quantity fertilization added 2 g N m –2 using LP30 (coated urea-N). This was adjusted at 4 wk after seeding to seedling 

establishment at 22 shares m –2 , 5 seedlings per share, and 111 seedlings m –2 . Data are means of 3 sections. Yield is the weight of brown rice graded by grain size (>1.85 

mm in thickness) at 15% moisture. Pushing resistance was measured at 20-cm height at 18–20 d after heading. * indicates a significant difference as opposed to multipleseed-pellet 

seeding at 0.05 according to the Dunnet method (Morita et al 2004). 

Developing the granulation method 

using clay-powder 

To granulate more simply, we developed an alternative production 

method for multiple seed pellets. With this method, 

seeds and clay-powder were rounded together between a wet 

cellulose spongy sheet and a sponge-plastic mat on which hemispherical 

hollows were arranged. Seeds and clay-powder were 

made to condense and a spherical pellet was fabricated (Sekiya 

et al 2004b). The granulation instrument made as an experiment 

can be continuously granulating seed pellets by moving 

conveyor belts with a spongy sheet and a sponge plastic mat. 

Clay-powder for ceramic art is mixed with an oxygen-generating 

product (Calper) used for a binder by a weight ratio of 1:1. 

This granulation method is as efficient as the clay-rod cutting 

method. The pellet was about 9 mm in diameter and contained 

4 seeds (CV 49%) on average. This method is expected to be 

less costly than the clay-rod cutting method. 

Developing the seeding method of multiple seed pellets 

We developed a precision hill-seeding cultivation method using 

multiple seed pellets. Multiple seed pellets are sown by a 

planter, which is an improved commercial inclined-belt-type 

planter attached to a 4WD tractor or a transplanter (Togashi et 

al 2002). The inclined-belt-type planter drops multiple seed 

pellets on puddled soil with about 15-cm interrow spacing and 

30-cm row spacing. The multiple seed pellets were precisely 

hill-seeded at about 5-mm depth. Planting efficiency of an eightrow 

planter is about 0.35 ha h –1 . After seeding, water is drained 

for 1 to 2 wk depending on soil conditions. Then, water is 

refilled and the usual cultural management is practiced. This 

planter can be used for soybean and wheat seeding, and seeding 

accuracy is good. 

Multiple seed pellets could be broadcast by a duster or a 

radio-controlled helicopter (Togashi et al 2004a). However, 

the broadcasting method sometimes made a block of seeds and 

cracked pellets. In direct seeding by a duster, the unevenness 

of multiple seed pellets was also large compared with singlecoated 

seeds. The granulation method needs to be improved 

for the broadcasting method. 

Field experiments 

In precision cultivation tests in which the number of seedlings 

to be established was arranged, the resistance to lodging of 

multiple seed pellets was improved compared with broadcast 

seeding or drilling (Table 1, Morita et al 2004). In a large number 

of fertilization conditions, the difference between multipleseed-pellet 

seedling cultivation and drilling cultivation was 

remarkable. For labor-saving fertilizing work, we experimentally 

produced a sidedressing fertilizer for a planter and carried 

out cultivation tests by a single application of the fertilizer 

(Sekiya et al 2004a). A sidedressing fertilizer application 

showed that sufficient yield was obtained with a single fertilizer 

application by this machine. 

In 6 years of field cultivating, results of multiple-seedpellet 

seeding in Ohta test fields in Akita Prefecture indicated 

a stable brown rice harvest of 4.58–6.84 t ha –1 and less lodging 

using Akitakomachi (Sekiya et al 2004b). Since the amount 

of fertilization is decreasing now, there were small lodging 

resistance differences in seeding multiple seed pellets and drilling, 

and there were also small yield differences. When culm 

length was comparable in multiple-seed-pellet seeding, lodging 

was mitigated compared with drilling. 

As the farmers introduced the multiple-seed-pellet direct-seeding 

system, labor expenses, seedling materials, and 

fertilizer expenses could be reduced. However, the entire materials 

expense had only a small difference in multiple-seedpellet 

seeding and transplanting with the increase in the expense 

of Calper or additional herbicide. The merit of the multiple-seed-pellet 

seeding system for farmers was that it was as 

easy to manage as transplanting. On the other hand, farmers 

seldom felt a difference in resistance to lodging between the 

multiple-seed-pellet seeding system and the drilling system, 


according to the improvement in water and fertilizer management 

techniques, and the direct-seeding managerial technique. 

Conclusions 

We established the direct-seeding system, which employed the 

characteristics of multiple seed pellets efficiently. Farmers who 

use a type of direct seeding that is weak in lodging resistance 

can be provided with stable direct-seeding technology by introducing 

the precision hill-seeding cultivation method using 

multiple seed pellets (Yaji et al 2001). However, this cultivation 

method has not spread up to now. The reasons are considered 

to be that there is still no commercial granulation instrument, 

there is little cost reduction, the seeding efficiency of an 

inclined-belt-type planter is somewhat low, and there are few 

differences in resistance to lodging with a drilling method for 

the present low yield level. Farmers who cooperated in cultivation 

tests mentioned that direct planting with very high seeding 

efficiency and high resistance to lodging was done by introducing 

multiple seed pellets into the broadcast-seeding system. 

Now, we are continuing to develop a simple and inexpensive 

granulation instrument suitable for the broadcast-seeding 

system and broadcast-seeding cultivation system tests using 

multiple seed pellets. 

References 

Morita S, Sekiya H, Yaji Y, Ogiwara H. 2004. Evaluation of lodging 

resistance by the chain method and improvement of the cultivation 

method. Report of integrated research for agriculture 

(NARCT). 18:32-33. 

Ogiwara H, Otani R, Yaji Y, Sekiya H. 2003. Emergence and seedling 

establishment of rice multiple seed pellet. Jpn. J. Tohoku 

Crop Sci. 46:7-8. 

Sekiya H, Kimura S, Ogiwara H, Otani R, Nisiwaki K, Watanabe H. 

2004b. Development of granulation method of the multiple 

seed pellet using clay-powder. Jpn. J. Farm Work Research 

39 (extra issue) 1:47-48. 

Sekiya H, Ogiwara H, Morita S, Suzuki Y, Yaji Y, Watanabe H. 2004a. 

Result of a direct-seeding cultivation examination at Ohta test 

field in Akita prefecture. Report of integrated research for 

agriculture (NARCT). 18:38-44. 

Togashi T, Amaha K, Nishiwaki K, Yaji Y, Kimura S, Kikuchi Y. 

2002. Development of the granulation and seeding technology 

of rice multiple seed pellet. Part 5. Jpn. J. Tohoku Agric. 

Implem. Mach. 49:15-18. 

Togashi T, Ito N, Nishiwaki K, Kimura S, Amaha K, Matuo K, Otani 

R, 2004a. Development of seeding method of multiple seed 

pellet. Report of integrated research for agriculture (NARCT). 

18:50-53. 

Togashi T, Ito N, Nishiwaki K, Kimura S, Amaha K, Otani R. 2004b. 

Development of simple granulation method of multiple seed 

pellet at low cost. Report of integrated research for agriculture 

(NARCT). 18:45-49. 

Yaji Y, Togashi T, Morita S, Sekiya H. 2001. Rice direct-seeding 

system using multiple seed pellet in a cold district. Jpn. J. 

Agric. Hort. 76:46-54. 

Notes 

Authors’ addresses: Hiroyuki Sekiya, Shoichi Kimura, and Ryuji 

Otani, National Agricultural Research Center for Tohoku Region 

(NARCT), NARO, Japan, e-mail: sekiya@affrc.go.jp; 

Hitoshi Ogiwara, Agriculture, Forestry, and Fisheries Research 

Council; Yukio Yaji, headquarters of NARO; Satoshi Morita, 

National Agricultural Research Center for Kyushu Okinawa 

Region (KONARC); Tatsushi Togashi and Hiroaki Watanabe, 

National Agricultural Research Center (NARC). 


This session covered specific concerns related to the improvement 

of efficiency through innovations in mechanization for boosting 

rice production. Some 90% of the total rice production is 

produced from Asian countries. Invited speakers came from Korea, 

China, and the Philippines. 

Professor N. Ito discussed the important role of rice as a 

key resource to save our planet. He said that the world faces 

various global issues, or tetralemma. The rice world requires a 

sufficient food supply, particularly including rice, because the world 

population is increasing markedly. On the other hand, globalscale 

dependency on fossil fuel must be minimized to prevent 

global warming. These problems can be solved only by improving 

rice production; hence, mechanization plays a vital role in this 

endeavor as a force multiplier compensating for the labor shortage. 

In addition, N. Ito introduced new, unique technologies: a 

combine husker, combine harvester equipped with a pivot turn 

mechanism, brown rice dryer, and others. Finally, he stressed 

that rice has good potential to reduce tetralemma as an environmentally 

friendly fuel. 

W.P. Park reviewed the current status of agricultural mechanization, 

which progressed rapidly in paddy rice production. As a 

result, Korean agricultural policy has contributed to self-sufficiency 

in food and improvement of welfare in rural society. But, Korean 

farmers have some problems to be solved because of the small 

farming scale and aging society. To solve these problems, Korean 

rice farming needs new strategic studies of advanced technology 

development to lower production costs, enhance the quality of 

products, and use precision agriculture and farm robots. 

Professor Y. Li analyzed the development of rice mechanization, 

and noted that mechanization of planting and harvesting 

is relatively less since farm mechanization does not go well with 

agronomy because of vast territory with complex agricultural situations. 

China will concentrate on developing rice planting and 

drying mechanization techniques, especially for new equipment, 


including the high-speed transplanter and small-quantity directsowing 

machine for hybrid rice. 

E. Bautista explained the rotary rice reaper to be developed 

in the PhilRice-JICA collaborative project. Rice harvesting 

using the traditional sickle remains a laborious activity. Owing to 

its simple design and ease of manufacture and maintenance, 

the new reaper design has high potential in the Philippines and 

in other Asian countries. Though the engine output is higher than 

that of imported units, it is the most realistic selection when 

considering repairs. It was compatible with other technologies 

that have been proven effective for farmers and friendly to the 

rice communities in the village. He also mentioned the challenge 

to develop a tractor-mounted type. 

Following the invited speakers, four Japanese researchers 

introduced newly developed technologies. 

H. Kitagawa explained the long-mat seedling culture system 

for labor-saving and comfortable work. The weight of the 

long-mat is around one-fifth that of a conventional soil seedbed. 

Accordingly, the long-mat system eliminates the need to carry 

nursery boxes many times from sowing to transplanting. Longmat 

technology makes loading and transplanting processes for 

seedlings easy. The required working hours for transplanting decreased 

to one-half, and this enabled a one-person operation. 

Y. Nagasaka discussed an automated rice transplanter and 

autonomous guidance system equipped with a GPS and fiberoptic 

gyro-sensor to achieve more efficient unmanned operation. 

Accuracy was enough for rice transplanting operations, but was 

not always enough for doing other work. A newly designed transplanter 

could be applicable, but, for seedling recovery and traveling 

on the road, and especially production costs, the adoption of 

a commercialized transplanter is recommendable. 

Y. Nishimura explained a new rice seed-coating machine 

and high-precision drilling machine. As for the direct-sowing machine, 

the precision drill seeder equipped with a soil-surface hardness 

sensor was able to stabilize seeding depth. These two machines 

have already been commercialized and they helped to 

extend the direct-sowing area. 

Finally, a general discussion covered direct sowing of rice, 

precision farming, autonomous technologies, and energy use of 

rice. The following points were highlighted: 

1. The proportion of direct-sowing area is about 6–7% in 

Korea, and that of Japan is less than 1%. China will 

tend to increase this area. However, the development 

of a new precision drilling machine or demand for cost 

reduction will probably promote a further increase in 

direct-sowing area. 

2. In Korea, precision farming (PF) research is conducted 

in collaboration with a U.S. university as well as counterparts 

in Japan. The effectiveness of PF had already 

been demonstrated on large-scale farms like those in 

the U.S., whereas the compatibility of environmental 

conservation and profitability for farmers was not entirely 

clarified in small-scale farming, including that of 

Japan. It is better to cooperate among Asian countries 

to create an Asian model for enhancing value-added 

productivity. 

3. For robot research, most participants, except Korea and 

Japan, have less concern about autonomous technologies. 

But, these technologies are expected as measures 

to cope with the labor shortage, or to eliminate laborious 

work in agriculture. Research on and development 

of this type of agricultural machinery will be accelerated 

in Korea and Japan. 

4. In Japan, rice policy reform is being steadily promoted 

since rice consumption per capita is decreasing. Although 

the use of rice as an energy crop was not discussed 

in the session, it seems to be one alternative 

besides an adjustment in rice production. 

5. From a global viewpoint, the amount of rice consumption 

exceeds that of production, which causes an imbalance 

between supply and demand. In addition, because 

of meteorological disasters such as droughts and 

floods, we need to realize the importance of stabilizing 

rice supply and enhancing productivity. Mechanization 

will play a major role in this in the future as well. 

As a whole, this session examined the wide diversity in 

Asian rice farming. This discussion proved to be a good opportunity 

to exchange information for international collaborative research 

activities. We appreciate the cooperation of Professor 

Koike, the chairperson from the Japanese Society of Agricultural 

Machinery, and all the participants. 


SESSION 8 

Improving rice quality 

CONVENER: U. Matsukura (NARO

The chemical basis of rice end-use quality 

Kshirod R. Bhattacharya 

Rice is among the top food grains in the world and is the most 

important one as human food. It is grown widely on all continents 

and under all agroclimatic conditions. This wide adaptation 

has led to the evolution of thousands of varieties of rice 

having diverse cooking, eating, and product-making characteristics. 

Rice palatable to one ethnic group or suitable for one 

product may not be so for others. One striking example of this 

diversity is the unique geographical shift in the quality of rice 

in Asia, the home of the world’s rice. Rice grown in northern 

and eastern Asia cooks soft and sticky, while that in southern 

Asia is completely different, for it cooks firm and fluffy, with 

that in Southeast Asia coming in between. 

Research on divergence in rice quality 

This wide varietal divergence in rice quality has attracted the 

attention of researchers for close to a century. In fact, this subject 

has been one of the most intensively investigated topics in 

cereal chemistry. The work can be divided into three phases. 

The first phase, up to the end of the 1940s, was merely a period 

of initiation. The work, carried out mostly in India, was 

sporadic and confined largely to studying the absorption of 

water by rice during cooking. 

The second phase, the period of data accumulation, 

started soon after World War II, when there was a sudden spurt 

of research activity in the subject in the United States, Japan, 

Germany, Italy, Spain, and France. This was followed shortly 

afterward by intensive and sustained research at the newly set 

up International Rice Research Institute (IRRI) in the Philippines 

as well as at the Central Food Technological Research 

Institute (CFTRI) in Mysore, India. 

This work led to a mass of data. The main conclusion 

was that the amylose starch content of rice was the single largest 

determinant of its end-use quality. The greater the amylose, 

the firmer was the rice texture when cooked and viceversa. 

However, it was also realized that amylose, though necessary, 

was not sufficient; other secondary factors were involved. 

In fact, pinpointing these secondary factors was the 

main brunt of research during this phase. Many factors were 

suggested: hydration power, starch-iodine blue value, gelatinization 

temperature (GT), alkali digestion score, Brabender 

viscograph pattern, mobility of an alkaline rice-flour gel (gel 

consistency), as well as hot-water-insoluble amylose. This last 

item has an interesting history. While investigating the starchiodine 

blue value of rice varieties, my school at CFTRI realized 

that it was nothing but the water-soluble amylose of rice. 

Researchers then observed that the solubility of amylose of 

rice varieties fell into three distinct groups. The Taiwanese 

high-amylose, low-GT, dwarfing-gene donors and many of their 

progenies had an amylose solubility of only about 35%; many 

high-amylose, intermediate-GT varieties of South Asia had 

about 45%; all other varieties showed about 55% solubility. 

The calculated insoluble amylose content of rice on this basis 

(total minus soluble amylose) showed a striking correlation 

with rice texture, particularly in the high-amylose group, which 

was thereby divided into three distinct subgroups showing distinct 

texture after cooking (Bhattacharya et al 1978). 

Taking the above secondary factors into consideration, 

it appeared by the mid-1980s that the Brabender paste breakdown, 

gel consistency, and insoluble amylose, along with amylose 

content of rice, could characterize its end-use quality very 

well (Sowbhagya et al 1987). But the chemical basis was unclear. 

Evidently, these approaches had reached the limit of their 

potential and a new approach was necessary. 

A new approach 

The third phase started at this juncture in the mid-1980s with a 

complete shift in the prevailing paradigm. Chemists’ understanding 

of the structure of starch, especially amylopectin, has 

been all along going through a metamorphosis. The chain structure 

of amylopectin came under serious reexamination from 

the end of the 1970s with the use of various starch-hydrolyzing 

enzymes and of gel permeation chromatography (GPC). 

Biliaderis et al (1981), for example, showed that the chain profile 

of legume starches varied significantly from one species 

to another. Again, Professors Fuwa and Hizukuri in Japan 

showed significant variation in amylopectin chain structure 

among various starches. Chinnaswamy in my laboratory at 

CFTRI at this time observed with surprise that the GPC-separated 

amylopectins of rice starch gave a varying shade of blue 

color with iodine. What is more, the absorption maxima (λmax) 

of these iodine complexes showed excellent correlation with 

the texture of the corresponding rice varieties after cooking 

(Fig. 1). These results showed that (1) unlike as so far believed, 

amylopectin also bound iodine, (2) this binding was 

apparently caused by the presence of some very long chains in 

it, (3) rice varieties therefore apparently differed in their amylopectin 

chain structure, and (4) the apparent amylose content 

of starch as determined by reaction with iodine was always an 

overestimation, part of the value being contributed by amylopectin. 

Hizukuri in collaboration with IRRI scientists shortly 

thereafter showed that the chain profile of rice amylopectin 

indeed differed from variety to variety, with high-amylose rice 

having more very long chains than low-amylose rice, with intermediate-amylose 

rice coming in between (Hizukuri et at 

1989). Radhika Reddy et al (1993) later confirmed these findings 

and also showed that high insoluble-amylose rice amylopectin 

not only had more long chains but also that these chains 

seemed to be largely located externally in the molecule and 


lmax of fraction 1 (iodine) (nm) 

r = 0.983*** 

600 

580 

A 

Jaya 

(I) 

0 min 60 min 0 min 60 min 

I 

B 

560 

S317 

(III) 

III 

540 

Intan 

(V) 

V 

520 

0 5 10 15 20 

Insoluble amylose in rice (% d.b.) 

Cooked-rice hardness 

T65 

(VII) 

VII 

Fig. 1. Variation in λmax of rice amylopectiniodine 

complex. Thirteen rice varieties having 

different insoluble amylose contents were fractionated 

over Sepharose 2B and the gel-excluded 

fraction 1 (amylopectin) was reacted 

with iodine. The absorption maxima of the blue 

complexes were read. d.b. = dry basis. (Adapted 

from Chinnaswamy and Bhattacharya 1986.) 

Fig. 2. Photomicrographs of starch granules in 12% rice-flour pastes 

of four varieties (types I, III, V, VII, having from very high to low 

insoluble amylose). Flours were pasted by heating up to 95 °C. (A) 

Light microscopy, (B) scanning electron microscopy. For each, left 

column: just pasted (0 min); right column: maintained at 95 °C for 

another 60 min. After 60 min, the granules are hardly affected in 

type I but are completely degraded in type VII. (Reproduced from 

Sandhya Rani and Bhattacharya 1995.) 

vice-versa. Work in many laboratories worldwide (Korea, UK, 

Japan, Taiwan-China) subsequently confirmed these findings. 

The influence of amylopectin 

Clearly, the texture of cooked rice was closely correlated with 

the relative abundance of long chains in its amylopectin. The 

more long chains, the firmer was the rice when cooked and 

vice-versa. The rationale was explained by rheological and 

microscopical studies in my laboratory. It was shown that the 

presence of abundant long chains in the amylopectin led, apparently 

by mutual interaction, to strong and resilient starch 

granules that resisted swelling and breakdown, whereas their 

absence led to weak and fragile granules that broke down easily 

during cooking (Fig. 2). 

Thus, after more than 50 years of intensive research, the 

texture of cooked rice and other rice end-use quality has been 

attributed primarily to the relative abundance of long chains 

in its amylopectin. Interestingly, other work has shown that 

the bulk of rice starch consists of amylopectin, true amylose 

being not more than about 10% in it (Ramesh et al 1999). 

Other factors 

Very recent work, primarily in Japan, has shown that the GT 

of rice starch, too, is correlated with the chain structure of 

amylopectin. The proportion of very short chains among the 

overall short chains in amylopectin is negatively correlated 

with the GT (Umemoto et at 2002). However, the implication 

of this finding in terms of rice quality is uncertain. From all 

accounts, rice GT does not seem to have a significant influence 

on its cooking-eating characteristics. 

A few other factors may affect rice-use value in a small 

way. Protein content of rice has long been thought to influence 

rice-use quality. This was first proposed by Spanish workers 

in the 1950s and 1960s. Subsequently, Japanese and other 

workers supported this contention. Recently, Ohtsubo and his 

group at the Food Research Institute in Japan have shown that 

the surface hardness of cooked rice especially is related to protein 

(Okadome et al 1999). American scientists have shown 

that rice protein may have some relationship to its paste properties. 

In addition, the content of free amino acids and sugars 

in rice has been suggested by Japanese scientists to affect the 

palatability and flavor of rice after cooking. 

References 

Bhattacharya KR, Sowbhagya CM, Indudhara Swamy YM. 1978. 

Importance of insoluble amylose as a determinant of rice quality. 

J. Sci. Food Agric. 29:359-364. 

Biliaderis CG, Grant DR, Vose JR. 1981. Structural characterization 

of legume starches. 1. Studies on amylose, amylopectin and 

beta-limit dextrins. Cereal Chem. 58:496-502. 

Chinnaswamy R, Bhattacharya KR. 1986. Characteristics of gel-chromatographic 

fractions of starch in relation to rice and expanded 

rice-product qualities. Staerke 38:51-57. 

Session 8: Improving rice quality 247

Hizukuri S, Takeda Y, Maruta N, Juliano BO. 1989. Molecular structures 

of rice starch. Carbohydr. Res. 189:227-235. 

Okadome H, Toyoshima H, Ohtsubo K. 1999. Multiple measurements 

of physical properties of individual cooked rice grains 

with a single apparatus. Cereal Chem. 76:855-860. 

Radhika Reddy K, Ali SZ, Bhattacharya KR. 1993. The fine structure 

of rice-starch amylopectin and its relation to the texture 

of cooked rice. Carbohydr. Polym. 22:267-275. 

Ramesh M, Ali SZ, Bhattacharya KR. 1999. Structure of rice starch 

and its relation to cooked-rice texture. Carbohydr. Polym. 

38:337-347. 

Sandhya Rani MR, Bhattacharya KR. 1995. Microscopy of rice starch 

granules during cooking. Staerke 47:334-337. 

Sowbhagya CM, Ramesh BS, Bhattacharya KR. 1987. The relationship 

between cooked-rice texture and the physicochemical 

characteristics of rice. J. Cereal Sci. 5:287-297. 

Umemoto T, Yano M, Satoh H, Shomura A, Nakamura Y. 2002. 

Mapping of a gene responsible for the difference in amylopectin 

structure between japonica-type and indica-type rice varieties. 


Notes 

Improving rice grain quality in Thailand 

Kunya Cheaupun, Sunantha Wongpiyachon, and Ngamchuen Kongseree 

Author’s address: Rice Research and Development Centre, 2633, II 

Main Road, V.V. Mohalla, Mysore 570002, India, e-mail: 

krb@rrdc.net. 

Rice, the most important crop in Thailand, was expected to 

alleviate the economic problem of jobless people in 1998. 

During 1998-2003, rice production obtained from main and 

second crops was 22–26 million tons annually. About 6.8–7.3 

million t of milled rice were exported in 1999-2003. Approximately 

50% of the commodity was used domestically, and the 

surplus exported to foreign markets. In 2003, Thai rice exports 

made up about 27% of the world market. Among this 

proportion of exports, good-quality rice represented 46%. This 

included Thai aromatic rice or Jasmine rice. Thus, Thailand’s 

rice program is projected to enhance production, sufficient for 

self-consumption and for export. Rice with good grain quality 

must aim to meet the requirements of both domestic and foreign 

consumers. 

Improving quality 

Rice quality is considered to have two general aspects: 

1. Milling, cooking, and processing qualities, which refer 

to suitability of the grain for a particular end-use. 

2. Physical quality, which means cleanliness, soundness, 

and being free from foreign materials. Usually, physical 

quality and milling, cooking, and processing qualities 

are interrelated in that certain market grades of 

rice are more suitable than others for consumer products. 

The quality of Thai rice has long been well known in 

international markets. The quality of rice is mainly defined as 

long and slender. Breeding selection has incorporated grain 

quality into varietal improvement with other agronomic characters. 

Rice with long, slender, and translucent grain is preferable. 

Different varieties produce a different quality of cooked 

rice; glutinous rice is very sticky, and steaming is always applied 

to prepare the cooked product. Nonglutinous rice may 

be soft and slightly sticky before cooking, and rather soft or 

hard in its texture when cooked. The main factor influencing 

this property is amylose content. The low-amylose type (less 

than 20% amylose) always has a soft texture and is sticky when 

cooked. The intermediate-amylose type (21–25%) produces 

rather soft cooked rice, whereas the high-amylose type has a 

hard texture. To obtain an optimum cooked-rice quality, highamylose 

milled rice requires more cooking water than that 

having a lower amylose content. To cook rice, the cooking 

time depends on the gelatinization temperature of the starch 

(gel. temp.) in milled grain. On the basis of gel temperature, 

rice varieties can be classified into three groups: low, intermediate, 

and high. Thai rice has a great variation in cooked-rice 

(Table 1). We have glutinous rice and various types of 

nonglutinous rice, providing an advantage to satisfy the preferences 

of different groups of rice consumers. Aroma is a special 

characteristic and is considered to be a special quality. 

Quality testing is therefore primarily a matter of determining 

whether the rice is suitable for a particular use and whether it 

meets specific requirements of cleanliness and purity. 

Furthermore, milling yields are based on the amount of 

whole milled kernels and all sizes of broken kernels obtainable 

from a unit of rough rice. Many factors influence milling 

yield; the miller must also judge the quality of rice for milling 

on the basis of variety, moisture content, the presence of red 

rice, peckiness (parboiled kernels with obviously black or dark 

brown area) and other kernel damage, foreign seeds and other 

material, chalky kernels, and mustiness or other undesirable 

odors (Webb and Stermer 1972). 

The world rice market classifies the commodity into six 

basic types: (1) predominantly indica, high-quality, long-grain 

raw milled rice; (2) predominantly indica, medium-quality, 

long-grain raw milled rice; (3) japonica short- or medium-grain, 

raw milled rice; (4) parboiled rice; (5) aromatic (fragrant) rice; 

and (6) waxy (glutinous) rice. Among these types, five have a 

significant role in the international market. These groups can 

also be subdivided depending on consumer preferences. Each 

type has different markets and, within each type, there are spe- 


Table 1. Grain characteristics of some Thai rice varieties. 

Varieties Grain length % amylose Gelatinization 

(mm) 

temp. 

Soft-sticky cooked rice 

KDML 105 a 7.4 12–18 Low 

RD 15 a 7.5 14–18 Low 

RD 21 7.3 17–19 Low 

Khao’ Jow Hawm 7.8 18–19 Low 

Khlong Luang1 a – – – 

Khao’ Jow Hawm 7.7 18–19 Low 

Suphan Buri a – – – 

Pathum Thani1 a 7.6 14–19 Low 

Rather soft cooked rice 

Khao Pahk Maw 148 7.7 24–26 Intermediate 

RD23 7.3 23–27 Intermediate 

Suphanburi 60 7.5 20–26 Low 

Hard cooked rice 

Leuang Pratew 123 7.4 28–32 Low-intermediate 

Suphanburi 1 7.3 29 Intermediate 

Suphanburi 90 7.4 27–30 Low-intermediate 

Chai Nat 1 7.4 27–30 Intermediate 

Pathum Thani 60 a 7.5 27–32 Low 

a 

Aromatic rice. 

cial quality demands for specific uses. In general, consumers 

view rice in an uncooked or unprocessed state and make judgments 

on its quality. To meet market standards, major rice exporters 

and importers have established official grades to identify 

the relative quality of the commodity. 

Thai rice standards 

As the major rice exporter, Thailand established Thai Rice 

Standards B.E. 2540. These standards were revised in 1997 

through a notification by the Ministry of Commerce. These 

rice standards described the minimum specifications of rice in 

each type and grade for domestic and international trade. They 

include white rice (milled rice), cargo rice (Loonzain rice, 

brown rice, or husked rice), white glutinous rice (milled waxy 

rice), and parboiled rice. The judgment is based on physical 

examination of the grain considering the following categories. 

Rice classification 

Classes of rice refer to the unbroken kernels of various length, 

which are a mixture in accordance to the proportion. They can 

be divided into four groups as follows: 

Class 1 long grain more than 7.0 mm 

Class 2 long grain 6.6–7.0 mm 

Class 3 long grain 6.2–6.6 mm 

Short grain Less than 6.2 mm 

High-quality rice is composed of class 1 long grain in a 

greater proportion than that of the lower grades. The percentage 

of this grain becomes less as the grade becomes lower. 

Grain composition 

Grain composition is whole kernels, head rice, and broken rice. 

Whole kernels are unbroken kernels and kernels retaining 

9/10 of the grain unbroken. The size of the broken rice depends 

on the rice grade. 

High-quality rice not only contains a higher proportion 

of class 1 long grain than that of other lower grades, but the 

sizes of head rice and broken kernels are also longer. Small 

broken rice (C1) that passed through a round-hole metal sieve 

no. 7 (0.79 mm in diameter and 1.75 mm in thickness) is not 

allowed to mix with 100% class white rice. 

Rice and other matter 

Some allowances are specified for each grade of rice, as follows. 

Red kernel: rice having red bran covering the kernel 

wholly or partly. 

Yellow kernel: rice having some parts of the kernels 

obviously turned yellow. 

Chalky kernel: nonglutinous rice having an opaque 

area of more than 50% of the kernel. 

Damaged kernel: rice kernels obviously damaged by 

moisture, heat, insects, or something else. 

Foreign matter: including underdeveloped grain, immature 

kernels, excess kernels, other seeds, rice husk, 

and bran. 

Black kernels (whole kernels), partly black kernels 

(more than 25% of the kernel becomes black), and 

pecked kernels (less than 25% become pecked) are 

limited in the parboiled rice standard. 

Paddy: glutinous rice in nonglutinous rice or 

nonglutinous rice in glutinous rice is also specified in 

all types of rice standards. 

Milling degree 

This is divided into four levels: extra well milled, well milled, 

reasonably well milled, and ordinarily well milled. 

Moisture content 

The moisture content of rice of all types and grades is specified 

as not exceeding 14%. 

The adulteration problem of Thai Jasmine rice 

The popularity of Thai aromatic or Jasmine rice increased remarkably 

from 1992 to 1997. The annual export quantities of 

aromatic rice were as high as 1.06–1.45 million t. It represented 

20–27% of the total quantity exported. The commodity 

was produced from varieties KDML 105 and RD15. Production 

was only 3–4 million t of paddy, which was not enough 

for market demand. This brought about an increase in the paddy 

price. The grains of these two varieties result in soft and sticky 

cooked rice that is too sticky for some consumers. The high 

price and this unfavorable cooked-rice texture brought about 

a blending of grain of aromatic rice with that of other varieties 

having a similar grain appearance. In general, these practices 


are in accordance with agreements between buyers and sellers. 

The grain of rice varieties used for blending comes from 

RD23 (intermediate amylose content, intermediate gelatinization 

temperature) and Chai Nat 1 (high amylose content, intermediate 

gel temperature). Blended rice resulted in changes in 

cooked-rice quality and introduced complaints from some customers 

who expected to have soft-texture aromatic cooked rice. 

To satisfy Thai aromatic rice consumers, the government took 

the following action. 

Establishment of Thai Hom Mali Rice Standards B.E. 2544 

Thai aromatic rice is always exported using the name Thai 

Hom Mali Rice, Jasmine Rice, or Thai Fragrant Rice. The commodity 

was certified in accordance with Thai Rice Standards 

through physical measurements. It is difficult to classify aromatic 

rice grain separate from other white rice. For the benefit 

of Thai exporters, the Ministry of Commerce, with cooperation 

from the Ministry of Agriculture, cooperatives, and the 

private sector, established a notification on the subject: Thai 

Hom Mali Rice Standards B.E. 2544. These standards comprised 

both physical and chemical analyses. The name Hom 

Mali was chosen and is well known among Thais in the domestic 

market. The standards disclose the name of Thai Hom 

Mali to the public, like Basmati of India and Pakistan. The 

standards concern the following categories: 

1. Thai Hom Mali Rice means cargo rice and white 

rice obtained from paddy of nonglutinous fragrant rice 

varieties that are produced in Thailand, and certified 

by the Ministry of Agriculture and Cooperatives, such 

as Khao Dowk Mali 105 and RD15. This rice has a 

natural fragrant aroma depending on whether it is new 

or aged rice. The cooked rice kernels have a tender 

texture. 

2. Characteristics and size of rice kernels. The rice 

kernels are long grain: the average length of unbroken 

kernels shall not be less than 7.00 mm, with a 

length/width ratio not less than 3.2. 

3. Chemical properties of the grain indicate its amylose 

content, 13–18%, at 14% moisture content. 

4. Types of rice. There are two types: white rice and 

cargo rice. 

5. Quality of Thai Hom Mali Rice. If it has a mixture 

of other rice, the other rice shall not exceed 8.0% by 

weight. 

6. Grades of rice are defined in accordance with a notification 

from the Ministry of Consumer Subjects: 

Rice Standards 2540, except for grain classifications, 

which are divided as follows: 

White rice: 100% grade A, 100% grade B, 100% 

grade C, 5%, 10%, and 15% broken rice A1 extra 

super, and broken rice A1 super. 

Cargo rice: 100% grade A, 100% grade B, 100% 

grade C, and 5%, 10%, and 15%. 

7. White rice and cargo rice that have mixtures of 

other rice exceeding 8.0% by weight shall not be 

considered as Thai Hom Mali Rice. Using other or 

previous trade names, exporters can still do business. 

The advantage of these standards is to ensure that both 

exporters and importers meet the standards for rice quality of 

Thai Hom Mali. Amylose content for low-amylose rice is 15– 

19%. This refers to soft and sticky cooked rice. Testing for 

other rice based on alkali spreading value is used to detect 

grain of RD 23 and Chai Nat 1. 

Bibliography 

Agricultural statistics of Thailand, crop year 2003/2004. Center for 

Agricultural Information, Office of Agricultural Economics, 

Ministry of Agriculture and Cooperatives, Bangkok, Thailand. 

www.oae.go.th. 

Department of Foreign Trade 2545. Thai Rice Standards. Notification 

of Ministry of Commerce. The Royal Gazette, Vol. 114, 

Section 31D,17 April, B.E. 2540, 1997. 

Department of Foreign Trade 2545. Thai Hom Mali Rice Standards. 

Notification of Ministry of Commerce. The Royal Gazette, 

Vol. 119, Special Part 55d, 21 June, B.E. 2545, 2002. 

Webb BD, Stermer RA. 1972. Criteria of rice quality. In: Houston 

DF, editor. Rice: chemistry and technology. St. Paul, Minn. 

(USA): American Association of Cereal Chemists. p 102-139. 

Notes 

Authors’ addresses: Kunya Cheaupun and Sunantha Wongpiyachon, 

Pathumthani Rice Research Center, Thanyaburi, Pathumthani 

12110, Thailand; Ngamchuen Kongseree, Office of the Senior 

Experts, Department of Agriculture, Bangkok 10900, 

Thailand, e-mail: ckunya@hotmail.com. 


New tools for understanding starch synthesis 

Melissa Fitzgerald, Jeffrey Castro, Rosa Paula Cuevas, and Robert Gilbert 

Starch is the major component of rice grain, and its structure 

contributes significantly to the “quality” of cooked rice. Starch 

structure can be described on several levels—ranging from 

the organization of compound starch granules (Whistler and 

BeMiller 1997) within the cells of the endosperm down to the 

connectivity of the chains in the individual molecules. 

Amylopectin and amylose are accumulated in the developing 

grains of all cereals. Amylopectin is deposited in semicrystalline 

growth rings, in alternating zones of crystalline and 

amorphous regions (French 1984). It is unclear whether amylose 

is interspersed within the amorphous lamella of the semicrystalline 

rings, or if it is deposited within the amorphous 

growth rings. The amorphous growth rings are known to contain 

carbohydrates linked by both alpha 1-4 and alpha 1-6 bonds 

(Atkin et al 1999), but the exact contents of those growth rings 

are unknown. 

The semicrystalline growth rings of starch are understood 

much more than are the amorphous growth rings. For amylopectin, 

crystalline lamellae alternate with amorphous lamellae. 

The synthesis of amylopectin occurs through the agency 

of a suite of enzymes, many of which have been identified, 

cloned, and sequenced, showing that the biosynthesis of starch 

has been strongly conserved throughout the plant kingdom. 

However, the exact function of each enzyme is not so well 

understood. Attempts to understand function have been mostly 

undertaken by comparing the molecular weight distribution 

(MWD) of the amylopectin chains of wild types with that of 

mutants or antisense lines that lack the activity of one enzyme 

in the pathway of starch synthesis (Myers et al 2000). Here we 

introduce a new technique for understanding the full MWD of 

debranched amylopectin, using methods that have been developed 

for quantitatively interpreting MWDs in synthetic polymers 

(Clay and Gilbert 1995). In brief, if the only events occurring 

during synthesis are random chain growth and stoppage, 

then the number of chains of degree of polymerization 

N, P(N), has the form ln[P(N)] ∝N, where the constant of proportionality 

is the ratio of the stoppage and growth rates. This 

starting point suggests that mechanistic inferences can be made 

from a plot of lnP(N) against N. Castro et al (submitted) have 

described the validation of the application of this model to 

starch, and mathematical proof of the model. 

Methods 

Starch was isolated from rice, wheat, barley, maize, and potato 

by established techniques. The starches were gelatinized, 

debranched, and labeled with 8-amino-1,3,6-pyrenetrisulfonic 

acid (APTS) and analyzed by capillary eElectrophoresis 

(O’Shea et al 1998). 

To investigate the structure of the amorphous material, 

flour (250 mg) from 16 waxy varieties of rice was mixed with 

25 g MilliQ water. The slurries were gelatinized with a Rapid 

Visco Analyser (RVA, Newport Scientific) (AACC method 61- 

02). The mixture was centrifuged immediately (10,000 g, 10 

min). The supernatant contained the hot water soluble (HWS) 

starch. The HWS material was debranched, labeled with APTS, 

and analyzed as described above. 

Results 

Figure 1A shows the logarithmic plots for the range of species. 

All the higher plants studied exhibit similar logarithmic 

MWD plots: an initial increasing nonlinear region (which we 

define as Region 1), followed by two distinct decreasing linear 

regions (Regions 2 and 4) separated by a nonlinear region 

(Region 3). Seven metrics of the plots have been defined (Fig. 

1B): four slopes and three intersects of the slopes. The slopes 

and position of the regions on the logarithmic plots are independent 

of the normalization of the original data. Comparing 

the various species, the slopes of the linear regions (2 and 4) 

clearly vary, but all the intersects are similar (Fig. 1A). 

Figure 2 shows the logarithmic plot of the debranched 

HWS starch and full flour for one representative from the collection 

of 16 varieties of waxy rice. All intersects differ and 

the slope of both Regions 2 and 4 differs. The logarithmic plots 

of the full flour and HWS starch for all 16 varieties showed 

the same differences, but these are not shown here. 

Discussion 

Starch synthesis is a multistep process, containing both random 

and nonrandom elements (Thompson 2000, Caldwell and 

Matheson 2003). Plotting the MWDs of the debranched starch 

as logarithmic plots (Fig. 1) allows the identification of these 

random and nonrandom elements. Monotonically decreasing 

functions (Regions 2 and 4) indicate random elements (Clay 

and Gilbert 1995). The monotonically increasing nonlinear 

regions of the lnP curves (Regions 1 and 3) are of particular 

interest—if all chains began at DP 1 and grew until a random 

stopping event, then P(N), and hence lnP(N), must necessarily 

be a decreasing function. Two decreasing functions added together 

still produce a decreasing function. This implies that 

some nonrandom event must be involved in generating Regions 

1 and 3. Region 1 chains are between DP 6 and 10 and 

are probably remnants of branching enzyme events (chain stoppage) 

rather than originating from controlled synthesis. 

Regions 2 and 4 appear to define two populations of 

chains. The longer chains (Region 4) must originate from the 

shorter chains (Region 2), as there is no evidence for a mecha- 


InP(N) + offset 

3 

A 

2 

InP(N) 

18 

17 

HWS 

Full flour 

1 

16 

0 

–1 

–2 

–3 

15 

14 

13 

12 

11 

–4 

10 

0 20 40 60 80 

–5 

0 

InP(N) 

20 40 60 80 100 

N 

Fig. 2. lnP(N) plots of the debranched hot water soluble (HWS) 

and full flour starch of chains from DP 6 to DP

fer between the plots, and the difference is consistent among 

the collection of 16 varieties. This shows that the kinetics of 

the synthesis of these HWS molecules differ, suggesting 

strongly that the HWS molecules composed of branches between 

DP 6 and 70 (consistent with being amylopectin) are 

actually synthesized by different processes from amylopectin 

molecules. Further analysis of the metrics of these plots is required 

to determine whether these HWS molecules account 

for the chains between DP 33 and 47, but we cautiously support 

the possibility of a third fraction of starch that could exist 

in the amorphous growth rings. Further, we suggest that this 

molecule could account for all the reasons behind the term 

“apparent amylose” and account for the alpha 1-6 linkages 

observed by Atkin et al (1999) in the amorphous growth rings. 

We propose that these molecules are an intrinsic part of the 

starch granule and, to reflect this, we propose that this fraction 

be called amyloglycan. 

References 

Atkin NJ, Cheng SL, Abeysekera RM, Robards AW. 1999. 

Localisation of amylose and amylopectin in starch granules 

using enzyme-gold labeling. Starch-Starke 51:163-172. 

Chinnaswamy R, Bhattacharya KR.1986. Characteristics of gel-chromatographic 

fractions of starch in relation to rice and expanded 

rice-product qualities. Starch-Starke 38:51-57. 

Clay PA, Gilbert RG. 1995. Molecular weight distributions in freeradical 

polymerizations. 1. Model development and implications 

for data interpretation. Macromolecules 28:552-569. 

French D. 1984. Organization of the starch granules. In: Whistler 

RL, Bemiller JN, Pasehall JF, editors. Starch: chemistry and 

technology. Orlando, Fla. (USA): Academic Press. 

Hizukuri S, Takeda Y, Maruta N, Juliano BO. 1989. Molecular structures 

of rice starch. Carb. Res. 189:227-235. 

Imberty A, Buleon A, Vinh T, Perez S. 1991. Recent advances in 

knowledge on starch structure. Starch-Starke 43:375-384. 

Myers AM, Morell MK, James MG, Ball SG. 2000. Recent progress 

toward understanding biosynthesis of the amylopectin crystal. 


O’Shea MG, Samuel MS, Konik CM, Morell MK. 1998. 

Fluorophore-assisted carbohydrate electrophoresis (face) of 

oligosaccharides: efficiency of labelling and high-resolution 

separation. Carb. Res. 307:1-12. 

Thompson DB. 2000. On the non-random nature of amylopectin 

branching. Carb. Polymers 43:223-229. 

Whistler RL, BeMiller JN. 1997. Carbohydrate chemistry for food 

scientists. St. Paul, Minn. (USA): American Association of 

Cereal Chemists, Inc. 

Notes 

Authors’ addresses: Melissa Fitzgerald and Rosa Paula Cuevas, International 

Rice Research Institute, DAPO Box 7777, Metro 

Manila, Philippines; Melissa Fitzgerald and Jeffrey Castro, 

CRC for Sustainable Rice Production, PMB, Yanco 2703, 

NSW, Australia; Melissa Fitzgerald, NSW Department of Primary 

Industries, PMB, Yanco 2703, NSW, Australia; Jeffrey 

Castro and Robert Gilbert, Key Centre for Polymer Colloids, 

School of Chemistry, The University of Sydney 2006, NSW, 

Australia, e-mail: m.fitzgerald@cgiar.org. 

Textural differences between indica 

and japonica varieties in cooked rice 

Keiko Hatae, Sonoko Ayabe, and Midori Kasai 

Indica rice varieties are popular worldwide. Cooked indica rice 

is hard but not sticky. However, Japanese people are fond of 

japonica cooked rice because of its moderate elasticity and 

stickiness. It is generally said that the amylose content contributes 

to the hardness and stickiness of cooked rice. Japonica 

rice has lower amylose content and moderate elasticity and 

stickiness. 

This study was carried out to confirm the textural difference 

between cooked indica and japonica rice and to know the 

reasons for the difference. Among indica varieties, Khao Dawk 

Mali is known to contain less amylose; thus, Khao Dawk Mali 

and a higher amylose Thai rice and japonica Nipponbare were 

used to compare their texture and some physicochemical properties. 


Two cultivars of Oryza sativa L. indica (Khao Dawk Mali 105 

and unknown high-amylose Thai rice) obtained from Thailand 

in 2001 were used. They were stored at 4 °C. Japonica rice, 

Oryza sativa L. japonica (Nipponbare), grown in Shiga Prefecture, 

Japan (2001), was used. Khao Dawk Mali 105 and 

Nipponbare were 90% milled in a polishing machine before 

use. 

Milled rice grain was ground with an electric miller, and 

the resulting flour was passed through a 100-mesh sieve. Starch 

was isolated from milled rice flour by using cold 0.2% NaOH 

according to the alkali method of Yamamoto et al (1973) with 

modifications. The precipitate was washed with distilled water 

until the pH decreased to 7.0. After rinsing with ethanol, 

the starch was air-dried and passed through a 100-mesh sieve. 

Milled rice grains were washed three times with distilled wa- 


Table 1. Textural properties of cooked rice from three varieties. a 

How cooked Textural property Nipponbare Khao Dawk Mali High-amylose rice 

Cooked with 1.4 Hardness (kgf) 2.19 ± 0.22 a 2.45 ± 0.15 a 3.74 ± 0.38 b 

times water Stickiness (kgf) 0.34 ± 0.04 a 0.10 ± 0.01 b 0.03 ± 0.04 c 

Cooked with 1.9 Hardness (kgf) 1.98 ± 0.19 a 2.18 ± 0.30 a 3.17 ± 0.16 b 

times water Stickiness (kgf) 0.41 ± 0.03 a 0.14 ± 0.06 b 0.04 ± 0.01 c 

a 

Mean ±SD of ten measurements. For letters a, b, and c, the same letter in the same row is not significantly 

different (P λ max ≥600 nm; Fr. II: 600 nm > λ max 

≥525 nm; Fr. III: 525 nm > λ max). 


Proximate analysis 

The moisture contents of the three rice flours ranged from 

13.4% to 14.5% and crude protein from 5.34% to 7.62%. Apparent 

amylose contents were almost the same (16%) in Khao 

Dawk Mali and Nipponbare, which were different from highamylose 

Thai rice (27%). This confirmed that indica variety 

Khao Dawk Mali and japonica variety Nipponbare have the 

same level of amylose. 

Textural properties 

Table 1 shows the hardness and stickiness of rice cooked with 

1.4 times water or 1.9 times water. The hardness of Nipponbare 

and Khao Dawk Mali was the same, but the stickiness was 

different. And those of high-amylose Thai rice were quite different 

from those of other rice varieties. Thus, the textural difference 

of cooked rice could not be explained by only the difference 

in amylose content. 

Degree of gelatinization 

The degree of gelatinization was almost 100% in Nipponbare 

and Khao Dawk Mali, both cooked with 1.4 and 1.9 times 

water. In contrast, the value of the high-amylose Thai rice was 

84% even when cooked with 1.9 times water. The degree of 

gelatinization is not enough to explain the textural difference 

of cooked rice. 


Total saccharide ratio ( ) 

12 

Nipponbare 

40 

12 

Khao Dawk Mali 

40 

Average degree of polymerization (d.p.) 

12 

High-amylose rice 

40 

10 

10 

10 

30 

30 

30 

8 

8 

8 

6 

20 

6 

20 

6 

20 

4 

4 

4 

10 

10 

10 

2 

2 

2 

0 

0 0 

0 0 

0 

25 30 35 40 45 50 55 25 30 35 40 45 50 55 25 30 35 40 45 50 55 

Fraction no. 

Fraction no. 

Fraction no. 

Fig. 1. GFC patterns of extracts from three kinds of cooked rice surface after isoamylase treatment. 

Weight of precipitates and sugar and protein 

contents of supernatant of cooked-rice extracts 

The weight of precipitates was the largest in Nipponbare, followed 

by Khao Dawk Mali, and the smallest in high-amylose 

Thai rice. The same order in the contents of total sugar, reducing 

sugar, and protein of supernatant was obtained. It seemed 

that the large amount of sugar in the supernatant was an important 

factor, especially in the stickiness of cooked rice. 

FT-IR spectrometry 

FT-IR is a useful tool to determine intermolecular interaction. 

This time, the difference in intermolecular interaction of the 

three kinds of rice starches (flour with 50% w/w of water) was 

measured. A wave length of 1,000–1,100 cm –1 is known to 

denote the interaction between C and O and C and C, and 1,020 

cm –1 is also known to show the absorption of sugar. Thus, we 

used a wave length of around 1,000 cm –1 . 

Three FT-IR patterns for each three kinds of rice were 

obtained, that is, raw rice starch, cooked rice, and extracts of 

the surface of cooked rice. Moreover, a difference in IR spectrum 

of the cooked rice and the extracts was found. From the 

results, the difference in IR spectrum of this approximate wave 

length was negatively larger in Nipponbare, followed by Khao 

Dawk Mali. On the other hand, high-amylose Thai rice was 

positively larger. This negatively large difference in cooked 

rice and the extracts means that a larger amount of sugar was 

in the surface layer than inside the cooked rice. This may be 

one reason to explain the textural difference of the cooked 

rice. 

Gel permeation chromatography 

Figure 1 shows the gel permeation patterns of rice starch after 

iso-amylase treatment. For raw rice, the patterns of Nipponbare 

and Khao Dawk Mali were similar and that of high-amylose 

Thai rice was different. The long-chain fraction was largest in 

high-amylose Thai rice. 

When three kinds of rice were cooked, Nipponbare decreased 

its amylose fraction, but its short-chain (about 15 dp) 

amylopectin fraction increased slightly. The pattern of the 

Nipponbare extracts was almost the same as with cooked rice. 

In Khao Dawk Mali, the amylose fraction decreased to the 

same level as Nipponbare when cooked. However, this amylose 

fraction increased in the extracts. In high-amylose rice, 

the amylose fraction decreased largely after cooking, whereas 

the short-chain fraction increased. And, in the extracts, the longchain 

fraction (amylose fraction) increased even much more 

than raw starch. 

From these findings, we assumed that the distribution of 

amylose and amylopectin was almost equal in Nipponbare, but 

not in the indica varieties. 

Conclusions 

It is suggested that the stickiness of cooked rice is related to 

the sugar content in the surface layer of cooked rice extracts. 

These extracts also contained a lower amount of amylose fraction 

and larger amount of short-chain amylopectin fraction. 

These components are once eluted from the rice during cooking 

and finally absorbed into the surface layer, thus giving 

stickiness to rice. 

References 

Hizukuri S, Takeda Y, Yasuda M, Suzuki A. 1981. Multi branched 

nature of amylose and the action of debranching enzymes. 

Carbohydr. Res. 147:342-347. 


Inouchi N, Glover DV, Takaya T, Fuwa H. 1987. Chain length distribution 

of amylopectin of several single mutants and the normal 

counterpart, and sugary-1 phytoglycogen in maize (Zea 

mays L.). Starch 39:259-266. 

Itoh T, Yoshio N, Teranishi K, Hisamatsu M, Yamada T. 1994. Saccharides 

extracted from rice grains during cooking process 

and effect of monoglyceride on them. Nippon Sholuhin Kogyo 

Gakkaishi 41:871-877. 

Juliano BO. 1971. A simplified assay for milled rice amylose. Cereal 

Sci. Today 16:334-338,340,360. 

Lowry OH, Rosenbrough GJ, Farr AL, Randall RJ. 1951. Protein 

measurement with the folin phenol reagent. J. Biol. Chem. 

193:265-275. 

Matsunaga A, Kainuma K. 1981. Studies on the retrogradation of 

starchy foods (Part 1). Retrogradation of cooked rice. J. Home 

Econ. Jpn. 32:653-659. 

Okabe M. 1977. Research on the taste of cooked rice. New Food 

Industry 19:65-71. 

Yamamoto K, Sawada S, Onogaki T. 1973. Properties of rice starch 

prepared by alkali method with various conditions. Denpun 

Kagaku 20:99-104. 

Notes 

Authors’ address: Department of Food and Nutrition, Ochanomizu 

University, e-mail: hatae@cc.ocha.ac.jp. 

Biological efficacy of consuming rice biofortified with iron 

J. Haas, J. Beard, A. Del Mundo, G. Gregorio, L.M. Kolb, and A. Felix 

The World Health Organization estimates that 3.5 billion people 

in the developing world are affected by iron-deficiency anemia. 

Given the need for a sustainable solution to reduce this 

number, biofortified rice (variety IR68144) was used to test 

the feasibility of high-iron rice as a source of dietary iron. We 

conducted a double-blind longitudinal (9 months) intervention 

study with religious sisters from 10 convents in and around 

Manila, Philippines. Some 317 religious sisters were randomly 

assigned to consume either an iron-enhanced (high-iron) rice 

or a local variety of rice (control). Rice preparation and distribution 

were standardized. The sisters consumed their assigned 

rice at every meal and food intakes were weighed on 3 days 

every 2 weeks for each woman. Blood samples were collected 

at baseline, mid-point, and end-point and used to determine 

serum ferritin, the transferring receptor, and hemoglobin, 

among other indicators of iron status. The efficacy trial indicated 

that the high-iron rice provided an additional 1.41 mg of 

iron per day, representing a 17% increase in dietary iron in the 

diets of these women. Ferritin and body iron levels were sig- 

nificantly different at the end of the 9-month trial for the 

nonanemic women who consumed the high-iron rice instead 

of the control after controlling for baseline values, amount of 

rice consumed, and convent. The greatest changes in ferritin 

and body iron were observed in the women who consumed the 

most biofortified rice and the calculation of the amount of iron 

transferred from the diet to iron stores is consistent with the 

observed changes in body iron. Therefore, it is possible to 

conclude that there was a significantly positive effect of consuming 

biofortified rice on the iron status of women with an 

iron-poor diet. 

Notes 

Authors’ addresses: J. Haas, J. Beard, and L.M. Kolb, Cornell University, 

USA; G. Gregorio, International Rice Research Institute, 

Los Baños, Philippines; A. Del Mundo and A. Felix, 

University of the Philippines Los Baños, Los Baños, Philippines. 

Radical-scavenging activity of red and black rice 

Tomoyuki Oki, Mami Masuda, Saki Nagai, Miwako Take’ichi, Mio Kobayashi, Yoichi Nishiba, Terumi Sugawara, Ikuo Suda, and Tetsuo Sato 

Rice (Oryza sativa) is widely consumed in the world, and the 

most common type (>85%) has a white pericarp. Other types 

have a colored pericarp, and the most common are green, black, 

and red. The black and red varieties are planted mainly in South 

Asia and other countries, such as Italy, Greece, and the United 

States (Simmons and Williams 1997). Rice with a colored pericarp 

has long been consumed in Japan and China and is considered 

to be a healthy food. We are especially interested in 

the antioxidative and radical-scavenging properties of rice 

because of their potential to provide health protection against 

reactive oxygen species and free radicals, which have been 

implicated in more than 100 diseases (Halliwell 1992). 

On the other hand, there have also been some reports 

concerning the antioxidative compounds found in rice. Gammaoryzanol, 

which is a mixture of ferulate esters of triterpene 

alcohol, is well known as an antioxidant in rice bran (Xu et al 


2001). In black rice, cyanidin-3-glucoside (Cy-3-Glc) has been 

reported to be one of the major antioxidant compounds (Osawa 

1999). However, little information exists on the antioxidative 

effects of red rice, although the existence of catechin and tannin 

has been reported (Nawa and Ohtani 1992). 

In this report, extracts from rice with pericarp of three 

different colors (white, black, and red) were prepared by using 

solvents with various polarities and subjected to a radicalscavenging 

assay. Furthermore, we determined the major radical-scavenging 

components in black and red rice. 



Three rice cultivars, Hinohikari, Saikai-225, and Beniroman, 

which have white, black, and red pericarp, respectively, were 

harvested at KONARC (Chikugo Branch, Fukuoka, Japan) in 

1998. The authors also used Beniroman, Tsukushiakamochi, 

Tsushimazairai, and Tanegashimazairai, which have red pericarp. 

Each cultivar was milled with an ultra centrifugal mill 

(type ZM1000, Retsch GmbH & Co., KG, Germany). The 

milled rice was stored at 4 °C and used within 2 days of milling. 

Preparation of rice extract 

Extraction 1. Milled rice powder (2.0 g) was sequentially extracted 

with six different polar solvents already reported by 

Oki et al (2002a). Extraction was done using 5.0 mL of solvent 

and the following sequence of solvents with increasing 

polarity: n-hexane, diethylether, ethyl acetate, acetone, methanol, 

and deionized water. After each extraction, the residue 

was dried under reduced pressure at ambient temperature for 

1 hour to remove the residual solvent. Each extract was filled 

up to 5.0 mL with the extraction solvent. An aliquot (1.0 mL) 

was dried under reduced pressure at 35 °C, and the residue 

was redissolved in 1.0 mL of dimethyl sulfoxide. 

Extraction 2. Milled rice powder (1.0 or 2.0 g) was extracted 

twice with 5.0 mL of a 70% acetone/0.5% acetic acid 

solution (red rice) and a 5% acetic acid solution (black rice) 

for 16 hours at ambient temperature in the dark. The supernatant 

obtained by centrifugation (1,200x g, 10 min, 4 °C) was 

combined and filled up to 10 mL with the same extraction solution. 

For black rice, an aliquot (2.0 mL) of the extract was 

applied on an Amberlite XAD-2000 column (10 × 40 mm) 

and eluted stepwise with 15 mL of distilled water and a 70% 

ethanol solution. The ethanol fraction was evaporated to dryness 

under reduced pressure and redissolved in 2.0 mL of deionized 

water. 

Assay for DPPH • and t-BuOO • scavenging activity 

The 1,1-diphenyl-2-picrylhydrazyl radical (DPPH • ) and the 

tert-butyl hydroperoxide radical (t-BuOO • ) scavenging activities 

were measured by the methods already reported (Oki et al 

2002a, 2002b). The radical-scavenging activity was expressed 

as a Trolox, water-soluble alpha-tocopherol equivalent. 

Determination of polyphenol, anthocyanin, 

and proanthocyanidin content 

The polyphenol and anthocyanin content were detemined by 

the methods already reported (Oki et al 2002b). The polyphenol 

and anthocyanin content were expressed as a gallic acid 

and Cy-3-Glc equivalent, respectively. 

The proanthocyanidin content was determined by the 

vanillin-sulfuric acid assay already reported (Oki et al 2002a). 

The proanthocyanidin content was expressed as a (+)-catechin 

equivalent. 


Comparison of radical-scavenging activity 

among three different-colored pericarps 

Rice powder was sequentially extracted with n-hexane, diethyl 

ether, ethyl acetate, acetone, methanol, and deionized water, 

and the DPPH • scavenging activity of each extract is shown in 

Figure 1A. The total DPPH • scavenging activity, which was 

calculated from the sum of the activity in each extract, varied 

according to the pericarp color. The highest activity was observed 

in red rice (2.77 µmol-Trolox equiv. mL –1 ), followed 

by black (0.92) and white (0.26) rice. In all varieties, the extracts 

prepared with high polar solvents, methanol, and deionized 

water showed radical-scavenging activity. In rice with 

black and white pericarp, the sum of the activities in two extracts 

corresponded to more than 90% of the total DPPH • scavenging 

activity. In red rice, high radical-scavenging activity 

(1.22 µmol-Trolox equiv. mL –1 ) was observed in the acetone 

extract in addition to the methanol and deionized water extracts. 

The t-BuOO • scavenging activity was estimated by the 

decrease in luminance measured with an analyzer equipped 

with a charge-coupled device camera (type CLA-IMG2, 

Tohoku Electronic Industrial Co., Sendai, Japan). As the activities 

were calculated as a Trolox equivalent per assay, it could 

be observed that the extracts prepared with high polar solvents 

exhibited higher t-BuOO • scavenging activity, which was similar 

to the results of the DPPH • scavenging activity (Fig. 1B). 

The total t-BuOO • scavenging activity, which was calculated 

from the sum of the activity in each extract, decreased in the 

order of rice with red pericarp (4.15 nmol-Trolox equiv. assay 

–1 ), black (2.83), and white (2.09). In addition, the acetone 

extract from red rice exhibited a high t-BuOO • scavenging activity, 

while no such activities of the other cultivars (white and 

black rice) were detected. 

Radical-scavenging components in red rice 

To obtain useful information about the radical-scavenging components 

in the acetone extract prepared from red rice, the extract 

was subjected to hydrochloric acid hydrolysis and a vanillin-sulfuric 

acid assay. In both reactions, the tested solution 

turned red, indicating the presence of proanthocyanidin in the 

extract. In the acid hydrolysate, only cyanidin was identified 

as an anthocyanidin by reverse-phase high-performance liquid 

chromatography (RP-HPLC). This result revealed that the 


DPPH • scavenging activity (mmol-Trolox equiv. mL –1 ) 

1.5 

White rice 

1.0 

0.5 

A 

t-BuOO • scavenging activity (nmol-Trolox equiv. assay –1 ) 

2.0 

White rice 

B 

1.5 

1.0 

0.5 

0 

ND 

ND 

0 

ND 

ND 

ND 

1.5 

Black rice 

2.0 

Black rice 

1.0 

1.5 

1.0 

0.5 

0.5 

0 

ND 

ND 

ND 

0 

ND 

ND 

ND 

1.5 

Red rice 

2.0 

Red rice 

1.0 

1.5 

1.0 

0.5 

0.5 

0 

n-hexane 

Diethylether 

ND 

Ethyl acetate 

Acetone 

Methanol 

Deionized water 

0 

n-hexane 

ND 

Diethylether 

ND 

Ethyl acetate 

Acetone 

Methanol 

Deionized water 

Fig. 1. DPPH • (A) and t-BuOO • (B) scavenging activities of extracts prepared from rice with red, black, 

and white pericarp by using solvents with various polarities. ND = not detected. 

proanthocyanidin in the extract from red rice was the 

procyanidin type, that is, (+)-catechin and/or (–)-epicatechin 

derivatives. The gel permeation chromatography of the acetylated 

proanthocyanidin revealed that the average molecular 

weight is about 5,000, in a range of about 500 to 18,000 after 

proanthocyanidin was purified from the acetone extract by 

using a Sephadex LH-20 column chromatography. 

We also examined DPPH • scavenging activity, polyphenol 

content, and proanthocyanidin content in various rice cultivars 

with red pericarp. The DPPH • scavenging activity varied 

according to the cultivars (1.4 to 46.7 µmol-Trolox equiv. 

g –1 ), and Tanegashimazairai, a traditional wild type of red rice 

in Japan, showed very high radical-scavenging activity. The 

DPPH • scavenging activity was highly correlated to polyphenol 

content and proanthocyanidin content, with correlation 

coefficients of 0.982 and 0.944, respectively (Fig. 2). This result 

indicated that the major radical-scavenging component in 

red rice was proanthocyanidin. 

Radical-scavenging components in black rice 

The black rice extracts were prepared from eight varieties purchased 

at the local market in the Kyushu and Okinawa regions. 

DPPH • scavenging activity of the extracts ranged from 3.1 to 

10.8 µmol-Trolox equiv. mL –1 of the extract. The radical-scavenging 

activity increased with the increase in anthocyanin content, 

and the correlation coefficient was 0.987. By RP-HPLC 

of the extracts, two major peaks were observed in all eight 

black rice varieties, and these peaks were identified as peonidin- 

3-glucoside (Pn-3-Glc) in addition to Cy-3-Glc. A slight compositional 

difference in anthocyanin was observed, however. 

The sum of Cy-3-Glc and Pn-3-Glc constituted more than 85% 

of the total anthocyanin. Therefore, it was revealed that the 

dominant radical-scavengers in black rice were Cy-3-Glc and 

Pn-3-Glc. 

We have already found that two anthocyanins (Cy-3-Glc 

and Pn-3-Glc) were detected in rat plasma after oral administration 

of the anthocyanin-rich concentrate prepared from black 

rice (BR-ANT) (Oki et al 2002c). Furthermore, the BR-ANT 


DPPH • scavenging activity (mmol-Trolox equiv. g –1 ) 

50 

A 

B 

40 

30 

20 

10 

R = 0.982 

R = 0.994 

0 

0 10 20 30 40 50 

Polyphenol content 

(mmol-gallic acid equiv. g –1 ) 

0 10 20 30 40 50 

Proanthocyanidin content 

(mmol-(+)-catechin equiv. g –1 ) 

Fig. 2. Correlation of polyphenol content (A) and proanthocyanidin content (B) with DPPH • scavenging 

activity in various rice cultivars with red pericarp. 

suppressed hemolysis and the oxidation of LDL (Oki et al 

2002c). The BR-ANT also inhibited lipid peroxidation induced 

by four different radical generation systems (Oki et al 2002c). 

These results suggested that dietary consumption of black rice 

might prevent many diseases related to lipid peroxidation 

caused by free radicals. 

On the other hand, there is currently no satisfactory explanation 

for how the polymeric proanthocyanidin is absorbed 

into the bloodstream. Ling et al (2001) have reported that, when 

a hypercholesterolemia rabbit had red rice as a part of its diet, 

the total antioxidative properties of the serum and liver increased. 

They suggested that constituents with antioxidative 

activity could be responsible for the increased antioxidative 

properties in the body; however, they did not identify the antioxidant 

compounds in red rice. This report suggests that, after 

proanthocyanidin in red rice is consumed, it could be absorbed 

by the body directly and/or indirectly (i.e., structural change 

or degradation), which would elevate the antioxidative capacity. 

Now, we are investigating the physiological functions of 

red rice in vivo and elucidating the mechanisms at work in 

these functions. 

References 

Halliwell B. 1992. The role of oxygen radicals in human disease, 

with particular reference to the vascular system. Haemostasis 

23(suppl. 1):118-126. 

Ling WH, Cheng QX, Wang JMT. 2001. Red and black rice decrease 

artherosclerotic plaque formation and increase antioxidant 

status in rabbits. J. Nutr. 131:1421-1426. 

Nawa Y, Ohtani T. 1992. Property of pigments in rice hulls of various 

colors. Food Industry 11:28-33. (In Japanese.) 

Oki T, Masuda M, Kobayashi M, Nishiba Y, Furuta S, Suda I, Sato 

T. 2002a. Polymeric procyanidins as radical scavenging components 

in red-hulled rice. J. Agric. Food Chem. 50:7524- 

7529. 

Oki T, Masuda M, Furuta S, Nishiba Y, Suda I. 2002b. Involvement 

of anthocyanins and other phenolic compounds in radical-scavenging 

activity of purple-fleshed sweet potato cultivars. J. Food 

Sci. 67:1752-1756. 

Oki T, Masuda M, Ikuo S. 2002c. Antioxidative activity of black 

rice commercially available in Kyushu and Okinawa region. 

In: Noguchi T, editor. The 56th Annual Scientific Meeting, 

19-21 July 2002. Hokkaido University (Japan): The Japanese 

Society of Nutrition and Food Science. p 230. (In Japanese.) 

Osawa T. 1999. Protective role of rice polyphenols in oxidative stress. 

Anticancer Res. 19:3645-3650. (In Japanese.) 

Simmons D, Williams R. 1997. Dietary practices among Europeans 

and different South Asian groups in Coventry. Br. J. Nutr. 

78:5-14. 

Xu Z, Hua N, Godber JS. 2001. Antioxidant activity of tocopherols, 

tocotrienols, and γ-oryzanol components from rice bran against 

cholesterol oxidation accelerated by 2,2′-azobis (2- 

methylpropionamidine) dihydrochloride. J. Agric. Food Chem. 

49:2077-2081. 

Notes 

Authors’ addresses: Department of Crop and Food Science, National 

Agricultural Research Center for Kyushu Okinawa Region 

(KONARC), National Agriculture and Bio-oriented Research 

Organization (NARO), 2421 Suya, Nishigoshi, Kikuchi, 

Kumamoto 861-1192, Japan, e-mail: tomooki@affrc.go.jp. 


A rice mutant with enhanced amylose content in endosperm 

derived from low-amylose variety Snow Pearl: isolation 

and characterization 

Yasuhiro Suzuki, Hiro-Yuki Hirano, Yoshio Sano, Kazuo Ise, Ushio Matsukura, Noriaki Aoki, and Hiroyuki Sato 

Amylose content of endosperm starch is an important characteristic 

of rice in determining eating and cooking quality. Amylose 

content is genetically controlled by a major gene, the Waxy 

(Wx) gene (Nelson and Pan 1995). In addition, amylose content 

is also affected by several modifying genes and environmental 

factors such as temperature. The Wx gene encodes the 

granule-bound starch synthase that is responsible for amylose 

synthesis in the endosperm. Rice strains having the Wx b gene 

contain 15–20% amylose, whereas rice strains having the Wx a 

gene contain 20–25% amylose (Sano 1984). The du loci modify 

amylose content by controlling expression of the Wx b gene 

and mutations at the du loci cause a reduction in amylose content 

(Okuno et al 1983, Hirano 1993). Although the amyloseextender 

(ae) mutant shows increasing levels of apparent amylose 

content, this increase is caused by the altered amylopectin 

structure, and the ae mutation is found in the gene for starchbranching 

enzyme IIb (Nishi et al 2001). While there are many 

mutants in starch composition, there is no report about a gene 

whose mutation gives rise to an increase in amylose content in 

the endosperm. In this paper, we describe the isolation and 

characterization of a rice mutant with enhanced amylose content 

in the endosperm that does not affect the amylopectin structure. 

To screen mutants with enhanced amylose content, we 

used a low-amylose rice variety, Snow Pearl (Higashi et al 

1999). In this variety, amylose content was very low (about 

5%) and the endosperm appeared chalky when seeds matured 

at a normal temperature (26 o C) (Table 1). On the other hand, 

when seeds matured at a cool temperature (20 o C), amylose 

content was about 18% and the endosperm appeared 

semitranslucent. Because amylose content in rice seed influences 

the degree of transparency (Suzuki et al 2002), a mutant 

with enhanced amylose content in (semitranslucent) seeds 

matured at a normal temperature could be easily distinguished 

from the original variety with low-amylose (chalky) seeds. 

Dry seeds of Snow Pearl rice were treated with 1 mM 

sodium azide in 0.1 M sodium phosphate buffer (pH 3.0) for 6 

h at room temperature, following washing with tap water. The 

first screening used 7,200 M 1 plants that matured at an average 

temperature of 24–26 o C. Although the panicles of almost 

all of these plants contained exclusively chalky seeds, one of 

these plants contained a mixture of chalky and semitranslucent 

seeds, 41 and 6 seeds, respectively. Next, semitranslucent M 2 

seeds were chosen as a candidate (designated as SP14) for a 

mutant with enhanced amylose content. The M 2 seeds were 

Table 1. Effect of temperature on the accumulation of amylose during 

seed development. a 

Strain 20 o C 26 o C Difference (20–26 o C) 

Snow Pearl 18.3 ± 0.8 5.2 ± 0.5 13.1 

SP14 24.7 ± 0.3 11.7 ± 0.6 13.0 

Koshihikari 25.1 ± 0.8 17.0 ± 0.3 8.1 

a 

Snow Pearl, SP14, and Koshihikari seeds matured at 20 or 26 o C. The amylose 

content of the samples was analyzed based on an iodine colorimetric assay (Juliano 

1971) with slight modifications (Suzuki et al 2002). Data (amylose content, %) are 

mean values ± SE from 5 trials. Twenty mg of white rice powder was used for each 

trial. 

seeded, and amylose content of SP14 (M 3 seeds) was compared 

with that of Snow Pearl and Koshihikari using a temperature-controlled 

growth chamber. The amylose content in 

mature seeds of SP14 was about 12% at 26 o C and 25% at 20 

o C (Table 1), indicating that amylose content of SP14 is responding 

to maturation temperatures and is 6–7% higher than 

that of Snow Pearl. Furthermore, to clarify whether amylose 

content in SP14 was increased by the ae locus or not, the amylopectin 

chain-length distribution in rice endosperm was analyzed 

by high-performance anion exchange chromatography 

with a pulsed amperometric detector, as previously described 

(Suzuki et al 2003). The chain-length distributions of amylopectin 

were almost the same between the two strains matured 

at both temperatures (data not shown). In addition, while 

endosperm appearance of EM10 (ae mutant) was floury, that 

of SP14 as well as that of Snow Pearl and Koshihikari was not 

floury (i.e., wild type = normal). These results indicated that 

the mutation in SP14 would not be the ae mutation, but a new 

mutation that enhances amylose content in rice endosperm. 

It should also be noted whether SP14 is derived from 

Snow Pearl or not. As Snow Pearl has a single nucleotide polymorphism 

(SNP) on the 3rd exon of the Wx gene (Nippon 

Suisan Kaisha, Ltd. 2000), a part of the SP14 Wx gene (344 

base pairs) containing the SNP was compared with that of Snow 

Pearl and Koshihikari. SNP of SP14 was identical with that of 

Snow Pearl, but not with that of Koshihikari, indicating that 

SP14 is a mutant derived from Snow Pearl. Furthermore, the 

nucleotide sequences of the wx locus (6,800 base pairs: Wx 

gene and promoter region) of SP14 and Snow Pearl were identical. 

These results suggested that the mutation from Wx b to 

Wx a in SP14 would not occur and that the mutation in SP14 

would not be on the wx locus. 


Table 2. Segregation of amylose content of F 2 seeds from selfed F 1 rice plants. a 

Seed number 

Chi-square value 

Strain Amylose content Total for 15:1 ratio Probability 

< 8% ≥ 8% 

SP14/Snow Pearl 92 8 100 0.523 0.4 

Snow Pearl 10 0 10 

SP14 0 10 10 

a 

Snow Pearl, SP14, and F 2 rice seeds (SP14/Snow Pearl) matured at 26 o C. The amylose content of white rice 

seeds (15–25 mg) was analyzed based on an iodine colorimetric assay (Juliano 1971) with slight modifications 

(Suzuki et al 2002). 

Next, the mode of inheritance of the character was analyzed 

by crossing SP14 with Snow Pearl. The seeds of F 1 , Snow 

Pearl, and SP14 plants matured at 26 o C. As segregation of 

amylose content in F 2 seeds derived from the cross fit the ratio 

of 15 Snow Pearl-type seeds:1 SP14-type seed, the mutation 

is controlled by two recessive genes (Table 2). Thus, we have 

isolated a new rice mutant (SP14) in which amylose content in 

the endosperm was enhanced compared with the original variety. 

This mutant should be useful in breeding programs designed 

to produce rice of a high quality and in understanding 

genetic and molecular mechanisms for amylose synthesis. 

References 

Higashi T, Saitoh S, Takita T, Yamaguchi M, Sunohara Y, Yokogami 

N, Ikeda R, Tamura Y, Oyamada Z, Kowata H, Inoue M, 

Matsumoto S, Kataoka T. 1999. Breeding of a new rice cultivar 

with low amylose contents, “Snow pearl”. Bull. Tohoku 

Natl. Agric. Exp. Stn. 95:1-12. 

Hirano H-Y. 1993. Genetic variation and gene regulation at the wx 

locus in rice. Gamma-Field Symp. 24:63-79. 

Juliano BO. 1971. A simplified assay for milled-rice amylose. Cereal 

Sci. Today 16:334-340. 

Nelson O, Pan D. 1995. Starch synthesis in maize endosperms. Annu. 

Rev. Plant Physiol. Plant Mol. Biol. 46:475-496. 

Nippon Suisan Kaisha, Ltd. 2000. Method to distinguish a specific 

low-amylose-rice variety. Japan patent exhibition 2000- 

201679. 

Nishi A, Nakamura Y, Tanaka N, Satoh H. 2001. Biochemical and 

genetic analysis of the effects of amylose-extender mutation 

in rice endosperm. Plant Physiol. 127:459-472. 

Okuno K, Fuwa H, Yano M. 1983. A new mutant gene lowering 

amylose content in endosperm starch of rice, Oryza sativa L. 

Jpn. J. Breed. 33:387-394. 

Sano Y. 1984. Differential regulation of waxy gene expression in 

rice endosperm. Theor. Appl. Genet. 68:467-473. 

Suzuki Y, Sano Y, Hirano H-Y. 2002. Isolation and characterization 

of a rice mutant insensitive to cool temperatures on amylose 

synthesis. Euphytica 123:95-100. 

Suzuki Y, Sano Y, Ishikawa T, Sasaki T, Matsukura U, Hirano H-Y. 

2003. Starch characteristics of the rice mutant du2-2 Taichung 

65 highly affected by environmental temperatures during seed 

development. Cereal Chem. 80:184-187. 

Notes 

The role of the water-soluble fraction 

in rice pasting behavior 

Tadashi Yoshihashi, Eizo Tatsumi, Vipa Surojanametakul, Patcharee Tungtrakul, and Warunee Varanyanond 

Authors’ addresses: Yasuhiro Suzuki, Ushio Matsukura, Noriaki 

Aoki, and Hiroyuki Sato, National Institute of Crop Science, 

Kannondai, Tsukuba, Ibaraki 305-8518; Hiro-Yuki Hirano, 

Graduate School of Sciences, The University of Tokyo, Yayoi, 

Bunkyo-ku, Tokyo 113-8657; Yoshio Sano, Graduate School 

of Agriculture, Hokkaido University, Kita 9, Nishi 9, Kita-ku, 

Sapporo 060-8589; Kazuo Ise, Japan International Research 

Center for Agricultural Sciences, Ohwashi, Tsukuba, Ibaraki 

305-8686, Japan, e-mail: suzuyasu@affrc.go.jp. 

Pasting behavior was widely used to estimate the physical 

properties of various cereals or starch-based products, including 

rice products. The components of rice, such as amylose 

content, protein, and enzymes, especially amylase activity, 

were known as the factors affecting pasting behavior. There 

were several reports on the effect of amylase activity on rapid 

visco analyzer (RVA) profiles in nonglutinous and glutinous 

rice varieties. Especially, the inhibition of amylase activity 

using copper sulfate or mercury chloride in Japanese glutinous 

rice varieties increased peak viscosity and could express 

the variety difference of glutinous rice varieties (Kikuchi 1998). 

However, the effects of amylase inhibitor in Thai high-amy- 


lose varieties were opposite compared to those in glutinous 

rice varieties: lower peak viscosity and higher final viscosity 

(Tungtrakul and Yoshihashi 2001). Also, Matsukura et al (2004) 

reported that endogenous amylase activity among genetically 

isolated lines between nonglutinous and glutinous rice varieties 

did not show clear differences. Therefore, these changes in 

pasting behavior might be caused by an unknown factor or 

component in glutinous rice varieties, which are sensitive for 

amylase activity. The existence of water-soluble polysaccharide 

in rice has been reported (Horino et al 1988), and the 

amount of the polysaccharide was reported to be higher in glutinous 

rice varieties. 

Since polysaccharide contains a water-soluble fraction 

in rice, information about the water-soluble fraction in pasting 

behavior is important. However, the report showed only its 

existence, and there is no information on its properties or role 

in rice properties. Generally, it is difficult to digest or react 

with raw starch from rice flour, without gelatinization for the 

usual amylase. Thus, we expected that the water-soluble fraction, 

which contains water-soluble polysaccharide, might be a 

novel factor in rice pasting behavior. Therefore, this study examines 

such an approach to understand the role of water-soluble 

polysaccharide in rice flour in rice pasting behavior through 

amylase inhibition. 


Four rice varieties, low-amylose variety Koshihikari (japonica, 

Japan), high-amylose variety Chainat 1 (indica, Thailand), and 

two glutinous rice varieties (RD6, indica, Thailand, and 

Hakuchoumochi, japonica, Hokkaido, Japan), were ground 

using an Ultracentrifugal Mill (ZM100, Retch, Germany) attached 

with a 0.5-mm screen. Rice flour was stored in a refrigerator 

until used. 

The effect of the water-soluble fraction on the pasting 

behavior of rice flour was examined using an RVA (RVA-super 

3, Newport Scientific, New South Wales, Australia) with a 

temperature program of AACC method No. 61-2. Various treatments 

with the rice flour used the following procedures: 

G1—the flour suspensions were prepared by adding 

25 mL of distilled water or 10 mM CuSO 4 to rice 

flour (3 g) in an RVA canister. The mixture underwent 

RVA analysis. 

G2—the flour suspension prepared as G1 was held at 

37 ºC for 60 min with stirring at 160 rpm to allow 

time for extraction of the water-soluble fraction by 

the RVA, and then the mixture immediately underwent 

an RVA analysis. 

G3—the water-soluble fraction was removed by centrifugation 

after the G2 procedure and the precipitate 

was then added to 25 mL of distilled water. Then, the 

mixture underwent RVA analysis. 

G4—the water-soluble fraction was removed following 

the G3 procedure and the precipitate was resuspended 

in 25 mL of distilled water or 10 mM CuSO 4 

solution. Then, the mixture underwent RVA analysis. 

Rice flour suspensions were prepared by weighing rice 

flour (3 g) in an RVA canister and adding 25 mL of distilled 

water or 10 mM CuSO 4 as amylase inhibitor. Also, suspensions 

were prepared from heated rice flour (110 ºC, 1 h). The 

suspensions were held at 37 ºC for 30 min with stirring at 160 

rpm by the RVA to allow extraction of the water-soluble fraction 

and digestion with inherent rice amylase. A suspension 

was added with 1M CuSO 4 (250 mL) to inhibit further amylase 

digestion. Each water-soluble fraction was removed by 

centrifugation and supernatants were subjected to gel permeation 

chromatography (GPC). GPC was conducted with the 

Tosoh PS-8020 system equipped with the evaporative lightscattering 

detector (ELSD; Sedex-55, Sedex, France). The 

column used was a Tosoh G6000PWXL (250 × 6 mm I.D.). 

For GPC analysis, 20 mL of water-soluble fraction was injected 

and eluted with 1 mL min –1 of water. The ELSD conditions 

were 70 °C for inlet temperature, 2.3 bars for nitrogen 

flow, and gain 8. 

Milled broken rice was prepared by breaking milled rice 

by a hammer. Broken rice was heated (110 °C, 1 h) and 3 g of 

sample was suspended into water (25 mL). The suspension 

was stirred by a magnetic stirrer for 2 h at room temperature. 

The supernatant was separated by centrifuging (5 ºC, 8,000 

rpm, 30 min). GPC analysis was performed in the same manner. 

Ethanol precipitate of the water-soluble fraction (10 mg) 

and water (50 mL) were prepared in a sealed silver pan for 

differential scanning calorimetry (DSC) and kept in a refrigerator 

overnight for reaching equilibrium. The sealed sample 

was scanned by a differential scanning calorimeter (SII 

DSC5100, Japan) from 30 to 120 °C at 0.5 °C min –1 . Also, 

water-insoluble powder of rice was analyzed in the same manner. 

Ethanol precipitate of the water-soluble fraction was dissolved 

into 90% formic acid (1% w/v) in a screw-capped vial 

and heated at 100 °C for 3 h. Hydrolysates were dried up in 

vacuo and resuspended into 10% trifluoroacetic acid and heated 

at 100 °C for 3 h. Complete hydrolysates were dried up in 

vacuo, then reduced by NaBH 4 and acetylated by acetic 

anhydrate. The obtained alditol acetates underwent GC-MS 

analysis to obtain constituent sugars. 


Effect of the water-soluble fraction 

on pasting properties of rice flour 

Under conventional RVA conditions (G1), Hakuchoumochi had 

a lower peak viscosity. When the soaking time was prolonged 

(G2), peak viscosity was the lowest. The presence of copper 

sulfate, as an amylase inhibitor, dramatically increased the peak 

viscosity of glutinous varieties. However, Koshihikari was less 

affected by the addition of copper sulfate. The results suggested 

that an endogenous component, which was sensitive 

for inherent amylase activity in rice, affected rice pasting behavior. 

In glutinous varieties, the amylase-sensitive component 

might be present in a larger amount. Removal of the wa- 


Viscosity (cP) 

G1 with CuSO 4 

2,400 

1,800 


1,200 


G4 with water 

G1 with water 

600 

G3 with water 

0 

0 3 6 9 12 

Time (min) 

Fig. 1. RVA profile of rice flour (with and without water-soluble fraction) in water and in CuSO 4 solution. 

ter-soluble fraction (G3 and G4) showed no peak increasing 

even with an amylase inhibitor. Therefore, it is suspected that 

the water-soluble fraction contains the substance, which affects 

pasting properties and is sensitive for inherent amylase 

activity. 

Effect of amylase activities on the water-soluble 

fraction 

Figure 2 shows an example of a GPC profile of the watersoluble 

fraction. The water-soluble fraction contained an expected 

peak as water-soluble polysaccharide, when inherent 

amylase activities were inhibited by copper sulfate or heating. 

The peak disappeared when suspensions were allowed to mix 

for 30 min by RVA. However, when an amylase inhibitor such 

as copper sulfate was added after 10 min, the peak remained, 

which corresponds to the water-soluble polysaccharide. These 

results suggested that water-soluble polysaccharide was extremely 

sensitive to inherent rice amylase. Since the heating 

procedure inhibited inherent amylase activities, high-temperature 

drying might control the digestion of water-soluble 

polysaccharide through denaturation of inherent amylase. 

Elution of water-soluble polysaccharide 

from broken rice 

Even less physical force was applied on the rice kernel for the 

preparation of broken rice. Water-soluble polysaccharide was 

eluted in GPC analysis. Therefore, water-soluble polysaccharide 

was not categorized as “damaged starch,” which might be 

produced during longer grinding (Chen et al 1988, 1999). 

Obtained DSC thermograms of water-soluble polysaccharide, 

water-insoluble powder, and original rice flour of 

Hakuchoumochi indicated that the water-soluble fraction never 

gelatinized, and the thermal properties of rice may be represented 

by water-insoluble powder. Water-insoluble powder is 

considered to represent “wet-milled” rice flour. Therefore, the 

wet-milling process might be a process to remove water-soluble 

polysaccharide, which might affect gelatinization (Chen et al 

1988, 1999). 

Constituent analysis of water-soluble 

polysaccharides 

All samples showed constituents as glucose only. Thus, watersoluble 

polysaccharide in rice was expected as a homoglycan 

built from glucose. 

With these properties, water-soluble polysaccharide was 

easily digested by inherent rice amylase, then released oligosaccharides, 

which show lower viscosity in RVA and “stickiness” 


Water-soluble polysaccharide 

Oligosaccharide 

Copper sulfate 

Water 

V 0 

CuSO 4 , 10 min 

CuSO 4 

Heated 

0 2.5 5.0 7.5 10.0 12.5 

Min 

Fig. 2. GPC profile of water-soluble fraction on amylase digestion (Hakuchoumochi). 

on the surface of products. Therefore, the polysaccharide might 

work like “cement” in rice gel, which is considered as “concrete.” 

These results might be a clue to understanding pasting 

of rice, and this is needed to understand rice gel, not only starch, 

but also polysaccharides. However, further studies are needed 

to elucidate more detailed properties of polysaccharide. 

References 

Chen JJ, Lu S, Lii CY. 1988. Thermal characteristics and microstructure 

changes in waxy rice flours by different milling 

methods. Food Sci. 25:314-330. 

Chen JJ, Lu S, Lii CY. 1999. Effects of milling on the physicochemical 

characteristics of waxy rice in Taiwan. Cereal Chem. 

76:796-799. 

Horino T, Ondona A, Kusumoto K, Soegusa T, Mori Y. 1988. Novel 

water-soluble polysaccharide from rice. Research highlights 

on food research, NFRI. (In Japanese.) 

Kikuchi H. 1998. Studies on chemical components in rice kernel for 

breeding. Hokkaido Pref. Agric. Res Report. 1998:1-68. 

Matsukura U, Suzuki Y, Iwai Y, Monma M, Kaneko N. 2004. Comparison 

of α-amylase activity in rice grains of nonglutinous 

and glutinous varieties and effect of α-amylase activity on 

rice pasting property. Nippon Shokuhin Kagaku Kougaku 

Kaishi. 51(10):554-558. (In Japanese.) 

Tungtrakul P, Yoshihashi T. 2001. Elucidation of amylase effect on 

amylopectin chain distribution in various rice flour. Report 

submitted to JIRCAS. 8 p. 

Notes 

Authors’ addresses: Tadashi Yoshihashi and Eizo Tatsumi, Japan 

International Research Center for Agricultural Sciences 

(JIRCAS), Tsukuba 305-8686, Japan; Vipa Surojanametakul, 

Patcharee Tungtrakul, and Warunee Varanyanond, Institute of 

Food Research and Product Development (IFRPD), Kasetsart 

University, 50 Phahonyothin Road, Chatuchak, Bangkok 

10900, Thailand, e-mail: tadashi@jircas.affrc.go.jp. 



Session 8 featured six oral presentations on rice quality. Four 

related to the chemistry of rice, grain quality, and eating quality, 

and two related to human health. 

K.R. Bhattacharya from the Rice Research and Development 

Center, India, covered “The chemical basis of rice end-use 

quality.” He gave an overview and history of research on rice 

quality worldwide. Amylose was the single largest determinant of 

rice end-use quality, but was not sufficient. The texture of cooked 

rice and other rice end-use qualities is now attributed mainly to 

the relative abundance of long chains in its amylopectin. The 

author said that the true amylose content in rice starch was less 

than 10%. But B. Juliano said that amylose content was at least 

15%. So, we need further investigation on true amylose content. 

The length (degree of polymerization) of the long chains of amylopectin 

was also discussed. 

The second speaker, K. Cheaupun from the Pathumthani 

Rice Research Center, Thailand, discussed “Rice grain quality 

improvement in Thailand.” She explained the classification and 

standards of Thai rice. The standards for white rice are grades, 

grain classification (length), grain composition, defects, and milling 

degree. For Thai aromatic rice (Thai Hom Mali Rice), the standards 

are Thai Hom Mali Rice content, average grain length, 

amylose content, and alkali spreading value. She explained the 

recommended rice varieties and their main production areas. 

Physical grain quality is determined by grain dimension, chalkiness, 

and milling quality. Cooking quality is determined by amylose content, 

gelatinization temperature, and aroma. Finally, she explained 

inspections and postharvest management. 

M. Fitzgerald from IRRI explained “New tools for understanding 

starch synthesis.” She told of the starch synthesis that 

affects starch structure. She analyzed the molecular weight distribution 

of debranched starch and explained the synthesis of 

amylopectin chains (control of chains and random growth) and 

amylopectin architecture (amorphous and crystalline lamellae). 

In addition, she suggested that there was a third starch that 

differed from amylopectin in its size and its spatial separation in 

the amorphous growth rings. She proposed calling this starch 

“amyloglycan.” After her talk, T. Yoshihashi said that he had found 

a similar new starch component susceptible to amylase. 

Professor K. Hatae from Ochanomizu University, Japan, 

discussed “Textural differences between indica and japonica varieties 

in cooked rice.” She used a japonica variety, Nipponbare, 

an indica variety, Khao Dawk Mali, and an unknown high-amylose 

Thai rice that had a different texture of cooked rice. The 

results indicated that the higher the solid content and amylopec- 

tin of the extract from the surface layer of the cooked rice, the 

more sticky the cooked rice. Moreover, the distribution of amylose 

in the rice grain was also an important factor in the stickiness 

of cooked rice. 

These four presentations indicated that amylose is an important 

factor but not enough to explain the eating quality of rice. 

The structure of amylopectin (chain length distribution) is also an 

important factor. Additionally, the existence of a third component, 

amyloglycan, is suggested and it is interesting to consider 

how amyloglycan contributes to rice quality. 

The next two presentations related to human health. Professor 

J. Haas from Cornell University (USA) covered the “Biological 

efficacy of consuming rice biofortified with iron.” The authors 

conducted a double-blind, longitudinal (9 months), and intervention 

study with religious sisters in the Philippines. The provision 

of additional dietary iron from biofortified rice contributed 

to a 20% increase in daily iron intake for diets that were ironpoor. 

The additional iron from biofortified rice was insufficient, 

but it significantly reduced the number of women who had deficient 

iron intakes. There was a biologically significant effect of 

consuming rice biofortified with iron on increases in plasma ferritin 

and total body iron for women who were initially not anemic. 

This was the first demonstration that plant breeding for enhanced 

micronutrient quality had a positive biological impact on the individuals 

who consumed the rice. One question addressed whether 

milling affected iron content. The answer was that milling did 

reduce iron content because most of the iron was in the bran 

that was removed by milling and polishing. The high-iron variety 

seemed to retain more iron in the endosperm than other varieties. 

The final presentation was given by T. Sato from the National 

Agricultural Research Center for Kyushu Okinawa Region, 

Japan, on “Radical-scavenging activity of red and black rice.” 

The author told us that red rice contained the highest activity. 

Results from sequential extraction with six different polar solvents 

showed that the highly polar solvents, methanol, and deionized 

water extracts from all red, black, and white rice had 1,1- 

diphenyl-2-picrylhydrazyl radical- and tert-butyl hydroperoxyl radical-scavenging 

activity. In addition, the acetone extract from only 

red rice had the highest activity. The major component responsible 

for the radical scavenging in the acetone extract from red 

rice was proanthocyanidin. It would be interesting to study whether 

consuming red rice is good for human health. Now, the authors 

are investigating the physiological functions of red rice in vivo 

and elucidating the mechanisms at work in these functions. 


SESSION 9 

Developing new uses of rice 

CONVENER: K. Ohtsubo (NFRI)

Overview of rice and rice-based products 

Bienvenido O. Juliano 

Rice (Oryza sativa L.) and rice-based products derived from 

rice grain and rice flour include parboiled rice; quick-cooking 

rice and ready-to-eat convenience foods; rice flours; rice starch; 

cakes and puddings; baked breads and crackers; breakfast cereals 

and expanded rice products; extrusion-cooked and puffedrice 

snacks; noodles, paper, and pasta; baby/weaning foods; 

fermented foods and beverages; pet foods; and bran products 

(Juliano 2003). The rice ingredient is preferably aged to have 

predictable, stable functional properties, and freshly milled and 

well-milled for minimal initial fat rancidity and long shelf life. 

The rice ingredient in most of these products has the same 

apparent amylose content (AC) type as that of the preferred 

table rice in the country. Milled rice AC (on a dry-weight basis) 

is classified into waxy (0–2% dry basis) and nonwaxy: 

low (12–20%), intermediate (20–25%), and high (25–33%). 

Waxy and low-AC rice are used for alcoholic beverages because 

of the higher enzymic conversion of starch to glucose 

than high-AC rice, and for products requiring starch-gel stability 

of wet products. High-amylose rice is preferred for products 

requiring an intact cooked product, resistant to disintegration, 

although it may be harder in texture. 

Traditional rice products in developing countries are 

small-scale and poorly packaged, and have a short shelf life. 

Improvement of value-added traditional rice products, selected 

based on marketability and profitability, will create jobs in rural 

areas and make a country’s products compete globally as import 

substitutes or as exports to niche markets. 

Semiwet-milled rice flour 

The availability of a semiwet (semidry)-milled rice flour of 

specified AC (waxy, intermediate-AC, and high-AC) is crucial 

to the development of the rice products industry in the 

Philippines, since a dry-milled flour cannot be used in many 

baked and steamed products requiring a fine flour with little 

starch damage. By contrast, wet-milled and semiwet-milled 

rice flours can be used in all rice products and are whiter than 

dry-milled flour. However, wet-milling is tedious and a source 

of environmental pollution. Yeh (2004) considered semidry-/ 

semiwet-milled rice flour as that produced from steeped milled 

rice (30% water) with a roller mill, a stamp mill, and a pin 

mill, without water added during milling. Its starch is reported 

to be 40% gelatinized from heat generated during milling, 

which may affect its functional properties. It has lower contents 

of protein, lipids, ash, and reducing sugars than dry-milled 

flour. The development and production of semiwet rice flour 

of uniform properties (specified AC) to replace wet-milled flour 

should improve the efficiency of small and medium rice enterprises 

since the tedious wet-milling step can be dispensed with. 

Starch properties 

Thermal properties of starch of importance to product quality 

include gelatinization temperature (GT) and degree of gelatinization, 

amylose-lipid complex I (melting at temperatures 

100°C), and staledamylopectin 

melting (45–60 °C). Final GT represents the temperature 

when at least 90% of the starch granules have swollen 

in hot water with loss of crystallinity. Residual ungelatinized 

starch may affect product properties. Parboiling studies showed 

that low-GT samples tended to have amylose-lipid complex I, 

and intermediate-high-GT rice tended to have amylose-lipid 

complex II in parboiled rice (Biliaderis et al 1993). The presence 

of amylose-lipid complex II in pressure-parboiled rice 

(>100 °C) may explain the hard texture of its cooked rice since 

amylose-lipid complex II does not melt during rice cooking. 

Rice noodles and corn starch noodles have a similar texture 

when freshly cooked. However, the cooked rice noodles 

are more stable, even on reheating, and have a better shelf life, 

whereas the corn starch noodles become mushy on reheating 

and readily spoil. Is this because of the greater stability of starch 

gels of high-AC rice relative to nonwaxy corn Or is it because 

of the presence of more protein in milled rice (7%) that 

is reported to contribute to the stability of rice starch gel (C.R. 

Mitchell, California Natural Products, 1994, unpublished 

data) Corn starch has only 0.7–1.0% protein. 

The glycemic index is the relative increase in plasma 

glucose within 3 h after a fasted subject ingests 50 g of carbohydrate, 

with white bread or glucose taken as 100%. It is used 

as a guide for diets of noninsulin-dependent diabetes mellitus 

patients. The glycemic index tends to be higher for cooked 

waxy and low-AC rice than for cooked intermediate- and high- 

AC rice (Table 1) (Panlasigui 1989, Foster-Powell and Brand- 

Miller 1995, Juliano 2003). Cooking processes that allow 

amylose staling such as parboiling and noodle extrusion further 

reduce the glycemic index of intermediate- to high-AC 

rice. Processes that suppress amylose staling such as puffing 

and precooking increase the glycemic index of all rice. However, 

varietal differences in glycemic index are shown in 

molded puffed brown rice cake. 

Filipino consumers (laborers) prefer rice that is more 

filling (Unnevehr et al 1992), for example, high-AC PSBRc10. 

Satiety is defined as the state of being fed or gratified to capacity. 

Boiled rice and molded puffed brown rice cake with a 

higher glycemic index and lower AC tended to have less perceived 

satiety and greater subsequent food intake than high- 

AC rice (Holt and Brand Miller 1995, Tetens et al 2003). The 

satiety situation for waxy and intermediate-AC rice should be 

determined. The glycemic index of intermediate-AC rice tended 


Table 1. Effect of apparent amylose content (AC) and processing on the glycemic index in humans of rice 

and rice products (Panlasigui 1989, Foster-Powell and Brand-Miller 1995, Juliano 2003). 

Rice product 

Glycemic index a (% of white bread) 

Waxy/low AC (0–20%) Intermediate/high AC (20–33%) 

Milled rice 116 ± 23, 126 ± 16/126 ± 4 (2) 81 ± 3 (13)/83 ± 11 (2) 

Brown rice 112 ± 11/116 ± 10 (2) –/94 ±10, 83 ± 12 

Rice bran –/27 ± 4 –/– 

Parboiled milled rice 133 ± 12/124 ± 10 68 ± 4 (13)/69 ± 9, 66 ± 8 

Noodles/pasta –/131 ± 11 –/50 ± 10, 58 ± 12, 83 ± 5 

Precooked/instant rice –/128 ± 10 –/132 ± 10 

Rice bubbles/krispies –/126 ± 10 (2), 117 ± 5 –/– 

Puffed brown rice cake, molded –/117 ± 16, 128 ± 10 –/85 ± 7 

a 

Mean ± standard deviation. Numbers of samples in parentheses, if mean of more than one sample. The glycemic index based on 

glucose is 0.70 of the glycemic index based on white bread. 

Table 2. In vivo resistant starch in cooked rice and rice products of various amylose 

(AC) and gelatinization temperature (GT) types as measured in five growing rats (Eggum 

et al 1993) and in an ileostomate (Jenkins et al 1987). 

AC type Resistant starch a (%) 

Rice product (%) 

Low GT 

Interm.-high GT 

Boiled rice Waxy and low (0–20%) R 0.0 a 

Intermediate (20–25%) R 1.5 b, H 3.9 

High (25–33%) R 0.4 a R 3.4 c 

Very high (40%) 

R 1.7 b 

Parboiled (100 °C) High (25–33%) R 3.7 c R 1.7 b 

Very high (40%) 

R 1.2 b 

Pressure-parboiled Intermediate (20–25%) H 3.6 ± 0.9 (4) 

High (25–33%) R 4.9 d R 2.1 b 

Instant rice Intermediate (20–25%) H 5.1 

Rice noodles High (25–33%) R 2.7 bc 

Rice Chex Low (10–20%) H 2.9 

a R = rat, H = human. Values followed by the same letter are not significantly different at the 5% level by 

Duncan’s multiple range test. Mean ± standard deviation. 

to overlap with that of high-AC rice, and waxy and low-AC 

rice have a similar glycemic index (Table 1). 

The resistant starch (staled amylose) in boiled rice is 

low at 0–4% in milled rice and tends to increase with AC, GT, 

and cooking time (Table 2) (Jenkins et al 1987, Eggum et al 

1993). The total dietary fiber (TDF)/neutral detergent fiber is 

0.7–2.3% of milled rice and 2.9–4.0% of brown rice. Brown 

rice has a lower TDF content than whole-wheat grain, with a 

minimum TDF of 5.6%. The increase in TDF during cooking 

of nonwaxy rice is due to the formation of resistant starch. The 

IR36-based amylose extender mutant with 40% AC can still 

be cooked in boiling water and was not lower in glycemic index 

than the high-AC intermediate-GT parent. It contains both 

elongated (B type) and normal polyhedral starch granules. Its 

in vivo resistant starch in rats was surprisingly lower than that 

of high-AC intermediate-GT rice (Table 2). Its TDF was also 

lower (7.5%) than that reported for high-amylose corn mutants 

(30–40%) that are cooked only by autoclaving (>100 °C). 

Other nutritional concerns 

The effects of food processing on protein quality, mycotoxin 

level, acrylamide content, and level of antioxidants have to be 

monitored to ensure maintenance of the nutritive value of the 

rice product (Juliano 2003). The amino acids lysine and cysteine 

are sensitive to heat, particularly during toasting of expanded 

rice. Hydrogen sulfide is produced during extrusion 

cooking of rice. Alkaline treatment may induce the formation 

of lysinoalanine, a carcinogen. Mycotoxin is a problem with 

wet grain, including delayed drying after harvest and parboiled 

rice. It is concentrated in the bran layer. 

Acrylamide, a carcinogen, was analyzed in 2002 because 

of its presence in high levels in baked and fried cereals and 

potato, but absence in boiled cereals, including rice (NFA 

2002). Acrylamide is produced by the Maillard reaction of free 

asparagine with free sugars. Developments in research on high 

acrylamide levels in foods should be closely monitored as they 

relate to baked, fried, and toasted rice products. 

Session 9: Developing new uses of rice 269

The high level of antioxidants in rice bran—γ-oryzanols, 

tocopherols, and tocotrienols—as reflected in high (3–7%) 

unsaponifiable matter in bran oil contributes to the 

hypocholesterolemic effect and other health benefits of fullfat 

bran. γ-oryzanols are a group of ferulic acid esters of sterols 

and triterpenoid alcohols (4,4-dimethylsterols). Rice bran 

is also rich in the water-soluble antioxidant phytic acid. Significant 

amounts of antioxidants are lost during the various 

steps of rice-bran oil processing. With the popularity of virgin 

coconut oil in the region, the production of virgin rice-bran oil 

with superior levels of antioxidants as a functional food/ 

nutraceutical could be considered as a cottage industry near 

rice mills. However, oil extraction may be a problem since 

rice bran has a lower fat content (15%) than coconut meat 

(33%). Virgin rice-bran oil may be blended with virgin coconut 

oil with a high content of medium-chain fatty acids (8:0 to 

12:0) to combine the healing properties of coconut oil with the 

antioxidant properties of rice-bran oil. 

These are important considerations in the formulation 

of new rice products. 

References 

Biliaderis CG, Tonogai JR, Perez CM, Juliano BO. 1993. 

Thermophysical properties of milled rice starch as influenced 

by variety and parboiling method. Cereal Chem. 70:512-516. 

Eggum BO, Juliano BO, Perez CM, Acedo EF. 1993. The resistant 

starch, undigestible energy and undigestible protein contents 

of raw and cooked milled rice. J. Cereal Sci. 18:159-170. 

Foster-Powell K, Brand-Miller J. 1995. International tables of glycemic 

index. Am. J. Clin. Nutr. 62:869S-893S. 

Holt SHA, Brand Miller J. 1995. Increased insulin response to ingested 

foods is associated with lessened satiety. Appetite 24:43- 

54. 

Jenkins DJA, Cuff D, Wolever TMS, Knowland D, Thompson L, 

Cohen Z, Prokipchuk E. 1987. Digestibility of carbohydrate 

foods in an ileostomate: relationship to dietary fiber, in vitro 

digestibility, and glycemic response. Am. J. Gastroenterol. 

82:709-717. 

Juliano BO. 2003. Rice chemistry and quality. Muñoz, Nueva Ecija 

(Philippines): Philippine Rice Research Institute. 480 p. 

NFA (National Food Administration). 2002. Analytical methodology 

and survey results for acrylamide in foods. Uppsala (Sweden): 

NFA. 2 p. 

Panlasigui LN. 1989. Glycemic response to rice. PhD dissertation. 

Toronto, Ontario (Canada): Department of Nutritional Sciences, 

University of Toronto. 171 p. 

Tetens I, Kabir KA, Parvin S, Thilsted SH. 2003. Differential rates 

of energy release from rice and effects on satiety. In: Mew 

TM, Brar DS, Dawe D, Hardy B, editors. Rice science: innovations 

and impact for livelihood. Proceedings of the International 

Rice Research Conference, 16-19 September 2002, 

Beijing, China. Los Baños (Philippines): International Rice 


Unnevehr LJ, Duff B, Juliano BO, editors. 1992. Consumer demand 

for rice grain quality. Terminal Report IDRC Projects. National 

Grain Quality (Asia) and International Grain Quality 

Economics (Asia). Manila (Philippines): International Rice 

Research Institute, and Ottawa (Canada): International Development 

Research Centre. 248 p. 

Yeh A-I. 2004. Preparation and applications of rice flour. In: Champagne 

ET, editor. Rice chemistry and technology. 3rd ed. St. 

Paul, Minn. (USA): American Association of Cereal Chemists. 

p 495-539. 

Notes 

Author’s address: Philippine Rice Research Institute, Los Baños, 

College Laguna 4031, Philippines, e-mail: 

bjuliano@laguna.net. 

Sterilization effect of electrolyzed water on rice food 

Seiichiro Isobe, Chang-yong Lee, and Kyoichiro Yoshida 

Recently, a lot of instant foods have been made from rice. The 

main categories are aseptic packaged rice and rice cookies. In 

this food process, the control of microorganisms, especially 

heat-resistant spores from raw materials to products, is the most 

important point to confirm safety in product flow. If the heatresistant 

spores can be controlled by pretreatment before cooking, 

excess heat should be omitted to make long shelf-life rice 

products, thus improving the quality of the products. Several 

papers have demonstrated the sterilization effects of electrolyzed 

water on food ingredients (Koseki et al 2004a,b). 

In this paper, we try to confirm the sterilization effect on 

rice using electrolyzed water and check the quality of rice during 

pretreatment. 


Electrolyzed water is produced by electrolyzing tap water with 

the addition of a small quantity of NaCl. Acidic electrolyzed 

water (AcEW) created at the anode has been observed to have 

sterilization effects on microorganisms, and alkaline electrolyzed 

water (AlEW) created at the cathode has been observed 

to have a rinsing effect on organic compounds. In Japan, AcEW 

already was approved as an indirect food additive in 2002. 

The principle of electrolyzed water is shown in Figure 1. 

We used a flow-type electrolyzed water producer (Rox- 

20TA: Hoshizaki Electric Co., Japan). This apparatus generates 

electrolyzed water by the electrolysis of a dilute (0.1%) 

saline solution in an electrolytic cell separated into an anode 

and cathode region with a diaphragm. The current passing 


Acid-electrolyzed 

water for 

sterilization 

Alkalineelectrolyzed 

water 

Survival number (Log cfu g –1 ) 

5 

HCl 

HClO 

NaOH 

+ 

Cl 2 

H 2 

– 

H + OH – 

Cl – Na + 

~Chemical reaction~ 

: 

2OH – ®H 2 

O+1/2O 2 

+ 

2e – 

2Cl – ®Cl 2 

+2e – 

Cl 2 

+H 2 

®HClO+HCl 

(low pH condition) 

HClO®H + +ClO – 

(high pH condition) 

: 

2H + +2e – ®H 2 

4 

3 

2 

1 

NaCl 

Tap water 

with NaCl 

NaCl 

Separation 

membrane 

Fig. 1. Principle of acidic electrolyzed water. Left: process 

flow of apparatus, right: chemical reaction during electrolyzing 

process. 

0 

A B C D 

Fig. 2. Sterilization effects of combination of 

electrolyzed water on rice preparation. (A) Control 

rice, (B) washing with AlEW (5 min), (C) 

soaking with AcEW (30 min) after B treatment, 

(D) soaking with D/W (30 min) after C treatment. 

through the electrolysis apparatus and voltage between the electrodes 

were set at 14A and 18V, respectively. AcEW was prepared 

within the anode region of the electrolytic cell, and AlEW 

was prepared within the cathode region. The physicochemical 

properties of electrolyzed water are as follows: 

AcEW: pH 2.7, oxidation reduction potential (ORP) 

1,481 mV, available chlorine concentration 51.5 ppm. 

AlEW: pH 11.6, ORP –576 mV. 

We used distilled water as a control processing solution. 

Polished rice (Kinuhikari) was purchased and used as a 

sample. Prepared Bacillus subtillis PCI219 was used as a heatresistant 

spore sample. 

Some 1 mL of spore solution (10 9 cfu mL) was added to 

9 mL of AcEW and mixed. The sample was collected from the 

mixed solution every 1 min and the survival number counted 

in each sample after cultivation. 

Some 100 g of polished rice were prepared for each 

preparation test. We devised a preparation process with 3 stages 

(washing: 5 min, soaking 1 ½ min, soaking 2 ½ min). We used 

AcEW, AlEW, and distilled water, respectively, as each stage 

solution. pH and color of rice were measured by a conventional 

pH meter and color meter. 

To confirm the microorganism control effect, we added 

the B. subtillis spore into the rice and we counted the survival 

number in each sample of the preparation test. 


We examined several tests to prove the influence on the microorganism 

and the change in the raw rice. 

AcEW and AlEW changed the color and pH of rice rapidly. 

But, when the rice was treated with a combination of electrolyzed 

water (washing with AlEW, soaking with AcEW, and 

soaking with distilled water), pH and color of the treated rice 

were no different from conventional washing and soaking with 

distilled water. 

AcEW showed a strong effect on microorganism control, 

even on heat-resistant spores in vitro. However, in the 

rice preparation test, AcEW could not kill B. subtillis spores 

on the rice completely. But, after washing by AlEW, AcEW 

could kill completely. So, finally, we set the optimum method 

of combination of electrolyzed water as follows. To remove 

rice bran powder and other foreign materials that coat the surface 

of the raw rice, at first we washed the rice with AlEW for 

5 min and soaked it with AcEW to sterilize microorganisms 

for 30 min. Then, to remove the odor of chlorine and to revert 

the pH level, we resoaked the rice with distilled water for 30 

min. Figure 2 shows the effects against microorganisms using 

the combination process of each electrolyzed water. 

In this experiment, we could confirm the effect of the 

combination of electrolyzed water on microorganism control. 

This effect can be considered the result of available chlorine 

concentration and the rapid change in pH and it raises the sen- 


sibility on the microorganism. From this experiment we can 

see the importance of the contact condition of the microorganism 

with the electrolyzed water, which increased the effect of 

sterilization. 

We concluded that the combination of electrolyzed water 

for rice preparation before cooking (washing and soaking) 

is efficient in reducing heat-resistant microorganisms; therefore, 

this process will be able to help make rice products with 

a long shelf-life without excess heat treatment to keep the quality 

of these products. 

In this experiment, we did not evaluate the quality (texture, 

flavor, chemical components, and so on) of the cooked 

rice. We must consider this quality before introducing this pretreatment. 

References 

Koseki S, Yoshida K, Isobe S, Itoh K. 2004a. Efficacy of acidic 

electrolyzed water for microbial decontamination of cucumbers 

and strawberries. J. Food Prot. 67(6):1247-1251. 

Koseki S, Yoshida K, Kamitani Y, Isobe S, Itoh K. 2004b. Effect of 

mild heat pre-treatment with alkaline electrolyzed water on 

the efficacy of acidic electrolyzed water against Escherichia 

coli O157:H7 and Salmonella on lettuce. Food Microbiol. 

21:559-566. 

Notes 

Authors’addresses: Seiichiro Isobe, National Food Research Institute, 

2-1-12 Kannodai, Tsukuba, Ibaraki 305-8642, Japan, e- 

mail: seiichi@nfri.affrc.go.jp; Chang-yong Lee, Cheiljedang 

Corporation Korea; Kyoichiro Yoshida, Hoshizaki Electric 

Co., Japan. 

Current status of varietal improvement 

and use of specialty rice in Korea 

Hae Chune Choi 

Korea has maintained self-sufficiency in rice production since 

1975, except during severely cold years. However, self-sufficiency 

in total cereal food supply, including other cereals, is 

still limited to about 54% and to about 30% for total consumption, 

including livestock feeds. Rice products, the staple food 

in Korea, cover 23% of farmers’ income and 41% of total agricultural 

resources, and provide 35% of caloric intake and 

21% of protein intake per capita at present. 

Historically, the continuous improvement of Tongil-type 

high-yielding rice cultivars greatly contributed to self-sufficiency 

in rice production through the so-called Green Revolution 

during the 1970-1980s. Moreover, the development of 

high-yielding and high-quality japonica rice cultivars during 

the 1980-1990s played a major role both in ensuring self-sufficiency 

of rice production and in enhancing the competitiveness 

of rice products under free trade. We believe that we have 

the most advanced techniques in rice breeding and maintain 

the highest level of rice yield potential per hectare while producing 

the best-quality rice in the world. However, Korea is 

still lagging behind in the international competitiveness of rice 

goods primarily because the size of rice farm land is so small 

and the production cost is still very high (Choi 2001). It should 

also be noted that per capita rice consumption in Korea has 

decreased dramatically from 120 kg in 1990 to 83 kg in 2003. 

The development of high-quality and specialty rice varieties 

should continue to be an important element in rice research. 

By developing a variety of rice products other than ordinary 

cooked rice, we can increase rice consumption and international 

competitiveness. 

Achievement of varietal improvement in specialty rice 

The target of rice breeding in Korea is to develop high-quality 

and high-yielding rice cultivars suitable for labor-saving, lowcost, 

and safe grain production under different environmental 

conditions. The main direction of the country’s rice breeding 

program was on how to increase yield potential and safe grain 

production and how to improve marketability and palatability. 

Also, efforts to develop various specialty rice varieties started 

in the mid-1980s to support an increase in rice consumption 

and use through food processing (Choi 2001). 

Since 1990, 23 specialty rice cultivars, including glutinous 

rice, have been developed. The first specialty rice cultivar 

was a large-kernel variety, Daeribbyeo 1, developed in 

1993. It has about 1.7 times heavier 1,000-grain weight of 34.8 

g than ordinary ones. It showed a good usability for popping 

and brewing. In the same year, high-yielding and semidwarf 

indica scented rice, Hyangmibyeo 1, was developed. It has a 

considerably strong aroma and a good flavor in cooked rice or 

processed rice foods such as traditional rice cakes and saccharified 

rice beverage. A high-quality japonica scented rice, 

Hyangnambyeo, was developed in 1995. Another high-yielding 

Tongil-type scented rice, Hyangmibyeo 2, was developed 

in 1996. Since then, two scented glutinous rice varieties, 

Aranghyangchalbyeo and Seolhyangbyeo, a premium scented 

rice, Mihyangbyeo, and a blackish purple-scented rice, 

Heughyangbyeo, were developed successively (Table 1). 

In 1994, a medium-sized, chalky-kernel rice variety having 

suitability for brewing, Yangjobyeo, was developed. An 

opaque nonglutinous rice, Seolgaeng, a 9% low-amylose dull 


Table 1. Development of specialty rice cultivars suitable for food processing. a 

Bred Culm Milled 1,000- Chalkiness Color of 

Cultivar in Heading length rice yield kernel L/W (WC/WB) seed Aroma Amylose 

year date (cm) (t ha –1 ) weight ratio (0–9) coat (0–9) (%) 

(g) 

Daeribbyeo 1 1993 15 Aug 88 4.45 34.8 1.94 1/0 YW 0 19.5 

Hyangmibyeo 1 1993 15 Aug 72 4.93 20.6 2.46 1/1 YW 5 18.3 

Hyangnambyeo 1995 11 Aug 82 5.03 21.3 1.81 0/0 YW 3 17.7 

Hangmibyeo 2 1996 4 Aug 77 6.14 22.8 2.44 1/2 YW 3 19.0 

Yangjobyeo 1994 14 Aug 71 5.11 25.4 1.78 7/0 YW 0 20.2 

Aranghyangchalbyeo 1997 13 Aug 88 5.37 20.5 1.90 – YW 3 0.0 

Heugjinjubyeo 1997 25 July 80 4.05 17.0 2.22 – BP 1 15.1 

Heugnambyeo 1997 13 Aug 73 4.97 23.5 2.14 – BP 1 16.7 

Mihyangbyeo 1998 16 Aug 79 5.57 22.6 1.86 0/0 YW 3 19.0 

Seolhyangchalbyeo 1999 8 Aug 89 5.23 24.2 2.10 – YW 3 0.0 

Goamibyeo 2000 18 Aug 85 5.38 20.8 1.77 0/2 YW 0 26.7 

Heughyangbyeo 2000 23 Aug 77 5.28 23.1 1.96 – BP 3 18.0 

Jeogjinjubyeo 2000 26 July 81 5.54 22.1 1.79 – RB 0 18.3 

Baegjinju 2001 21 Aug 77 5.18 19.2 1.62 Dull YW 0 9.1 

Seolgaeng 2001 19 Aug 83 5.27 20.3 1.63 Opaque YW 0 19.3 

Yeongan 2001 13 Aug 83 5.45 22.4 1.79 0/1 YW HL 18.9 

Goami 2 2002 22 Aug 75 4.24 14.6 1.89 4/4 YW HF 28.1 

Manmi 2002 17 Aug 86 4.46 19.9 1.84 0/0 YW 0 12.9 

Heugkwang 2003 14 Aug 89 5.05 14.8 1.73 – BP 0 18.0 

a 

L/W ratio = length/width ratio of brown rice, YW = yellowish white, BP = blackish purple, RB = reddish brown, HL = high lysine, HF = high fiber. 

Source: Choi (2004). 

mutant rice, Baegjinju, and a high-fiber and high-amylose rice, 

Goami 2, were developed through mutation breeding by MNU 

(N-methyl-N-nitrosourea) treatment on a premium-quality 

japonica rice variety, Ilpumbyeo. Another semiglutinous rice, 

Manmi, which has 12.9% amylose content, was developed by 

a conventional breeding method using a dull mutant as a parent. 

Some colored rice varieties having reddish brown or 

blackish purple pericarp were also developed (Table 1). In 

2001, a premium-quality rice variety with high lysine content, 

Yeongan, was developed. Some specialty rice varieties, such 

as blackish purple or reddish brown glutinous rice and giantembryo 

glutinous rice, were bred by mutation or from weedy 

rice collections (Choi 2004). 

The cultivation area of specialty rice, including glutinous 

rice, is only about 1.5% of the total rice cultivation area 

since its usability and processing have not been fully developed. 

Use of specialty rice 

The large-kernel and low-amylose rice varieties revealed both 

improved popping and brewing characteristics. Their softer 

gel consistency and lower amylose content were closely related 

to the higher ratio of popped rice grains and also to the 

lower bulk density of popped grain in either brown or milled 

rice (Choi 2002). 

Of the high-amylose and high-protein rice varieties, some 

showed a desirable quality of processed rice noodles or rice 

bread. The acceptability for noodle formation and eating qual- 

ity of cooked noodles might be due to some factors independent 

of each other in rice. The potassium and magnesium contents 

of milled rice were negatively related to the gross score 

of noodle making when mixed with wheat flour half and half 

(Choi 2002). 

Batters for making brown rice bread had more volume 

expansion, better loaf formation, and more springiness in rice 

bread. The higher the protein content in rice, the more moist 

milled rice bread resulted. The springiness of rice bread was 

significantly correlated with high amylose content and hard 

gel consistency. A considerably different tendency of acceptability 

for rice bread baking among tested rice varieties was 

revealed between brown and polished rice (Kang et al 1997). 

There was a large varietal difference in suitability for koji 

(malted rice) fermentation by either Aspergillus oryzae or 

Monascus anka. The nonglutinous rice having opaque endosperm, 

Seolgaeng, showed better rooting density of mycelia 

and higher saccharogenic power in A. oryzae-fermented 

rice, and higher pigment concentration in Monascus anka-fermented 

rice when compared with dull or ordinary nonglutinous 

rice. The chalky and large-grain rice showed better suitability 

for fermentation and brewing than less chalky and smaller rice 

(Choi 2002). 

Colored rice can be used as natural functional pigments. 

The anthocyanin pigment of blackish purple rice showed significantly 

high antioxidant and antimutagenic reaction (Kang 

2002). The pigments of colored rice are very useful for traditional 

colored rice cakes or for brewing colored rice wines. 

The extracted anthocyanin pigment can be used for diverse 

processed rice foods or for beauty products (Choi 2002). 


The mutant rice, Goami 2, which has extreme contrasts 

in cooking quality and physicochemical and ultrastructural 

properties from those of its original variety, Ilpumbyeo (Kang 

et al 2003, Kim et al 2004), revealed a considerable “health 

effect” by reducing blood glucose and insulin levels in both 

normal and obese or diabetic participants after feeding the rice 

orally (Lee and Shin 2002). In addition, Goami 2 also demonstrated 

an ability to lower the levels of serum triglycerides and 

C-peptides in obese persons. However, the diabetic participants 

showed no significant reduction in plasma glucose and 

BMI (body mass index) after 4 weeks of eating the fiber-rich 

Goami 2 rice diet (Lee and Shin 2002, RDA 2004). Goami 2 

has low-digestible and high-amylose starch, and has approximately 

three times more crude fiber and three-plus times higher 

arabinose and xylose contents than the original variety. It has 

a significantly low proportion of short-chain frequency in amylopectin 

structure compared with its original variety (Kang et 

al 2003, Cheong et al 2004). It also showed significantly low 

activity in branching enzymes, but remarkably high activity in 

ADP-glucose pyrophosphorylase and total starch synthase as 

well as UDP-glucose pyrophosphorylase in rice grain during 

the rapid filling period (Cheong et al 2004). On the basis of 

these data, it is proposed that Goami 2 should be an excellent 

candidate for processing a variety of low-digestible, low-calorie 

“healthy” food products other than ordinary cooked rice. 

Regarding the soaking or germination treatment of brown 

rice, the production of γ-aminobutyric acid (GABA), a chemical 

that has been shown to contribute diverse health-related 

functions to the human body, increased two to three times more 

than that in untreated brown rice. Especially, brown rice with 

giant-embryo generally exhibited 6–8 times increases in GABA 

concentration compared with nonsoaked giant-embryo brown 

rice when soaked for 4 hours. This soaked and/or germinated 

giant-embryo brown rice may be useful for producing various 

foods processed with high “healthy” elements (Nemoto 2001). 

Glutinous rice varieties revealed significantly larger 

variations in viscogram characteristics, amylopectin structure, 

and physical as well as retrogradation properties of cooked 

rice. They were classified into nine different varietal groups 

based on 11 physicochemical and structural characteristics of 

rice endosperm (Fig. 1). Significant positive or negative correlations 

were noted in water absorption rates of rice grain, 

physical properties of cooked rice, and viscogram characteristics 

of rice flour. Especially in japonica glutinous rice, the 

breakdown and setback viscosities of rice flour were closely 

related to alkali digestibility of milled rice and the stickiness 

of cooked rice. The frequency ratio of short glucose chains 

(A-chains) to intermediate chains (B-chains), the ratio of B- 

chain to long-chain (C-chain) fractions, and the relative frequency 

of A- or B-chain fractions in debranched rice starch 

were closely related to the breakdown and setback viscosities 

of rice flour. There were some close associations between some 

of the grain properties in glutinous rice and the respective suitability 

for processing various kinds of traditional rice foods 

(Choi et al 1999). 

1 

0 

–1 

I 

III 

II 

IV 

Z 2 

VI 

V 

–1 0 1 2 

Fig. 1. Classification of glutinous rice varieties using principal component 

analysis, based on physicochemical and structural characteristics 

of rice endosperm (11 quality properties). = indica, 

= japonica, = javanica. I–IX = varietal groups, Z 1 and Z 2 are 

1st and 2nd principal components, respectively, derived from 11 

grain quality properties by principal component analysis. (Based 

on Choi et al 1999.) 

Prospects for specialty rice breeding 

Unstable environmental conditions caused by the global greenhouse 

effect and unexpected local meteorological disasters may 

be continuous problems in rice production. The deterioration 

of paddy soil is also expected with the increased air and water 

pollution, and the shortage of irrigation water may become 

more serious in the future. It is obvious that the greenhouse 

effect will result in increased disease incidence and insect pests. 

In addition, the pressure of the rice market opening from some 

export countries may become increasingly serious, resulting 

in an unfavorable situation for maintaining self-sufficiency in 

domestic rice supply and demand. 

It is urged, therefore, that rice breeders must come to 

terms with the outlook of the rice industry and be ready for 

unexpected changes in environmental and socioeconomic situations 

as well as the trend in world trade. Fortunately, an effective 

selection and introduction of useful agronomic target 

genes has been made possible recently through the use of DNA 

markers linked with target genes and genetic engineering, following 

the biochemical pathway and mechanism of the gene 

expression being elucidated. Moreover, the possibility of desirable 

recombination and the enhancement of useful gene 

expression may be gradually realized through the identification 

of target gene loci, gene cloning, and transformation. The 

genetic variation of rice breeding materials may become further 

diversified through the introduction of various useful alien 

genes by using biotechnological procedures in the future, and 

these may be further helpful for increasing the stability of rice 

production and enhancing the healthy competitiveness of rice 

products. 

The yield potential of milled rice is expected to increase 

to 6.5 t ha –1 in specialty japonica rice and to 10.0 t ha –1 in 

super-yielding specialty rice in the next ten years. Breeding 

efforts for diversification and physicochemical characteristics 

VII 

VIII 

IX 

Z 1 


in specialty rice grain will be continuously intensified to enhance 

the usability for various food-processing and health-related 

purposes. Specifically, breeding efforts to enhance healthrelated 

functions will focus on developing specialty rice suitable 

for particular needs, such as low-allergen rice helpful for 

atopic dermatitis patients, low-protein rice beneficial to nephritis 

patients, high-lysine or high-sulfur-containing aminoacid 

rice, high procarotenoid rice, giant-embryo rice, etc., for 

other health-related subjects (Choi 2004). 

Multiple resistance to major pests and environmental 

stresses suitable for each relevant region must be gradually 

strengthened in specialty rice by introducing new gene sources 

from wild species and through in situ selection breeding adaptable 

to global climatic change. 

References 

Cheong MY, Park YM, Choi HC, Park YI, Kim SI, Chung YH, Kim 

SH. 2004. A rice mutant with altered polysaccharides in grain. 

(Unpublished.) 

Choi HC, Hong HC, Kim YG, Nahm BH. 1999. Varietal variations 

in physicochemcial characteristics and amylopectin structure 

of grain in glutinous rice. Korean J. Crop Sci. 44(3):207-213. 

Choi HC. 2001. Achievements and advanced technology of rice production 

in Korea. In: Rice culture in Asia: KCD and ICID, 

Seoul. p 55-80. 

Choi HC.2002. Current status and perspectives in varietal improvement 

of rice cultivars for high-quality and value-added products. 

Korean J. Crop Sci. 47(S):15-32. 

Choi HC.2004. Principal characteristics of specialty rice cultivars 

and regarding cultivation management. Lecture note for advanced 

course of elite farmers. 17 p. 

Kang HJ, Hwang IK, Kim KS, Choi HC. 2003. Comparative structure 

and physicochemical properties of Ilpumbyeo, a highquality 

japonica rice, and its mutant Suweon 464. J. Agric. 

Food Chem. 51:6598-6603. 

Kang MY, Choi YH, Choi HC. 1997. Comparison of some characteristics 

relevant to rice bread processing between brown and 

milled rice. Korean J. Soc. Food Sci. 13(1):64-69. 

Kang MY. 2002. Functional components and efficacy of rice. In: 

Illumination of hygienic function in rice and strategy for functional 

rice foods. Proceedings of 2002 Spring Symposium of 

the Korea Industrial Food Engineering Society. p 35-50. 

Kim KS, Kang HJ, Hwang IK, Hwang HG, Kim TY, Choi HC. 2004. 

Comparative ultrastructure of Ilpumbyeo, a high-quality 

japonica rice, and its mutant Suweon 464: scanning and transmission 

electron microscopy studies. J. Agric. Food Chem. 

52:3876-3883. 

Lee C, Shin JS. 2002. Effects of different fiber content of rice on 

blood glucose and triglyceride levels in normal subjects. J. 

Korean Soc. Food Sci. Nutr. 31(6):1048-1051. 

Nemoto H. 2001. Current status and perspectives of high value-added 

and specialty rice production in Japan. Book Series No. 11 of 

the Korea Rice Technical Working Group. p 97-104. 

RDA (Rural Development Administration). 2004. Effect of newly 

developed naturally fiber-rich rice (Goami 2) on body weight 

and lipid metabolism in Korea. Final report. 41 p. 

Notes 

Author’s address: National Institute of Crop Science, Rural Development 

Administration, 209 Seodun, Suwon 441-857, Republic 

of Korea, e-mail: hcchoi@rda.go.kr. 

Processed novel foodstuffs from pregerminated brown rice 

by a twin-screw extruder 

Ken’ichi Ohtsubo, Tomoya Okunishi, and Keitaro Suzuki 

Rice is one of the most important cereals in the world in addition 

to wheat and maize. Rice production is more than 500 

million tons (as paddy per year) in the world and about 90% of 

it is produced and consumed as a major staple food in the 

densely populated Asian countries. 

Processing of rice grains in Japan 

Rice is processed and used as various kinds of foodstuffs besides 

direct food use, such as parboiled rice, fermented rice 

wine, rice noodles, rice crackers, rice cakes, rice snacks, rice 

flour, and other fermented rice products. 

Traditional rice-based convenience foods in Japan 

One of the most famous traditional rice-based convenience 

foods in Japan is sushi. Sushi originated as a preservative food 

of fish using natural fermentation (narezushi). In this case, 

cooked rice was used as a substrate with salt for Lactobacillus 

bacteria. Nowadays, nigirizushi was invented as an easily prepared 

rice dish by pressing cooked rice balls manually and 

topping them with sliced raw fish in Tokyo in the 1800s. 

(Cooked rice becomes tastier by adding salt, sugar, and vinegar 

after cooking.) Soy sauce and grated horseradish (wasabi) 

are necessary to enjoy the original taste of sushi. 

Onigiri or omusubi is another traditional rice-based convenience 

food in Japan. These cooked rice balls wrapped with 

bamboo leaves were used as a box lunch for many years. 

Breakfast cereals from rice 

Breakfast rice cereals are made from rice grains, milled rice 

flour, or cooked rice dough. These rice materials are precooked, 

dried, flaked, then expanded or puffed and toasted. Examples 


are puffed rice, rice flakes, and shredded rice cereals. Various 

kinds of breakfast cereals were developed and consumed in 

the United States, China, Korea, Thailand, Vietnam, and many 

other Pacific countries. 

Retort rice 

Retort rice was developed in the early 1970s in Japan. Rice 

and water are packed in a laminated plastic container and pasteurized 

at 120 ºC. Consumers soak this in hot water for 15 

min or heat it in a microwave oven for a couple of minutes. Its 

shelf-life is more than half a year without refrigeration. The 

price is reasonable. Its problems are off-flavor by the excess 

heating and the texture of cooked rice grains. Retort rice represented 

22,000 tons in Japan in 1996. 

Canned rice 

Canned rice has a long history of more than 50 years. Milled 

rice and water are placed in tin cans, steamed for 30 min, and 

sealed and sterilized in a retort at 112 ºC for 80 min. Consumers 

heat it in hot water for 15 min before eating. Canned rice 

can be stored for several years under natural conditions. 

Pregelatinized rice 

Cooked and dried rice or pregelatinized rice is prepared by 

usual cooking followed by abrupt drying. As its moisture content 

is very low, it can be preserved for several years under 

natural conditions. It is easy to cook as its starch is 

pregelatinized and prevented from retrogradation because of 

its low moisture content. 

Instant rice, such as Cup Rice, is a kind of high-quality 

pregelatinized rice. Consumers can eat it by only adding hot 

water and keeping it warm for several minutes. 

Recently, pregelatinized and packaged rice without excess 

drying has been developed in Japan (hayadakimai or 

“quick-cooking rice”). Consumers only have to add the tasty 

cooking soup or water and cook for about 15 minutes. They 

can omit the time-consuming washing of rice, soaking it, and 

keeping it warm after cooking. Compared with “instant rice,” 

quick-cooking rice is improved in terms of taste and texture. 

Although its moisture content is more than 35%, it can be stored 

for several months without refrigeration by pasteurization or 

by the oxygen absorber. Quality evaluation of these various 

kinds of quick-cooking rice, sensory test, gelatinization properties 

test, and physical properties measurement were carried 

out in our laboratory as a collaborative research activity with 

Kiewpie Co. Ltd. 

Frozen cooked rice 

The market for frozen cooked rice in Japan expanded to 

138,000 tons in 1996. Frozen cooked rice is convenient to prepare 

with a microwave oven at home and its high quality is 

preserved in a freezer for a long time. Its price is higher than 

that of other processed rice, but frozen roasted cooked rice 

balls or frozen pilaf are very popular among Japanese consumers. 

In cooperative research work with a manufacturer of 

freezer systems, Mayekawa Co. Ltd., we developed a new 

freezing method, “medium-rate freezing,” for frozen cooked 

rice balls. Frozen cooked rice balls prepared at a medium rate 

had better palatability than those frozen at a slow rate or quick 

rate. In addition, their physical properties, hardness, and stickiness 

were maintained better than those of others. 

Aseptic cooked rice 

Aseptic cooked rice was developed within one decade in Japan. 

Rice grains are washed well to remove bacteria, followed 

by cooking and packaging. Because contamination by microorganisms 

is very low, this product can be stored for half a 

year under natural conditions. As the rice is not heated excessively 

as is retort-pouched rice, its eating quality is very good. 

Cooked rice distributed at low temperature 

Cooked rice and packaged rice in a plastic container are sometimes 

shipped and distributed under low temperature (lower 

than 18 ºC). This is recommended for sanitation, but the problem 

is the retrogradation of rice starch. Its market is expanding 

year by year. 

“New characteristic rice” research project 

The Ministry of Agriculture, Forestry, and Fisheries supported 

the research project “New characteristic rice” from 1990 to 

1995. To enhance rice consumption, Japanese breeders tried 

to develop various kinds of new rice varieties, such as big 

grains, long grains, aromatic rice, pigmented rice, low-amylose 

rice, etc. In addition to national institutes, many universities, 

prefectural research institutes, and private companies took 

part in the project. 

Using pressurization in rice processing 

High-pressure treatment was applied to the manufacturing of 

rice crackers and rice cakes. It is well known that rice crackers 

are made in many processing steps. High-pressure treatment 

made it possible to simplify the manufacturing procedure. 

According to Dr. A. Yamasaki, starch of nonglutinous rice was 

gelatinized by high-pressure treatment of 700 MPa at 35 ºC. 

He also succeeded in developing low-allergenic rice, which is 

useful for preparing low-allergenic bread and low-allergenic 

cooked rice. These rice products are promising because many 

people suffer from allergy. 

Pregerminated brown rice and extrusion cooking 

Pregerminated brown rice has become popular in Japan. It is 

believed that pregerminated brown rice is good for health. It 

was reported that the content of gamma-aminobutyric acid 

(GABA) in pregerminated brown rice is higher than in ordinary 

milled rice or ungerminated brown rice. The development 

of a novel foodstuff from pregerminated brown rice by a 

twin-screw extruder was investigated. The characteristics of 

the components of the materials and the products are as follows. 


1. Pregerminated brown rice was prepared by soaking 

in water for 72 h at 30 ºC followed by drying to 13% 

to 15% moisture content at 15 ºC in a low-humidity 

artificial weather-control room. Total dietary fiber, 

total ferulic acid, and GABA contents of the 

pregerminated brown rice were higher than those of 

ordinary brown rice or polished rice. 

2. The pregerminated brown rice was processed with a 

twin-screw extruder. The puffed pregerminated brown 

rice contained more oryzanol, inositol, total ferulic 

acid, and total dietary fibers than the unpuffed polished 

rice. And the product prepared by the co-extrusion 

of pregerminated brown rice (90%) and beer 

yeast (10%) contained more free amino acids, such 

as GABA, glycine, alanine, aspartic acid, and glutamic 

acid, than polished rice, brown rice, and puffed 

pregerminated brown rice. 

3. Extrusion cooking was shown to sterilize the germinated 

brown rice by the incubation test, which would 

lead to the development of consumer-oriented rice 

products following food safety procedures. 

4. Wheat bread prepared with 30% puffed pregerminated 

brown rice contained more GABA and free sugars, 

such as maltose, than ordinary wheat bread. The 

extrudate bread was shown to be sweeter (P

Ohtsubo K, Nakamura S, Imamura T. 2002. Development of primer 

sets for identification of a rice cultivar, Koshihikari, by PCR. 

Nippon Nogeikagaku Kaishi 76:388-397. (In Japanese with 

English summary.) 

Notes 

Authors’ address: National Food Research Institute, Tsukuba, Ibaraki 

305-8642, Japan, e-mail: kenohtsu@nfri.affrc.go.jp. 

High-pressure food processing of rice and starch foods 

Rikimaru Hayashi 

The use of pressure (P) in addition to temperature (T) (Hayashi 

1987, Hayashi et al 1987) has been accepted in the field of 

food science and technology over the past 15 years, and research 

and development are now under way in the food industry, 

universities, and government institutes (Hayashi 2002). 

Several commercial products are now on the market that are 

prepared using high-pressure techniques. 

This paper describes the principle of high-pressure treatment 

and the effects of high pressure on foods, with emphasis 

on the high-pressure effects on starches. Finally, recent successes 

of the Echigo Seika Company in the application of highpressure 

techniques to rice and rice products are discussed. 

Principle and method 

High pressure means high hydrostatic pressure generated by 

the compression of water. A pressure of 100 MPa or higher 

(usually lower than 1,000 MPa) is used under temperatures 

below 100 °C. A unit of pressure is expressed in kg cm –2 , bar, 

or pounds in –2 , but now the international Pascal unit is commonly 

used. Therefore, 1,000 bar or 1,000 kg cm –2 almost 

equals 100 MPa. 

Food, which is contained in a plastic bag and sealed by 

carefully removing air, is placed in a pressure vessel filled with 

water and high pressure is then generated in the vessel by a 

pressure pump. 

Pressurization of food 

An egg is not crushed by compression at 600 MPa, but the egg 

white and yolk are coagulated. The color of egg yolk of a pressurized 

egg is naturally yellow, whereas a boiled egg changes 

to a faded yellow. The color difference is attributed to changes 

associated with pressurization: pressure affects only 

noncovalent bonding and coagulates proteins without splitting 

covalent bonds, thus keeping the color and smell intact. 

High pressure induces protein denaturation in the same 

way as high temperature. 

Meat protein is also denatured by pressure treatment at 

400 MPa, preserving the properties of raw meats. An example 

of prawns or shrimp is interesting: although a boiled prawn 

turns red and the meat coagulates, the appearance of a pressurized 

prawn is the same as that of raw shrimp, but the meat 

coagulates after pressurization at 400 MPa for 10 min. 

High pressure also has an effect on starches: a thick suspension 

of rice starch forms a ball by standing against its own 

weight after pressurization at 700 MPa, indicating that starches 

are gelatinized by the pressure treatment. 

Versatile utility of high pressure 

A high-pressure treatment generally coagulates protein, thereby 

inactivating the enzymes, gelatinizing the starches, and killing 

microorganisms. Thus, the use of high pressure promises to be 

a versatile process in food science and technology. Pressure 

produces a new texture in meats and starch-based foods, while 

keeping the original nutrients, flavor, and color. 

High-pressure effects on pure starches 

High-pressure-induced gelatinization of starch 

Effect of pressure on amylase digestibility of starches. The 

starches of potato, maize, and wheat are gelatinized by pressure 

treatment at warm conditions of 45–50 °C. The pressurization 

produces unique properties that are different from those 

of heat-gelatinization: heat-treatment destroys starch granules, 

resulting in a transparent solution, but a pressure-treatment 

swells the granules while maintaining the granular structure. 

Nevertheless, amylolytic enzymes such as α-, β-, and glc-amylases 

digest the pressurized starches well, being similar to the 

phenomenon in which the pressure-treatment of proteins increases 

protease digestibility (Hayashi and Hayashida 1989). 

The pressure-induced gelatinization of starches exhibits 

a sigmoid curve, suggesting that a two-state transition is involved, 

as in heat-induced gelatinization (see below). 

Effect of pressurization time on amylase digestibility of 

starches. To attain full amylase digestibility of starches, pressurization 

under warm conditions for 2 to 6 h is necessary. 

Interestingly, pressurization of starches for a longer time 

makes amylase digestion difficult: the amylase digestibility of 

starches decreased by 20–50% after pressurization for 17 h 

compared with the maximum digestibility obtained after pressurization 

for 2 to 6 h. 

These observations suggest that pressure induces gelatinization 

of starches, similar to heating, but prolonged pressurization 

produces a new stable structure of starches, which is 

not susceptible to attack by amylase (Ezaki and Hayashi 1992). 

Birefringence of starches after pressurization. The birefringence 

of starches is lost as increasing pressure is applied. 


Wheat starch is sensitive to pressure and birefringence is lost 

at 200 MPa. 

In the pressurization of starches, the number of granules 

exhibiting complete birefringence decreases without increasing 

the incomplete birefringent granules. This loss of birefringence 

shows that the crystalline structure is destroyed by high 

pressure as well as by high temperature and that the high-pressure-induced 

gelatinization of starches follows a two-state transition 

without an intermediate state of destruction. 

Physical properties of pressurized starches 

When a 50% water suspension of starches is pressurized at 

100 to 500 MPa and at 45 °C for 1 h and air-dried, followed 

by analyses of their physical properties, the results are as follows. 

According to X-ray crystal analysis, the crystalline structure 

of potato, waxy maize, and maize starches decreases with 

an increase in pressure, although the crystalline structure of 

potato starch does not change up to 500 MPa. The decrease in 

the high-pressure-induced crystalline structure is parallel with 

the increase in amylase susceptibility. 

Amylograms show that the transition temperature of pressurized 

starches is elevated, thus decreasing the viscosity. 

Interestingly, differential-scanning calorimetry (DSC) of 

pressurized starches shows that the peak temperature of the 

DSC patterns increases while the peak area decreases, indicating 

that some pressure-induced structural state of the pressurized 

starches is perturbed by low energy at higher temperatures. 

Summary of pressure-induced changes in starches 

The structure of pressurized starches changes accompanying 

the increase in amylase digestibility, a loss of birefringence, 

and a loss of crystallinity. These unique structural properties, 

which are somewhat different from those of heat-gelatinized 

starches, should be analyzed in detail for effective applications 

of high pressure to starch and related foods. 

Rice-based foods produced by high-pressure processing 

Dr. Akira Yamazaki, of Echigo Seika Company Ltd., studied 

the properties of pressurized rice in detail to introduce the use 

of high pressure to the food industry (Kinefuchi et al 1999, 

Sasagawa and Yamazaki 2002). He equipped several large highpressure 

machines in his own factory and succeeded in sending 

rice and rice products to the market by introducing the 

high-pressure technique to the manufacturing process of rice 

products, cooked rice (gohan), rice crackers (osembei), and 

rice cakes (omochi). 

High-pressure effects on rice grains 

Traditionally cooked-rice grains change their shape and develop 

some cracks, but rice grains after high-pressure pretreatment 

followed by heat-cooking swell, and their original shape 

is maintained without cracks. 

High-pressure cooked rice for microwave ovens 

Bread is purchased in the store and toasted just before eating. 

This is a typical life style, especially on a busy morning. However, 

45 min are required to cook rice. Cooked rice tastes best 

and has its best texture just after steaming. Warming of cold 

cooked rice makes the taste worse. In general, heating a starchbased 

food twice leads to an unpalatable food. In history, instant-rice 

is a dream of Japanese consumers. 

Dr. Yamazaki succeeded in producing oven-cooked rice 

with a good taste and texture for consumers, although the production 

scale is only enough to fulfill the requirement of a 

small city. He also succeeded in producing instant rice containing 

miscellaneous cereals. These instant cooked-rice cereals 

exhibited good taste and good texture by a 3-min heating 

in a microwave oven. 

When high-pressure pretreated and successively heatcooked 

rice is compared with traditionally cooked rice, its properties 

are as follows: first, it is more gelatinized; second, it is 

more slowly retrograded; and third, it is gelatinized to a greater 

extent by heating just before serving. Japanese consumers who 

are particularly sensitive to rice accept these properties. 

High-pressure rice cakes 

During New-Year days, the Japanese public is freed from the 

kitchen and enjoys the New-Year cerebrations, eating rice cakes 

(omochi) every day, which are a preserved food. The company 

introduced the high-pressure technique for producing the traditional 

rice cake and a special kind of omochi, which contains 

herbs with a natural color, smell, and taste, and this is 

now available on the market. 

Merits of high-pressure food processing 

High-pressure processing is useful not only for producing highquality 

food, but also for improving the manufacturing process. 

The production time for rice crackers is shortened by 

introducing a high-pressure technique into the traditional processing 

system and the total energy cost is decreased by 10% 

and the labor force by 56%. 

Conclusions 

The use of high pressure for food processing and cooking, in 

addition to heating and cooling, is now in our hands. Two factors, 

T and P, are useful for the manufacturing of good foods, 

including starch-based foods. Although T and P are used independently, 

their combined use is also important for the optimal 

use of high pressure. For example, heat-tolerant bacterial 

spores are inactivated by pressurization under elevated temperature. 

References 

Ezaki S, Hayashi R. 1992. High-pressure effects on starch: structural 

change and retrogradation. In: Balny C, Hayashi R, 

Heremans K, Masson P, editors. High pressure and biotechnology. 

France: John Libbey Eurotext. p 163-165. 


Hayashi R. 1987. Possibility of high-pressure technology for cooking, 

sterilization, processing, and storage of foods. Food Dev. 

(Shokuhin to Kaihatu) 22(7):55-62. 

Hayashi R, Kawamura Y, Kunugi S. 1987. Introduction of high pressure 

to food processing: preferential proteolysis of â-lactoglobulin 

in milk whey. J. Food Sci. 52:1107-1108. 

Hayashi R, Hayashida A. 1989. Increased amylase digestibility of 

pressure-treated starch. Agric. Biol. Chem. 53:2543-2544. 

Hayashi R, editor. 2002. Trends in high-pressure bioscience and biotechnology. 

Netherlands: Elsevier. 

Kinefuchi M, Watanabe K, Komiya S, Yamazaki A, Yamamoto K. 

1999. Characteristic of retrogradation of pressure-treated 

cooked rice. J. Appl. Glycosci. 46:31-38. 

Sasagawa A, Yamazaki A. 2002. Development and industrialization 

of pressure-processed foods. In: Hayashi R, editor. Trends in 

high-pressure bioscience and biotechnology. Netherlands: 

Elsevier. p 375-384. 

Notes 

Author’s address: Food Science and Technology, College of 

Bioresource Science, Nihon University, Kamenoi 1866, 

Fujisawa, Kanagawa 252-8502, Japan, e-mail: 

hayashi2@brs.nihon-u.ac.jp. 

Acknowledgments and sympathy: The author expresses his thanks 

to Miss Atsuko Hayashida, Kyoto University, and Dr. Souji 

Suzuki, Ajinomote Company, who performed a part of the 

experiments of high-pressure effects on starches, and sincere 

gratitude to Dr. Akira Yamazaki, manager of Echigo Seika 

Food Company, who gave me permission to use his data on 

high-pressure rice and rice products. 

On the evening of 23 October 2004, a huge earthquake 

struck Niigata Prefecture, especially Ojiya City. This city, the 

location of Echigo Seika Company Ltd., was the seismic center. 

I express my sincere sympathy to the people there, and, as 

it is now a very busy season for the coming New Year, I regret 

the damage to the rice-processing systems of the company, 

including the high-pressure machines, resulting from this earthquake. 

Developing novel processes for incorporating 

the unique nutritional and functional properties 

of rice into value-added products 

Elaine T. Champagne, Harmeet S. Guraya, Frederick F. Shih, and Ranjit S. Kadan 

An expanding global rice market and diet diversification have 

presented the rice industry in the United States and other developed 

countries challenges and opportunities to develop new 

markets for rice. Global trade has expanded with advances in 

technologies leading to record yields of high-quality rice. Countries 

that at one time were importers or had only a small share 

of the export market are now major exporters. The United 

States, once the lead exporter of high-quality rice, has been hit 

hard by this shrinking export market. Outside of the U.S., countries 

such as Brazil, China, Japan, Taiwan, and India have been 

faced with declining per-capita consumption because of diet 

diversification. Value-added rice products that capture the 

unique nutritional and functional attributes of rice offer a means 

to keep rice as a mainstay of the diet and to expand domestic 

markets. This paper will highlight new product introductions 

in the U.S. and technologies recently developed at the USDA- 

ARS Southern Regional Research Center for value-added rice 

products. 

New product introductions 

Domestic usage of rice in the United States has tripled in the 

last ten years in response to consumer demand for more healthful 

foods, more convenient products, and a growing interest in 

ethnic foods. The market has been flooded with new product 

introductions. Foods that use rice now include baby food, bak- 

ery goods, breakfast cereals, candy, desserts, beverages and 

dairy items, side dishes, package mixes, entrees (frozen, shelfstable, 

chilled), meats and meat substitutes, sauces and spreads, 

snacks, soups, and pet foods. 

Rice usage in entrees has increased 10-fold in the last 20 

years. Interest in ethnic foods has boosted this growth. Frozen 

ethnic entrees represented a $2.2 billion industry in 2000 

(Brandt 2001). A popular new addition is bowl meals—a frozen 

entree in a microwaveable bowl usually containing individually 

quick-frozen rice, meat or poultry, and vegetables. 

They have contributed largely to the growth of the entree market. 

Package rice mixes to serve as side dishes combine rice 

with seasonings, dehydrated vegetables, and other ingredients. 

The use of rice in packaged mixes doubled during the 1990s 

(USA Rice Federation 2000). Mixes for Louisiana Cajun and 

Creole foods, such as jambalaya, gumbo, and etouffee, and 

difficult-to-make dishes, such as risotto, are popular. Rice for 

mixes is either regular or parboiled, with a low moisture content 

around 6.5%. For a product in which distinct grains are 

important, such as a pilaf, or when a sauce will add moisture, 

parboiled rice often performs better than regular rice. 

Snack bars using rice include granola, breakfast, and 

energy bars. Energy bars, introduced in the 1990s, account for 

80 to 100 new product launches a year and represented about 

$500–700 million in sales in 2000 (Kreuzer 2001, Mintel In- 


ternational Group Ltd. 2003). Snack bars target the healthconscious 

consumer and are often enriched. A breakfast bar 

introduced in 2003 contains milk infused into the grains composing 

the bar to give one the taste of a bowl of cereal. Crisped 

rice is used in granola, breakfast, and energy bars. High-pressure 

extrusion processing is generally used for the manufacture 

of crisped rice for these snack bars. 

Desserts with rice emerged in the late 1990s with waxy 

rice serving as a fat replacer in ice cream. The addition of 

1.5% waxy rice starch improves the creamy mouthfeel and 

overall texture, allowing the product to mimic premium (higher 

fat) ice cream. The excellent freeze-thaw stability of waxy rice 

starch helps reduce iciness during storage (Bakal 1994, 

Wilkinson and Champagne 2004). Crisped rice provides texture 

to a number of chocolate dessert products. 

Rice incorporated into a meat product can help reduce 

fat content, for example, in the manufacture of “light” sausages. 

Rice starch can be used to bind water in poultry products 

during vacuum tumbling or injection. Rice and rice protein 

are used in meat analogs to help bind the other ingredients. 

Rice has also found new applications in dairy alternatives 

and electrolyte-replacement drinks. Rice offers a lactosefree, 

hypoallergenic alternative to cow and soy milk products. 

In soy milk, rice flour and syrup can be used to help mask 

beany flavors and keep insoluble proteins suspended in solution 

(Burrington 2002). In electrolyte-replacement drinks, the 

long-chain carbohydrates and proteins of rice enhance the absorption 

of fluid and electrolyte salts into the body’s cells 

(Greenough 1998). 

Novel processes developed at the Southern Regional 

Research Center 

The postharvest rice research program at the Southern Regional 

Research Center is directed at (1) obtaining a fuller understanding 

of component structure-function relationships in rice 

and its coproducts that contribute to their unique nutritional 

and functional properties and (2) developing new technologies 

for converting rice components, in situ and isolated, into 

high-value, high-demand products that capture these unique 

properties. Technologies resulting from this research are highlighted. 

Physical process for rice starch 

and protein separation 

Rice starch has not been produced in the United States since 

1943, except on a small scale in recent years. Seventy-five 

percent of the world’s 25,000 metric tons of rice starch produced 

annually is made in Europe. The classical, commercial 

rice starch operation employed in Europe requires extensive 

soaking in dilute sodium hydroxide prior to the separation of 

starch and protein. This wet-milling process is water, energy, 

and time intensive and requires costly wastewater treatments. 

The usage of alkali results in a salt disposal problem and can 

impart bitterness to the protein component. The environmental 

and energy issues associated with this process have hindered 

the comeback of a rice starch industry in the U.S. 

Dr. Harmeet Guraya under a Cooperative Research and 

Development Agreement (CRADA) and Small Business Initiative 

Research (SBIR) grant with Sage V Foods (Los Angeles, 

California) has developed a physical process for separating 

rice flour into starch and protein. In this process, proteinstarch 

agglomerates of rice are physically disrupted in the presence 

of water by the use of high-pressure homogenization, as 

achieved by the use of a microfluidizer. The deagglomeration 

leads to better density-based separation of rice starch and protein. 

This process yields starch with low damage (5–6% versus 

12% in commercial starches) and low protein content 

(< 0.5%). The protein has high solubility compared with that 

of commercially available protein concentrates, making it suitable 

for beverage applications. Sage V Foods has been granted 

an exclusive license for U.S. Patent 6,737,099 (18 May 2004), 

“Process for the Deagglomeration and Homogeneous Dispersion 

of Starch Particles.” The company has scaled-up the process 

and will shortly begin manufacturing the starch product. 

This technology has the potential of reducing imports of rice 

starch and increasing profits for the U.S. rice industry. 

Physical process for quick-cooking brown rice 

Brown rice is rich in minerals and vitamins, making it a nutritionally 

valuable food. Consumption of brown rice, however, 

is low. A major drawback for brown rice is its long cooking 

time (45–50 min) because of the slow rate of hydration. This 

long cooking time produces an undesirable sticky, soft texture 

on the surface of the kernel. Existing commercial methods for 

producing quick-cooking brown and white rice involve precooking 

the rice, followed by drying. These methods require a 

significant input of water and energy, which, in turn, creates 

significant expense. 

Dr. Guraya has invented a dry instantization process that 

reduces the cooking time of brown and wild rice from 45–50 

min to that of white rice (20 min). This process “sandblasts” 

brown rice with parboiled rice flour. The bombardment creates 

microperforations in the waxy layer of the bran, allowing 

the kernel to absorb water more readily and cook more rapidly. 

Visually, the uncooked kernel looks the same as an untreated 

kernel. The cooked kernel has a texture similar to that 

of white rice and is not soft and mealy like that of untreated 

brown rice. The U.S. Patent 6,586,036 (1 July 2003), “Process 

for Increasing Rate of Hydration of Food Crop Seeds,” 

has been licensed by three companies. One licensee, Progressive 

Technologies (Grand Rapids, Michigan), has manufactured 

a continuous system for production of these products by 

the other two licensees. The process cost is approximately 2 to 

4 cents per kg of brown rice. The invented process reduces the 

cost of processing to make quick-cooking rice, reduces environmental 

pollution, and will make nutritious brown rice more 

appealing to consumers. 


Gluten-free rice bread 

Economical rice bread products are needed for consumers with 

Celiac Sprue disease and other disorders that prevent consumption 

of gluten-containing grain (e.g., wheat) products. Dr. Ranjit 

Kadan has developed formulations for rice bread using a home 

bread machine. The prototype bread has desirable flavor and 

texture comparable with those of wheat bread. A U.S. patent 

application has been submitted. The developed process will 

allow consumers to readily and economically prepare glutenfree 

bread. Ingredients cost $0.65 kg –1 loaf –1 , with marketing 

cost 2–3 times this cost. This compares with $4–5 loaf –1 for 

commercial wheat bread. 

Low-oil-uptake rice batter and donuts 

Fried batters may enhance the sensory quality of the coated 

food, but they may also contain high amounts of oil and contribute 

to oil-related health problems such as obesity and heart 

disease. Therefore, in spite of their popularity, fried foods with 

excessively elevated oil are undesirable and should be avoided. 

Dr. Fred Shih and Ms. Kim Daigle (1999) found rice flour to 

have better oil resistance than wheat flour when formulated as 

a batter. However, rice-flour slurries, at concentrations such 

that the viscosity is appropriate for a batter, become brittle 

and hard during frying. These observations led them to discover 

that gelatinized long-grain rice flour and phosphorylated 

long-grain rice starch esters can be effective in enhancing both 

the viscosity and oil-lowering properties of rice-flour batters. 

These batters, when applied to chicken, absorb up to 60% less 

oil than a traditional wheat-based batter. U.S. Patent 6,224,921 

(1 May 2001), “Rice flour-based low oil-uptake frying batters,” 

was issued. 

Following an approach similar to that taken for the batter 

development, Shih and Daigle (2002) developed low-oiluptake 

rice donuts. The product formulated with long-grain 

rice flour and pregelatinized long-grain rice flour (30%) absorbed 

as much as 54% less oil than traditional wheat-flour 

donuts. 

Fungi-free rice straw 

Japan desires to import rice straw from the United States for 

cattle feed and requires that the straw be free of fungi. Dr. 

Guraya has invented an economical process for fungi-free rice 

straw. Research is under way, in collaboration with an industry 

partner through a CRADA, to make the process continuous 

and scale it up for a commercial operation. This process will 

allow the U.S. to export rice straw to Japan, which currently 

imports 2 million tons of forage other than that from rice. 

New research thrusts 

Current research efforts are directed at developing novel technologies 

for health-beneficial products from rice bran and hulls. 

These technologies include processes for (1) protein concentrates 

and isolates for infant formulas, beverages, and ingredient 

applications; and (2) fractions (rice wax, wax-rich fractions, 

hull and bran extracts) with cholesterol-lowering and 

antioxidative activity for various food applications. Technologies 

for new low-oil-uptake rice and sweet potato–rice products 

that suit the need of health-conscious consumers are also 

being developed. All of these applications target the unique 

nutritional and functional attributes of rice co-products and 

their components. The processes being pursued to achieve these 

products are efficient, environmentally friendly, and commercially 

viable. 

References 

Bakal A. 1994. The lowdown on formulating lowfat. Prep Foods 

163:75-78. 

Brandt LA. 2001. Entrees take the orient express. Prep Foods 170:65- 

68. 

Burrington KJ. 2002. More than just milk. Food Prod. Design 12:37- 

64. 

Greenough WB. III. 1998. Oral hydration therapy: something new, 

something old. Infec. Dis. Clin. Pract. 7:97-100. 

Kreuzer H. 2001. High-octane bars and beverages. Food Prod. Design 

11:33-58. 

Mintel International Group Ltd. 2003. Global new products database: 

category review. At www.gnpd.com. (See categories 

snack bars, snack mixes, and energy bars; cold cereal; rice; 

soup.) 

Shih FF, Daigle KW. 1999. Oil uptake properties of fried batters 

from rice flour. J. Agric. Food Chem. 47:1611-1615. 

Shih FF, Daigle KW. 2002. Preparation and characterization of low 

oil uptake rice cake donuts. Cereal Chem. 79:745-748. 

USA Rice Federation. 2000. U.S. rice distribution patterns, 1999- 

2000 report. The Federation, Houston, TX. (Also 1996 and 

other rice distribution pattern reports through 2001.) 

Wilkinson HC, Champagne ET. 2004. Value-added rice products. 

In: Champagne ET, editor. Rice chemistry and technology. 

3rd edition. St. Paul, Minn. (USA): American Association of 

Cereal Chemists. p 473-493. 

Notes 

Authors’ address: USDA-ARS Southern Regional Research Center, 

New Orleans, Louisiana, USA, e-mail: 

etchamp@srrc.ars.usda.gov. 


Dehydrin proteins in rice bran 

Michiko Momma 

Dehydrin is a group of late embryogenesis-abundant (LEA) 

proteins, which accumulate in the maturing seed of many plants. 

Dehydrins are reported to protect protein or membranes in plant 

tissue under abiotic stress such as desiccation or low temperature 

(Close 1997). Besides the biological function, application 

for preservation of food quality is expected, similar to 

antifreeze proteins in fish (Breton et al 2000). Since rice bran 

could be a good source of dehydrin for nonfood usage, we 

tried to characterize rice dehydrin in this study. Rice bran protein 

was analyzed by two-dimensional (2D) electrophoresis, 

and dehydrins in the protein fraction were detected by 

immunoblotting using antibody against the conserved motifstructure 

in order. A major rice dehydrin was partially purified 

from rice bran and its cryoprotective activity was estimated to 

assess the function of the protein. 


Fractionation and partial purification of proteins 

Rice bran was obtained in the process of refining 1 kg of rice 

(cv. Koshihikari) to 90%. Fifty grams of the rice bran was defatted 

with 500 mL of hexane three times and kept at –20 o C 

before use. The defatted rice bran (20 g) was mixed with 20 

mM HEPES-NaOH, pH 7.0, for 3 min by a hiscotron mixer 

(NS-50, Nichi-On), centrifuged twice at 8,000x g for 20 min. 

Supernatant was collected as a water-soluble fraction of rice 

bran. The heat-tolerable protein fraction was prepared by heating 

the water-soluble fraction in boiling water for 10 min and 

removing precipitate by centrifugation at 12,000x g for 20 min. 

Rice dehydrin was partially purified by modifying the 

method of purification of soybean dehydrin (Momma et al 

2003). From 20 g of defatted rice bran, a 10–50% saturated 

ammonium sulfate fraction was isolated, dialyzed, heated in 

boiling water, and then centrifuged. Supernatant was put onto 

an S Sepharose Fast Flow column (2 × 35 cm), and eluted with 

a linear gradient of 0–300 mM NaCl. The dehydrin fraction 

was collected based on detection by SDS-PAGE and 

immunoblotting with antiserum against the conserved lysinemotif 

of dehydrin. The water-soluble fraction, heat-tolerable 

fraction, and partially purified dehydrin were dialyzed against 

distilled water and freeze-dried and used to assess cryoprotective 

activity. 

Electrophoresis of rice bran proteins 

SDS-PAGE was carried out on 5–20% polyacrylamide gradient 

gel (PAGEL, NPG-520L, ATTO) by the method of Laemmli 

(1970). Gels were stained with Coomassie Brilliant Blue (CBB) 

R-250. Two-dimensional electrophoresis of protein fractions 

isolated from rice bran was conducted with the Multiphore II 

System (Amersham Pharmacia Biotech) following the 

supplier’s instructions. Protein fractions were mixed with 

sample buffer (3-10NL) and IPG strips (pH3-10NL, 7 cm) were 

soaked in the sample solution for 15 h. The IPG strips were set 

on the system, and electrophoresis was carried out at 200– 

3,500 V for 90 min and at 3,500 V for 65 min. IPG strips were 

then equilibrated in 50 mM Tris-HCl, pH 8.8, containing 6 M 

urea, 30% (w/v) glycerol, 2% SDS, 10 mg mL –1 dithiothreitol, 

and a trace of bromophenol blue for 20 min. Second-dimensional 

electrophoresis (SDS-PAGE) was carried out on a 5– 

20% polyacrylamide gradient gel. Gels were stained with CBB 

R-250. 

Immunoblot analysis 

Electrophoresis gels were incubated in a blotting buffer (25 

mM Tris-HCl, pH 9.5, containing 40 mM ε-aminocaproic acid, 

20% methanol, 0.05% SDS) for 20 min, and then protein on 

the gel was blotted onto a PVDF membrane (Immun-Blot, 

Biorad) at 2.5 mA for 30 min. The PVDF membrane was kept 

overnight in 3% BSA-TBS (Tris-buffered saline, 20 mM Tris- 

HCl, pH 7.5, 500 mM NaCl, containing 3% bovine serum albumin). 

Dehydrin protein was detected by polyclonal antiserum 

raised against conserved lysine-motif sequence 

(DQNEKKGIMDKIKEKLPGGH) conjugated with ovalbumin. 

The PVDF membrane was incubated for 60 min in the 

antiserum of rabbits diluted to 1:200 with 1% BSA-TBS, then 

washed with TBS-t (Tris-buffered saline containing 0.05% 

tween 20) for 20 min three times, and with TBS for 20 min. 

The washed membrane was incubated with peroxydase-conjugated 

antirabbit goat antibody diluted to 1:2,000 with 1% 

BSA-TBS, washed with TBS-t for 20 min three times, and 

washed with TBS for 20 min. Reacted polypeptides were detected 

by a detection kit (Immunostain HRP1000, Konica). 

Cryoprotective activity on lactate dehydrogenase 

Cryoprotective activities of dehydrins in comparison with bovine 

serum albumin (BSA, A4378, Sigma) on freeze/thaw inactivation 

of lactate dehydrogenase were assayed following 

the method of Lin (1992). Lactate dehydrogenase from rabbit 

muscle (L5132, Sigma) was dissolved in 10 mM sodium phosphate 

buffer, pH 7.5, at a concentration of 2.5 µg mL –1 . The 

enzyme solution (0.1 mL) was mixed with equal amounts of 

test compounds dissolved in the same buffer at a concentration 

of 2 × 10 2 – 2 × 10 –4 µM. The solutions were frozen at 

–20 o C for 24 h and thawed at room temperature for 5 min. 

Enzyme activity was measured with the change of absorbance 

at 340 nm at 25 o C using an assay kit (DG1340K, Sigma). All 

samples were assayed twice in triplicate. Residual activity was 

shown as the percentage of the control activity assayed immediately 

after mixing of enzyme and protein solution. 


A 

– 3-10NL + 

B 

– 3-10NL + 

C 

– 3-10NL + 

97.4 

106.0 

97.4 

66.2 

80.0 

66.2 

45.0 

49.5 

45.0 

31.0 

21.5 

14.4 

kDa 

32.5 

27.5 

18.5 

kDa 

Fig. 1. Two-dimensional electrophoresis and immunoblotting. (A) 2D electrophoresis of water-soluble fraction, (B) immunoblot of watersoluble 

fraction, (C) 2D electrophoresis of heat-tolerable fraction. = spots reacted with antibody. 

31.0 

21.5 

14.4 

kDa 


In the results of SDS-PAGE analysis, two bands of dehydrinlike 

protein were detected in water-soluble and heat-tolerable 

fractions of rice bran by immunoblotting for the conserved 

motif-sequence of dehydrin. Two major bands of 44 kDa and 

23 kDa were detected with similar intensity in this experiment. 

Still et al (1994) had investigated the accumulation of dehydrin 

in the course of seed development of rice using an antiserum 

similar to the one we used. They detected a polypeptide estimated 

to be 21 kDa as a major dehydrin-like protein in the 

seeds of 15–39 DAF along with a small amount of 38-kDa 

polypeptide, which also cross-reacted with the antibody. It had 

been known that dehydrin with low molecular weight (about 

20 kDa) also accumulated in rice seedlings under desiccation 

stress. They concluded that the 21-kDa polypeptide was the 

major dehydrin protein in rice seed. As for the dehydrin-like 

proteins with higher molecular weight, besides the 38-kDa 

polypeptide in developing seed, Jayaprakash et al (1998) found 

a trace amount of 60-kDa and 45-kDa polypeptides in desiccated 

rice seedlings. Dehydrin cognates with a higher molecular 

weight than the major one were detected in other plant seeds 

such as pea or wheat. They have often been found to lack the 

N-terminal motif sequence consisting of DEYGNP (Danyluk 

et al 1994). From findings in previous publications, the 23- 

kDa polypeptide detected by the immunoblot analysis was 

supposed to be the major dehydrin in rice seed and seedlings 

under abiotic stress, and that 45-kDa polypeptide would be 

one of the dehydrin-cognate proteins. 

Two-dimensional electrophoresis profiles of rice bran 

proteins are shown in Figure 1. In the results of immunoblotting 

for dehydrin motif, 3 spots around 44 kDa and 5 spots around 

23 kDa were detected. Though no clear difference was observed 

in the SDS-PAGE pattern of the water-soluble and heattolerable 

protein fraction, their 2D-electrophoresis patterns 

were considerably different. In general, the number of detected 

spots decreased with the heat treatment, but some of the spots 

on the acidic side seemed to be concentrated in the heated 

sample. In the heat-tolerable fraction, cross-reacted spots were 

shifted to the acidic side and the separation of the spots became 

obscure, implying that they might interact with other 

components in rice bran. 

Remaining activity (%) 

120 

100 

80 

60 

40 

20 

Bovine serum abumin 

Sucrose 

Ovalbumin 

Rice dehydrin 

Soybean dehydrin 

0 

10 –6 10 –4 10 –2 1 10 2 10 4 

Concentration (mg m –1 ) 

Fig. 2. Cryoprotective activity of 23-kDa dehydrin. 

The cryoprotective activity of partially purified 23-kDa 

dehydrin from rice bran on the freeze/thaw inactivation of lactate 

dehydrogenase was estimated (Fig. 2). The CP 50 values, 

protein amounts necessary to keep 50% activity of lactate dehydrogenase, 

of the water-soluble fraction, heat-tolerable fraction, 

and dehydrin were 43, 29, and 15.6 µg mL –1 (0.78 µM), 

respectively. In our previous study, we assayed the activities 

of two types of purified soy dehydrins (Momma 1997, 2003). 

The CP 50 of soybean 26-kDa dehydrin had been 5.2 µg mL –1 

(0.2 µM). Compared with the soy protein and other commercially 

available proteins, rice 23-kDa dehydrin had a level of 

cryoprotective activity similar to that of ovalbumin and lacto 

albumin, and about one-third the activity of 26-kDa soybean 

dehydrin. 

Lin and Thomashow (1992) showed that in vitro-translated 

COR15 from Arabidopsis had cryoprotective activity on 

freeze/thaw inactivation of lactate dehydrogenase. This was 

the first report indicating a function of LEA-related proteins. 

This cryoprotective activity was found in Chlorella HIC6, a 

group 3 LEA protein (Honjoh et al 2000), and a protein from 


an ice-nucleating bacterium and others as well. It is presumed 

that the extreme hydrophilic property of LEA proteins relates 

to their protective function on protein denaturation. The cryoprotective 

activity is considered as an index of function of protein 

induced by stress such as low temperature or desiccation. 

Breton et al (2000) suggested the possibility that the cryoprotective 

activity of LEA protein would be applicable to preserve 

the quality of frozen food or prevent frostbite. In this 

study, we found multiple dehydrin proteins in water-soluble 

and heat-tolerable proteins of rice bran. In the results of in 

vitro assay, rice dehydrin was found to preserve enzyme activity 

against freeze/thaw damage. The fact that the rice dehydrin 

has cryoprotective activity is encouraging for novel usage of 

rice bran protein, but further characterization of the protein 

would be necessary for the development of applications. 

References 

Breton G, Danyluk J, Ouellet F, Sarhan F. 2000. Biotechnological 

applications of plant freezing associated proteins. Biotechnol. 

Annu. Rev. 6:59-101. 

Close TJ. 1997. Dehydrins: a commonality in the response of plants 

to dehydration and low temperature. Physiol. Plant. 100:291- 

296. 

Danyluk J, Houde M, Rassart E, Sarhan F. 1994. Differential expression 

of a gene encoding an acidic dehydrin in chilling 

sensitive and freezing tolerant gramineae species. FEBS Lett. 

344:20-24. 

Honjoh K, Matsumoto H., Shimizu H, Ooyama K, Tanaka K, Oda Y, 

Tanaka R, Joh T, Suga K, Miyamoto T, Iio M, Hatano S. 2000. 

Cryoprotective activities of group 3 late embryogenesis abundant 

protein from Chlorella vulgaris C-27. Biosci. Biotechnol. 

Biochem. 64:1656-1663. 

Jayaprakash TL, Ramamohan G, Krishnaprasad BT, Ganeshkumar, 

Prasad TG, Mathew MK, Udayakumar M. 1998. Genotypic 

variability in differential expression of lea2 and lea3 genes 

and proteins in response to salinity stress in fingermillet 

(Eleusine coracana Gaertn.) and rice (Oryza sativa L.) seedlings. 

Ann. Bot. 82:513-522. 

Laemmli UK. 1970. Cleavage of structural proteins during the assembly 

of the head of bacteriophage T4. Nature 227:680-685. 

Lin C, Thomashow MF. 1992. A cold-regulated Arabidopsis gene 

encodes a polypeptide having potent cryoprotective activity. 

Biochem. Biophys. Res. Commun. 183:1103-1108. 

Momma M, Haraguchi K, Saito M, Chikuni K, Harada K. 1997. 

Purification and characterization of the acid soluble 26-kDa 

polypeptide from soybean seeds. Biosci. Biotechnol. Biochem. 

61:1286-1291. 

Momma M, Kaneko S, Haraguchi K, Matsukura U. 2003. Peptide 

mapping and assessment of cryoprotective activity of 26/27- 

kDa dehydrin from soybean seeds. Biosci. Biotechnol. 

Biochem. 67:1832-1835. 

Still DW, Kovach A, Bradford KJ. 1994. Development of desiccation 

tolerance during embryogenesis in rice (Oryza sativa) 

and wild rice (Zizania palustris) dehydrin expression, abscisic 

acid content, and sucrose accumulation. Plant Physiol. 

104:431-438. 

Notes 

Author’s address: National Food Research Institute,2-1-12 Kannodai, 

Tsukuba, Ibaraki 305-8642, Japan. 

Physicochemical properties of modified rice flour 

and its use for processed food 

T. Takahashi, M. Miura, N. Ohisa, K. Mori, and S. Kobayashi 

Rice is an important grain. Most of it is consumed as milled 

rice. Interest in rice-based products and their processing technologies 

has increased despite higher costs than for either maize 

or wheat. Modified (by a physical or chemical treatment) 

starches and cereal flours have become important in processed 

food because the functional properties of the starches and flours 

are improved over the gelatinization properties of native 

starches and cereal flours. Numerous studies have explored 

the physical properties of rice flour from treated kernels, but 

few studies have addressed modified rice flour’s characteristics 

of cooking and processing. Our study aims to clarify the 

effects of heat treatments on gelatinization property and cooking 

quality. 


Preparation of heat-treated rice flour 

The moisture content of milled rice (japonica, nonglutinous 

type) was 12.4% (w/w), determined by the drying method. 

Amylose content was determined to be 17.7% (w/w) by the 

method of Juliano (1971). The milled rice was heated at 120 

°C for 60 min by an autoclave. The moisture content was 20.6% 

after autoclaving and 12.9% after drying at room temperature 

(25 ± 1 °C) for 24 h. The milled rice was heated at 160 °C for 

60 min by an air-oven. The moisture content was 2.2% after 

oven treatment. Untreated rice flour (UTR), heat-moisturetreated 

rice flour (HMR), and heat-dry-treated rice flour (HDR) 


Table 1. Dynamic viscoelastic characteristics of 10.0% (w/w) rice flour suspensions during gelatinization. a 

Samples G´o/°C G´p/Pa G´95 /Pa G2 BD /Pa tan δ/- 

UTR 63.9 ± 0.8 a 1396 ± 55 c 611 ± 55a b 784 ± 69 b 0.12 ± 0.01 a 

HDR 65.5 ± 0.4 b 778 ± 40 b 660 ± 30 b 117 ± 29 a 0.13 ± 0.01 a 

HMR 79.8 ± 2.1 c n.d. 519 ± 26 a n.d. 0.26 ± 0.01 a 

a G´o, onset temperature; G´p, peak G´; G´95 , G´ at 95 °C; G´ BD , breakdown (= G´ p – G´ 95 ). n.d., not detectable. Mean ± 

standard deviation (n = 3). Values followed by the same letter in the same column are not significantly different (P85 

°C) increased slightly. HMR paste exhibited weak gel because 

the tan δ of HMR was greater than that of UTR and HDR. It is 

recognized that a rigid starch granule formed with the interaction 

of both amorphous and crystallite regions during heatmoisture 

treatment. The rheological behavior of rice flour paste 

was primarily attributed to interaction in the system, such as 

contact between the granules directly and between the granules 

and glucose chains (Hoover and Vasanthan 2000). 

Table 2 shows the DSC characteristics of 30% (w/w) 

rice flour suspensions. The T o of HMR was 8 °C higher than 

that of UTR and HDR (Table 2). Other endothermic peaks 

(T p2 and T p3 ) in the DSC corresponding to amylose-lipid complexes 

were observed at temperatures of 100–120 °C (Table 

2). The peak transition for HMR was at a higher temperature 

(112 °C) than for UTR and HDR (102 °C). Lai (2000) reported 

that increases in gelatinization (T o ) and the peak of 

amylose-lipid complex (T p2 ) temperature were observed in DSC 

curves of parboiled rice flour. It was also reported that the 

peak of the amylose-lipid complexes shifted to a higher temperature 

with heat-moisture treatment. Changes in DSC parameters 

of HMR indicate that rearrangement of starch chains 

occurs during heat-moisture treatment, and leads to the formation 

of crystallites of different thermal stabilities. Therefore, 

the modification of G´ p and T o of HMR is attributed to more 

thermally stable crystallites in starch after heat-moisture treatment. 

The intensities of the major peaks of HMR were weaker 

than that of UTR and HDR measured by X-ray diffractometry. 

However, the diffraction peak of HMR at 19.5° (2θ) became 

much sharper and another feature peak at 13° (2θ) was also 


Table 2. Effect of heat treatments on differential scanning calorimetry characteristics of rice flours in excess water. a 

Samples 

Gelatinization 

Amylose-lipid complex 

T o /°C T p1 /°C T c /°C ∆H/kJ·kg -1 T p2 /°C ∆H 2 /kJ·kg -1 T p3 /°C ∆H 3 /kJ·kg -1 

UTR 60.2 ± 0.3b 68.1 ± 0.1 b 78.4 ± 0.3 b 8.6 ± 0.4 b 101.4 ± 0.6 a 0.5 ± 0.0 a – – 

HDR 58.1 ± 0.3 a 66.5 ± 0.2 a 76.8 ± 0.2 a 9.6 ± 0.2 c 102.4 ± 1.9 a 0.8 ± 0.1 b – – 

HMR 68.3 ± 0.0 c 76.2 ± 0.2 c 86.7 ± 0.5 c 5.2 ± 0.1 a – – 111.9 ± 0.0 0.7 ± 0.0 

a 

T o , onset temperature; T p , peak temperature; T c , conclusion temperature; ∆H, gelatinization enthalpy. Values followed by the same letter in the same column are not significantly 

different (P

Introducing soybean β-conglycinin genes into rice 

to improve nutritional and physiological value 

Takayasu Motoyama, Nobuyuki Maruyama, Takahiko Higasa, Masaaki Yoshikawa, Fumio Takaiwa, and Shigeru Utsumi 

In advanced countries, the aging of society is becoming a serious 

problem. Especially in Japan, elderly people over age 65 

account for 14.5% of the population, and it is estimated that 

they will increase to 24.7% by 2025 and 32.3% by 2050. In 

such an ultra-aged society, lifestyle-related diseases such as 

hypertension, hyperlipemia, and diabetes could spread. Further, 

our immune system weakens against various viruses as 

we age. Therefore, demand for food that can prevent lifestylerelated 

disease and stimulate the immune system will become 

larger and larger in advanced countries. 

Soybean β-conglycinin, composed of α, α′, and β subunits, 

is reported to have many physiological functions such 

as cholesterol-lowering activity in human serum (Sirtori et al 

1995), serum triglyceride-lowering activity (Aoyama et al 

2001), LDL-cholesterol-lowering activity (Manzoni et al 1998), 

and phagocytosis-stimulating activity in the α′ subunit (Tsuruki 

et al 2001). On the other hand, rice seed storage proteins, glutelin, 

prolamin, and globulin, do not have any significant physiological 

functions. Furthermore, they are deficient in lysine, 

whereas β-conglycinin, especially the α′ subunit, is rich in 

lysine. Therefore, in this study, we attempted to develop 

transgenic rice containing the α′ subunit for conferring high 

nutritional value, the modified β subunit having improved phagocytosis-stimulating 

activity (Maruyama et al 2003) for 

physiological property, and the β subunit for control of phagocytosis-stimulating 

activity. 

Transformation of rice with chimeric α′ and β genes 

β-conglycinin α′ and β cDNAs were connected with GluB-1 

and GluB-2 promoter regions in the pGTV-HPT vector (Becker 

et al 1992), respectively. The resultant pGTV-HPT/α′ and 

pGTV-HPT/β were introduced into rice calli by Agrobacterium 

tumefaciens-mediated transformation. Eleven and 22 lines 

expressing α′ and β were regenerated, respectively. Total proteins 

extracted by SDS buffer from T 1 seeds were spotted onto 

nitrocellulose membranes and the accumulation levels of α′ 

and β were estimated immunologically. The accumulation level 

of α′ was basically higher than that of β. The average and highest 

accumulation levels of α′ and β were 3.9% and 1.9%, and 

8.5% and 4.3%, respectively. The lines with the highest accumulation 

underwent subsequent analyses. To investigate the 

reason for this difference, we compared the transcription level 

of α′ and β in rice seeds. The transcription level of α′ was 

close to that of β. This suggests that the difference in accumulation 

level between α′ and β did not depend on the transcriptional 

step, but depended on the posttranscriptional step. 

Interaction of α′ and β with rice seed storage proteins 

To investigate whether α′ and β interacted with rice seed storage 

proteins glutelin and prolamin, sequential extraction of 

proteins from transgenic rice seeds and western blotting using 

anti-α′ and β sera were conducted (Fig. 1-I). β contains no 

Cys residue, whereas α′ contains four Cys residues in the pro 

region and one near the N-terminus of the mature region. Most 

of the β was extracted with buffer without 2-mercaptoethernol, 

2-ME (fraction 1), but not by buffer with 2-ME (fraction 5) 

(Fig. 1-IB). On the other hand, α′ was extracted by buffer without 

and with 2-ME (Fig. 1-IA). In this case, fraction 5 contained 

acidic polypeptide of glutelin (data not shown), suggesting 

that α′ interacted with glutelin by disulfide bonds. Fractions 

9–16 containing glutelin extracted by lactic acid gave 

bands of α′ and β, corresponding to 19% and 20% of their 

total amounts, respectively, and there was no difference in their 

distributions in each fraction. Therefore, α′ and β in fractions 

9–16 probably interacted with glutelin by hydrophobic interaction. 

Fraction17 containing prolamin extracted by SDS buffer 

also gave the bands of α′ and β, corresponding to about 10% 

of their total amounts. This indicates that α′ and β interacted 

with prolamin by disulfide bonds or hydrophobic interaction. 

Subcellular localization of β-conglycinin 

in mature transgenic seeds 

Subcellular localizations of α′ and β in transgenic rice endosperm 

were analyzed by immunocytochemical analysis using 

immunoelectron microscopy (IEM). Rice seeds have two 

types of protein bodies (Tanaka et al 1980). Protein body I 

(PB-I) is derived from ER, mainly accumulating prolamin, and 

protein body II (PB-II) is derived from vacuole, mainly accumulating 

glutelin and globulin. A morphological difference was 

not observed between nontransgenic and α′-transgenic rice 

seeds. When ultra-thin sections of α′-transgenic rice seed were 

labeled with anti-α′ antiserum conjugated with gold particles, 

these particles were mainly detected in the peripheral region 

of PB-II (Fig. 1-IIA), and nonspecific adsorption was not observed 

in the nontransgenic rice seeds under the same conditions. 

On the other hand, there was an obvious morphological 

difference between nontransgenic and β-transgenic rice seeds 

(Fig. 1-IIB). The electron density of PB-II was uniform in 

nontransgenic rice seed, but PB-II in β-transgenic rice seed 

contained a slightly low electron density region. When this 

seed was reacted with anti-β serum, the gold particles were 


I 

–2-ME +2-ME Lactic acid 

A 

M 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 

–2-ME +2-ME Lactic acid 

B 

M 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 

II 

A 

C 

E 

B 

D 

F 

Fig. 1. Accumulation of α′ and β-conglycinin in rice seeds. (I) Sequential extraction of α′ (A) and β (B) from individual transgenic rice 

seeds. At first, proteins were extracted from ground rice seeds four times by buffer without 2-ME (lanes 1 to 4), then four times by 

buffer with 2-ME (lanes 5 to 8), subsequently eight times by lactic acid (lanes 9 to 16), finally by SDS buffer (lane 17). (II) Electron 

micrographs of α′- (A, B, C) and β- (B, D, F) transgenic rice seeds. A and B, mature seeds; C to F, developing seeds. 

observed only in the low electron density regions, and no nonspecific 

adsorption was observed in the nontransgenic rice 

seeds. Anti-glutelin gold particles were not observed in this 

region. These results indicate that β and glutelin were separately 

accumulated in a different region of PB-II. Around 10% 

of α′ and β expressed in the rice seeds was accumulated in 

PB-I. 

Subcellular localization of β-conglycinin in developing seeds 

Rice storage proteins are basically synthesized on the ER, then 

assembled in the ER lumen, transported to the Golgi apparatus, 

and sorted into the vacuoles (Krishnan et al 1986). To 

investigate the trafficking pathway of α′ in rice seeds, the developing 

seeds of α′-transgenic rice were observed immunologically 

by EM. There was no morphological difference between 

α′-transgenic and nontransgenic rice seeds. In the early 

stage of PB-II formation, α′ was already localized in the peripheral 

region (Fig. 1-IIC). Vesicles budding from the Golgi 

apparatus were observed, and anti-α′ gold particles recognized 

these vesicles (Fig. 1-IIE). This suggests that the α′ subunit 

was transported to the vacuoles via the Golgi apparatus. 

In the β-transgenic rice seeds, some PB-II had two regions, 

high and low electron densities (Fig. 1-IID). The low 

electron density region of PB-II was not seen in immature α′- 

transgenic and nontransgenic rice seeds. Anti-β gold particles 

mainly existed in low electron density regions of PB-II. On 

the other hand, anti-glutelin gold particles did not exist in the 

low electron density regions of PB-II. Moreover, morphologically 

different vesicles were observed (Fig. 1-IIF, arrow). 

Though the size of these vesicles was analogous with that of 

conventional vesicles (Fig. 1-IIE, arrowhead), with a diameter 

of 100 to 500 nm and uniform smooth density surface (Oparka 

and Harris 1982), the morphologically different vesicles were 

surrounded by ribosome membrane-like precursor-accumulating 

(PAC) vesicles, observed in pumpkin seeds (Hara- 

Nishimura et al 1998). It has been reported that PAC vesicles 

serve to carry storage proteins from the ER to the vacuole, 

bypassing the Golgi apparatus. So, rice seeds may possess the 

same trafficking pathway. To investigate whether the PAC-like 


A 

I 135 - VNPHDHQNLKIIWKAIPVNKPGRYDDFFLSSTQAQQ - 170 

II 135 - VNPHDHQNLKMIWLAIPVNKPGRYDDFFLSSTQAQQ - 170 

B 

1 2 3 

C 

Phagocytotic index 

4.0 

3.5 

3.0 

2.5 

2.0 

1.5 

1.0 

0.5 

0 0 0.01 0.1 1.0 10 

Concentration (mM) 

Fig. 2. Development of rice to stimulate human phagocytotic activity. (A) Substitution of amino 

acids of β subunit to confer the phagocytotic activity. I, β subunit; II, I122M/K124W. (B) SDS- 

PAGE of proteins extracted by buffer from β and I122M/K124W transgenic rice seeds. Lane 1, 

nontransgenic rice; lane 2, β transgenic rice; lane 3, I122M/K124W-transgenic rice. (C) Phagocytotic 

assay of I122M/K124W and β. Solid line, I122M/K124W; dotted line, β. 

vesicles are derived from the ER, we observed immunologically 

with anti-serum against binding protein (BiP) the ERlocated 

chaperon protein. Consequently, BiP was located in 

the PAC-like vesicles (data not shown), suggesting that the β 

subunit was transported to the vacuoles not only from the Golgi 

apparatus but also from the ER, bypassing the Golgi apparatus. 

Developing rice to stimulate human phagocytotic activity 

We also aimed to develop rice to stimulate human phagocytotic 

activity. For this, we had substituted two amino acid residues 

of the β subunit to confer the phagocytotic activity (Fig. 

2A) to produce I122M/K124W (Maruyama et al 2003). Subsequently, 

we connected I122M/K124W cDNA with the GluB- 

2 promoter, and introduced it into rice seeds. The accumulation 

level of I122M/K124W was similar to that of native β. 

We extracted and purified I122M/K124W and native β from 

individual transgenic rice seeds (Fig. 2B), and subjected them 

to a phagocytotic assay. At first, neutrophils were purified from 

heparinized human blood. Fluorescent beads were opsonized 

with human serum, and these beads were then added into neutrophils. 

The intensity of phagocytosis was shown as the phagocytic 

index, which is presented as the number of incorporated 

fluorescent beads per 100 neutrophils. The phagocytotic 

activity of human neutrophils of I122M/K124W after digestion 

by trypsin was about three times higher than that of native 

β (Fig. 2C). 

Conclusions 

In this study, we succeeded in developing transgenic rice accumulating 

α′ and β. α′ accumulated in seeds about twice as 

much as β. It was presumed that the difference in the accumulation 

level was dependent on whether they interact with glutelin 

through disulfide bonds or not. When we observed these 

transgenic rice seeds by EM, it was suggested that α′ and β 

were transported to the vacuole through the Golgi apparatus. 

In addition, a part of β was transported from the ER to the 

vacuoles, bypassing the Golgi apparatus. 

I122M/K124W was accumulated similarly to native β. 

The phagocytotic activity of I122M/K124W was about three 

times higher than that of native β. 

References 

Aoyama T, Kohno M, Saito T, Fukui K, Takamatsu K, Hashimoto T, 

Yamamoto Y, Hirotsuka M, Kito M. 2001. Reduction by 

phytate-reduced soybean beta-conglycinin of plasma triglyceride 

level of young and adult rats. Biosci. Biotechnol. 

Biochem. 65:1071-1075. 

Becker D, Kemper E, Schell J, Masterson R. 1992. New plant binary 

vectors with selectable markers located proximal to the 

left T-DNA border. Plant Mol. Biol. 20:1195-1197. 

Hara-Nishimura I, Shimada T, Hatano K, Takeuchi Y, Nishimura M. 

1998. Transport of storage proteins to protein storage vacuoles 

is mediated by large precursor-accumulating vesicles. 

Plant Cell 10:825-836. 


Krishnan HB, Franceschi VR, Okita TW. 1986. Immunochemical 

studies on the role of the Golgi complex in protein body formation 

in rice seeds. Planta 169:471-480. 

Manzoni C, Lovati MR, Ggianazza E, Morita Y, Sirtori CR. 1998. 

Soybean protein products as regulators of liver low-density 

lipoprotein receptors. II. α-α′ rich commercial soy concentrate 

and α′ deficient mutant differently affect low-density lipoprotein 

receptor activation. J. Agric. Food Chem. 46:2481- 

2484. 

Maruyama N, Maruyama Y, Tsuruki T, Okuda E, Yoshikawa M, 

Utsumi S. 2003. Creation of soybean β-conglycinin β with 

strong phagocytosis-stimulating activity. Biochim. Biophys. 

Acta 1648:99-104. 

Oparka KJ, Harris N. 1982. Rice protein-body formation: all types 

are initiated by dilation of the endoplasmicreticulum. Planta 

154:184-188. 

Sirtori CR, Lovati MR, Manzoni C, Monetti M, Pazzucconi F, Gatti 

E. 1995. Soy and cholesterol reduction: clinical experience. 

J. Nutr. 125:598S-605S. 

Tanaka K, Sugimoto T, Ogawa M, Kasai Z. 1980. Isolation and characterization 

of two types of protein bodies in the rice endosperm 

glutelin and prolamin. Cereal Chem. 48:169-181. 

Tsuruki T, Takagata K, Yoshikawa M. 2001. Soymetide, an 

immunostimulating peptide derived from soybean β- 

conglycinin, is an fMLP agonist. FEBS Lett. 540:206-210. 

Notes 

Authors’ addresses: Takayasu Motoyama, Nobuyuki Maruyama, 

Takahiko Higasa, Masaaki Yoshikawa, and Shigeru Utsumi, 

Graduate School of Agriculture, Kyoto University, Japan, e- 

mail: mtaka@kais.kyoto-u.ac.jp; Fumio Takaiwa, Department 

of Plant Biotechnology, National Institute of Agrobiological 

Sciences, Japan. 


B.O. Juliano of the Philippine Rice Research Institute gave an 

“Overview of rice and rice-based products.” He pointed out the 

importance of the amylose content of rice as table rice, for example, 

in relation to palatability and glycemic index. For materials 

for baked and steamed products, the milling method was 

shown to be important. He indicated that the effect of food processing 

on protein quality, level of antioxidants, mycotoxin level, 

and acrylamide content has to be monitored. Finally, he concluded 

that the high level of antioxidants in rice bran, oryzanol, 

tocopherols, and tocotrienols, as reflected in high unsaponifiable 

matter in bran oil, contributes, in addition to phytic acid, to the 

hypocholesterolemic effect and other health benefits of full-fat 

bran. 

S. Isobe of the National Food Research Institute, Japan, 

discussed the “Sterilization effect of electrolyzed water on rice 

food.” He pointed out that controlling microorganisms from raw 

materials to products and confirming safety in product distribution 

are most important for enhancing rice consumption in the 

development of new rice products, such as aseptic packaged 

rice and rice cookies. He found that the combination of acidicelectrolyzed 

water and alkaline-electrolyzed water is very effective 

for controlling microorganisms without a change in color and 

pH. He reported the effect of electrolyzed water on preventing 

raw rice from microorganism infestation by an in vitro test using 

Bacillus subtillis spores. 

H.C. Choi of the National Institute of Crop Science, Korea, 

covered the “Current status of varietal improvement and use of 

specialty rice in Korea.” He mentioned various kinds of newly 

developed specialty rice varieties, such as low-amylose dull, 

opaque endosperm, or high-fiber mutants, and high-amylose and 

high-lysine rice in Korea. He explained the relationship between 

the physicochemical properties and processing suitabilities for 

rice snacks, rice noodles, rice bread, and fermented products. 

High-fiber and low-digestible rice was revealed to have a considerable 

dietetic effect and to lower the blood-sugar level. He concluded 

that breeding to improve the utility and nutritional function 

in rice for the future should focus on diversification in the 

morphological, physicochemical, and nutritional characteristics 

of rice grain suitable for processing various value-added rice foods. 

K. Ohtsubo of the National Food Research Institute, Japan, 

discussed “Processed novel foodstuffs from pregerminated 

brown rice by a twin-screw extruder.” He mentioned various kinds 

of Japanese processed rice products, such as retort-pouched 

rice, dried cooked rice, aseptic cooked rice, and pregerminated 

brown rice. He developed a novel foodstuff from pregerminated 

brown rice by a twin-screw extruder. The pregerminated brown 

rice was rich in gamma-aminobutyric acid (GABA). The extruded 

pregerminated brown rice contained more oryzanol, inositol, ferulic 

acid, and dietary fibers than polished rice. Wheat bread prepared 

with 30% of the extruded pregerminated brown rice contained 

more GABA, free sugars, and amino acids than ordinary bread. 

These results showed that a novel foodstuff from extruded 

pregerminated brown rice would be acceptable to consumers or 

the food industry as a promising foodstuff that contains more 

bio-functional components than ordinary rice products. 

Professor R. Hayashi of Nihon University covered “Highpressure 

food processing of rice and starchy foods.” He pointed 

out that high pressure is used for food processing at high temperature. 

When starch is pressurized at 200 MPa or higher, it 

changes its physical properties, including loss of birefringence 

and increases in amylase digestibility and crystalline properties. 

Thus, high-pressure food processing is now applied in rice and 

its related food industry in Japan. He explained the high-pressure 

effects on starches and mentioned the recent development 

of the high-pressure technique in the food industry. 


E. Champagne of the Southern Regional Research Center, 

USDA/ARS, USA, reported on “Developing novel processes for 

incorporating the unique nutritional and functional properties of 

rice into value-added products.” She pointed out that, with more 

efficient production and milling and with value-added applications 

for rice, its components and co-products are necessary to 

prosper in both the domestic and export arena. She highlighted 

novel processes developed by Southern Regional Research Center 

scientists, such as (1) a cost-effective, environment-friendly 

rapid technology for separating rice starch and protein using physical 

means, (2) a dry instantization process for quick-cooking brown 

rice, (3) rice batter that absorbs 60% less oil than wheat batter, 

(4) rice donuts with textural properties of wheat donuts, but substantially 

lower oil uptake, and (5) rice bread developed for preparation 

using a home bread machine. 

Posters included ones such as “Dehydrin proteins in rice 

bran” by M. Momma of the National Food Research Institute, 

“Physicochemical properties of modified flour and its use for processed 

food” by T. Takahashi of the Akita Prefectural Food Research 

Institute, and “Introduction of soybean beta-conglycinin 

into rice to improve nutritional and physiological values” by T. 

Motoyama of Kyoto University. Other topics covered were the 

breeding of novel-pigmented rice, giant-embryo rice, high-yielding 

rice for cattle feed, and new technology for the preparation of 

pregerminated brown rice. 


SESSION 10 

Postharvest technology for efficient processing 

and distribution of rice 

CONVENER: T. Kimura (Univ. of Tsukuba) 

CO-CONVENER: J. Rickman (IRRI)

Postharvest technology for rice in India: a changing scenario 

Pallab Kumar Chattopadhyay 

India produces about 93 million tons of rice, which is about 

one-fifth of world production. India is the second-largest riceproducing 

country in the world. More than 200 commercial 

rice varieties of indica subspecies are grown on an area of 

about 45 million hectares, with 25% irrigated area. Yield averages 

2,086 kg ha –1 (FAI 2003). Rice provides an occupation 

for more laborers than any other industry in India. About twothirds 

of the production is retained by producers for their own 

consumption and for seed. Of the total quantity of rice available, 

about 10% is processed into products such as beaten rice, 

puffed rice, puffed paddy, and deep-fried products such as 

papads and chakli, and fermented products such as idli, dosa, 

and uthapam. India exports rice valued at about US$1.4 billion, 

of which Basmati is $420 million and non-Basmati is 

$980 million (MFPI 2003). 

Development of a modernization program 

In 1955, the government of India set up a committee to examine 

the problems concerning the development of a rice-milling 

industry. It recommended that preference be given to shellertype 

mills over the existing metallic huller. On the basis of 

these recommendations, the government promulgated the Rice 

Milling Industry (Regulation and Licensing) Act and the Rules 

thereof for regulating the industry. In 1963, the government, 

under the Intensive Agricultural Development Program, along 

with the Ford Foundation, undertook a study on the rice-milling 

industry. Seven modern rice mills of rubber-roll sheller 

type were imported from Japan and Germany and were set up 

in seven rice-growing states. The remaining components, 

namely, parboiling plants, mechanical dryers, silos, and mechanical 

handling and conveying equipment, were constructed 

and fabricated in the country. A report submitted in 1968 (MA 

1971) showed that modern mills gave a substantial increase in 

total rice outturn and head-rice yield for both raw and parboiled 

paddy over existing sheller mills and a much higher 

increase over existing huller mills (Table 1). Further, modern 

mills yielded by-products (husk and bran) separately for better 

end uses. Mechanical drying increased yield by 1–2%, and 

the improved technique of parboiling in conjunction with mechanical 

drying increased yield by 2%. With silo storage, estimated 

additional yield ranged from 2% to 4%. 

The government of India recognized that there was vast 

potential for improving the entire postharvest technology of 

the paddy/rice system. Accordingly, the Rice Milling Industry 

(Regulation) Act, 1958, and Rice Milling Industry (Regulation 

and Licensing) Rules, 1959, were amended in 1968 and 

1970, respectively. In the initial phase, a battery of hullers, 

huller-cum-sheller combinations, and shellers in the organized 

sector was brought under the purview of modernization. The 

Table 1. Mean average percentages of additional outturns of total 

rice and head rice obtained in modern mills versus conventional 

mills for raw and parboiled paddy. 

Mean average percentage of additional outturns 

of rice in modern mills versus conventional mills 

Type of paddy 

milled Total rice Head rice 

Sheller Huller Sheller Huller 

Raw 2.5 6.6 6.1 15.1 

(0.8–4.2) a (1.8–12.5) (1.9–12.9) (6.9–24.7) 

Parboiled 0.8 1.6 1.3 4.1 

(0.0–1.3) (0.3–2.5) (0.8–2.5) (1.0–8.5) 

a 

Numbers in parentheses represent the variations of actual additional outturns from 

different varieties of paddy and over different mills. Source: MA (1971). 

single hullers were left out for the pending development and 

supply of modern equipment of the same capacity (250 to 300 

kg h –1 ) in the country. 

Apart from the seven modern rice mills initially set up 

in the pilot study, the Food Corporation of India established 

25 modern rice mills without including silo storage. In addition, 

the cooperative and private sector together established 

55 modern rice mills and modernized 51 existing mills. 

Modernization of huller rice mills 

Over the years, several designs of low-cost mini rice mills became 

available in the country. In July 1976, by an amendment 

to the Rice Mill Industry Rules, provision was made for the 

gradual modernization of single-huller mills by adding rubber-roll 

shellers or centrifugal dehuskers, paddy cleaners, and 

separators. A huller subsidy scheme with 50% of the cost of 

modernization provided to each beneficiary was being implemented 

in six states to encourage the modernization of huller 

mills. 

Modernization of paddy parboiling 

During the 1950s, the Indian Council of Medical Research 

sponsored research at the Central Food Technological Research 

Institute (CFTRI), Mysore, and Jadavpur University, Calcutta, 

for improving the parboiling and drying processes as well as 

nutritional and cooking quality of rice. In 1957-58, the Ministry 

of Food and Agriculture recommended the new technique 

of parboiling developed by CFTRI for popularization. 

Modernization of paddy/rice storage 

In 1965-66, the “Save Grain Campaign” was organized by the 

government, covering 19 states/union territories to popularize 

an effective method of grain storage among farmers, traders, 


Number of modern mills 

30,100 

32,200 

33,557 34,110 34,668 34,668 34,668 35,088 

25,000 

1991 1992 1993 1994 1995 1996 1997 1998 1999 

Year 

Fig. 1. Growth of modern rice mills. Source: MFPI (2003). 

etc. The Indian Grain Storage Institute was established at Hapur, 

with two substations at Ludhiana and Bapatla for conducting 

applied research, development, and training in grain storage 

with the assistance of the United Nations Development 

Programme. Storage structures made of steel, concrete, and 

metal-plastic combinations at the farmers’ level were being 

popularized, initiated through assistance received from the 

United Kingdom and Netherlands Freedom From Hunger Campaign, 

and USAID. 

Training, research, and extension 

The government established in 1970 a full-fledged Rice Process 

Engineering Centre (presently, Post Harvest Technology 

Centre) at the Indian Institute of Technology, Kharagpur, to 

provide training facilities at the Bachelor, Masters, and Ph.D. 

levels, and short-term training courses for rice mill engineers, 

managers, and operators. The Ministry of Food Processing 

Industries (MFPI) has been sponsoring various research and 

development activities through this Centre as well as the Paddy 

Processing Research Centre, Thanjavur. MFPI has also set up 

a technical cell to assist the industry by providing technical 

assistance and consultancy service. In addition, extension and 

training work was undertaken through a network of ten regional 

extension service centers. Grant-in-aid is being provided to 

various institutions to promote research and development, extension, 

and training work on postharvest technology for rice. 

Development of modern equipment 

Realizing that the country would require modern rice-milling 

equipment on a large scale, three manufacturers were initially 

given licenses to manufacture machinery with suitable modifications 

in collaboration with the companies from where seven 

modern mills were imported for the pilot study. 

To encourage the production of good-quality rice mills 

and allied machinery by various manufacturers, four testing 

centers were set up in four different regions of the country to 

undertake testing of mill machinery according to standards 

prescribed by the Bureau of Indian Standards. 

Present position of the industry 

Rice postharvest technology in India has come a long way over 

the past three decades. Now, more than 50% of the overall rice 

production is processed by modern mills, with a steady growth 

in numbers (Fig. 1), about 40% by traditional mills, and about 

10% by hand pounding. There are now 35,088 modern/modernized 

mills (0.5–4 t h –1 ), 4,538 under-runner disc shellers, 

8,385 hullers-cum-shellers, and 91,287 metallic hullers (MFPI 

2003). 

Fifty percent of the total paddy production is parboiled. 

The CFTRI method of parboiling is used mostly by modern 

mills to drastically reduce soaking time and improve rice quality. 

However, the age-old practices of premilling treatments at 

the village level and traditional parboiling methods are still 

being followed in some commercial mills. 

Drying of paddy is carried out by sun drying in open 

drying yards and/or by mechanical drying in L.S.U.-type dryers 

in commercial mills. 

The estimated losses in storage and handling are about 

10%. About 70% of the produce is stored by farmers in small 

indigenous storage structures with some modifications and 

made of locally available materials. The commercial-scale use 

of bag storage of paddy/rice in godowns is common. 

Out of a potential availability of 1 million t of bran oil, 

only 0.5 million t is produced from 3.4 million t of bran, twothirds 

of which is of edible grade and the rest is of industrial 

grade. A part of the edible quality is consumed via root blending 

or hydrogenation. About 70,000–80,000 t of the oil are 

used for direct cooking purposes. De-oiled bran is mainly used 

for cattle feed. 

Rice husks are primarily used as fuel in husk-fired furnaces 

to produce steam through a boiler for parboiling, run- 

Session 10: Postharvest technology for efficient processing and distribution of rice 295

ning the mill using a steam engine and hot air for drying. Producer 

gas is being produced from husks as fuel for a diesel 

engine to run the mill and to generate electricity. Husks are 

also being used commercially to produce furfural. 

At present in India, a large number of trained technical 

personnel are available. Many research workers are actively 

involved in rice postharvest technology research. India manufactures 

and also exports all types of rice-processing machinery 

and equipment. India also provides training to foreign personnel. 

Constraints and concerns 

There have been several problems with the modern large-capacity 

rice mills: the inadequate supply of paddy for continuous 

operation throughout the year, controls and restrictions 

imposed by the government with regard to price, a levy, and 

movement for paddy and rice. The high capital cost for silos 

prohibits the use of integrated units. Because of the unavailability 

of good-quality rice bran and its inadequate timely collection 

from the mills, the use of bran is very poor. A genuine 

need exists for efficient low-cost mini rice mills to replace the 

existing metallic hullers. 

Even though sufficient systematic development has been 

made and experience has been gained, much remains to be 

done for rice postharvest technology to play an important role 

in the proper use of valuable resources and in contributing 

substantially toward Indian economic growth. 

References 

FAI (Fertilizer Association of India). 2003. Fertilizer statistics. New 

Delhi (India): FAI. 

MA (Ministry of Agriculture, Government of India). 1971. Modern 

versus traditional rice mills—a performance study. Delhi (India): 

MA. 

MFPI (Ministry of Food Processing Industries, Government of India). 

2003. Annual report. New Delhi (India): MFPI. 

Notes 

Author’s address: Professor, Post Harvest Technology Centre, Agricultural 

and Food Engineering Department, Indian Institute 

of Technology, Kharagpur 721302, India. 

Development of a far-infrared radiation dryer for grain 

Yasuyuki Hidaka, Kotaro Kubota, and Tomohiko Ichikawa 

Drying, especially for rice, is an important postharvest process 

in monsoon climate areas such as Japan, where there is 

much precipitation during the harvest period. The Japanese 

Ministry of Agriculture, Forestry, and Fisheries has taught farmers 

to dry crops within 8 hours after harvest. 

Currently in Japan, heated-air (HA) drying by burning 

kerosene is regularly used for artificial grain drying. However, 

to maintain grain quality, the combustion heat is mixed with 

ambient air and its temperature is lowered to about 40 to 50 

°C. Heat energy cannot be used effectively because of the need 

to maintain grain quality. Although the HA dryer has evolved, 

its capability is limited from the energy viewpoint. 

The applicability of far-infrared radiation (FIR) to grain 

drying was investigated earlier by Harry and David (1959) and 

then by Faulkner and Wratten (1969). FIR radiation can efficiently 

transfer energy to an object. Moreover, the wavelength 

of FIR is 2.5 to 1,000 µm, and grain has an infrared absorption 

wavelength in this range. For this reason, FIR can be used to 

heat grain efficiently. Heat emission causes very little energy 

loss in transmitting drying energy to the grain. 

In Japan, the possibility of FIR drying for rough rice 

was also evaluated by Bekki (1991) and Matusoka (1990). Itoh 

et al (1994) and Mouri and Huoqing (1996) also explored the 

possible application of FIR to other agricultural products. In 

their studies, they acquired basic drying data for rough rice 

using an electric heater. However, they needed to change the 

dryer structure and modify it to use an FIR electric heater, so 

that it was difficult to develop commercial models. 

We decided to use a design including the FIR body with 

a kerosene burner in the recirculating batch dryer because it is 

widely used in Japan. 

Schematic of the FIR dryer 

The concept of an FIR dryer and an HA dryer is shown in 

Figure 1. The HA dryer heats the air to 40 to 50 °C by mixing 

the flame and the ambient air, and passes the heated air through 

the grain layer. The FIR dryer converted thermal energy by 

heating an FIR body with a kerosene burner. Moreover, exhaust 

heat was reused for drying to increase drying efficiency. 

The FIR body was made from a stainless-steel pipe coated 

with silicone resin. FIR body surface areas were 1.6 to 3.6 m 2 . 

The rate of radiation was 0.9, and the FIR body surface temperature 

was 300 to 500 °C. The center of the wavelength was 

4.3 to 5.2 µm as determined by Wien’s displacement law. An 

electromagnetic wave containing far-infrared radiation is generated. 

These spectra are the absorption band of water. Moreover, 

30–50% of the input energy was converted into FIR. 

The FIR body was useful wherever installed. Two types 

of dryers are used, one with the FIR body installed in the drying 

chamber and the other with it installed in the space be- 


FIR dryer 

Ambient air 

HA dryer 

Ambient air 

Exhaust 

heat 

Heated air 

Fig. 1. Concept of the FIR dryer. 

FIR body 

Grain layer 

tween the rotary valve and the under-screw conveyer, considering 

development costs. 

FIR 

Burner 

Suction fan 

Burner Heated air Grain layer Suction fan 

the ratio that compared FIR and HA represented the rate of 

energy reduction. 

Rough rice was husked by hand before and after drying, 

and the cracked grain was judged visually using a grain scope 

(Shizuoka Seiki Co. Ltd., DC-50). The difference in the cracking 

rate before and after drying represented the increase in 

cracking rate. 

Noise measurements. The A-weighted sound pressure 

level was measured based on JIS Z8731. The sound level meter 

was located 1.0 m from the dryer at a height of 1.2 m. The 

average level of noise for a fixed period of drying time was 

measured. We measured the A-weighted sound pressure level 

when the difference between the background noise and the real 

A-weighted sound pressure level was 10 dB or more. 

Quality tests. The germination tests were performed 

based on the Japanese Food Agency method. The palatability 

tests were performed using the rice dried by HA and FIR. 

Twenty panelists evaluated palatability based on six criteria: 

appearance, flavor, taste, viscosity, hardness, and total evaluation. 

The palatability tests were performed by the Japan Grain 

Inspection Association. 


Materials 

Nine varieties of rough rice (Hinohikari, Koshihikari, 

Akitakomachi, Haenuki, Kinuhikari, Yumeminori, 

Tukinohikari, Akiroman, and Asanihikari) and five varieties 

of wheat (Nourinn61, Kitakami, Chikugoizumi, Chihoku, and 

Bandouwase) were used. 

Initial moisture content (MC) was 18% w.b. to 29% w.b. 

for rough rice and 15% w.b. to 39% w.b. for wheat. Final MC 

was 15% w.b. for rice and 12.5% w.b. for wheat. Grain weight 

was 1.2 to 6 t, with an average of 3.5 t. 

Methods 

Drying tests. To compare drying performance, we prepared 

FIR and HA dryers with identical capacities as in the drying 

test. We used grain from the same field so that drying characteristics 

did not change. When grain was taken into or discharged 

from the dryer, we took 20 samples that were 1 kg per 

sample. And we measured MC and grain quality using these 

samples. 

Drying time was calculated from start time to finish time, 

when the moisture meter attached to the dryer measured 15% 

w.b. The true MC was measured by the air-oven method and 

drying rate was calculated. 

Heated-air temperature, exhaust-air temperature, FIR 

body surface temperature, grain temperature, atmosphere temperature, 

and humidity were automatically measured at 10-min 

intervals during drying using a data logger and thermocouple. 

The amount of water removed was measured automatically by 

load cells in the bottom of a dryer. Fuel consumption was calculated 

automatically by a balance and the amount of electricity 

used was measured by a wattmeter. The moisture extracted 

by 1 kg kerosene and 1 kWh electricity was calculated, and 


Drying tests 

The average drying rate was 0.75% h –1 , the same as for the 

HA dryer. Under some conditions, the drying rate was 1.2% 

h –1 . Furthermore, the increase in cracking rate of rice was satisfactory 

at 2% or less. Although the cracking rate of rice dried 

by HA was increased at drying rates exceeding 1% h –1 , there 

was no increase in cracking rate by FIR drying. We thus demonstrated 

the possibility of high-speed drying using the FIR 

dryer. 

The FIR dryer’s energy efficiency averaged 5.11 MJ 

kg –1 H 2 O, which is 10% better than that of the HA dryer. To 

perform more practical evaluations, specific moisture extraction 

rates by kerosene and electricity were averaged and found 

to be 9.43 kg H 2 O kg –1 and 14.6 kg H 2 O kWh –1 . We could 

thus reduce electricity consumption by 30% and kerosene consumption 

by 10% vis-à-vis the HA dryers (Table 1). 

Noise measurements 

Because the kerosene was burned in the FIR body with a burner, 

combustion noise was low. Moreover, the FIR dryer could 

operate with a 10% reduction in fan RPM compared with the 

HA dryer. Therefore, the A-weighted sound pressure level was 

reduced by an average of 3 dB. 

Drying is actually performed at night in many cases, and 

so a low noise level is highly desirable. 

Quality tests 

The germination rate averaged 98.3%. FIR drying therefore 

did not decrease the germination rate compared with HA drying. 

Palatability testing revealed that rice dried by the FIR 

dryer tended to have greater viscosity and better palatability. 


Table 1. Results of drying tests. 

Item 

Unit 

Rough rice 

Wheat 

FIR a HA FIR HA 

Moisture content w.b. Before 21.7 21.7 21.8 21.8 

After 14.8 14.9 12.4 12.3 

Grain temperature °C 30.5 28.8 33.8 33.0 

Heated air temperature °C 48.4 43.4 52.8 52.7 

Exhaust air temperature °C 29.4 27.6 32.0 31.1 

Ambient air: Temperature °C 20.3 22.3 

Humidity % 79.9 71.3 

Air flow rate m 3 s –1 t –1 0.36 0.51 0.27 0.34 

Drying rate % h –1 0.72 0.69 0.78 0.79 

Consumption of kerosene Kg 27.2 29.2 40.7 43.1 

Consumption of electricity kWh 18.5 25.4 24.0 28.6 

Energy efficiency MJ kg –1 H 2 O 4.98 5.65 5.24 5.71 

FIR/HA ratio % 88.1 91.8 

Specific moisture extraction kg H 2 O kg –1 9.72 8.86 9.14 8.47 

rate by kerosene 

FIR/HA ratio % 110 108 

Specific moisture extraction kg H 2 O kWh –1 14.6 10.4 14.6 12.0 

rate by electricity 

FIR/HA ratio % 140 122 

a FIR = far-infrared radiation, HA = heated air. 

Although the mechanism is not yet perfected, we demonstrated 

the potential of high-quality drying using an FIR dryer. 

Conclusions 

The FIR dryer developed has satisfactory drying, energy efficiency, 

and energy conservation compared with the HA dryer. 

The reduced noise level will be comfortable to work with. In 

addition, palatability tended to improve. Thus, the potential 

for high-quality drying was demonstrated. 

These dryers have been produced with commercial models 

from cooperating research companies since 1998, and the 

34,000 FIR dryers so far produced were distributed among 

individual farmers in Japan. We now have a series of FIR dryers 

with a capacity of 1 to 30 t for a drying facility. 

References 

Bekki E. 1991. Rough rice drying with a far-infrared panel heater. 

JSAM 53(1):55-63. 

Faulkner MD, Wratten FT. 1969. The Louisiana State University 

infrared preheat rice drier. 61st Annual Progress Report, Rice 

Research Station, Louisiana State University Agricultural 

Center, Crowley, Louisiana. 

Harry WS, David WR. 1959. Drying rough rice with infra-red radiation. 

Rice J. June, p 16-38. 

Itoh K, Chung SH. 1994. Drying of agricultural products using long 

wave infrared radiation. Part 1. Fundamental heating characteristics 

of long wave infrared radiation. SASJ 25(1):39-45. 

Matsuoka T. 1990. Drying characteristics of rough rice by far-infrared 

radiation heating. SASJ 21(2):85-93. 

Mouri K, Huoqing L. 1996. Study on the drying of agricultural products 

by using far-infrared ray. Part 1. The performance of the 

drying device and the heating properties of agricultural products. 

SASJ 27(2):57-63. 

Notes 

Authors’ address: Bio-oriented Technology Research Advancement 

Institution (BRAIN), Institute of Agricultural Machinery 

(IAM), 1-40-2 Nisshin, Kita-ku, Saitama, 331-8537 Japan, e- 

mail: yhidaka@affrc.go.jp. 

Acknowledgments: We are very grateful for the cooperation of the 

research companies Iseki Co., Ltd.; Kaneko Agricultural Machinery 

Co., Ltd.; Satake Corporation; Shizuoka Seiki Co., 

Ltd.; and Yamamoto Co., Ltd. 


Status of rice milling and use of by-products 

Naoto Shimizu, Yuji Katsuragi, and Toshinori Kimura 

In addition to energy, materials can be recovered from biomass, 

and it is clear that the world must, from now on, substantially 

increase the use of biomass as a source of both energy 

and materials by optimizing biomass production, conversion, 

and use technologies. 

Rice milling and processing need to be part of this effort, 

given that the rice-milling process generates by-products, 

such as husks, bran, and broken kernels. Here, we report on 

several aspects of rice milling and processing: (1) the use industry 

and new rice products, (2) the application of the dosa 

fermentation process to indica-type rice use, (3) a finishing 

system for manufacturing prewashed rice, and (4) an overview 

of rice milling and processing in developing countries. 

The use industry and new rice products 

Figure 1 presents uses of rice and by-products in the rice industry. 

In Japan, 1,200,000 tons of rice are used for processing 

raw materials. This is about 12% of the total consumption. 

The main application is for sake (600,000 t), rice confectionary 

(200,000 t), miso (180,000 t), and other purposes (200,000 t). 

The product of processed rice consists mainly of whole-milled 

rice grain and its directly processed materials. Sake, miso, 

packed rice-cake, and arare are produced directly by kojimaking 

and rice-cake making from whole-milled rice grain. 

Flour and starch-making rarely occur by milling and wet-milling 

such as are done with wheat and maize. A very small amount 

of rice flour and starch has been distributed and used. 

Paddy 

Parboiling 

Parboiled rice 

Husking process 

Brown rice 

Rice milling 

Husk 

SteamingBrewing 

Pressurization/steamingDrying 

Roasting 

Rice bran 

Sake 

NEO-unpolished brown rice 

Brown rice tea 

Grading 

Broken rice 

Animal feed, for beer, noodles, confectionery 

White rice 

Prewashed rice 

Enriched rice-vitamin, mineral 

Steaming 

Fermentation 

Brewing 

Sake, beer, vinegar, alcohol 

Roasting 

Roasted rice 

For milling 

For confectionery 

Puffing 

Puffed rice 

Pressurization puffing 

SteamingFlaking 

Puffed rice crackers 

Rice flakes 

Roasting 

Drying 

Flour milling 

Pulverize 

Flour for confectionery 

Flour for confectionery 

Flour milling 

CookingFreezing 

Drying 

Solidifying 

Steamed rice 

For bread, dosa 

Frozen food 

Instant rice 

Sushi/rice, ball/rice, burger 

Fig. 1. Uses of rice and by-products in the rice industry. 


Rice grits and their processing by wet-milling 

Rice grits or broken rice kernels are used for brewing. Generally, 

wet-milled rice flour is superior to dry-milled flour for 

making baked products. Adjuncts (usually 20–40% of the total 

ingredients) are used to obtain fermentable sugars at a lower 

cost, and play a very important and specific role in giving a 

characteristic quality to beer to attract consumers. All-malt beer 

is richer in flavor and has more fullness of body and a darker 

color. On the other hand, beer with adjuncts has a lighter color 

and taste, so that consumers drink more because of its less 

satiating taste. Also, beer with adjuncts has a higher colloidal 

stability because of its lower nitrogen content (Yoshizawa and 

Kishi 1985). 

Application of the dosa fermentation process 

to indica-type rice use 

Dosa is a spongy pancake that has a desirable flavor, texture, 

and sourness. The Japanese are receptive to the characteristics 

of dosa. A new food product is expected to have a desirable 

flavor, texture, and function through fermentation. We examined 

the lactobacillus identification and physical-biochemical 

parameters of dosa and tofu-by-product-additive dosa batter. 

Dosa and tofu-by-product-additive dosa were measured using 

the creep compliance test, and antioxidant activity was measured 

by the DHHP method. The fermentation characteristics 

and texture of tofu-by-product-additive dosa (additive ratio 

less than 60%) were similar to those of dosa produced by conventional 

fermentation. Tofu-by-product-additive dosa had 

improved antioxidant activity, although the tofu-by-productadditive 

ratio was more than 60%, yielding a poor texture 

(Shimizu et al 2003). 

The finishing system 

Facility for manufacturing prewashed rice 

A new manufacturing facility aims to polish conventional milled 

rice to a condition good enough to be cooked without prior 

washing. Today, facilities for manufacturing prewashed rice 

can be broken down broadly into three types: (1) a dry-polishing 

and finishing system, (2) a wet-milling and finishing system, 

and (3) a special processing and finishing system. Facilities 

of all these types finish the milled rice to a prewash rice 

state by removing the surface portion of rice by 1–2% by weight 

of the rice. These facilities are designed to process japonica 

rice. 

The dry-polishing and finishing system 

Kapika and Rifure are major ones of this type. Kapika has a 

long roller, 1.7 times longer than the roller of an ordinary friction 

rice-milling machine, with three extrusions on the mixing 

roller. The screen has a dodecagon shape. The milled rice flows 

downward from the top of the machine and rice grains are 

milled by grain-to-grain friction. The machine comes in four 

models, with capacities of 0.3, 1, 1.8, and 3 t h –1 . 

Rifure was developed by converting a rice-milling machine 

into a dry-polishing machine. The machine has brush 

knives, on which are plated fine nylon fiber 0.275 mm across, 

containing 800-mesh carbon silicate particles at 15% weight. 

Rice is fed into the lower portion of the machine and is pushed 

upward to the roller, where rice is polished by the brushes 

embedded on the roller. Rifure comes in two capacities, 1 and 

1.5 t h –1 , the former in two lines and the latter in three lines. 

The wet-milling and finishing system 

The Super jiff rice (SJR) consists of two stages of processing, 

the first being wet-milled and the second being finishing. The 

milled rice is fed from the top of the wet-milling section, or 

the first stage, and rice flows downward to the milling roller, 

where the rice is wet-milled. The retention time of the rice in 

the first stage is about 3 seconds. Water is added at about 50% 

by weight to the feed rice, and heated to 20 to 30 °C by the 

heat exchanger on the water tank. Subsequently, rice undergoes 

dehydration by centrifugation, and then is fed into the 

finishing process, where the rice is finished by air heated to 

about 45 °C. The water discharged from the wet-milling section 

is sent to a dryer, where water is evaporated by the heated 

air from the boiler, which then leaves flakes of suspended solids. 

The special processing and finishing system 

The Neo Tasty White Process (NTWP), a new special processing 

and finishing system, has been adopted by large rice 

milling plants since 2000. This system consists of three sections: 

the moist pressurized milling section, the heated adhesion 

low-pressure mixing section, and the drying and separation 

system. The moist pressurized milling section mixes the 

grain after adding water at a rate of 5% of the milled rice. The 

retention time of milled rice in this section is about 10 seconds. 

The subsequent secondary process in the heated adhesion 

low-pressure mixing section adds tapioca starch, heated 

to about 85 °C, to the grain at a rate of 50% of the milled rice. 

Here, tapioca starch acts as a heated adhesive agent. The tapioca 

starch removes the remaining bran layer from the milled 

rice while being mixed with the latter. The retention time of 

the milled rice in this section is about 10 seconds. The tertiary 

process separates the tapioca starch agent from the bran by the 

bag filter and the separator installed separately, and tapioca 

starch is recycled back to the system after being heated by a 

boiler. 

Overview of rice milling and processing in the Philippines 

The following factors have been highlighted by many reports 

as the main causes of loss through the various stages from harvest 

to consumption. 

1. Characteristics of the rice variety. High-yielding varieties, 

especially IR64, are the main rice crop these 

days. These varieties are characterized by a high shattering 

loss because of the delayed harvesting and 

threshing time. 


2. Equipment and facilities. The use of machinery is still 

very limited at all stages. The lack of dryers among 

farmers especially causes a considerable loss of paddy. 

3. Unsuitable milling equipment. This is the cause of 

low yield among custom mills in rural villages. 

4. Inadequate facility management. Poor management 

of small and medium storage facilities in particular 

results in a considerable loss to vermin and spoilage 

by microorganisms. 

5. Socioeconomic factors. Poor roads, including farm 

roads, cause losses during handling and transportation 

(Yoshizaki 2002). 

Table 1 shows the level of postproduction losses in the 

Philippines. A postproduction assessment was conducted by 

NAPHIRE (present-day BPRE) in 1994-95 with the cooperation 

of 13 universities using a common survey methodology. 

The total loss (percentage) is therefore considered to be higher 

than these numbers indicate. 

References 

Shimizu N, Seimiya H, Toyoki A, Kimura T. 2003. Application of 

the dosa fermentation process to tofu-byproduct utilization in 

food industry. 5th International Food Convention, 5-8 December 

2003. Mysore (India): Central Food Technological 


Yoshizaki S. 2002. Report on the local survey on rice distribution 

and management in the Philippines. ODA Project for Improving 

Rice Distribution in Asia. Japan Grain Inspection Association. 

p 9-10. 

Yoshizawa K, Kishi S. 1985. Rice in brewing. In: Juliano BO, editor. 

Rice: chemistry and technology. St. Paul, Minn. (USA): 

American Association of Cereal Chemists, Inc. p 638. 

Table 1. Postproduction losses in the Philippines (%). 

Process 1994-95 1988 1974 

(NAPHIRE) (NAPHIRE) (IRRI) 

Harvesting Trace–4.85 (1.81) a Trace–4.8 1–3 

Piling Trace–1.77 (0.54) Trace–1.0 2–7 

Threshing and winnowing 0.04–5.09 (2.17) 0.1–5.4 2–6 

Drying 0.74–8.70 (4.50) Trace–0.7 1–5 

Storage 0.35–5.20 (2.72) 2.6–5.0 2–6 

Milling 0.00–6.33 (3.10) 6.3–8.3 2–10 

Total loss 1.13–31.94 9.1–23.0 10–37 

Average 14.84 16 23.5 

a Numbers in parentheses show the average. 

Notes 

Authors’ addresses: Naoto Shimizu and Toshinori Kimura, Graduate 

School of Life and Environmental Sciences, University of 

Tsukuba, Tennodai 1-1-1, Tsukuba 305-8572, Ibaraki-ken, 

Japan; Yuji Katsuragi, Japan Rice Millers Association, 

Kojimachi 3-3-6, Chiyoda-ku 102-0083, Tokyo, Japan, e-mail: 

shimizu@sakura.cc.tsukuba.ac.jp. 

Acknowledgments: We express our gratitude to the Ministry of Agriculture, 

Forestry, and Fisheries, Japanese Grain Inspection 

Association, and KOKKEN. Especially, the overview of rice 

milling and processing in the Philippines is a result of the 

ODA project on “Improving Rice Distribution in Asia.” 

Advanced application technology of rice bran: 

preparation of ferulic acid and its applications 

Hisaji Taniguchi, Eisaku Nomura, and Asao Hosoda 

The production of brown rice in the world is about 560 million 

tons a year. Japan produced about ten million tons. Brown 

rice is usually polished before eating, and 10% of its weight is 

discharged as rice bran. Thus, 1 million tons of rice bran are 

discharged in Japan. About 40% of the rice bran is used to 

produce rice edible oil in Japan, by extraction. In the course of 

production of rice edible oil, rice bran pitch is discharged. This 

is a dark and viscous oil and it is a waste material. We prepared 

ferulic acid from rice bran pitch (Taniguchi et al 1994). 

Ferulic acid was produced from petroleum before. Petroleum, 

however, is predicted to be used up in the future, after about 

40 years. So far, the cost of ferulic acid has been about 

US$2,500 per kg. No one could use this as a chemical material 

because of its expensiveness. In the method we have developed, 

the cost of ferulic acid is about $100 per kg. Thus, 

ferulic acid has become a very important chemical material. 

The carbon resource of ferulic acid is carbon dioxide in the 

air. Further, ferulic acid will be obtained forever as long as 

humankind keeps on producing rice. Therefore, ferulic acid 

can become a renewable resource of the future chemical industry. 

This development is an environmentally friendly one. 

Figure 1 illustrates the concept of renewable resource usage 

with a system flowchart of the organic chemical industry using 

ferulic acid obtained from rice bran. 

Preparation of ferulic acid 

Ferulic acid and rice bran pitch 

It is well known that ferulic acid can be prepared by the condensation 

reaction of vanillin with malonic acid (Adams 1942). 


Sun 

Renewable resource 

Light 

Rice plant Brown rice Rice bran Rice salad oil 

Synthetic methods 

— Organic chemistry 

— Biochemistry 

Evaluation studies 

Cancer prevention 

Antioxidant activity 

Rice bran pitch 

CO 2 , 

water 

Ferulic acid 

Preparation of new 

compounds 

Ferulic acid is a starting 

material 

Compounds 

to maintain health 

Fossil resource 

(Petroleum-coal) 

Consumers 

Organic chemical industry 

Public demand 

Ecorecycle 

CO 2 , water 

Fig. 1. A system flowchart for the organic chemical industry using ferulic acid obtained from rice bran. 

This produces ferulic acid at a high yield, but takes as long as 

3 weeks. In addition, the product is a mixture of trans- and 

cis-isomers. A vigorous search has therefore been made for a 

commercial method for manufacturing high-purity trans-ferulic 

acid. 

In the manufacture of rice edible oil, a blackish brown 

waste oil having a high viscosity, called rice bran pitch, an 

alkaline oil cake rich in oil components, called soap stock, 

and a by-product rich in crude fatty acids, called dark oil, are 

discharged. These waste materials are known to contain useful 

components but they have generally been disposed of as useless 

industrial by-products, for example, by burning, because 

of the lack of techniques for effective use. 

We have investigated the components of rice bran pitch 

and found by chromatography that about 30% by weight is γ- 

oryzanol itself. Because other components are contained in 

large amounts, it is very difficult to recover γ-oryzanol alone, 

leading to a high manufacturing cost. Thus, it is economically 

not feasible to recover γ-oryzanol from rice bran pitch. However, 

hydrolysis under certain conditions must make it possible 

for us to produce ferulic acid, as illustrated in Figure 2. 

Yet, because γ-oryzanol was included within various oil 

components of rice bran pitch, hydrolysis under usual conditions 

did not proceed at all. It should be noted that rice bran 

pitch is quite immiscible with water so that a suitable alcohol 

must be applied as a solvent. With a simple alcohol having a 

small number of carbon atoms, which can be infinitely dissolved 

in water, the alkaline compound and rice bran pitch can 

also be dissolved without difficulty. 

In our procedure, hydrolysis was carried out at 90– 

100 °C for 8 hours under an atmospheric pressure, resulting in 

crude ferulic acid with a purity of about 70–90%. The crude 

ferulic acid was purified by the recrystallization from a cosolvent 

of ethyl alcohol and water. Ferulic acid was obtained 

in a 100% yield based on γ-oryzanol that existed in rice bran 

pitch (Taniguchi et al 1999). 

EGMP and ferulic acid, which act as anticarcinogens 

The synthesis of EGMP 

It is known that many kinds of chemopreventive agents are 

present in a diverse range of edible plants and herbs. It is difficult, 

however, to isolate or extract chemically pure substances 

from plants because these substances are present in very small 

amounts in plants. 

However, ferulic acid can now be obtained easily in a 

large amount from rice bran. This development prompted us 

to synthesize potential cancer chemopreventive agents. On 

searching for an appropriate target compound that should be 

synthesized by the reaction of ferulic acid with other reagents, 

we found that aurapten, a citrus coumarin, is an excellent cancer 

chemopreventive agent. Recently, Koshimizu et al (1997) 

have isolated aurapten from citrus natudaidai (natsumikan in 

Japanese) and showed that aurapten inhibited 12-Otetradecanoylphorbol-13-acetate 

(TPA)-induced skin tumor 

promotion in mice. 

Aurapten consists of 7-hydroxycoumarin and a geranyl 

group. 7-hydroxycoumarin has one phenolic hydroxyl group, 

one carbon-carbon double bond, and one carbonyl group in its 

molecule. Ferulic acid also has these functional groups in its 

molecule. Therefore, we synthesized many compounds in which 

ethyl 3-(4-geranyloxy-3-methoxyphenyl)-2-propenpate 

(EGMP) was contained, resembling the chemical structure of 

aurapten using ferulic acid. Especially, for EGMP, the geranyl 

group is attached to the phenolic hydroxyl group of ethyl 

ferulate. 


Extraction 

Rice edible oil 

Brown rice 

Rice bran 

Rice bran pitch 

COOH 

HO 

H 3 CO 

COO 

Ca. 30% 

MOH 

ROH 

OH 

OCH 3 

γ-oryzanol 

Ferulic acid 

Fig. 2. Preparation of ferulic acid. 

Anticolon carcinogenesis of ferulic acid and EGMP 

Among compounds we synthesized, ferulic acid itself and 

EGMP showed anticarcinogenesis. The anticarcinogenesis of 

these compounds has been studied by Tsuda et al (1999). 

The inhibitory influence of ferulic acid and EGMP on 

the postinitiation stage of azoxymethane (AOM)-induced colon 

carcinogenesis was investigated in male F344 rats given 

two s.c. injections of AOM (15 mg kg –1 body weight) during 

week 1. Diets containing ferulic acid or EGMP at doses of 

0.1% or 0.2% were then fed for 3 weeks from weeks 2 to 5, 

when the animals were sacrificed. The numbers of aberrant 

crypt foci (ACF) and aberrant crypts (AC) per rat in the group 

given 0.2% ferulic acid decreased significantly (P

is an excellent UV absorbent, it is not stable at high temperatures. 

Thus, we have developed a facile preparation method of 

the ferulic acid dimer, that is, bis-4,4′-(2-ethoxycarbonyl-1- 

ethenyl)-2,2′-methoxyphenoxymethane, which can become a 

better UV absorbent than available UV absorbents such as 

benzophenone and salicylic acid. The compound absorbs light 

at about 260–370 nm and exhibits two absorption maxima 

(λ max ) at 321 nm (ε max : 3.28 × 10 4 ) and 292 nm (ε max : 3.21 × 

10 4 ). The decomposition temperature is 359 °C. This temperature 

is far higher than that of ferulic acid, which decomposes 

a 

t 

176 °C. Based on these results, the compound is able to become 

a thermally stable absorbent of UV light for plastics. 

Other applications 

Ferulic acid is used as the starting material for the preparation 

of reagents having antibacterial activity, fragrances, food additives, 

control of germination, and cosmetics (Taniguchi at al 

2003). 

References 

Nomura E, Kashiwada A, Hosoda A, Nakamura K, Morishita H, 

Tsuno T, Taniguchi H. 2003. Synthesis of amide compounds 

of ferulic acid, and their stimulatory effects on insulin secretion 

in vitro. Bioorg. Med. Chem. 11:3807-3813. 

Taniguchi H, Hosoda A, Tsuno T, Maruta Y, Nomua E. 1999. Preparation 

of ferulic acid and its application for the synthesis of 

cancer chemopreventive agents. Anticancer Res. 19:3757- 

3762. 

Taniguchi H, Nomura E, Hosoda A. 2003. Useful application of ferulic 

acid produced from rice bran. Yukigouseikagakukyoukaishi 

61:310-321. 

Taniguchi H, Nomura E, Tsuno T, Mimami S, Kato K, Hayashi C. 

1994. Method of manufacturing ferulic acid. US Patent No. 

5,288,902. Japanese Patent No. 2095088. 

Tsuda H, Park CE, Takasuka N, Baba-Toriyama H, Sekine K, Moore 

MA, Nomura E, Taniguchi H. 1999. Influence of ethyl 3-(4- 

geranyloxy-3-methoxyphenyl)-2-propenpate (EGMP) on early 

stage colon cacinogenesis in rats treated with azoxymethane 

(AOM). Anticancer Res. 19:3779-3782. 

Tsuno T, Maruta Y, Hosoda A, Nomura E, Taniguchi H. 2001. Preparation 

of a novel thermally stable UV absorbent from natural 

resources. ITE Letters on Batteries, New Technologies & 

Medicine 2C6:808-811. 

Notes 

New movements regarding the safety of rice 

and residual agrochemical inspection 

Yukio Hosaka 

Authors’ address: Industrial Technology Center of Wakayama Prefecture, 

60 Ogura, Wakayama 649-6261, Japan, e- 

mail:taniguti@wakayama-kg.go.jp. 

Recently, there have been new movements in the Japanese rice 

market to inspect residual agrochemicals before harvesting, 

that is, the crops are checked for residual agrochemicals for 

each farm or each harvesting lot. If agrochemicals are not detected, 

the crop will be harvested and sent to a large-scale drying 

facility, but, when agrochemicals are detected, harvesting 

will be postponed until the sun and rain dissolve the chemicals, 

and the crop will be harvested later, after another residual 

agrochemical inspection. 

The reason for this preharvest inspection is that, when 

crops are sent to a large-scale drying facility, they are mixed 

with other crops within a storage silo. If a part of the crop has 

residual agrochemicals, the whole grains will be contaminated 

by the agrochemicals and it will be impossible to separate sound 

grains and contaminated grains afterward. 

Raw paddy is collected from standing plants and threshed 

by simple threshing equipment to make a 130-g sample, and 

then taken to a large-scale drying facility to be processed with 

a test dryer and test husker to make brown rice, 25 g of which 

are used for checking residual agrochemicals. This method was 

introduced to all of Fukuoka Prefecture in Japan starting in 

2004, under the guidelines of JA Central Fukuoka. 

Background of the new movements 

I will explain the background of these new movements. For 

each agrochemical, the maximum permissible level is determined 

by a law, which is set based on studies of the medical 

effect of agrochemicals on the human body. However, in the 

market, it is not so simple to judge that residual agrochemicals 

under the standard are acceptable and that residual agrochemicals 

over the standard are not. 

Suppose that there are two rice products, one of which 

is almost certain not to exceed the legal level of residual agrochemicals 

from past experience and the way of cultivation but 

not confirmed with measuring equipment, and the other is 

checked for confirmation. Which rice will customers buy if 

other commercial factors such as price, taste, and appearance 

are the same At least, in a culturally mature society, the latter 

will be chosen. Or, suppose two samples are checked by an 


ACA2000 analyzing flow 

Activate toxicity 

of enzyme 

Add enzyme 

Preprocessing 

2–3 hours 

ChE 

Grind Shake Filter Column Evaporate 

chromatography 

Dissolve 

Measurement 

Acetylcholine 

Add substrate 

15 min 

Measuring principal enzyme inhibition method 

H 2 O 

ChE 

Agrochemical 

2H 2 O 

(Oxidant) 

POD 

(Reductant) 

M 

(Red.) 

e – 

Current detection 

Choline ChOD H 2 O 2 

H 2 O + O 2 

Fig. 1. Measuring procedures. 

POD 

(Red.) 

M 

(Ox.) 

M: electron-transfer 

Electrode mediator 

POD: peroxidase 

agrochemical inspector, one of which has near but under the 

regulatory level of residual agrochemical, and the other contains 

no residual agrochemicals. As before, if other factors are 

the same, the latter rice will be chosen. Even when the level of 

residual agrochemicals is under the standard, the product’s 

value still differs widely in human society because of actually 

checking the residues or confirming whether the level of residue 

is near the standard or almost zero. Therefore, the issue of 

residual agrochemical in the recent market is not that of a legal 

or a medical problem, but more of a cultural or human 

nature matter. 

In recent rice trading in the market, this cultural or human 

nature aspect is claimed in competitions. This is clear 

because agrochemical-free production is promoted in the market. 

Another reason that rice producers are becoming sensitive 

to residual agrochemicals is that, if their rice products are 

found to contain residues over the standard, because of the 

abovementioned cultural and human nature aspects, the brand 

name could suffer critical damage through news reports. 

The new movements regarding residual agrochemical inspection 

started from this background, and are attempting to 

achieve a higher level of residue-free rice production. 

A new residual agrochemical detector to study the trend 

Because of this new trend, we need to check residual agrochemicals 

for each farm or harvesting lot in a short period of 

time; thus, a large number of agrochemical detectors must be 

placed among rice-producing areas. Therefore, a low-cost detecting 

system is indispensable. 

Moreover, it is also requested that the system be able to 

process a large amount of samples in a short period of time, 

and that it produce test results rapidly, since the test is conducted 

right before harvesting. 

Also, the system must be able to check not only that residues 

are under the regulatory level, but it must have a high 

sensitivity to confirm that residual agrochemicals do not really 

exist. To meet these demands, we have developed a measuring 

system using the enzyme inhibition method. 

Many agrochemicals have the effect of inhibiting the 

neural transmission of insects, so this system measures the 

degree to which the components of agrochemicals inhibit cholinesterase, 

an important factor for neural transmission function. 

Acetylcholinesterase (ACHE) is known as a hydrolytic 

enzyme of excitatory neurotransmitter acetylcholine that exists 

widely from invertebrates to Pisces and mammals, and 

activity is inhibited by organophosphorus compounds and carbamate 

compounds. 

Figure 1 shows the measuring procedures. First is a preprocessing 

to collect the possibly contained residual agrochemicals, 

and then the toxicity of agrochemicals to the enzyme is 

activated before adding the enzyme cholinesterase. After adding 

the enzyme, acetylcholine is added as a substrate. This 

acetylcholine is transformed to choline by the effect of cholinesterase. 

The amount of the generated choline is measured 

by transforming the choline into hydrogen peroxide and check- 


Measuring principle 

A method to measure the amount of agrochemicals from the enzyme inactivities by 

agrochemicals = enzyme inhibition method 

Reference 

Current/nA 

300 

250 

The generation of H 2 O 2 is fast since 

cholinesterase is intact 

Rapid increase of electric current 

Without agrochemical 

H 2 O 2 

Acetylcholine 

ChE 

Acetylcholine 

Choline 

ChE Agrochemical 

ChOD 

Choline 

200 

V R 

H 2 O 2 

ChOD 

150 

V S 

Sample 

100 

The generation of H 2 O 2 is slow since 

cholinesterase is affected 

Slow increase of electric current 

50 

With agrochemical 

0 

2 3 4 5 6 7 8 

Time (min) 

Fig. 2. The measuring principle. 

ing by a hydrogen peroxide sensor. The resulting output is in 

the form of an electrical current. During this process, the 

amount of generated choline is influenced by the inhibition of 

the activity of cholinesterase by agrochemicals, which in turn 

affect the amount of hydrogen peroxide and change the electrical 

current. The amount of electrical current is determined 

by the degree to which residual agrochemicals inhibit the effect 

of cholinesterase. Therefore, the electrical current is affected 

by the amount of agrochemicals, so the residual agrochemicals 

can be measured by using a hydrogen peroxide sensor. 

To be more specific, as shown in Figure 2, if there are 

no agrochemicals, the generation of hydrogen peroxide is fast, 

so the increase in electrical current is faster. On the other hand, 

if there are residual agrochemicals, the generation of hydrogen 

peroxide is slow, thus the increase in electrical current is 

slower. Therefore, the amount of residual agrochemicals can 

be measured by the differential V R or V S . This method cannot 

specify the type of agrochemical component, but it can detect 

the agrochemicals that inhibit the cholinesterase as a whole. 

In the case of DDVP (dichlorovinyl dimethyl phosphate), 

which is an organophosphorus insecticide, sensitivity is high 

enough compared with the legal standard. 

Therefore, this method can be considered as a high-accuracy 

screening method. According to data of several years 

ago, residual agrochemicals are actually detected in one out of 

300 samples in Japan, but they contain only a minute amount 

of agrochemicals and the residues barely exceed the legal standard. 

In the conventional legal method, all 300 samples must 

be checked by liquid chromatography and mass spectrometer, 

etc. However, this legal method is high in cost and the measuring 

efficiency, the number of samples that can be checked per 

day, is low. What if screening tests are done by using our simple, 

inexpensive, high-efficiency and low-measuring-cost method, 

and the legal method is performed only on the samples that are 

picked up by this screening test If the legal method detects 

one out of 300 samples, our high-accuracy screening will detect 

about two samples that contain residual agrochemicals. 

Therefore, while the conventional method needed the inefficient 

legal method to be performed on all 300 samples, this 

high-accuracy screening allows you to use the legal method 

on only two samples. The measuring efficiency and the cost 

for measuring residual agrochemicals are significantly improved 

in this total system with the new technology and legal 

method combined. 

The idea of screening before the legal method is not new; 

however, the point is how far you can reduce the number of 

samples using the screening method. If the degree of narrowing 

down by a screening is small, and the dependence on the 

legal method is still high, the screening will make the measurement 

inefficient. However, because of its high measuring 

sensitivity, our new screening method can significantly reduce 

the number of samples to be checked by the legal method, and 

can well be called a high-accuracy screening method. 

The theory of agrochemical measurement by the enzyme 

inhibition method existed from before, although this may be 

the first practical system that allows for a high-accuracy measurement. 


Furthermore, our newly developed system can be called 

an expanded enzyme inhibition method, which means that it 

uses not only the traditional cholinesterase inhibition theory, 

but also tries to apply a new principle of transforming the agrochemical 

component that originally did not inhibit enzyme 

activity to inhibit the enzyme, and widens the variety of agrochemicals 

that can be measured. Because of the development 

based on this new principle, some of the agrochemicals that 

could not be detected by conventional methods now can be 

detected by the system, and, furthermore, not only insecticides 

but a part of the bactericides can also be measured. We will 

make further progress in this area, while also considering the 

use of other enzymes. 

Notes 

Author’s address: e-mail: y-hosaka@satake-japan.co.jp. 

Impact of infrastructure on profitability and global 

competitiveness of rice production in the Philippines 

Rowena G. Manalili and Leonardo A. Gonzales 

One of the constraints to the growth of Philippine rice production 

is the lack of appropriate and adequate infrastructure. This 

includes marketing and distribution infrastructure such as farmto-market 

roads; limited access to credit, processing, and storage 

facilities; and the lack of effective irrigation systems. Several 

policy studies had pointed out the importance of these 

facilities, but studies that look into the impact of infrastructure 

on profitability and global competitiveness are still lacking. It 

is therefore important to evaluate and measure the impact of 

irrigation and road infrastructure on rice profitability. 

Our study evaluated the impact of infrastructure (roads 

and irrigation) development on farm productivity, rural income, 

technology adoption, and transaction costs in rice farming, and 

compared the profitability and global competitiveness of rice 

farmers who have access to good infrastructure and those who 

do not. 

Methodology 

The study used the 1996 wet season (WS) and 1997 dry season 

(DS) farm-level and barangay-level data of the Regular 

Monitoring of Rice-Based Farm Household Survey (RBFHS) 

project of the Philippine Rice Research Institute’s (PhilRice) 

Socioeconomic Division (SED) and the Bureau of Agricultural 

Statistics (BAS). The farm-level data included production 

costs and returns and socioeconomic and technological 

profiles of 2,500 rice farmers from 281 barangays (villages) 

of 30 selected provinces in the Philippines. Sample barangays 

were classified according to their state of road infrastructure 

development. A barangay with good infrastructure was characterized 

by its year-round accessibility owing to asphalt or 

concrete roads and sufficient transport facilities. On the other 

hand, a barangay with poor road infrastructure was characterized 

by a lack of all-weather roads, making it barely accessible, 

and having insufficient transportation vehicles. 

The methods employed in the analysis were the costs 

and returns analyses, competitive advantage analysis using do- 

mestic resource cost (DRC), and the seemingly unrelated regression 

(SUR) analysis. 

The rice-based farm households across ecosystems, seasons, 

and infrastructure with yield as the major indicator were 

regressed to measure the impact of infrastructure indicators 

on total productivity, output prices, input prices, technology 

adoption, and net farm income. The models were estimated 

using the SUR analysis. The models were as follows: 

1. Output = f (production inputs, infra indicators) 

2. Output price = f (yield, infra indicators) 

3. Input price = f (inputs, infra indicators) 

4. Technology adoption = f (inputs, infra indicators) 

5. Farm income = f (inputs, infra indicators) 

The explanatory variables for the regression model were 

the inputs of production that include farm size (ha), seed use 

(kg ha –1 ), fertilizers—nitrogen, phosphorus, and potassium (kg 

ha –l ), pesticides (active ingredients in kg), and preharvest labor 

(person-days), as well as infrastructure indicators such as 

road classification, distance and travel time to major wholesale 

markets, transport cost of rough rice from farm to major 

wholesale market, availability of transportation vehicles, number 

of input/output dealers, number of business establishments, 

and number of communication facilities. 


Yield and profitability 

Irrigated versus nonirrigated farms. Table 1 shows the yield 

performance and relative profitability of farmers from irrigated 

and nonirrigated environments. Data in both seasons showed 

a substantial yield advantage of irrigated over nonirrigated 

farms. The average yield for irrigated farms during the WS 

was 3.54 t ha –l , whereas the yield for nonirrigated rice was 

2.55 t ha –l . Similarly, in the DS, yield of irrigated rice farms 

averaged 3.66 t ha –l , while yield of nonirrigated farms averaged 

2.34 t ha –l , or around a 1 t ha –l (56%) yield difference. 


Table 1. Average costs and returns of rice production in the Philippines, 1996 

WS and 1997 DS by agroecological system. 

Item Wet season Dry season 

Irrigated Nonirrigated Irrigated Nonirrigated 

Yield (t ha –1 ) 3.54 2.55 3.66 2.34 

Price of paddy ($ t –1 ) 290 290 300 300 

Gross returns ($ ha –1 ) 1,027 740 1,098 702 

Production costs ($ ha –1 ) 

Seeds 47 43 47 42 

Fertilizer 86 60 75 52 

Pesticides a 50 47 48 46 

Irrigation fee 35 – 38 – 

Transportation 3 3 4 3 

Fuel and oil 18 11 20 12 

Labor costs 324 328 343 309 

Depreciation 23 19 23 18 

Land (rental value) 27 14 27 15 

Other costs b 38 33 30 25 

Total production costs ($ ha –1 ) 651 558 655 522 

Net farm income ($ ha –1 ) 375 182 443 180 

Net profit-cost ratio 0.58 0.33 0.68 0.35 

Cost per ton ($) 184 219 179 223 

a Includes herbicides, insecticides, fungicides, molluscicides, and rodenticides. b Includes food costs, 

interest expenses, and land tax. Official exchange rate: P26.30 = US$1. 

Source: Department of Agriculture-PhilRice Rice-Based Farm Household Survey, 1997. 

Costs and returns analysis showed that irrigated farms 

were more profitable than nonirrigated farms in both seasons. 

The higher production cost incurred by irrigated rice farms 

was offset by their higher yield gains. Net farm income realized 

from irrigated farms was higher at US$375 ha –l for the 

WS and $443 ha –l for the DS versus $182 ha –l for the WS and 

$180 ha –l for the DS obtained from nonirrigated farms. Moreover, 

net profit-to-cost ratios were better for irrigated farms. 

Good versus poor infrastructure. Table 2 shows the comparative 

productivity and profitability of farmers with access 

to good and poor road infrastructure. Rice yield of farmers on 

irrigated farms with access to good road infrastructure was 

higher at 3.61 t ha –l for the WS and 3.77 t ha –l for the DS than 

that of farms with poor infrastructure at 3.31 t ha –l for the WS 

and 3.24 t ha –l for the DS. Results of the analysis showed that, 

except for the nonirrigated farms during the DS, farms with 

good infrastructure appeared to be more profitable as indicated 

by their higher gross returns, net income, and net profitto-cost 

ratios. Net incomes from farms with good road infrastructure 

were higher by 14–40% for all seasons. 

Competitive advantage analysis 

Results of the competitive advantage analysis showed that the 

Philippines has no competitive advantage under the export trade 

scenario. The country has no competitive advantage in domestically 

producing rice for export as indicated by the resource 

cost ratios greater than 1. A major reason for this export 

noncompetitiveness was the high per unit cost of rice production 

shown in the costs and returns analyses. 

Under the import trade scenario, there is a competitive 

advantage in domestically producing rice only on irrigated 

farms. For infrastructure development, farmers with access to 

good infrastructure had higher levels of competitive advantage. 

Seemingly unrelated regression analysis 

Results of the SUR analysis showed that nitrogen fertilizer, 

pesticides, and irrigation were positively correlated with yield 

and showed a very high level of significance. These indicate 

that an increase in one of these explanatory variables will positively 

and significantly affect rice yield. Road structure affected 

the price of urea fertilizer. Urea fertilizer was more expensive 

in areas with poor road conditions owing to higher transportation 

costs incurred in the purchase of this input. 

Another major finding of the study is that road structure 

and vehicle density positively and significantly affected the 

quantity of nitrogen fertilizer used in rice production. Farmers 

would apply more nitrogen fertilizer to their crops because of 

its lower price brought about by lower transaction costs, good 

road conditions, and the availability of sufficient transport facilities 

in the area. 

The major determinants of farm income were the major 

inputs of production such as nitrogen fertilizer, pesticides, 

preharvest labor, as well as infrastructure indicators such as 

irrigation, road structure, and transportation costs. 


Table 2. Average costs and returns of rice production in the Philippines, 1996 WS and 1997 

DS by state of infrastructure development. 

Wet season 

Dry season 

Item Irrigated Nonirrigated Irrigated Nonirrigated 

Good Poor Good Poor Good Poor Good Poor 

Yield (t ha –1 ) 3.61 3.31 2.61 2.39 3.77 3.24 2.26 2.32 

Price of output ($ t –1 ) 290 290 290 290 300 300 300 300 

Gross returns ($ ha –1 ) 1,047 960 757 693 1,131 972 678 696 

Production costs ($ ha –1 ) 

Seeds 49 42 46 40 48 43 43 38 

Fertilizer 87 79 53 52 90 80 58 47 

Pesticides a 52 45 48 38 49 43 45 40 

Irrigation fee 34 0 – – 37 40 – – 

Transportation 3 3 3 3 3 3 3 3 

Fuel and oil 18 19 10 13 21 19 10 14 

Labor costs 339 353 289 285 346 336 296 297 

Depreciation 24 24 20 17 24 23 17 17 

Land (rental value) 27 27 14 14 27 27 15 15 

Other costs b 38 35 32 27 34 34 28 23 

Total production costs ($ ha –1 ) 670 629 515 490 679 648 514 493 

Net farm income ($ ha –1 ) 377 331 242 203 452 324 164 203 

Net profit-cost ratio 0.56 0.53 0.47 0.41 0.67 0.50 0.32 0.41 

Cost per metric ton ($ t –1 ) 186 190 197 205 180 200 227 213 

a 

Includes herbicides, insecticides, fungicides, molluscicides, and rodenticides. b Includes food costs, interest expenses, 

and land tax. 

Source: Department of Agriculture-PhilRice Rice-Based Farm Household Survey, 1997. 

Policy implications 

Results of the study generally imply that good infrastructure 

improves farm profitability and productivity. Therefore, there 

is a need to evolve an infrastructure development strategy in 

irrigation and farm-to-market roads for the rice sector. 

Public investment policy for irrigation 

Irrigation increases the cropping intensity of farmers. The 

Department of Agriculture (DA) should fast-track the construction 

of additional irrigation systems because of their high interaction 

effect in terms of efficiency in the use of water and 

other inputs such as fertilizer and seeds to sustain productivity 

and reduce risks in the rice sector. Irrigation also provides a 

safety net from distortion caused by natural agro-climatic phenomena 

such as drought, as indicated by the higher net farm 

income of farmers in irrigated areas. 

Public investment policy for farm-to-market roads 

Inadequate farm-to-market roads generally increase the transaction 

costs of inputs and outputs. According to Gonzales 

(2000a,b), inadequate infrastructure in marketing and distribution 

accounts for 35–40% of the total costs of production 

from the farm to the wholesale market in the grain sector. It 

also restricts the adoption of technology because of the relatively 

higher cost on-farm of yield-shifter inputs such as fertil- 

izer and certified seeds. The restriction in the use of these yieldenhancing 

inputs results in lower farm productivity and consequently 

lowers farm income. 

The dismal condition of rural roads has been cited as the 

major reason (1) traders are able to dictate farm-gate prices, 

(2) transportation costs are exorbitant, and (3) agricultural 

growth did not lead to growth of the nonagricultural sector as 

it did in the newly industrializing economies (AGRICOM 

1997). Given the long gestation period in investment in public 

infrastructure, strategic alliances among agricultural stakeholders 

should always advocate the inclusion of public infrastructure 

investment in the budgetary process. 

The open trade regime in agricultural products has made 

the Philippines vulnerable to globalization. Given the state of 

noncompetitiveness of the rice sector, there is a need to focus 

on key result areas, some of which are proposed by Gonzales 

(2000a,b), following the agribusiness framework to cushion 

the impact of globalization. 

Accelerate farmers’ access to good-quality seeds 

Results of this study showed that only a very small percentage 

of Philippine rice farmers use good-quality seeds. Since goodquality 

seeds enhance yield by 10–15% over ordinary seeds, 

full adoption of good-quality seeds can partly enhance productivity 

in the rice sector. 


Improvements in postharvest facilities 

Postharvest losses account for a substantial portion in the production-distribution 

chain and should therefore be minimized, 

if not totally eradicated, to enhance productivity. 

Promote a higher rate of fertilizer use 

The complementarity of fertilizer use, availability of irrigation 

water, and adoption of high-yielding varieties should be 

considered in promoting higher rates of fertilizer use. Provision 

of other services such as credit and marketing services as 

well as improvement in infrastructure such as irrigation and 

transportation will encourage farmers to apply the right amount 

of fertilizer. 

References 

AGRICOM (Congressional Commission on Agricultural Modernization). 

1997. Modernizing agriculture. Manila (Philippines): 

Congress of the Philippines. 

Gonzales LA. 2000a. Philippine agriculture in the next millennium: 

some strategic issues and directions. Proceedings of the Aventis 

Launch Conference, 4 February 2000, Makati City, Philippines. 

Gonzales LA. 2000b. The Philippine rice industry: strategic issues 

and directions. STRIVE Foundation, Putho, Tuntungin, Los 

Baños, Laguna, Philippines. 

Notes 

Authors’ addresses: Rowena G. Manalili, Philippine Rice Research 

Institute, Maligaya, Science City of Muñoz, 3119 Nueva Ecija, 

Philippines, e-mail: rgmanalili@philrice.gov.ph; Leonardo A. 

Gonzales, SIKAP/STRIVE Foundation, One Tepeyac Place, 

Gov. San Luis Road, Putho, Tuntungin, Los Baños, Laguna, 

Philippines, e-mail: lag@strivefoundation.com. 

Development of an on-farm rice storage technique 

using fresh chilly air and preservation of high-quality rice 

Shuso Kawamura, Kazuhiro Takekura, and Kazuhiko Itoh 

There are two commercial brown-rice storage systems in Japan: 

an environment-temperature storage system, in which the 

temperature during storage is not controlled, and a low-temperature 

storage system, in which the temperature is maintained 

below 15 °C during storage. The low-temperature storage system 

minimizes insect activities and mold growth, and fumigants 

are therefore not required during storage. However, this 

system requires electricity to cool the stored brown rice. 

A previous basic study (Kawamura et al 1997) revealed 

that the quality of rice stored at a temperature below ice point 

is comparable with that of newly harvested rice. Kawamura et 

al (2004) reported that rice with moisture content of less than 

17.8% does not freeze at a temperature of –80 °C and that a 

temperature below ice point minimizes the physiological activities 

in rice and hence minimizes the deterioration of rice 

quality. Rice storage at a temperature below ice point was 

named “super-low-temperature storage” by Kawamura because 

the storage temperature was much lower than that of low-temperature 

storage. 

Based on the results of the studies mentioned above, we 

conducted an on-farm experiment to develop a new rice storage 

technique using ambient fresh chilly air in winter and to 

preserve the quality of rice. 


Storage structure 

Two of 12 silos of a grain elevator constructed in Hokkaido, 

Japan, were used for the on-farm experiment. Each silo had a 

diameter of 7.4 m and a height of 23.2 m. Each silo was made 

of steel with a 75-mm insulation layer on the outside of the 

wall. An automatic system was installed for aeration from the 

bottom through to the top of the silo, and an automatic system 

was installed for ventilation in the upper vacant space of the 

silo. 

Rice samples 

Kirara397 and Hoshinoyume varieties were used for the storage 

experiment. The moisture content of each rough rice sample 

was 15.4%. 

Storage conditions 

Five hundred tons of Kirara397 rough rice was stored in one 

silo and 494 t of Hoshinoyume rough rice was stored in the 

other silo. The storage period was from the end of November 

1999 until the end of July 2000. The rough rice in the two silos 

was simultaneously aerated in January 2000. Aeration was 

automatically carried out when the temperature of the fresh air 

was below –7 °C and it was continued until the cooling front 

had moved through all of the rough rice in the silos. Total aerating 

time (fan time) was 91 h. 


Grain temperature (ºC) 

30 

Mean temperature in parentheses 

20 

Room-temperature storage (20.2 ºC) 

Silo storage 

10 cm from silo wall (5.8 ºC) 

10 

Aeration 

Low-temperature 

storage (7.6 ºC) 

0 

Silo storage in 

center of silo (1.5 ºC) 

Minus 5 ºC storage (–5.0 ºC) 

–10 

10 Nov 10 Dec 10 Jan 9 Feb 11 Mar 10 Apr 11 May 10 Jun 11 Jul 10 Aug 

Period (1999-2000) 

Fig. 1. Rice grain temperatures during on-farm silo storage and control storage (Kirara397). 

Three control storage experiments were also carried out 

at the same time: a room-temperature storage, in which rice 

samples were stored in a laboratory room; a low-temperature 

storage, in which rice samples were stored in a commercial 

rice warehouse and kept at a temperature below 15 °C; and a 

storage experiment at –5 °C, in which rice samples were stored 

in a refrigerator and kept at –5 °C. About 15 kg each of rough 

rice and brown rice were stored in polyethylene containers in 

each control experiment. 

Temperature measurement and sampling 

The temperatures of rough rice in the center of the silo and at 

10 cm from the silo wall were measured by thermocouples set 

at 2.2-m intervals from the bottom to the top of the silo. The 

temperatures of rough rice and brown rice in each container in 

the control storage experiments were also measured. 

Rice was sampled and tested for quality before, during, 

and after storage. Rough rice samples were taken from the 

center of the silo and at four points 15 cm from the silo wall 

(north, west, south, and east in the silo) at the end of the storage 

period. Sampling depths below the surface of the rough 

rice were 0.1, 0.5, 1.0, 2.0, and 4.0 m. 

Quality assessment 

Moisture content, germination rate, free-fat acidity, and 

texturogram property were determined to assess rice quality. 


Grain temperature during storage 

The range of grain temperatures in the vertical direction in 

each silo was less than 3 °C, and there was no tendency in the 

grain temperature distribution. The temperatures recorded in 

the center and near the wall of each silo were averaged, and 

the average values were used as indicators of changes in grain 

temperature during on-farm storage (Fig. 1). 

The grain temperature was 10 °C at the beginning of 

storage. The temperature of rice grains near the wall gradually 

decreased as the ambient temperature fell. The minimum grain 

temperature near the wall in the middle of February was 

–2 °C. From the end of March until the end of the storage 

period (at the end of July), the grain temperature near the wall 

gradually increased. The maximum grain temperature near the 

wall in the middle of July was 21 °C. The grain temperature in 

the center of the silo remained constant (10 °C) at the beginning 

of storage and fell to –2 °C when the silo was aerated at 

the end of January. The grain temperature in the center of the 

silo remained below ice point until the end of July. After aeration, 

the grain temperature throughout the silo remained below 

ice point until the end of March. These results indicate 

that super-low-temperature storage of rice in a farm-scale silo 

can be achieved by using aeration and chilly fresh air in winter. 


The thermal conductivity of rough rice (about 0.09 W 

m –1 K –1 ) is nearly equal to that of lumber (0.15 W m –1 K –1 ) 

and glass wool (0.04 W m –1 K –1 ). This means that rough rice 

is a thermal insulating material. On the other hand, the specific 

heat of rough rice (about 1.7 J K –1 g –1 ) is larger than that 

of lumber (1.3 J K –1 g –1 ) and concrete (0.8 J K –1 g –1 ). This 

means that rough rice is a refrigerant material. Because of these 

physical properties of rough rice, the grain temperature in the 

center of the silo remained below ice point until the end of 

July despite the increase in outside temperature in summer. 

Quality of rice sampled from different parts 

of the silo 

Rice samples were collected from different parts of the silo at 

the end of the storage period and were examined for quality. 

The germination rates of all samples were more than 97%. 

Free-fat acidities of all samples ranged from 12 to 15 mg. A 

germination rate of rice grains of more than 90% and free-fat 

acidity of rice grains of less than 20 mg mean that there has 

been no deterioration in the quality of the rice. The temperature 

of rice grains near the silo wall increased to about 20 °C 

in July. However, the temperature of rice grains near the wall 

was below ice point during winter, and the mean temperature 

of rice grains near the silo wall during storage was 5.8 °C (Fig. 

1). The results of quality measurements showed that there was 

no deterioration in the quality of rice grains near the silo wall. 

Quality assessment of rice before and after storage 

The germination rates of rice subjected to the silo, –5 °C, and 

low-temperature storage were more than 98% after storage, as 

high as that of samples taken before storage. On the other hand, 

the germination rate of rice subjected to room-temperature storage 

decreased to about 50%, indicating that the rice had lost 

vitality during room-temperature storage. The free-fat acidities 

of rice increased in all of the storage experiments. However, 

there were differences in the rates of increase in free-fat 

acidity: the rate of increase in rough rice during storage was 

less than that of brown rice. Moreover, the rate of increase in 

free-fat acidity was highest for rice stored at a room temperature, 

next highest for rice stored at a low temperature and in 

the silo, and lowest for rice stored at –5 °C. These results indicate 

that the decomposition process of fat is slower in rough 

rice during storage and in rice during storage at lower temperatures. 

Changes in the texture of cooked rice were minimized 

by silo storage and –5 °C storage and by storage of 

rough rice. 

The results of quality assessment indicate that storage 

of rice at temperatures below ice point (silo storage and –5 °C 

storage in this study), which is called super-low-temperature 

storage, preserves the quality of rice at a much higher level 

than that of rice stored at higher temperatures. 

Storage capacity of rough rice (000 t) 

120 

100 

80 

60 

Number of newly 

constructed facilities 

Each year 

Total 

47 

64 

79 

104 

111 

115 

40 

(3) (6) 

(4) 

19 

20 

(2) 

28 (5) (4) 

25 

8 

17 

0 

16 

(1) 

11 

(2) 

6 5 

0 

1995 1996 1997 1998 1999 2000 2001 2002 2003 

Year of construction 

Fig. 2. Storage capacity of rough rice in Hokkaido, Japan. 

use of fresh chilly air enables the quality of rice to be preserved 

at a high level without requiring a cooling unit or electricity. 

The use of the super-low-temperature storage technique 

has been increasing in cold regions of Japan in recent years. In 

Hokkaido, the northernmost island in Japan, 26 grain-elevators 

have been constructed since 1996. The storage capacity 

of rough rice in Hokkaido was 115,000 t at the end of 2003 

(Fig. 2). 

References 

Kawamura S, Natsuga M, Kouno S, Itoh K. 1997. Super-low temperature 

storage for preserving rice quality. Proceedings of 

the Joint International Conference on Agricultural Engineering 

& Technology, Volume III, 15-18 December 1997, Dhaka, 

Bangladesh. p 820-824. 

Kawamura S, Takekura K, Itoh K. 2004. Rice quality preservation 

during on-farm storage using fresh chilly air. Proceedings of 

the 2004 International Quality Grains Conference, 19-22 July 

2004, Indianapolis, Ind. (USA). p 1-17. 

Notes 

Authors’ addresses: Shuso Kawamura, Graduate School of Agricultural 

Science, Hokkaido University, Kita 9, Nishi 9, Sapporo 

060-8589, e-mail: shuso@bpe.agr.hokudai.ac.jp; Kazuhiro 

Takekura, National Agricultural Research Center, 3-1-1 

Kannondai, Tsukuba, Ibaraki 305-8666, e-mail: 

takekura@affrc.go.jp; Kazuhiko Itoh, Graduate School of 

Agricultural Science, Hokkaido University, Kita 9, Nishi 9, 

Sapporo 060-8589, Japan, e-mail: 

kazu@bpe.agr.hokudai.ac.jp. 

Extension of super-low-temperature storage of rice 

A new technique for storing rice at a temperature below ice 

point using fresh chilly air in winter was developed. A combination 

of rice storage at a temperature below ice point and the 


Life-cycle inventory analysis of local parboiling processes 

Poritosh Roy, Naoto Shimizu, Takeo Shiina, and Toshinori Kimura 

Parboiled rice is the staple food in some developing countries, 

including Bangladesh. In Bangladesh, per capita consumption 

is reported to be 168 kg year –1 (FAO STAT 2001). Kar et al 

(1999) reported that one-fourth of the world’s paddy is parboiled. 

Parboiled rice has been produced by both traditional 

and modern methods. Modern methods are energy- and capital-intensive, 

and are not suitable for small-scale operation at 

the village level. In rural areas, various methods are being used 

to produce rice and these consume different amounts of energy. 

The commonly used local parboiling processes are vessel 

(0.5–1.2 t batch –1 ), small-boiler (2–4 t batch –1 ), and medium-boiler 

(5–10 t batch –1 ). Biomass, that is, rice husk, is the 

main source of energy for parboiling. In Bangladesh, 63% of 

the total energy consumption is met by biomass fuel and 37% 

is commercial fuel. The household sector consumes 80% of 

total biomass energy and rural households use it almost exclusively 

for cooking (Bari et al 1998). It is reported that biomass 

combustion contributes as much as 20–50% of global greenhouse 

gas emissions, of which one-third may come from households. 

This has an adverse effect on human health and the environment 

(Smith 1999). With the growing concern about environmental 

pollution and health risks, it is very important to 

find a sustainable rice-processing method. Therefore, this study 

attempts to evaluate the environmental effects of different riceprocessing 

methods using the life-cycle assessment (LCA) 

methodology. 

Methodology 

LCA is a methodology to assess all the environmental impacts 

associated with a product, process, or activity throughout its 

life cycle by accounting for and evaluating resource consumption 

and emissions. This concept consists of four major steps: 

goal definition and scoping, inventory analysis, impact assessment, 

and improvement assessment (SETAC 1993). 

Goal definition and scoping 

This stage defines the purpose of the study, the expected product 

of the study, the boundary conditions, and the assumption. 

Furthermore, a reference unit (functional unit, FU), to which 

all the environmental impacts are related, has to be defined. 

The goal of this study was to investigate the life cycle of rice 

produced by the vessel, small-boiler, medium-boiler, and untreated 

process. Figure 1 shows the life cycle of rice under 

different processing methods and the boundary of this study, 

which is encircled by a broken line. 

It has been reported that agricultural LCAs often exclude 

production processes of medicine and insecticides, machines, 

buildings, and roads because of a lack of data (Cederberg and 

Mattsson 2000). In this study, environmental releases related 

to the cultivation of paddy and construction of the parboiling 

facilities were not considered because of the unavailability of 

data. Usually, the paddy and rice are marketed at the nearby 

market or at the mill-gate in local areas. Manually operated 

three-wheeled rickshaw-vans are the main transportation used 

in local parboiling processes. Therefore, energy consumption 

and environmental releases from transportation were also not 

considered. The FU has been defined as the mass of the product, 

that is, 1 ton of head rice. 

Inventory analysis 

The life-cycle inventory (LCI) analysis quantifies resource use, 

energy consumption, and environmental releases associated 

with the system being evaluated. The energy consumption in 

the parboiling processes, dehusking, milling, and cooking, was 

taken from our own study (Roy 2003). Energy consumption 

during drying of parboiled paddy was derived from the literature 

(Palipane et al 1988). Energy consumption in dehusking, 

milling, and cooking was considered to be the same for parboiled 

rice produced by different processing methods. 

Seedling 

Cultivation 

Parboiling 

PresteamingSoakingSteaming DryingDehuskingMilling 

Parboiled 

Harvesting 

Paddy 

Fresh 

Untreated DehuskingMilling Rice 

CookingConsumption 

Parboiling 

PresteamingSoakingSteaming Drying DehuskingMilling 

Parboiled 

Fig. 1. Life cycle of rice and the system boundary of this study (marked by the broken line). 


Table 1. Inventory results (inputs and outputs) of the rice life cycle produced by different 

processes. a 

Inputs and outputs per t 

of head rice 

Process 

Vessel Small-boiler Medium-boiler Untreated 

Input Paddy (kg) 1,517.5 1,445.1 1,501.5 1,666.7 

Biomass energy (kg) 1,275.7 1,286.2 1,170.0 885.1 

Diesel energy (L) – – 0.1 – 

Output Biomass energy (kg) 440.1 419.1 435.4 483.3 

Atmospheric emissions (kg) 

CO 2 1,349.0 1,360.0 1,237.9 938.2 

CO 94.1 94.2 93.3 82.8 

CH 4 3.0 3.0 2.8 2.0 

TSP 18.3 18.6 15.3 8.9 

SO x 4.1 4.1 3.7 2.8 

NO x 3.6 3.6 3.5 2.9 

VOC – – 0.00077 – 

Water emissions (g) 

BOD 426.8 406.5 422.3 – 

COD 1,114.4 1,061.3 1,102.7 – 

Amino nitrogen 113.6 108.2 112.4 – 

Phenol 70.0 67.0 69.0 – 

Solid waste (ash) (kg) 261.7 263.8 241.4 185.1 

a The head-rice yield of the vessel, small-boiler, medium-boiler, and untreated process was 65.9%, 69.2%, 

66.6%, and 60%, respectively. 

It was assumed that the energy requirement in the life 

cycle of rice was met by biomass energy and that biomass (rice 

husk) was the source of primary energy for all types of energy 

consumed in the rice life cycle, except diesel energy. The biomass 

used in parboiling and electricity generation is considered 

to be the industrial use of biomass and an improved domestic 

cook-stove was used for cooking. It has been reported 

that the electricity efficiency of IGCC (integrated gasification 

combined cycles, biomass option) is 43% (Gustavsson 1997). 

The efficiency of an improved cook-stove (ASTRA) is reported 

to be 30% (Bhattacharya et al 1999). Based on these factors, 

the primary energy, atmospheric emissions (CO 2 , CO, TSP, 

CH 4 , NO x , SO x , and VOC), water emissions (amino nitrogen, 

phenol, BOD, and COD), and production of final solid waste 

(ash) were determined. 


Although the inventory results consist of an exhaustive list of 

parameters, the only parameters discussed in this study are 

resource use, energy use, environmental emissions, and solid 

waste (ash) production. The inventory results (inputs and outputs) 

of the rice life cycle are reported in Table 1. 

Energy consumption 

The rice-processing industry consumes some energy and at the 

same time produces some energy in the form of a by-product 

(rice husk). Based on this assumption, an energy inventory was 

worked out and expressed in terms of primary energy (rice 

husk). In the case of parboiled rice, the energy inventory was 

lower for the medium-boiler than for the vessel and small-boiler 

process. Fresh rice consumes less energy than parboiled rice, 

but it consumes more resource (paddy) because of low headrice 

yield (Table 1). 

Atmospheric emissions 

The environmental load was found to be lower for untreated 

rice than for parboiled rice. Parboiled rice produced by the 

medium-boiler process had a lower environmental load than 

that produced by the vessel and small-boiler. The atmospheric 

emission inventory indicates a need to switch methods to reduce 

air emissions. 

Water emissions 

The excess soak-water drains after the soaking treatment in 

the parboiling process and is the main source of water emissions 

in the life cycle of parboiled rice. There is no soaking 

treatment in the untreated process and hence no water emissions 

were reported for fresh rice. 

Solid waste (ash) 

Untreated rice produces the lowest amount of solid waste compared 

with the other processes. For parboiled rice, solid waste 

was lower for the medium-boiler process (Table 1). 

LCI analysis of rice reveals that all the processes have a 

negative effect on the environment and the environmental load 

depends on the production process. The results showed a 

gradual decrease in inventory (energy consumption, atmospheric 

emissions, and solid waste) from the small-boiler to 

the untreated process (small-boiler>vessel>mediumboiler>untreated) 

and there are no waterborne emissions in 

the case of the untreated process. The environmental load is 


lowest for untreated rice; however, it consumes more resource 

(paddy) because of the lowest head rice. This study indicates 

that a switch in rice production processes is required to reduce 

energy and resource consumption, and environmental pollution. 

If fresh rice is considered to be a sustainable energy consumption 

option (the energy shortage would be overcome by 

agri-residues, animal wastes, and tree leaves and twigs), then 

the other processes might be responsible for deforestation. 

Considering the head-rice yield and consumption pattern, it 

would be wise to produce parboiled rice by adopting the medium-boiler 

process even though it requires a higher initial 

installation cost than the vessel and small-boiler process, and 

consumes more energy than the untreated process. 

Conclusions 

This study makes it possible to compare the environmental 

load of different types of rice and it reveals that all the processes 

are responsible for environmental pollution, although 

the intensity of pollution varies from process to process. Thus, 

a change in the rice production process and consumption pattern 

would reduce energy consumption, atmospheric emissions, 

waterborne emissions, and solid waste in the rice life cycle. A 

nominal incentive, motivation, and awareness of the environment 

and health are required for method switching. Method 

switching would reduce environmental pollution, deforestation, 

and global warming. 

References 

Bari MN, Hall DO, Lucas NJD, Hossain SMA. 1998. Biomass energy 

use at the household level in two villages of Bangladesh: 

assessment of field methods. Biomass Bioenergy 15(2):171- 

180. 

Bhattacharya SC, Attalage RA, Augustus Leon M, Amur GQ, Abdul 

Salam P, Thanawat C. 1999. Potential of biomass fuel conversion 

in selected Asian countries. Energy Conver. Manage. 

40:1141-1162. 

Cederberg C, Mattsson B. 2000. Life cycle assessment of milk production: 

a comparison of conventional and organic farming. 

J. Cleaner Prod. 8:49-60. 

FAOSTAT. 2001. Food and Agriculture Organization of the United 

Nations, Rome Italy, FAO statistical databases, available online 

at http://apps.fao.org/. 

Gustavsson L. 1997. Energy efficiency and competitiveness of biomass 

based energy systems. Energy 22(10):959-967. 

Kar N, Jain RK, Srivastav PP. 1999. Parboiling of dehusked rice. J. 

Food Eng. 39:17-22. 

Palipane KB, Adhikarinayake TB, Watson WR, Thorpe GR, 

Ilangantilleke SG, Senanayake DP. 1988. Fluidized bed drying 

of parboiled rice. Sri Lankan J. Post Harvest Technol. 

1(1):1-9. 

Roy P. 2003. Improvement of energy requirement in traditional parboiling 

process. PhD thesis. University of Tsukuba, Japan. 

(Unpublished.) 

SETAC. 1993. A conceptual framework for life-cycle impact assessment. 

Workshop (1-7 February 1992) Report, The Society of 

Environmental Toxicology and Chemistry, Pensacola, Fla. 

(USA). 

Smith KR. 1999. Fuel emission, health and global warming. Regional 

Wood Energy Development Programme in Asia (FAO). Wood 

Energy News 14(3):4-5. 

Notes 

Authors’ addresses: Poritosh Roy and Takeo Shiina, Distribution 

Engineering Laboratory, National Food Research Institute, 

Tsukuba, Japan; Naoto Shimizu and Toshinori Kimura, Graduate 

School of Life and Environmental Sciences, University of 

Tsukuba, Japan, e-mail: shimizu@sakura.cc.tsukuba.ac.jp. 

Using genetics to create changes in rice cooking, 

processing, storage, and health-beneficial properties 

Christine Bergman 

A great deal of improvement in rice grain end-use quality has 

occurred over the past decades using postharvest technology. 

For example, a method has been developed to fabricate rice 

flour into grain-like shapes possessing superior mineral content. 

Preharvest technology, however, has been used to a lesser 

degree to affect rice end-use quality. To date, only traditional 

breeding methods combined with molecular markers and relatively 

simple physical and biochemical tools have been used 

to change the end-use quality of rice cultivars. The few traits 

that have been modified using these tools are amylose content, 

gelatinization temperature, kernel morphology, bran color, 

chalkiness, and milling quality. Changing rice end-use quality 

characteristics using preharvest techniques has an advantage 

over postharvest methods because they create permanent 

change. On the contrary, changes made using postharvest methodology 

must be made to each rice harvest. 

Advancement in genetics methods and biochemical separation 

and identification tools have lagged behind the technological 

advancements underpinning postharvest methods used 

to modify rice end-use quality. But, the gap between the tools 

of pre- and postharvest technology is closing. Consequently, 

some research programs across the world are using these tools 

to evaluate the potential of developing cultivars with modified 

traits such as storage, cooking, processing, and health-beneficial 

properties. These programs consist of collaboration among 

geneticists, breeders, food scientists, nutritionists, and chemists. 

Below is a review of research efforts using the tools of 


genetics to modify rice end-use quality, including traditional, 

mutation, and molecular breeding techniques. 

Traditional breeding 

To use traditional breeding to modify rice traits, genetic diversity 

for the trait must exist. Nearly 420,500 Oryza samples are 

maintained in germplasm collections across the world. Most 

of this genetic material has yet to be explored for its variation 

in any aspects of end-use quality. There is an increasing interest 

by some breeding programs around the world in using these 

genetic resources to develop rice cultivars with modified kernel 

characteristics. As an example of this, work by my colleagues 

and me at the U.S. Department of Agriculture is discussed 

below. 

A germplasm collection was chosen to contain as much 

genetic diversity as possible. More than 200 rice cultivars and 

exotic accessions from over 25 countries were grown using 

the same cultural management practices during two growing 

seasons. The collection has been used to evaluate the potential 

for developing cultivars with a reduced need for postharvest 

bran stabilization and enhanced levels or ratios of bran 

phytochemicals using genetic diversity and cross-pollination 

techniques. 

Oil content (twofold) and fatty acid profiles in the collection 

varied a great deal (Goffman et al 2003). Significant 

genotypic differences in bran E vitamer (i.e., tocopherols and 

tocotrienols) content (twofold) and gamma-oryzanol (threefold) 

content were also found (Chen and Bergman, unreported 

data). Total gamma-oryzanol content and the three primary 

gamma-oryzanol compounds were not correlated with the level 

of E vitamers across the germplasm collection. The three 

gamma-oryzanol compounds were not correlated with each 

other. Nor was the tocopherol content correlated with the 

tocotrienol content. The lines with white or light brown bran 

had low phenolic contents, whereas those with darker bran 

(dark brown, red, and black) showed a much greater range in 

phenolic concentration (23-fold) and antiradical efficiency 

(Goffman and Bergman 2003). The rate of hydrolytic rancidity 

in this collection was well correlated with esterase activity, 

which varied approximately fourfold (Goffman and Bergman 

2003). These results suggest that it will be possible to use traditional 

breeding methods to increase the oil and various antioxidant 

fractions of rice bran, as well as to modify the bran’s 

fatty acid profile and reduce its susceptibility to becoming rancid. 

Mutation breeding 

When an end-use quality trait is desired, but does not exist in 

rice germplasm or it varies little, then mutation or molecular 

breeding must be used. Mutation breeding becomes the method 

of choice when the genes controlling the trait have not been 

cloned, or if transgenic technology is not desired. Mutation 

breeding has recently been used successfully to create new 

rice kernel traits. A few examples are described below. 

Located primarily in the aleurone layer, phytic acid, when 

consumed by humans, is thought to reduce mineral 

bioavailability. Also, when fed to animals, the phosphorus in 

this molecule is excreted in their feces and rain takes it to waterways, 

where the algae growth that occurs as a result depletes 

the oxygen and other aquatic life dies. Thus, bran low in 

phytic acid is desired, but little genetic variation for this trait 

has been found in rice germplasm. 

A low-phytic-acid mutant has been reported by Larson 

et al (2000). Gamma radiation mutagenesis was used to create 

the nonlethal low-phytic-acid mutant. The mutation was designated 

low phytic acid 1-1 (lpa1-1). Homozygosity for rice 

lpa-1 changed the wild-type portion of seed phosphorus in 

phytic acid from 71% to 39%. The inorganic portion of phosphorus 

increased from 5% to 32%, whereas the total phosphorus 

content of the seeds was unchanged. 

Cool temperatures during grain maturation can interact 

with some genotypes to cause greater amylose content and thus 

firmer rice may ensue. Variability in textural properties is a 

problem for the food-processing industry. Therefore, cultivars 

that are insensitive to low growing temperatures would be advantageous. 

Suzuki et al (2002) set out to create such a genotype 

by mutating a temperature-sensitive low-amylose type. 

This genotype had a mutation at the du2 locus and, as a result, 

the seeds could be visibly distinguished depending on whether 

they matured at cool or warmer temperatures. A mutant that 

was insensitive to cool temperature was identified based on its 

semitranslucent phenotype. Amylose synthesis in the mutant 

and wild type was not significantly different. The authors are 

now confident that the effect of the mutation is not limited to 

the du2 mutant background. They are now incorporating the 

mutation into locally adapted germplasm (personal communication). 

As with other cereal grains, rice is lower in the essential 

amino acid lysine than other sources of protein consumed by 

humans. Elevated lysine levels in rice could help reduce humans’ 

reliance on animal protein, which is in shorter supply 

than grain protein and its production is considered to be less 

sustainable. Kim et al (2004) irradiated calli of a cultivar with 

gamma-rays and mutant cell lines resistant to growth inhibition 

by S-(2-aminoethyl)-cysteine (AEC) were selected. The 

AEC-resistant M 3 lines had lysine levels from 1.58 to 2.66 

times higher than the wild type. 

Molecular plant breeding 

Some characteristics that are desired in rice kernels will have 

to be created using genetic transformation techniques. These 

are generally traits that do not exist, vary too little or too much 

in rice germplasm, or are unlikely to be found using mutation 

methods. The creation of enhanced levels of iron, zinc, and 

beta-carotene in rice using molecular breeding techniques has 

been well described (Beyer et al 2002, Vasconcelos et al 2003). 

Described below are examples of other efforts to use transformation 

technology to change rice end-use quality. 


A family of plant phenolic compounds that are derivatives 

of the neurotransmitter, serotonin, has been hypothesized 

to have an array of health-beneficial effects, such as antioxidant 

activity. Kang et al (2004) transformed rice to have enhanced 

levels of pepper hydroxycinnamoyl-CoA:serotonin N- 

(hydroxycinnamoyl) transferase. The transformed seeds had 

on average ninefold higher amounts of serotonin derivatives 

than the wild type. The radical scavenging activity of the 

transgenic rice was also higher and showed concentration dependence 

on the serotonin derivatives. 

Puroindolines are thought to be involved in creating 

variation in wheat grain texture, but are absent in rice. Mutations 

in the genes that code for puroindolines are associated 

with wheat kernel hardness. Transgenic rice was produced with 

the puroindoline genes under the control of a maize promoter 

(Krishnamurthy and Giroux 2001). Textural analysis of the 

transgenic rice seeds provided evidence that the expression of 

these genes is able to reduce rice grain hardness. After milling, 

flour prepared from these softer seeds had reduced starch 

damage and an increased percentage of fine flour particles. 

The authors suggested that these characteristics could have 

utility in designing new rice-based products. Interestingly, this 

transformation also resulted in increased rice plant resistance 

to several fungal diseases. 

A fifth or more of stored grain in less developed countries 

is lost because of insect and disease damage. The primary 

source of storage loss (i.e., weight and sensory quality) 

for rice is the rice weevil (Sitophilus oryzae). These insects 

depend on serine proteinases to metabolize protein and, consequently, 

consumption of proteinase inhibitors decreases their 

survival rate. Alfonso-Rubi et al (2003) reported the transformation 

of rice with the Itr1 gene, which encodes the barley 

trypsin inhibitor. The gene was controlled by its own promoter 

that confers endosperm specificity and the maize ubiquitin 

promoter. The authors reported a significant reduction in the 

survival rate of rice weevils fed the transgenic rice seeds expressing 

the inhibitor compared with the controls. This confirmed 

the potential utility of this proteinase inhibitor gene to 

decrease rice storage losses caused by the rice weevil. 

Conclusions 

None of the research efforts described above have yet resulted 

in the release of a new cultivar. The traits need to be moved 

into superior-yielding germplasm and its agronomic performance 

evaluated. Also, producers must be convinced of a 

cultivar’s agronomic superiority before they will replace the 

cultivars they are currently planting. An alternative is for an 

interested party to have a cultivar with a unique end-use property 

grown under contract, regardless of its agronomic quality. 

In some instances, mechanisms also need to be in place to segregate 

harvests of the modified cultivar from those of other 

cultivars. When these hurdles are overcome, consumers will 

be closer to having access to rice with a plethora of enhanced 

end-use quality characteristics. 

References 

Alfonso-Rubi J, Ortego F, Castanera P, Carbonero P, Diaz I. 2003. 

Transgenic expression of trypsin inhibitor CMe from barley 

in indica and japonica rice confers resistance to the rice weevil 

Sitophilus oryzae. Transgenic Res. 12:23-31. 

Beyer P, Al-Babili S, Ye X, Lucca P, Schaub P, Welsch R, Potrykus 

I. 2002. Golden Rice: introducing the ß-carotene biosynthesis 

pathway into rice endosperm by genetic engineering to 

defeat vitamin A deficiency. J. Nutr. 132:506S-510S. 

Goffman FD, Pinson SR, Bergman CJ. 2003. Genetic diversity for 

lipid content and fatty acid profile in rice bran. J. Am. Oil 

Chem. Soc. 80:485-490. 

Goffman FD, Bergman CJ. 2003. Relationship between hydrolytic 

rancidity, oil concentration and esterase activity in rice cultivars. 

Cereal Chem. 80:446-449. 

Kang K, Jang S-M, Kang S, Back K. 2004. Enhanced neutraceutical 

serotonin derivatives of rice seed by hydroxycinnamoyl- 

CoA:serotonin N-(hydroxycinnamoyl) transferase. Plant Sci. 

(In press.) 

Kim DS, Lee IS, Jang CS, Lee SJ, Song HS, Lee YI, Seo YW. 2004. 

AEC resistant rice mutants induced by gamma-ray irradiation 

may include both elevated lysine production and increased 

activity of stress related enzymes. Plant Sci. 167:305-316. 

Krishnamurthy K, Giroux MJ. 2001. Expression of wheat 

puroindoline genes in transgenic rice enhances grain softness. 

Nat. Biotechnol. 19:162-166. 

Larson SR, Rutger JN, Young KA, Raboy V. 2000. Isolation and 

genetic mapping of a non-lethal rice (Oryza sativa L.) low 

phytic acid 1 mutation. Crop Sci. 40:1397-1405. 

Suzuki Y, Sano Y, Hiro-Yuki H. 2002. Isolation and characterization 

of a rice mutant insensitive to cool temperatures on amylase 

synthesis. Euphytica 123:95-100. 


Krishnan S, Margarida O, Goto F, Datta S. 2003. Enhanced 


gene. Plant Sci. 164:371-378. 

Notes 

Author’s address: University of Nevada, Las Vegas, Departments of 

Food and Beverage and Nutrition Sciences, 4505 Maryland 

Parkway, Las Vegas, NV 89154, USA, e-mail: 

bergman5@unlv.nevada.edu. 


The role of proteins in textural changes in aged rice 

Toshihisa Ohno, Takahiro Kaneko, and Naganori Ohisa 

The texture of cooked rice from aged grains is generally hard 

and nonsticky. Japanese people prefer cooked rice from newly 

harvested grains because of its highly sticky and soft texture. 

Many factors have been identified as the cause of these changes 

in cooking properties, including the increase in free fatty acids 

during storage and their inhibitory effect on the gelatinization 

of rice starch, changes in the physicochemical properties of 

rice starch itself, and changes in the structure-maintaining components 

(Shibuya and Iwasaki 1982). 

Some investigators have specifically studied the proteins 

associated with these changes. Moritaka and Yasumatsu (1972) 

suggested that disulfide bond formation is responsible for the 

textural changes in cooked rice during storage, and Arai and 

Watanabe (1993) reported that denatured proteins were associated 

with these changes. Furthermore, Hamaker and Griffin 

(1990) reported that reducing agents influence the texture of 

cooked rice. 

We therefore aimed to clarify the role of proteins in the 

textural changes in cooked rice seen after storage. 


Experimental materials 

A japonica-type rice (Oryza sativa L. japonica, cultivar 

Akitakomachi) harvested in Akita Prefecture was used in the 

experiments. After brown rice grains harvested in 2001 were 

polished to wash-free grade, they were used as New Rice A. 

New Rice A grains, stored for 2 months at 30 °C in a closed 

aluminum pouch, were used as Aged Rice A. Brown rice grains 

harvested in 2001 were normally polished and used as New 

Rice B. After the brown rice grains harvested in 1999 had been 

stored for 3 years at 15 °C in a paper bag, they were normally 

polished and used as Aged Rice B. Thereafter, all samples were 

stored at below –5 °C until use in the experiments. 

Cooking and texture measurement 

Ten-gram samples of polished rice were not washed and then 

soaked in 16 mL of distilled water for 1 h in an aluminum cup 

covered with aluminum foil. Next, they were cooked in an electric 

rice cooker (National, SR-W100, with 75 mL of water 

added to the outer pan) for about 12 min. After steaming for 

30 min, the cooked rice was transferred to a closed dish and 

kept at 25 °C for 90 min. 

The hardness (H) and stickiness (S) of the individual 

cooked rice grains were then measured by 90% deformation 

with a compression tester (Taketomo, Tensipresser, TTP- 

50BX2). The most important parameter, stickiness/hardness 

(S/H) ratio, was worked out. More than 40 cooked rice grains 

of every type were measured. Conditions for the measurement 

were as follows: plunger and stage, aluminum; bite speed, 2 

mm sec –1 ; and sample temperature at the time of measurement, 

25 °C. 

The influence of the addition of an oxidizing or reducing 

agent on the S/H ratio was also investigated. Instead of 

distilled water, an aqueous solution of 5 mM potassium iodate 

(an oxidizing agent) or 8 mM sodium sulfite or 10 mM cysteine 

(reducing agents) was used as the cooking water. The 

method of cooking and measuring S/H ratio was the same as 

described above. 

Removal of external layer of polished rice 

After the rice samples were further polished in a mill (Grain 

Testing Mill, Satake or HS-4 Mill, Tiyoda Engineering), they 

were cooked. The rate of removal of the external layer of the 

polished rice was 7–20%. The method of cooking and measuring 

S/H ratio was the same as described above. 

Electrophoresis 

The rice flour derived from the external layer of the rice grains 

was mixed with distilled water or an aqueous solution of 5 

mM potassium iodate for 30 min. After the mixture had been 

centrifuged, the pellet was rinsed in distilled water and centrifuged 

again. The precipitate was mixed with Solution A (10 

mM sodium hydroxide, 1% SDS) and incubated for 2 h. 

The mixtures were then centrifuged and the resulting supernatants 

subjected to electrophoresis. The native-PAGE using 

0.1% SDS was performed on 5–20% gradient polyacrylamide 

gels. Coomassie brilliant blue was used as the staining 

agent. 


We evaluated the textures of cooked rice. Figure 1 shows the 

S/H ratios of the various types of cooked rice. The S/H ratio of 

aged rice cooked in water was low. Since Japanese people generally 

prefer cooked rice with a high S/H ratio (Fuke et al 1991), 

Aged Rice A and B were inferior to New Rice A and B. These 

changes have been observed in numerous previous reports (e.g., 

Fuke et al 1991). 

Next, the influence of the addition of an oxidizing or 

reducing agent on the S/H ratio was investigated (Fig. 1). The 

addition of 5 mM potassium iodate to the cooking water lowered 

the S/H ratio. On the other hand, the addition of a reducing 

agent to the cooking water raised it. The S/H ratio of aged 

rice was low as described above, and the addition of the oxidizing 

agent brought the S/H ratio of New Rice down. We therefore 

estimated that the textural changes in the aged rice are 

attributable to oxidation. 


S/H ratio 

0.5 

0.4 

Cooked in water 

Cooked in water with oxidizing agent 

Cooked in water with reducing agent 

The rate of removed external layer to the polished rice was 7%, cooked in water 

The rate of removed external layer to the polished rice was 1% or 20%, cooked in water 

0.3 

0.2 

0.1 

0 

New Rice A Aged Rice A New Rice B Aged Rice B 

Fig. 1. The S/H ratio of various types of cooked rice. 

After the polished rice had been further polished in a 

mill, it was cooked. Figure 1 also shows the S/H ratio of 

samples. After the removal of 7% of the external layer from 

the polished rice, the S/H ratio of Aged Rice A increased to 

slightly higher than that of New Rice A. Distinct differences 

were seen between the S/H ratio of Aged Rice B and that of 

New Rice B after the removal of 7% of the external layer, but 

these differences disappeared completely after the removal of 

1% or 20% of the external layer. Since the S/H ratio of Aged 

Rice B was not sufficiently improved by the removal of 7% of 

its external layer, Aged Rice B is likely to have been in an 

advanced stage of deterioration. These improvements to the 

S/H ratio in aged rice by the removal of the external layer 

suggest that the external layer of the aged rice caused the decreased 

S/H ratio. 

Finally, the native-PAGE results for rice flours treated 

with the oxidizing agent are shown in Figure 2. The native- 

PAGE of rice flour treated with the oxidizing agent revealed a 

decrease in low-molecular proteins (21, 23, and 32 kDa) and 

an increase in high-molecular proteins (larger than 46 kDa). 

These facts suggest that intermolecular disulfide linkages of 

rice proteins were easily formed in the presence of an oxidizing 

agent. 

Our results indicate that proteins may be polymerized 

by disulfide linkage in the external layer of aged rice, and that 

some of these proteins influence its texture. Further investigations 

need to be made to clarify this supposition. 

Positions of molecular 

weight markers 

97.0 k 

66.0 k 

45.0 k 

30.0 k 

20.1 k 

14.4 k 

1 2 

Fig. 2. Native-PAGE of rice flours treated with water or 

oxidizing agent. Lane 1: treated with distilled water; 

lane 2: treated with aqueous solution of 5 mM potassium 

iodate. 


References 

Arai E, Watanabe M. 1993. Enzymatic improvement of the cooking 

quality of aged rice: a main mode of protease action. Biosci. 

Biotech. Biochem. 57:911-914. 

Hamaker BR, Griffin VK. 1990. Changing the viscoelastic properties 

of cooked rice through protein disruption. Cereal Chem. 

67:261-264. 

Moritaka S, Yasumatsu K. 1972. The effect of sulfhydryl groups on 

storage deterioration of milled rice: studies on cereals (part 

10). Eiyo Syokuryo 25(2):59-62. 

Shibuya N, Iwasaki T. 1982. Effect of the enzymatic removal of endosperm 

cell wall on the gelatinization properties of aged and 

unaged rice flours. Starch/Staerke 34:300-303. 

Notes 

Authors’ address: Akita Research Institute of Food and Brewing, 4- 

26, Aza-sanuki, Araya-machi, Akita 010-1623, Japan, e-mail: 

ohno@arif.pref.akita.jp. 

Acknowledgments: This work was financially supported by the Iijima 

Food Science Foundation. 

Postharvest technology of rice: the role of farm women 

in storing grains with different storage practices 

P. Sumathi and M.N. Budhar 

India has increased rice production markedly during the past 

50 years by enhancing production from 22 to 93 million t. On 

the other hand, because of severe and unprecedented drought 

conditions during 2002-03, the total annual rice production 

had a shortfall of 17 million t. However, rice demand in 2010 

and 2025 will be 100 and 140 million t, respectively (Mishra 

2004). Hence, the shortfall in unexpected years and the projected 

demand to feed the growing population have to be met 

by proper handling of produced rice besides increasing production. 

The importance of storage is often ignored in rice 

production ventures. Inadequate storage facilities and improper 

storage can cause considerable losses to rice farmers. Total 

postharvest losses of food grains account for 9.3% and the 

losses during storage alone account for 6.5% (Gandhi 1983). 

Unless losses of food grains in the postharvest phase are satisfactorily 

minimized, the problem of feeding hungry millions 

could continue even with a substantial increase in production. 

Thus, farmers must learn the proper way of storing rice for a 

transitory period on the farm and the local market and for a 

seasonal/national reserve or buffer stock, besides protecting 

grains against weather and insects. Preventing grain discoloration 

and a bad odor to maintain rice quality for feeding and 

marketing is essential. 

The majority of the world’s agricultural producers are 

women: they produce more than 50% of the food that is grown 

worldwide. Women are usually responsible for food processing 

and make a major contribution to food storage. Women 

farmers play a distinct and well-accepted role in all activities 

of rice cultivation and 93% of the farm women are actively 

involved themselves in storing rice grains (Sumathi and Budhar 

2003). Women farmers receive less than 5% of extension services 

worldwide and the priorities of women farmers are rarely 

reflected in agricultural research or national policies, and the 

role of women as agricultural producers is still largely unrecognized 

and has not been addressed (LEISA 2002). Hence, we 

conducted a study among women farmers to find out their role 

in storing paddy grains with different storage practices and 

also to find out their needs for effective storing of grains. 


The study was conducted in Vellore District of Tamil Nadu, 

India, where rice is grown on about 72,000 ha and there are 20 

blocks in this district. Among these blocks, Tirupattur block 

was randomly selected for collection of data. This block consists 

of 32 revenue villages wherein agriculture is the main 

source of livelihood and rice is grown in three distinct seasons: 

the dry (Jul-Nov), wet (Dec-Mar), and summer season 

(Feb-May). Employing a proportionate sampling method, 100 

respondents (farm women) representing small farms (0.1 to 1 

ha) and big farms (>1 ha) were selected, out of which 50 were 

small-farm women and another 50 big-farm women. Thus, a 

sample of 100 women farmers (20 from each village) was selected 

randomly from five randomly selected villages. Most 

of the women farmers cultivate relatively small farms (0.1 to 1 

ha) and some who have more land (>1 ha) live in clusters. 

Through a well-structured interview schedule, the data on adoption 

of women farmers for storing rice grains in different storage 

methods and feedback on suggestions to overcome problems 

in adoption were collected. The relevant data gathered 

were tabulated, processed, and subjected to statistical tests of 

percentage analysis. 


Adoption level of storage methods 

Among 12 practices, four (Table 1)—drying grains before storage, 

keeping bags horizontally, keeping bags on planks and 

leaving passages, and using metal traps for rodent control— 

were found to have been adopted by a majority of the farm 

women. Upadhay and Gupta (1987) also stated that the majority 

of rural women possessed sufficient knowledge of grain 


Table 1. Distribution of respondents according to their practice in the adoption 

of storage methods. 

Small-farm Big-farm Total 

Recommended practices women women 

(no.) (%) (no.) (%) (no.) (%) 

Drying grains before storage 50 100 50 100 100 100 

Impregnated bags – – 6 12 6 6 

Polythene-lined bags 26 52 25 50 51 51 

Keeping bags on planks and 36 72 40 80 76 76 

leaving passages 

Pretreatment of storage 5 10 15 30 20 20 

structures 

Keeping bags horizontally 50 100 50 100 100 100 

Metal storage bins 15 30 3 6 18 18 

Tin-cone plates for storage 9 18 21 42 30 30 

structures 

Fumigation by ethylene 1 2 4 8 5 5 

dibromide ampules 

Metal traps for rodent control 43 86 45 90 88 88 

Anticoagulants for rat control 7 14 18 36 25 25 

Prebaiting and poison baiting 

for rats with zinc phosphide 21 42 34 68 55 55 

storage. All these practices required less cost and skill, which 

might be the reason for their better adoption. Nearly 50% of 

the farm women followed the use of polythene-lined bags and 

prebaiting and poison baiting with zinc phosphide against rats. 

This practice is dangerous to domestic animals, has a high cost 

of technology, and has no technical guidance. The rest of the 

practices had a very low level of adoption because they were 

capital-intensive and demanded skill in practicing them. A 

considerably higher percentage of nonadoption was found in 

other storage practices for paddy grains because these practices 

had a high cost, were complex, a local substitute was 

available, and knowledge on storage structures and technical 

guidance were lacking. 

The big-farm women had a higher percentage of adoption 

of storage practices such as impregnated bags, pretreatment 

of storage structures, tin-cone plates for storage structures, 

and anticoagulants for rat control since the big-farm 

women can afford the costs of these storage methods. However, 

the percentage of usage by small-farm women of metal 

storage bins was high since these structures were given to the 

small-farm women only at the subsidized rate. 

Suggestions for increasing the efficiency 

of storing grains 

The majority of the small-farm women (76%) suggested providing 

financial assistance on a credit basis for the purchase of 

better storage structures such as metal bins. Also, 72% of the 

farm women suggested having subsidized facilities for the same 

purpose and 34% of the respondents wanted special campaigns 

on storage structures (Table 2) since the percentage of suggestions 

was more than with big-farm women. On the other hand, 

the big-farm women suggested an adequate (70%) and timely 

supply of materials (64%). Half of the farm women needed 

timely technical guidance on postharvest technology of paddy 

(52%) since these women farmers are economically stable and 

hence need a timely supply of materials as well as technical 

guidance. 

In general, most of the farm women suggested financial 

assistance on a credit basis (54%) and a timely supply of materials 

(52%), followed by providing a subsidized facility 

(48%), an adequate supply of materials (47%), and timely technical 

guidance (44%). Less than 30% of the farm women 

wanted the research system to develop low-cost storage structures 

to prevent grain losses from infestation and arrange campaigns 

on scientific storage structures. 

The results imply that, in many parts of the world, there 

is an increasing trend in the involvement of farm women in 

agriculture and other allied sectors and these women farmers 

have more responsibilities, especially in the Indian agricultural 

system. Rural women who are contributing to the rural 

economy should be supported by the government (Narayana 

2002). Hence, for a balanced and sustainable growth of rural 

India, women’s role in the developmental process and activities 

should be recognized and they should be adequately supported 

as were the men farmers to achieve gender democracy 

and guarantee food security. 

References 

Gandhi NK. 1983. Stepping up rural warehousing. Kurukshetra 

31(22):18-20. 

LEISA. 2002. Women in agriculture. Magazine on Low External 

Input Sustainable Agriculture (LEISA) India 4(4):4-5. 

Mishra B. 2004. Exploring new opportunities: survey of Indian agriculture. 

The Hindu. p 29-31. 


Table 2. Distribution of respondents according to their suggestions to overcome 

problems in adoption. 

Small-farm Big-farm Total 

Suggestions a women women 

(no.) (%) (no.) (%) (no.) (%) 

Financial assistance on a credit 38 76 16 32 54 54 

basis 

Providing a subsidized facility 36 72 12 24 48 48 

Timely supply of materials 20 40 32 64 52 52 

Timely technical guidance 18 36 26 52 44 44 

Arranging special campaigns 17 34 8 16 25 25 

on storage structures 

Adequate supply of materials 12 24 35 70 47 47 

Developing low-cost storage 10 20 13 26 23 23 

structures 

a Multiple responses. 

Narayana RL. 2002. Women in agriculture. Magazine on Low External 

Input Sustainable Agriculture (LEISA) India 4(4):30. 

Sumathi P, Budhar MN. 2003. Extent of women’s involvement in 

Indian rice cultivation. Int. Rice Res. Notes 28(1):74-76. 

Upadhay M, Gupta P. 1987. Adoption of selected home making by 

rural women. Indian J. Extn. Edn. 23(1&2):76-78. 


Session 10 included a wide range of interests from conditioning 

or primary processing of rice to safety and environmental matters. 

Professor P. Chattopadhyay from India reported on the 

changing scenario in postharvest technology in India, whose rice 

production and economy are growing rapidly. From Japan, the 

latest technological information on drying, milling, and storage 

facilities was introduced. Relatively new topics such as the use 

of by-products, residual agrochemical inspection, life-cycle inventory 

analysis, and so forth were also presented. To meet these 

global needs, the importance of new concepts for rice breeding 

was pointed out from the U.S. R.G. Manalili from the Philippines 

mentioned that the infrastructure surrounding rice production was 

essential for improving international competitiveness in rice production 

as well as in technological renovations. 

Session participants reached the following conclusions: 

1. The covering area of postharvest or postproduction technology 

is becoming wider. This includes safety, environmental, 

and socioeconomic issues. 

Notes 

Authors’ address: Regional Research Station, Tamil Nadu Agricultural 

University, Paiyur 635112, Krishnagiri Dist., Tamil Nadu, 

India, e-mail: mnbudhar@rediffmail.com. 

2. Postharvest technology can improve production efficiency 

more easily than other preharvest measures such 

as the development of new arable land or a new variety. 

3. Therefore, more emphasis should be given to postharvest 

technology and related matters to achieve well-balanced 

production after examining productivity, quality, safety, 

economic, and environmental issues. 

Two outstanding posters were chosen: “Postharvest of rice: 

the role of farm women in storing grains by different storage practices” 

by P. Sumathi from India and “The role of proteins in textural 

changes of cooked rice during rice storage” by T. Ohno from 

Japan. The first dealt with women’s labor in postharvest practices 

during the storage of rice in rural areas of India. It was 

understood that the role of women’s labor would strongly affect 

the future improvement of postharvest practices in developing 

countries. 


SESSION 11 

Enhancing the multifunctionality 

of rice systems 

CONVENER: Y. Tsutsui (NIRE) 

CO-CONVENER: K. Yamaoka (NIRE)

Multifunctional roles of paddy irrigation in monsoon Asia 

Takao Masumoto 

Approximately 54% of the world’s population is thought to 

live in the regions known as humid Asia or the Asian monsoon 

region, which constitute only approximately 14% of the world’s 

land area. The majority of Asia’s massive population is supported 

by intensive paddy rice cultivation, which originates 

from the warm and humid environment and offers high land 

productivity because of irrigation. 

Paddy rice cultivation in the Asian monsoon region is 

not only an excellent form of agriculture offering high land 

productivity; it is also seen as a sustainable and environmentally 

friendly economic activity that suits the climatic and topographical 

conditions of this region. This economic activity 

has continued to evolve for thousands of years across various 

regions, as witnessed by the archaeological traces of 7,000- 

year-old rice cultivation obtained in China. Even today, paddy 

rice cultivation is a unique way of life supported by the endeavors 

of people living in harmony with water. Moreover, 

paddies also make life more convenient for city dwellers by 

reducing floods, fostering groundwater, etc. These “products” 

cannot be sold in any market, and their functions have been 

referred to as “the multifunctional roles of paddy irrigation” 

in this study. 

In this paper, the multifunctional roles of paddy irrigation 

are illustrated by comparing the hydrological environments, 

forms of irrigation practiced, and characteristics of paddies 

in humid regions with those in the arid and semiarid regions. 

Further, the paper examines the quantitative or economic 

methods for evaluating these multifunctional roles and formulates 

policy proposals for Japan and the rest of the world with 

a view toward solving water-related problems. 

Characteristics of paddy irrigation in monsoon Asia 

Definition of Monsoon Asia and characteristics 

of climatologic and hydrologic environment 

Musiake (2001) asserts that regions such as monsoon Asia 

should be characterized by the region’s uniqueness (including 

the nature of human activity), in addition to the climatic conditions 

that have been emphasized in the past. Thus, this paper 

follows his new classification. The Asia monsoon region embraces 

the Indian Ocean to the south, the expansive region of 

Tibet, the Himalayan mountain mass and continental China to 

the north, and the Pacific Ocean to the east. Most of it consists 

of high-precipitation warm regions that have annual rainfall in 

excess of 1,500 mm, influenced by low pressure and monsoons 

accompanied by westerly winds. Meanwhile, the water balance 

(calculated by subtracting annual potential evapotranspiration 

from annual precipitation) generally exceeds 500 mm. 

Characteristics of paddy irrigation in monsoon Asia 

The principal grain crop in monsoon Asia is rice. In fact, nearly 

90% of the world’s rice is produced in the countries of this 

region. Monsoon Asia could be called a virtually homogeneous 

region in that paddy rice cultivation extends over almost the 

whole of its area. Generally speaking, regions with annual precipitation 

of less than 400 mm are classified as perennial irrigation 

zones, those with precipitation from 400 to 1,000 mm 

as unstable irrigation zones, and those with more than 1,000 

mm as replenishment irrigation zones. Of these, the whole of 

monsoon Asia belongs to the last classification. What is more, 

the Asian monsoon region is also characterized by large seasonal 

and short-term fluctuations in the supply of water resources, 

as evident in the distinct dry and rainy seasons. 

Paddy irrigation in arid and semiarid regions 

Overview 

As characteristics of paddy irrigation in the world’s arid and 

semiarid regions, cases from the United States and Australia 

are discussed. In the U.S., paddy zones extend along the Mississippi 

River and in five southern states as well as California. 

The rice cultivated area changes from year to year, from 

300,000 to 1,500,000 ha. In Australia, about 100,000 ha of 

paddy fields are concentrated in the Murray and Murrumbidgee 

River basins. With annual precipitation of about 400 mm, paddy 

irrigation is practiced there. 

Examples 

The state of California in the U.S. is a semiarid region with 

rainfall in winter but little in summer. The average annual precipitation 

is 575 mm, and agriculture there would be impossible 

without irrigation facilities. Paddies and water-supply 

channels are used as retarding basins and diversion channels 

for floodwater. Paddies also provide sanctuaries for migrating 

birds as well as for duck hunting. Although some features of 

paddy rice culture in California are similar to those of monsoon 

Asia, the California paddies and water distribution and 

the drainage system do not have the all-encompassing multifunctional 

roles of paddy irrigation in monsoon Asia. In the 

rice cultivation zones along the Murray River in Australia, 

meanwhile, some floodwater is stored during high flows. However, 

paddies are not seen as having a function for relief from 

flooding. Furthermore, groundwater percolation from paddies 

causes an accumulation of soil salinity in surrounding nonrice 

arable land because the shallow groundwater is saline. In terms 

of the sustainability of agriculture, this could be seen as an 

example of a negative impact being exerted on a region’s water 

resources. 


Paddy storage capacity, S (mm) 

200 

Catherine typhoon 

Status quo 

S 0 

D 0 

Irrigation period 

Fukuoka headwork 

Category 

150 

1/100-y return period 

L 

Block by gates 

B 1 and drops 

Characteristics 

100 

50 

0 

0 

1/50-y RP 

1/10-y RP 

2,000 

4,000 

Drainage (conveying) capacity, D (m 3 s –1 ) 

Fig. 1. Drainage capacity (urban area) and 

storage capacity of paddies (Kinu River), 

(Masumoto 2002). 

St.2 

G 

C 

U 1 

I 3 

S mv 

A 

Main irrigation 

channel 1 

G s A Ecosystem in 

plateaus 

G 

C 

U 1 

D 3 

of channel unit 

S mv 

A 

5 4 3 2 1 0 

Ecosystem 

evaluation by 

hydrologic 

indices 

Evaluated points 

U 5 

I 3 

Methods of quantifying the multifunctional roles 

of paddy irrigation 

Kokai 

River 

U 2 

D 3 

S mv 

A 

Branch 

channel 

C 

A 

Definition of multifunctional roles 

In 2001, the Science Council of Japan delivered a report on 

the multifunctional roles of agriculture and forests to the Minister 

of Agriculture, Forestry, and Fisheries. The report included 

a list of categories connected with the multifunctional roles of 

paddy irrigation. Based on this, the multifunctional roles of 

paddy field irrigation are defined as (1) water cycle control 

functions (flood prevention, groundwater recharge, prevention 

of soil erosion), (2) environmental load control functions (water 

purification, processing of organic waste, climate modification), 

(3) nature formation functions (biodiversity, landscapes), 

and (4) social culture formation functions (health and 

recreation, participatory learning). 

Specific examples 

Each function mentioned above is summarized in the report 

(Masumoto 2004). Discussions on the multifunctional roles of 

agriculture are also under way and attempts are being made 

not only to define multifunctional roles qualitatively but also 

to evaluate them quantitatively. Here, some of the latest results 

are introduced below. 

Flood prevention function. When changes in runoff volume 

due to the abandonment of paddy farming were measured 

in mountainous parts of the Hokuriku region, an increase in 

total runoff and peak runoff for all observed floods was observed 

at times when paddy fields were wet (Masumoto et al 

1997). Meanwhile, as examples in which low-lying paddy zones 

contribute to flood prevention in river basins, research has been 

done on the Ogura basin in Kyoto, the Tomari basin of the 

Natsu River system in Miyagi Prefecture, and the Mekong River 

basin (Masumoto 1998, 2003a, Masumoto et al 2002). 

Attempts have been made to evaluate the 

abovementioned functions at the basin level to propose meth- 

Oka 

headwork 

St.1 


channel 3 

Drainage river 

(Nakadori River) 

U 3 

G 

A 

Branch 

irrigation 

channel 

Iatami 

gate 

Branch 

drainage 

channel 

Fig. 2. Evaluation of irrigation and drainage network for fish ecosystem 

(Fukuoka headwork area, irrigation period), (Masumoto et 

al 2004). 

ods of evaluating flood prevention functions at the basin level 

in macro scales. The relationship between discharge in urban 

rivers and flood prevention capability of suburban paddies can 

be defined as the relation between drainage and storage capabilities 

(Masumoto 2003b). Figure 1 shows an example of this 

evaluation method applied to the basin of the Kinu River, a 

tributary of the Tone River in the north of the Kanto region. 

Ecosystem preservation function. A survey has been 

made of the ecosystem maintenance functions inherent in paddy 

irrigation in the Kokai River of the Tone River system in Japan 

(Masumoto et al 2004). For this, a network map for ecosystem 

evaluation was created using, among other indicators, 

the state of connection of each channel in a network of irrigation 

and drainage channels (Fig. 2). The condition of fish in 

the target area was also evaluated by using time and space 

hydrologic indices proposed. 

D 

B 1 

G 

A 


channel 2 

U 4 

D 1 Ecosystem 

in 

C m 

B 1 plateaus 

Session 11: Enhancing the multifunctionality of rice systems 325

Economic evaluation of the multifunctional roles 

of paddy irrigation 

The benefits brought about by the external economy intrinsic 

to the economic activity of agriculture are public properties to 

be enjoyed by all the people of a nation. Methods applied to 

evaluate the scale of these benefits in economic terms are the 

substitution method, CVM (contingent valuation method), the 

hedonic method, the travel cost method, and the direct method. 

When evaluation is difficult to achieve by these methods, the 

tendency has been for the CVM and other methods to be 

adopted. 

So far, the multifunctional roles of paddies, upland fields, 

and rural areas throughout Japan were evaluated and their results 

produced values ranging from 4.1 to 11.9 trillion yen for 

functions that can be quantified. The following functions, not 

included in these calculation results, are calculated through 

economic evaluation. 

As a method of economic evaluation for flood prevention 

functions, attempts have been made to evaluate the flood 

storage capability of a paddy area by substituting it with the 

volume of a retarding basin. Using a method based on questionnaire 

surveys (stated choice method), meanwhile, methods 

of analysis are moving toward greater refinement by increasing 

the targeted range (functions), and there is a progressive 

shift from CVM toward choice experiments (choice-based 

conjoint analysis: contingent ranking method). With choice 

experiments, however, several situations are evaluated for comparison. 

Policy proposals for Japan and the international community 

Development, economic activity (e.g., agriculture), 

and environmental preservation 

Agriculture, of course, is a type of economic activity, but the 

water supply and drainage systems that exist in rural regions 

could also help protect a region’s ecosystems, for example, by 

providing nursery grounds for fish in connecting rivers. Many 

examples are demonstrating that agricultural development is 

not necessarily in opposition to environmental preservation. 

Rather, we would assert that several points still require improvement 

in terms of protecting ecosystems. The reason for 

this is that drainage channels fed by urban drainage continue 

to flow throughout the year, while water supply channels have 

no flow of water outside the irrigation period. 

Sustainable water-recycling agriculture 

The difference between paddy field agriculture in arid regions 

and that in monsoon Asia needs to be clarified. Namely, there 

are huge differences in paddy irrigation between those regions. 

For paddy fields in the arid and/or semiarid regions, excessive 

irrigation to cope with arid climates causes salt damage due to 

a rise in groundwater levels and this in turns leads to external 

diseconomy. In Japan and other highly humid zones, on the 

other hand, the supply of groundwater from paddy fields can 

be evaluated as an external economy in that it builds up groundwater 

resources. 

Introduction of external economic evaluation 

(basin management) and sharing of water 

management costs 

Studies should be made from the perspective of the basin as a 

whole, rather than just being aimed at the direct beneficiaries 

to agriculture. If this, a case of external economic evaluation, 

were studied in terms of the basin as a whole, it could also be 

considered from the perspective of the internal economy. 

Numerous facilities have been developed in agricultural 

and rural development projects to date. In the past, facilities 

for water management in regions and basins mainly concerned 

with agriculture used to be developed upon applications from 

agricultural concerns. The development was mainly undertaken 

by farmers and land improvement districts, which also bore 

the maintenance and management costs. However, urbanization 

has opened its way to all parts of the country in recent 

years. In many cases, farmers also bear the costs for operating 

and maintaining district drainage and drainage channels in 

suburban agricultural zones. These costs should ordinarily be 

borne by urban residents. In other words, there should be a 

sharing of management costs commensurate with the manifestation 

of multifunctional roles of paddy irrigation, partly as a 

consideration for providing comfort to local residents in urbanized 

areas. Legal support systems also need to be developed 

to address this. 

Basin management during excessive flooding 

In paddy-dominant basins where pump drainage is practiced 

in low-lying areas, flood plains for its connecting main river 

are designed for the capacities on a 100-year probability scale. 

The maximum drainage volume for agricultural facilities is 

normally assumed on a 10-year probability scale (in some recent 

cases, 20–30 years). Thus, runoff that exceeds the drainage 

capacity on a scale of 10 years cannot drain outside the 

basin, and is forcibly stored in drainage channels and/or paddies. 

In other words, farmlands that include agricultural drainage 

channels and paddies have the function of storing floodwater 

for drainage rivers, and, as a result, could be said to help 

reduce the risk of flooding in main rivers downstream where 

urban areas extend. 

References 

Masumoto T. 1998. Paradigm shift in the evaluation of water storage 

function of paddies and in watershed management. J. Jpn. 

Soc. Hydrol. Water Res. 11(7):711-722. (In Japanese with 

English abstract.) 

Masumoto T. 2003a. Flood prevention function of paddies in monsoon 

Asia. Proceedings of Japan/OECD Expert Meeting on 

Agriculture and Land Conservation: Developing Indicators 

for Policy Analysis, Kyoto, Japan, May 2003. p 91-98. 


Masumoto T. 2003b. Indices to evaluate flood prevention function 

of paddies. Proceedings of Japan/OECD Expert Meeting on 

Agriculture and Land Conservation: Developing Indicators 

for Policy Analysis, Kyoto, Japan, May 2003. p 159-164. 

Masumoto T. 2004. Multi-functional roles of paddy irrigation in 

monsoon Asia. J. JSIDRE. 72(7):11-16. (In Japanese.) 

Masumoto T, Kubota T, Matsuda S. 2002. An integrated approach to 

use multifunctional roles of paddies in basin-wide flood and 

water-use management. Electric Transactions of the 18th International 

Congress on Irrigation and Drainage (ICID). 

Q51R12.07:1-16. 

Masumoto T, Kubota T, Matsuda S, Takagi, A. 2004. An evaluation 

method of water resources in paddy regions to preserve ecological 

environments. Technical Report of the National Institute 

for Rural Engineering. No. 202:81-90. (In Japanese with 

English abstract.) 

Masumoto T, Takaki K, Yoshida S. Adachi K. 1997. Effects of abandoned 

paddies in hilly rural areas on runoff. Trans. JSIDRE. 

189:59-68. (In Japanese with English abstract.) 

Musiake K. 2001. Hydrology and water resources in monsoon Asia: 

a consideration of necessity to organize “Asian Association 

of Hydrology and Water Resources.” In the Proceedings of 

the Symposium on Innovation Approaches for Hydrology and 

Water Resources Management in Monsoon Asia, Tokyo, Japan. 

p 1-11. 

Notes 

Author’s address: National Institute for Rural Engineering, 2-1-6 

Kan-nondai, Tsukuba, Ibaraki 305-8609, Japan, e-mail: 

matsumoto@nkk.affrc.go.jp. 

Accounting for culture in paddy cultivation: 

toward a broader definition of “livelihood” 

David Groenfeldt 

Paddy cultivation forms the basis of traditional Southeast Asian 

societies and the livelihoods of the people who comprise those 

societies. Historically speaking, paddy cultivation has always 

(at least for several millennia) been multifunctional—providing 

not only the raw material for subsistence and trade but also 

serving as the central focus for family and community life as 

well as spiritual and religious expression. While times have 

certainly changed, this paper suggests that the multifunctional 

nature of paddy cultivation continues to be important, and that 

our concept of rural “livelihood” should incorporate these cultural 

dimensions. 

Traditional versus modern paddy cultivation 

Twentieth-century economic developments changed the wellintegrated 

world of paddy farmers and imposed new pressures 

while offering new options from modern agriculture. In Bali, 

for example, farmers no longer had to follow the traditional 

cropping calendar imposed by the temple priests. Thanks to 

new crop varieties from the International Rice Research Institute 

(IRRI) and new infrastructure from the Asian Development 

Bank (ADB)-financed Bali Irrigation Project, farmers 

were encouraged to adopt modern methods that treat farming 

more as agri-business rather than agri-culture. 

The Balinese case has become a symbol of the dangers 

of ignorantly tinkering with a carefully evolved agricultural 

system. The disruption of coordinated planting times allowed 

insect populations to spike, causing severe crop losses (Lansing 

1991). The project-imposed diversion weirs were incompatible 

with customary water rights, resulting in conflicts among 

farmers (ADB 1998). The lesson from Bali, and from many 

other cases of externally imposed agricultural strategies, is that 

traditional farming systems are often highly evolved sociotechnical 

systems, and the introduction of any improvements 

must be attempted very carefully, taking into account the potential 

cascade of multiplier effects from seemingly small 

changes. 

Respect for the technical sophistication of traditional 

farming systems has now become an accepted tenet within the 

current paradigm of agricultural development. Indigenous technical 

knowledge even has its own acronym—ITK. And the 

concept of rural livelihoods goes even further in capturing the 

complex interactions of the farming system with the larger 

economic context of handicrafts, wage labor, remittances from 

abroad, etc. The legitimacy that the development profession 

has grudgingly accorded to the economic strategies of rural 

farm families, however, has not been extended to the overall 

cultural “lifeways” of the people themselves. Traditional beliefs 

in river spirits and rice gods are dismissed as superstitions. 

Local customs about diet, nutrition, and health are viewed 

with suspicion, if not ridicule. And traditional values referring 

to time and money are met with exasperation by Western advisers 

preaching the virtues of market-based capitalism. 

The conventional model 

According to the conventional model of development, traditional 

family and community-based modes of paddy cultivation 

are doomed to an eventual demise in the face of competition 

from industrial-style producers. While many observers of 

economic development—myself included—are uncomfortable 

with the social and cultural costs of economic change, the process 

is generally accepted as both necessary and inevitable. 


After all, the growing population of the world needs to be fed, 

and the most efficient arrangements for producing food must 

be followed. 

This conventional model of development is applied to 

paddy cultivation and rural livelihoods as a sort of economic 

determinism that requires changes in cultural values toward 

more economic thinking. Traditional societies are advised to 

“get with the program,” give up their traditional customs, and 

adopt Western-style materialism. The only alternative is poverty. 

Another approach: happiness 

The message of this paper is that there is a middle way, and it 

is a far more interesting way to live than the materialistic extremism 

advocated by conventional development experts. Life 

is more than money, and livelihoods are more than economics. 

Paddy cultivation is not only an economic strategy; it is a part 

of a way of life that needs to be considered in those very broad 

terms. 

The concept of gross national happiness (GNH), which 

forms the center-piece of Bhutan’s development policies, captures 

the broad dimensions of life that I feel need to be incorporated 

into our concept of livelihoods. There are four pillars 

to the concept of GNH as defined in the Bhutan context: (1) 

economic development, (2) cultural heritage, (3) preservation 

and sustainable use of the environment, and (4) good governance. 

It is significant that economic development is listed 

first. The dominance of economics is natural and justified, but 

as a means, and not an end in itself. The dimension I will focus 

on in this paper is that of cultural heritage. The dimension of 

environment is addressed in other conference papers. The governance 

dimension is also relevant to paddy cultivation (e.g., 

irrigator associations as a type of local governance) but is 

omitted from this discussion in the interest of maintaining focus. 

Let me now turn to cultural heritage, which I will gloss 

with the more general term “culture.” 

Culture and livelihoods 

How does paddy agriculture contribute to the cultural dimensions 

of rural life I will subdivide these dimensions into four 

categories: (1) rural social structure at the family and community 

level, (2) cultural identity, (3) spiritual and religious life, 

and (4) aesthetic beauty. 

Social structure 

The irrigation systems used for paddy cultivation largely determine 

the location of village settlements. In Sri Lanka, villages 

grew up around the reservoirs (tanks) constructed for 

irrigating paddy and other crops. In Bali, village lands are defined 

by irrigation canals that are linked to larger networks 

supervised by priests. In Bhutan, the villages typically take 

water from a stream diversion built and maintained by farmers 

from that village. The paddy irrigation infrastructure serves as 

a physical representation of the village community. Family and 

community life are organized around the cultivation cycle, with 

periodic labor exchange for transplanting and harvesting, and 

social gatherings tied to agricultural rituals. This dimension 

can be thought of as the social capital of paddy agriculture. 

The organizational depth of these social interactions serves 

multiple functions beyond paddy cultivation (e.g., credit circles, 

political action, etc.) while also providing social “happiness” 

directly through the satisfaction of social interaction. 

Cultural identity 

The interconnectedness of food, farming, and identity is fundamental 

to traditional societies as well as modern ones. An 

American Indian farmer explained this relationship in kinship 

terms: “The [traditional crops of] corn, beans, and squash plants 

are like our children. We don’t have to live with our children, 

but life is much richer and happier when we do.” Traditional 

cultural identities are often seen as a luxury that modernity can 

no longer support, or even as a threat to national integration. 

The importance of traditional identity is typically appreciated 

only under the looming cloud of cultural extinction. Yet, from 

a psychological perspective, cultural identity is fundamental 

to emotional balance and personal happiness. Traditional foods 

and crop varieties are a key part of this happiness. What is a 

Sri Lankan—or Bhutanese, or Balinese—meal without rice 

And, more specifically, local varieties of rice, prepared in certain 

traditional ways Even in many industrial societies (such 

as Japan and France), where the farming population has been 

reduced to a fraction of the work force, the identity with particular 

foods and farming practices continues to be valued. 

Religion 

The tasks of farming itself can be a spiritual practice. The interaction 

with plants and animals, the preparation of the soil, 

diverting of water, all have meanings that are systematized 

through rituals and ceremonies throughout Asia. The Water 

Temples of Bali are famous examples of merging religion and 

paddy cultivation, but less dramatic illustrations are everywhere, 

from the Hanuman deity guarding a tubewell in India 

to prayers on the threshing floor in Sri Lanka. The daily experience 

of spiritual life is tied to the cultivation process. 

Aesthetic 

The landscape value of paddy fields is important to both rural 

producers and urban consumers. In some cases, these values 

may be harnessed through eco-tourism (as some resorts in Bali 

have attempted to do), but, in most other cases, the agrarian 

beauty of the landscape is a type of common resource provided 

free to the observer. Another type of aesthetic benefit 

provided by traditional paddy cultivation is the rice itself, which 

may be valued for its culinary qualities (taste, texture) as well 

as for its cultural heritage value (as a symbol of a particular 

region or ethnic group). This is why farmers often grow one 

variety of rice for selling, while growing lower-yielding traditional 

varieties for home consumption. 


Implications for agricultural policy 

Both farmers and urban food consumers share an interest in 

preserving certain dimensions of traditional paddy cultivation, 

for the simple reason that these aspects contribute to their happiness. 

Is this a valid basis for selecting among development 

options The idea that development is exclusively the domain 

of economics has emerged for historical reasons, not logical 

ones. What would development policies look like if psychologists, 

or priests, were employed as the advisers The policies 

would probably be quite similar to Bhutan’s policy of gross 

national happiness, which closely reflects the Buddhist heritage 

of that country. 

Happy paddy farmers Happy rice consumers Are these 

realistic objectives for agricultural policy Perhaps we should 

turn the question around: If agricultural policy is not oriented 

toward increasing general happiness, then what is its purpose 

And if policies are indeed oriented to increasing happiness, 

then the dimensions of happiness need to be addressed. 

References 

ADB (Asian Development Bank). 1998. Re-evaluation of the Bali 

Irrigation Sector Project in Indonesia. December. 

Lansing SJ. 1991. Priests and programmers. Princeton, N.J. (USA): 

Princeton University Press. 

Notes 

Author’s address: Independent Consultant, Santa Fe, New Mexico 

(USA), e-mail: DGroenfeldt@aol.com. 

The need to keep irrigated rice culture sustainable 

Nyoman Sutawan 

Rice is the most important staple crop for many people in Asia, 

especially in East and Southeast Asia. It has a long history of 

cultivation, probably for more than seven millennia. Its cultivation 

plays a significant role in Asia’s customs and religions. 

In the year 2000, the world rice harvest area was around 152.0 

million hectares. Of this, Asia contributed nearly 90%. Global 

rice production was 597.2 million tons of rough rice and Asia 

as a whole produced more than 91% (Kim 2001). 

For many countries, rice is not merely an economic good; 

it is also a social, cultural, religious, and even political good. 

A rice shortage could create national instability. Moreover, the 

world rice market is highly volatile and relying upon such an 

unstable rice market could be quite risky for national food security. 

No wonder that many countries still insist on adopting 

a rice self-sufficiency policy. This implies that preserving rice 

culture seems to be of primary importance. 

This paper tries to (1) propose a concept of sustainable 

irrigated rice culture, (2) raise arguments for preserving irrigated 

rice culture, and (3) recommend relevant policy measures 

to preserve irrigated rice culture. 

Proposed concept of sustainable irrigated rice culture 

There are a lot of definitions or concepts of sustainable agriculture. 

Pakpahan (1995) has cited several definitions of sustainable 

agriculture but those will not be repeated here. Suffice 

it to say that most of the definitions seem to merely emphasize 

more the environmental and agro-technological aspects. 

Since land and irrigation water are indispensable for 

irrigated agriculture, it seems necessary to see sustainability 

of agriculture, especially irrigated rice culture, from a rather 

different point of view. 

Irrigated rice culture getting water from an irrigation 

system can be viewed to possess several interrelated elements. 

Sustainable irrigated rice culture should be perceived here as 

a prolonged existence and functioning of several important 

interrelated elements of irrigated rice culture. In other words, 

sustainability of irrigated rice culture should encompass the 

sustainability of the following: (1) the irrigators’ association 

or institution (institutional sustainability), (2) irrigation networks 

(technical sustainability), (3) agricultural production 

(economic sustainability), (4) irrigated land ecosystem (ecological 

sustainability), and (5) socio-cultural values linked to 

rice cultivation (socio-cultural sustainability). 

However, the sustainability of those five elements of irrigated 

rice culture also largely depends on the local natural 

environment, especially the upstream watershed and the quality 

of river water upstream (environmental sustainability), 

which are external to the irrigated rice culture concerned. An 

individual irrigation system for rice cultivation is also an element 

or subsystem of the entire system along a river course. 

This implies that the performance of the individual irrigation 

system or irrigated paddy culture getting water from a lower 

stream may also be affected by the system getting water from 

the upper stream. The sustainability of the local environment 

of the system along the water course as well as the individual 

irrigation system is also affected by various external forces 

such as demographic, social, economic, cultural, political, industrial 

development, tourism, government policy, etc. 


Arguments for preserving irrigated rice culture 

Multifunctionality of irrigated agriculture, 

particularly irrigated rice culture 

Irrigated agriculture, irrigated paddy cultivation in particular, 

possesses multifunctional roles with positive externalities. Irrigated 

agriculture produces not only food and fiber but also 

other intangible “goods” that are quite difficult to evaluate in 

economic terms. The multifunctional roles of irrigated agriculture 

are the provision of agricultural products to ensure food 

security, flood preservation, soil erosion control, groundwater 

recharge, water purification, and air-cooling effect; provision 

of habitat for various tiny living creatures that can create ecological 

balance or biodiversity preservation; provision of beautiful 

landscape that can have good potential for eco-tourism; 

provision of drinking water for rural people and domesticated 

animals; provision of an additional source of water plant and 

animal protein; provision of a place for duck raising; provision 

of a place for religious rituals and rural festivals; and many 

others (Mizutani 2002, Kwun 2002, Groenfeldt 2003). 

We cannot imagine how much money we need to combat 

flood and soil erosion, to purify water, to clean the air, and 

to solve the numerous problems that could emerge if the next 

generation abandons rice farming. Kwun (2002) provides important 

information on estimates of economic values of several 

multifunctional benefits of irrigated agriculture in Korea 

and Japan. For example, just for the flood control function of 

paddy farming, the value ranges from US$16.27 billion to 

$23.99 billion in Japan and from $1.11 billion to $1.32 billion 

in Korea. Thus, rice farmers deserve to be honored and praised 

as national heroes in environmental and biodiversity preservation. 

This is because they produce unmarketable multifunctional 

benefits that are reaped and enjoyed by other people not 

directly involved in rice-farming activities. 

Local wisdom and socio-cultural values attached 

to irrigated rice culture 

Local wisdom or indigenous knowledge and socio-cultural 

values are inherent in rice cultivation. Rural traditions and 

various kinds of rituals and festivals associated with rice farming 

can bring social stability and social harmony. Before the 

advent of the Green Revolution, irrigated agriculture in many 

developing countries was rich in a wide variety of genetic resources. 

Local rice varieties were numerous. Modern agriculture 

has replaced local varieties with their associated local 

wisdom by modern rice varieties requiring expensive agrochemical 

inputs, which in many cases are beyond the capacity 

of local farmers to purchase. 

Moreover, the trade-related intellectual property rights 

recently implemented through international trade arrangements 

have pushed the farmers of developing countries into further 

difficulties. They would not have access to such costly technology 

crucial to enhancing agricultural development in their 

countries. 

Technical, managerial, and financial constraints 

faced by rice farmers 

After the implementation of an irrigation management transfer 

program in many countries, the farmers must be responsible 

for shouldering the cost burden for repairs, operation, 

and maintenance of the systems that were formerly managed 

by the government. However, many rice farmers lack technical, 

managerial, and financial capacity. They still need external 

support to be able to manage the irrigation system properly. 

Threats to the sustainability of irrigated rice culture 

Several conditions or factors may severely threaten the 

sustainability of irrigated rice culture: (1) declining interest of 

rural youth to work as farmers, particularly as rice farmers, (2) 

declining irrigated land area because of conversion for nonagricultural 

uses, (3) increasing conflict in the use of water resources, 

and (4) deforestation and pollution of irrigation water. 

Recommended measures for maintaining irrigated rice 

culture sustainable 

Minimizing irrigated paddy land conversion 

This can be done through (1) careful spatial and land-use planning 

taking water resource availability into consideration, and 

(2) creation of a legal framework prohibiting the use of paddy 

land within the prescribed zone for nonfarming activities, with 

strict law enforcement in its implementation. 

Narrowing the rural-urban gap 

This can be achieved through (1) pro-paddy farmers’ agricultural 

policies that ensure the enhancement of farmers’ income 

and terms of trade of farm products, (2) agricultural-based rural 

industrial development efforts for rural employment and 

income generation, and (3) improvement of rural infrastructure. 

Empowering irrigators’ associations (IAs) 

The following methods can be employed: (1) provision of support 

services such as agricultural credit, market information, 

and so on; (2) provision of training and education to increase 

farmers’ skills and knowledge in various fields; (3) facilitating 

and motivating the potential IAs to perform income-generating 

activities beyond irrigation management; (4) external support 

for selected IAs that are badly in need of major rehabilitation 

through a participatory approach; (5) government recognition 

for IAs as a legal entity so that they can make economic 

transactions and receive credit from financial institutions. 

Reducing water conflict 

This can be endeavored through (1) creating a legal framework 

for clearly defined water rights for different users, (2) 


promoting good coordination among existing IAs both within 

a single “large-scale” irrigation system and intersystem coordination 

along the watershed for equitable water allocation and 

distribution, (3) mobilizing and organizing dialog among water 

users of different sectors to develop a mutual understanding 

on how to use water as a common property for the benefit 

of the community as a whole, and (5) promoting more efficient 

use of the available water. 

Protecting upstream watershed and water quality 

from further degradation 

This can done through the following measures: (1) imposing 

strict punishment for water polluters and illegal woodcutters 

in the protected watershed; (2) refusing the issuance of a permit 

for any project investment for which, from an environmental 

impact, assessment criteria are not feasible; (3) implementing 

a “polluters pay principle”; (5) strengthening the roles 

of existing “community-based” forestry; (6) exploring the possibility 

of transferring government-managed forestry to local 

communities; (7) reducing the excessive use of chemical farm 

inputs and applying organic farming; and (8) strengthening 

interagency coordination in handling water problems. 

Closing notes 

Sustainability of irrigated rice culture should encompass technical, 

institutional, economic, ecological, socio-cultural, and 

environmental sustainability. In this way, it would enable us to 

(1) better understand the location-specific nature of greatly 

diverse irrigated rice culture in a rather holistic perspective, 

and (2) formulate irrigation and rice policy best suited to the 

specific conditions of a given country. This concept of 

sustainability, however, may be applicable mainly to irrigated 

agriculture getting water from farmer-managed irrigation systems 

but might not quite be relevant for those receiving water 

from government-managed irrigation systems. 

One of the greatest threats to the sustainability of irrigated 

rice culture is the rapid conversion of paddy fields to 

nonagricultural uses. As rice culture plays multifunctional roles 

with positive externalities, it must be preserved. Effective and 

appropriate policy measures need to be taken to restrict the 

changing uses of irrigated paddy fields. 

References 

Groenfeldt D. 2003. Multi-functional roles of irrigation with special 

reference to paddy cultivation. Proceedings of Sessions on 

Agriculture, Food, and Water for the Third World Water Forum 

(WWF3), 19-21 March 2003. Kyoto (Japan): World Water 

Council 3rd World Water Forum. p 71-82. 

Kim KH. 2001. Rice production and consumption. In: Kwun S-K, 

Choi HC, Chung S-O, Kim J-C, Kim TC, Heu M-H, Lee BW, 

Lee K-H, editors. Rice culture in Asia. Gyeonggi-do (Korea): 

Korean National Committee on Irrigation and Drainage, and 

International Commission on Irrigation and Drainage. p 14- 

19. 

Kwun S-K. 2002. Multi-functional roles in paddy fields and on-farm 

irrigation. Proceedings of the Pre-Symposium for the Third 

World Water Forum (WWF3), 20-21 March 2002. Otsu, Shiga 

(Japan): World Water Council 3rd World Water Forum. p 55- 

68. 

Mizutani M. 2002. Multi-functional roles of paddy field irrigation 

in the Asia monsoon region. Proceedings of the Pre-Symposium 

for the Third World Water Forum, 20-21 March 2002. 

Otsu, Shiga (Japan): World Water Council 3rd World Water 

Forum. p 37-54. 

Pakpahan A. 1995. Fundamental issues toward sustainable agriculture. 

Perencanaan Pembangunan. Nomor 03/1995:83-91. 

Notes 

Author’s address: Professor Emeritus, Udayana University, Bali, 

Indonesia, e- mail:stw_trisula@yahoo.com. 

Multifunctional roles of irrigation water 

and rice fields in Dujiangyan, China 

Liu Yulong, Yamaoka Kazumi, and Ren Yonghuai 

As a developing country, China has made outstanding progress 

in economic growth during the past 55 years. Especially in the 

last two decades, China has emerged as one of the fastest-growing 

economies in the world (Anderson and Peng 1998). From 

1949 to 2003, grain production in China increased from 113.2 

million to 430.7 million tons and peaked at 504.5 million tons 

in 1996. Meanwhile, the Chinese population has increased from 

0.4 billion to 1.3 billion. China feeds 25% of the world’s people 

with 7% of the total cultivated land in the world. 

However, since 1998, Chinese total rice production has 

declined continuously, from 200 million to 170 million tons. 

Although huge changes in agriculture with the development of 

a social economy and a population peak of 1.6 billion are expected 

in the next two decades, China faces increasing food 

demand and limited arable land and water resources (Liu et al 

2001). From 1996 to 2003, Chinese total cereal production 

declined by 14.6%. The main reasons influencing this decline 

in Chinese cereal production are attributed to a shift in peasants’ 

willingness to cultivate grain crops instead of other economic 

crops, a price fluctuation in the international grain market, 

and the limited land and water. China affects the international 

grain market because of its huge population and 


economy. China cannot rely on the international grain market 

for grain supply, whereas other developing countries can. 

Efficient ways to increase the production of the staple 

food crop rice are to enlarge the planted area and increase 

yield by improving arable land with affluent water resources. 

This environment provides the water-loving rice plant with the 

fundamental conditions necessary to achieve a high yield when 

combined with crossbreeding and gene-mapping technologies. 

More attractive and efficient ways to use water are now crucial 

for farmers in order to encourage them to plant rice because 

rapid economic growth in China has led to high competition 

among water users and requires the introduction of effective 

systems for water charging. Farmers need to wisely 

develop multifunctional uses of paddy fields and water to harvest 

more rice based on hydrologic conditions. Economic externalities 

generated by the multifunctional uses of paddy fields 

and water should be incorporated when the efficiency and 

sustainability of water use is addressed in rice paddy agriculture 

in the Asian monsoon region (Yamaoka et al 2004). 

This paper emphasizes the multifunctional roles of paddy 

fields and irrigation water, and introduces the multifunctional 

social roles of wise water management in the Dujiangyan irrigation 

area. 

Multifunctional roles of paddy fields and irrigation water 

Multifunctional roles are typically involved with fisheries in 

two aspects in the Dujiangyan irrigation area. One aspect is 

the role of irrigation water in fisheries. In most of southern 

China, especially in the Huai, Yangtze, and Pearl river basins, 

farmers have developed irrigation water for both paddy fields 

and fisheries. In the process of irrigation and drainage, farmers 

at the hamlet level have cooperatively built bypassed water-cycle 

courses adjacent to the channels with permission from 

rural authorities and used fishery drainage water to irrigate 

paddy fields (see Fig. 1.). 

These fisheries do not require the digging of ridges in 

the paddy field, but only require the building of fishponds beside 

the irrigation channel. In those areas of Dujiangyan where 

paddy fields are influenced by unstable water supplies because 

of the hilly topography, farmers build fishery ponds upstream 

beside the channel and irrigate rice fields downstream. 

The main construction and management steps are to (1) 

survey the scope of the fishpond beside the channel and calculate 

the differences in water levels between upstream and downstream, 

(2) design the size of the fishpond with the connecting 

water channels, gates, and pipes, (3) lay clay or concrete on 

the bottom and sides of the fishpond to prevent water leakage, 

(4) disinfect the fishpond, and (5) fill the fishpond with water 

and monitor the quality of the water body. 

The fishery pond also works as a retaining reservoir, 

jointly owned by a farmers’ group, for rice paddy fields downstream 

in case of water shortage such as during transplanting. 

A person who wants to manage the fishery does not pay a water 

charge but pays a rental fee for the fishery pond, about 

Channel 

Gate 

Gate 

Gate 

Fig. 1. Fishery beside the channel. 

Gate 

Fishpond 

Irrigated water 

Rice field 

US$400 ha –1 year –1 , to the farmers’ group. This person must 

follow the group’s instructions on water management. In fact, 

this multifunctional water use improves the efficiency of the 

irrigation water and also helps to strengthen farmers’ cooperative 

water management autonomy. 

Another aspect of the role of paddy fields and irrigation 

water is reconstruction of the paddy field with ridges and fossae. 

The ridges need to be about 70 cm wide. The depth of the 

fossae must be in the range of 20–25 cm and all the fossae 

need to be connected. Paddy rice grows on the surface of the 

ridges and fish live in the water in the fossae (see Fig. 2). 

The main processes for accomplishing this role are reconstructing 

the rice field, disinfecting the water and field, 

feeding fish fry, monitoring water volume and quality, managing 

the rice field, and harvesting. 

The fisheries beside the irrigation channels mainly feed 

herring, mandarin fish, crucian, carp, catfish, etc. In the paddy 

fields, the multifunctional technologies mainly provide breeding 

places for the rice field eel, loach, shrimp, crabs, giant 

spiny-frogs, ducks, and golden-fish. Statistical data show that, 

with these wise skills, the farmers get both high rice production 

and a relevant effective income. Compared with 

nonbreeding fields, multifunctional fields harvest more rice, 

approximately 750 kg ha –1 , and produce an economic commodity 

of fish, about 450 kg ha –1 . All these technologies alleviate 

grain shortages and produce extra income for the farmers. 

In addition, farmers are encouraged to be enthusiastic about 

planting rice. 


Soil 

Rice Rice Rice 

Ridge Ridge Ridge 

Fishing water 

Fig. 2. Fishery among rice fields. 

Soil 

Multifunctional social roles of the Dujiangyan irrigation 

system 

Besides technologies for irrigation water management and 

cultivated paddy fields, the irrigation water also fulfills multifunctional 

social roles in the irrigation areas. The Dujiangyan 

Irrigation System has been designated as a World Heritage by 

UNESCO, and its multifunctional social roles provide many 

revelations. This system is a gigantic water project built 2,260 

years ago on the upper reaches of the Minjiang River. The 

Dujiangyan diversion dam lies only 57 km from Chengdu, the 

capital of Sichuan Province, which has a population of approximately 

9 million, and it provides water for cities as well 

as for agriculture. This system is a good example of the development 

of efficient and sustainable water-use management 

through multifunctional social roles from ancient to modern 

times. 

The multifunctional social roles are outlined below: 

1. Economic benefit. At present, this system boasts a 

673,333-ha gravitation irrigation area, not only providing 

the irrigation water to produce 31% of the total 

provincial rice production, but also providing functions 

for urban living, industry, flood prevention, fishery, 

etc., for 30% of the total population of approximately 

85 million in Sichuan Province. The 

Dujiangyan Irrigation System brings prosperity to the 

Chengdu Plain and protects this plain against drought 

and waterlogging. This system has contributed to the 

development of Sichuan’s economy. 

2. Cultural development. The Dujiangyan Irrigation System 

and Mount Qingcheng have been put on the World 

Heritage List. Mount Qingcheng was the birthplace 

of Taoism, which is celebrated in a series of ancient 

temples (UNESCO 2000). It is well known that Taoism 

is the traditional Chinese religious culture. Even 

now, Taoism still influences the behavior of many 

people. In addition, Taoism has provided water management 

ideas and the apotheosis of ancient water 

treatment experiences that have become the base of 

the sustainable modern water management institution. 

This civilization was based on irrigation water and 

reflects the long-term social functions of irrigation 

water. People realized that the “unification of humans 

and nature” is a rule of cultural development. 

3. Biodiversity maintenance. After thousands of years 

of development, the Dujiangyan area formed five 

types of ecosystems: urban, agricultural, forest, high 

and cold, and river. In this area are 3,284 species, 

13% of the total species of Chinese advanced plants 

(Zhuang et al 2004). The unique humid environment 

throughout the vast alluvial fan allows many species 

of animals and plants to survive only in this area. This 

provides a rare biodiversity and gene pool in the 

Dujiangyan area, even for the world. The irrigation 

water provides the environment for this biodiversity. 

On the other hand, the biodiversity also maintains the 

whole water system to produce benefits and a civilization 

resulting in the unification of humans and nature. 

4. Sustainable development revelation. Many world 

heritages are no longer active in history and their thrilling 

stories are reserved in a book for the memory of 

humankind, but Dujiangyan is still active. The combination 

of ancient wisdom and modern technology 

is the sustainable condition for this system. The coordination 

of use, construction, and conservancy is important 

for the development of water resources and 

environmental protection. The administrative authority 

of the Dujiangyan Irrigation System has the ambition 

to expand its irrigation area to 1 million ha. This 

will bring sustainable developmental challenges to the 

irrigation system. 

References 

Anderson K, Peng C. 1998. Feeding and fueling China in the 21st 

century. World Dev. 26(8):1413-1429. 

UNESCO. 2000. Report of twenty-fourth session, World Heritage 

Committee, Cairns, Australia, 27 November-2 December 2000. 

p 45. 

Yamaoka K, Horikawa N, Tomosho T. 2004. Water productivity and 

economic externalities of rice paddy agriculture in the Asian 

monsoon region. The collection of theses of the International 

Academic Forum for 2,260th Anniversary of the Founding of 

the Dujiangyan Irrigation System. p 303-311. 

Yulong L, Yukuo A, Minjian C, Yibin C. 2001. Multi-regression 

analysis of China’s grain production from 1949 to 2050. 

Biosyst. Stud. 1(1):29-36. 

Zhuang Ping, Gao Xianming, Feng Jinbo, Jing Changwei. 2004. 

Biodiversity research and conservation, Dujiangyan. In: General 

situation and characteristics of biodiversity in Dujiangyan. 

Sichuang Science and Technology Press. www.wwfchina.org/ 

csis/shengwudyx2/djy/px12.htm. 


Notes 

Authors’ addresses: Liu Yulong, Department of Water Resources, 

China Institute of Water Resources and Hydropower Research 

(IWHR), 20 Chegongzhuang Xilu, Beijing 100044, China, 

liuyl@iwhr.com; Yamaoka Kazumi, Laboratory of Agricultural 

Water Management, National Institute for Rural Engineering 

(NIRE), 2-1-6 Kan-non-dai, Tsukuba, Ibaraki 305- 

8609, Japan, kyamaoka@nkk.affrc.go.jp; Ren Yonghuai, 

Laboratory of Agricultural Water Management, National Institute 

for Rural Engineering (NIRE), 2-1-6 Kan-non-dai, 

Tsukuba, Ibaraki 305-8609, Japan, ren@nkk.affrc.go.jp. 

Decommissioning of paddy lands in the Wet Zone 

of Sri Lanka: some effects on food security and ecosystems 

Kusum Athukorala and Missaka Hettiarachchi 

Rice is the staple food and main agricultural commodity of Sri 

Lanka, where 17.7% of the gross domestic product comes from 

the agricultural sector. Highly water-intensive rice cultivation 

consumes more than 70% of the total water allocated for food 

production in the country. 

Paddy lands in Sri Lanka can be categorized into three 

major types depending on the water usage: 

Major and medium irrigated systems—264,602 ha in 

1999-2000, 50% of the total area sown 

Minor irrigated systems—129,604 ha in 1999-2000, 

 

23% of the total area sown 

Rainfed systems—155,040 ha in 1999-2000, 28% of 

the total area sown 

Out of this, the rainfed systems are mainly found in the 

Wet Zone of the country. The Wet Zone is a climatic subdivision 

in the southwestern part of the country, which gets up to 

2,000–2,500 mm of annual rainfall. It constitutes about onethird 

of the country’s total area and accommodates about 63% 

of the population. The major part of heavily urbanized areas, 

including the cities of Colombo, Kandy, and Galle, is situated 

in the Wet Zone. Because of the high population density and 

high level of urbanization, demand is overwhelming for land 

in the Wet Zone. The population density in Colombo District 

is 3,305 persons per km 2 versus 957 persons per km 2 in the 

Wet Zone. In the Dry Zone, this decreases to 163 persons per 

km 2 . Competition for land is therefore strong and, because of 

the availability of other employment, labor costs are increasingly 

high. 

In recent years, rice or paddy farmers (as they are termed 

in Sri Lanka) were either attracted by more lucrative alternatives 

or demotivated by the competition from rice imported 

from Pakistan, Vietnam, or Thailand, and tended to abandon 

rice farming. Table 1 shows rice imports by quantity and as a 

percentage of total national consumption. 

The inevitable consequence of this pressure is that most 

of the traditional Wet Zone farmers are pushed toward selling 

their paddy lands in lucrative deals for real estate development. 

This has significantly decreased the total area of rainfed 

paddy cultivation schemes during recent times (see Fig. 1). 

From the statistics, it is apparent that the percentage of 

rainfed paddy fields, which was as high as 45% of the total 

cultivated area in the early 1960s, had dropped to about 28% 

by 2000. 

The impact of the loss of rainfed paddy land 

on national food security 

A study of rice imports in the past few years showed a significant 

growth of rice imports in 1996-2000. This could be mainly 

attributed to the severe shortfall of rain that the country had to 

face intermittently during this period. Importing water-intensive 

goods to a country where the water stress is high from 

another country where the water stress is comparatively less is 

a form of virtual water trade, and it is a practice encouraged 

by most of the modern water ideologues since the Dublin-Rio 

conference onward. 

But, unfortunately in the Sri Lankan situation, most paddy 

lands affected by a shortage of rain are in the irrigated systems 

in the Dry Zone, whereas water stress in the Wet Zone is much 

less even during drought; therefore, the rainfed systems in the 

Wet Zone could be cultivated to a certain extent even during 

relatively dry years. Because of the loss of paddy cultivation 

Table 1. Rice imports by year. 

Item 

Year 

1994 1995 1996 1997 1998 1999 2000 2001 2002 

Rice imports in million SLR 655 122 5,118 4,331 2,621 3,290 288 969 1,732 

Imports as a percentage 2 0.3 16.5 13.7 6.2 7.5 0.5 2.0 3.3 

of total local production 

Source: www.statistics.gov.lk. 


Sown area (ha) 

250,000 

200,000 

Season 1 

Season 2 

150,000 

100,000 

50,000 

0 

1961-62 

1963-64 

1965-66 

1967-68 

1969-70 

1971-72 

1973-74 

1975-76 

1977-78 

1979-80 

1981-82 

1983-84 

Years 

1985-86 

1987-88 

1989-90 

1991-92 

1993-94 

1995-96 

1997-98 

1999-2000 

Fig. 1. Variation in total sown area in the rainfed system from 1961 to 2002. 

External debt and trade balance (billion US$) 

10 

8 

External debt 

Trade balance 

6 

4 

2 

0 

–2 

–4 

1990 1991 1992 19931994 1995 1996 1997 1998 1999 2000 2001 2002 

Year 

Fig. 2. External debt and trade balance. 

in the Wet Zone, Sri Lanka is pushed toward dependency on 

imported rice, hence the import of virtual water, placing a heavy 

burden on the already precarious balance of payments while a 

significant amount of real water is being lost as runoff because 

of the limitation of cultivable land in the Wet Zone. Therefore, 

this is more akin to a trade of virtual land than virtual water. In 

this process, not only is there a loss of water, but also food 

security and the trade balance of the country are at stake. Because 

of the competition from cheap rice imports, small-scale 

farmers in the rainfed system are further discouraged and are 

tempted by high land prices to decommission their paddy land, 

thus creating a vicious circle (Fig. 2). 

The impact of the loss of rainfed paddy land 

on the water sector 

From the above section, it is clear how the decommissioning 

of rainfed paddy lands contributes to wastage of a significant 

volume of water that could have been used for cultivation. In 

all the recent environmental resolutions from Dublin-Rio to 

Rio+10 conferences, it has been stipulated that uses of water 


are further defined as water for people, and water for food and 

ecosystems. The immediate repercussion of decommissioning 

a paddy land will be an impact on water for the food sector. 

With further investigation, it becomes apparent that there are 

negative effects on water for people and water for ecosystems, 

too. 

There are two types of rainfed paddy lands in Sri Lanka: 

the lowland rainfed paddy lands and the up-country terraced 

paddy fields. Because of the cultivation methods used traditionally 

for centuries in Sri Lanka, paddy lands act as sponges 

that absorb and retain water, thus enhancing groundwater recharge. 

When paddy lands are sold as real estate, where the 

land will be filled, separated into small blocks, and sold for 

urban housing, more than 70% of the land is paved, which 

reduces groundwater recharge significantly. Housing also encroaches 

on the neighboring watersheds. Groundwater reduction 

for this reason is mainly felt in the up-country mountainous 

regions where many water supply schemes are supplied 

by small streams. In valley areas such as Peradeniya, Gelioya, 

and Kundasale in the Central Province, where there had been 

significant filling of traditional paddy lands, there are major 

flow fluctuations in small streams and springs that used to be 

perennial (e.g., Nanu Oya, Kandy District, Central Province). 

This has a direct negative effect on drinking water since water-supply 

authorities have begun to experience difficulties in 

maintaining the minor stream-fed supply points. 

The result could be that the water-supply authorities 

would have to pump water from a distance with greater treatment 

and conveyance costs. This would be a major economic 

implication for the nation because of the loss of groundwater. 

The effects in the lowlands take on a different twist. Filling of 

paddy lands increases runoff; hence, in sudden storms, runoff 

surpasses the flood limit, quite frequently causing severe flash 

floods. This is exacerbated in the high-intensity rains currently 

experienced because of El Niño. Economic implications in this 

case are much more intense than merely the loss of groundwater. 

In low-lying areas of Colombo District such as Kotte and 

Kollonnawa, where a significant amount of paddy land was 

filled for urban housing, flash floods have become a major 

and regular problem. Even the Parliament of Sri Lanka in the 

low-lying area of Sri Jayewardenepura Kotte has been flooded 

twice. The Greater Colombo Flood Control and Environmental 

Improvement Project (1993-2004) has identified the filling 

of paddy land as one of the major causes of urban floods. The 

total cost of the project is about 15,000 million SLRs (US$150 

million) and it has had to resettle about 900 families, creating 

another set of significant socioeconomic and environmental 

problems. 

Remedies 

From the above facts and statistics, it is obvious that there are 

significant economic implications of decommissioning rainfed 

paddy lands that need to be reviewed. Since most of these lands 

are used for urban housing, a direct prohibition on selling paddy 

land might cause conflicts in other sectors. Therefore, it is 

necessary that mitigatory measures such as local authorities 

imposing an extra tax on paddy land transactions or filling be 

considered. This could be based on the expense of commissioning 

a similar piece of land in the Dry Zone irrigated systems. 

It should also include an allowance for damage from 

groundwater loss and flash floods. The estimation for this total 

loss is about $1,700–2,000 per ha; policymakers could arrive 

at their rates based on the above value. A participatory 

awareness program, introducing alternate high-value crops 

(such as orchids) to encourage Wet Zone farmers, must be put 

in place immediately to continue with the maintenance of rice 

paddies in an ecologically appropriate manner. This could be 

useful in mitigating the negative impacts of decommissioning 

paddy lands. 

References 

Official Web site of the Central Bank of Sri Lanka. 

www.centralbanklanka.org. 

Department of Census and Statistics, Sri Lanka Web site, 

www.statistics.gov.lk. 

Greater Colombo Flood Control and Environmental Improvement 

Project (1993-2004): Report on the Badovita Resettlement 

Project. 

Notes 

Authors’ addresses: Kusum Athukorala, Netwater, No. 7, St. Mary’s 

Lane, Colombo 15, Sri Lanka, e-mail: kusum@itmin.net; 

Missaka Hettiarachchi, Department of Earth Resources, University 

of Moratuwa, Moratuwa, Sri Lanka, e-mail: 

nadalochana@yahoo.co.uk. 


True agro-biodiversity depending on irrigated rice cultivation 

as a multifunction of paddy fields 

Kazumasa Hidaka 

As one of the multifunctions of irrigated rice cultivation, biological 

diversity in rice paddies has been a critical issue in 

Japan and some Asian countries under free marketing of the 

WTO. Especially, in Japan, conservation of biodiversity and 

ecosystems in agriculture has been an important issue in agricultural 

and environmental policy since 1999, when a changed 

agricultural foundation law stimulated progress by the Japanese 

government and EU. Agro-biodiversity has now become 

a popular password in many nations, and the Internet now has 

30,000 articles on it. 

According to the literature before, about 2,000 species 

of plants and animals associated with rice paddy fields have 

been recorded in Japan (Hidaka 1998). This biodiversity in 

species richness can be regarded as a general characteristic in 

the rice paddy ecosystem. Actually, 645 animal species were 

recorded in the Philippines (Cohen et al 1994) and 765 species 

of arthropods were examined in Indonesia (Settle et al 

1996), for example. For rice, 212 species were recorded in the 

Philippines (Heong et al 1991). 

We must not ignore, however, that there are actually two 

types of agro-biodiversity in rice paddies: one is “apparent 

agro-biodiversity” and the other is “true agro-biodiversity.” 

Usually, we have been evaluating agro-biodiversity as a value 

for the number of species collected in arable land, but not all 

species of plants and animals truly depend on the agricultural 

ecosystem. In the case of rice paddy fields, some species may 

be stranger species or may depend on other wetland habitats. 

There are many kinds in natural or managed ecosystems with 

the exception of rice paddy fields: ponds, lakes, rivers, streams, 

and reservoirs in the limnological viewpoint. Furthermore, 

species components of agro-biodiversity in rice paddies may 

vary depending on the preserved situation of natural wetlands. 

Approaches toward evaluating true agro-biodiversity 

To discriminate the true species depending on rice cultivation 

as true agro-biodiversity, there are three ways to examine this: 

natural history, review work, and field work in ecology and 

systematics: 

1. Review work in the literature (classical natural history). 

2. Collection of specimens attached to a living habitat 

(new natural history) (Mineta, Hidaka, and Enomoto, 

in preparation). 

3. Ecological field work of each species population associated 

with agricultural practices (agroecological 

field work). 

In this presentation, some typical studies, which we or 

other pioneers have conducted, will be reviewed from the viewpoint 

of agroecological principles. 

Case studies 

Examining natural history 

More than 200 plant species, 20 species of amphibians and 

reptiles, 30 species of fish and birds, and more than 1,000 species 

of arthropods can be reviewed as apparent agrobiodiversity 

in rice paddies (Hidaka 1998). A few collecting 

studies were conducted with scientifically appropriate methods 

of sampling. Kobayashi et al (1973) examined in detail 

the biodiversity of arthropods using sweeping methods in rice 

paddies of several sites of Tokushima Prefecture in the late 

1950s. More than 450 species of insects, spiders, and mites 

could be recorded in spite of the methodological faults of 

sweeping methods that can mainly collect arthropods only on 

rice plants. Nowadays, if we had a chance to examine arthropod 

fauna in the same fields, it might be difficult to collect a 

closed number of species in rice paddies such as in the 1950s. 

An example of this examination for species number must be 

significant and important to evaluate the agro-biodiversity level 

as a first step. But many species are stranger species in rice 

paddies, which the authors pointed out. We have to examine 

the fauna in other wetland ecosystems as well as rice paddies 

to evaluate true agro-biodiversity. Recently, Saijo (2001, 2002) 

tried to examine aquatic insect fauna in both rice paddies and 

irrigated ponds in Shimane Prefecture to make a clear reproduction 

of each species site. These studies are significant in 

the ecological sciences to reveal the dependency of each species 

on rice paddy ecosystems and to evaluate true agrobiodiversity. 

For botanists, Kasahara (1947) examined a list of rice 

paddy weed specimens in 25 prefectures over Japan and recorded 

174 plant species. All botanical species could be classified 

as serious harmful weeds (96 species) and semiwild 

weeds (78 species) in rice paddy ecosystems. 

Habitat records of specimens 

Recording habitat when collecting specimens in natural history 

work is important information for evaluating the agricultural 

dependency of each species as true-biodiversity. Although 

there are very few cases that recorded the habitat of collected 

specimens, the same specimen wetland plants had details on 

the collecting place such as rice paddy, pond, irrigation way, 

or natural wetland. Some botanists had been recording habitats, 

such as Kasahara’s collection and Fujii (2001). So many 


25 

L. indica subsp. 

incnopnyiia 

N. coreana 

75 

25 

M. quadrifolia 

R. extorris G. japonica 

50 

R. cantoniensis A. japonica 

50 

M.korsakowii 

V. undulata 

A. imbricata 

R. nipponicus 

Rice paddy fields 

H. oliganthum 

(%) 

100 R. leptopetala var. littorea 

R. pusilla, M. minima 

A. humidum 

N. indica 

Irrigation ponds 

P. chinense 

S. natans Blyxa spp. 

U. minor 

Isoetes spp. 

0 

100 75 50 25 

100 

0 

Food plains in rivers 

Fig. 1. Frequency of habitat records among endangered aquatic 

weeds in Okayama Prefecture according to botanical specimen 

collections (Mineta, Hidaka, and Enomoto, in preparation). 

data were accumulated so that dependency on rice paddies 

could be demonstrated in weed species in irrigated rice cultivation 

(Fig. 1; Mineta, Hidaka, and Enomoto, in preparation). 

Constructing a database of specimen habitat records using a 

GPS system is a key methodology for evaluating true agrobiodiversity. 

75 

Agroecological study of local populations 

and communities 

The most evidential and practical methods among these three 

approaches are agroecological field work at the population or 

community level for each species. For the community level, 

generally, examinations of an assembly of biological species 

on grown rice plants such as Kobayashi et al (1973), Heong et 

al (1991), and Une et al (1989) are correctly regarded as agrobiodiversity. 

There is another assembly of aquatic animals following 

the irrigation system in rice paddies for agrobiodiversity 

such as Katano (2001) in fish or Saijo (2002) in 

insects. 

At the population level, it is difficult for all species, one 

by one or gene by gene, to reveal their dependency on rice 

cultivation by analyzing basic data, which used to require tasks 

such as genetic analysis of one local population. If we could 

make their dependency on paddy fields clear for a particular 

species population, we could use the most practical tactics for 

the population in agricultural policy. Therefore, it is reasonable 

to start from an emergent case such as endangered species, 

which need to conserve their population soon in an environmental 

policy (Hidaka 1998). Recently, some ecological 

field workers have started detailed studies of population biology 

for RDB animals and plants to demonstrate their habitat 

use quantitatively. Some species, including endangered populations, 

depend on agricultural practices in rice paddy fields. 

To make agricultural dependency clear, all wetlands in not only 

rice paddies but also in ditches, ponds, or streams and rivers 

and other natural wetlands should be examined in detail over 

several years. Many species may use several kinds of ecosystems 

to round out their life cycles as their indispensable habitats. 

Our examinations for the giant water bug Lethocerus 

deyrollei (Hemiptera) and the diving beetle Cybister japonicus 

(Coleoptera), representative endangered aquatic insects, have 

been conducted in hotspots to reveal their life history and agricultural 

dependency on rice paddy fields since 2000. These 

local populations of each species depend on irrigated rice fields 

and agricultural practices for their reproduction, according to 

results from our field work (Fig. 2). Furthermore, L. deyrollei 

populations depend on frogs, which strongly depend on the 

rice paddy ecosystem as a reproductive habitat (Hirai and 

Hidaka 2002). Their dependency on irrigated rice cultivation, 

however, might not be permanent. It is necessary for their agricultural 

dependency to conduct similar field examinations in 

different places and different climatic conditions. Often, the 

habitat use pattern of a species population has changed in a 

fluctuating environment. For more adequate conservation management, 

it is necessary to conduct ecological field work that 

can evaluate the habitat use of each local population quantitatively, 

such as ornithologists’ work (Elphick 2000, Fujioka et 

al 2001, Maeda 2001). 

Integrating conservation and agricultural production 

For endangered species in rural areas where rice has been cultivated, 

some conservation practices are necessary immediately 

to recover or sustain the population level so that it won’t 

become threatened. For other conservation practices, we have 

to sustain agricultural production. Especially in the case of 

conserving agriculturally dependent species such as the giant 

water bug, maintaining agricultural activity is necessary in some 

hotspots of true agro-biodiversity. Therefore, integration of the 

conservation of biodiversity and agricultural production is 

important and necessary. Recently, a new concept, IBM (integrated 

biodiversity management), has been proposed in Japan 

(Kiritani 2002). Practically, to manage biodiversity well in 

agroecosystems, finding and making devices that can buffer a 

conflict between biodiversity and food production are desirable 

for both hardware and software. 

In our research field, some IBM trials have started to 

manage an endangered population of the giant water bug associated 

with biodiversity and sustainable agriculture in several 

practices of agricultural, educational, and scientific research. 

For our scientific researchers in agroecology, quantitative 

analyses of population dynamics can produce optimal results 

for the management of endangered species populations. Analyzing 

an endangered population by genetic methods at the 

molecular level, furthermore, must result in not only management 

of the population but mitigation for its appropriate conservation. 

In addition to studies of an intrapopulation, impor- 


Conventional rice cultivation practices 

Tillage and 

puddling fertilizer 

Transplanting and 

pesticides 

Drainage 

Pesticide 

with IPM 

Harvest 

Tillage 

MAR APR MAY JUN JUL AUG SEP OCT NOV DEC JAN 

Overwintering Immigration In rice paddy 

Movement among paddy fields 

Mating and reproduction 

Overwintering 

Almost disappearing 

from aquatic habitats 

Dispersal and repeated reproduction 

Larval development 

Emigration with drainage 

Other wetland habitats 

(ditches, ponds, rivers) 

Water level of irrigation 

Higher 

Irrigation 

start 

New adults 

Dispersal 

Emigration 

Life stages of the paddy bug 

Fig. 2. A schematic description of the life cycle of the giant water bug, Lethocerus deyrollei, rice 

cultivation practices, and irrigation water level in a hotspot field of agro-biodiversity. 

tant research at the interpopulation level is necessary for the 

conservation of endangered species. Generally, each species 

coexists with more than 20 species, at least for sustaining their 

lives. Therefore, examinations of population dynamics should 

be included in community ecology for optimal conservation 

planning. It was demonstrated that diverse amphibian species 

are important for supporting food resources for a carnivorous 

species of the giant water bug (Hirai and Hidaka 2002). On 

the other hand, educational work must be important for villagers 

and consumers in the city to promote the significance of 

biodiversity in agriculture. In 1960s Japan, some fish, snails, 

and insects such as grasshoppers in rice fields were sold and 

eaten. Good devices often lurk behind our traditional life style. 

Recently, several ecological workers have been conducting 

a detailed population analysis in a residual population of 

threatened animals and plant species and conservation with 

villagers in Japan. However, conservation biologists are not 

enough to produce a successful regeneration of an endangered 

species, although it is in emergency for its conservation. 

Conclusions 

To understand and manage true agro-biodiversity in rice paddies, 

we need to organize an integrated ecosystem management 

system such as IBM as soon as possible. First of all, it is 

necessary to answer a question about what kinds of species 

depend on the rice paddy ecosystem in not only negotiations 

with the WTO but also in building up a consensus of many 

people in villages on management practices for agrobiodiversity. 

For this, the integration of scientific research and 

education must always be the most practical. 

References 

Elphick CS. 2000. Functional equivalency between rice fields and 

seminatural wetland habitats. Conserv. Biol. 14(1):181-191. 

Heong KL, Aquino GB, Barrion AT. 1991. Arthropod community 

structures of rice ecosystems in the Philippines. Bull. Entomol. 

Res. 8:407-416. 

Hidaka K. 1998. Biodiversity conservation and environmentally regenerated 

farming system in rice paddy fields. Jpn. Ecol. 

48:167-178. 


Hirai T, Hidaka K. 2002. Anuran-dependent predation by the giant 

water bug Lethocerus deyrollei (Hemiptera: Belostomatidae), 

in rice fields of Japan. Ecol. Res. 17:661-665. 

Kasahara Y. 1947. Studies on the characteristics of weeds and their 

geographical distribution (1). Nogaku Kenkyu 37(1):8-9. 

Kiritani K. 2000. Integrated biodiversity management in paddy fields: 

shift of paradigm from IPM toward IBM. Int. Pest Manage. 

Rev. 5:175-183 

Kobayashi T, Noguchi Y, Hiwada T, Kanayama K, Maruoka N. 1973. 

Studies on the arthropod associations in paddy fields, with a 

particular reference to insecticidal effect on them. 1. General 

composition of the arthropod fauna in paddy fields revealed 

by net-sweeping in Tokushima Prefecture. Konchu 41(3):359- 

373. 

Maeda T. 2001. Patterns of bird abundance and habitat use in rice 

fields of the Kanto Plain, central Japan. Ecol. Res. 16(3):569- 

585. 

Saijo H. 2002. The role of loach-farming paddy on the life cycle of 

aquatic insects inhabiting lentic habitats. Jpn. J. Ecol. 51(1):1- 

12. 

Une Y, Hidaka K, Akamatsu T. 1989. A field guide of paddy insects 

for reduced farmers. Tokyo (Japan): Nobunkyo Pub. Co. 

86 p. 

Notes 

Author’s address: College of Agriculture, Ehime University, Japan, 

e-mail:sunhwkaz@agr.ehime-u.ac.jp. 

Enhancing the multifunctionality of floating-rice farming 

in the Chao Phraya delta of Thailand 

Yuyama Yoshito, Ogawa Shigeo, and Ueda Tatsuki 

Paddy farming in the Asian monsoon area is closely connected 

to the rural social system. Paddy farming has important roles 

and functions. Recently, energetic efforts have been made to 

spread such recognition worldwide and to evaluate it quantitatively 

(Japanese National Committee of ICID 2003, MAFF 

2003, JIID 2003). In this paper, we introduce the 

multifunctionality of floating rice farming in the Chao Phraya 

delta of Thailand and its possibility for enhancement by water 

management. 

Farming and water management in the Chao Phraya delta 

The Chao Phraya River basin has a catchment area of 162,000 

km 2 , including 1.34 million ha of low-lying delta area. The 

Greater Chao Phraya Project drastically changed farming and 

water management in the delta. Only one crop per year was 

possible there before the project. The original purpose of the 

project was supplementary irrigation in the rainy season. Its 

purpose has now expanded to irrigation in the dry season and 

domestic water supply, among others. The delta area has become 

one of the largest rice granaries in the Asian monsoon 

area. 

The present irrigable area in the delta is 1.08 million ha. 

Rice cultivation for more than five times per two years is possible 

where water accessibility is good. New trials of rice cultivation 

in the dry season started only about 30 years ago. The 

rice cultivation area increased rapidly, but soon reached its 

limit because of a shortage of water resources and insufficient 

capacity of existing irrigation facilities. Farming and land use 

changed rapidly again in the last 15 years. Water management 

became more complicated than before because water demand 

increased and competition among water users became more 

serious (Francois et al 1999, 2001). The problems of water 

shortage (Roongrueng et al 1996) and flooding (Pramote 1999) 

have to be examined in parallel. 

Water management is mainly carried out under the authority 

of the Royal Irrigation Department (RID). To prevent 

water conflict, water allocation in the dry season from January 

to June is very important. Hereafter, the period from January 

to June is referred to as “WM (water management) dry season,” 

while the period from July to December is referred to as 

“WM wet season.” 

Multifunctionality of floating-rice farming 

Floating-rice farming in the delta has played important roles. 

It can be summarized as having low input and low yield but 

sustainable farming. Cultivated floating-rice area decreased 

from 228,000 ha in 1987 to 114,000 ha in 1997 (CTI et al 

1999). Floating rice can grow flexibly according to irregular 

increases in water level. It grows in a dry field with weeds at 

the beginning stage. Then, the stem increases in length from 2 

to 5 m according to water conditions. The cultivation calendar 

matches the natural water conditions. It takes 7–9 months from 

planting to harvest. The harvest starts in December or January 

after the stored water in the field is drained. 

In the past, people in the delta had to follow natural water 

conditions. Recently, the water level and flow conditions of 

the main rivers and canals have been controlled to some degree 

by the operation of diversion dams and regulators along 

both the main irrigation and drainage canals. A unit of floating 

rice area is from a few thousand to 10,000 ha. Each unit has a 

so-called “drainage box” that has a drainage regulator to control 

the water level upstream. Newly constructed big roads 

function as embankments. Therefore, the location of flooding 

areas is also partially regulated by water operation. Scattered 


Specific volume of standing floodwater 

Paddy fields located in low-lying area 

Paddy fields located in comparatively 

elevated area 

Residential area located in low-lying 

area 

Forest 

Allowable limit 

for paddy fields 

1 2 5 10 20 50 100 200 

Flood prevention function 

Damage from flood 

Return period of rainfall (y) 

Flood mitigation function that 

paddy fields potentially have 

Design return period for 

drainage of paddy fields 

Flood mitigation function 

Paddy rice where drainage improvement was implemented 

Residential area with paddy fields in upstream side 

Residential area without paddy fields in upstream side 

1 2 5 10 20 50 100 200 

Return period of rainfall (y) 

Fig. 1. Volume of standingflood water and flood disaster damage 

in respective land use under different rainfall magnitudes. 

inundation prevents serious flood damage. Figure 1 represents 

the general image of flood mitigation function by the paddy 

field. 

Problems of flooding in the delta often occur in October 

and November. A paddy field cultivating a high-yielding variety 

cannot receive too much water during the flood. However, 

floating-rice area has a great possibility to receive surplus water, 

which contributes to flood mitigation at the regional level. The 

storage volume of the floating-rice area can be estimated at 

2,280 million m 3 by assuming that the area is 114,000 ha and 

the water depth is 2.0 m. The volume would be almost the 

same as the storage in the remaining paddy fields in the WM 

wet season, assuming the water depth were 0.2 m. 

If we could convey surplus water to the floating-rice area 

to decrease the peak flood discharge with a depth of 25 cm (5 

cm day –1 for 5 days), the total storage capacity would become 

285 million m 3 . We can call this the buffer function. That volume 

is equal to the inflow discharge of 660 m 3 s –1 for 5 days. 

This is compared with the storage volume of dams and the 

discharge of water allocation in the WM dry season. It is impossible 

to protect metropolitan Bangkok from flooding without 

paddy fields. 

In addition, the river and canal network also has a buffer 

function for flood mitigation. If the rivers or canals upstream 

and midstream can store surplus water temporarily, they help 

to limit a peak discharge. If a river or canal with a length of 

100 km and width of 200 m could increase the water level by 

30 cm without any damage from overflow or breaking of an 

embankment, the storage volume would become 6 million m 3 . 

This volume is equivalent to 70 m 3 s –1 , but it is effective on 

only one day. That is achieved by conveying surplus water in 

the Chao Phraya River to the Lopburi River, the Noi River, 

and the main irrigation and drainage canals. This function to 

control the timing of release of surplus water to the downstream 

of the Chao Phraya River is significant because the 

lower delta is a more serious flood-prone area affected by high 

tide. Moreover, the function of the river and canal network 

that conveys surplus water to the floating-rice area is important. 

Water released from the floating-rice area can be used 

in downstream areas and it contributes to decreasing the salinity 

concentration at the beginning of the WM dry season. EC 

(electrical conductivity) in the irrigation and drainage canals 

of the upper east bank of the delta ranged from 15 to 30 mS 

m –1 . Compared with the EC standard for irrigation of 70 mS 

m –1 , the measured values were relatively low. The released 

water also functions to increase the water level and discharge 

in the Chao Phraya River. It is important for navigation and 

the tap-water supply. In other words, it can decrease the release 

discharge from the Chao Phraya diversion dam and save 

water resources at upstream storage dams. The floating-rice 

field is also a living place for fish. It provides aquatic products. 

Ducks are sometimes kept in the floating-rice field after 

it is harvested. There is little soil erosion in this kind of farming. 

It produces oxygen. The landscape of the floating-rice area 

provides opportunities for eco-tourism. 

Enhancing multifunctionality by water management 

The positive multifunctionality of floating-rice farming can 

be enhanced by water management because the floating-rice 

area is a buffer space in terms of water flow. The RID is daily 

monitoring hydrology, meteorology, and water operation information 

at many stations and facilities. Monitored information 

should be effectively used for decision making on water 

operations. In the decision support system for water operations, 

the monitoring of existing conditions, their summary, 

and the judging of facility operations learned from past records 

and experience are essential. Many proposals at the planning 

stage of water allocation and at the practical water operation 

stage as well as remarks on flood control were made (Yuyama 

et al 2003a,b). Table 1 shows the methods to enhance three 

functions of floating-rice farming and possible impact. New 

methods to enhance the existing function must match technology 

development and public agreements. A quantitative study 

is needed to clarify the relationship among release discharge 

from the dams, tidal level, water levels along the river, and 

salinity concentration. Moreover, we have to keep in mind that 

one preferential enhancement function often has a negative 

impact on other functions. 


Table 1. Methods to enhance the functions of floating-rice farming and possible impacts. 

Preferential 

enhancement function 

Flood mitigation Development of water resources Food production 

Method (how) 

Induced 

positive 

impact 

What 

When 

Where 

For whom 

Induced negative impact 

or remarks 

Preliminary release of standing water 

from floating-rice area 

Conveyance of floodwater into floatingrice 

area 

Insurance contract between government 

and farmer for receiving surplus floodwater 

Heightening of embankment 

Advanced operation of drainage regulator 

Development of decision support system 

of infrastructure for information and 

communication technology 

Peak period of flood discharge 

Flooding period (October, November) 

Downstream delta, especially Bangkok 

People in downstream delta 

Increase in risk in farming and harvest 

yield 

Release of standing water considering the 

timing of water use downstream 

Change of drainage point to use water 

resources more effectively 

Arrangement of new facility such as 

pump, regulator, and regulating pond 

Modification of cultivation pattern in 

downstream fields 

Development of a decision support system 

Provision of new water resources 

From middle of December to end of January 

Downstream fields 

Whole basin 

Farmers of downstream fields 

People in the basin 

Increase in salinity in the Chao Phraya 

River 

Decrease in water level in the downstream 

Chao Phraya River (it is sometimes 

unsuitable for navigation and intake 

for tap water) 

Introduction of new farming technology 

for dry-season cultivation 

Drainage improvement 

Arrangement of irrigation facility 

New water allocation 

New farm products 

From February to May 

Floating-rice field 

Farmers 

Consumers 

Deterioration of water quality and soil 

Increase in water demand 

New water conflict might occur 

Change in ecological environment 

Conclusions 

Whether the people are aware or not, multifunctionality of floating-rice 

farming in the Chao Phraya delta of Thailand really 

exists. The functions of flood mitigation and cultivation of water 

resources for downstream areas are obvious. If we could limit 

the flood peak discharge of 285 million m 3 , new dam construction 

would not be needed. In addition, we could obtain 

new water resources of 2,280 million m 3 . This is equivalent to 

the water demand for a rice cultivation area of 11,000–22,000 

ha. This would help to increase the flexibility of water operations. 

Other functions such as water quality conservation, contribution 

to salinity control, maintaining navigation, and provision 

of eco-tourism opportunities are also important. 

Multifunctionality would be enhanced by water management 

accompanied by a decision support system for water operations. 

References 

CTI Engineering Co., Ltd. and INA Corporation. 1999. The study 

on integrated plan for flood mitigation in Chao Phraya River 

basin. Final Main Report, RID, JICA. 

Francois M, Sripen D, Chatchom C, Alexandre J, Yuphaa L. 1999. 

Improvement of rice cultivation and water management in the 

flooded area of the Central Plain of Thailand. DORAS Center. 

Francois M, Chatchom C, Thippawal S, Jesda K. 2001. Dry-season 

water allocation and management in the Chao Phraya Delta. 

DORAS Center. 

Japanese National Committee of ICID. 2003. Proceedings of WWF3 

Agriculture, Food, and Water. 

JIID. 2003. A message from Japan and Asia to the world water discussions. 

MAFF. 2003. The global diversity of irrigation. 

Pramote M. 1999. Development and achievements in flood control 

and management in Thailand, regional cooperation in the 

twenty-first century on flood control and management in Asia 

and the Pacific. New York, N.Y. (USA): ESCAP, UN. p 59- 

111. 

Roongrueng C, Charoon K, Siripong H, Natha H. 1996. Master planning 

of water resources development in Thailand (1994-2006). 

Trans. of 16th ICID Congress. Cairo, Q. 47, R.1.05. p 55-67. 

Yuyama Y, Pongsak A, Phonchai K, Chatchom C, Athaporn P, Shioda 

K. 2003a. Strategy to improve water management in the Chao 

Phraya delta. Rural Environ. Eng. 44:42-59. 

YuyamaY, Pongsak A, Shioda K, Onimaru T, Nakazawa N, Fujisaki 

T. 2003b. Improvement of water allocation planning and practical 

operation in the upper east bank of the Chao Phraya delta. 

Technical Report of the National Institute for Rural Engineering. 

No. 201. p 93-124. 


Notes 

Authors’ address: National Institute for Rural Engineering, 2-1-6, 

Kannondai, Tsukuba, Ibaraki 305-8609, Japan, e-mail: 

yuya@nkk.affrc.go.jp. 

Acknowledgments: The authors worked for the Modernization of 

Water Management System Project (MWMS) at RID. This 

project is under implementation facilitated by the Japan International 

Cooperation Agency (JICA). The authors greatly appreciate 

all the people involved in the project. 

Demonstration program on multifunctionality of paddy fields 

in the northeastern region of Thailand 

Chatchai Boonlue 

The principal grain crop in the Lower Mekong River Basin 

(LMB) is rice. In fact, nearly 90% of the world’s rice is produced 

in the Asian countries including this region. It is commonly 

recognized that “paddy fields” have many functions and 

advantages or disadvantages but consume a large amount of 

water. The value of irrigation water encompasses more than 

the net economic returns of the rice produced. The multiple 

values of irrigation are particularly evident in paddy cultivation 

where water supports a sustainable and ecological water 

cycle control, which is the basis for the characteristic sociocultural 

system of Southeast Asia. 

Northeast Thailand has an area within the LMB of 

513,113 km 2 and runoff showing as much fluctuation as 80– 

85% in the wet season but very scarce in the dry season. Accordingly, 

water control functions (flood prevention, groundwater 

recharge, prevention of soil erosion) are essentially required 

to control flood and drought problems. 

In this connection, we need to know more precise information 

about the multifunctional roles and multipurpose uses 

of paddy-field irrigation. For this region, the Demonstration 

Program on Multifunctionality of Paddy Fields over the 

Mekong River Basin (DMPF) is developed by the Mekong 

River Committee (MRC) member countries to collect data and 

information on the area that help to demonstrate the 

multifunctionality of paddy fields in the LMB. 

Thailand has expected that the outputs of DMPF will 

provide useful information for the effective use of water to 

achieve high production in paddy fields under appropriate 

water, soil, and crop management. 

The overall concept of the DMPF aims at promoting 

agriculture and irrigation development within the MRB in a 

sustainable and environmentally sound manner, and with good 

participation and cooperation of all concerned. This will serve 

as a practical and integrated approach for agricultural development 

at the farm, subbasin, and basin levels. Achievement 

of this program will benefit local people, scientists, and managers 

of relevant agencies. Local farmers would learn how to 

cultivate rice with an optimum or minimum use of water and 

how to manage the consequent impact from rice growing. Scientists 

would gain further insight into such aspects as the rice 

cultivation system and the possible impact of actual water use 

for paddy growing, the balance of water use, surface and un- 

derground water quality, and the relationships of various factors 

in paddy growing. Results at the farm level will be extrapolated 

to the subbasin (or catchments) and, ultimately, to 

the whole basin. The expected outputs will provide useful information 

for the optimum use of water to obtain higher paddy 

yields under appropriate soil, crop, and water management. 

Summary study of the DMPF 

The total timeframe of the DMPF is proposed for 5 years. In 

the first and second year, activities will concentrate on data 

collection in the subbasin and experimental plot. A demonstration 

model for the multifunctionality of paddy fields will 

be established in the third year and will be improved and applied 

to a wider area in the fourth year. Finally, analysis and 

demonstration will be done. 

Demonstration will be done through simulation of a logical 

model and a small experimental study. Through the study 

for the demonstration of paddy fields in northeastern Thailand, 

two types of models will be developed, at the field level 

and at the subbasin level. 

Field-level model 

The field-level model shows mainly the inflow/outflow of water 

in the field for agricultural land (paddy and upland). In addition, 

if possible, it is better to reflect some other aspects, such 

as field conditions (topography, soil, meteorology), farming 

practices, and infrastructure (irrigation/drainage, roads). 

The preliminary field-level model will be studied, based 

on the schematic model shown in Figure 1, and on other factors. 

To evaluate the multifunctionality of paddy fields at the 

sub-/entire basin levels, it is necessary to consider the impact 

of upland crops. The model can be applied for both paddy and 

upland by changing the coefficients of several factors. 

Subbasin level 

In this study, the subbasin is defined as one of the areas assuming 

a unit block from the viewpoints of conditions of hydrology, 

topography, land use, water use, flooding, and environment. 

It may include one or more rivers/tributaries of the 

Mekong. The model at the subbasin level (Fig. 2) will be in 

conflict because of factors such as other water use (urban, 


Evaporation 

Methane emissions 

Rainfall 

Landscape/amenity 

Bund 

Flood 

mitigation 

Soil erosion 

control 

Irrigation 

water 

Biodiversity 

Ponded wetland ecology 

Return flow to the river 

Nitrate 

Groundwater recharge 

Groundwater 

Fig. 1. Basic idea of preliminary model for field level. 

Measuring station 

Headwork 

Paddy field 

Water intake 

Return flow 

Paddy field 

Paddy field 

Other water use 

Fig. 2. Basic idea of preliminary model for the subbasin level. 

domestic, industrial), environmental impact on the area beyond 

the field, flood damage, etc. In the case of including reservoirs 

in a subbasin, it is necessary to consider the operation 

of the reservoirs and conditions for hydropower generation. 

To accommodate the model subbasin-wide, three options 

are proposed for data collection. The subbasin consisting of 

rice ecosystems as designed will be selected. These ecosystems 

are rainfed and the irrigated wet and dry seasons. Within 

the selected subbasin, a farmer-field experiment will be conducted 

to collect data as needed. 

Proposed data to be collected 

Subbasin level 

Subbasin conditions. The data on subbasin boundary, administrative 

conditions, infrastructure, irrigation, and land use are 

selected for basic items on the respective subbasin conditions. 

These data are available for designing the basic framework of 

the subbasin model and for clarifying the present conditions 

of the subbasin. The entire LMB levels will be clarified by 

compiling the information on the subbasin. 

Natural conditions. The data on climate, topography, 

geology, hydrogeology, soil, hydrology, drainage, and flood 

are selected for the natural conditions in the subbasin. These 

data are basic calculation data of the entire LMB and the 

subbasin model and actual observation data will be used for 

calibration and improvement of both models. 

Rice ecosystem. The data on rice ecosystem, paddy area, 

soil moisture, rice crop, and agro-economy are selected for 

the rice ecosystem of the subbasin. These data will be used 

mainly for analysis of direct impacts of the multifunctionality 

of paddy fields. 

Other conditions. The data related to socioeconomics, 

water use, community, and environment are selected for the 


other conditions of the respective subbasin. Indirect impacts 

will be analyzed mainly based on these data. 

On-farm level 

General conditions. The data on irrigation, rice area, landholdings, 

infrastructure, nearby drainage, topography, soil, land 

use, and climate are selected for the general conditions for the 

on-farm level. These data are required for designing the basic 

framework of the on-farm level model and for clarifying the 

present conditions at the field level. These data are also basic 

calculation data of the field-level model and actual observation 

data will be used for calibration and improvement of the 

model. 

Rice ecosystem. The data on the rice ecosystem, rice crop, 

and farming practices are selected for the rice ecosystem at 

the on-farm level. These data will be used mainly for analysis 

of direct impacts of the multifunctionality of paddy fields. 

Water and other conditions. The data on water inflow, 

water outflow, water table, underground water, and environment 

are selected for the water and other conditions of the onfarm 

level. These data will be used for analyses of water inflow 

and outflow and the main multifunctionality. 

Progress at the subbasin level 

The data on irrigation projects and reservoirs, soil and land 

use, and rice crops to be collected are basically in the form of 

an attribute table that will be linked subsequently with existing 

point data in a GIS file. 

The following results have been obtained. 

Irrigation data There are 8,764 irrigation projects, 

as follows: 5,125 reservoir projects, 1,347 weir 

projects, 1,192 pumping projects, and 1,100 other 

headwork projects (those are 17 large-, 1,335 medium-, 

and 7,412 small-scale projects). 

Land-use data. The total area of the Mae Kong 

subbasin is 186,221 km 2 . The major land uses are 

paddy fields (45%), forest (18%), and other agricultural 

land (20%). 

Rice crop data. Rainfed lowland rice on bunded fields 

occupies a large portion of the rice area. In northeastern 

Thailand, it is grown on rainfed area of about 3.3 

million ha and on irrigated area of about 0.3 million 

ha. 

Progress of the on-farm experiments 

The on-farm experiment in the dry season was selected at two 

different locations, as site I in Udon Thani and site II in Roi 

Et. Site I covered 2.5 ha, elevation 395 m, located at the lateral 

canal of the Huai Luang Irrigation Project. Site II covered 

4 ha, using water from the Seoi Yai Irrigation Project. 

Both on-farm experiments are preferably nearby the 

meteorological station, which is the site of the main source of 

data for the site: rainfall, temperature, and other climatic data 

needed to be collected. For the wet season, two more meteorological 

stations for on-farm experiments in the rainfed area 

will be established and an automatic rain gauge is already installed 

at the sites. 

The land use in the on-farm and surrounding sites was 

evaluated and soil samples were collected. The data on water 

inflow, water outflow, water depth, and water table are recorded 

daily by the staff of the Royal Irrigation Department (RID). 

Actually, the daily data record began in mid-March 2004. 

Some results 

A hook gauge is used for measuring the water level from an 

installed apparatus such as an N-type and water-requirement 

measuring apparatus. Daily data recorded by the N-type apparatus 

indicated a decrease in value because of the rate of water 

consumption by both percolation and evapotranspiration. The 

rate of percolation in the experimental area at the Huai Luang 

site is about 3–5 mm d –1 (average 3.57 mm d –1 ), evapotranspiration 

is 7–15 mm d –1 (average 11.13 mm d –1 ), and evaporation 

is 5–7 mm d –1 (average 5.77 mm d –1 ). The water-depth 

fluctuation in the paddy field was about 110–160 mm in the 

on-farm experiment. 

The on-farm experiment for the wet season started in 

August. There are two sites in irrigation projects and two sites 

in the rainfed area. 

Bibliography 

Mekong River Commission. 2003. Inception report. Program to 

Demonstrate Multifunctionality of Paddy Fields. 

Thai National Mekong Committee. 2004. Interim report. DMPF. 

Notes 

Author’s address: Royal Irrigation Department, Bangkok, Thailand, 

e-mail: ffpd@mail.rid.go.th. 


Remediation effect of rice terraces for strong acidic 

and nitrate-rich effluent water from tea bush areas 

in the Makinohara plateau of Shizuoka, Japan 

Kiyoshi Matsuo, Yuhei Hirono, and Kunihiko Nonaka 

To meet the demand for high-quality green tea products, Japanese 

tea farmers have applied excess nitrogen fertilizer (over 

1 t N ha –1 y –1 ) for a long time. As a result, heavy groundwater 

pollution by nitrate occurred in tea areas. Nitrate pollution is 

associated with strong acidity because of severe soil acidification 

in tea fields. Therefore, we evaluated the bioremediation 

effect of traditional small-scale rice terraces on effluent water 

from tea fields. Many small rice terraces have been used by 

farmers for a long time to cultivate rice for their home consumption 

near tea plantation areas in the Makinohara plateau 

of Shizuoka, Japan. The water remediation effect of those rice 

terraces on strong acidic and nitrate-rich effluent water from 

tea areas was investigated by a field survey and a field experiment. 

For the survey in two rice terrace areas in a traditional 

land-use region in Kurasawa, water quality (water temperature, 

pH, nitrate concentration, EC) was investigated from 

upstream to downstream. Farmers used those rice terraces for 

rice cultivation. Results are summarized in Table 1. 

A fallow rice terrace at Tann-no-pond, Ogasa-cho, 

Shizuoka, was used for an experiment without rice plants. The 

effect of paddy on water quality was investigated from June to 

October 2001. The experimental area was 180 m 2 and vegetation 

was paddy weed = no rice plant. Water management was 

submergence and total irrigated water to the plot was 1,946 t 

(surface effluent water was only 56 t). The effluent water from 

tea bushes was poured into the experimental paddy plot. Water 

quality (water temperature, pH, nitrate concentration, EC) 

was monitored once a week and some mineral elements ware 

measured by ICP or AAS. Table 2 summarizes the results. 

Original effluent water from tea bush areas is strongly 

acidic at pH 4 to 5 and has high nitrate content, 20 ppm or 

more as nitrogen. Occasionally, aluminum was as high as 7 

ppm. Rice terraces had a marked effect on remediating water 

quality. After passing the paddy soil, water pH was neutralized 

and nitrate was reduced to half. Aluminum was reduced 

below 0.5 ppm. Some field surveys were conducted at the same 

Table 1. Water quality in rice terraces farmed at Kurasawa, 

Shizuoka. 

Sampling point pH Nitrate N a Al Mg 

Rice terraces A (in) 4.44 33.5 2.2 17.5 

Rice terraces A (out) 6.09 23.1 0.4 15.8 

Rice terraces B (in) 5.44 29.3 0.5 16.9 

Rice terraces B (out) 6.85 13.8 0.5 14.8 

Original groundwater 4.28 27.2 5.7 17.4 

a Nitrate, aluminum, and magnesium concentrations are in ppm. Data were 

gathered in mid-summer 2000. Rice terrace A is about 2,500 m 2 with 7 

terraces, including slopes, and rice terrace B is about 5,000 m 2 , with 13 

terraces. The rice terraces were continuously irrigated. The original groundwater 

was measured at a water spring on the upper slope of Makinohara 

plateau. 

site in Makinohara. We confirmed the remediation effect of 

rice terraces for improving local water quality in heavily fertilized 

tea plantation areas in Shizuoka, Japan. 

Notes 

Table 2. Water remediation effect of rice terrace 

fields (experimental data in a fallow field). 

Sampling point pH Nitrate a Aluminum 

Influent water 4.70 24.6 2.74 

Paddy water 5.77 17.4 0.99 

Groundwater 6.60 9.9 0.31 

a Nitrate and aluminum concentrations are in ppm. Aluminum 

was measured on 6 different dates, pH and nitrate were measured 

once a week from 27 June to 27 October in 2001. 

Authors’ address: National Institute of Vegetables and Tea Science, 

e-mail: matuok@affrc.go.jp. 


Where are the hot spots of RDB plants in rice paddy fields 

A GIS-based analysis using information accumulated 

at a natural history museum 

Takuya Mineta, Kenji Ishida, and Takashi Iijima 

Rice paddy fields that occupy artificial wetland maintained by 

agricultural practices continue to occupy about 7% of the land 

area in Japan. In recent years, rice paddy fields were expected 

to be an alternative natural wetland for rapidly decreasing marsh 

and moorland caused by development. Numerous aquatic organisms, 

which depend on wetland habitat, are threatened with 

extinction. There are 191 species of 43 families, including 

natural wetland plants with origin in paddy fields, in the entire 

region (Kasahara 1954). Now, 49 species of 25 families in 

paddy fields are registered with the Red Data Book (RDB) in 

the Kanto region as endangered aquatic plants. These plants 

are considered as species dependent on the rice culture environment 

(= agro-dependent species), although information is 

lacking about characteristics of the habitat and distribution. 

We should be enhancing multifunctionality such as the 

preservation of both biodiversity and productivity in paddy 

fields to ensure a sustainable rice-cropping system in the world. 

On the other hand, Myers et al (2000) suggested “biodiversity 

hot spots,” where exceptional concentrations of endemic species 

are undergoing an exceptional loss of habitat, to support 

the most species. To enable us to pinpoint conservation priorities, 

we need knowledge of hot spots where endangered and 

agro-dependent species are concentrated in rural areas. We 

attempt to find a hot spot of endangered aquatic plants using 

information accumulated at a natural history museum on geographic 

information systems (GIS). 

This paper attempts to provide useful information on sustainable 

agriculture for the harmonious co-existence of 

biodiversity. 

Extraction of a hot spot of 

RDB plants 

GSIS process 

Analysis of characteristics 

of the habitat of each species 

Selection of endangered plants 

Specimen data input into database 

Creation of positional information 

(Low accuracy) 

Meshing 

Mapping 

Extraction of hot spots 

(High accuracy) 

Pointing 

Habitat information on specimens 

(Digital National Land Information) 

Land-use chart 

Extraction of 

paddy use 

Developing a database 

Habitat classification of plants 

Soil classification 

Extraction of 

ill-drained paddy 

How can we find a hot spot 

Figure 1 describes the process to find a hot spot of aquatic 

endangered plants and to understand the characteristics of their 

habitat. A detailed procedure follows. 

A survey was conducted in Tochigi Prefecture, which is 

located in the northern Kanto region (within 100 km of Tokyo) 

and occupies approximately 640,000 ha. Tochigi Prefecture 

has been accumulating geographic information on wildlife 

in standard grid squares (approximately 1 × 1 km). First, 

31 species of endangered hygrophilous plants were targeted in 

this study: they are selected based on records on paddy environment 

and natural wetland although their population has been 

decreasing rapidly. Second, we obtained a database with information 

such as collection location, date, and habitat information 

from botanical specimens stored at the Tochigi Prefectural 

Museum (at Ustunomiya City, Tochigi Prefecture). Then, 

Grasp of distribution pattern 

Fig. 1. Flow chart to find hot spots of endangered plants, and to 

analyze distribution pattern based on GIS, using the habitat data 

of botanical specimens. 

the database of a collection location of select RDB plants was 

arranged by standard grid squares to be able to analyze the 

distribution in detail. Furthermore, if we know a site where 

RDB plants were collected with higher precision, we could 

obtain point data and analyze habitat with more information 

about endangered species. 

From information on RDB plants in Tochigi Prefecture, 

we were able to make a map of 197 grids, which observed the 

RDB plants on GIS. Multiple endangered species were observed 

on 75 grids, particularly at the Watarase Retarding Ba- 


Paddy field 

dominance type 

River 


Hill-bottom 


Endangered species 0 20 40 60 80 100 (%) 

Rotala leptopetala 

Blyxa echinosperma 

Azolla japonica 

Eusteralis stellata 

Utricularia multispinosa 

Ceratopteris thalictroides 

Rorippa cantoniensis 

Eriocaulon cinereum 

Microcarpaea minima 

Salvinia natans 

Tillaea aquatica 

Gratiola japonica 

Monochoria korsakowii 

Inula britannica 

Rumex nipponicus 

Salvia plebeia 

Penthorum chinense 

Deinostema adenocaulum 

Najas foveolata 

Blyxa japonica 

Sagittaria aginashi 

Najas japonica 

Eriocaulon hondoense 

Ottelia japonica 

Isoetes japonica 

Utricularia australis 

Polygonum taquetii 

Ixeris makinoana 

Rotala pusilla 

Ludwigia greatrexii 

Eusteralis yatabeana 

Paddy fields Upland fields Rivers or lakes 

Forests Residence Others 

Fig. 2. Land use of areas around which RDB plants appeared. *Among plants in each type, 

significant differences were found as to their habitat (χ 2 test, χ 2 = 161.74, P

1. Paddy field dominance type (the majority of land is 

used for paddy fields). 

2. River or lake dominance type (consists mainly of riverside 

land and riverbanks). 

3. Hill-bottom dominance type (consists mainly of forests 

and small paddy fields). 

These data suggest that each group of RDB plants has a 

different habitat. The majority of hot spots were observed in 

the hill-bottom-type area, where forests accounted for more 

than 50% of the land. We call such small rice fields yachida or 

yatsuda spreading out over the hill-bottom. 

Discussion 

From the overlying data on land use, plants collected, and performed 

agricultural development projects, we understood that 

RDB plants were preserved under stable land use. In addition, 

we were able to confirm a few RDB plants such as Isoetes 

japonica, Eusteralis stellata, and Microcarpaea minima in the 

area where agricultural development took place. For harmony 

with an improvement in infrastructure and conservation of 

biodiversity in the rural environment, we could obtain the necessary 

information by detailed analysis of a habitat in the future. 

However, most current hot spots are places where agricultural 

development projects have not yet been carried out. 

We can use information about the hot spots of threatened 

aquatic plants that have different habitats depending on 

land use as indices for agricultural development projects that 

consider the environment. 

When using specimen information, data on botanical 

specimens should be handled carefully because information 

about collected areas and years may be biased, or there might 

be a lack or nonuniformity of information about the collection 

location and the habitats of plants. The following are issues. 

To use specimen information, it is important to develop a database 

with enormous amounts of information on specimens 

stored in a museum of the whole country. In addition, information 

on latitude, longitude, and altitude has to be unified. Detailed 

habitat descriptions are also important for site information, 

since analysis on habitat and distribution needs various 

sources of information. 

We describe how to find a hot spot of endangered plants 

by using the habitat data of botanical specimens in combination 

with GIS. Then, by overlapping a land-use chart with the 

GIS, we can understand the characteristics of their habitat. 

References 

Kasahara Y. 1954. Studies on the weeds of arable land in Japan, with 

special reference to kinds of harmful weeds, their geographic 

distribution, abundance, life-length, origin and history. Ber. 

Ohara Inst. Landw. Forsch. 10:72-116. 

Myers N, Mittermeier RA, Mittermeier CG, da Fonseca GAB, Kent 

J. 2000. Biodiversity hotspots for conservation priorities. 

Nature 403:853-858. 

Notes 

Authors’ address: National Institute for Rural Engineering, Japan, 

e-mail: minetaku@nkk.affrc.go.jp; ishida@nkk.affrc.go.jp; 

tiijima@affrc.go.jp. 


Session 11 featured discussion on the various values and roles 

in our lives as a whole fulfilled by rice cultivation with paddy fields 

and irrigation under the designated theme. A welcome was given 

by H. Satoh, president of the National Institute for Rural Engineering 

(NIRE), K. Miyamaoto, executive director of NIRE, and Y. 

Tsutsui, convener. The session, with about 200 participants from 

various countries, discussed the subject of inquiry raised by Professor 

R. Nakamura in his earlier keynote speech. K. Yamaoka, 

co-convener, proposed recalling the following arguments summed 

up by Professor Nakamura. 

1. Water is indispensable to rice cultivation. Sustainable 

irrigated rice paddy agriculture is crucial for developing humans’ 

future. 

2. Increasing water-use efficiency, namely, water productivity, 

is in fashion and seems to be our universal mission. 

3. However, we have to note that in humid regions it is 

rational to use ample water to substitute for farmers’ labor, 

whereas in arid regions constantly valuable water cannot be replaced 

by other resources. (In humid regions, developing appropriate 

water resource allocation systems for occasional dry spells 

is most important to increase real water-use efficiency.) 

4. Ample agricultural water in humid regions generates vast 

externalities that cannot be priced through market mechanisms. 

5. Significance of the externalities has been maintained 

but has shifted its weight from benefits to farmers to those for 

the public (town dwellers) according to the socioeconomic development 

of the country. 

6. Policy implications for publicly facilitating sustainable 

irrigated rice paddy agriculture in restoring a natural balance should 

be further promoted in various development stages of countries 

in the Asian monsoon region. 

K. Yamaoka also presented the background and purpose 

of the session. It was expected to feature discussions to endorse 

the arguments and to sum up the main points of the arguments 

to be contributed to discussions at the OECD, the Fourth World 

Water Forum, and relevant international conferences. Eight contributors 

made presentations from various interdisciplinary aspects. 

The first contributor, T. Masumoto from Japan, stated that 

the geographical, climatological, and hydrological characteristics 


of the region have led to paddy rice farming having an advantage 

in environmentally friendly characters as well as crop productivity 

over other types of farming anywhere in the Asian monsoon region. 

He contrasted paddy rice farming with farming in upland 

fields in an arid climate and reviewed the major methodologies 

and results for assessing various socioeconomic externalities 

generated by paddy rice-farming activities, including quantitative 

and monetary assessment. 

D. Groenfeldt from the United States stressed the importance 

of intangible benefits to social, cultural, spiritual, and aesthetic 

life contributed by paddy cultivation to rural livelihoods in 

the broadest sense. He criticized the negative effects of the Green 

Revolution on cultural ways of life in rice paddy cultivation introduced 

by modern materialism and professional aspirations and 

identified the roles of paddy rice cultivation in rural quality of life 

as social structure, cultural identity, religion and spirituality, and 

aesthetic beauty. He concluded that a participatory methodology 

for studying livelihood happiness should be developed for describing 

people’s vision of their life, taking into consideration four 

pillars of happiness: economic development, cultural heritage, 

environment, and governance. 

N. Sutawan from Indonesia emphasized the need to preserve 

irrigated rice culture because of its multifunctional roles 

with invaluable positive externalities. He pointed out the rapid 

conversion of paddy fields to other nonfarming usage as one of 

the greatest threats to the sustainability of irrigated rice culture. 

He concluded that sustainability depends on sustainable water 

circulation at a basin scale, including several irrigation systems 

as well as sustainability of irrigation institutions and networks, 

rice production, ecosystems, and socio-cultural values featured 

by each independent irrigation system. 

L. Yulong from China mentioned that current Chinese food 

production, including rice, has declined by 15% and he pointed 

out the difficulty in expanding farm land because of the limitation 

of land and water resources. He illustrated the multifunctional 

roles and technologies of the Dujiangyan Irrigation Area, China’s 

oldest and biggest irrigation area, including a case study for aquaculture 

in and beside paddy fields, contributing to poverty alleviation, 

the regional economy, indigenous culture, biodiversity, and 

system sustainability. 

K. Athukorala from Sri Lanka described the terrible situation 

brought about by the decommissioning of rainfed paddy fields 

in southern Sri Lanka with abundant precipitation. She mentioned 

that the soil of the decommissioned paddies acted like a sponge 

in absorbing moisture. She concluded that the situation causes 

groundwater aquifers to deteriorate for water of city dwellers and 

increases floods downstream even around the parliament building 

in the capital city as well as threatening food security of people 

in Sri Lanka. 

K. Hidaka from Japan proposed a definition of true 

biodiversity depending upon rice ecosystems as different from 

that of conventional apparent biodiversity. He also proposed integrated 

research for organizing the three methods of collecting 

specimens, reviewing records in the literature, and ecological fieldwork. 

He emphasized a need for integration between the conservation 

of biodiversity and agricultural production, with a new concept 

of integrated biodiversity management for managing agricultural 

activity in hotspot villages for species such as the giant 

water bug. 

Y. Yuyama from Japan introduced floating-rice cultivation 

in regularly inundated areas in Thailand. He assessed its potential 

function for mitigating flood and providing water resources in 

Bangkok, located downstream, taking into consideration the development 

of facilities such as diversion dams and drainage regulators 

controlling water level in the area. He concluded that the 

monitoring of water flow, water operations, and cultivation conditions 

is essential and the decision support system for water operations 

is a key technology. 

Finally, C. Boonlue from Thailand introduced a new challenging 

program used to measure and improve the 

multifunctionality of paddy fields in the northeastern region of 

Thailand. He mentioned the intensive experiments to collect data 

in paddy fields on evaporation, transpiration, percolation, and 

infiltration as well as data on runoff at the subbasin and wholebasin 

level. He anticipated the development of a model to describe 

water use of paddy cultivation and its effects on surface 

and underground water and return flow in the basin as well as on 

various rice ecosystems. 

Many valuable comments were made after each presentation. 

The discussions are summarized as follows. 

The significance of rice systems in the Asian monsoon region 

was attributed to centuries of sustainability of paddy farming, 

which has produced different socioeconomic externalities as 

well as food for people. The excellent power of rice farming to 

feed people and to form sustainable and harmonious societies is 

evidenced by the fact that, in this region, 54% of the world’s 

population survives on only 14% of the world’s land area. People’s 

ceaselessly striving to produce rice under extreme water conditions 

such as frequent flood and occasional water shortage has 

molded their stable and mutually dependent communities as well 

as their basic lifestyle and culture closely connected with water. 

Paddy rice farming in the Asian monsoon region now accounts 

for about 90% of the world’s rice production and about half of 

the world’s freshwater use. Centuries of sustainability of paddy 

farming with broad socioeconomic externalities have gone by with 

massive water use. 

However, the broad positive externalities of rice paddy farming 

and irrigation for such ecosystems, sound water circulation, 

and socioeconomic stability have deteriorated with the development 

of the market economy because nobody pays compensation 

for the externalities through market mechanisms. Though it 

is important to improve water-use efficiency, namely, economic 

water productivity in farming in water-deficient regions, we should 

also promote enhancing the multifunctionality of rice systems, 

not only for sustainable rice paddy farming in the Asian monsoon 

region but also for sustaining and making optimal use of global 

water resources. 

We were satisfied with the interdisciplinary discussions with 

the selected speakers from various fields in international research. 

We are strongly convinced that we need to promote further research 

activities in this area. 


SESSION 12 

Conservation of soil, water, 

and environment in rice cultures 

CONVENER: M. Saito (NARO) 

CO-CONVENERS: S. Itoh (NIAES) and T. Nozoe (IRRI)

Paddy soils around the world 

K. Kyuma 

The world’s annual rice production reached 589 million tons 

in 2003 (FAO 2004). It exceeded that of wheat and amounted 

to 28.4% of the world’s total cereal output. Maize production 

has recently surpassed that of rice, but its increase is mainly to 

meet the demand for animal feed and not for immediate human 

consumption. Thus, rice as the staple for the human diet 

is becoming increasingly important in Asia and in some parts 

of Africa and the Americas, feeding nearly half of the world’s 

population today. 

According to IRRI (2002), some 11% of the total area 

cultivated for rice worldwide is in uplands, the rest being in 

wetlands, either as rainfed, deepwater, or irrigated paddies. 

As upland rice is generally much less productive, rice production 

in paddy soils should definitely exceed 90% or even 95% 

of the total global output. This justifies my confining this discussion 

to only paddy soils. 

Natural settings of paddy lands 

There are some 135 million ha of paddy lands in the world, of 

which 126 million ha, or 93%, are in monsoon Asia, or in humid 

East, Southeast, and South Asia (IRRI 2002). Climate and 

landforms are the two determinants for this strongly biased 

distribution of paddy lands (Kyuma 2004). Monsoon climate 

is characterized by a yearly inversion of wind direction and 

this accompanies alternating rainy and dry seasons. Therefore, 

rainfall is concentrated during the rainy season, often bringing 

more than 1,000 mm of rain in less than half a year, thus enabling 

cultivation of a rice crop. In terms of landforms, monsoon 

Asia features an exceptional abundance of lowlands (Table 

1). Being in the region of active orogeny and volcanism, in 

combination with high rainfall, monsoon Asia undergoes severe 

erosion in its high mountainous lands and deposition in 

its riparian and coastal lowlands, producing extensive floodplains 

and deltas in the middle and lower reaches of gigantic 

rivers such as the Yangtze, Mekong, Brahmaputra, and Ganges. 

These lowlands are naturally inundated with monsoon rains 

during the rainy season. 

As depicted above, monsoon Asia, with its high seasonal 

concentration of rainfall and its exceptionally extensive area 

of lowlands, provides Oryza sativa, a plant species native to 

the region, with the most adapted natural habitat. Thus, rice 

culture originated as an adaptation to the given natural settings 

of monsoon Asia. Later expansion of paddy lands has 

been made possible mainly by the provision of irrigation. 

A small share of indigenous rice culture with O. 

glaberrima is known to exist in West Africa, but it has been 

narrowly confined to its place of origin throughout history, for 

the climate and landforms of the region are not favorable for 

its propagation. 

Table 1. Importance of alluvial land (in 10 6 ha). 

Region 

Land area 

Characteristics of paddy soils 

Alluvial soil area 

Total Potentially Total Potentially 

arable 

arable 

World 13,000 3,152 588 316 

Asia a 2,704 620 – 192 

Tropics 4,893 1,652 365 172 

Tropical Asia 987 344 168 114 

a 

Excluding the former USSR. 

Source: White House (1967). 

Any soil that is used for growing aquatic rice can be called a 

paddy soil. Paddy soil in this definition is related directly to 

land use, but not to any particular type of soil in a pedological 

sense. Paddy soils occur basically in lands with an aquic (note: 

this is a technical term used in U.S. Soil Taxonomy and widely 

recognized as a term to denote hydromorphic conditions) moisture 

regime, in floodplains, deltaic plains, fans, and terraces. 

Fluvisols, Gleysols, and Cambisols are the most common pedological 

members of paddy soils, but some other members, 

such as Acrisols and Luvisols, occur among paddy soils on 

older land surfaces. Eswaran et al (2001) also listed all the soil 

orders of soil taxonomy except Gelisols as soils used for aquatic 

rice cultivation, but they also indicated that Inceptisols and 

Ultisols are the most frequently used. 

Irrespective of their pedological nature, all paddy soils 

are submerged for at least a few months a year, either naturally 

or artificially. Managing paddy soils under water is entirely 

different from managing other soils used for upland crops, and 

this produces important differences, particularly in chemical 

and biochemical or microbiological processes. However, paddy 

soils are also kept drained for the rest of the year, again naturally 

or artificially. This cyclic change in micro-environmental 

conditions exhibits properties not encountered in other soils 

and differentiates paddy soils from most other soil systems. 

The most important change upon submerging a soil is a 

decrease in oxygen partial pressure, which accompanies a lowering 

of redox potential (Eh), reduction of chemical species 

[O 2 →H 2 O, NO 3 – →N 2 , Mn(IV)→Mn(II), Fe(III)→Fe(II), 

SO 4 2– →S 2– , CO 2 →CH 4 ], elevation of soil pH, and changes in 

biotic composition (aerobic to anaerobic) (Table 2). All these 

changes further induce enhanced nitrogen availability and its 

fixation, solubilization of soil phosphates, and so forth. 

Draining a paddy soil reverses most of the above processes: 

Eh rises, chemical species are oxidized, soil pH decreases, 

and aerobic soil biota recover. However, some of the 


Table 2. Reduction processes and microbial metabolism in submerged soils. 

Stage of Chemical Initial Eh 7 Expected pattern of Formation of Formation of Formation of Hypothetical pattern of Redox state 

reduction transformation in soil, V energy metabolism NH 4 -N CO 2 organic acids organic matter decomposition 

First Disappearance of 0.5 Aerobic respiration Rapid process Rapid process Usually not accumulated Aerobic and semianaerobic Oxidized 

molecular oxygen without fresh organic matter decomposition process 

Disappearance of nitrate 0.4 Nitrate reduction Weakly reduced 

Formation of Mn II 0.4 (Mn III, IV reduction) Moderately 

reduced 

Formation of Fe II 0.2 (Fe III reduction) 

Second Formation of S –II 0 Sulfate reduction Slow process Slow to stagnant Early stage: rapid accumulation Anaerobic decomposition Strongly reduced 

process process 

Formation of methane –0.2 Methane fermentation Advanced stage: rapid decrease 

Formation of hydrogen –0.2 Fermentation Formation and decomposition 

of formic acid 

Source: Takai (1978, partly modified). 

changes are irreversible, for example, “ferrolysis” proceeds in 

every cycle of submergence and drainage, impoverishing soil 

basic cations and leading to Al-interlayering and eventual destruction 

of smectitic clay lattices (Kyuma 2004). 

Intrinsic merits and demerits of the paddy soil/rice system 

In comparison with the soils used for growing upland crops, 

paddy soils cropped to aquatic rice have many intrinsic merits 

and some demerits. Among the merits, the following can be 

enumerated: (1) a higher natural supply of nitrogen, bases, and 

silica; (2) higher availability of soil phosphorus; (3) relative 

indifference to soil physical properties; (4) detoxification of 

excessive nutrients; (5) detoxification of many agrochemicals; 

(6) perfect resistance to soil erosion; (7) relative ease of weeding; 

(8) tolerance for monoculture; and (9) carbon sequestration 

(Kyuma 2004). 

Some of these merits contribute to higher land productivity 

of rice relative to upland cereals, and others contribute 

to higher stability of rice farming. It deserves a special mention 

that routine production techniques such as land leveling 

and bund construction for submerging the rice field simultaneously 

serve to protect the land from erosion and degradation. 

All in all, these intrinsic merits confer upon the paddy 

soil/rice system the highest sustainability among existing food 

production systems in the world. 

Some demerits, however, are also intrinsic to paddy rice 

cultivation, such as potential pollution hazards for water and 

air. In countries with a high fertilizer use, paddy fields are considered 

to be the largest potential nonpoint source of pollutants, 

particularly nitrogen and phosphorus, for water bodies. 

Globally, paddy soils are said to be responsible for 11% of 

total annual methane emissions (IPCC 1995). For researchers 

of the paddy soil/rice system, it is imperative to strive to lessen 

these local as well as global environmental hazards incurred 

by paddy rice cultivation. 

Conclusions 

The world population is still increasing and it is forecast to 

reach nine billion by 2050. If demand for rice increases proportionally, 

the total production will have to increase by as 

much as 50% by mid-century. Taking into consideration the 

requirements for climate and landforms in making paddy lands, 

prospects will not be great for further expansion of rice area. 

The only way to attain the target would be to develop infrastructure 

for irrigation and drainage. Even in many traditional 

rice-growing countries, irrigated paddies are often less than 

half of the total. Provision of better irrigation and drainage 

would both enhance the productivity of rice and help to negate 

the major demerit of methane emissions from paddy lands. 

Session 12: Conservation of soil, water, and environment in rice cultures 353

References 

Eswaran H, Moncharoen P, Reich P, Padmanaban E. 2001. Rice, 

land and people: the faltering nexus in Asia. Proceedings of 

the 5th Conference of the East and Southeast Asia Federation 

of Soil Science Societies, Krabi, Thailand. Bangkok (Thailand): 

Department of Agriculture. p 38-66. 

FAO (Food and Agriculture Organization). 2004. FAOSTAT. Rome 

(Italy): FAO. 

IPCC (Intergovernmental Panel on Climate Change). 1995. Climate 

change 1994, radiative forcing of climate change and an evaluation 

of the IPCC 1992 IS92 emission scenarios. New York, 

N.Y. (USA): Cambridge University Press. 

IRRI (International Rice Research Institute). 2002.World rice statistics, 

2002. Los Baños (Philippines): IRRI. 

Kyuma K. 2004. Paddy soil science. Kyoto (Japan): Kyoto University 

Press. 280 p. 

Takai Y. 1978. Redox processes in the soil under submergence. In: 

Kawaguchi K, editor. Suiden-dojyogaku (Paddy soil science). 

Tokyo (Japan): Kodan-sha. p 23-55. (In Japanese.) 

White House. 1967. World food problem: a report of the president’s 

Science Advisory Committee. Vol. II. Report of the Panel on 

the World Food Supply. Washington, D.C. (USA): White 

House. 

Notes 

Author’s address: Professor Emeritus, Kyoto University and the 

University of Shiga Prefecture, 1-79 Nagatanicho, Iwakura, 

Sakyoku, Kyoto 606-0026, Japan. 

Sustainability of paddy soil fertility in Vietnam 

Ngo Ngoc Hung, Nguyen Bao Ve, Roland J. Buresh, Mark Bayley, and Takeshi Watanabe 

Although the Lower Mekong Delta (LMD) covers only 12% 

of Vietnam’s total land area, it is the main source of agricultural 

products of the country, supplying more than 50% of the 

staple food. Among the diverse range of agricultural products 

in the Delta, rice is the main crop. It is planted in irrigated 

lowlands with three major cropping seasons: winter-spring 

(November to February), spring-summer (March to June), and 

summer-autumn (July to October), which produce about 56% 

of the nation’s total production. Some soils in the Mekong Delta 

receive high amounts of fertilizer, either because of the production 

of three crops per year or because soils such as acid 

sulfate soils require much fertilizer to support production. 

Yield of summer-autumn rice in the LMD is usually low 

(3–4 t ha –1 ). Therefore, the government and farmers are interested 

in diversifying continuous lowland rice culture to multiple 

cropping. Urea is the main N fertilizer applied to lowland 

rice in the Mekong Delta. Phosphorus fertilizers have been 

used on some soils in Mekong Delta agriculture for more than 

30 years. Especially, along the main rivers, soils have received 

P fertilizer for more than 20 years. 

The objectives of this research are to (1) determine the 

effects of the previous crop on rice yield and the economic 

merits of crop rotation, (2) determine phytoplankton activity 

in nitrogen fertilizer-use efficiency, and (3) assess the effects 

of long-term P fertilization in intensive rice production on cadmium 

contamination of soils. 

Effects of crop rotations on rice 

Field experiments to assess crop rotations with three consecutive 

crops per year were conducted on alluvial soils in 2002 

and 2003 at Can Tho. Five crops (rice, soybean, maize, sweet 

potato, and sesame) were grown in spring-summer and rice 

was then grown in winter-spring and summer-autumn seasons. 

The experiment was a randomized complete block design 

with four replications. All crops received 60 kg P 2 O 5 

ha –1 as single superphosphate. The application of N and K 2 O 

fertilizer for each crop was 100-30 for rice, 30-60 for soybean, 

100-60 for maize, 50-90 for sweet potato, and 60-30 for 

sesame. Nitrogen was applied as urea in three equal splits: 1/3 

at 10 days after planting (DAP), 1/3 at 20 DAP, and 1/3 at 45 

DAP. Potassium was applied as KCl. Rice variety IR66 was 

used. Soil properties in the top 25 cm were pH = 4.2, organic 

C = 2.8%, total N = 0.25%, CEC = 16 cmol c kg –1 , clay = 42%, 

and silt = 56%. 

A rotation of rice with other crops in nearly all cases 

significantly improved rice yield (Table 1). Yield and the apparent 

recovery of N fertilizer by summer-autumn rice were 

the greatest following sweet potato, and they were greater for 

rice following soybean and maize than following rice. Nitrogen 

fertilizer-use efficiency was higher following sweet potato 

than rice. The decomposition of sweet potato residues in 

raised beds after harvesting provided N and potentially improved 

soil physical, biological, and chemical processes. In 

addition, a symbiotic combination of N 2 -fixing microorganisms 

of Azospirillum brasilense with sweet potato roots was 

found (Hill et al 1983). 

The profit for each cropping system was determined from 

grain yield and market price. The profit (US$ ha –1 year –1 ) of 

the cropping systems was rice-soybean-rice = 644, rice-sweet 

potato-rice = 612, rice-sesame-rice = 586, rice-maize-rice = 

503, and rice-rice-rice = 421 (Ngo Ngoc Hung and Nguyen 

Bao Ve 2004). Additional research conducted at four locations 

during summer-autumn 2001 and 2002 showed that, in all the 

cases, rice yield was higher after a 3-month fallow (average: 

4.1 t ha -1 ) than with continuous rice cropping (average: 3.8 t 

ha –1 ) (Nguyen Bao Ve et al 2002). 


Table 1. Effect of crop rotation on rice grain yield and N efficiency for 100 kg N ha –1 

applied to rice at Can Tho, southern Vietnam. a 

Previous crops Summer-autumn Summer-autumn season (2003) 

(spring-summer season (2002) 

season) (rice grain yield, Rice grain AFRN b Ndff % c NUE d 

t ha –1 ) yield (t ha –1 ) 

Rice 3.4 a 4.4 a 23 15 19 

Soybean 4.2 bc 5.5 c 37 – – 

Maize 3.9 ab 4.9 b 33 – – 

Sweet potato 4.6 c 6.3 d 52 32 29 

Sesame 3.6 ab 4.7 ab 26 – – 

LSD 5% 0.67 0.56 

a 

Numbers followed by the same letter in a column are significant at the 5% level. b Apparent fertilizer N 

recovery. c Plant N derived from fertilizer. d Nitrogen fertilizer-use efficiency. 

Phytoplankton activity and nitrogen losses in rice fields 

The design was a randomized complete block with four replications. 

Floodwater pH, urea, ammonium, and biomass of phytoplankton 

were monitored after N fertilizer application. Fertilizer 

was applied to rice as urea or ammonium sulfate, superphosphate, 

and potassium chloride. 

The biomass of phytoplankton (dominated by Diatom, 

Chlorophyta, Cyanophyta, and Euglenophyta) reached a maximum 

within 2–4 days after urea application. During summerautumn, 

at 3 days after N fertilizer application, the biomass 

reached 8.0 g m –2 in urea plots versus 5.0 g m –2 in ammonium 

sulfate plots. Because of daytime consumption by algae of CO 2 

dissolved in the floodwater, the floodwater pH rose to > 8 on 

some days at the first and second N application. However, there 

was no significant difference in biomass of phytoplankton or 

pH of floodwater between these two treatments after the second 

fertilizer application (Watanabe et al 2003). This was conducive 

to a rapid conversion of NH 4 

+ 

to gaseous NH 3 . 

15 Nitrogen balance research during the dry season of 

1990 at Can Tho showed that the fraction of added urea-N not 

recovered in the plant and soil, and presumably lost, depended 

on the timing of urea application. The estimated N loss for N 

applied at 10 DAS was very high (49% of applied N). It decreased 

to 25% of the applied N at 20 DAS and 11% of the 

applied N at 44 DAS. The distribution of 15 N in the soil plus 

roots suggested that leaching was not an important loss mechanism, 

and therefore the N loss in this case was presumed to be 

due to NH 3 volatilization and denitrification (Ngo Ngoc Hung 

et al 1995). 

Cadmium content in flooded rice soils 

Soil samples from the top 20 cm, which covered three major 

soil groups (alluvial soils, acid sulfate soils, and saline soils) 

in the LMD, were collected. In addition, P fertilizers used in 

the LMD were collected. 

Cadmium content in P fertilizers ranged from 0.02 to 

2.76 mg kg –1 . The highest content was in imported NPK fertilizer 

(2.76 mg kg –1 ), whereas some domestic P fertilizers had a 

lower Cd content (1.73 mg kg –1 ). Cadmium content in LMD 

rice soils ranged from 0.01 to 0.56 mg kg –1 . Generally, the 

maximum value of Cd (0.56 mg kg –1 ) in rice soils was lower 

than the standard limit of Vietnam (2.0 mg kg –1 ), the EC Directive 

(1–3 mg kg –1 ), and Canada (1.4 mg kg –1 ) for Cd in 

agricultural soils. 

The correlation between duration of P fertilizer use and 

Cd content in soil is well recorded (Fig. 1). Cadmium content 

was highest in alluvial and acid sulfate soils and relatively lower 

in saline soils. This can be attributed to the more intensive 

crop cultivation on alluvial soils (3–4 crops year –1 ) than on 

saline soils (1–2 crops year –1 ). Although crop cultivation is 

not as great on acid sulfate soils (usually 2 crops year –1 ) as on 

alluvial soils, the rate of P fertilizer application is higher on 

acid sulfate soils (60–90 kg P 2 O 5 ha –1 ) than on alluvial soils 

(50–60 kg P 2 O 5 ha –1 ). 

In serious cases, increasing soil Cd content could decrease 

soil health, reduce crop yield (Steve 1994), and affect 

humans by Cd contamination in food. Recent research has 

shown that the Cd uptake in crops depends largely on soil and 

climatic factors, plant genotype, and agronomic management. 

While plant breeding and agronomic management can minimize 

soil-plant transfer of Cd, and maximize concentrations 

of antagonists to Cd assimilation in humans, implications of 

the studies are that inputs of this metal to soil can be minimized 

(McLaughlin et al 1999). 

Conclusions 

The rotation of rice with other crops in nearly all cases significantly 

improved rice yield. Rice yield was highest after sweet 

potato, and the N fertilizer-use efficiency of rice was higher 

following sweet potato (29%) than rice (19%). The rotation of 

rice with soybean attained the highest profit because soybean 

had a high market price. 

Floodwater pH is a major factor affecting NH 3 volatilization, 

and photosynthetic aquatic biomass has a key role in 

increasing floodwater pH. The biomass of phytoplankton, 

dominated by Diatom, Chlorophyta, Cyanophyta, and 

Euglenophyta, reached a maximum within 2–4 days after urea 


Soil Cd content (mg kg –1 ) 

0.6 

0.5 

y = 0.0066x + 0.0306, R 2 = 0.32* 

0.4 

0.3 

0.2 

0.1 

0.0 

0 10 20 30 40 

Duration of P application (y) 

Fig. 1. The correlation between years of P fertilizer application and soil Cd content. 

application. Management practices to reduce floodwater N 

concentration can help minimize N losses. 

Some soils in the Mekong Delta received high amounts 

of P fertilizer, which can contain varying degrees of contaminants, 

including cadmium. The concentrations were lower than 

critical values considered by Vietnamese and some international 

levels, but the results indicate an anthropogenic impact 

on soil cadmium levels. Concerns about food-chain contamination 

should be evaluated. 

References 

Hill WA, Bacon-Hill P, Crosman SM, Steven C. 1983. Characterization 

of N 2 -fixing bacteria associated with sweet potato roots. 

Can. J. Microbiol. 29:860-862. 

McLaughlin MJ, Parker DR, Clarke JM. 1999. Metals and micronutrients: 

food safety issues. Field Crops Res. 60:143-163. 

Ngo Ngoc Hung, Singh U, Vo Tong Xuan, Buresh RJ, Padilla JL, 

Tran Thanh Lap, Truong Thi Nga. 1995. Improving nitrogenuse 

efficiency of direct-seeded rice on alluvial soils of the 

Mekong River Delta. In: Denning GL, Vo Tong Xuan, editors. 

Vietnam and IRRI: a partnership in rice research. Los 

Baños (Philippines): International Rice Research Institute. 

p 137-149. 

Ngo Ngoc Hung, Nguyen Bao Ve. 2004. Effect of crop rotation on 

nitrogen use efficiency and yield of summer-autumn rice. Vietnam 

Sci. Technol. J. Agric. Rural Dev. 5(2004):634-636. 

Nguyen Bao Ve, Ngo Ngoc Hung, Nguyen Thanh Hoi, Pham Duc 

Tri, Nguyen Van Nhieu Em. 2002. Effect of soil fertility and 

farming practices on rice growth and yield of summer-autumn 

rice in the Mekong Delta. Vietnam Soil Sci. J. 16(2002):76- 

83. 

Steve PM. 1994. Case study 6: effects of heavy metals from sewage 

sludge on soil microbes in agricultural ecosystems. In: Ross 

SM, editor. Toxic metals in soil-plant systems. John Wiley 

and Sons. 247 p. 

Watanabe T, Ngo Ngoc Hung, Nguyen Do Chau Giang, Vo Thi Yen 

Phi, Tran Chan Bac. 2003. Changes of nitrogen in floodwater 

after fertilizer application. In: Development of new technologies 

and their practice for sustainable farming systems in the 

Mekong Delta. Proceedings of the Final Workshop of JIRCAS 

Mekong Delta Project, 25-26 November 2003. Can Tho (Vietnam): 

College of Agriculture, Can Tho University. p 373-382. 

Notes 

Authors’ addresses: Ngo Ngoc Hung and Nguyen Bao Ve, Can Tho 

University, Vietnam; Roland J. Buresh, International Rice 

Research Institute (IRRI), Philippines; Mark Bayley, University 

of Aarhus, Denmark; Takeshi Watanabe, Japan International 

Research Center for Agricultural Sciences (JIRCAS), 

Japan, e-mail: ngochung@ctu.edu.vn. 

Acknowledgments: This research was supported in part by grants 

from the projects of CAULES, funded by DANIDA (Danish 

International Development Agency), and JIRCAS (Japan International 

Research Center for Agricultural Sciences). 


Managing soil fertility for sustainable rice production 

in northeast Thailand 

Kunnika Naklang 

Thailand is divided geographically into four main regions: 

northern, northeast (NE), Central Plain, and southern. Northern 

Thailand, with about 22% of the total rice area, contributes 

about 26% of the rice production. The NE has 57% of the 

total area and 46% of the total production. In the Central Plain, 

the rice-growing area is 17% and rice production is 24%. In 

southern Thailand, the corresponding percentages are only 4% 

and 4%. 

Rainfed rice is grown on approximately 4.66 million 

hectares or about 92% of the total rice-growing area in northeast 

Thailand. The average paddy yield in the 2003 crop season 

was 1.81 t ha –1 , versus 3.21 t ha –1 for irrigated lowland 

rice in the Central Plain. Among the four regions, rice yield in 

the northeast is the lowest because of at least two major constraints, 

poor soil and the unpredictable amount and distribution 

of rainfall. Low soil fertility recurs every year in the rainfed 

lowlands in this region, but the problem of drought does not 

occur in the same manner every year (Fukai et al 1995, Ragland 

1997). 

The quantity of organic matter in the soil is a major indicator 

of its quality. To achieve higher grain yield, we should 

apply chemical and organic fertilizer. For sustainability, it is 

important to incorporate organic matter into the soil. Koopmans 

and Goldstein (2001) stated that to improve soil quality we 

should treat our organic matter like a bank account. A bank 

account lets us deposit, save, and withdraw something we value. 

For sustainability, it is important to deposit in the account of 

active organic matter in the soil on a regular basis, thereby 

building cultural fertility. Poulton (1995) summarized that longterm 

experiments are essential in determining the factors affecting 

soil fertility and sustainable production. 

Experiments 

Effects of chemical fertilizer and farmyard manure 

on grain yield of rice in NE Thailand 

Twenty-six experiments were conducted at eight sites in NE 

Thailand from 1995 to 1997 to determine the effects of farmyard 

manure (FYM) and chemical fertilizers on rice yield in 

rainfed areas. Out of eight sites, the experiments were separated 

into two experimental plots: one was rainfed and the other 

under irrigated conditions. The soil texture was loamy sand at 

three sites, sandy loam at three sites, and clayey loam and clay 

at one site each. The soil’s pH varied from 4.0 to 5.7. Organic 

matter varied from 0.39% to 1.57% and Bray-II P ranged from 

7 to 38 mg kg –1 . Nonglutinous, photosensitive, and tall cultivars 

RD 15, harvested in mid-November, and KDML 105, 

harvested in late November, were used in these experiments. 

The fertilizer treatments were nil, FYM (cattle dung at 6.25 t 

ha –1 ), PK, NPK, CR-NPK (controlled-release N), all (NPK + 

micronutrients), N, and FYM + NPK. The rates of N, P 2 O 5 , 

and K 2 O were all 50 kg ha –1 . Rice seedlings were transplanted 

at a spacing of 25 × 25 cm at almost all sites except in Tung 

Kula Ronghai 1996, where dry seed was broadcast at 62.5 kg 

ha –1 . 

Soil properties indicate that soil productivity was controlled 

by many factors, including soil fertility, environment, 

and cultural management. The decline in grain yield under 

rainfed loamy sand soil at Ubon was lower than under irrigated 

conditions. Late dry-seeded sowing and late transplanting 

reduced the growth and yield of rice (Naklang et al 1998). 

The results indicated that the most effective fertilizer was N, 

which made the average grain yield increase over all sites by 

39%. This result reflected that N was the most limiting nutrient 

in NE Thailand. The application of PK was effective only 

when it was applied with N. The application of NPK gave a 

higher grain yield than N applied alone: but only 8%. In loamy 

sand soil, the addition of some micronutrients together with 

NPK (all) increased grain yield from 2% to 21% compared 

with applied NPK. The application of FYM increased grain 

yield in many experiments. Average grain yield of FYM was 

only 7.5% less than the yield of NPK (mean yield of NPK was 

2,425 kg ha –1 ). The additional role of FYM in contributing 

NPK was found in some sandy soils. The grain yield of rice 

was higher in clay soil than in loamy sand soil when chemical 

fertilizer was applied (Table 1). 

Long-term effects of rice straw compost 

and chemical fertilizer on rice yield 

and their residual effects 

An experiment has been conducted at the Surin Rice Research 

Center since 1976. Rice straw compost (RSC) at 0, 3.125, 6.25, 

9.375, and 12.5 t ha –1 with and without chemical fertilizer (CF) 

at the rate of 50-50-50 kg of N-P 2 O 5 -K 2 O ha –1 was applied for 

25 consecutive years (1976-2000). During the first 25 years, 

the rice variety used was RD 7, short, nonglutinous, and 

nonphotoperiod-sensitive. From 2001 to 2003, residual effects 

of 25 years’ application of RSC and CF were also investigated 

without any addition of RSC or CF. The rice variety used was 

KDML 105. 

The results showed that average rice grain yield over 25 

years of application of RSC with CF was higher than grain 

yield of the application of RSC without CF. Rice yield increased 

when the rate of RSC increased. The residual effects also 

showed that the higher the RSC rate applied, the higher the 

rice grain yield obtained (Table 2). 


Table 1. Rice grain yield average for 3 years (kg ha –1 ) as affected by the application of chemical fertilizer and farmyard manure in different locations in northeast 

Thailand. a Loamy sand soil Sandy loam soil Clay soil 

Treatment Ubon Ubon Tung Kula Sakon Khon Udon Surin Chumpae Phimai Treatment 

irrigated rainfed Ronghai Nakorn Kaen mean 

pH 4.2, pH 4.1, pH 4.6, pH 7.1, pH 4.5, pH 5.7, pH 4.0, pH 4.2, pH 4.6, 

OM 0.52% OM 0.60% OM 0.77% OM 0.96% OM 0.39% OM 1.32% OM 0.59% OM 1.18% OM 1.57% 

Nil 1,316 b 857 cd 1,154 b 1,632 de 1,310 c 2,674 2,271 c 1,993 d 1,426 c 1,634 d 

FYM 2,234 a 928 cd 2,016 a 2,488 ab 1,669 bc 2,636 3,192 ab 2,876 c 2,371 b 2,264 c 

PK 1,423 b 689 d 1,272 b 1,428 e 1,312 c 2,605 2,314 c 2,206 d 1,639 c 1,655 d 

NPK 2,408 a 1,318 abc 1,602 ab 1,981 c 2,224 ab 3,005 2,896 ab 3,473 a 3,168 a 2,425 bc 

CR-NPK 2,129 a 1,458 ab 1,638 ab 1,907 cd 2,286 a 3,020 3,045 ab 3,012 bc 3,262 a 2,385 bc 

All 2,454 a 1,661 a 2,005 a 2,341 b 2,195 ab 2,773 3,244 a 3,350 ab 3,225 a 2,559 ab 

N 2,204 a 1,158 bc 1,252 b 1,990 c 2,152 ab 2,957 2,626 bc 3,077 abc 2,992 a 2,240 c 

FYM+NPK 2,624 a 1,424 ab 2,221 a 2,696 a 2,467 a 2,812 3,290 a 3,503 a 3,288 a 2,680 a 

Location mean 2,099 1,186 1,645 2,058 1,952 2,810 2,859 2,936 2,681 2,230 

a At each site, means followed by a common letter are not significantly different at the 5% level by Duncan’s multiple range test. 

Source: Naklang et al (1998). 

Table 2. Rice grain yield (kg ha –1 ) as affected by rice straw compost (RSC) with and without 

chemical fertilizer (CF) and their residual effects (RSC and CF were applied annually from 

1976 to 2000 and their residual effects were evaluated from 2001 to 2003). a 

Effects of RSC and CF applied Residual effects of RSC and CF 

for 25 years on RD 7 grain yield on KDML 105 grain yield 

RSC rate 

(kg ha –1 ) CF rate (kg of N-P 2 O 5 -K 2 O ha –1 ) CF rate (kg of N-P 2 O 5 -K 2 O ha –1 ) 

0-0-0 50-50-50 RSC mean 0-0-0 50-50-50 RSC mean 

0 2,409 3,308 2,858 e 2,498 2,961 2,728 c 

3,125 2,831 3,877 3,354 d 3,040 3,215 3,127 b 

6,250 3,161 3,999 3,580 c 2,983 3,410 3,197 b 

9,375 3,344 4,188 3,766 b 3,436 3,675 3,555 a 

12,500 3,527 4,360 3,944 a 3,542 3,656 3,598 a 

CF-mean 3,054 B 3,946 A 3,099 B 3,383 A 

a In each column, means followed by a common letter are not significantly different at the 5% level by Duncan’s 

multiple range test. 

Naklang and Rojanakusol (1992) also reported the longterm 

effects of compost, green manure (sunnhemp: Crotalaria 

juncea Linn.), and dry rice straw incorporated into soil at 0, 

6.25, 12.5, 18.75, and 25 t ha –1 for 8 years and 5 years without 

any organic or inorganic fertilizer. The rice varieties were 

KDML 105 and RD 7. Grain yield of KDML 105 increased 

with an increasing rate of compost, whereas RD 7 gave the 

highest response to green manure followed by compost. The 

application of green manure up to 18.75 t ha –1 gave a higher 

grain yield of KDML 105 and then dropped at the rate of 25 t 

ha –1 . Dry rice straw gave the lowest yield among three materials. 

The residual effects of dry rice straw and compost gave no 

difference in grain yield and obtained a higher rice yield than 

the residual effects of green manure. 

The annual application of 1.5 kg dry matter ha –1 of leaf 

litter differed in the breakdown rate from a legume crop and 

three kinds of trees for five seasons resulted in an increase in 

rice grain yield of 23–48% above the no-leaf-litter control 

(Whitbread et al 1999) and increased the total and labile carbon 

pool in the soil by 24–37% (Naklang et al 1999). A higher 

rate of chemical fertilizer did not result in increased soil carbon. 

Crop residue, leaf litter, and green manure with slow breakdown 

rates are needed to rehabilitate soil C and nutrients, which 

will benefit the following crop. In the NE Thailand climate, 

there is significant potential for a rapid breakdown of any re- 


maining residue during the first rain of the next season, but 

before the crop can use the nutrients that are released (Naklang 

et al 1999). 

Supapoj et al (1998) reported on some long-term experiments 

conducted in the sandy soils of NE Thailand. The 

application of rice husks at 3.13 t ha –1 annually for 17 years 

showed that the application of rice husks alone and plus chemical 

fertilizer produced grain yield about 15% and 74% higher 

than the no-fertilizer control treatment (1.97 t ha –1 ). In dryseeded 

rice cultivation, straw mulching gave better grain yield. 

The effects of integrating mungbean (Vigna radiata) with rice 

(mungbean seeds were broadcast together with rice seed) on 

rice yield were less because mungbean did not establish well 

under water-saturated acid infertile soil. 

The importance of both organic and inorganic (chemical 

or synthetic) fertilizer has been recognized for the supply 

of plant nutrients to ensure efficient crop production. These 

fertilizers have some weak and strong points. 

Although the application rate of organic materials was 

higher than that of chemical fertilizer and they needed more 

labor to prepare and apply to or incorporate into the soil, they 

decomposed and served as a very important source of plant 

nutrients and also accumulated as organic matter to maintain 

long-term soil fertility. To avoid hard work and the transportation 

of organic fertilizer, field crops, trees, and shrub legumes 

could be introduced to grow in rice fields for producing biomass. 

Recently, many countries have become interested in organic 

agriculture. Also, the Thai government announced Surin 

as an Organic Farming Model Pilot Project in November 2001. 

Some of the principles of organic crop production are to protect 

and minimize all forms of pollution of soil water and air 

as a result of agricultural practices, to develop and implement 

a soil improvement program and maintain long-term soil fertility, 

to maintain biodiversity within the farming system, to 

recycle plant materials and resources within the farm, and to 

handle products and processing methods to maintain the organic 

integrity and vital qualities of produce from harvesting 

to consumption. 

Many research results can be applied and introduced to 

farmers and can support some principles in sustainable and 

organic crop production. In the future, research will focus on 

green manure and organic fertilizer management, develop cropping 

systems in organic rice fields, and evaluate the effects of 

organic rice production on soil quality. 

Conclusions 

Rice productivity and paddy soil fertility in northeast Thailand 

can be improved by using chemical fertilizer and organic 

fertilizer such as compost, farmyard manure, rice straw, rice 

husk, and leaf litter. Apart from management of the cropping 

system, the use of perennial shrubs or tree legumes as a source 

of leaf litter has been recommended by planting these trees 

near (or on the bund of) the paddy field. 

References 

Fukai S, Rajatasereekul S, Boonjung H, Skulkhu E. 1995. Simulation 

modeling to quantify the effect of drought for rainfed 

lowland rice in northest Thailand. In: Fragile lives in fragile 

ecosystems. Proceedings of the International Rice Research 

Conference, 13-17 Feb. 1995. Manila (Philippines): International 

Rice Research Institute. 976 p. 

Koopmans C, Goldstein W. 2001. Soil organic matter budgeting in 

sustainable farming with applications to southeastern Wisconsin 

and northern Illinois. Bulletin No. 7. Michael Fields Agricultural 

Institute. 39 p. 

Naklang K, Rojanakusol S. 1992. Long-term effect of compost, green 

manure and rice straw on rice yield and their residual effects. 

Rice Res. J. 1:31-41. (In Thai, English abstract.) 

Naklang K, Harnpichitvitaya D, Wade LJ, Jearakongman S, Ekasit 

Skulkhu E, Romyen P, Phasopa S, Suriya-arunroj D, 

Khunthasuvon S, Kraisorakul D, Youngsuk P. 1998. Management 

× environment interaction and quantification of factors 

limiting productivity of rainfed lowland rice in Northeast 

Thailand. In: Annual rice research report for 1998. Bangkok 

(Thailand): Rice Research Institute. p 639-666. 

Naklang K, Whitbread A, Lefroy R, Blair G, Wonprasaid S, Konboon 

Y, Suriya-arunroj D. 1999. The management of rice straw, 

fertilizers and leaf litter in rice cropping systems in Northeast 

Thailand. 1. Soil carbon dynamics. Plant Soil 209:21-28. 

Poulton PR. 1995. The importance of long-term trials in understanding 

sustainable farming systems: the Rothamsted experience. 

Austr. J. Exp. Agric. 35:825-834. 

Ragland JL. 1997. Managing soil acidity in northest Thailand. In: 

Fukai S, Cooper M, Salisbury J, editors. Breeding strategies 

for rainfed lowland rice in drought-prone environments. Proceedings 

of an International Workshop held at Ubon 

Ratchathani, Thailand, 5-8 November 1996. ACIAR Proceedings 

No. 77. p 286-299. 

Supapoj N, Naklang K, Konboon Y. 1998. Using organic material to 

improve soil productivity in rainfed lowland rice soils. In: 

Ladha JK, Wade LJ, Dobermann A, Reichardt W, Kirk GJD, 

Piggin C, editors. 1998. Rainfed lowland rice: advances in 

nutrient management research. Proceedings of the International 

Workshop on Nutrient Reserch in Rainfed Lowlands, 12-15 

Oct. 1998, Ubon Ratchathani, Thailand. Manila (Philippines): 


Whitbread A, Blair G, Naklang K, Lefroy R, Wonprasaid S, Konboon 

Y, Suriya-arunroj D. 1999. The management of rice straw, 

fertilizers and leaf litter in rice cropping systems in northeast 

Thailand. 2. Rice yield and nutrient balance. Plant Soil 209:29- 

36. 

Notes 

Author’s address: Surin Rice Research Center, Surin, Thailand, 

32000, e-mail: kunnika_naklang@hotmail.com. 


Site-specific nutrient management and the sustainability 

of phosphorus and potassium supply in irrigated rice soils 

of Asia 

C. Witt, A. Dobermann, R. Buresh, S. Abdulrachman, H.C. Gines, R. Nagarajan, S. Ramanathan, P.S. Tan, and G.H. Wang 

The management of soil phosphorus (P) and potassium (K) is 

receiving greater attention in intensive irrigated lowland rice 

systems of Asia because of concerns that fertilizer P and K 

rates are not optimally adjusted to long-term needs (Greenland 

1997, Johnston and Syers 1998). Breeding offers limited opportunities 

to change plant nutrient requirements or uptake 

efficiencies so that long-term management strategies must focus 

on overcoming immediate nutrient deficiencies and maintaining 

adequate nutrient balances in the top-soil layer. Current 

fertilizer P and K strategies in Asia are still mostly based 

on soil tests that have shown little correlation with the effective 

nutrient supply to the irrigated rice crop (Dobermann et al 

2003b). Site-specific nutrient management (SSNM) allows an 

effective management of indigenous P and K supplies by estimating 

fertilizer requirements based on yield level, yield response 

to fertilizer P and K application, and a nutrient balance 

model (Witt et al 2002, 2004). In that approach, fertilizer P 

and K maintenance rates are commonly developed based on 

only two seasons of on-farm experiments, which may be insufficient 

to develop long-term strategies. In this paper, we 

evaluate SSNM strategies to prevent soil nutrient depletion 

using data from five long-term experiments established from 

1968 to 1995 in China, India, Indonesia, the Philippines, and 

Vietnam. 

Indigenous nutrient supplies 

Nutrient supplies from indigenous sources can be estimated 

by measuring plant nutrient uptake or grain yield in nutrient 

omission plots, which receive all nutrients applied as fertilizer 

except for the omitted nutrient of interest (Dobermann et al 

2003a). In a large multinational research project, omission plots 

have been used for several seasons to estimate the spatial and 

temporal variation in indigenous nutrient supplies in farmers’ 

fields (Dobermann et al 2004). These on-farm experiments have 

confirmed for a wide range of soils and growing environments 

that farmers’ current fertilizer P and K management causes 

slightly positive P balances at a number of sites, while fertilizer 

K application is commonly insufficient to balance K removal 

with grain and straw. Apart from unavoidable inaccuracies 

in the estimation, nutrient balances probably supply limited 

information regarding the risk of nutrient depletion, unless 

combined with estimates of medium- or long-term changes 

in indigenous nutrient supplies. In our past on-farm research, 

however, the location of the omission plots was changed every 

season in individual farmers’ fields to avoid residual effects 

that occur when fertilizer is omitted for longer periods. In the 

following, we will re-evaluate this strategy. 

Long-term P and K management strategies 

In the SSNM approach, fertilizer P and K rates are based on 

the yield difference between treatments with optimal fertilizer 

use of N, P, and K (NPK) and omission plots that receive all 

nutrients except for the omitted one (0P, 0K). Recommended 

fertilizer rates would increase with an increasing yield gap 

between NPK and 0P or 0K treatments to overcome greater 

nutrient limitations. Maintenance fertilizer P and K rates are 

suggested where a yield response to P or K application is lacking, 

and fertilizer rates would increase with increasing yield 

level to balance greater nutrient removal with grain and straw 

(Witt et al 2002, Witt and Dobermann 2004). The NPK, 0P, 

and 0K treatments were also included in five long-term experiments 

with two rice crops per year, and the cumulative 

yield differences between NPK and omission plots for a period 

of up to 15 seasons are plotted in Figure 1. Differences 

were substantial in short- and long-term yield responses to fertilizer 

P and K application among sites. Yields were not affected 

by K application in India or P application in China in 

the first two seasons, and initial yield responses to fertilizer 

application were generally small except for Vietnam (P) and 

China (K). However, yield responses developed within a few 

seasons at all sites except for K application at Omon in the 

Mekong River Delta of Vietnam, where annual flooding supplies 

a large K load through sedimentation (Nguyen 2003). 

Yield responses developed linearly except for China, where 

yield responses to P but not K application developed exponentially 

within eight seasons or four years. Long-term fertilizer 

requirements would generally be underestimated if fertilizer 

requirements were based only on short-term yield responses 

without considering nutrient removal with grain and 

straw and the overall input-output balance. 

In the absence of a yield response to fertilizer P and K 

application, SSNM recommends fertilizer P and K maintenance 

rates that are calculated based on a nutrient input-output model 

(Witt and Dobermann 2004). Since initial yield responses appear 

to be generally small based on the data presented from 

long-term experiments, fertilizer requirements with SSNM 

could, to a larger extent, follow maintenance strategies. At issue 

is whether SSNM recommendations would need to be adjusted 

if yield responses to fertilizer P and K application were 

evaluated for longer periods. Table 1 shows fertilizer require- 


SD Grain yield (t ha –1 ) 

20 

Philippines 

(PhilRice) 

S. Vietnam 

(CLRRI) 

Indonesia 

(IIRR) 

15 

10 

5 

0 

20 

1 3 5 7 9 11 13 15 

India 

(TNRRI) 

1 3 5 7 9 11 13 15 

China 

(ARS) 

1 3 5 7 9 11 13 15 

Season 

15 

(NPK) – (OP) 

(NPK) – (OK) 

10 

5 

0 

1 3 5 7 9 11 13 15 1 3 5 7 9 11 13 15 

Season 

Season 

Fig. 1. Cumulative grain yield increase in fully fertilized plots (NPK) over plots without fertilizer P (0P) and K (0K) application 

in long-term experiments at five sites in Asia, 1995-2002. PhilRice = Philippine Rice Research Institute, Muñoz, Philippines; 

CLRRI = Cuu Long Rice Research Institute, Omon, Vietnam; IIRR = Indonesian Institute for Rice Research, Sukamandi, 

Indonesia; TNRRI = Tamil Nadu Rice Research Institute, Tamil Nadu, India; ARS = Agricultural Research Station, Zhejiang, 

China. 

ments calculated with SSNM based on data from two and eight 

cropping seasons in the long-term experiments. Note that the 

actual fertilizer P and K rates were several times higher than 

the recommended rates with SSNM. High fertilizer P and K 

rates in long-term experiments are usually chosen to ensure 

that these nutrients are not limiting. 

Considering two seasons of data, recommended fertilizer 

rates ranged from 21 to 32 kg P 2 O 5 ha –1 and from 20 to 48 

kg K 2 O ha –1 . Suggested rates were highest where both yield 

and yield responses to fertilizer application were high (Philippines, 

India, China). Fertilizer rates changed at some but not 

all sites when eight instead of two seasons of data were considered 

in the calculation of fertilizer rates. Fertilizer rates 

needed little adjustment at sites where yield responses developed 

steadily and yields in NPK treatments remained more or 

less constant (e.g., Philippines). Adjustments in fertilizer rates 

were needed where yield levels in the first year were not representative 

for a longer time period, as in Indonesia, where 

yields increased with time. Adjustments were also needed 

where average yield responses to fertilizer application were 

either lacking or very strong over the four-year period. A lacking 

yield response over a long period indicates a strong soil 

nutrient-supplying power, while a strong yield response developed 

within a few seasons is a sign of a low soil buffering 

capacity for nutrient depletion. For example, four years of data 

from Omon revealed greater P but lower K requirements than 

the rates calculated after one year. Despite these adjustments, 

the initial estimates of fertilizer requirements showed a good 

congruence with the fertilizer rates that were based on several 

seasons of data. 

Conclusions 

Two-season estimates of indigenous P and K supplies provide 

a first, sufficiently robust estimate of fertilizer P and K requirements. 

However, on-farm monitoring of indigenous nutrient 

supplies for longer time periods is essential to fine-tune 

fertilizer P and K recommendations based on historical trends 

in soil fertility. Omission plots embedded in farmers’ fields 

should be continued at the same location for several seasons if 

data from representative long-term experiments are not available 

to identify the response pattern that develops with time. 


Table 1. Average fertilizer rates and fertilizer requirements with SSNM (Witt et al 

2002, 2004) based on the average yield increase (response) in N-, P-, and K-fertilized 

plots (NPK) over plots without fertilizer P (0P) and K (0K) application during two and 

eight cropping seasons in five long-term experiments in Asia. 

Parameter 

Unit 

Sites 

Philippines Vietnam Indonesia India China 

(PhilRice) (CLRRI) (IIRR) (TNRRI) (ARS) 

Fertilizer rates 

Starting year 1968 1995 1995 1995 1997 

Fertilizer P 2 O 5 kg ha –1 60 57 57 57 57 

Fertilizer K 2 O kg ha –1 60 90 120 120 120 

Short-term fertilizer requirements 

Number of seasons 2 2 2 2 2 

Grain yield (NPK) t ha –1 6.3 5.0 5.2 7.0 6.0 

Yield response to P t ha –1 0.5 0.8 0.1 0.5 0.0 

Yield response to K t ha –1 0.2 0.2 0.4 0.0 0.7 



Long-term fertilizer requirements 

Number of seasons 8 8 8 8 8 

Grain yield (NPK) t ha –1 6.0 4.7 6.0 6.3 5.6 

Yield response to P t ha –1 0.7 1.5 0.2 0.5 1.5 

Yield response to K t ha –1 0.4 0.1 0.3 0.5 0.8 



References 

Dobermann A, Witt C, Abdulrachman S, Gines HC, Nagarajan R, 

Son TT, Tan PS, Wang GH, Chien NV, Thoa VTK, Phung CV, 

Stalin P, Muthukrishnan P, Ravi V, Babu M, Simbahan GC, 

Adviento MA. 2003a. Soil fertility and indigenous nutrient 

supply in irrigated rice domains of Asia. Agron. J. 95:913- 

923. 

Dobermann A, Witt C, Abdulrachman S, Gines HC, Nagarajan R, 

Son TT, Tan PS, Wang GH, Chien NV, Thoa VTK, Phung CV, 

Stalin P, Muthukrishnan P, Ravi V, Babu M, Simbahan GC, 

Adviento MA, Bartolome V. 2003b. Estimating indigenous 

nutrient supplies for site-specific nutrient management in irrigated 

rice. Agron. J. 95:924-935. 

Dobermann A, Witt C, Dawe D, editors. 2004. Increasing productivity 

of intensive rice systems through site-specific nutrient 

management. Enfield, NH (USA) and Los Baños (Philippines): 

Science Publishers, Inc., and International Rice Research Institute 

(IRRI). p 1-410. 

Greenland DJ. 1997. The sustainability of rice farming. Wallingford 

(UK) and Manila (Philippines): CAB International and International 


Johnston AE, Syers JK, editors. 1998. Nutrient management for sustainable 

crop production in Asia. Wallingford (UK): CAB International. 

p 1-394. 

Nguyen MH. 2003. Soil potassium dynamics under intensive rice 

cropping: a case study in the Mekong Delta, Vietnam. 

Wageningen University, Netherlands. p 1-203. 

Witt C, Balasubramanian V, Dobermann A, Buresh RJ. 2002. Nutrient 

management. In: Fairhurst T, Witt C, editors. Rice: a practical 

guide for nutrient management. Singapore and Los Baños: 

Potash and Phosphate Institute & Potash and Phosphate Institute 

of Canada and International Rice Research Institute. 

p 1-45. 

Witt C, Buresh RJ, Balasubramanian V, Dawe D, Dobermann A. 

2004. Principles and promotion of site-specific nutrient management. 

In: Dobermann A, Witt C, Dawe D, editors. Increasing 

productivity of intensive rice systems through site-specific 

nutrient management. Enfield, NH (USA) and Los Baños 

(Philippines): Science Publishers, Inc., and International Rice 


Witt C , Dobermann A. 2004. Toward a decision support system for 

site-specific nutrient management. In: Dobermann A, Witt C, 

Dawe D, editors. Increasing productivity of intensive rice systems 

through site-specific nutrient management. Enfield, NH 

(USA) and Los Baños (Philippines): Science Publishers, Inc., 

and International Rice Research Institute. p 359-396. 


Notes 

Authors’ addresses: C. Witt and R. Buresh, International Rice Research 

Institute (IRRI), DAPO Box 7777, Metro Manila, Philippines; 

A. Dobermann, University of Nebraska, Lincoln, NE 

68583-0915; S. Abdulrachman, Indonesian Institute for Rice 

Research (IIRR), Sukamandi, West Java 41256, Indonesia; 

H.C. Gines, Philippine Rice Research Institute (PhilRice), 

Muñoz, Nueva Ecija 3119, Philippines; R. Nagarajan, Soil 

and Water Management Research Institute (SWMRI), 

Thanjavar, Tamil Nadu 613501, India; S. Ramanathan, Tamil 

Nadu Agricultural University (TNAU), Coimbatore 641003, 

India; P.S. Tan, Cuu Long Delta Rice Research Institute 

(CLRRI), Omon, Vietnam; G.H. Wang, Zhejiang Agricultural 

University, Hangzhou, Zhejiang 310029, China. Current address: 

C. Witt, Potash & Phosphate Institute/Potash & Phosphate 

Institute of Canada and International Potash Institute, 

Southeast Asia Program, 126 Watten Estate Road, Singapore 

287599, e-mail: cwitt@ppi-ppic-ipi.org. 

Acknowledgments: The International Rice Research Institute (IRRI), 

the Swiss Agency for Development and Cooperation (SDC), 

the Potash and Phosphate Institute/Potash and Phosphate Institute 

of Canada (PPI/PPIC), and the International Potash 

Institute (IPI) provided funding for this research. 

Ecological engineering for sustainable rice production 

and the restoration of degraded watersheds in West Africa 

Toshiyuki Wakatsuki, Md. Moro Buri, and Oluwarotimi O. Fashola 

During 1970-2000, although rice production increased from 

2.4 to 7.4 million tons in West Africa (WA), consumption per 

capita increased from 15 to 30 kg and imports increased from 

0.7 to 4.3 million tons. Yield, however, has been stagnant at 

1.3–1.7 t ha –1 during the past 30 years in this region. More 

than 100 high-yielding varieties for various rice ecologies are 

available and another 6 NERICA varieties are in the pipeline 

for upland areas. The lowland elite varieties under the sawah 1 

system with good management could yield up to 6 t ha –1 . But 

these potentials are never achieved in farmers’ fields because 

of poor water control and bad soil management systems. This 

is why a renewed call for a concerted effort needs to be made 

for better water and soil management systems, referred to in 

this paper as the sawah system (Wakatsuki et al 1998), and 

correct policies for rice development. 

Why has the Green Revolution not yet occurred in West 

Africa in spite of its successes in Asia in the 1960s The Green 

Revolution laid the foundation for the rapidly growing economies 

of Asia today. As our team revealed, the ecological environment 

of soil and water conditions in this region is very severe. 

Lowland soil fertility in WA may be the lowest among 

the major tropical areas in the world (Issaka et al 1997, Buri et 

al 2000, Hirose and Wakatsuki 2002). The main cause of the 

present agricultural and environmental crises in WA, however, 

may be the underdevelopment of lowland agriculture. Environmentally 

creative technology, or ecological engineering 

technology, such as lowland sawah farming, is not traditionally 

practiced in WA. Sawah is a multifunctional constructed 

wetland, which is the prerequisite for realizing the Green Revolution 

as well as for preserving and even restoring ecological 

environments. Irrigation and drainage without farmers’ sawah 

farming technologies have proved inefficient or even damaging 

because of accelerated erosion processes. Thus, the development 

of irrigation has been slow. In the absence of water 

control, fertilizers cannot be used efficiently. Consequently, 

the high-yielding varieties perform poorly and soil fertility 

cannot be sustained. Hence, the Green Revolution cannot take 

place. 

The sawah system and integrated watershed approach 

The upper part of Figure 1 shows a concept of macroscale 

ecological engineering, that is, watershed ecological engineering. 

The soils formed in uplands and the nutrients released 

during rock weathering and soil formation processes in uplands 

are accumulated in lowlands (geological fertilization). 

If the sawah system exists in the lowlands, it can store and 

effectively use the nutrient-rich water and fertile topsoils eroded 

from the uplands. Watershed agroforestry through the integration 

of upland forestry and lowland sawah systems in a unit 

watershed as seen in the upper part of Figure 1 is a typical 

model of watershed ecological engineering. Optimum landuse 

patterns and landscape management practices optimize 

geological fertilization through optimum hydrology in a given 

watershed. This is an eco-environmental basis for the longterm 

intensive sustainability of sawah-based rice farming in 

Asia. 

1 The sawah system is the rice farmers’ basic infrastructure for intensive and 

sustainable rice production. The term “sawah” refers to a leveled rice field 

surrounded by bunds with inlet and outlet connections to irrigation and drainage 

canals. The term originates from Malayo-Indonesian. The English term 

“paddy or paddi” also originates from the Malayo-Indonesian term “padi,” 

which means rice plant. However, the term “paddy” refers to rice grain with 

husk in West Africa as a whole. The paddy field is almost equivalent to the 

sawah for Asian scientists. However, the term “paddy fields” refers to just a 

rice field, including an upland rice field in West Africa. Therefore, to avoid 

confusion among the terms “rice plant,” “paddy,” and the improved manmade 

rice-growing environment through ecological engineering, the authors 

propose to use the term “sawah.” 


(A) Optimum land-use patterns and landscape management practices optimize geological 

fertilization through the control of optimum hydrology in the watershed 

Rainfall 

Watershed 

agroforestry 

Application of various types 

of humified organic matter 

Intensive sustainability in uplands 

= combination with stock raising 

Uplands: 1,000–5,000 ha 

Nutrient cycle 

Rock weathering, soil formation, 

and nutrient release 

Topsoil 

Geological fertilization 

Nutrient-rich water 

Topsoil erosion 

Village 

Small irrigation canal 

Sedimentation 

Stream discharge 

Small reservoir 

Fish culture pond 

Lowlands = about 5% 

Supply of 

nutrients by 

animals 

Sawah hypothesis: 1 ha of sawah fields 

is equivalent to 10–15 ha of uplands 

Geological 

fertilization 

A few mm 

to 2 cm 

10–20 cm 

O 2 + N 2 CH 4 

CO 2 

emission 

Technology development for nitrogen fixation through 

integrated management of soil, water, algae, and rice 

N 2 

biological 

nitrogen 

N 2 O N2 (N 2 O, NO) 

fixation Algae 

Submerged 

Algae 

Nitrificaton Denitrificaton water 

CH 4 →CO 2 NH 3 →NO 2 →NO 3 +400–500 mV Oxidized 

layer 

Organic N 2 

NO 3 N 2 (N 2 O, NO) 

NO Mn (IV) 

3 ←NO 2 ←NH 4 -N 

Mn (II) –100 mV 

CO 2 CH 4 

Fe (III) 

SO 4 

N 2 NH 4 

CO 2 CH 4 

CH 3 COOH CH 4 

Fe (II) 

H 2 S 

–250 mV 

(Eh) 

Reduced 

layer 

Plow layer 

O 2 , N 2 

CH 4 →CO 2 

Accumulating layer Fe 2 CO 3 ⋅ nH 2 O 

Accumulating layer MnO 2 ⋅ nH 2 O 

(B) The sawah system as multifunctional constructed wetlands 

Fig. 1. Macro- and microscale ecological mechanisms of intensive sustainability of the lowland 

sawah system: (A) geological fertilization through watershed ecological engineering and 

(B) multifunctional constructed wetlands for enhanced supply of N, P, Si, and other nutrients. 

Although the rate of soil formation, and thus also soil 

erosion, is much higher in Asia than in Africa, suppose that the 

soil formation rate in uplands, which make up 95% of the total 

area in the example of a watershed shown in Figure 1, is 1 t 

ha –1 y –1 . In a stable ecosystem in a watershed, the rates of soil 

formation and erosion should be well balanced; therefore, the 

topsoils formed in uplands—which account for 95% of the 

area—and the nutrients produced in the process will be concentrated 

in lowlands, which make up 5% of the area. Thus, 

the soil formation rate in the lowlands equals 20 t ha –1 y –1 

(Hirose and Wakatsuki 2002). Though it will be impossible to 

use all of the rich soils and nutrient-rich water from the uplands 

effectively, the sawah farming system could artificially 

reinforce the geological fertilization processes. The quantitative 

scientific evaluation of the geological fertilization processes 

in a watershed will be an important future research subject. 

The excessive decomposition of organic matter in tropical 

soils is another problem. Therefore, organic matter management 

through agroforestry and cover-crop systems is important 

for sustaining soil fertility. Sustainability under intensive 

cultivation of uplands is possible through integration with 

stock raising. Animals can accumulate nutrients on upland 

farms and lowlands. However, except for lowland sawah farming, 

these options are not successful enough to make tropical 

upland farming sustainable under intensive cultivation. The 

sawah system can also control such excessive organic matter 

decomposition (Kyuma 2003). Another possibility to restore 


upland soil fertility is the application of various refractory humified 

materials to soil (Wakatsuki et al 2003). 

When the unit yields of upland slash-and-burn rice cultivation 

are compared with that of lowland sawah rice cultivation, 

the latter (2.5 t ha –1 ) is approximately 2.5 times higher 

than the former (1 t ha –1 ) under conditions of no-fertilizer application. 

In addition, the planting of rice in rainfed upland 

areas must be followed by a fallow period of at least 4–5 years 

to allow restoration of soil fertility, that is, to sustain 1 ha of 

upland rice cultivation based on slash and burn, 4–5 ha of upland 

are necessary. In comparison, continuous cultivation is 

possible with sawah fields. 

Therefore, when these two types of cultivation are compared 

for a long period of 10-plus years, which are necessary 

to sustain a complete cycle of slash-and-burn cultivation, taking 

the above facts into consideration, the difference in 

sustainability per unit of productivity can be more than tenfold, 

that is, the yield difference (2.5) times the difference in 

required area (4–5 times) for sustainable production. 

Accordingly, the development of 1 ha of sawah field 

enables the conservation or regeneration of 10 ha or more of 

forest area. Sawah fields can therefore contribute to not only 

increased food production but also conservation of the forest 

environment as well as soil and water conservation in the whole 

watershed, resulting in enhanced sustainability of an intensive 

lowland sawah field system. Furthermore, these fields can contribute 

to the alleviation of global warming and other global 

environmental problems through the fixation of carbon to forests 

and forest soil. 

The sawah system as multifunctional wetland 

The lower part of Figure 1 shows the microscale mechanisms 

of intensive sustainability of the sawah system. The sawah system 

can be managed by multifunctional constructed wetlands. 

It is well known that weeds can be controlled by means of 

water. But it is not well evaluated that the nitrogen fixation 

amount of soil microbes under a submerged sawah system 

reaches 20–100 kg ha –1 year –1 in Japan and 20–200 kg ha –1 

year –1 in the tropics depending on the level of soil fertility and 

water management (Greenland 1997, Kyuma 2003, Hirose and 

Wakatsuki 2002). This amount is comparable with the nitrogen 

fixation amount of leguminous plants. Rainfed upland farming 

has no such option but to rely on the use of leguminous 

plants, animal dung, other organic fertilizer, and/or chemical 

fertilizer. Under submerged conditions, because of reduction 

of ferric iron to ferrous iron, phosphorus availability is increased 

and acid pH is neutralized; hence, micronutrient availability is 

also increased (Kyuma 2003). These are the other benefits of 

the sawah system. These eutrophication mechanisms not only 

encourage the growth of the rice plant but also encourage the 

growth of various algae that increase nitrogen fixation. The 

quantitative evaluation of nitrogen fixation in the sawah system, 

including the role of algae, will also be an important future 

research topic. Under nitrate-rich submerged water conditions, 

the sawah system encourages the denitrification process. 

Easily decomposable organic materials become substrate 

of various denitrifiers. Another function of the sawah system 

is the purification of nitrate-polluted water. 

State of rice cultivation in West Africa 

Rice is the only crop that can be grown in all the agroecological 

zones of WA: (1) rainfed upland, (2) rainfed lowland, (3) irrigated 

lowland, and (4) deepwater/mangrove swamp. According 

to a recent survey by WARDA (1988, 1999), the percentage 

of area under rice by ecology is as follows: 31% for upland, 

47% for lowland, 16% for irrigated, and 5% for deep 

water. The percentages of upland rice area and production 

decreased dramatically in 1990-2000, 57% to 31% and 42% 

to 25%, respectively. Rice cultivation in WA was traditionally 

an extension of upland farming. However, following the pioneering 

technical cooperation activities of Taiwanese teams 

regarding widespread and intensive sawah-based farming for 

some 10 years in the 1960s and 1970s, the number of rice 

farmers who are consciously conducting water management 

has steadily increased over the last 30–40 years (Hsieh 2001). 

Such management involves the introduction of bunding, leveling, 

construction of dams, dikes, and weirs, and extension of 

water canals. Consequently, there are now many types of rice 

cultivation in WA in terms of water control, ranging from upland 

rice cultivation to irrigated lowland sawah (Fig. 2). 

Restoration of degraded inland valley watersheds 

in West Africa: a case study of an ecological engineering 

project 

In the past 16 years, the authors’ group carried out various 

surveys and trials for on-farm research and participatory as 

well as farmers’ self-support development at two benchmark 

watersheds in the Guinea Savanna Zone (Nigeria) and Forest 

Transitional Zone (Ghana). The major target was to develop 

sustainable eco-technologies to increase food, especially rice, 

and at the same time restore the degraded watershed. The results 

showed that various sawah-based rice farming activities 

in lowlands are a key technology (Wakatsuki et al 1998, 2001, 

Hirose and Wakatsuki 2002). 

In 2003, a new but continuing five-year research project 

had started under the title “Watershed Ecological Engineering 

for Sustainable Increase of Food Production and Restoration 

of Degraded Environment in West Africa” (Wakatsuki 2004). 

The major outcome will be to consolidate the long-term comprehensive 

plan for sustainable development of 10–20 million 

ha of the lowland sawah system in WA. Based on the sustainable 

increase in food production through sawah development, 

our final target is to regenerate 100–200 million ha of forest in 

WA. 

Reference 

Buri MM, Ishida F, Kubota D, Masunaga T, Wakatsuki T. 2000. Sulfur 

and zinc levels as limiting factors to rice production in West 

Africa lowlands. Geoderma 94:23-42. 


Forest destruction induced 

savannization 

Upland-lowland 

continuum 

Intensified rainfed to irrigated 

lowland sawah system 

NERICA with 

bunding, leveling 

Water table 

Geological 

fertilization 

Drought 

Flood 

Some level of water control 

Functional constructed 

wetland 

Nitrogen fixation 

(20–200 kg ha –1 y –1 ) 

Geological fertilization and enhanced 

supply of Si, Ca, Mg, K, P 

Spring irrigation 

Bunding 

Evaporation 

Stabilized 

Fragile: 

no sustainable upland 

rice including NERICA 

Leveling 

Leaching runoff 

Humid zone 

Sahel zone 

Upland 

Soil conservation 

Slope upland 

No soil conservation 

Hydromorphic 

slopes (fringe) 

Lowland 

(valley bottom) 

Intensified 

lowland 

Irrigated 

lowland 

Main water supply: 

rainfall 

Rainfall Rainfall + 

water table 

seepage in 

spring 

Rainfall + 

water table + 

floodwater in spring 

Regulated 

floodwater in spring 

Irrigation dams, 

wells, pumps 

Agroecological zone: 

Guinea savanna to 

humid forest zone 

Guinea savanna to 


Guinea savanna 

to humid forest 

zone 

Sudan savanna to 


Sudan savanna to 


Sahel to 


Rudimentary sawah Rainfed sawah Irrigated sawah 

Rainfed upland 

Rainfed lowland 

Fig. 2. Rice ecologies along a continuum of inland valley watersheds and floodplains in West Africa (excluding the ecologies of deepwater 

and mangrove swamp rice). 

Irrigated 

Greenland DJ. 1997. The sustainability of rice farming. Wallingford 

(UK): CAB International, in association with the International 

Rice Research Institute. 273 p. 

Hirose S, Wakatsuki T. 2002. Restoration of inland valley ecosystems 

in West Africa. Tokyo (Japan): Norin Tokei Kyokai. 600 

p. 

Hsieh S-C. 2001. Agricultural reform in Africa, with special focus 

on Taiwan-assisted rice production in Africa, past, present and 

future perspectives. Tropics 11:33-58. 

Issaka RF, Ishida F, Kubota D, Wakatsuki T. 1997. Geographic distribution 

of selected soil parameters of inland valleys in West 

Africa. Geoderma 75:99-116. 

Kyuma K. 2003. Paddy soil science. Kyoto (Japan): Kyoto University 

Press. 305 p. 

Wakatsuki T. 2004. Watershed ecological engineering for sustainable 

increase for food production and resoration of degraded 

environment in West Africa. www.kindai-ecotech.jp. 

Wakatsuki T, Shinmura Y, Otto E, Olanian G. 1998. African-based 

sawah system for the integrated watershed management of 

small inland valley in West Africa. Water Reports 17. Institutional 

and technical opinion in the development and management 

of small-scale irrigation. Rome (Italy): FAO. p 56-79. 

Wakatsuki T, Otto E, Andah WEI, Cobbina J, Buri MM, Kubota D, 

editors. 2001. Integrated watershed management of inland 

valley in Ghana and West Africa: ecotechnology approach. 

Final Report on JICA/CRI joint study project, CRI, Kumashi, 

Ghana, and JICA, Tokyo. 337 p. 

Wakatsuki T, Matsui K, Shibata K, Matsuoka K, Masunaga T. 2003. 

A method for organic fertilizer production. Japanese Patent 

Application No. 2002-338185. 

WARDA (West Africa Rice Development Association). 1988. Strategic 

plan. Bouaké (Côte d’Ivoire): WARDA. 

WARDA (West Africa Rice Development Association). 1999. African 

rice initiative report. Bouaké (Côte d’Ivoire): WARDA. 

Notes 

Authors’ addresses: Toshiyuki Wakatsuki, Faculty of Agriculture, 

Kinki University, 3327-204 Nakamachi, Nara 631-8505, Japan, 

e-mail: wakatuki@nara.kindai.ac.jp; Md. Moro Buri, Soil 

Research Institute-CSIR, Kumasi, Ghana, e-mail: 

sawahksi@ghana.com; Oluwarotimi O. Fashola, Watershed 

Initiative of Nigeria, WIN2001, International Institute of Tropical 

Agriculture, Ibadan, Nigeria, e-mail: r.fashola@cgiar.org. 


Nitrogen cycling under the rice-wheat rotation 

and environmental effects 

Jian-guo Zhu, Xiaozhi Wang, Zu-cong Cai, Ren Gao, and Yasukazu Hosen 

Nitrogen fertilizer efficiency in flooded rice fields is usually 

poor. Upland crops frequently use 40–60% of the applied N, 

whereas flooded rice crops typically use only 20–40% (Vlek 

and Byrens 1986). This means that about 60% of the applied 

nitrogen in the rice season is released from the soil to a water 

body (surface water and groundwater) and to the atmosphere. 

For these reasons, sustainable forms of agricultural systems 

that are more stable and friendly to the environment are now 

gradually recognized by the government as a long-term policy. 

Meanwhile, to develop corresponding technologies and evaluate 

their environmental impact have become important for 

agroenvironmental researchers. Much research has been done 

in the area of NO 3 

– 

movement, N 2 O emission, and NH 3 volatilization, 

but generally only one or a few components are involved 

in one experiment. As we know, the processes in N 

cycling are correlated and they interact. Therefore, there is a 

need to determine as much as possible measurable items for N 

cycling at one time in one experiment. 

A field experiment on monolithic lysimeters was carried 

out in the 2001 rice season in the Tai-hu Lake region. The 

objective of our studies was to investigate nitrogen cycling 

data for urea applied in paddy fields managed according to the 

prevailing practices of local farmers, including losses through 

surface runoff, percolation loss, and ammonia volatilization, 

and to evaluate their impact on the environment. 


The experiments were carried out at the Changshu Agro-ecological 

Experiment Station (31º32′45″N and 120º41′57″E). The 

soil was classified as a Wushan soil, a waterlogged soil developed 

from a parent material of lacustrine deposit, with the following 

characteristics in the surface 0.15-m layer: pH 7.36 

(H 2 O); total N 1.79 g kg –1 ; organic carbon 20.0 g kg –1 ; texture 

silty sandy clay loam; bulk density 1.22 g cm –3 ; and CEC 20.20 

cmol kg –1 . 

Ten profile-undisturbed lysimeters each had an 800-mm 

inner diameter and 1,000-mm height. One percolate outlet was 

drilled at the center of the bottom. Each lysimeter was installed 

with two runoff outlets in the column wall, one at 60 mm above 

and one at 100 mm below the top surface. 

The experiment included three treatments: U550 (urea 

at 300 and 250 kg N ha –1 for rice and wheat, respectively; conventional 

level), U275 (urea at 150 and 125 kg N ha –1 for rice 

and wheat, respectively); and a control (no N). The treatments 

(3 replicates) and control (4 replicates) were arranged randomly. 

For U275 and U550 treatments, readily available urea 

was applied as basal, tillering, and booting fertilizer, with 40%, 

20%, and 40% of N applied in the whole rice season, respectively. 

All treatments received 100 kg K 2 O ha –1 and 100 kg 

P 2 O 5 ha –1 as basal fertilizer. Fertilizers for basal application 

were scattered evenly on the wet soil surface and flooded and 

puddled for transplanting. The other two applications adopted 

the topdressing method. 

For each irrigation, the soil was kept flooded at 5-cm 

depth, and supplemental irrigation water was added according 

to local practices. Meanwhile, the percolation rate was artificially 

controlled at 5 mm day –1 continuously with a medical 

transfusion apparatus. Some 7.5 L of percolate was collected 

every 3 days with a 10-L pail that contained 10 mL of 2 M 

H 2 SO 4 . 

The runoff samples were collected in the rice season. 

The rain sample was collected with a rain gauge. All water 

samples were stored in the dark at 4 °C until analyzed. 

Plant samples were collected for straw and grain nutrient 

analysis when the crop was harvested. 

The ammonia volatilization rate was measured using a 

continuous airflow enclosure method (Kissel et al 1977). The 

rate of airflow, generated by a pump, was adjusted using valves 

to 15–20 times the chamber volume per minute (Liao 1983). 

Measurement continued every day (except for a heavy rain 

day) until there were no significant differences between the 

NH 3 volatilization rates from the N treatments and the control. 

NH 4+ , NO 3– , and total N in water samples were measured 

colorimetrically by the indophenol (NH 4+ ), ultraviolet spectrophotometry 

(NO 3– ), and K 2 S 2 O 8 oxidation ultraviolet spectrophotometry 

(total N) methods. 


Ammonia volatilization potential 

The cumulative NH 3 losses from U treatments were significantly 

higher than from the control. The average cumulative 

NH 3 losses in three rice seasons were 39.43 and 100.41 kg N 

ha –1 from the U275 and U550 treatments, which corresponded 

to 24.06% and 32.34% of the applied N, respectively (Table 

1). 

The cumulative NH 3 losses in the wheat season in 2002- 

03 were 20.99 and 40.10 kg N ha –1 from the U275 and U550 

treatments, which corresponded to 11.94% and 13.61% of the 

applied N, respectively (data not shown). 

The ratio of ammonia volatilization to nitrogen applied 

increased, along with the increase in application rate in urea 


Table 1. The amount and ratio of N loss by NH 3 volatilization potential and runoff in the rice season. 

Treatments 

N loss (kg N ha –1 ) Loss ratio of N applied (%) 

2001 2002 2003 2001 2002 2003 Av 

NH 3 volatilization 

Control 0 ± 0 C a 6.92 ± 0.36 c 3.10 ± 0.85 c 

U275 24.11 ± 6.25 b 48.91 ± 4.40 b 45.27 ± 7.49 b 16.08 b 28.00 28.11 24.06 

U550 77.98 ± 11.20 a 110.67 ± 7.34 a 112.68 ± 13.34 a 25.99 34.50 36.53 32.34 

Runoff 

Control 4.1 ± 0.4 c 2.2 ± 0.4 c 4.4 ± 1.8 c 

U275 20.4 ± 4.5 b 12.3 ± 3.5 b 25.4 ± 2.0 b 10.9 6.7 14.0 10.53 

U550 44.7 ± 7.1 a 18.8 ± 3.7 a 42.7 ± 1.6 a 13.5 5.5 12.8 10.60 

a Average values in the same row without a common letter differ significantly at P = 0.01. b Loss ratio of N applied is N loss through NH 3 emissions 

minus the control value divided by N rate. 

treatments. But, the ratio of NH 3 loss to N applied in the wheat 

season was much lower than that in the rice season. 

Runoff loss 

Surface runoff occurred 5, 7, and 2 times for the 2001, 2002, 

and 2003 rice seasons, respectively. Nitrogen loss by runoff 

varied significantly with different treatments at different stages. 

Significant variance in both runoff total N (TN) concentrations 

and TN losses by runoff between different years occurred. 

Nitrogen loss by runoff was 20.4 and 44.7, 12.3 and 18.8, 

and 25.4 and 42.7 kg N ha –1 for U275 and U550 in 2001, 2002, 

and 2003, respectively. The ratios of nitrogen loss by runoff to 

N applied were 5.5–14.0% for U treatments. There was a significant 

positive correlation between nitrogen runoff losses and 

applied N amount in urea treatments; the ratio of runoff N loss 

to N applied was similar (insignificant difference at P 150 kg N ha –1 and < 300 kg N ha –1 for 

common urea under the rice-wheat rotation in this region. 

Conclusions 

The following conclusions can be drawn: 

1. Urea treatments for ammonia volatilization were remarkable 

in this region in both the rice and wheat 

seasons. 


Table 2. Field nitrogen balance under two rice-wheat 

rotations in 2001-03 (kg N ha –1 ). 

Treatments Control U275 U550 

Total income 55.3 606.1 1,156.3 

Rainfall 51.6 51.6 51.6 

Irrigation 3.7 4.5 4.7 

Fertilization 0.0 550.0 1,100.0 

Bio-uptake 190.0 448.8 630.1 

Total measured loss 23.2 176.7 483.6 

NH 3 volatilization 6.1 112.7 334.8 

NO 2 emissions 0.0 0.1 0.6 

Runoff and drainage 12.0 58.3 139.8 

Percolation 5.1 5.6 8.4 

Balance –157.9 –19.4 42.7 

2. The interval between rainfall and urea fertilization was 

the key factor in nitrogen runoff losses. The likelihood 

of N loss by runoff is not great if heavy rain 

hasn’t occurred within 5 days after fertilizer application. 

The N loss by runoff will also decrease markedly 

if the runoff outlet is heightened properly. 

3. TN loss by percolation has not been rising in parallel 

with the increasing N fertilizer application rate, and 

NO 3 – -N in groundwater is not directly related to N 

percolation from N applied to the paddy field. 

4. The N input by rainfall should be considered in research 

on nitrogen balance in the agroecosystem and 

eutrophication. 

5. A suitable urea application rate should be > 150 kg N 

ha –1 and < 300 kg N ha –1 under the rice-wheat rotation 

in this region. 

References 

Kissel DE, Brewer HL, Arkin GF. 1977. Design and test of a field 

sampler for ammonia volatilization. Soil Sci. Soc. Am. J. 

41:1133-1138. 

Liao XL. 1983. Method of nitrogen gas loss research. Prog. Soil 

Sci. 15(5):49-55. (In Chinese.) 

Nelson DW. 1982. Nitrogen in agricultural soils. Madison, Wis. 

(USA): Soil Science Society of America. p 327-363. 

Vlek PLG, Byrens BH. 1986. The efficacy and loss of fertilizer N in 

lowland rice. Fert. Res. 9:131-147. 

WHO (World Health Organization).1984. Guidelines for drinking 

water quality. Vol 1. Recommendations. Geneva (Switzerland): 

WHO. 

Zhu JG, Liu G, Han Y, Zhang YL, Xing GX. 2003. Nitrate distribution 

and denitrification in the saturated zone of paddy field 

under rice/wheat rotation. Chemosphere 50:725-732. 

Notes 

Authors’addresses: Jian-guo Zhu, Xiaozhi Wang, Zu-cong Cai, and 

Ren Gao, Institute of Soil Science, Chinese Academy of Sciences, 

Nanjing, Jiangsu 210008, China; Xiaozi Wang, Environment 

Science and Engineering College, Yangzhou University, 

Yangzhou, Jiangsu 225009, China; Ren Gao, Fujian Normal 

University, Fuzhou, Fujian 350007, China; Yasukazu 

Hosen, Japan International Research Center for Agricultural 

Sciences, 1-2 Ohwashi, Tsukuba, Ibaraki 305-8686, Japan, e- 

mail: jgzhu@issas.ac.cn. 

Urea deep placement increases yield and saves nitrogen 

fertilizer in farmers’ fields in Bangladesh 

W.T. Bowen, R.B. Diamond, U. Singh, and T.P. Thompson 

An innovative yet simple technology that increases nitrogenuse 

efficiency, and which is being widely disseminated for 

managing urea in flooded rice systems in Bangladesh, involves 

the deep placement of urea supergranules or briquettes into 

puddled soil shortly after transplanting rice. The briquettes are 

made by compressing prilled or granular urea in small machines 

with indented pocket rollers that, depending on the size 

of the pocket, produce individual briquettes varying in weight 

from 0.9 to 2.7 g. Within a week after transplanting rice, the 

briquettes are inserted into the puddled soil by hand, being 

placed to a depth of 7–10 cm in the middle of alternating 

squares of four hills of rice. Often referred to as urea deep 

placement (UDP), this technology improves N-use efficiency 

by keeping most of the urea N in the soil close to plant roots 

and out of the floodwater, where it is more susceptible to loss 

as gaseous compounds or runoff (Mohanty et al 1999, Savant 

and Stangel 1990). 

The UDP technology as applied to flooded rice in 

Bangladesh offers a compelling example of an effective approach 

to managing urea fertilizer that results in not only improved 

efficiency but also greater yield with less urea fertilizer. 

Farmer awareness of the technology has been increased 

through training and promotional activities, including on-farm 

comparisons of UDP and broadcast urea in trials managed by 

farmers. For this report, data from on-farm trials are used to 

examine changes in grain yield, urea fertilizer use, and N-use 

efficiency attributed to the use of UDP technology. 


Methods 

UDP grain yield (t ha –1 ) 

Rice is grown throughout the year in Bangladesh, with two 

principal cropping seasons: boro (irrigated dry season) and 

aman (wet season). The boro crop is transplanted in January- 

February and harvested in May-June, while the aman crop is 

transplanted in July-August and harvested in November-December. 

Yields are usually greater during the boro season because 

of more favorable temperatures and solar radiation for 

crop growth. 

9 

8 

7 

6 

5 

Boro 

Aman 

On-farm trials 

During the 2000-04 boro and aman seasons, farmers in seven 

districts (Bogra, Chandpur, Jessore, Kishoreganj, Mymensingh, 

Pabna, and Tangail) conducted 531 on-farm trials to measure 

rice yields obtained in side-by-side comparisons of UDP versus 

the current farmers’ practice (FP) of applying split applications 

of broadcast urea. Both treatments (UDP and FP) were 

managed equally except for urea, and all farmers used semidwarf 

high-yielding varieties. At harvest, grain weights and 

moisture content were determined by taking yield cuts from 

two subplots of 11.2 m 2 in each of the UDP and FP main plots. 

Grain yields are reported as paddy (unmilled rice) at a standard 

moisture content of 14%. 

Fertilizer inputs were recorded in 304 of the comparison 

trials (222 during the boro and 82 during the aman). The 

amount of N applied in the FP treatment was chosen by each 

farmer according to his own usual practice or local recommendations. 

For the FP boro, the average was 149 kg N ha –1 

(range from 71 to 239 kg N ha –1 ), and for the FP aman it was 

95 kg N ha –1 (range from 35 to 253 kg N ha –1 ). The amount of 

N applied by deep-point placement of urea briquettes (UDP 

treatment) was substantially less; the UDP boro received an 

average of 79 kg N ha –1 (range from 69 to 91 kg N ha –1 ), while 

the UDP aman received an average of 59 kg N ha –1 (range 

from 53 to 64 kg N ha –1 ). Phosphorus and K were applied 

according to local recommendations, with an average application 

of 21 kg P ha –1 and 47 kg K ha –1 . 

Parameters evaluated 

The effectiveness of UDP technology was assessed through 

the following parameters: changes in grain yield (∆Grain yield; 

kg ha –1 ), changes in fertilizer N applied (∆N applied; kg N 

ha –1 ), and changes in the efficiency of fertilizer N use as measured 

by the partial factor productivity of N (∆PFP-N; kg grain 

yield per kg N applied). These parameters were calculated for 

each individual on-farm comparison of UDP and FP as follows: 

∆Grain yield = UDP grain yield – FP grain yield 

∆Applied N = UDP applied N – FP applied N 

∆PFP-N = PFP-N for UDP – PFP-N for FP 

4 

3 

2 

1:1 line 

1 

1 2 3 4 5 6 7 8 9 

Farmers’ practice grain yield (t ha –1 ) 

Fig. 1. Rice yields (paddy) obtained in Bangladesh using deeppoint 

placement of urea briquettes (UDP) plotted against yields 

obtained using broadcast urea (farmers’ practice) in side-by-side 

comparisons in farmers’ fields during 2000-04. 

Results 

Figure 1 shows the relationship between UDP and FP grain 

yields. The 1:1 line in the figure provides a reference by which 

the performance of UDP can be compared directly with that of 

the current practice of broadcast urea (FP); hence, points to 

the left of the line indicate UDP dominance, whereas points to 

the right indicate FP dominance. Clearly, in side-by-side comparisons 

of UDP versus FP, grain yields were consistently 

greater with deep-point placement of urea briquettes. 

Two aspects of the comparison shown in Figure 1 are 

revealing. First, the yield increase with UDP was achieved using 

much less urea fertilizer than with FP. On average, the amount 

of N applied to the UDP boro crop was only 53% of that applied 

to FP boro. Likewise, the UDP aman crop received only 

62% of the N applied to FP aman. The second aspect of interest 

is the consistency of the UDP yield benefit across a wide 

range of yield environments. The regression of UDP yields on 

FP yields provided a highly significant linear relationship defined 

by the following equation: UDP = 1.02 + 1.00 FP (r = 

0.92, SE of regression coefficient = 0.019). The slope equal to 

one indicates that the average yield benefit of 1.02 t ha –1 (SE 

= 0.08) was the same for all farmers regardless of whether 

they were at the low end or the high end of the yield gradient. 

Figure 2 shows the comparative distributions for changes 

in grain yield (∆Grain yield), N fertilizer use (∆Applied N), 

and the partial factor productivity of applied N (∆PFP-N) for 

boro and aman crops. Invariably, UDP was superior to FP for 

all parameters, providing greater yield with less applied N, 

which meant that the PFP of applied N increased as well. For 

the boro season, the UDP yield benefit was greater than 700 


kg grain ha –1 

2,000 

1,500 

1,000 

500 

0 

–500 

Boro Aman Boro Aman Boro Aman 

kg grain per kg N applied 

100 

50 

0 

–50 

–100 

DGrain yield DApplied N DPFP-N 

–150 

Fig. 2. Comparative distributions showing the changes in grain yield (∆Grain yield), N 

fertilizer use (∆Applied N), and partial factor productivity of applied N (∆PFP-N) attributed 

to UDP technology versus the farmers’ practice (FP) for the boro and aman rice 

crops. The box plots display the 10th, 25th, 50th, 75th, and 90th percentile observations 

along with the mean (solid marker). 

kg grain ha –1 in 75% of the on-farm comparisons, increasing 

to more than 1,400 kg grain ha –1 in as many as 25% of the 

comparisons. At the same time, 75% of the farmers realized a 

reduction in applied N of more than 45 kg N ha –1 , while as 

many as 25% of the farmers achieved a savings that exceeded 

90 kg N ha –1 . Furthermore, in 75% of the comparisons, the 

increased yield and savings in applied N translated into increases 

in the PFP-N of more than 30 kg of grain per kg of 

applied N. Similar distributions, but with somewhat smaller 

values, were obtained for all three parameters during the aman 

season. Except for a few cases where the use of applied N was 

greater for UDP aman than for FP aman (i.e., ∆Applied N was 

positive), all distributions showed that UDP provided greater 

benefits in both seasons, and there was essentially no risk of 

obtaining a less favorable outcome using UDP versus FP. 

The average UDP yield benefit over the current practice 

was 1,120 kg grain ha –1 (SE = 32.4) during the boro season 

and 890 kg grain ha –1 (SE = 32.5) during the aman. Importantly, 

the savings in applied N amounted to 70 kg N ha –1 (SE 

= 2.4) during the boro season and 35 kg N ha –1 (SE = 9.1) 

during the aman. As impressive as these numbers appear, they 

are certainly realistic and within the range of values cited by 

others for both on-farm and on-station experiments (Daftardar 

et al 1997, Mohanty et al 1999, Pasandaran et al 1999, Savant 

and Stangel 1990). 

Discussion and conclusions 

The current practice of broadcasting urea represents a tremendous 

waste because much of the applied N is lost before it can 

be effectively used by the plant. Greater adoption of UDP by 

rice farmers could improve the use efficiency of applied N 

and successfully reduce losses to the environment. If the numbers 

obtained from the on-farm studies are taken to be representative 

for all of Bangladesh, then the 7 million hectares of 

rice (4 million boro, 3 million aman) grown annually in the 

country could potentially produce about 7 million t more rice, 

and do so with about 0.4 million t less applied N using UDP 

technology. 

The efforts to improve N-use efficiency using UDP technology 

in Bangladesh and other Asian countries have also 

helped to identify situations where the effectiveness of UDP 

may be less than expected and where adoption may also be 

limited (Mohanty et al 1999). For example, UDP would not be 

as effective on soils with high internal drainage. Additionally, 

it is important to recognize that UDP will be most effective 

when using high-yielding varieties and good management practices. 

For flooded rice systems, few practical improvements 

in fertilizer management or products prove to be as cost-effective 

as UDP in increasing fertilizer-use efficiency and addressing 

food production and environmental concerns. As UDP technology 

continues to be improved and adopted in Bangladesh, 

it promises to significantly reduce N losses to the environment 

while also increasing grain yield and improving farmer income. 

References 

Daftardar S, Wagle S, Savant N. 1997. Agronomic performance of 

urea briquettes containing diammonium phosphate in rainfed 

transplanted rice on farmers’ fields. J. Agric. Sci. Cambridge 

128:291-297. 

Mohanty SK, Singh U, Balasubramanian V, Jha KP. 1999. Nitrogen 

deep-placement technologies for productivity, profitability, and 

environmental quality of rainfed lowland rice systems. Nutr. 

Cycl. Agroecosyst. 53:43-57. 

Pasandaran E, Gultom B, Sri Adiningsih J, Apsari H, Rochayati S. 

1999. Government policy support for technology promotion 

and adoption: a case study of urea tablet technology in Indonesia. 

Nutr. Cycl. Agroecosyst. 53:113-119. 


Savant N, Stangel P. 1990. Deep placement of urea supergranules in 

transplanted rice: principles and practices. Fert. Res. 25:1- 

83. 

Notes 

Authors’ addresses: W.T. Bowen, resident representative, International 

Fertilizer Development Center-Asia Division, Road 

54A-House #2-Apt. #6, Gulshan 2, Dhaka 1212, Bangladesh; 

R.B. Diamond, senior scientist-retired, IFDC; U. Singh, senior 

scientist, Resource Development Division; T.P. Thompson, 

senior scientist, Market Development Division, IFDC, 

P.O. Box 2040, Muscle Shoals, Alabama 35662, USA. 

Acknowledgments: We wish to acknowledge the significant contributions 

made by M.M. Islam, S.A.M. Hossain, and K.M. Elahi 

in carrying out the work that has been presented. Funding for 

the on-farm studies was provided by the International Fund 

for Agricultural Development. 

New fertilizer management to maximize yield 

and minimize environmental effects in rice culture 

Masahiko Saigusa 

In paddy fields, leaching and denitrification of applied fertilizer 

nitrogen result in a low fertilizer-use efficiency. The average 

recovery rate of basally applied rapidly available fertilizer 

(RAF) nitrogen has been clarified to be about 30% in the 

tillage transplanting system using 15 N-labeled fertilizer. These 

losses of fertilizer nitrogen lead to increased nitrous oxide in 

the troposphere, and contribute to both global warming and 

destruction of the stratospheric ozone layer. Compared with 

basally applied nitrogen, topdressed N showed a fairly high 

recovery rate (50–60%). Therefore, multiple split application 

of N fertilizer reduces N losses, but increases operating costs. 

On the other hand, the invention of controlled-availability fertilizer 

(CAF) whose nutrient release rate meets plant nutrient 

demand allowed contact application of fertilizer with plant 

seeds or roots (co-situs application). Consequently, CAF contributes 

to the improvement of fertilizer-use efficiency to a great 

extent, mitigation of environmental loading from fertilizer, and 

the development of innovative fertilizer application methods 

such as a single basal application of total fertilizer, a single 

basal nursery application of nutrients, the application of an 

aimed form of nutrients directly to plant roots, etc., and the 

establishment of innovative farming systems such as no-tillage 

transplanting rice culture, no-tillage direct-seeding rice 

culture, etc. (Shoji and Gandeza 1992). 

The objectives of this paper are to show the concept of 

co-situs application of CAF and to summarize the innovative 

field work that has been done by our research group in the last 

two decades using this technique. 

The concept of co-situs application 

Rapidly available fertilizers such as ammonium sulfate, urea, 

etc., applied at the same site as seeds or in the intensive rooting 

zone often cause salt and/or ammonia injuries to plants 

because of their high solubility. Therefore, farmers generally 

have to adjust the soil intervening between fertilizer and seeds 

or roots, or to apply only small amounts of RAF as a contactplacement 

to avoid these injuries. Phenomena such as nitrification, 

denitrification, immobilization, and fixation of nutri- 

ents in the soil are caused by the intervention of soil between 

nutrients and plant roots as shown in Figure 1. However, the 

invention of controlled-availability fertilizer has made possible 

the contact application (co-situs application) of fertilizer with 

seeds or roots. The term co-situs, which means the existence 

of both fertilizer and seeds or roots at the same site (situs), 

was proposed by Shoji and Gandeza (1992) to distinguish the 

seed-placement or contact-placement of small amounts of RAF. 

Therefore, co-situs application of CAF can directly supply 

fertilizer components to plant roots, and thus increase fertilizer-use 

efficiency to a great extent by minimizing the reaction 

and interaction of fertilizer components with the soil (Fig. 

1). Consequently, it is also highly possible to mitigate the environmental 

loading of N 2 O and NO 3 

– 

caused by fertilizer N 

applied. 

No-tillage direct-seeding culture of rice with CAF 

In Japan, rice cropping has been facing serious problems, including 

overproduction, trade liberalization, the advanced average 

age of farmers, and a small number of successors, and 

thus a reduction in labor and production costs in rice cultivation 

is required. The most desirable rice farming system is direct 

seeding by no-tillage and a single basal fertilization using 

CAF, which can also reduce the growth process of rice seedlings. 

Single basal co-situs application of total fertilizer nitrogen 

with CAF, polyolefin-coated urea of the 100-day type 

(POCU-100), in no-tillage direct-seeding culture was conducted 

in comparison with a conventional culture using ammonium-sulfate 

(AS) for both basal and topdressing. Nitrogen 

recoveries of basal POCU-100, basal-AS, and topdressed AS 

fertilizer were 63.2%, 8.5%, and 41.5%, respectively. Reflecting 

the N-use efficiency, the brown rice yields of the POCU- 

100 plot were 33–55% superior to those in the conventional 

AS plot (Sato et al 1993). 

This no-tillage direct-seeding system is being used basically 

in well-drained paddy soil. However, in Tohoku (northeastern) 

District, the main rice-producing zone in Japan, more 


Conventional application 

Co-situs application 

N 2 

O N 2 

Denitrification 

(paddy soil) 

Rapidly available 

fertilizer (RAF) Plant root 

Controlled 

availability 

fertilizer 

Plant root 

N 2 

H 

+ 

4 

(CAF) 

Paddy 

NO 3 

– 

NH 4 

+ 

Soil 

NO 3 

– 

Upland 

Soil 

NH 4 

+ 

Soil 

NH 4 + , (NH 2 

) CO 

Nitrification 

NO 3 

– 

NO 3 

– 

Leaching 

(upland soil) 

Direct 

absorption 

Fig. 1. Dynamics of applied fertilizer in soil. 

than half of the paddy fields are occupied by poorly drained 

soil. Therefore, the development of no-tillage direct seeding 

of rice on Y-shaped furrows without covering soil using CAF 

was studied in a heavy-textured Hachirogata soil and resulted 

in the same level of brown rice yield as with the conventional 

transplanting system (Tashiro et al 2000). 

Single co-situs application of CAF to a nursery box 

In the last two decades, about one-third of the total area under 

rice cultivation in Japan was converted to other crops such as 

soybean, wheat, and forage crops by political control of rice 

cultivation. Therefore, for farmers, the quality of rice is becoming 

a greater issue than the quantity because of its much 

higher price. At the same time, farmers are forced to reduce 

labor and production costs in rice cultivation. The development 

of a sigmoid type of CAF, which shows a delayed N release 

for a specified number of days, signified a single cositus 

application of CAF to nursery boxes (Sato and Shibuya 

1991). The controlled release of nitrogen from sigmoid 

POCUs-100 held by rice roots, which takes place after transplanting, 

remarkably improved the N-use efficiency of fertilizer 

(Saigusa et al 1996, 2004). 

No-tillage transplanting cultivation 

with co-situs application of CAF to a nursery box 

No-tillage transplanting rice cultivation is expected to maintain 

the grain quality of rice while saving both labor and production 

costs by eliminating plowing and puddling. However, 

in this system, farmers had to apply RAF on the surface of the 

soil as basal and topdressing, and thus fertilizer-use efficiency 

was extremely low (< 10%). Therefore, the technique of growing 

seedlings by the single co-situs application of CAF to a 

nursery box was introduced in this system and both the recovery 

of nitrogen and brown rice yield were examined in three 

different types of paddy soils in comparison with those in the 

conventional transplanting system (Saigusa et al 1996, 2004). 

The recoveries of POCUs-100 in the no-tillage system, basal- 

AS, and the 1st and 2nd topdressed AS in the conventional 

tillage system were 61.4–82.7%, 31.1–43.1%, 40.0–76%, and 

43–82.7%, respectively. The recoveries of fertilizer in each 

treatment seemed to be in the following order: light clay alluvial 

soil > clay loam Andisol > sandy loam alluvial soil, and 

also basal POCUs-100 > 2nd topdressed AS > 1st topdressed 

AS > basal AS. These results indicate that the co-situs application 

of sigmoid CAF in a nursery box for the no-tillage transplanting 

system increases the N-use efficiency of fertilizer compared 

with AS in the conventional transplanting system, and 

thus could increase both the growth and yield of rice plants in 

each soil and reduce environmental degradation caused by fertilizer 

N (Saigusa et al 2004). 

Supply of the aimed form of nitrogen 

using co-situs application of CAF 

In paddy soil, nitrogen applied as nitrate is much more unstable 

than ammonium because of denitrification caused by its 

reductive condition (Fig. 1). Therefore, it is almost impossible 

to supply the nitrate form of nitrogen for rice plants using RAF. 

On the other hand, the single co-situs application of CAF may 

directly supply the fertilizer component to plant roots (Fig. 1). 

Therefore, the supply of the nitrate form of nitrogen to rice 

plants was examined by the co-situs application of polyolefincoated 

calcium nitrate (POC-CN) (Saigusa et al 2001). 

As Figure 2 shows, the recoveries of calcium nitrate (CN) 

at the harvesting stage were only about 2% in both clay loam 

Andisol and light clay alluvial soil because of denitrification, 

whereas those of POC-CN were 21–26%, which were almost 

the same as those of AS (22–24%). These results may show 

that rice roots can absorb nitrate-N from POC-CN particles 

because they can come into contact with the CAF without any 


Recovery (%) 

80 

Clay loam Andisol 

60 

40 

20 

80 

60 

40 

20 

Light clay alluvial soil 

0 

0 

CN AS POC-CN POCU 

CN AS POC-CN POCU 

Fig. 2. Recoveries of different forms of nitrogen fertilizer by the rice plant. CN = calcium nitrate, 

AS = ammonium sulfate, POC-CN = polyolefin-coated calcium nitrate, POCU = polyolefin-coated 

urea. 

burning. The highest recoveries were obtained in POCU (49– 

50%). Reflecting the recovery of each fertilizer, the grain yields 

of rice were in the following order: POCU > AS = POC-CN > 

CN. These results imply that the supply of the aimed form of 

nutrient directly from fertilizer to plant roots is highly feasible 

through the co-situs application of CAF. Similar results were 

obtained in supplying NH 4 + -N in upland soil, Fe 2+ in alkaline 

paddy soil, Cu 2+ in high-humus Andisol, and P in Andisol by 

the co-situs application of CAF containing these components. 

From the foregoing, I conclude that co-situs application of CAF 

with seeds or plant roots is a unique technology and is highly 

feasible in developing an innovative farming system. 

References 

Saigusa M, Hossain Md Z, Shibuya K. 2004. No-tillage transplanting 

system of rice with controlled availability fertilizer in a 

nursery box. J. Integ. Field Sci. 1:67-73. 

Saigusa M, Hossain Md Z, Tashiro T, Shibuya K. 1996. Maximizing 

rice yield with controlled availability fertilizer and no-tillage 

culture in controlling environmental degradation. Proceedings 

of the symposium on “Maximizing sustainable rice yields 

through improved soil and environmental management,” Khon 

Kaen, Thailand. p 799-804. 

Saigusa M, Ombodi A, Owashi T, Watanabe H. 2001. Supply of 

aimed form of nitrogen using controlled availability fertilizer. 

Fertilization in the Third Millennium: Fertilizer, Food Security 

and Environmental Protection, 12th World Fertilizer Congress 

of CIEC. p 315-321. 

Sato T, Shibuya K. 1991. One-time application of total nitrogen fertilizer 

at nursery stage in rice culture. Rep. Tohoku Br. Crop 

Sci. Soc. Jpn. 34:15-16. 

Sato T, Shibuya K, Saigusa M, Abe T. 1993. Single basal application 

of total nitrogen fertilizer with controlled-release coated 

urea on non-tilled rice culture. Jpn. J. Crop Sci. 62:408-413. 

Shoji S, Gandeza AT. 1992. New concept of controlled release fertilization. 

In: Shoji S, Gandeza AT, editors. Controlled release 

fertilizer with polyolefin resin coating. Sendai (Japan): Konno 

Printing Co. p 1-7. 

Tashiro T, Sugawara O, Naganoma H, Chiba K, Saigusa M. 2000. 

Direct seeding of rice on Y-shaped furrow without tillage and 

covering soil in poorly drained paddy soil. Jpn. J. Crop Sci. 

69:547-553. 

Notes 

Author’s address: Field Science Center, Graduate School of Agricultural 

Science, Tohoku University, 232-3, Yomogida, 

Oguchi, Naruko, Tamatukuri, Miyagi 989-6711, Japan, e-mail: 

masa@bios.tohoku.ac.jp. 

Does anaerobic decomposition of crop residues impair soil 

nitrogen cycling and yield trends in lowland rice 

D.C. Olk, K.G. Cassman, M.M. Anders, K. Schmidt-Rohr, and J.-D. Mao 

Soil organic matter (SOM) plays a central role in the storage 

and release of essential soil nutrients into plant-available forms. 

Its behavior in nutrient cycling is thought to be affected by its 

chemical nature, although evidence is sparse. Recent work in 

irrigated lowland rice (Oryza sativa L.) soils in the Philippines 

suggests that the chemical nature of SOM depends in 

part on whether its main parent material, crop residues, decomposes 

anaerobically or aerobically, raising the question 

whether anaerobic decomposition contributed to a long-term 

yield decline. In addition, scattered reports from lowland rice 

in other regions and from other agrosystems in which crop 

residues also decompose anaerobically indicate recurring nutrient 

deficiencies, which can also reduce yield. The specific 

nutrients that become deficient are all prone to chemical binding 

with SOM. Hence, the question arises whether anaerobic 

decomposition of crop residues inherently alters the formation 

of new SOM and its chemical composition, which results 

in greater binding of new SOM with nutrients, inhibiting their 

release to available forms. This report will summarize existing 

knowledge on the occurrences of reduced nutrient avail- 


Table 1. Rice crop uptake (kg ha –1 ) of fertilizer N and soil N at three crop growth 

stages for two crop rotations in Stuttgart, Arkansas, 2002. Standard errors are in 

parentheses. 

N pool Rotation Green ring 50% heading Preharvest 

(83 DAE) a (103 DAE) (126 DAE) 

Fertilizer Rice-soybean 59 (6) 52 (3) 54 (4) 

Rice-rice 52 (3) 42 (4) 40 (4) 

Difference 7 10 14 

Soil Rice-soybean 50 (2) 69 (11) 91 (13) 

Rice-rice 44 (8) 44 (8) 67 (5) 

Difference 6 25 24 

a DAE = days after emergence. 

ability in relation to the chemical nature of SOM in flooded 

soils, with emphasis on lowland rice cropping in tropical Asia 

and the southern U.S. rice belt. 

Intensive lowland rice cropping in tropical Asia 

A long-term decline in grain yield has been noted in some field 

trials under double- and triple-cropping of irrigated lowland 

rice (Cassman et al 1995), which involves regular anaerobic 

decomposition of crop residues and long-term submerged conditions. 

The principal cause of the yield decline appears to be 

decreased availability of soil organic nitrogen (N, Dobermann 

et al 2000), especially during late-season crop growth stages; 

the quantity of soil N did not decrease during the years of declining 

yields and crop uptake efficiency of fertilizer N changed 

little. Hence, research efforts have focused on changes in the 

chemical nature of soil organic N. 

An accumulation of phenolic compounds has been associated 

with this intensive cropping (Olk et al 1996). The phenols 

are found in incompletely decomposed lignin residues from 

the woody tissues of crop roots and stubble. Studying the young 

mobile humic acid (MHA) fraction that was extracted from 

both a triple-cropped rice soil and a nearby rice field maintained 

under aerobic conditions at the International Rice Research 

Institute (IRRI), Schmidt-Rohr et al (2004) demonstrated 

that the triple-cropped fraction had an agronomically 

significant excess (55 kg N ha –1 ) of organic N bound covalently 

in aromatic compounds, mainly with phenolic lignin residues 

in an anilide molecule. This form would cycle only slowly under 

flooded conditions. 

In a double-cropped rice experiment conducted at three 

sites in the Philippines, phenols accumulated in the MHA fraction 

despite large differences among the sites in the degree of 

soil drying during two 2-month fallows (Olk et al 1998). Grain 

yields declined at the two sites with either the wettest or the 

driest fallows, and decomposition of crop residues was anaerobic 

at all three sites. Hence, the key feature of continuous rice 

cropping that may promote phenol accumulation and decreased 

availability of soil N may be anaerobic decomposition of crop 

residues, not the total duration of annual flooding. 

Soil phenol accumulation and N mineralization were 

measured in a 4-year study at IRRI that compared anaerobic 

decomposition of crop residues with aerobic decomposition 

for double-cropped rice. In the third wet-season crop, urea fertilizer 

labeled with 15 N was applied at transplanting. Two weeks 

later, the amount of 15 N that became immobilized in the MHA 

was similar for both treatments. During the rest of the season, 

45% of the immobilized 15 N was remineralized in the aerobic 

decomposition treatment, but only 8% was remineralized in 

the anaerobic decomposition treatment (Olk and Cassman 

2002). In-season mineralization of total N from the MHA was 

22 kg N ha –1 less with anaerobic decomposition than with aerobic 

decomposition. Phenols accumulated in the MHA under 

anaerobic decomposition, in tandem with the inhibition of N 

mineralization. The phenol accumulation was also associated 

with inhibited mineralization of soil N across crop rotations 

(rice-rice versus rice-maize, Zea mays L.), humic fraction, N 

fertilizer rate, and year. A crop response to the N inhibition 

could not be demonstrated because N fertilizer was applied in 

synchrony with crop N demand. 

Continuous rice rotation in Arkansas, U.S. 

The conventional crop rotation in eastern Arkansas is rice-soybean 

(Glycine max (L.) Merr.), with one crop per year. Pending 

market changes may promote continuous production of 

rice in fields near water bodies. In a 4-year field trial, however, 

grain yield of a continuous rice rotation was 19% (1,350 

kg ha –1 ) less than for rice yield following soybean (Anders et 

al 2004). Similar to the IRRI yield decline, agronomic symptoms 

attribute this yield gap to a late-season N deficiency. In a 

15 N-microplot study conducted in 2002, crop uptake at midto 

late-season crop growth stages of unlabeled N, presumably 

mineralized from SOM, was less for continuous rice than for 

rice following soybean (Table 1). The difference between crop 

rotations for uptake of labeled fertilizer N was less severe, 

suggesting an inhibition of soil N mineralization in the continuous 

rice rotation. In-season degradation of soil phenols also 

slowed under the continuous rice rotation, as their enrichment 

compared to rice following soybean was greater near harvest 

than at an early growth stage (Fig. 1). This study is ongoing. 


Phenol content (g kg –1 organic C) 

18 

Rice-soybean 

Continuous rice 

18 

12 

Early season (55 DAE) 

12 

Preharvest (126 DAE) 

6 

6 

0 

p-Hydroxyl Cinnamyl 

Vanillyl Syringyl 

Phenol class 

Total 

0 

p-Hydroxyl Cinnamyl 

Vanillyl Syringyl 

Phenol class 

Fig. 1. Soil contents of four phenol classes for rice-soybean and continuous rice rotations at two 

sampling times in Stuttgart, Arkansas, 2002. Units are g phenol kg –1 soil organic carbon (C). DAE is 

days after emergence. Standard errors are indicated above each bar. 

Total 

Rice-wheat rotation in Japan 

A rice sterility problem has occurred in rice-wheat (Triticum 

sp.) rotations in Osato, Saitama, Japan. In a greenhouse study, 

Noguchi et al (1997) associated the sterility with a copper (Cu) 

deficiency that occurred with anaerobic decomposition of crop 

residues from either rice or wheat. Anaerobic soil conditions 

without incorporation of crop residues did not cause rice sterility. 

Rice uptake of manganese, and to a lesser extent of other 

nutrients, was also less with anaerobic decomposition than with 

anaerobic conditions without crop residue incorporation. Copper 

and manganese have a high affinity for binding with humic 

molecules in laboratory conditions (Schnitzer and Hansen 

1970). Rice crop uptake of Cu and manganese was also diminished 

by anaerobic decomposition of crop residues in the IRRI 

field comparison of anaerobic decomposition with aerobic 

decomposition as described above (data not shown). Similarly, 

Cu deficiencies in organic soils are well known; the Cu is 

thought to be bound to SOM by oxygenated functional groups, 

such as phenolic hydroxyls. 

No-tillage in temperate regions 

In no-tillage systems of temperate regions, crop residues decompose 

to a larger extent during the springtime than under 

conventional tillage, which by contrast typically promotes autumn 

decomposition through postharvest tillage. In many temperate 

regions, abundant springtime precipitation saturates the 

soils for extended periods of time. Consequently, decomposition 

of crop residues is more often anaerobic, or partially anoxic, 

with no-tillage than with conventional tillage. In several 

North American studies involving 15 N-enriched fertilizer, crop 

uptake of unlabeled N, presumably mineralized from SOM, 

was significantly less with no-tillage than with conventional 

tillage (Doran and Smith 1987). At some locations, the decrease 

in soil N uptake was associated with a loss of grain 

yield. 

Discussion 

Evidence from multiple agrosystems is consistent with the 

hypothesis that anaerobic decomposition of crop residues promotes 

an enrichment of phenolic lignin residues in newly forming 

SOM, which can bind specific soil nutrients into less available 

forms. This hypothesis would account for the decreased 

in-season N mineralization that was observed in rice soils of 

both the Philippines and Arkansas. Phenolic compounds are 

also capable of binding with Cu, which also becomes deficient 

with anaerobic decomposition. 

Demonstration of strong nutrient binding, at least for N, 

does not itself prove that anaerobic decomposition leads to 

yield loss. Even strongly bound nutrient forms, most notably 

the anilide N found by Schmidt-Rohr et al (2004), must mineralize 

at some point; otherwise, total soil N would increase indefinitely 

with anaerobic decomposition, which is not the case. 

Future research will explore the hypothesis that mineralization 

of strongly bound nutrients is so slow as to become 

unsynchronized with crop nutrient demand, and as the strongly 

bound forms accumulate in soil, the soil nutrient supply also 

becomes increasingly unsynchronized, potentially diminishing 

the yield. Possible mitigation options include timely soil 

aeration, which would promote decomposition of the lignin 

residues and release of the bound nutrients. For example, rice 

fields in Japan are commonly drained during later crop growth 

stages to promote soil N mineralization. 

How relevant is chemical stabilization of soil nutrients 

to yield trends for Asian rice farmers Available data sets do 

not allow comment on whether on-farm yields or productivity 

(ratio of yield to inputs) have declined since lowland rice cropping 

was intensified 30+ years ago. A definitive database would 


include on-farm grain yields since the onset of intensive rice 

cropping for fields in which fertilizer rates and other management 

strategies have remained relatively constant. This database 

does not exist. 

Chemical binding of nutrients by phenolic lignin residues 

following anaerobic decomposition may most likely affect 

yield trends when yield is primarily limited by N, when 

initial yield levels are relatively high, and when N fertilizer is 

not applied in synchrony with crop N demand. At high yields, 

large amounts of crop lignin are incorporated into the soil, 

which may enhance N binding. In other situations where chemical 

binding may not perceptibly affect yield levels, its occurrence 

would nevertheless cause less efficient use of soil N, 

requiring increased N fertilizer inputs to maintain yields. 

References 

Anders MM, Olk D, Harper T, Daniel T, Holzhauer J. 2004. The 

effect of rotation, tillage, and fertility on rice grain yields and 

nutrient flows. In: Proceedings, 26th Southern Conservation 

Tillage Conference for Sustainable Agriculture, 8-9 June 2004. 

Raleigh, N.C. (USA): North Carolina Agricultural Research 

Service. CD format. 

Cassman KG, de Datta SK, Olk DC, Alcantara J, Samson M, 

Descalsota J, Dizon M. 1995. Yield decline and the nitrogen 

economy of long-term experiments on continuous, irrigated 

rice systems in the tropics. In: Lal R, Stewart BA, editors. 

Soil management: experimental basis for sustainability and 

environmental quality. Boca Raton, Fla. (USA): CRC/Lewis 

Publishers. p 181-222. 

Dobermann A, Dawe D, Roetter RP, Cassman KG. 2000. Reversal 

of rice yield decline in a long-term continuous cropping experiment. 

Agron. J. 92:633-643. 

Doran JW, Smith MS. 1987. Organic matter management and utilization 

of soil and fertilizer nutrients. In: Follett RF, Stewart 

JWB, Cole CV, editors. Soil fertility and organic matter as 

critical components of production systems. Madison, Wis. 

(USA): Soil Science Society of America. p 53-72. 

Noguchi A, Hasegawa I, Shinmachi F, Yazaki J. 1997. Possibility of 

copper deficiency and impediment in ripening in rice from 

application of crop residues in submerged soil. In: Ando T, 

Fujita K, Mae T, Matsumoto H, Mori S, Sekiya J, editors. 

Plant nutrition for sustainable food production and environment. 

Proceedings of the XIII International Plant Nutrition 

Colloquium. Dordrecht (Netherlands): Kluwer Academic Publishers. 

p 799-800. 

Olk DC, Cassman KG, Randall EW, Kinchesh, P, Sanger LJ, Anderson 

JM. 1996. Changes in chemical properties of organic 

matter with intensified rice cropping in tropical lowland soil. 

Eur. J. Soil Sci. 47:293-303. 

Olk DC, Cassman KG, Mahieu N, Randall EW. 1998. Conserved 

chemical properties of young humic acid fractions in tropical 

lowland soil under intensive irrigated rice cropping. Eur. J. 

Soil Sci. 49:337-349. 

Olk DC, Cassman KG. 2002. The role of organic matter quality in 

nitrogen cycling and yield trends in intensively cropped paddy 

soils. In: Proceedings, 17th World Congress of Soil Science, 

14-20 August 2002. Bangkok: Thailand Ministry of Agriculture 

and Cooperatives. p 1355(1)-1355(8). 

Schmidt-Rohr K, Mao J-D, Olk DC. 2004. Nitrogen-bonded aromatics 

in soil organic matter and their implications for a yield 

decline in intensive rice cropping. Proc. Nat. Acad. Sci. USA 

101:6351-6354. 

Schnitzer M, Hansen EH. 1970. Organo-metallic interactions in soils. 

8. An evaluation of methods for the determination of stability 

constants of metal-fulvic acid complexes. Soil Sci. 109:333- 

340. 

Notes 

Authors’ addresses: D.C. Olk, USDA-ARS, National Soil Tilth Lab, 

2150 Pammel Drive, Ames, Iowa 50011; K.G. Cassman, Department 

of Agronomy and Horticulture, University of Nebraska, 

Lincoln, Nebraska 68583; M.M. Anders, University 

of Arkansas Rice Research and Extension Center, 2900 Hwy 

130E, Stuttgart, Arkansas 72160; K. Schmidt-Rohr, Department 

of Chemistry, Iowa State University, Ames, Iowa 50011; 

J.-D. Mao, Dept. of Chemistry, Rocky Mountain College, 1511 

Poly Drive, Billings, Montana 59102, USA. 

Influence of the paddy-upland rotation on soil 

physicochemical properties and crop growth 

in the Honam Plain of Korea 

Lee Deog-Bae, Yang Chang-Hyu, Ryu Chul-Hyun, Lee Kyeong-Bo, and Kim Jae-Duk 

Rotation cultivation of rice and upland crops in paddies has 

been adopted for the multiple use of land and has affected the 

nutrient imbalance in the plant and soil. Multiple use of paddies 

has been required to adjust the balance between demand 

and supply of rice in some countries. Crop growth primarily 

depends on soil moisture content, and secondarily on nutrient 

content. The distribution of soil moisture and nutrients greatly 

depends on soil texture, groundwater level, and cropping system. 

Nutrient contents depend on fertilization and the application 

of farming materials. Rotation cultivation affects these 

properties. Therefore, the influence of the paddy-upland rotation 

on soil physicochemical properties, and crop growth and 

productivity, was studied in the Honam Plain area of the Republic 

of Korea. 


Table 1. Change in organic matter content by cropping system (g 

100 g –1 ). 

Cropping system First Second Third Fourth Fifth 

year year year year year 

Continuous rice 2.7 2.3 2.5 2.2 2.6 

Rice-barley 2.8 2.5 2.5 2.6 2.3 

Soybean-barley 2.6 2.2 2.4 2.3 2.3 

Soybean-rice 2.3 2.2 2.6 2.6 2.4 

Soybean 2 y-rice 3.9 2.1 2.5 2.2 2.3 

Soybean 3 y-rice 3.0 2.3 2.1 2.0 2.2 

Table 2. Change in available phosphate content by cropping system 

(mg kg –1 ). 

Cropping system First Second Third Fourth Fifth 

year year year year year 

Continuous rice 188 170 148 137 139 

Rice-barley 105 145 161 134 131 

Soybean-barley 127 175 168 176 274 

Soybean-rice 154 153 145 144 189 

Soybean 2 y-rice 171 117 168 151 206 

Soybean 3 y-rice 157 148 159 171 192 


The experimental field was installed with a sand-gravel drainage 

trench with 2-m intervals, 60-cm width, and 30-cm height 

at 1-m soil depth in 1982. After installation of the drainage 

system, rice was cultivated in 1988. Treatments were continuous 

rice cultivation, rice-barley cropping, soybean-barley cropping, 

soybean-rice rotation, soybean for two years-rice for one 

year rotation, and soybean for three years-rice for one year 

cropping. Applied amounts of N-P 2 O 5 -K 2 O were 11-7-8 kg 

ha –1 in rice, 4-7-68 kg ha –1 in soybean, and 15-10-88 kg ha –1 

in barley. 

Results 

Rice yield increased after the rotation with soybean, and this 

effect increased gradually with two- and three-year continuous 

cultivation of soybean. This resulted from the increased 

number of panicles and number of grains per m 2 by the high 

content of NH 4+ -N in the soil. Meanwhile, rice yield decreased 

by 7% more in the barley-rice cropping system than in continuous 

rice cultivation. 

Soybean yield was higher in the annual rotation than in 

the biennial and three-year terms: 3.92 t ha –1 in annual, 3.56 t 

ha –1 in biennial, and 3.72 t ha –1 in the three-year interval. This 

effect resulted from the changing number of pods and grains 

per plant as time passed. Park et al (1993) reported that soybean 

yield decreased greatly in 4-year continuous cultivation 

in sandy loam paddy and 3-year continuous cultivation in silt 

loam paddy. Motomatsu (1990) also reported that soybean yield 

decreased by 10% in 2-year continuous cultivation and by 20% 

in 3-year continuous cultivation because of the reduction in 

soil N content and root activity and increment in disease and 

pest injury with continuous cropping. 

Barley yield increased by 4–7% in the soybean-barley 

cropping system vis-à-vis the rice-barley cropping system. 

Therefore, this was a positive effect of soybean culture on barley 

yield. Meanwhile, the annual yield of barley decreased 

gradually in both cropping systems as time passed. 

Upland crop cultivation increased soil porosity as time 

passed. The effect was higher in 2-year and 3-year cultivation 

of upland crops than with the annual rotation of rice. The increasing 

soil porosity was in the order of soybean for three 

years-rice rotation, soybean-barley cropping system, soybean 

for two years-rice rotation, and soybean-rice rotation. Iwata et 

al (1977) reported that soil porosity increased annually in the 

paddy-upland rotation with the development of aggregate. 

There was little difference in organic matter content with 

the continuous rice cultivation, rice-barley cropping system, 

and rice-soybean cropping system. But organic matter content 

decreased drastically with soybean cultivation for two years 

and three years (Table 1). Kitada et al (1992) mentioned that 

soil organic matter decreased gradually with easy mineralization 

of organic matter in upland paddy conditions. 

Soil available phosphate content decreased gradually 

with continuous rice cultivation, but increased with rotation 

cultivation of an upland crop in the paddy (Table 2). This result 

was not related to fertilization amount but to phosphate 

dynamics. There was little difference in the amount of phosphate 

fertilization among the crops. Phosphate is released in 

aerobic conditions and fixed in anaerobic conditions. This increment 

of phosphate in soybean cultivation resulted from 

phosphate fixation due to aerobic conditions. Kitada et al 

(1992) also reported that availability of phosphate decreased 

in upland conditions because of fixation to Fe and Al. 

There was no change annually in Ca 2+ of soil with continuous 

rice cultivation, but the content increased after upland 

cultivation. Meanwhile, soybean cultivation caused an increase 

in Ca 2+ content in double cropping and rotation cultivation. 

Exchangeable K + content showed a small decrease in continuous 

rice cultivation, but an increase in rotation cultivation and 

a clearer increase in soybean-barley and soybean two yearsrice 

cultivation rotation. This stable content of Ca 2+ and K + in 

continuous rice cultivation implied that submerged rice cultivation 

had a positive effect on the balance of Ca 2+ and K + in 

the soil. 

The content of ammonium nitrogen was high in the order 

of soybean-rice, soybean two years-rice cultivation, ricebarley 

cropping system, soybean-barley cropping system, and 

continuous rice cultivation. The released amount of inorganic 

nitrogen in the rotational paddy was influenced by the soil texture, 

organic matter content, period of rotation, and cropping 

system (Watanabe 1985, Kogano 1987). Mineralized nitrogen 

was higher in upland culture than in continuous rice culture. 


Conclusions 

There are great differences in field conditions between rice 

and upland crops, mostly anaerobic and aerobic conditions, 

and irrigation of surface water and capillary migration of 

groundwater. These two factors could influence mass balance 

greatly. For a stable crop yield and soil environment in the 

paddy, rice cultivation can play a more positive role than upland 

crops. Therefore, annual rotational cultivation of rice and 

soybean was recommended. 

References 

Ikeda M, Harada I, Tamura K. 1958. Studies on three phases of soils. 

Part 4. Three phases of soil horizons and plant growth. Jpn. 

Soc. Sci. Soil Manure 29(2):84-86. 

Iwata H, Sawada M, Ookami M, Hukada K, Katou T. 1977. Redevelopment 

on method of paddy field land usage in lowland: 

change of soil physico-chemical properties in rotational paddy 

soil based on the ripening of upland soil. Bull. Aichi Natl. 

Agric. Exp. Stn. A9:125-130. 

Kitada K, Kamekawa K, Akiyama Y, Shimoda H, Yamagata M. 1992. 

Estimation of nutrient movement in rotational paddy soil based 

on the ripening of upland soil. Jpn. J. Soil Sci. Plant Nutr. 

63(3):349-351. 

Kogano K. 1987. Technique of stable high yield on paddy rice in 

rotational paddy field. Soil Field (Jpn.) 220:62-69. 

Motomatsu T. 1990. Research plan to increase the land utility in 

Japan. In: Han KH, Jo In Sang, editors. Strategic development 

in future and results of farmland cultivation project. Proceedings 

of symposium. Suwon (Korea): Rural Development 

Administration. p 161-183. 

Park CY, Kang UG, Hwang GS, Jung YT. 1993. Changes of crop 

yields according to cropping systems and fertilizing levels in 

paddy-upland rotation soils. J. Agric. Sci (S&F) 35(1):281- 

288. 

Watanabe K. 1985. Method of fertilization and soil characteristic on 

rotational paddy field in Hokkaido Center Region. Jpn. J. Soil 

Sci. Plant Nutr. 512-514. 

Notes 

Authors’ address: Honam Agricultural Research Institute, National 

Institute of Crop Science, Rural Development Administration, 

381 Songhak Dong Iksan 570-080, Republic of Korea, e-mail: 

leedb419@rda.go.kr. 

A decrease in soil fertility and crop productivity 

by succession of the paddy-upland rotation 

Hirokazu Sumida 

In Japan, rice production has tended to exceed domestic demand 

since around 1970, and the rice production adjustment 

by the government has already continued for more than 30 years. 

In 2004, the area of rice cultivation is restricted to about 60% 

of all paddy fields (i.e., 1.6 million ha). The government has 

recommended several types of alternative paddy field uses with 

subsidiary payments to farmers, especially alternative cropping 

of soybean, wheat, or fodder crops to increase self-sufficiency 

rates of these crops. 

It is known that the yields of soybean and/or wheat decrease 

when these crops are continuously cultivated in the upland 

fields that have been converted from paddy fields for many 

years. However, this injury from continuous cropping can be 

avoided by a paddy-upland (irrigated paddy rice and upland 

crop) rotation and the rotation leads to high yields of both rice 

and upland crops. Therefore, the paddy-upland rotation has 

been encouraged. Several previous studies on the combination 

of upland period and paddy period showed that more than 2–3 

years of paddy period is required for 2–3 years of upland period 

to keep crop productivity and soil nutrient dynamics suitable 

(Takahashi 1983, Hanai 1987, Tsukuda 1990, Kitada et al 

1993). Recently, it has been pointed out that soybean productivity 

decreased at some production sites, especially in the warm 

region of Japan, where the rotation cycles are one year of upland 

for one to several years of paddy. Although there is no 

clear evidence, this decrease in productivity could be attributed 

to a decrease in soil fertility. Therefore, to promote paddyupland 

rotation farming, sustainability of crop production in 

the successive paddy-upland rotation system should be investigated 

from a long-term view. 


The changes in soil fertility, soybean productivity, and paddy 

rice productivity were investigated by long-term field experiments 

under single cropping in a year on gray lowland soil in 

the cool region of Japan (National Agricultural Research Center 

for Tohoku Region, NARCT, Omagari, Akita). 

In experiment A, a field experiment in combination with 

different paddy/upland periods and organic matter application 

has been conducted since 1990. In a short-term upland rotation 

treatment, soybean (Glycine max (L.) Merr.) and paddy 

rice (Oryza sativa L.) were grown with a cycle of 1–2 years of 

upland and 1–3 years of paddy, respectively. A cycle of 3–4 

years of upland and 1–2 years of paddy was adopted in a medium-term 

upland rotation treatment. As a control, continuous 


paddy rice cropping was done in a continuous paddy treatment. 

Rice straw was applied at 0 or 6 t ha –1 in each treatment 

after harvesting every autumn. 

In experiment B, the experimental paddy field was converted 

to an upland field in 1982 and soybean was cultivated 

during 18 years. After that, the field was returned to paddy and 

paddy rice was cultivated for three years. Then, in 2003, it was 

reconverted to an upland field for cultivating soybean. As a 

control, a continuous paddy treatment was prepared and paddy 

rice was cultivated there. Rice straw compost was applied at 0 

or 20 t ha –1 to both treatments before plowing every spring. 

The average of total nitrogen (N) content of the applied rice 

straw compost was 5.5 mg g –1 (fresh-weight basis) for the 20 

years, and it was 3.5 mg g –1 for the last several years. The C/N 

ratio of the rice straw compost was about 20. 

For soil fertility, soil samples were collected from the 

plow layer in 1982 (only in experiment A), 1991, and 1999- 

2003. All the soil samples were air-dried and sieved (< 2 mm). 

Total carbon (C) and N content of the soil were determined by 

using the CN-Analyzer (Yanaco MT-700). The available N in 

the soil was measured by the soil incubation method (under 

submerged conditions at 30 ºC for 4 wk). 


Changes in soil fertility 

In the continuous paddy fields of experiments A and B, the 

total C, total N, and available N of soil were maintained at the 

initial levels without organic matter application, and were increased 

by organic matter application. 

When the successive paddy-upland rotations were continued 

for 10 years or more, whether the rotation was for the 

medium term or for the short term, a decrease in total C of the 

soil was observed and it wasn’t made up for by rice straw application 

at 6 t ha –1 . For N fertility, although the total N and 

available N of the soil were maintained roughly at their initial 

levels by rice straw application for the short-term upland rotation, 

the decrease in total N and available N of the soil wasn’t 

fully mitigated by the rice straw application in the mediumterm 

upland rotation (Fig. 1). 

Cultivating soybean for 18 years in the upland field converted 

from paddy field (long-term upland conversion) decreased 

the available N of the soil more than the paddy-upland 

rotation. The available N of the soil in the long-term upland 

conversion, even with application of rice straw compost at 20 

t ha –1 , was markedly lower than that in the continuous paddy 

field where no organic matter was applied. By returning to a 

paddy field from long-term upland conversion and keeping it 

as a paddy field for three years, the available soil N recovered 

up to about 25% of the decrease caused by the 18-year longterm 

upland conversion. 

Changes in crop production related to soil fertility 

As Figure 2 shows, soybean yield in the successive paddyupland 

rotation and long-term upland conversion decreased 

by about 5–25% compared with that in the field where soy- 

Available soil nitrogen (index) 

140 

120 

100 

80 

60 

40 

127 

Continuous P/+RS 

100 

–RS 

92 

75 

13 years after starting 

(after SB-SB-SB (or R)-SB) 

71 

Short-term U/+RS 

–RS 

Medium-term U/+RS 

Starting year 

9 years after starting 

(after rice-rice) 

Fig. 1. Changes in available soil nitrogen in the successive paddyupland 

rotation system. Available soil nitrogen (index): relative % 

of that in the continuous paddy field without rice straw application 

in the same year. P = paddy, U = upland, RS = rice straw, SB = 

soybean. 

bean was planted after a sufficient paddy period (the yield trial 

by the Soybean Breeding Lab, NARCT). In contrast, the yield 

of paddy rice in both cases (paddy-upland rotation and longterm 

upland conversion) was higher than that in the continuous 

paddy field. 

Rice straw application in the successive paddy-upland 

rotation produced no increase in soybean yield, whereas the 

rice straw compost application in long-term upland conversion 

increased soybean yield slightly. According to the experiment 

on the topdressing of N fertilizer, the N topdressing at 70 

kg ha –1 scarcely improved the yield of soybean planted in the 

upland field where the decrease in soil N fertility occurred. 

These results implied that, in response to a decrease in 

soil N fertility, soybean yield in the successive paddy-upland 

rotation system decreased. In contrast, in spite of the considerable 

decrease in soil N fertility, the yield of paddy rice in the 

paddy field just returned from an upland field became higher 

than that in the continuous paddy field. 

References 

Hanai Y. 1987. Cropping systems in paddy field for multiple use. 

Res. J. Food Agric. 10(9):28-32. (In Japanese.) 

Kitada K, Shimoda H, Kamekawa K, Akiyama Y. 1993. Changes in 

soil nutrients affected by rotation of upland and paddy crop in 

gray lowland soil and search for the most suitable term of 

rotation. Jpn. J. Soil Sci. Plant Nutr. 64:154-160. (In Japanese 

with English summary.) 

72 

–RS 


Yield index of soybean and paddy rice 

150 

125 

Short-term upland (Exp. A) 

Long-term upland (Exp. B) 

Medium-term upland (Exp. A) 

(gray: soybean, white: rice 

100 

75 

50 

Exp. A: av with and without rice straw application 

Exp. B: without rice straw compost application 

1990 

(1) 

[9] 

1992 

(3) 

[11] 

1991 

(2) 

[10] 

1993 

(4) 

[12] 

1994 

(5) 

[13] 

1996 

(7) 

[15] 

1995 

(6) 

[14] 

1997 

(8) 

[16] 

1998 

(9) 

[17] 

2000 

(11) 

[19] 

1999 

(10) 

[18] 

2001 

(12) 

[20] 

2002 

(13) 

[21] 

2003 

(14) 

[22] 

Year 

(years after starting Exp. A) [years after starting Exp. B] 

Fig. 2. Crop productivity in the successive paddy-upland rotation system and long-term 

upland conversion system. Control of soybean: soybean planted after sufficient paddy 

period (the yield trial by the soybean breeding lab, NARCT). Control of paddy rice: paddy 

rice planted in continuous paddy field (experiments A and B). 

Takahashi H. 1983. Land use by group farming and paddy-upland 

rotation system. Jpn. J. Farm Manage. 20(3):14-22. (In Japanese.) 

Tsukuda K. 1989. Suitable paddy-upland rotation cycle: introduction 

of experiments in Kanto-Tokai district. Agric. Hortic. 

65:385-388. (In Japanese.) 

Notes 

Author’s address: Headquarters, National Agriculture and Bio-oriented 

Research Organization, 3-1-1 Kannondai, Tsukuba, 

Ibaraki 3058517, Japan, e-mail: mizuki@affrc.go.jp. 

Acknowledgments: The author wishes to thank his colleagues, Dr. 

N. Kato and Mr. M. Nishida, for their helpful discussion. 

Promising technologies for reducing 

cadmium contamination in rice 

Satoru Ishikawa 

Cadmium (Cd) is principally dispersed in natural and agricultural 

environments through human activities. Japanese arable 

lands, especially paddy fields, are contaminated to some extent 

by Cd through irrigation with river water originating from 

mines or through emissions from metal-smelters. Recently, the 

joint FAO/WHO Expert Committee on Food Additives has 

established a Provisional Tolerable Weekly Intake (PTWI) of 

Cd at 7 µg kg –1 body weight per week. Based on this risk assessment 

of Cd, the Codex Alimentarius Commission will propose 

new criteria of Cd concentration in foods, especially in 

staple crops. The standard of Cd for rice has been discussed as 

being from 0.2 to 0.4 mg kg –1 the weight of polished rice. 

According to a survey of Cd contamination in rice by the Ministry 

of Agriculture, Forestry, and Fisheries (MAFF) of Japan, 

3.3% or 0.3% of rice cultivated in Japan exceeded a limit of 

0.2 mg kg –1 or 0.4 mg kg –1 , respectively. Although our diet is 

below the PTWI of Cd (4 µg kg –1 ), rice is the largest source of 

dietary intake of Cd for the Japanese people, so it is urgent to 

reduce the Cd concentration of rice. In this paper, current ameliorating 

techniques being applied in Cd-polluted paddy fields 

and promising techniques for reducing Cd contamination in 

rice are discussed. 


Current techniques for reducing Cd contamination in rice 

Soil dressing 

In 1971, the Japanese government enforced the law that paddy 

fields that had produced brown rice beyond 1 mg kg –1 were 

subject to preventive measures against soil pollution such as 

soil dressing. There are two types of soil dressing: (1) removing 

the Cd-polluted soil layer and replacing it with nonpolluted 

soils and (2) covering Cd-polluted soils with nonpolluted ones. 

According to a survey of rice Cd concentration after soil dressing 

by which the Cd-polluted soil layer with more than 25 cm 

was replaced, it did not exceed 0.03 mg kg –1 over more than 

20 years (Kuwana et al 2003). Thus, soil dressing is a permanent 

technique for reducing Cd contamination in rice, but an 

alternative approach is needed because of its enormous cost 

($300,000–500,000 per ha). 

Water management 

The solubility of Cd depends on the redox potential in the soil. 

Bioavailable Cd in the soil decreases to a great extent under 

submerged conditions because of the formation of less-soluble 

cadmium sulfide (CdS). Pot and field experiments conducted 

by many researchers showed that keeping a submerged condition 

for 3 weeks before and after heading time is a very useful 

method for reducing the Cd concentration of rice. Even with 

such conditions, however, rice grains are always exposed to 

the threat of Cd contamination because high levels of Cd still 

remain in the soil. Besides, the effect of submergence varies 

among years depending on weather conditions (e.g., rainfall). 

Application of soil amendments 

The application of alkaline amendments is one of the practicable 

methods for reducing the Cd concentration of rice. The 

combined use of calcium-silicate and fused magnesium phosphate 

for a basal dressing showed positive effects on suppressing 

Cd uptake by paddy rice (Takijima and Katsumi 1973). 

However, no effects or negative effects of alkaline amendments 

on Cd uptake in rice were also reported. Currently, the combination 

of water management and application of soil amendments 

for suppressing Cd contamination of rice is widely practiced 

in Cd-polluted paddy fields in Japan. 

Promising techniques for reducing Cd contamination in rice 

Phytoremediation 

Phytoremediation, which is the use of plant systems to 

remediate contaminated environments, has drawn attention as 

an environmentally friendly and cost-effective technique. The 

most important point for developing a successful 

phytoremediation (phytoextraction) technology in Cd-polluted 

soils is to select promising cleanup plants. Hyperaccumulator 

plants such as Thlaspi caerulescens (Brown et al 1994) and 

Brassica juncea (Salt et al 1995) have been exploited as ideal 

plants for phytoremediation because they possess several preferable 

characteristics for phytoremediation: the ability to take 

up and translocate an exceedingly large amount of Cd to their 

Table 1. Estimation of Cd removal from polluted 

soils using a high Cd-accumulating rice variety, 

Milyang 23. 

Item 

Amount 

Cd concentration of polluted soil 0.7 mg kg –1 

Average of Cd concentration 0.3 mg kg –1 

of nonpolluted soils 

Quantity of Cd to be removed per ha 400 g ha –1 

Cd concentration of shoots 15 mg kg –1 

Dry matter weight 800 kg ha –1 

Cd uptake by shoots 120 g ha –1 

Years for phytoremediation to 3–4 years 

be reached in nonpolluted soils 

shoots and hypertolerance of Cd. However, it is questionable 

whether hyperaccumulator plants are applicable for 

phytoremediation in Cd-polluted paddy fields in Japan. The 

primary reason is that appropriate agricultural practices such 

as planting, fertilization, water management, and mechanical 

harvesting are not well established for hyperaccumulator plants. 

In addition, the main targeted areas in Japan for 

phytoremediation have slight and moderate contamination of 

paddy fields, which might potentially produce paddy rice or 

upland crops contaminated with Cd. Under such lands, it is 

unknown whether hyperaccumulator plants can have an extraordinary 

ability for Cd uptake. Our research group found 

that some indica-japonica cross varieties cultivated under upland 

conditions accumulated a substantially high concentration 

of Cd in grains and shoots. When the shoot Cd concentration 

was compared between several rice varieties and B. juncea 

grown in two types of Cd-polluted soils, its concentration was 

2 to 4 times higher in rice varieties than in B. juncea. Moreover, 

Cd uptake in the shoots was approximately 6 times higher 

in rice varieties than in B. juncea. A remarkable decrease in 

Cd content in the rhizosphere was found after harvesting rice. 

There are many merits for using rice as a cleanup plant, in 

addition to its superior ability of Cd uptake. First, the cultivation 

technique for rice is well established. Second, farmers 

can use their own machines to plant or harvest high Cd-accumulating 

rice varieties without any new investment. Third, they 

can plant a popular commercial variety immediately after ricebased 

phytoremediation. 

Based on the results of a pot experiment, estimation of 

Cd removal from a low-level polluted soil (0.7 mg kg –1 of 0.1 

M HCl extractable Cd in the soil) using Milyang 23, which is 

one of the high Cd-accumulating rice varieties and also a highyielding 

variety, is shown in Table 1. If the shoot biomass of 

Milyang 23 is estimated at 0.8 t ha –1 , the Cd concentration of 

the soil could be reached at close to 0.3 mg kg –1 , which is the 

average Cd concentration of nonpolluted soils, by successive 

planting for 3 to 4 years. A high removal efficiency of Cd by 

Milyang 23 was also confirmed by a field experiment, in which 

approximately 190 g ha –1 of Cd was removed from a paddy 

field polluted with 2.5 mg Cd kg –1 soil. These results suggest 

that high Cd-accumulating rice is an ideal plant for 


Grain Cd concentration (mg kg –1 ) in upland conditions 

4.5 

4.0 

SL217 

3.5 

3.0 

2.5 

2.0 

1.5 

1.0 

SL218 

SL204 

SL205 

SL219 

SL225 

Kasalath 

SL238 SL212 

SL233 

SL202 SL230 SL232 

SL206 

SL239 SL235 

SL220 SL231SL234 

SL221 

SL214 

SL228 

SL213Koshihikari 

SL226 SL227 

SL222 

SL229 

SL203 

SL209 

SL224 

SL211 

SL236 

SL223 

SL207 

SL208 

SL215 

Y = 7.62x + 1.07 

R 2 = 0.129 

0.5 

0.0 

0.00 0.05 0.10 0.15 0.20 

Grain Cd concentration (mg kg –1 ) in paddy conditions 

Fig. 1. Cd concentration of grains in 39 chromosomal substitution lines carrying Kasalath (an 

indica rice variety) chromosomal segments in a Koshihikari (japonica) background. These 

lines were grown on Cd-polluted soils (1.8 mg Cd kg –1 soil) under upland and continuous 

paddy conditions. The concentration of Cd in grains was analyzed by inductively coupled plasmamass 

spectrometry after digesting with an acid mixture. 

phytoremediation of Cd-polluted paddy fields in Japan. For a 

successful decontamination by phytoremediation, it is imperative 

to establish postharvest technology for handling cleanup 

plants, which contain a large amount of Cd. One way is to 

incinerate the cleanup plants, turning them into ash, then recover 

the Cd from the ash. The other way is to use the plant 

biomass for a bio-energy such as bio-methanol after recovering 

the Cd. The National Institute for Agro-Environmental 

Sciences (NIAES) has organized a research consortium for a 

rice-based phytoremediation project. 

Introduction of low Cd-accumulating rice varieties 

Grain Cd concentration varies greatly among rice varieties 

(Arao and Ae 2003). This suggests the possibility of breeding 

varieties with less grain Cd concentration. In durum wheat, 

Clarke et al (1997) reported that the low grain Cd trait is highly 

heritable and is controlled by a single dominant gene. Recently, 

their work suggested that the Cd uptake locus resides on chromosome 

5B. A DNA marker, which was linked in repulsion 

with a low-Cd allele, has been developed for an efficient wheat 

breeding program. Although the genetic analysis of rice is much 

more developed than that of other crops, there is no useful 

information on the low-Cd trait of rice. We tried to identify the 

quantitative trait loci (QTLs) controlling the Cd concentration 

of rice grain. Substitution lines (SLs) carrying Kasalath (an 

indica rice variety) chromosomal segments in a Koshihikari 

(japonica) background were used to identify the QTLs. A se- 

ries of SLs has covered the entire genome with overlapping 

introgressed segments of each line, and it is a powerful tool to 

simply detect the QTLs. The parents and 39 SLs were grown 

on Cd-polluted soils (1.8 mg Cd kg –1 soil) under upland or 

continuous paddy conditions. Figure 1 shows the Cd concentration 

of their grain. The average grain Cd concentration was 

20 times higher in upland conditions (1.67 mg kg –1 weight of 

brown rice) than in paddy conditions (0.08 mg kg –1 ). The grain 

Cd concentration in Koshihikari was significantly lower than 

that of Kasalath in both conditions. Although there is no correlation 

between the grain Cd concentration in rice grown in 

paddy conditions and rice in upland conditions, several SLs 

(SL 207, SL 224, and SL 223) showed significantly lower grain 

Cd concentration than Koshihikari under both conditions. 

Based on graphical genotypes of SLs, the putative QTLs controlling 

low grain Cd were detected on chromosomes 3 and 8. 

As the detected SLs represent more than 90% of the Koshihikari 

genetic background, the development of a new Koshihikari 

with low grain Cd should be feasible in the near future. 

References 

Arao T, Ae N. 2003. Genotypic variations in cadmium levels of rice 

grain. Soil Sci. Plant Nutr. 49:473-479. 

Brown S, Chaney R, Angle J, Baker AJM. 1994. Phytoremediation 

potential of Thlaspi caerulescens and bladder campion for 

zinc- and cadmium-contaminated soil. J. Environ. Qual. 

23:1151-1157. 


Clarke JM, Leisle D, Kopytko GL. 1997. Inheritance of cadmium 

concentration in five durum wheat crosses. Crop Sci. 37:1722- 

1726. 

Kuwana T, Aoyama Y, Tsudaka H, Yoshikura J, Adachi I. 2003. A 

survey of paddy field past to 20 years of soil dressing. Abstr. 

Annu. Meet. Jpn. Soc. Soil Sci. Plant Nutr. 41:288. (In Japanese.) 

Salt DE, Prince RC, Pickering IJ, Raskin I. 1995. Mechanisms of 

cadmium mobility and accumulation in Indian mustard. Plant 

Physiol. 109:1427-1433. 

Takijima Y, Katsumi F. 1973. Cadmium contamination of soils and 

rice plants caused by zinc mining. IV. Use of soil amendment 

materials for the control of Cd uptake by plants. Soil Sci. Plant 

Nutr. 19:235-244. 

Notes 

N uptake inhibition of the rice plant 

in flooded soils receiving wheat straw 

Fukuyo Tanaka 

Author’s address: National Institute for Agro-Environmental Sciences, 

3-1-3 Kannondai, Tsukuba, 305-8604, Japan, e-mail: 

isatoru@niaes.affrc.go.jp. 

The application of fresh plant residue, such as wheat straw, 

often inhibits the growth and N uptake of paddy rice (Oryza 

sativa L.) plants in the establishment phase. This inhibition is 

generally attributed to a deficiency of available N in soils with 

straw added because of considerable N assimilation by the soil 

microbes. This phenomenon is called biological N immobilization. 

In our previous study, however, we reported some cases 

in which much mineral N existed in paddy soils with added 

wheat straw, even though the growth of rice plants was delayed 

(Watanabe et al 1988). Here, N uptake activities of both 

rice plants and soil microbes were measured with 15 N tracing 

to validate that the widely accepted biological N immobilization 

theory was precisely the cause of rice growth inhibition 

by fresh wheat-straw application. Furthermore, the inhibitory 

substances were identified and their concentration determined 

in soil solutions under various conditions. 

Validation of the biological N immobilization theory 

To analyze the behavior of soil mineral N that existed at a 

given time, 15 N tracer experiments were conducted (Tanaka 

and Nishida 1996). Rice seedlings were transplanted into a 

pot with flooded soil, to which was applied wheat straw or no 

wheat straw. At 17 days after transplanting (DAT), tracer 15 N, 

15 NH 4 Cl solution, was added to the soil using a syringe. The 

soil and rice plants were sampled at 24 h after application of 

the tracer to avoid remineralization of the immobilized tracer 

15 N. The plants were divided into tops and roots. The N content 

and 15 N abundance were analyzed on plant and soil 

samples. The behavior of mineral N in the soil upon the application 

of the tracer can be estimated by using the ratio of distribution 

of 15 N in plants and soil. 

Figure 1A shows that the uptake of 15 N of rice in a plot 

receiving straw was about half of that in the control plot, even 

though the mineral 15 N was larger. In this experiment, the total 

weights of rice plants, N uptake, and N content in the straw 

plot were smaller than those in the control. A decrease in 15 N 

Distribution of tracer 15 N at 24 h after application 

Without straw 

With straw 

0 20 40 60 80 100 

Distribution (%) 

Nontracer N uptake of rice plants 

Without straw 

With straw 

Root Top Mineral N in soil 

0 10 20 30 40 50 

Nontracer N uptake (mg pot –1 ) 

Tracer and nontracer mineral N in soil 

Without straw 

With straw 

Root 

Top 

0 20 40 60 80 

Mineral N soil (mg pot –1 ) 

Nontracer 

Tracer 

Fig. 1. Distribution of tracer and nontracer N in the rice plant and 

soil. Rice plants and soils were sampled at 17 days after transplanting. 

Tracer N represents the behavior of mineral N at the 

application of tracer (A), nontracer N represents that which was 

derived from basal fertilizer and soil (B), tracer and nontracer mineral 

N in soil (C). 

A 

B 

C 


uptake per unit of dry matter was also found. These results 

indicate that the decrease in N uptake by rice plants was not 

associated with biological N immobilization but with the inhibition 

in N uptake. Furthermore, the tracer content of tops decreased 

by 51%, which corresponded to a decrease of only 

13% for roots. This suggests that the translocation of 15 N from 

roots to tops was also inhibited. 

However, nontracer mineral N (basal fertilizer and soil 

N) in the plot with straw applied was 83% of that in the control 

plot (Fig. 1C). Considering that nontracer N uptake by 

plants in the plot with straw application was 52% of that in the 

control (Fig. 1B), only a 17% decrease in mineral N in the plot 

with straw applied was not enough to explain the inhibition. 

This finding indicates that biological N immobilization was 

stimulated during the initial days and this might affect rice 

growth to some extent. 

Regarding the effects of 15 N tracer experiments on the 

behavior of N in flooded soil with wheat straw (Tanaka 2001), 

biological N immobilization was stimulated for the initial few 

DAT, although growth inhibition appeared during 10 to 20 DAT. 

In addition to this time mismatch between N immobilization 

and growth inhibition, the amount of mineral N in soil in the 

plot with straw applied was often larger than that in the control. 

Biological N immobilization in the presence of wheat 

straw may inhibit rice growth during the initial few DAT, but it 

may not be a main cause for growth inhibition. Presumably, 

inhibitors in the soil might cause this inhibition. 

Rice-growth-inhibitors formed in paddy soil 

It was suggested that N uptake inhibition was one of the main 

causes of a delay in growth of rice plants in the presence of 

straw in the early stage. Though various substances had been 

reported as growth inhibitors of rice plants, conclusive ones 

have not been found yet. To find growth inhibitors in flooded 

soils, some experiments were conducted. 

Soil solutions were collected from flooded soil receiving 

wheat straw (Tanaka et al 1990). The solutions were divided 

into acidic or basic fractions using diethyl ether extraction. 

As the acidic fraction showed an inhibitory effect on rice 

root elongation, the acids in the fraction were identified one 

by one with gas chromatograph-mass spectrometry and highperformance 

liquid chromatography. Aromatic acids such as 

benzoic acid (BA), phenylacetic acid (PA), 2-phenylpropanoic 

acid (2PPA), 3-phenylpropanoic acid (3PPA), 4-phenylbutyric 

acid (4PBA), and salicylic acid (SA) were detected (Tanaka et 

al 1990, Tanaka 2002). This was the first report that 2PPA and 

3PPA were detected from paddy field soils as a free form. 

The inhibitory activities of these aromatic acids on rice 

root elongation were assayed. Figure 2 shows that the aromatic 

acids have inhibitory effects on root elongation. Especially, 

2PPA showed the highest toxicity, which inhibited rice root 

elongation at a concentration of less than 0.5 µM. PA showed 

severe toxicity at a concentration higher than 40 µM. 3PPA 

Root elongation (%) 

120 

100 

80 

40 

2-phenylpropanoic acid 

20 

Phenylacetic acid 

3-phenylpropanoic acid 

Benzoic acid 

0 

0.1 1 10 100 1,000 

Concentration (mM) 

Fig. 2. Inhibition curve of the effect of aromatic acids on rice root 

elongation. Root elongation was compared with elongation of the 

control (100%). Twenty seedlings (root length was 5–7 mm) were 

used for each treatment. 

also inhibited root elongation at 30.7% of that in the control at 

40 µM. 

Inhibition of N uptake by the rice plant and N transport 

from roots to tops were also examined by hydroponics using 

tracer 15 N (Tanaka and Nishida 1998). The effects of BA (100, 

1,000 µM), 2PPA (1, 10 µM), and 3PPA (100, 1,000 µM) on 

N uptake activity were evaluated. Activity was defined as the 

ratio of 15 N uptake to plant N present before the treatments. 

All the treatments showed inhibition of 15 N uptake. The 

partitioning of 15 N to shoots, indicating the transport of 15 N 

from roots to shoots, was also inhibited by BA and 3PPA at 

1,000 µM. Though N transport inhibition by aromatic acids 

was reported for the first time in this paper, the toxic level was 

higher than that in soil solutions of paddy fields. It was rather 

unrealistic to consider that acids were the causative agents for 

the N transport inhibition caused by straw application. Further 

study on N transport inhibition in association with aromatic 

acid and other aqueous soil components is required. 

Behavior of aromatic acids in paddy fields 

The concentrations of aromatic acids in soil solution from 

paddy fields were determined during cropping periods. In 1989, 

the highest concentration was 59, 1, 11, and 12 µM for BA, 

PA, 2PPA, and 3PPA, respectively. 2PPA was over the critical 

level. Additionally, squeezes of small clusters of wheat straw 

sampled from paddy fields contained a high concentration of 

acids, for example, 122 and 194 µM for BA and 3PPA, respectively. 

This finding indicated that growth inhibition was 

not expressed uniformly over the field but was severe near the 

straws clusters. 

Further studies with field experiments showed that 

growth inhibition of the rice plant by the application of wheat 

straw was noticeable for the first 1–2 years after straw application 

and that it diminished gradually. The accumulation of 

aromatic acids also diminished over time. Some characteristics 

of the soil or field, such as poor water permeability, low- 


air phase, and low temperature, were closely related to the 

increase in the accumulation of aromatic acids. These trends 

agreed with the trends in the appearance of growth inhibition 

of rice plants. Furthermore, it is noteworthy that the aromatic 

acids described here were not decomposition fragments derived 

from lignin of wheat straw but metabolites from soil 

microbes under anaerobic conditions. 

In conclusion, rice growth inhibition in the presence of 

wheat straw cannot be associated with biological N immobilization 

but with N uptake inhibition. Aromatic acids detected 

from soil solution with wheat straw inhibited rice root elongation 

and N uptake. Their concentration and accumulation patterns 

provide an important explanation for rice growth inhibition. 

References 

Tanaka F. 2001. Growth inhibition of rice (Oryza sativa L.) plants in 

flooded soils with added fresh wheat straw: the delicate interaction 

between rice plant and soil microorganisms. 

Nougyougijutu 56:506-510. (In Japanese.) 

Tanaka F. 2002. Formation of aromatic acids and growth inhibition 

of rice (Oryza sativa L.) plants in flooded soils with wheat 

straw added. Bull. Natl. Agric. Res. Cent. Kyushu Okinawa 

Reg. 40:33-78. (In Japanese.) 

Tanaka F, Nishida M. 1996. Inhibition of nitrogen uptake by rice 

after wheat straw application determined by tracer NH 4 + - 15 N. 

Soil Sci. Plant Nutr. 42:587-591. 

Tanaka F, Nishida M. 1998. Inhibitory effects of aromatic acids on 

nitrogen uptake and transport in rice (Oryza sativa L.) plants 

cultured on hydroponics. Soil Sci. Plant Nutr. 44:691-694. 

Tanaka F, Ono S, Hayasaka T. 1990. Identification and evaluation of 

toxicity of rice root elongation inhibitors in flooded soils with 

added wheat straw. Soil Sci. Plant Nutr. 36:97-103. 

Watanabe F, Ono S, Hayasaka T. 1988. Effect of wheat straw application 

on accumulation of organic acid in paddy soil and rice 

growth in early stage. Kyushu Agric. Res. 50:87. (In Japanese.) 

Notes 

Stable isotope ratios of hydrogen and oxygen 

in paddy water affected by evaporation 

Yohei Hamada, Shiho Yabusaki, Norio Tase, and Ichiro Taniyama 

Author’s address: National Agricultural Research Center, 3-1-1 

Kannondai, Tsukuba City, Ibaraki 305-8666, Japan, e-mail: 

fukuyot@affrc.go.jp. 

Evaporation and transpiration from a paddy field are very important 

for regional water balance, and therefore irrigation 

management for rice cultivation, especially in monsoon Asia. 

Stomatal transpiration is accompanied by CO 2 uptake for photosynthesis 

and is essential for plant growth, whereas evaporative 

loss of ponded water would moderate diurnal variations 

of temperature and moisture around the paddy field. These 

two vapor fluxes into the atmosphere are often integrated and 

called evapotranspiration; nevertheless, they are essentially 

different processes. One of the reasons is the difficulty of determining 

these fluxes separately by micrometeorological methods 

such as eddy correlation, which has been widely used to 

directly measure evapotranspiration from various terrestrial 

ecosystems. 

As another approach, isotope hydrological techniques 

could divide evapotranspiration into evaporation and transpiration 

because only evaporation can affect the isotope composition 

of ponded water. For example, Simpson et al (1992) 

successfully calculated the proportion of evaporation occupied 

in total evapotranspiration using isotope analysis, at a 

rice paddy field located in the semiarid region of southeast 

Australia. In their study, the water balance of the paddy field 

was considerably simple: a small amount of rainfall less than 

10% of irrigation, no surface drainage, and a constant ponding 

depth. In the moist Asia region, in contrast, much rainfall oc- 

curs and traditional and complicated irrigation systems are 

widely developed. Such conditions may disturb the formation 

process of the isotope composition of paddy water, but few 

studies have been carried out to investigate the relationship 

between the isotope composition and evaporative loss and to 

confirm the possibility to separate evaporation and transpiration 

from a paddy field in the moist region. In this study, we 

investigated stable isotope ratios of paddy water samples and 

evaluated the effect of evaporation to examine the potential of 

the use of environmental isotope analysis. 

Methods 

Field observations were made in a rectangular lot of a paddy 

field (36°03′N, 140°01′E, 15 m above sea level; 100 m long 

and 54 m wide) located in Mase, Tsukuba, Japan. Rice is usually 

cultivated from May to August in the paddy field; the rest 

of year has no cultivation. We collected paddy water samples— 

irrigation water, drain water, and ponded water—three times 

during cultivation: in mid-May, early in June, and at the beginning 

of August in 2002. The ponded water samples were 

collected at six places distributed in the paddy field to detect 

isotopic change during surface water movement. Stable isotope 

ratios of hydrogen (δD) and oxygen (δ 18 O) of the paddy 

water samples were determined with a dual-inlet-type isotope 


atio mass spectrometer, MAT-252 (Finnigan MAT), equipped 

with an automatic equilibration system at the University of 

Tsukuba. The precisions of the isotope measurement are ±1‰ 

and ±0.1‰ for δD and δ 18 O, respectively. Components of water 

balance of the paddy field—the amount of rainfall, irrigation, 

evapotranspiration, drainage, and infiltration, and the depth of 

ponded water—were also measured and recorded, as well as 

the parameters related to plant growth such as leaf area index 

(LAI). 

d 18 O ( ) 

2 

0 

–2 

15 May 

7 June 

1-3 August 

Flow direction 

Results 

–4 

The change of δ 18 O along the flow direction of paddy water is 

shown in Figure 1. Ponded water samples collected at the six 

places are arranged consistent with the water movement in the 

paddy field (PW #1 to #6). As the paddy water flowed, the 

value of δ 18 O gradually increased. In May and June, the differences 

between irrigation water (IR) and drain water (DR) 

reached 6.1‰ and 5.3‰, respectively. In August, however, 

the increase in δ 18 O was little and the difference between IR 

and DR was only 0.6‰. The pattern of δD change along the 

flow direction was similar to that of δ 18 O. 

Figure 2 is a δ 18 O vs δD plot for the paddy water samples. 

A local meteoric water line (LMWL), which was determined 

from a long-term event-basis water collection by Yabusaki et 

al (2003) and indicates that the isotope composition of rainwater 

around the study area will be plotted along the line, is 

also shown. The paddy water samples were plotted almost linearly 

for each month. The result of linear regression analysis 

suggested good correlation between δ 18 O and δD (the correlation 

coefficients were larger than 0.98), and the slopes of the 

regression lines were 4.3, 4.3, and 5.8 for May, June, and August, 

respectively, showing smaller values than the slope of 

LMWL (7.4). 

–6 

–8 

IR 

PW #1 

PW #2 

PW #3 

Fig. 1. Change in δ 18 O of paddy water samples along the flow direction. 

IR = irrigation water, DR = drain water, PW = ponded 

water, and #n = the order of PW samples arranged along the water 

movement. 

PW #4 

PW #5 

PW #6 

d 18 O ( ) 

–8 –6 –4 –2 0 

LMWL (dD = 7.4 d 18 O + 8.5) 

–20 

DR 

Discussion 

The results of the stable isotope analysis illustrated in Figures 

1 and 2 clearly indicate that the isotope composition of paddy 

water was intensively affected by evaporation, especially in 

May and June. When water vaporizes, water molecules that 

consist of lighter isotopes (H and 16 O) tend to become vapor 

more than those including heavier isotopes (D and/or 18 O). 

This causes isotope fractionation between liquid and vapor 

phases; therefore, ponded water became more and more enriched 

isotopically as the water moved downstream under the 

influence of evaporation. In addition, the relationship between 

δ 18 O and δD of the water affected by evaporation is usually 

different from that of the water not affected. According to pioneering 

work by Craig (1961), most meteoric water samples 

collected all over the world were plotted linearly in the δ 18 O 

vs δD plot (the global meteoric water line, GMWL) and the 

slope of GMWL was 8. But the samples collected from lakes 

and rivers in East Africa, where isotope compositions were 

strongly affected by evaporation, showed another linear relationship 

and the slope was about 5. This line is generally called 

15 May 

7 June 

1-3 August 

dD ( ) 

–40 

–60 

Fig. 2. A δ 18 O vs δD plot for the paddy water samples. The local 

meteoric water line (LMWL) in Tsukuba after Yabusaki et al (2003) 

is also shown. 

an evaporation line (EL), and it results from the difference in 

isotope fractionation ratios ( 18 O/ 16 O vs D/H) between evaporation 

of surface water and condensation of meteoric water. 

The slopes of ELs obtained in the paddy field (4.3–5.8) are 

therefore evidence of the influence of evaporation. 


In August, the increase in δ 18 O was much smaller than 

in May and June, while the slope of EL was largest, although 

evapotranspiration determined by eddy correlation was relatively 

similar (3–5 mm day –1 for all months). Comparing the 

condition of plant growth in these months, the values of LAI 

in the paddy field were 0.05, 0.73, and 5.16 m 2 m –2 , respectively. 

The much larger LAI in August than in the other months 

implies that active transpiration and the depression of evaporative 

loss by the grown rice canopy would occur. Under such 

conditions, transpiration from the rice canopy must be predominant. 

By transpiration, indeed, isotope fractionation occurs 

between the water in a leaf and the transpired vapor, but 

the isotopically enriched leaf water is never brought back to 

the ponded water. Consequently, the isotope composition of 

paddy water in August changed little in contrast to May and 

June, when significant evaporation considerably advanced the 

isotopic enrichment. 

Conclusions and future studies 

As a result of this study, it was suggested that isotopic enrichment 

of paddy water is closely related to direct evaporation of 

ponded water, rather than the total evapotranspiration from 

the paddy field. This supports the possibility to separate evaporation 

and transpiration using the isotope hydrological approach, 

usually impossible by micrometeorological measurements. 

For more detailed and quantitative analysis, we conducted 

intensive field observations of water balance and the 

collection of water samples in the same paddy field during 

rice cultivation in 2004. Based on the data, we intend to examine 

the relationship among the isotope composition of paddy 

water, the components of water balance, and the stage of rice 

growth more precisely, and to construct an isotope hydrological 

model that can calculate the ratio of evaporation vs transpiration 

from a paddy field, considering the influences of rainfall 

and relatively complex irrigation management. 

References 

Craig H. 1961. Isotopic variations in meteoric waters. Science 

133:1702-1703. 

Simpson HJ, Herczeg AL, Meyer WS. 1992. Stable isotope ratios in 

irrigation water can estimate rice crop evaporation. Geophys. 

Res. Lett. 19:377-380. 

Yabusaki S, Tase N, Kihou N, Yuita K. 2003. Characteristics of oxygen 

isotopic ratios of soil water and groundwater at paddy, 

upland and grove fields in Tsukuba, Ibaraki Prefecture, Central 

Japan. J. Jpn. Assoc. Hydrol. Sci. 33:161-176. (In Japanese 

with English abstract.) 

Notes 

Erosion control by sawah in comparison 

to other land-use systems 

Fahmuddin Agus and Irawan 

Authors’addresses: Yohei Hamada, Terrestrial Environment Research 

Center, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, 305- 

8577 Japan; Shiho Yabusaki and Norio Tase, Graduate School 

of Life and Environmental Sciences, University of Tsukuba, 

1-1-1 Tennodai, Tsukuba 305-8572, Japan; Ichiro Taniyama, 

National Institute for Agro-Environmental Sciences, 3-1-3 

Kannondai, Tsukuba 305-8604, Japan, e-mail: 

hamada@suiri.tsukuba.ac.jp. 

Sawah, also known as a paddy field, has been sustainably producing 

rice, the staple food for most Indonesians, for hundreds 

or perhaps even thousands of years. Its other functions, especially 

in producing environmental services, are indispensable, 

yet have not been recognized or are ignored by most stakeholders. 

Sawah areas are characterized by high population density 

and perfectly flat plots. The sawah plot size is usually small, 

especially on sloping land. As such, sawah farming has low 

efficiency, resulting in low profitability. Low-incentive sawah 

farming results in accelerating conversion of sawah to nonagricultural 

uses. 

This study evaluated the environmental functions, especially 

soil loss and water-retaining capacity, of sawah relative 

to other land uses in Java in 2001. The results are important 

for advocating internalization of environmental services produced 

by sawah in national land-use-related policies, such that 

the existence of sawah with its multifunctionality can be maintained. 

Objective 

The objectives of this study were to assess soil loss and water 

retention capacity (as an indicator of flood mitigation function) 

of sawah relative to that of other land-use systems. 


Location 

Erosion from sawah was measured in Ungaran (07 o 20′S; 

110 o E), Central Java Province. The soil subgroup in the sawah 

area was classified as Typic Tropaquepts. A water retention 

study was conducted in the Upper Citarum Watershed 

(6 o 40′30″–7 o 15′00″S; 107 o 30′00″–107 o 55′00″E), West Java 


Province. The dominant soil groups in the watershed are 

Eutrudepts, Dystrudepts, and Hapludalfs in uplands and 

Endoaquepts and Dystrudepts in lowlands (sawah). 

Erosion and sedimentation 

For sawah, soil loss was measured in two rice seasons from 18 

plots of terraced sawahs, ranging in area from 12 to 360 m 2 , 

with a total area of 2,515 m 2 . The macro slope of these terraces 

is 22%. The average elevation difference between plots 

was 73 cm. 

V-notch weirs were installed at the outlets of selected 

plots to determine water level and discharge from the irrigation 

canal to the upper sawah plot, from one sawah to consecutive 

lower ones, and from the lowest sawah to the stream. Sediment 

concentrations were determined from water samples by 

the gravimetric technique. 

For the upland farming systems, sediment yield data were 

obtained from the literature. 

Water retention capacity 

Water retention capacity (WRC) is watershed capacity to absorb 

and hold (rain) water temporarily such that the portion of 

water does not flow as direct runoff (Nishio 1999). This includes 

water intercepted by plants’ canopy, ponded on soil surface, 

absorbed by soil pores, and additional water that could 

be stored by paddy fields and dams. 

Paddy fields surrounded by dikes temporarily store water 

at times of heavy rain, and discharge it gradually into downstream 

rivers and surrounding areas. In this way, they function 

as mini dams and thus hold water in the plots which would 

otherwise flow directly downstream, which might in turn contribute 

to flooding. 

Upland fields, on the other hand, store rainwater temporarily 

in the porous soil layer as well as intercept rainwater in 

the plant canopy. Some temporary ponding of water occurs on 

the soil surface because of its roughness. 

To understand the contribution of each land use, the WRC 

of each land use was assessed separately. For sawah, WRC 

was estimated by the difference between dike height (which is 

normally 12–15 cm) and normal water level (usually 0.05 m). 

For the upland systems, WRC was assessed by summing the 

canopy interception, soil surface ponding capacity, and soil pore 

absorption capacity, using the equation 

WRC = (TPS – FC)*AZ + PC + IC 

where TPS is the percentage of total soil pore space, FC is the 

percentage of soil water content at field capacity, AZ is the 

depth of absorption zone or rooting zone, PC is surface ponding 

capacity, and IC is plant canopy interception capacity. 

Results 

Erosion and sedimentation 

Table 1 summarizes the measurement of sediment transport. 

The two-season data show that there was a net sediment gain in 

Table 1. Amount of sediment entering and leaving 18 plots of terraced 

sawah. 

Variable 

First a 

Rice season 

Second b 

Total sediment entering sawah from 3.4 6.2 

irrigation canal (t ha –1 ) 

Sediment yield from sawah (t ha –1 ) 1.4 0.8 

Sediment yield from sawah during 0.7 0.6 

tillage (t ha –1 ) 

Net sediment deposition (t ha –1 ) 2.0 5.4 

a 31 Oct. 2001 to 31 Jan. 2002. b 16 Mar. 2002 to 1 July 2002. 

paddy fields as high as 2.0 t ha –1 in the first season and 5.4 t 

ha –1 in the second season. Sediment output from the plots is 

relatively high during the plowing and puddling periods but 

the sediment transported from one plot is mostly deposited in 

the next few lower plots and thus the net output of sediment at 

the end of the terraces is very low (about 2.2 t ha –1 y –1 during 

the two seasons). In comparison, sediment output was 10–20 t 

ha –1 y –1 from a 1.1-ha catchment planted to annual upland crops 

at a nearby location (Agus et al 2003). Similar research results 

in Indonesia (van Dijk 2002) suggested that sediment transported 

from rainforest and logged pine plantation forest is very 

low (3–7 t ha –1 y –1 ) compared to upland agricultural land uses, 

such as multistrata agriculture (10–12 t ha –1 y –1 ), annual cropbased 

agriculture on bench terraces (19–25 t ha –1 y –1 ), and 

vegetables on steep slopes with terraces (42–75 t ha –1 y –1 ). 

Therefore, it can be stated that the function of sawah in reducing 

soil erosion and sedimentation is more effective than, or at 

least is as effective as, forest. 

Water retention capacity 

The water retention capacity of different land uses at Citarum 

Watershed, West Java, is presented in Figure 1. The data show 

that forest and mixed cropping (tree-based multistrata system) 

in this study area had a WRC of about 0.15 and 0.09 m, respectively. 

Meanwhile, the WRC of paddy field, annual upland, 

and housing and industrial areas was 0.08, 0.06, and 0.02 

m, respectively. This means that during and shortly after heavy 

rain 1 ha of forest and sawah can store, respectively, about 

1,500 and 800 m 3 of water before runoff takes place. The capacity 

of sawah to retain water is much higher than that of 

annual upland and housing and industrial areas of 600 and 200 

m 3 ha –1 . Therefore, it can be stated that, if more sawah and 

upland agricultural lands are converted to housing and industrial 

development, the same amount of rainfall will cause greater 

runoff, and will finally cause a higher chance of flood. 

Similar research in Japan showed that the average WRC 

for forest, sawah, orchard, grassland, and dry cropland was 

0.18, 0.15, 0.11, 0.02, and 0.04 m, respectively (Nishio 1999). 


WRC (m) 

0.16 

0.14 

0.12 

0.10 

0.08 

0.06 

0.04 

0.02 

0.00 

Fig. 1. Water retention capacity (WRC) of sawah and other landuse 

systems. 

Conclusions 

Forest Sawah Mixed 

cropping 

Annual 

upland 

Interception capacity 

Ponding capacity 

Pore absorption 

Housing 

Industry 

Sawah is superior in controlling erosion to other land uses. 

Erosion mainly happens in sawah during and shortly after tillage 

operations. 

The capacity of sawah to retain water during and shortly 

after rainfall is comparable with that of the tree-based mixed 

cropping system and significantly higher than that of annual 

crop-based and housing and industrial areas. Policy measures 

should be implemented to maintain the existence of sawah considering 

the significant environmental services it can offer, in 

addition to its tangible role as a rice producer. 

References 

Agus F, Vadari T, Watung RL, Sukristiyonubowo, Valentin C. 2003. 

Effects of land use and management systems on water and 

sediment yield: evaluation from several micro catchments in 

Southeast Asia. In: Maglinao AR, Valentin C, Penning de Vries 

F, editors. Proceedings of IWMI-ADB Project Annual Meeting 

and 7th MSEC Assembly, Vientiane, Lao PDR, 2-7 Dec. 

2002. Bangkok (Thailand): International Water Management 

Institute, Southeast Asia Regional Office. p 135-149. 

Nishio M. 1999. Multifunction character of paddy farming. Paper 

presented at the Second Group Meeting on the Interchange of 

Agricultural Technology Information between ASEAN Member 

Countries and Japan, Jakarta, Indonesia, 16-18 February 

1999. 

van Dijk AJM. 2002. Water and sediment dynamics in bench-terraced 

agricultural steeplands in West Java, Indonesia. PhD 

dissertation. Vrij University, Netherlands. 363 p. 

Notes 

Authors’ address: Soil Research Institute, Jln. Juanda 98, Bogor 

16123, Indonesia, e-mail: f.agus@cgiar.org. 


Although it is recognized that the cultivation of paddy rice sustains 

soil fertility, there is a large area whose fertility of soil is still 

low. Therefore, it is necessary to increase both the sustainability 

and productivity of these soils urgently. One of the main themes 

of the conference is “the development of sustainable rice cultivation 

based on environment and food security.” In this session, 

this theme was discussed from the viewpoint of soil and water 

conservation. 

Participants entered into “the world of paddy soils” with 

the presentation of Prof. K. Kyuma (Japan). His keynote lecture 

was “Paddy soils around the world.” Based on his more than 40 

years’ experience in the field of paddy soil science, he emphasized 

the importance of diversity of paddy soil and its environment. 

Diversity and sustainability of paddy soils were discussed 

by N. Ngoc Hung (Vietnam), K. Naklang (Thailand), and T. 

Wakatsuki (Japan). N. Ngoc Hung reported on the current status 

of soil fertility in the Mekong Delta of Vietnam. He pointed out 

some problems in fertility such as nitrogen (N) loss through ammonium 

volatilization. In contrast to the relatively fertile Mekong 

Delta, the paddy soils in northeastern Thailand are very infertile. 

K. Naklang introduced her long-term field trials, and discussed 

the effect of organic matter application on the improvement of 

soil fertility. T. Wakatsuki reported that the area with lowland rice 

farming has increased in West Africa. He predicted that a Green 

Revolution with paddy rice farming would take place in this area 

with the proper integration of forestry, upland farming, and lowland 

rice farming. 

To optimize fertilizer management in such diverse paddy 

soils, C. Witt (Singapore) proposed an effective and site-specific 

management of nutrients on the basis of his field trials with fertilizer 

application. 

Recently in China, rice yield has increased dramatically with 

the increase in fertilizer application. However, the amount of application 

for rice exceeds the requirement. As a result of this 

imbalance, N pollution is becoming a serious environmental problem. 

J. Zhu (China) evaluated the environmental impact of N loss 

under the wheat-rice system in the region around Taihu Lake. To 

solve this problem, it is necessary to increase the recovery of N 

from fertilizer urgently. M. Saigusa (Japan) reported that the application 

of polyolefin-coated urea, which was used as a controlled-availability 

fertilizer (CAF), improved N recovery. The invention 

of CAF now enables a co-situs application (contact application 

of fertilizer with plant roots) and that is improving the efficiency 

of fertilizer, environmental loading from fertilizer, and so 

on. W. Bowen (Bangladesh) also reported a good recovery of N 

from fertilizer with the deep placement of supergranule urea for 

lowland rice in Bangladesh. 

D. Olk (USA) indicated that the yield decline caused by 

intensive rice cropping under continuously flooded conditions was 


associated partly with the decrease in soil N availability. He suggested 

that phenolic compounds from lignin residues in soil absorbed 

soil organic N, and that this binding was related to the 

decrease in mineralization of soil N late in the season. He emphasized 

that this anilide structure was recalcitrant and might 

release N to the rice plant only very slowly or upon field drainage 

and soil aeration. H. Sumida (Japan) and D.B. Lee (Korea) reported 

that N fertility of the soil was often exhausted when paddy 

soils were converted into upland crops in their countries. They 

reported a decrease in soil N availability under continuous upland 

conditions. The appropriate application of organic matter or 

re-conversion of rotated paddy fields into flooded rice was necessary 

to maintain fertility in flooded rice–upland crop rotation 

systems. 

Rice is one of the greatest sources of dietary intake of 

cadmium (Cd) in some Asian countries. The amounts of Cd in 

rice grain sometimes exceeded the critical level that was proposed 

by Codex. Therefore, it is urgently required to develop technologies 

for reducing Cd levels in rice grain. S. Ishikawa (Japan) 

showed current advances in ameliorating technologies to reduce 

Cd content in polluted fields and promising technologies for 

remediating Cd-polluted soils. 

The general discussion covered the diversity of paddy soils, 

maintenance of soil fertility, and conservation of environments. 

Based on the diverse paddy rice farming in Asia, West Africa, and 

other regions, it was important to recognize that ecological engineering 

improved the artificial environment of paddy soil as a 

whole. D. Olk focused on soil aeration in terms of the decrease 

in available N under continuous flooded conditions. To avoid 

changing available N into recalcitrant forms, how much soil aeration 

is necessary Rotation with upland crops is one method for 

aeration, and its frequency may be mainly a matter of economics; 

therefore, the relative profits for rice versus an upland crop 

should be calculated. Also, the presence of another decline in 

soil fertility by the paddy-upland rotation should be taken into 

account. Future research should examine the effects of soil aeration 

on the quality of organic matter as well as N availability. 

However, many other potential effects of timely soil aeration also 

deserve study. 

In conclusion, this session provided an excellent opportunity 

to facilitate the science of paddy soil among participants, 

and it shall contribute to improving our future. 


SESSION 13 

Farmers’ participatory approaches 

to facilitate adoption of improved technology 

CONVENER: T. Paris (IRRI) 

CO-CONVENER: J.S. Caldwell (JIRCAS)

How does a farmer accept a new technology 

Fujihiko Tozawa 

Ogata village, located in Akita Prefecture in northern Japan, 

was built on the reclamation of Hachirogata Lake, formerly 

the second-largest lake in Japan, during 1957-77. This was the 

biggest agricultural land reclamation in Japanese history. The 

new reclaimed land had an area of about 15,600 ha, with residual 

canals and lakes of about 5,000 ha. When the project 

was completed in 1977, Ogata was expected to be a “model of 

modern agriculture,” based on large-scale and productive farming 

systems. Some 589 farmers were settled in Ogata from all 

over the country. Each of them had a farm of 15 ha, regarded 

as large-scale in Japan, where the average farm size is about 1 

ha. 

Like other farmers of Ogata village, I am a full-time and 

large-scale rice farmer. In 1998, I established a company with 

some friends to sell our rice. As full-time farmers, we are always 

interested in technological innovations in farming and 

farm management because we always want to reduce costs and 

sell our rice under better conditions. Unlike other Japanese 

farmers, the majority of farmers of Ogata do not sell their rice 

to agricultural cooperatives, but we have multiple marketing 

channels. For instance, we can sell rice by individual direct 

marketing, or group direct marketing, or to rice wholesalers. It 

is common here that a single farmer has different sales channels 

and grows rice in different ways on his farm to respond to 

his different consumers. Farmers of Ogata are all independent 

farm owners who must decide what kind of rice they will grow, 

to whom they will sell, and how much. Therefore, it takes us a 

long time to consider whether we should accept a new technology 

because we should examine whether the new technology 

is compatible with other factors of our farm management. 

As for environmental problems, we have been suffering 

from a specific one: worsening of the water quality of the residual 

lake of the commonly called Hachiroko Lake, whose 

water quality (measured by COD) was listed as one of the worst 

five lakes in the country in 2001. Ogata farmers are respon- 

sible for the water quality deterioration of Hachiroko Lake 

because we use its water for agricultural irrigation and drainage. 

During the 40-year history of Ogata, we have been struggling 

to improve the water quality of Hachiroko Lake by stopping 

the sale of synthetic detergents in a store of the agricultural 

cooperative, banning pesticide spraying by airplanes, stopping 

the project of a golf resort planned within the village, 

developing various kinds of sustainable agricultural technologies 

from organic to no-till, etc. These efforts were partial and 

fragmentary at first, but they were expanded to the whole community 

and became systematic in the 1990s, when researchers 

of Akita Prefectural University and the Akita Agricultural Experimental 

Station started joining farmers to help them develop 

new technologies and social systems for sustainable farming 

on the reclaimed land. A series of collaborating programs between 

farmers and researchers began at that time, including 

Ogata Low-Input Sustainable Agriculture (O-LISA 1991), the 

Research Project for Sustainable Agriculture in Ogata (RPSAO 

1998), and the Declaration of Environmentally Creative Agriculture 

in Ogata (ECA 2001). In 2001, we established a voluntary 

organization to promote ECA, and I have been president 

of this organization. 

The above research has revealed that Ogata is the largest 

single area of sustainable rice production in Japan. Table 1 

shows examples of sustainable farming practices in Ogata. 

From the top, we see that the area of “Less chemical farming” 

is 5,110 ha and accounts for 74.1% of the total farmland of the 

village. “No chemical farming” (not the same as “organic”) is 

520.8 ha and the percentage of the total farmland is 7.5%. 

These lower input practices already cover 80% of the total 

farmland of the village, whereas conventional farming covers 

20%. As you can imagine, these figures are extraordinarily 

high in Japan in terms of total area, area per capita, and percentage 

of total farmland. We are also active in accepting “more 

friendly practices with Hachiroko,” namely, the “use of con- 

Table 1. Various sustainable farming practices in Ogata (1998). 

Area Number of Area per % of area in 

Farming practice (ha, a) farmers (b) capita the total farmland 

(ha, a/b) of Ogata 

Less chemical farming 5,110.0 – – 74.1 

No chemical farming 520.8 110 4.7 7.5 

Use of controlled-availability fertilizer 1,214.7 112 10.8 17.6 

Local fertilization 1,052.3 84 12.5 15.3 

No-puddling 132.7 14 9.5 1.9 

No-till 36.8 10 3.7 0.5 

Direct seeding 7.2 2 3.6 0.1 

Weeding with ducks 71.5 12 5.9 1.0 

Source: Taniguchi and Sato (2001). 


trolled-availability fertilizer,” “local fertilization,” “no-puddling,” 

“no-till,” “direct seeding,” and “weeding with ducks.” 

Now I return to the theme of this paper: how a farmer of 

Ogata accepts a new sustainable technology. To answer this 

question, I would like to compare organic farming and no-till 

farming. These two practices are useful when we investigate 

what makes a farmer accept a new technology. It is said that 

the first farmer who challenged organic farming in Ogata was 

Mr. Tsuneo Gotsu in 1982. As organic farming technologies 

were not developed at that time, he had to employ women living 

in neighboring communities to weed rice paddies by hand. 

His neighbors thought that he had become crazy because it 

was unthinkable not to use herbicide in such large-scale paddies 

as Ogata’s. These efforts were basically “ethical,” that is, 

based on goodwill for a better environment. But, in the 1990s, 

sustainable farming practices spread rapidly to the whole community, 

mainly because consumers’ increasing concern about 

food safety caused a sharp expansion of “safe food markets” 

all over the country. A lot of offers came from wholesalers and 

retailers. Many farmers of Ogata began growing rice in an organic 

or less chemical manner, not based on ethical intentions 

but on commercial interests. There were many technical difficulties 

with organic farming: providing organic fertilizer; seedling, 

weeding, disease, and insect control; and others. But farmers 

tried to solve these problems in various ways. Establishing 

a study group and experimenting with new technologies with 

other farmers are the most common ways of solving problems. 

Sometimes we invite diverse researchers to study new technologies 

or new materials, and we try them within our own 

paddy fields. It takes time to decide whether we accept a new 

technology or material because we need to see whether it actually 

works in the particular conditions of our own fields. With 

efforts for years, the major difficulties of organic farming were 

solved, but weeding is still very difficult. What has moved 

farmers forward in technological innovation in organic farming 

is, in my view, the economic incentive. How much farmers 

receive for growing rice organically varies. This depends on 

what kinds of customers they have. If they find wealthy consumers, 

they could receive double the standard price of rice, 

or more. So, information about customers is the “top secret” 

for farmers of Ogata. It sometimes happens that one Ogata 

farmer tries to take a customer from another farmer. 

On the other hand, no-till is not popular in our village 

mostly because we have to use chemical fertilizer and herbicide 

for no-till farming, so that we cannot expect a higher price 

in the current rice market, where the use of organic fertilizer 

and nonuse of chemicals are highly valued. Though we know 

that no-till is very effective in making water cleaner, only ten 

farmers among more than 500 are now practicing no-till. The 

great popularity of organic farming in Ogata, in contrast, has 

caused concern from researchers about water pollution caused 

by overusing organic fertilizer. Many organic farmers are aware 

of this problem but probably they would not change their farming 

style until a time when someone shows them an alternative 

that could provide profitability and sustainability compatible 

on a higher level. 

Reference 

Taniguchi Y, Sato S. 2001. An attempt of farmer-university collaboration 

in research on sustainable farming practices. Nogyo 

Oyobi Engei (Agric. and Hort.) 76-3:338. (In Japanese.) 

Notes 

Author’s address: Hanasaka Farm, 010-0442, Higashi 4-47-1, Ogata, 

Akita, Japan, e-mail: fugihiko@ogata.or.jp. 

Farmer participatory evaluation of nitrogen 

management technology: the case 

of the leaf color chart in West Bengal, India 

B. Bagchi, M.Z. Abedin, and S.K.T. Nasar 

West Bengal, the largest rice-growing state in India, has made 

respectable progress in rice production over the last two decades. 

Rice production increased from 11.2 million tons in 

1980-81 to 20.9 million t in 2001-02, mostly because of technological 

progress that raised yield from 2.16 to 3.39 t ha –1 . 

But the price of rice has increased at a much slower rate than 

the unit cost of cultivation, showing the strains of diminishing 

returns to farmers. One way to increase the profitability of rice 

cultivation is to reduce the unit cost by adopting efficient crop 

management technologies that can save modern inputs. 

Since the introduction of high-yielding varieties, farmers 

have been applying more fertilizer than the recommended 

dose. This reduces the agronomic and economic efficiency of 

N fertilizer and has negative consequences for the environment. 

The International Rice Research Institute (IRRI), through 

experiments on soil-plant development analysis, developed a 

simple decision-making tool called the leaf color chart (LCC) 

for the timely application of N fertilizer to the rice crop. The 

color shades of the LCC card are designed to match leaf colors 

of the rice plant from signs of nitrogen “hunger” to overfeeding 

of the plant. By matching the color on the LCC with 

Session 13: Farmers’ participatory approaches to facilitate adoption of improved technology 395

that of the rice plants, farmers can decide on the proper time 

and amount of N application. LCC validation experiments in 

Vietnam and other countries have shown that farmers could 

save a substantial amount of nitrogen without any reduction in 

yield (Balasubramanian et al 2000, Bijay Singh et al 2002). 

To validate the technology with farmers, social scientists 

emphasized conducting farmer participatory research that 

would help farmers assess not only the suitability of the technology 

to the farmers’ biophysical and agronomic environment, 

but also its acceptability to farmers given their socioeconomic 

environment and cultural background. Once the farming community 

sees the benefit of the technology through its own experimentation, 

demand is expected to grow and the process of 

adoption would become faster. Community participatory, 

farmer-managed trials were therefore conducted by farmers in 

West Bengal, India, during 2002-04 to validate the efficiency 

of the LCC in managing nitrogen in rice cultivation. 

The overall objectives were (1) to enable farmers to 

evaluate the effectiveness of the LCC in N management and 

(2) to assess the economic and other benefits associated with 

the use of the LCC. Specific objectives were to observe the 

effectiveness of a community participatory approach in technology 

assessment and its adoption. This paper reports the results 

of experiments in the dry season of 2002-03 and transplanted 

rice premonsoon and monsoon seasons in 2003 in West 

Bengal, India, by Bidhan Chandra Krishi Viswavidyalaya, 

IRRI, and a local NGO. 


Out of six agroclimatic zones of West Bengal, two zones, the 

new alluvial and old alluvial, were chosen for the dry season. 

Initially, six villages from six districts of the two zones were 

randomly selected for the study. As it was thought to be difficult 

to monitor intensively and interact closely with the farmers, 

a cluster of four neighboring villages in the premonsoon 

season and five villages in the monsoon season was identified 

purposively, during the subsequent seasons, based on representativeness 

of the flood-prone rice environment and the target 

group, and ease of monitoring and supervision. To do the 

work smoothly, partnership with a local NGO, the Nadia Zilla 

Farmers’ Development Organization, was established. Participating 

farmers were selected using a community participatory 

approach. Meetings were organized in each village to discuss 

the LCC and its role in the judicious use of nitrogen in rice 

cultivation. Farmers from each village selected 10 representatives 

from the willing farmers in the dry season. Farmers were 

informed of technical details for using the LCC. Since validation 

should be under farmers’ management conditions, they 

were informed that no material and financial inputs would be 

given to the participants so that their decision-making would 

not be affected or distorted. To establish an ongoing monitoring 

system at the community level, small groups of farmers 

were formed to share each LCC. Measurements were taken 

using the LCC matching the color of 10 uppermost fully expanded 

leaves from randomly selected plants per plot from 

the 14th day after transplanting (DAT) at 7-day intervals until 

flowering. For the pre-wet-season and wet-season crop, the 

critical value was 3.5, and for the dry-season crop it was 4. 

Yield was estimated through the crop-cutting method. 

Field days were organized before harvesting, enabling farmers 

to evaluate the technology. During field days, the farmers 

discussed among themselves and evaluated the performance 

by ranking the efficacy of the LCC using their own criteria. 


Farmers’ general perceptions were that the greener were the 

leaves, the higher yield would be. However, after testing, the 

participating farmers were of the opinion that, by using the 

LCC, the amount of nitrogenous fertilizer applied was reduced 

without affecting production. This saved farmers money and 

helped increase net returns. Nitrogen use decreased by 20.2, 

30.2, and 36.0 kg ha –1 over the farmers’ traditional practice in 

the pre-wet-, wet-, and dry-season rice crops, respectively. 

Participating farmers thus saved a substantial amount as fertilizer 

cost is an important part of direct costs. 

Farmers found out from their observations and from trials 

designed and conducted by some of them that using the 

LCC could significantly reduce the use of pesticides, too. They 

were of the opinion that higher doses of N fertilizer were responsible 

for increased pest and disease infestation. 

The real-time use of nitrogen affected the cost of cultivation 

and net returns of rice in different seasons. Farmers 

earned more profit per hectare in the LCC plots than in the 

farmers’ own plots. These amounts were more in LCC plots 

than in farmers’ plots by US$13.10 in pre-wet rice, $20.40 in 

wet rice, and $19.40 in dry rice (Table 1). 

During the field days, participating and nonparticipating 

farmers evaluated the performance of LCC technology. 

They found out that, in the LCC plots, the number of unfilled 

grains was lower than in farmer-managed plots and plants of 

LCC plots did not lodge despite strong winds, whereas plants 

in the farmers’ plots lodged. Some of the varieties also matured 

earlier by about 5 to 7 days when N was managed by the 

LCC than in farmers’ plots. 

These positive effects have significantly changed the perception 

of farmers that very dark green leaves are associated 

with higher grain yield. They now understand that N should be 

applied when the plant starts feeling “hungry” for nitrogen. 

The farmers’ rankings showed that the reduction in N 

use was most important and reduction in lodging was the least. 

The rankings were more or less uniform across communities, 

suggesting that farmers were equally concerned about the overuse 

of N. Influenced by the success of the LCC experiments, 

the number of participating farmers tripled by the third season. 

The rate of adoption has increased each season and there 

have been no reports of discontinuation of LCC use. 

The LCC technology was found to be eco-friendly. The 

reduction in N use has potential to reduce NO 3 pollution and 

minimize the use of pesticides, which would have a beneficial 

effect on human health, biodiversity, and soil and water qual- 


Table 1. Cost of cultivation and profit of rice cultivation in the leaf color chart (LCC) plot and farmers’ practice 

(FP) in different seasons of West Bengal. 

Cost a ha –1 Difference in Gross return ha –1 Profit ha –1 Difference 

Season cost ha –1 in 

LCC FP LCC FP LCC FP profit ha –1 

Early wet season 254.90 263.80 8.90 332.60 328.40 77.70 64.60 13.10 

Wet season 328.20 343.10 14.90 413.40 407.80 85.20 64.80 20.40 

Dry season 421.50 445.30 23.80 591.40 595.80 169.90 150.50 19.40 

a Values are in US$. 

ity. The participating farmers believed that this simple tool 

empowers them to make decisions on the time and amount of 

N to be applied. 

It was thought that the community participatory approach 

used strengthened the partnership among key stakeholders, 

which facilitated community action and technology adoption. 

The farmers believed that more and more frequent interaction 

with the scientists and stakeholders would help enhance their 

knowledge of improved farming practices. It appeared that an 

action group has been formed in this locality. The farmers of 

this group were very close to each other. Evidence from the 

participatory approach used showed that farmer-to-farmer 

transfer of technology was taking place, and the technology is 

likely to be adopted by the large group of farmers. This experiment 

also had an impact on the scientific community. The 

scientists who visited the experimental plots appreciated the 

capabilities of farmers in testing new technologies under their 

own management. 

Conclusions 

Real-time management of nitrogen with the LCC helped farmers 

minimize N use, reduce the cost of cultivation, decrease 

pest and disease incidence, minimize chaffy grains, and protect 

the crop from lodging. Farmers’ perceptions that the use 

of more nitrogenous fertilizer brings more yield changed. 

Community action in technology testing facilitated better 

and faster decision-making for N management. Small groups 

established to share each LCC facilitated such evaluation. Frequent 

interaction with the scientists and within the farming 

community improved the knowledge base of the farmers. This 

helped to strengthen the social capital. However, the trial confirmed 

that a varietal interaction does exist and the communities 

should be encouraged to evaluate this in partnership with 

researchers. 

References 

Balasubramaniam V, Morales AC, Cruz RT, Thingarajan TM, 

Nagarajan R, Babu M, Abulrachman S, Hai LH. 2000. Adoption 

of the chlorophyll meter (SPAD) technology for real-time 

nitrogen management in rice: a review. Int. Rice Res. Notes 

25(1):4-8. 

Bijay Singh et al. 2002. Chlorophyll meter- and leaf color chartbased 

nitrogen management for rice and wheat in northwestern 

India. Agron. J. 94. 

Notes 

Authors’ addresses: B. Bagchi and S.K.T. Nasar, Directorate of Research, 

Bidhan Chandra Krishi Viswavidyalaya, Nadia, West 

Bengal, India; M.Z. Abedin, International Rice Research Institute, 

Los Baños, Philippines, e-mail: drbbagchi@vsnl.net. 

A farmer participatory approach in the adaptation 

and adoption of controlled irrigation for saving water: 

a case study in Canarem, Victoria, Tarlac, Philippines 

F.G. Palis, M. Hossain, B.A.M. Bouman, P.A.A. Cenas, R.M. Lampayan, A.T. Lactaoen, T.M. Norte, V.R. Vicmudo, and G.T. Castillo 

Water plays a major role in rice cultivation. On average, some 

4,000 L of water are required to produce 1 kg of rice (Tuong 

and Bouman 2003). Thus, an inadequate supply of water from 

crop establishment to the reproductive stage of the crop generally 

leads to a significant yield reduction (Wopereis et al 

1996, Bouman and Tuong 2001). 

The water supply is increasingly becoming scarce, however, 

because of the world’s increasing population that brings 

along multiple competing demands, such as for agriculture, 

industry, and domestic use, and climatic changes such as El 

Niño phenomenon (IRRI 1995, Bouman and Tuong, 2001). 

The increasing water scarcity threatens not only general food 

security in Asia but also the livelihood of most Asian farmers 


ecause of the increasing irrigation costs borne by them. To 

meet the food demand of Asian consumers and producers, efficient 

water management practices are needed that maintain 

rice yields and production at a high level. 

To solve the problem of water scarcity, researchers have 

been developing water-saving irrigation technologies. In the 

Philippines and in 2001, IRRI began the project “Technology 

Transfer for Water Savings” (TTWS) in collaboration with the 

National Irrigation Administration (NIA) and Philippine Rice 

Research Institute (PhilRice) to facilitate farmers’ adaptation 

and adoption of controlled irrigation (CI) (Bouman et al 2002, 

Lampayan et al 2003). CI entails an irrigation schedule in 

which, contrary to the normal practice of continuous flooding, 

water is applied to the field a number of days after the disappearance 

of ponded water and yet high yields are maintained 

(Tabbal et al 2002, Belder et al 2004). The CI technology was 

piloted among farmers using irrigation water drawn from deepwell 

systems in Tarlac Province, where these deep wells are 

maintained and operated by farmers’ irrigation service cooperatives 

(ISC). 

The successful adoption of a new technology such as CI 

depends on not only its suitability and profitability but also on 

social and cultural factors such as farmers’ perceptions. Furthermore, 

where irrigation water use is organized in a communal 

way, the organizational structure and functioning play a 

key role as well. The objectives of this paper are (1) to determine 

farmers’ perceptions of CI as an effective water-saving 

technology in rice production and (2) to explore success factors 

for collective action in an organized irrigation management 

system that would facilitate and ensure eventual farmers’ 

adoption. 


We focus our study on the case of the deep well P-38 of the 

TTWS project, a reactivated system in Canarem, Victoria, 

Tarlac, about 70 km north of Manila. Eleven and 12 farmer 

volunteers in the 2002 and 2003 dry season (DS), respectively, 

participated in the CI experiment. Each farmer participant contributed 

two (neighboring) plots with a size of 500–1,000 m 2 

each, one representing the current farmers’ practice (FP) and 

one the CI technology. The water management of the CI plots 

was a joint affair among all project members and aimed at 

site-specific optimization of the technology (in terms of irrigation 

rotation, depth of application, number of days without 

ponded water, etc.). All activities and farm inputs and output 

were closely monitored in both seasons. An input-output survey 

of rice production for all 63 farmers was done in the 2003 

DS using a semistructured questionnaire. Focus-group discussions 

(FGDs) and key informant interviews were conducted 

among different stakeholders such as ISC officers and members, 

NIA staff, and IRRI water scientists. Field observations 

were likewise made and records of ISC meetings were documented 

throughout the dry-season rice cropping. 

Table 1. Farmers’ perceptions about controlled irrigation (CI) as a 

water-saving technology. 

Perceptions about CI (n = 15) Frequency Percentage 

1. Less time, less expensive, and saves labor 15 100 

2. It saves a lot of water 15 100 

3. No yield difference vis-à-vis farmers’ 

practice (FP) 14 93 

4. I learned a lot of new techniques in modern 

rice production such as fertilizer application, 

selection of pure seeds, when to spray, how 

to save water during irrigation, etc. 11 73 

5. If CI would be adopted by all members in 

P-38 ISC, the water rotation interval would 

be improved in timeliness, reliability, and 

equity of water distribution from one field 

to another. 11 73 

6. Grains in CI plots are rounded and heavier 5 33 

7. FP plots could not produce more tillers 

because of continuous flooding 3 20 

8. I observed that, in modern rice 

production techniques, straight planting 

requires more people 2 13 

9. Others 3 20 

Others: Crop stand in both FP and CI are the same, CI plots have no diseases. 

Farmers’ perceptions of CI 

The majority of the farmer-cooperators gave positive feedback 

about the effectiveness of CI as a water-saving technology, as 

shown below and in Table 1. 

No yield difference compared with farmers’ practice 

(FP). Most farmers perceived that yield in the FP plots was 

similar to that in the CI plots, regardless of less water being 

used in the CI plots. This perception agrees with the yields 

obtained from crop-cut samples of 2 × 2.5-m 2 areas as indicated 

in Table 2. In both 2002 and 2003, there was no significant 

yield difference under FP and CI. 

Saves water. Farmers recognized that CI saves water. 

Aside from alternate wetting and drying of the paddy field, CI 

maintains only a 2-cm depth of standing water in contrast to 

the usual farmers’ practice of maintaining a 5–8-cm water depth. 

On average, the amount of water saved in CI versus FP was 

16% in 2002 and 24% in 2003 (Fig. 1). The largest water savings 

were 24% in 2002 and 33% in 2003. The higher savings 

in 2003 reflect the effects of the learning process: at that time, 

farmers were already confident about the performance of CI 

and willing to take more risk in saving more water. 

Saves time, labor, and expenses. All farmer-cooperators 

acknowledged that CI saves time, labor, and expenses. It reduced 

costs by using 20–25% less fuel and oil (Table 2). It 

reduced labor as farmers spent fewer hours in irrigation, depending 

on the distance. Moreover, the ISC could abolish water 

delivery during nighttime. 

Better grain quality. Farmers observed that rice grains 

from CI plots were heavier, bigger, and in better shape because 

of soil aeration, whereas, in FP plots, rice grains were 

lighter and more slender, and sometimes unfilled, because of 


Table 2. Average yield, costs, and returns of rice grown under two water management practices 

for demonstration farmers in Canarem, Victoria, Tarlac, for the dry season of 2002 and 

2003. The last column gives mean data for 2003 of all 63 farmers. FP = farmers’ practice, CI 

= controlled irrigation, n = number of farmers. 

Item 

2002 dry season 2003 dry season All farmers of 

P-38 ISC 

PF CI PF CI 

(n = 11) (n = 12) (n = 63) 

Yield (kg ha –1 ) 5,400 5,400 6,230 6,080 4,550 

Gross returns (pesos ha –1 ) 51,300 51,300 62,348 60,785 44,226 

Costs (pesos ha –1 ) 

Materials 15,701 13,329 3,370 12,085 11,856 

Seed 1,950 1,950 600 600 1,114 

Fertilizer 2,400 2,400 5,362 5,632 3,711 

Pesticides 650 650 252 252 389 

Fuel and oil 9,841 7,469 6,769 5,484 6,640 

Others 860 860 387 387 – 

Labor 8,567 8,521 15,174 14,936 17,242 

Land preparation 2,500 2,500 3,089 3,085 733 

Crop establishment 2,700 2,700 3,658 3,658 2,010 

Crop care 867 821 1,496 1,496 347 

Postharvest labor 2,500 2,500 6,696 6,696 8,662 

Permanent labor – – – – 5,488 

Total production cost 24,268 21,849 28,545 27,021 29,098 

Net profit (pesos ha –1 ) 27,032 29,451 33,803 33,765 15,127 

the absence of soil aeration. In 2002, members of the TTWS 

project took grain samples from CI and FP plots for some quality 

analysis. The results showed no significant difference between 

CI and FP grains in the milling indicators, that is, head 

and broken rice, brewers’ rice, bran, chalky and immature 

grains, discoloration, and damaged grains. 

ISC functioning and impact of CI 

At the start of the TTWS project, tensions about water use 

were rising because of the increase in membership of the ISC 

(from 45 in 1998 to 63 at present), and because of subsequent 

increases in service area (from 50 to 70 ha) and the lengthening 

of the interval of water delivery (from 7 d to 12–14 d). The 

introduction of CI added to the tension because farmers were 

apprehensive of looking at their rice fields with dry land, especially 

when they saw the cracking of the soil. Many farmers 

doubted that CI would give them the same yields as or higher 

yields than their usual practice and, thus, some farmers started 

stealing water by placing holes underneath their paddy dikes, 

hidden from other farmers’ eyes. Conflicts were resolved 

through the intercession of village officials, particularly the 

village security officers. 

At the same time, many farmers were delinquent in paying 

their dues, resulting in a temporary stoppage of irrigation 

operations. Within 2 days, most farmers settled their bills, perhaps 

because water deliveries ceased when the crop was close 

to the flowering stage, when water is really much needed. Consequently, 

the ruling for having access to water was likewise 

changed from using NIA’s fuel and oil supply to using an individual 

supply. All farmers need to bring their own fuel and oil 

to the pump for irrigation of their fields, and this resulted in a 

remarkable 40% decrease in the total amount of fuel and oil 

used. During the time when the cooperative was using fuel 

provided by NIA, the farmers were quite extravagant in their 

use, but they were very prudent in their use when fuel was 

individually provided. 

After two years, the experiences with CI also helped ease 

tensions in water use. Farmers were no longer apprehensive of 

not having ponded water on their fields for some days as they 

had experienced that this did not reduce yields. Cohesive interaction 

among farmers also increased during the experimentation 

because everyone was curious about the performance of 

CI. And, when farmers saw the effectiveness and viability of 

CI, many were convinced that CI could reduce their water consumption. 

Since CI reduced water consumption, CI became a 

significant factor in solving operation problems and induced 

farmers to cooperate, that is, with the fuel scheme of bringing 

one’s own fuel and oil, and to be a continuing member of the 

deep-well system. 

Factors in success of collective action 

and adoption of CI 

Collective action is the basic foundation for facilitating adoption 

of CI in a deep-well irrigation system. Below are the factors 

identified as affecting the success of collective action. 

Group size. Groups should be small to minimize transaction 

costs, but cannot be too small or else the ISC won’t be 

able to cover the costs of operating and maintaining the sys- 


% savings P 38, Canarem 

35 

30 

25 

20 

15 

10 

5 

0 

High 

Middle 

Low 

Average 

High 

Middle 

Low 

Average 

2002 dry season 2003 dry season 

Fig. 1. Average water savings (%) in controlled irrigation over farmers’ practice in Canarem, 

Victoria, Tarlac, for 2002 and 2003 dry seasons, based on data from 11–12 farmers. 

tem. Relatively small group sizes are generally associated with 

homogeneity among members; in addition, communication 

among members takes less time and effort. In the experience 

of P-38, communication and management (particularly in terms 

of payment) were easier with less than 50 members. 

Service area size. The size of the service area should be 

small, just large enough to maintain a water delivery schedule 

that would not create water stress to rice plants. At P-38, when 

the service area increased from 45 to 78 ha, the interval between 

water deliveries increased to accommodate all the fields. 

A long time interval is not acceptable to most members, and 

they prefer a 7-day interval rather than 12- to 14-day intervals. 

Profitability. Relative profitability of adopting CI should 

be high enough to make collective action profitable. Based on 

the experiments in 2002 and 2003, the amount of water used 

in CI was lower than in the usual farmers’ practice (Fig. 1), 

and hence irrigation costs were lower as well. Water savings 

averaged 16–24% and costs savings 20–25% (Table 2). 

Strong leadership—strict enforcement of rules to solve 

free-rider problem. In the first four years of operation, fuel 

and oil were supplied by NIA as part of a loan. Records show 

that ISC members used fuel excessively, probably because the 

payment was the same for each member and because payments 

were not immediate but made only after harvest. The chairman, 

imbibing the Filipino values of smooth interpersonal relationships 

and a sense of shame, was not so strict in enforcing 

the policies and regulations, particularly those related to paying 

dues to the ISC. But, with the strict enforcement of rulings 

on payment (that had resulted in halting the operation of P-38 

for two days) and access to water (bringing one’s own fuel), 

everyone was frugal in the amount of fuel used. These new 

policies and the use of CI reduced total fuel consumption by 

40%. 

Excludability. Excludability (i.e., the degree to which 

entitled users can keep out free riders) is an important factor 

in the effective operation of an ISC. Excludability is high in P- 

38, as witnessed by the temporary suspension of operations in 

2002 and by the rule of providing one’s own fuel and oil, without 

which one does not have access to water. 

Institutional linkages. Partnership with local government 

is necessary. It reinforces cooperation and discipline among 

members where the village security officers intervened in resolving 

conflicts related to the stealing of water. Moreover, 

through close monitoring and mentoring by NIA of the ISC’s 

operation and maintenance of the system, farmers were able to 

receive guidance and technical support, which helped the P- 

38 ISC to be a stronger cooperative. 

Conclusions 

Controlled irrigation was tested through farmer participatory 

research and development, and found to be a viable technology 

for farmers’ use, as demonstrated by P-38 ISC members. 

Farmers perceived that CI saves significant amounts of water, 

time, labor, and cash, and reduces the costs of rice production. 

Furthermore, CI is perceived to give similar yield and produce 

more tillers, and bigger and heavier grains with good shape. 

Scientific evidence has corroborated the farmers’ perceptions 

on water savings and yield performance. Though CI is suitable 

for adoption by farmers based on their positive perceptions, 

good functioning and cooperation among members of 

an ISC seem to be another requirement for the adoption of CI 


in groundwater irrigation systems. The factors identified for 

successful collective action to facilitate CI’s implementation 

are group size, service area, profitability, high level of excludability, 

strong leadership to deal with free riders, and close 

linkages with local governments and the NIA. 

References 

Belder P, Bouman BAM, Cabangon R, Lu G, Quilang EJP, Li Y, 

Spiertz JHJ, Tuong TP. 2004. Effect of water-saving irrigation 

on rice yield and water use in typical lowland conditions 

in Asia. Agric. Water Manage. 65(3):193-210. 

Bouman BAM, Tuong TP. 2001. Field water management to save 

water and increase its productivity in irrigated lowland rice. 

Agric. Water Manage. 49:11-30. 

Bouman BAM, Tabbal DF, Lampayan RA, Cuyno RV, Quiamco MB, 

Vicmudo VR, Norte TM, Lactaoen AT, Quilang EJP, de Dios 

JL. 2002. Knowledge transfer for water-saving technologies 

in rice production in the Philippines. Proceedings of the 52nd 

Philippine Agricultural Engineering Annual National Convention, 

22-26 April 2002, Puerto Princesa City, Palawan, Philippines. 

p 14-30. 

IRRI (International Rice Research Institute). 1995. Water: a looming 

crisis. Program report. Los Baños (Philippines): International 


Lampayan RM, Bouman BAM, de Dios JL, Lactaoen AT, Quilang 

EJP, Tabbal DF, Llorca LP, Norte TM, Soriano J, Corpuz AA, 

Espiritu AJ, Malasa RB, Vicmudo VR. 2003. Technology transfer 

for water savings (TTWS) in Central Luzon, Philippines: 

results and implications. Proceedings of the 53rd Philippine 

Agricultural Engineering Annual National Convention, 21- 

25 April 2003, Davao, Philippines. p 239-251. 

Tabbal DF, Bouman BAM, Bhuiyan SI, Sibayan EB, Sattar MA. 

2002. On-farm strategies for reducing water input in irrigated 

rice: case studies in the Philippines. Agric. Water Manage. 

56:93-112. 

Tuong TP, Bouman BAM. 2003. Rice production in water-scarce 

environments. In: Kijne JW, Barker R, Molden D, editors. 

Water productivity in agriculture: limits and opportunities for 

improvement. Wallingfond (UK): CABI Publishing. p 53-67. 

Wopereis MCS, Kropff MJ, Maligaya AR, Tuong TP. 1996. Droughtstress 

responses of two lowland rice cultivars to soil water 

status. Field Crops Res. 46:21-39. 

Notes 

Authors’ addresses: F.G. Palis, M. Hossain, and P.A.A. Cenas, Social 

Sciences Division; B.A.M. Bouman and R.M. Lampayan, 

Crop, Soil, and Water Sciences Division, International Rice 

Research Institute (IRRI), DAPO Box 7777, Metro Manila, 

Philippines; A.T. Lactoen, T.M. Norte, and V.R. Vicmudo, 

National Irrigation Administration, Groundwater Irrigation 

System Reactivation Project, Tarlac, Philippines; G.T. Castillo, 

Consultant, IRRI. 

Companion modeling and multi-agent systems 

for collective learning and resource management 

in Asian rice ecosystems 

F. Bousquet and G. Trébuil 

It is widely admitted that poor coordination among stakeholders 

leads to inefficient resource use, economic and environmental 

damage, negative externalities, and social conflict. Diverse 

stakeholders use resources for different purposes, with 

differing perceptions of their dynamics, and adopt various strategies 

to cope with problems. Consequently, the number of social 

conflicts is increasing and they are frequently reported in 

the national and international media, for example, the cases of 

water sharing at rice transplanting among farmers and villages 

in Bhutan, and conflicts over land use between highlanders 

and lowlanders in northern Thailand uplands. To manage these 

problems, new legislative frameworks decentralizing the management 

of renewable resources are being introduced in many 

countries. Their success depends on the quality of the local 

coordination among stakeholders, who often lack tools, methods, 

and trained managers to achieve success. 

Thus, there is a demand for innovative approaches and 

tools to improve coordination processes among an increasing 

number of stakeholders using common resources at the rice 

agroecosystem level. Our hypothesis is that an understanding 

and modeling of the diversity of stakeholders’ perceptions, 

associated with participatory simulation sessions, can be used 

to improve resource management through better coordination 

of people’s actions in any type of rice ecosystem. 

The proposed companion modeling approach 

Usually, models are used to assemble scientific knowledge and 

to propose recommendations to a given decision-maker. In our 

case, we propose to use models to represent the different perceptions 

of various stakeholders in order to facilitate the coordination 

of their actions on a common resource. Some experiments 

were conducted by using geographic information systems 

(GIS) (Abbot et al 1998, Gonzalez 2000), but very few 

of them dealt with the use of simulations (Costanza and Ruth 

1998). The main lesson learned is that these tools, like many 


technological innovations, can both marginalize and empower 

people and communities. 

Companion modeling (ComMod) is a methodology for 

the collective implementation and use of simulation models, 

and, more precisely, multi-agent simulation systems (MAS) 

(Barreteau 2003). The ComMod approach proposes an iterative 

and evolving modeling process: participatory modeling is 

used to facilitate stakeholders’ interactions and to identify and 

scope resource management problems to be discussed and 

negotiated. Then, simulation models are used to collectively 

assess scenarios selected by stakeholders. This may lead to 

new questions, new discussions, changes in the model, and so 

on. 

ComMod combines the use of different tools such as 

agent-based modeling (ABM), GIS, participatory mapping, and 

role-playing games (RPG). From a methodology development 

point of view, the ComMod approach has been tested and used 

in several places, leading to concrete policy recommendations 

or collective actions by communities (Aquino (d’) et al 2002, 

2003, Etienne et al 2003). The use of ComMod implies the 

execution of the following set of activities in an iterative way: 

Framing. Preliminary diagnostic analysis at the system 

level. Identification and understanding of the 

system’s ecological and social dynamics and of key 

issues, concerns, and intervention points with stakeholders. 

Collection of relevant existing data. Identification 

of knowledge gaps to be filled through specific 

surveys. 

Prioritization and visioning. Participatory selection 

of a key concrete problem to be examined. Establishment 

of a common vision shared by all key stakeholders 

through RPG and participatory simulation 

workshops. Delimitation of initial areas of agreement, 

disagreement, uncertainty, and room for coordination 

and negotiation. Production of qualitative guidelines 

for monitoring and evaluation of the subsequent 

ComMod activities. 

Participatory field work and modeling. Implementation 

of an iterative, integrated, flexible, and userfriendly 

modeling approach combining participatory 

workshops and laboratory work. Joint validation of 

the simulation tool with all concerned stakeholders, 

followed by participatory identification of relevant 

resource management scenarios to be simulated and 

assessed collectively, taking into account the competing 

uses of resources by multiple users. 

Collective exploration and discussion of trade-offs 

displayed by the simulated scenarios. Choice and design 

of an action plan to be implemented to tackle 

the resource management problem under consideration. 

Assessment of local impact. Assessment of the local 

impact of the approach for participatory and integrated 

renewable resource management (IRRM). 

Case studies 

For the past three years, the ComMod approach has been used 

to examine various resource management problems in different 

Asian rice ecosystems (Bousquet et al, in press). 

Water sharing at rice transplanting 

in Lingmuteychu, Bhutan 

Lingmuteychu is a catchment covering 34 km 2 in west-central 

Bhutan, which is drained by a totally rainfed stream originating 

from a rock face at an altitude of 2,400 m. Five small irrigation 

systems formed of 12 canals irrigate about 200 ha of 

terraced paddies belonging to 121 households of six villages. 

These villages share irrigation water within a broadly respected 

customary regime evolved during a time when demands were 

lower. Under the current processes of market integration, decentralization, 

and resource conservation policies, and changes 

in villagers’ social needs, this customary water-sharing set of 

rules is not adapted to current farming conditions anymore and 

causes repetitive social conflicts, particularly at rice transplanting, 

as these conflicts remain unresolved. There is a definite 

contrast in perceptions of users on the water resource: some 

consider it an infinite resource, many consider this resource as 

an exclusively common pool with a free-access regime, whereas 

the state considers it as its property. The Draft National Water 

Policy pronounces water as a state property while it emphasizes 

integrated management. Under the new national community-based 

NRM policy (CB-NRM), this watershed has been 

selected as a pilot site to improve coordination among water 

users. It is located near the Bajo Renewable Natural Resources 

Research Center (RNR-RC), which is leading the national effort 

in the field of CB-NRM, and the Natural Resource Training 

Institute (NRTI) at Lobeysa, the principal institution for 

higher education in agriculture and resource management in 

the kingdom. Participatory land-use planning and rural appraisal 

activities were carried out by the RNR-RC Bajo team 

at this site. More recently, an analysis of existing water dynamics, 

water-sharing arrangements, and farming practices has 

been implemented and a first participatory workshop was held 

in the two upper villages (where the conflict over water use is 

acute) in May 2003 to test the proposed ComMod methodology 

for supporting farmers from the two villages to examine 

collectively the problem of water exchange at rice transplanting 

(Gurung, in press). Based on the successful outputs, another 

participatory workshop was held in December 2003 to 

modify water-sharing rules. Figure 1 shows the assessment of 

the ComMod process on the perceptions of the water-sharing 

issue by the local stakeholders who took part in both workshops. 

By using a MAS model reproducing the RPG played 

and validated by the farmers, simulations were carried out later 

to examine the relative effects of social networks, communication 

protocols, and climate conditions on water-use efficiency. 

Results indicate that communication protocols constitute 

the most sensitive factor. 


Respondents (%) 

80 

70 

60 

50 

40 

30 

20 

10 

0 

May 2003 

Dec 2003 

Farm 

activities 

Valuation 

of water 

Water 

shortage 

Share 

water 

Lessons learned 

Income 

Canal 

management 

Control 

water 

use 

Fig. 1. Lessons learned by the stakeholders after two participatory workshops. 

Soil and water conservation in diversifying highlands 

of upper northern Thailand 

Mae Salaep is a highland village in Mae Fah Luang District of 

Chiang Rai Province, where small-scale poor Akha farmers 

are being rapidly integrated into the market economy. While 

their former agrarian system based on swiddening is being replaced 

by semipermanent agriculture on steep slopes, the risk 

of increased land degradation through soil erosion by concentrated 

runoff is a major problem. The diversity of farmers (economic 

status, agricultural practices, etc.) is already extensive, 

and their economic and institutional environment is becoming 

more complex (Trébuil et al 2002). An increasing number of 

individual or collective stakeholders with differing land- and 

water-use strategies interact in the dynamics of diversifying 

sloping-land agriculture. In collaboration with Chiang Mai 

University, the Department of Public Welfare (a government 

development agency looking after highland ethnic minorities), 

and a local NGO, a spatially explicit multi-agent model linked 

to a small GIS was built to represent the key interacting ecological 

and agronomic dynamics (slope characteristics, rainfall, 

main cropping systems and succession of practices, farmer 

differentiation, etc.) to assess the risk of soil degradation. This 

MAS model was later simplified and translated into an RPG 

to be used with farmers to validate the understanding of the 

agricultural dynamics by the research team. In December 2002 

and May 2004, two participatory modeling workshops, focusing 

on land-use changes (particularly the transition from annual 

crops to perennial plantations) and the organization of 

the credit system at the village level, respectively, were held 

with representatives from all categories of farmers and development 

agencies. From the first to the second workshop, at 

the request of the participants, the emphasis shifted from 

achieving a shared representation of the IRRM problem and 

its causes to the use of ComMod for improving the allocation 

of new village funds to improve the current situation and especially 

the access of small farmers to plantation crops to limit 

erosion risk. MAS simulations were used to examine the respective 

influence of formal and informal credit on the welfare 

of different kinds of stakeholders (Barnaud 2004). 

Conclusions 

The ComMod approach seems to be well received by both 

scientists working in the field of IRRM and local stakeholders. 

It is important to distinguish between the use of this approach 

in two specific contexts: (1) to produce knowledge on 

complex rice systems and (2) to support evolving, iterative, 

and continuous collective decision-making processes for 

IRRM. Where the land policy favors a decentralized management 

of resources, such a MAS-based companion modeling 

approach has great potential to improve the collective management 

of rice lands. It can be used to facilitate dialogue, to 

mitigate conflicts, and to establish coordination mechanisms 

regarding multiple uses of the land by multiple stakeholders. 

It is also a powerful tool to integrate knowledge from different 

disciplines, sources, and levels of organization. ComMod facilitates 

the collective assessment of desirable scenarios and 

identification of suitable innovations to overcome current 

IRRM problems in rice ecosystems. 

References 

Abbot J, Chambers R, et al. 1998. Participatory GIS: opportunity or 

oxymoron PLA Notes 33:27-34. 

Aquino (d’) P, Le Page C, et al. 2002. A novel mediating participatory 

modelling: the ‘self-design’ process to accompany collective 

decision making. Int. J. Agric. Resources Governance 

Ecol. 2(1):59-74. 

Aquino (d’) P, Le Page C, et al. 2003. Using self-designed roleplaying 

games and a multi-agent system to empower a local 

decision-making process for land use management: the 

SelfCormas experiment in Senegal. J. Artif. Societ. Soc. Simul. 

6(3). 


Barnaud C. 2004. Erosion des sols et systèmes agraires dens les hautes 

terres de la Thailandes: une approche de la complexité par 

une modélisation d’accompagnement. Master Thesis, Mémoire 

de DEA, Géographie et pratique du développement. Paris, 

Université de Paris X-Nanterrep. 

Barreteau O. 2003. Our companion modelling. J. Artif. Societ. Soc. 

Simul. 6(1). 

Bousquet F, Trébuil G, Hardy B, editors. In press. Companion modeling 

and multi-agent systems for integrated natural resource 

management in Asia. Los Baños (Philippines): International 


Costanza R, Ruth M. 1998. Using dynamic modeling to scope environmental 

problems and build consensus. Environ. Manage. 

22(2):183-195. 

Etienne M, Le Page C, et al. 2003. A step-by-step approach to building 

land management scenarios based on multiple viewpoints 

on multi-agent system simulations. J. Artif. Societ. Soc. Simul. 

6(2). 

Gonzalez R. 2000. Platforms and terraces: bridging participation and 

GIS in joint learning for watershed management with the 

Ifugaos of the Philippines. PhD thesis. ITC-Wageningen University. 

186 p. 

Gurung T. In press. Companion modeling to examine water-sharing 

arrangements among rice-growing villages in west-central 

Bhutan: preliminary results. In: Bousquet F, Trébuil G, Hardy 

B, editors. Companion modeling and multi-agent systems for 

integrated natural resource management in Asia. Los Baños 


Trébuil G, Shinawatra-Ekasingh B, Bousquet F, Thong-Ngam C. 

2002. Multi-agent systems companion modeling for integrated 

watershed management: a northern Thailand experience. In: 

Jianchu X, Mikesell S, editors. Landscapes of diversity. 

Yunnan Science and Technology Press, China. Proceedings 

of the 3rd International Conference on Montane Mainland 

Southeast Asia (MMSEA 3), Lijiang, Yunnan, China, 25-28 

August 2002. p 349-358. At www.cbik.ac.cn/cbik/resource/ 

MMSEA_Index.asp, accessed 8 September 2003. 

Notes 

Authors’ addresses: F. Bousquet, IRRI-Cirad-DOA, Bangkok, Thailand, 

e-mail: f.bousquet@cgiar.org; G. Trébuil, IRRI-Cirad- 

DOA Project, Bangkok, Thailand, e-mail: guy.trebuil@cirad.fr. 

Participatory approaches for improving rice breeding 

in the Mekong Delta of Vietnam 

Nguyen Ngoc De and Kotaro Ohara 

Most conventional breeding programs have been set up and 

designed by breeders, neglecting the role of users: farmers and 

farming communities. Therefore, the dissemination process of 

so-called “technology transfer” was very slow and costly for 

both breeders and farmers. 

The use of participatory approaches assures the involvement 

of farmers at different levels in the whole process of crop 

improvement to overcome the shortcomings of conventional 

breeding approaches. 

Can Tho University, as the leading research institution 

for adapting the participatory approach in rice improvement, 

started on-farm breeding programs as early as 1975 by sending 

out its staff and students to work closely with farmers for 

local crop improvement (Xuan et al 1993). Later, with the inception 

of the Community Biodiversity Development and Conservation 

(CBDC) project in 1994, and the global in situ conservation 

of biodiversity project implemented by the International 

Plant Genetic Resources Institute (IPGRI) in 1998, participatory 

plant breeding (PPB) and participatory variety selection 

(PVS) approaches have been introduced to develop crop 

varieties specific to niche environments and farmers’ preferences. 

This paper reviews the achievements, problems, and lessons 

learned from these practices using new approaches in the 

Mekong Delta of Vietnam. 

Methods used in participatory crop improvement 

Witcombe and Joshi (1996) defined PPB as involving farmers 

in selecting genotypes from genetically variable segregating 

materials and PVS as involving the selection by farmers of 

nonsegregating characterized products from plant breeding 

programs. In practice, the use of PVS and/or PPB depends on 

farmers’ varietal needs and farmers’ breeding knowledge and 

technical skills. The PVS approach has been used to improve 

local landraces (crop cultivars/varieties that adapted and have 

been kept for a long time in local areas) and to evaluate the 

finished breeding materials from research institutions. When 

varietal options available to farmers through PVS are limited 

or exhausted, PPB begins. 

A successful PPB program should involve farmers as 

much as possible in the whole process of plant breeding (De 

2000). 

Participatory plant breeding involves the following steps 

and activities. 

Needs assessment and community organization 

Community meetings are organized to identify farmers’ problems 

and needs. By discussing with farmers and local officials, 

cooperating farmers are divided into three groups depending 

on their technical knowledge and skill. 


Farmer participation 

Collection 


Thailand 

Genebank 

Collaboration 

* 

* 

** 

Characterization 

Evaluation 

Conservation 

Use 

Breeding 

program 

Can Tho University 

Seed request 

IRTP/INGER 

Selection 

coding as 

Coordinating 

Receiving 

Selection 

coding as 

“L” lines 

* 

Preliminary observation 

“IR” lines 

* 

Replicated observation 

CBDC and IPGRI 

projects 

*** 

Coordinating 

Yield trial 

Naming 

as 

“MTL” lines 

**** 

Adaptation field test 

**** 

Seed multiplication 

**** 

New variety released 

Fig. 1. The participatory rice varietal improvement program in the Farming Systems Research 

and Development Institute, Can Tho University, Vietnam. MTL (Mien Tay Lua) = 

crossed, selected, and released by Farming Systems Institute; CBDC = community 

biodiversity development and conservation; IPGRI = International Plant Genetic Resources 

Institute with on-farm conservation project; farmer participation level = from little (*) to full 

(****) participation. 

 

 

 

Group 1 (2–3 farmers) will take responsibility for 

making crosses and selecting segregating breeding 

materials (PPB activities). 

Group 2 (5–10 farmers) will field-test stable lines and 

promising varieties (PVS activities). 

Group 3 (30–40 farmers) could be involved in seed 

multiplication and distribution of selected varieties 

for local use. 

Setting breeding objectives 

and identifying donor parents 

Breeders work closely with farmers to identify breeding objectives. 

Some examples of farmers’ breeding criteria for rice 

are high yield, short duration, resistance to major pests and 

diseases, good eating quality, and so on. Based on the breeding 

objectives, breeders then assist farmers in searching for 

suitable donor parents. These donors could be found among 

the available genetic materials at the local level or from research 

institutions. 

The level of farmers’ participation in the rice breeding 

process is described in Figure 1. 

Conducting field activities 

1. Participatory plant breeding (PPB) 

Making crosses. Group 1 farmers are given additional 

training on breeding and are assisted by 


Results 

 

breeders in making desirable crosses to meet local 

breeding objectives. 

Selecting from segregating materials. Early generations 

will be evaluated and selected together 

with segregating lines provided by research institutions. 

This process is repeated until stable 

lines are obtained. Farmers with adequate training 

can practice mass selection and pure-line 

selection. 

2. Participatory varietal selection (PVS) 

Genetic improvement of local adapted varieties. 

Farmers can improve their local landraces 

or common varieties for better genetic purity, 

higher yield, better quality, and resistance to 

major pests by pureline selection or mass selection 

methods. Sometimes, the local varieties collected 

from the community or other communities 

are reintroduced to the community after seed 

loss caused by a disaster. In addition, new varieties 

from research institutions are provided for 

testing together with finished products from PPB 

activities. 

Varietal testing for local adaptation. Stable lines 

(20–40) selected from the segregating materials 

or provided by research institutions are planted 

in an observation test with common local varieties 

as local checks. The promising varieties (5– 

10) selected from the observation test are then 

tested in larger plot sizes for a yield trial. Farmer 

field days (activities to assemble local farmers 

at demonstration plots for direct observation, 

evaluation, and interaction) are organized just 

before harvesting for a joint evaluation. 

Seed multiplication. Desirable varieties (usually 

2–3) from yield trials are then selected for seed 

multiplication by a larger group of farmers to 

use in the community. 

3. Monitoring and evaluation 

Farmers closely monitor and keep records on field conditions 

and crop performance for later analysis to determine 

suitable varieties. A major data set consists of 

growth duration, plant height, tillering capacity, grain 

yield, grain quality, and tolerance ability for major insects 

and diseases. Field visits and farmer field days involving 

breeders, extension workers, and farmers are 

the most appropriate tools for monitoring and evaluating 

PPB/PVS activities. 

Participatory varietal selection 

From 1975 to 2003, hundreds of promising rice varieties were 

tested in farmers’ fields under farmers’ management, and many 

varieties were identified and released as national recognized 

varieties using the participatory approach. Of these, some famous 

rice varieties are IR36 (named as NN3A), MTL2 

(NN6A), MTL30 (NN7A), MTL9 (NN2B), IR42 (NN4B), 

MTL58 (IR13240-108-2-2-3), and MTL87 (IR50404-57-2-2- 

3). These varieties have made a great contribution to the improvement 

of rice production in the Mekong Delta. Many farmers, 

such as Hai Huu (Long An Province), Hai Chung, Tu Tai, 

and Ba Chuong (Tien Giang Province), Ba Cung and Hai Triem 

(An Giang Province), Muoi Tuoc and Muoi Than Nong (Vinh 

Long Province), and many others are known as “rice selection 

kings.” Farmers are not only selecting suitable finished varieties 

but also improving the varieties formally released for better 

grain quality and adaptation to specific conditions in their 

areas by pure-line selection. This process increases both crop 

productivity and local crop diversity. 

Depending on local resource availability and management 

capacity, each community has received 7–30 varieties 

for testing and 3–4 varieties have been selected annually for 

seed multiplication and use (Table 1). 

Participatory plant breeding 

In the 1996-97 dry season, Can Tho University started providing 

63 segregating F 2 -F 3 populations of 12 crosses for four 

pilot communities. 

In 1998-99, L246-7-3-B and L247-1-5-B, the two promising 

farmer selections noted as SiC-1 (Soc Trang Selection, 

no. 1) and SiC-2 (Soc Trang Selection, no. 2), respectively, 

were purified by Ke Sach community (Soc Trang Province) 

using the bulk selection method. Mr. Canh (a farmer group 

leader) led the selection activities. Similarly, L246-10-1-B, a 

promising line selected by farmers in My Thanh community 

(Ba Tri District, Ben Tre Province), is also undergoing a yield 

test and seed multiplication. In 2003, Mr. Thanh in Cho Moi 

District (An Giang Province) selected two stable lines, TH1 

and TH2, from segregating breeding material of Can Tho University 

in a similar way. 

Discussion 

With technical training and field assistance, farmers can manage 

their own breeding programs. The total number of rice 

varieties in the communities is increasing through the introduction 

of new varieties and locally improved cultivars. As a 

result, land productivity and farm income have been improving. 

Problems 

Some problems have occurred. 

Farmers are more willing to multiply promising varieties 

(PVS) than to select from segregating materials 

or make crosses (PPB) because of the time and resources 

needed and the requirement of skills and intellectual 

input. 

The low education level of farmers limits the adoption 

of PPB; therefore, more training and field coaching 

are required. 


Table 1. Numbers of rice varieties tested and selected for local use by farming 

communities since 1994. 

Year Varieties tested Varieties selected Seed multiplication (t) 

Total Each Total Each Total Each 

community community community 

1994 20 6.7 5 1.6 3.2 0.8 

1995 197 39.4 25 4.2 4.3 0.9 

1996 182 26.0 22 3.1 12.6 2.1 

1997 58 7.3 34 3.4 18.3 2.3 

1998 78 8.7 30 3.0 35.6 3.9 

1999 90 8.2 36 3.6 39.2 3.9 

2000 131 9.4 44 3.7 121.4 9.3 

2001 192 8.7 66 3.5 316.3 16.6 

2002 280 10.8 107 4.7 512.4 21.3 

2003 275 12.0 103 4.5 647.7 28.2 

Av 150.3 13.7 47.2 3.5 171.1 8.9 

Source: Data from field survey by authors, 2003. 

 

Morris (2003) also pointed out three challenges of 

participatory breeding: 

— There is a need to develop varietal evaluation 

methods capable of generating credible data for 

widespread acceptability. 

— A level of participation that will ensure equitable 

compensation for participating farmers is needed. 

— National and international regulatory frameworks 

that govern the evaluation, approval, and release 

of new plant varieties from PPB are necessary. 

Lessons learned 

The following lessons can be drawn from these activities: 

Farmers conserve and improve biodiversity to meet 

their needs for home consumption, the market 

economy, and adaptation to local environments and 

farm resources. Therefore, farmers’ needs and perceptions 

should be taken into consideration to bring 

breeders’ objectives closer to farmers’ objectives. 

Support from local authorities and organizations in 

terms of organization, management, additional funds, 

and facilitation is very important to ensure a successful 

participatory approach. 

Cooperation with a group/community on PPB/PVS 

turns out better than that with individual farmers because 

group/community members can help each other. 

Farmer field schools and farmer field days are good 

ways to stimulate farmers’ participation and to train 

farmers with a lower education. 

Conclusions 

Participatory approaches could improve breeding for local 

adaptation, local genetic diversity, efficiency, and the empowerment 

of rural communities. The achievements in these activities 

have had a strong impact on national seed policies, 

which are under review and reform in Vietnam. However, a 

successful participatory approach in crop improvement requires 

a willingness to contribute time, land, labor, and other inputs, 

and to incur possible risk. It requires technical knowledge and 

skill from farmers, and a decentralized breeding strategy and 

technical assistance from breeders. It also requires favorable 

policies and support in organization and management. 

References 

De Nguyen Ngoc. 2000. Linking the national genebank of Vietnam 

and farmers. In: Friis-Hansen E, Sthapit B, editors. Participatory 

approaches to the conservation and use of plant genetic 

resources. Rome (Italy): International Plant Genetic Resources 


Morris ML, Bellon MR. 2003. Participatory plant breeding research: 

opportunities and challenges for the international crop improvement 

system. Euphytica (2003):1-15. 

Witcombe J, Joshi A. 1996. Farmer participatory approaches for 

varietal breeding and selection and linkages to the formal seed 

sector. In: Participatory plant breeding. Proceedings of a workshop 

on participatory plant breeding, 26-29 July 1995. 

Wageningen, Netherlands. p 57-65. 

Xuan Vo-Tong et al. 1993. Present status of agricultural extension in 

Vietnam. Paper presented at the first Southeast Asia workshop 

on formulation of project proposals on technology transfer 

for major food crop production, FAO and UAF, Ho Chi 

Minh City, Vietnam, 6-9 Dec. 1993. 

Notes 

Authors’ addresses: Nguyen Ngoc De, Mekong Delta Farming Systems 

Research and Development Institute, Can Tho University, 

Vietnam; Kotaro Ohara, Department of Sustainable Resource 

Sciences, Faculty of Bio-Resources, MIE University, 

Japan, e-mail: nnde@ctu.edu.vn, ohara@bio.mie-u.ac.jp. 


IRRI’s approach to participatory research for development: 

advances and limitations 

Thelma R. Paris and M. Zainul Abedin 

Why do participatory research While the traditional linear 

and top-down research-extension-farmer transfer of technology 

(TOT) model was successful in disseminating rice technologies 

for favorable rice environments, this approach no 

longer holds for rice research in unfavorable environments. In 

the traditional top-down approach, technologies developed 

were packaged for delivery to farmers and experiments with 

component rice technologies were conducted on-station. Despite 

the many available technologies generated at the station, 

the traditional research approach has had limited success in 

unfavorable rice environments, which have high biophysical, 

socioeconomic, and cultural diversity, and poor access to markets 

and infrastructure facilities. This created a mismatch between 

research recommendations and farmers’ needs and expectations 

and thus low or no adoption of technologies. Crop 

and natural resource management technologies require an 

analysis of how farmers make decisions, develop a corrective 

heuristic, frame it as a hypothesis, and become motivated to 

participate in an experiment to test it. Similarly, germplasm 

improvement requires an understanding of farmers’ preferences 

and criteria for rice varietal selection in fragile environments. 

We need to understand what “enabling inputs” may be required 

to accelerate and widen adoption. These requirements necessitate 

the use of participatory research approaches through 

which researchers and farmers can work as active partners in 

developing suitable technologies suited to their livelihood and 

ecosystems. Such active partnership creates synergistic effects 

as it integrates the knowledge, experience, and skills of farmers 

and scientists. 

What is participatory research 

“Participation” and “participatory” have become such fashionable 

terms recently that any kind of activity involving a 

group of people is termed “participatory.” As these terms embrace 

a multitude of meanings, and these meanings become 

correspondingly dilute, a serious threat is posed to the use of 

the term “participatory research.” The risk is that a catch-all 

definition of participatory research is destined to fall out of 

fashion and to be discarded as fashion changes, without even 

receiving the serious scientific evaluation of its potential that 

a rigorous but less trendy use of the term would invite. The 

term “participatory research” is a collection of approaches that 

enable participants to develop their own understanding of and 

control over the processes and events being investigated. It is 

loosely used to describe various types and levels of local involvement 

in and control of the research process (Ashby 1997). 

Participatory research has diverse approaches, perspectives, 

practices, and methods. At IRRI, participatory research 

differs according to various interpretations of the mode of research 

partnership or “participation” in research, the goal or 

rationale for encouraging participation, gender and stages in 

the project, the level of “control” or “ownership” that local 

people have over the research process, the scale of participatory 

activities and stakeholder involvement and the levels of 

management involved, and the level of disaggregation and representation 

of different stakeholders required for the research. 

Clearly, there is no formula for deciding which level of 

participation is the best. The level chosen will depend on the 

objectives of the activity. What is more important is to have 

well-defined research problems and well-defined technologies 

for a vigorous process whether with farmers or other stakeholders. 

It is important, however, to distinguish which of these 

levels of farmer or community participation we refer to when 

research is called “participatory.” 

Advances in participatory research at IRRI 

Through the years, progress has been made in participatory 

research at IRRI. Biological scientists in close interaction with 

social scientists have tested, refined, and adapted protocols 

for farmer/community participatory research in natural resource 

management (NRM) and participatory plant breeding (PPB). 

Participatory research (PR) has fast-tracked the adoption of 

innovations and in several cases has contributed to adoption 

on a large scale. Working with national agricultural research 

and extension systems (NARES) on projects using the PR approach 

has changed the attitudes, mind-sets, and research orientation 

of scientists and research managers. The specific advances 

in PR at IRRI in natural resource management and rice 

germplasm development are described below. 

Natural resource management 

In integrated pest management (IPM) in Vietnam, farmers 

“learn by doing” and decision rules are modified on the basis 

of direct experience. A radical change in the research sequence 

of IPM was made by starting with the farmers’ perspective 

during the problem definition or initial problem diagnosis. This 

approach provided a mechanism for scientists to learn about 

farmers’ decision constraints, determine research needs, use 

research information, and “distill” these into testable hypotheses 

for farmers (Heong and Escalada 2003, Huan et al 2004). 

In Bangladesh and in West Bengal, India, the use of the community 

participatory approach to research (CPAR) to evaluate 


the leaf color chart (LCC) for nitrogen management and a plastic 

drum seeder for direct wed seeding of rice facilitated adoption 

on a large scale (Bagchi et al, this volume). The formation 

of small groups of participating farmers around each LCC or 

drum seeder facilitated continued monitoring by participating 

farmers, which led to better decision making and adoption 

(Abedin 2004). 

Participatory varietal selection (PVS) 

Farmer participation in varietal selection improved the selection 

of suitable varieties for complex rainfed environments 

because farmers were given the opportunity to screen new varieties 

in their specific environments. Farmer-managed trials 

provided important information about the performance, quality, 

and overall acceptability of new rainfed rice varieties (Atlin 

et al 2002a, Gregorio et al 2004). Moreover, farmers’ selection 

criteria from different socioeconomic groups, including 

women, are now well understood by breeders (Paris et al 2001, 

2002). Farmer-preferred lines through PVS are spreading fast 

through farmer-to-farmer exchange. Lines selected from PVS 

are now being considered for national/formal release in eastern 

India (Paris et al 2002) and in Bangladesh. More and more 

plant breeders now practice PVS as a normal routine of the 

breeding process. PVS is being institutionalized with the Lao 

breeding program. PVS in the Thai national program is proceeding 

rapidly. Key ingredients from senior Thai breeders, 

the availability of training and methodology documentation 

from IRRI, and initial support from an external donor have 

contributed to this initial success (Atlin et al 2002b). 

Capacity development and institutionalization 

Several training workshops were conducted to enhance the 

capabilities of NARES partners in various aspects of PR in 

NRM and PVS. IRRI has incorporated PR in the curriculum 

of the new rice plant breeding course held every year. In-country 

training on PR and socioeconomic analysis has been conducted 

by IRRI and CIMMYT and will continue to further 

enhance the skills of NARES collaborators. 

Limitations of participatory research 

 

 

 

Although using PR has many advantages, some limitations need 

to be overcome. These are discussed below. 

Poor understanding of PR. The process of becoming 

a “participatory” researcher entails a holistic change 

in knowledge, attitude, and practice. Participatory research 

is often seen as a threat to classic research 

paradigms and not so much as complementary research. 

Moreover, the core capacities required for 

participatory research are often seen as being incompatible 

with dominant research norms and practices. 

There is some diversity regarding the understanding 

of demand-driven, client-oriented, or participatory 

research approaches by senior management. Its basic 

principles and strategic dimensions are not well understood 

by all. Commodity orientation, which still 

prevails, hinders a systems approach to understanding 

farmers’ circumstances under constrained environments. 

Traditional research sees the research findings 

or outputs, whereas PR views livelihood and 

positive environmental change as the product. 

Lack of support from management. Not enough senior 

researchers have experience in participatory research 

at research centers. Most researchers working 

with participatory approaches are young, are on “soft” 

money, and don’t have enough possibilities to stay. 

Problems with continuity and quality are the consequence. 

Turnover of scientists is high and this limits 

the building of social capital with farmers. The incentives 

or rewards for scientists to do farmer participatory 

research in the Consultative Group on International 

Agricultural Research (CGIAR) and 

NARES is still very much based on the production of 

data and number of scientific publications instead of 

on impact and process results. Researchers have very 

little incentive to do participatory research, with the 

risk of becoming marginalized within the scientific 

community. 

Lack of staff with training and experience on PR. Very 

few researchers (scientists and NARES) are trained 

on farmer participatory research and facilitating skills 

on PVS and NRM (Linquist et al 2004, Singh et al 

2002). 

Availability of social scientists. Social scientists are 

still a very marginal group in CGIAR centers and are 

not available in the agricultural research institutions 

that are collaborating with IRRI. In this small group, 

most social scientists are economists, leaving a 

vacuum for other pressing social science issues. Further, 

the few social scientists available are located in 

nonagriculture-based universities and logistical arrangements 

for field work are a problem. 

Necessary conditions to sustain/institutionalize farmer 

participatory research 

 

 

At the farm/community level. Farmers themselves 

must be interested in the problem or issue, while scientists 

should have knowledge and the right attitudes 

and skills. There should be NGOs or similar partners 

in or near the village to monitor experiments and assist 

farmers. There should be clear objectives and protocols 

that all stakeholders agree on. 

Research. Participatory research requires both institutional 

and personal commitments on a sustainable 

basis. There is a need to change the mind-sets of the 

scientists and research administrators engaged in onfarm 

trials. A strong interaction between social scientists 

and biological scientists with NARES is essential. 

Institutional arrangements need to be given due 

importance. 


At the institutional level. The integration of PR approaches 

into the main research and training programs 

at IRRI should be supported by management. IRRI’s 

comparative advantage lies in the application of participatory 

research to strategic and pre-adaptive research 

such as participatory research methodologies 

for use by NARES, NGOs, and others and in the management 

of natural resources, plant breeding for 

rainfed environments, IPM, geographic information 

systems, decision support tools for soil management 

and land-use planning, validation of technologies for 

rainfed environments, etc. There is a need to include 

farmer participatory research methods in IRRI’s and 

NARES’ training programs on conventional research. 

There should also be a mechanism to strengthen linkages 

and partnerships with NGOs and extension institutions. 

Donors, which pressure for more farmer 

participation, should make a commitment to the issue 

with a long-term perspective and multiyear funding. 

Conclusions and issues to study 

Not all kinds of research can be participatory and PR is not a 

substitute for traditional or conventional research. Participatory 

research will be more effective if backed by formal research 

or conducted in parallel with on-station research. Farmer 

participatory research is increasingly necessary as one moves 

toward more diverse and complex environments. As one moves 

in this direction, recommendations cannot be broadly applied 

and adaptive research becomes increasingly important. 

Farmer participatory research can be an effective tool to 

motivate farmers to “experiment” with new concepts, innovations, 

ideas, etc. It is by far the most powerful tool for changing 

attitudes and beliefs, especially deeply entrenched attitudes 

(“more seed and fertilizer will give higher yields; all insects 

are bad and must be killed”). 

Farmer participatory research helps breeders consider 

farmers’ needs and preferences in setting their breeding goals 

for a target environment. PVS, as a complement to conventional 

breeding, provides farmers with opportunities to evaluate 

and select the genotypes they want in their own fields under 

their level of management before they are formally released 

in the national systems. 

The appropriate scale and level of representation of different 

interest groups, the methods chosen, and the extent of 

local participation in and control over the research process 

will depend on the project goals, available resources, and scope 

of the research as well as the rationale for using the participatory 

approach. 

Strategies and methods for involving farmers, local 

people, extension people, research administrators, etc., will 

have to be followed on a case-by-case basis and may differ for 

germplasm improvement and NRM. Moreover, more women 

farmers should be included in participatory research because 

410 Rice is life: scientific perspectives for the 21st century 

they are increasingly becoming the de facto farm managers 

and potential users of new technologies. 

While most PR activities are strong in initial diagnosis, 

planning, design, and testing, very few have had impact. Experience 

in scaling up should be documented. 

There is a need to study the impact of participatory research 

and assess other enabling mechanisms such as matching 

funds to support partnerships among projects as well as 

resolutions, regulations (such as for the varietal release system), 

and official instructions given by departments and local 

governments. 

We also need to reflect on the following issues: 

To what extent should international research organizations 

and NARES be involved in scaling up to accelerate 

impact (Methodologies) 

What approaches to scaling up work and which do 

not (Methodologies) 

Do international research organizations have the 

“warm bodies” to do this with a reduction in nationally 

recruited staff and budget reductions (Institutionalization) 

How can we unite/complement formal traditional/conventional 

research and informal science (participatory 

research) (Research) 

How can international research organizations facilitate 

mainstreaming or institutionalizing participatory 

research by NARES How can the practice of PR be 

brought out of donor-driven project activities (Institutionalization) 

References 

Abedin Z. 2004. The emerging face of rice production in Bangladesh: 

direct wet seeding of rice using a plastic drum seeder to make 

rice production more profitable. Paper presented at the International 

Symposium on Rainfed Rice Ecosystems: Perspectives 

and Potential, held at Indira Gandhi Agricultural University, 

Raipur, Chhattisgarh, India, 11-13 October 2004. 

Ashby J. 1997. What do we mean by participatory research in agriculture 

In: New frontiers in participatory research and gender 

analysis for technology development: proceedings. International 

Seminar on Participatory Research and Gender Analysis 

for Technology Development, 1996. Cali (Colombia): 

Centro Internacional de Agricultura Tropical. p 15-22. 

Atlin G, Paris T, Courtois B. 2002a. Sources of variation in participatory 

varietal selection trials with rainfed rice: implications 

for the design of mother-baby trial networks. In: Bellon MR, 

Reeves J, editors. 2002. Quantitative analysis of data from 

participatory methods in plant breeding. Mexico, D.F.: Centro 

Internacional de Mejoramiento de Maíz y Trigo. p 36-43. 

Atlin G, Paris TR, Linquist B, Phengchang S, Chongyikangutor K, 

Singh A, Singh VN, Dwivedi JL, Pandey S, Cenas P, Laza M, 

Sinha PK, Mandal NP, Suwarno. 2002b. Integrating conventional 

and participatory crop improvement in rainfed rice. In: 

Whitcombe JR, Parr LB, Atlin GN, editors. Breeding rainfed 

rice for drought-prone environments: integrating conventional 

and participatory plant breeding in South and Southeast Asia. 

Proceedings of a DFID Plant Sciences Research Programme/ 

IRRI Conference, 12-15 March 2002, International Rice Research 

Institute, Los Baños, Laguna, Philippines. p 36-39.

Gregorio GB, Salam MA, Karim NH, Seraj ZI. 2004. Evaluation 

report of sub-project on development of high-yielding varieties 

for coastal wetlands of Bangladesh. Dhaka (Bangladesh): 


Heong KL, Escalada M. 2003. Farmer participatory experiments in 

pest management. In: Pound B, Snapp S, McDougall C, Braun 

A, editors. Managing natural resources for sustainable livelihoods: 

uniting science and participation. London (UK): 

Earthscan Publications Ltd., and Ottawa (Canada): International 

Development Research Centre. p 210-212. 

Huan NH, Thien LV, Chien HV, Heong KL. 2004. Farmers’ participatory 

evaluation of reducing pesticides, fertilizers and seed 

rates in rice farming in the Mekong Delta, Vietnam. Crop Prot. 

(In press.) 

Linquist B, Bounthanh K, Houemchitsavak S, Horne P. 2004. Upland 

research in Laos: experiences with participatory research 

approaches. In: Participatory research and development. Vol. 

3. CIP/UPWARD, Philippines. (In press.) 

Paris TR, Singh A, Luis J. 2001. Listening to male and female farmers 

in rice varietal selection: a case in eastern India. In: CIAT 

2000: an exchange of experiences from South and Southeast 

Asia. Proceedings of the International Symposium on Participatory 

Plant Breeding and Participatory Plant Genetic Resource 

Enhancement, Pokhara, Nepal, 1-5 May 2000. Cali 

(Colombia): Centro Internacional de Agricultura Tropical. 

Paris TR, Singh RK, Atlin GN, Sarkarung S, McLaren G, Courtois 

B, McAllister K, Piggin C, Pandey S, Sngh A, Singh BN, 

Singh ON, Singh S, Singh RK, Mandak NP, Prasad K, Sahu 

RK, Sahu VN, Sharma MK, Singh RKP, Thakur R, Singh 

NK, Chaudhary D, Ram S. 2002. Farmer participatory breeding 

and participatory varietal selection in eastern India: lessons 

learned. In: Whitcombe JR, Parr LB, Atlin GN, editors. 

Breeding rainfed rice for drought-prone environments: integrating 

conventional and participatory plant breeding in South 

and Southeast Asia. Proceedings of a DFID Plant Sciences 

Research Programme/IRRI Conference, 12-15 March 2002, 

International Rice Research Institute, Los Baños, Laguna, Philippines. 

p 32-36. 

Singh RK, Paris T, Thakur R, Singh VP. 2002. Institutionalizing 

participatory plant breeding within the national research systems: 

issues and future challenges. Paper presented at the International 

Symposium on Agronomy Conference in New 

Delhi. Organized by the Department of Extension, Indian 

Council of Agricultural Research (ICAR), New Delhi, India, 

November 2002. 

Notes 

Authors’ address: Social Sciences Division, International Rice Research 

Institute, DAPO Box 7777, Metro Manila, Philippines, 

e-mail: t.paris@cgiar.org. 

A participatory approach for building sustainable ricefarming 

systems in the reclaimed farmland of Ogata, Japan 

Yoshimitsu Taniguchi and Satoru Sato 

We present here a case study using a participatory approach in 

the regional development of sustainable rice-farming systems 

in Ogata village in northeastern Japan. After Mr. Tozawa’s interesting 

report as a farmer, we would like to present, from a 

scientific point of view, several principles of participatory research 

derived from our experience with the development process 

for two contrasting technologies, organic and no-till, in 

Ogata. This presentation is based on the on-farm research we 

have carried out over the past 7 years in collaboration with 

farmers of Ogata. The theme of this research is rural development 

based on sustainable agriculture. As Mr. Tozawa said, 

Ogata is the largest single area of sustainable rice production 

in Japan, and Ogata farmers are very active in using new technologies 

of sustainable farming. These research experiences 

have taught us that participation and interactiveness are the 

keys to success in the development of new sustainable technologies, 

for three reasons. 

First, technologies for sustainability are much more 

knowledge-intensive than conventional chemical-based technologies 

that farmers need much more time to learn, experiment, 

and adapt them before they can decide when and how to 

use them. This can be seen in the difference between organic 

fertilizer and chemical fertilizer. When farmers use chemical 

fertilizer, they have to think only about the necessary quantity 

of fertilizer and just place it on the ground. The fertilizer works 

and dissolves soon. But, when they use organic fertilizer, they 

have to consider a wider variety of factors: how long will it 

take organic fertilizer to dissolve under the ground how much 

of it will accumulate and when and how will it work on plants 

The reason organic fertilizer requires farmers to think more is 

that it works on plants through much more complex interactions 

among soil, water, and microorganisms than chemical 

fertilizer. Organic fertilization is more difficult because it is 

more site-specific than chemical fertilization. The effect of 

fertilizer may differ from one site to another. Thus, organic 

farmers have to have much knowledge of soil science, plant 

nutrition, ecology, etc. No-till farming is also knowledge-intensive 

in a different way from organic farming. Pioneering 

farmers of Ogata have developed new technologies as partial 

technologies of no-till farming. These technologies include a 

special type of transplanting machine for no-till and controlledavailability 

fertilizer for seedlings. Thus, no-till farmers have 

to know much about the relationship between water permeability 

and the nature of soil, the most effective use of con- 


trolled-availability fertilizer, and the cost-reducing effect of 

no-till and others. 

Second, since each sustainable agricultural technology 

is partial (not a total change of the production system), farmers 

need to make mutual adjustments to incorporate a new technology: 

sometimes modifying the technology to make it more 

compatible with their existing farm management, and sometimes 

modifying their management of other components to 

make them more compatible with the new technology. Farmers 

will use new technologies only when they reach the conclusion 

after experimentation that these are useful for their 

overall farm management. For instance, labor and agricultural 

machinery are important factors for mutual adjustments. One 

Ogata farmer doubled his farm size by purchasing extra farmland. 

He said he would be able to manage the farm of 30 ha 

with only labor of his family of four adults. Another Ogata 

farmer has been hesitating to begin no-till farming, mainly 

because he could not decide whether to make a new investment 

by purchasing a special transplanting machine for notill. 

He is afraid that he would not be able to repay the loan if 

the price of rice continues to decrease. 

Third, the innovation process of sustainable farming technologies 

is always accompanied by uncertainties: the effects 

of new technology are not well defined. Technical success does 

not necessarily mean economic rewards or social acceptance; 

short-term results may change existing long-term goals. These 

three types of uncertainty can be seen in the example of both 

organic and no-till farming. In organic farming, organic fertilizer 

is a cause of uncertainties. Its effects may differ according 

to weather conditions, for example. Especially during the period 

of seedlings, organic farmers are extremely sensitive as 

to whether organic fertilizer works properly. Even experienced 

farmers sometimes fail with seedlings using organic fertilizer. 

Then they have to rush to dispose of the failed soil and seeds, 

and to try again from the beginning. With no-till farming, a 

cause of uncertainty lies in the use of herbicide and chemical 

fertilizer. The present technological system of no-till farming 

requires the use of herbicide in early spring and after transplanting, 

and the use of chemical fertilizer as controlled-availability 

fertilizer. This means that rice grown by no-till farming 

is not regarded as “environmentally friendly” in the present 

market. So, farmers cannot expect a premium. As Mr. Tozawa 

said, this is the most important reason why no-till has not spread 

in Ogata. 

To avoid these uncertainties, farmers tend to take a “stepby-step” 

approach based on communication with other agents, 

within the community, in markets, and in the public sector. In 

the above example, as was reported by Mr. Tozawa, “establishing 

a study group and experimenting with new technologies 

with other farmers are the most common ways of solving 

problems.” There are 40 or 50 groups of this kind, and a farmer 

often joins several study groups to learn various technologies 

and materials. It takes time for Ogata farmers to use a new 

technology, but, once they accept it, that technology spreads 

rapidly all over the village. The use of controlled-availability 

fertilizer is a good example. In 1998, it covered 1,214 ha, but 

the latest statistics indicate that it covered more than 3,000 ha 

in 2003. It is also common for other agents to be involved in 

this process. As Mr. Tozawa said, Ogata farmers sometimes 

invite researchers to learn about new technologies and materials. 

In some cases, retailers and consumers participate in technological 

innovation by encouraging farmers and inspecting 

soil conditions and residual pesticide with rice. A participatory 

approach parallels this natural process of farmers, and 

enables research and extension support agents in the public 

sector to interact with this process. 

Notes 

Authors’ addresses: Y. Taniguchi, Department of Bioresource Sciences, 

Akita Prefectural University; S. Sato, Department of 

Bioresource Sciences, Akita Prefectural University, 010-0195 

Nakano, Shimoshinjo, Akita City, Akita, Japan, e-mail: 

tani@akita-pu.ac.jp, ssatoru@akita-pu.ac.jp. 

Rice farmers’ participatory research has played a key role 

in implementing the System of Rice Intensification 

Dandu Jagannadha Raju 

West Godavari District is the rice bowl of Andhra Pradesh 

State and the kingpin of aquaculture in India. Rice farmers 

have enjoyed the fruits of the Green Revolution with the introduction 

of new high-yielding rice varieties starting with IR8. 

They reached a peak and have attained the highest rice yields 

in the state for more than two decades. Later hybrid rice varieties 

did not increase yield. The regular new high-yielding rice 

varieties became a ray of hope but have been disappearing 

because of one problem or another, and continued stagnated 

rice yields frustrated farmers and made them switch to horticultural 

crops and aquaculture. In view of the existing soil and 

climatic conditions, the majority were still forced to adjust to 

rice cultivation and its stagnated yields. The farmers’ entrepreneurial 

behavior motivated them to turn to a new technique 

such as the System of Rice Intensification (SRI). Yet, the initial 

field problems and laborious technological skills involved 

in it have slowed down its introduction. But, to the surprise of 

everyone concerned, including the scientists, the farmers’ ini- 


Fig. 1. Farmers’ participatory research efforts have brought out new line markers and weeders to suit different situations for easy application 

of SRI technology. 

tiative in doing research in association with extension scientists 

with the new SRI technology in farmers’ own fields as 

well as at extension center farms helped to modify existing 

field situations and make them suitable for the new SRI technology. 

This has helped with the successful implementation of 

SRI in India. 

The SRI was first introduced in India during the 2003 

rainy season and it was simultaneously investigated at a university 

extension center such as the Krishi Vigyan Kendra 

(KVK), Undi, West Godavari District, and in farmers’ fields. 

Technologies were prescreened in association with selected 

farmer groups through discussion and interaction sessions. Participation 

of farmer groups was encouraged at different stages 

of crop growth. At the harvesting stage, yield profitability, the 

cost-benefit ratio, practicability, and likes and dislikes of farmers 

were assessed. Field days were conducted in farmers’ fields 

as well as at the extension center. Farmer conventions were 

organized, duly inviting resourceful farmers and those farmers 

who had already adopted the SRI technology. This helped 

with constructive interaction among the farmers and motivated 

others to adopt the technology on a large scale in the next season 

itself. 

Farmers as scientists have undertaken several experiments 

based on their own experiences. They themselves have 

evolved location-specific and cost-effective devices such as 

line markers and weeders. This has encouraged other farmers 

to modify accordingly. The findings reveal that real farmers’ 

participation right from the introduction of technologies has 

made farmers modify technologies according to their situations, 

thus inventing needed implements such as markers. This 

has made it easy to adopt different and wider spacings in rice 

transplanting, and the weeders were easy to operate even in 

critical field situations. The collective approach of farmers 

made it easy to manage water properly. Though at the initial 

stages more laborers were needed for leveling and transplanting 

young seedlings, this gradually became standardized with 

less labor and speedy transplanting. The additional crop yield 

of 20–30% has encouraged farmers to adopt this system on a 

large scale in their fields. 

Farmers locally manufactured small implements such as 

markers and weeders to suit their field conditions. They have 

made several modifications and developed the best of them 

with their own efforts, such as ropes with beads, ropes with 

knots, bamboo sticks with nails, fixed line markers, adjustable 

markers, and, ultimately, the multipurpose liner, which is now 

mostly used by farmers because it easily marks lines, both vertical 

and horizontal, at the same time with less labor. Similarly, 

weeders were locally manufactured, such as cono weeders 

and rotary weeders with bearings, and a weeder with a mesh, 

etc. (Fig. 1). 

The farmers also struggled for effective water management 

practices in the delta region, which is a crucial step of 

SRI cultivation under delta situations. These management practices 

included making a small pit in one corner of the main 

rice plots and pumping out the water, making inner new and 

small bunds and deep alleyways at a 2-m interval, etc. The 

farmers also experimented with the SRI method using various 

varieties in order to find out the most suitable ones (Fig. 2). 

Farmer participatory research 

An innovative and enthusiastic farmer, Mr. B. Sudhakar Reddy 

of Akividu village of West Godavari District, is known for his 

realistic approach to rice cultivation. He successfully implemented 

the newly introduced SRI technique on 10 ha during 

the kharif and rabi 2003-04 seasons with varied combinations 

of different varieties and spacings, with careful water and weed 

management practices. He achieved higher yields with all varieties. 

He has proved that the SRI technology can easily be 

implemented in any field situation. He has obtained record 

yields as well as formed rice plots as rice gardens by using 

wider spacing at 50 × 50 cm. He has also explained the technology 

to hundreds of farmers and correlated it to other farmers’ 

situations. He has convinced others to adopt it in their 


Fig. 2. Location-specific water management practices and different varieties tested by farmers themselves for successful implementation 

of SRI technology. 

own fields. He has become a role model for others to emulate 

to adopt the SRI technique easily. 

He himself designed an adjustable marker and used it 

for marking lines. This has helped him to follow different spacings 

(i.e., 25 × 25, 50 × 25, 50 × 50 cm). 

Weeding with a specially made rotary weeder was done 

at 7-d intervals to check further weed growth and to incorporate 

weeds in the soil. He could locally manufacture a weeder 

in collaboration with other farmers. 

As there was a drainage problem in his field, he made a 

small ditch at one corner of his field and drained the excess 

water from that field into the ditch and pumped the water outside 

with the help of a small oil engine. Deep alleyways and 

small bunds were made to drain out excess water. 

Small farmer Sri Tirupathi Srinivas has 1 ha of land. He 

belongs to Ballipadu village in Attili Mandal of West Godavari 

District. He has cultivated paddy under the SRI method on 

about 0.5 ha. KVK, Undi, provided paddy seed and technical 

guidance to him at regular intervals. He attended all training 

classes conducted on SRI technology at KVK. He followed all 

the water and weed management practices regularly. As rats 

were the major problem in that area, he installed a polythene 

sheet all around the field, erecting it to 1-m height. The paddy 

plants produced a high number of tillers and he got higher 

yields, with 3,000 kg ha –1 more yield with SRI technology than 

with the conventionally transplanted variety MTU 1071. During 

winter 2004, he followed the SRI technology on his entire 

farm of 1 ha by joining with his brothers and he motivated 8 

other young farmers of his village to adopt this new technology. 

He reaped the highest yields in the district. 

Young and dynamic farmer M. Ramesh from Kaikaram 

village of Unguturu Mandal of West Godavari District struggled 

with weeding practices. He modified different weeders to suit 

his conditions and found a way to overcome his chronic weeding 

problem. 

Large farmer Sri N.V.R.K. Raju belongs to 

Gummampadu village of West Godavari District. Motivated 

by the KVK’s SRI demonstrations during kharif 2003, he 

implemented SRI on 40 ha during rabi 2003-04 on what was 

originally a fish pond. When experiencing difficulty of paddy 

cultivation under an acute water shortage, he was forced to 

switch to SRI cultivation after learning about it. He modified 

several weeders and line markers to suit his field conditions 

apart from a channelized water management system. He 

pumped out the excess water in the field by using an oil engine. 

He was the first farmer to cultivate paddy under SRI in a 

larger area. He reaped a good harvest and motivated several 

others to follow this new technique. 

Many of these research efforts of a group of farmers in 

their own fields have motivated several other farmers of the 

same village and neighboring villages to join in sharing their 

innovative efforts. This type of farmers’ innovative research 

has become an eye opener even to our own researchers. This 

clearly shows that farmers are the real initiators of their own 

research efforts; hence, if farmers are made partners in doing 

research in agriculture, a better outcome will be possible. 

Notes 

Author’s address: Acharya N.G. Ranga Agricultural University, 

Andhra Pradesh, India, training organizer and head, Krishi 

Vigyan Kendra, Undi, West Godavari District, Andhra 

Pradesh, India, e-mail: Drdandu_jraju@yahoo.co.in. 



Session 13 introduced several innovations in the session organization, 

applying the principles of participatory research to the 

setting. First, presenters were selected to include diverse stakeholders 

in agricultural research, including the representative of a 

farmers’ organization working with a university, and the representative 

of a nongovernmental organization working with a research 

station. Second, participants were given the opportunity 

to ask questions after each presentation. Third, rather than a 

panel of experts, a small group discussion was used, giving all 

participants an opportunity to contribute to the output of the session. 

Summary of papers presented 

Three papers illustrated how multiple stakeholders with different 

interests and strategies can contribute to participatory research. 

F. Tozawa provided the perspective of farmers from the Ogata 

area in Akita Prefecture and Y. Taniguchi the perspective of researchers 

at Akita Prefectural University on collaborative research 

to reduce the environmental “footprint” from rice production in 

northwestern Japan. F. Tozawa showed examples of how farmers’ 

input was essential to make no-till and organic technology 

usable by farmers. He also explained how technology needs to 

enable farmers to produce what consumers want, and thus why 

consumer participation is essential in technology development. 

Y. Taniguchi explained how farmers are not users of technology, 

but the central agents in innovation that produces technology. A 

participatory approach enables researchers to enter the farmers’ 

innovation process. In an example from another part of Japan, Y. 

Iwabuchi explained how farmers, researchers, and wildlife preservation 

organizations jointly developed winter-flooded rice-field 

technology with monitoring of wildlife populations. Farmers themselves 

developed an improved method for monitoring frog diets. 

Two papers presented results of how farmers assessed researcherdeveloped 

technologies to determine if they would work successfully 

under their conditions. B. Bagchi explained how farmers 

tested the leaf color chart technology in West Bengal, India. This 

technology enabled farmers to save on fertilizer cost with no reduction 

in yield. F. Palis reported how farmers in the Philippines 

tested controlled irrigation, which alternates wetting and drying 

of paddy fields. Farmers determined that this technology saved 

water and cost with no reduction in yield. In both cases, the trial 

results provided a farmer-based verification for community-based 

extension to other farmers and villages. 

Two papers presented methods for going beyond traditional 

farmer participatory research in both space and time. G. Trébuil 

provided an introduction to companion modeling, an approach 

that combines participatory role-playing by multiple stakeholders 

with agent-based computer simulations using agroecosystems, 

biophysical, and socioeconomic data to collectively assess scenarios 

of changes. As seen in an example from the northern Thai- 

land highlands, this method can enable farmers and researchers 

to explore future change options over a longer time span and 

wider area. It can also facilitate intra- and intervillage communication 

for developing and testing options for change at the watershed 

level, as seen in a second example on irrigation water 

management from Bhutan. Nguyen De presented 10 years of 

results of participatory plant breeding and participatory varietal 

selection in Vietnam. An average of 3.5 varieties have been selected 

and used by 29 communities through this approach. 

T. Paris and Z. Abedin concluded the presentations with a 

review of IRRI’s participatory approaches. These have moved from 

contractual and consultative modes, in which researchers lead 

and farmers respond, to the collaborative mode, in which farmers 

and researchers are partners. Then, as participatory research 

becomes collegial, farmers have acquired the ability to begin research 

on their own and seek out researchers to support their 

own efforts. The authors concluded by pointing out difficulties in 

institutionalizing participatory research, and offered future directions 

to overcome these difficulties. These include 

Changing the “mind-set” of scientists and research administrators 

to see that participatory research complements 

conventional research, providing a mechanism 

for scientists to determine research needs and understand 

how farmers change technologies. 

Recognizing that participatory research has a higher 

start-up cost but shortens farmer decision time in making 

change, and thus is cost-effective over the technology 

generation process from needs identification to widespread 

farmer change. 

Including participatory research methods in training programs 

of both IRRI and national programs. 

Moving participatory research beyond donor-driven, 

project-based activities to become an integral part of 

national programs in technology generation, validation, 

and extension. 

Participant contributions: methods 

for improving participatory research approaches 

After the presentations, participants were divided into two groups, 

each organized around a series of questions. Each small group 

discussed the questions, selected key points, and summarized 

them in slides or transparencies. Then, the two groups came 

back together, each group presented its summary, and members 

of the other group asked questions and commented. 

Group A’s topic was “Working with farmers as researchers.” 

The group considered three questions and offered the following 

guidelines. In these guidelines, information in parentheses 

is taken directly from the groups’ report, while explanation of 

specific terms used in participatory research has been added in 

brackets for readers of this summary. 


Farmers’ 

advantages 

Researchers’ 

advantages 

Goals 

l cost reduction 

l premium quality 

l environment 

Farmer conditions 

Research must be interesting 

(for example, use role playing) 

Goal and 

hypothesis 

Fusion of 

l quantifiable goals 

l farmer participation 

in goal setting 

Methods 

l broadening scope of ideas to test 

l verifiable methodologies 

Fig. 1. Advantages (boxes) of participatory research for farmers and researchers (ovals). 

1. What methods can enable farmers to contribute as researchers 

in the field to experimentation for technology 

development 

assessment and prioritization of needs and opportunities 

matching appropriate interventions [innovations to 

test] 

 

 

 

rapport building 

assessment of what farmers have done so far (giving 

recognition to farmers’ own efforts and building 

on what they are doing already) 

technology introduction, farmer volunteerism, and active 

farmer/researcher partnership (researchers may 

supply critical inputs but risks should be shared) 

2. What mechanisms and types of farmer-researcher interaction 

can enable farmers to contribute their own 

ideas for scientific evaluation, rather than simply involving 

farmers to provide their evaluation of ideas from 

scientists 

consensus 

rapport building (cultural sensitivity) 

encouraging feedback between partners 

using community communication structures 

undertaking direct observation (in the field) and participant 

observation [researchers joining with farmers 

in carrying out trials in the field] 

3. How can farmer participation help biological scientists 

do better and more relevant science 

scientists gain access to farmers’ knowledge 

scientists learn from farmers (who have a broader 

perspective of the farming system) 

widens scientists’ horizons 

enables scientists to be practically oriented (praxis), 

thus refining technology 

ensures appropriate technology development 

Group B’s topic was “Broadening participation socially and 

spatially.” Before considering broadened participation, the group 

first summarized its understanding of the advantages of farmers 

and researchers participating together in on-farm research. Farmers 

bring their goals and conditions, while researchers bring methods 

that can help farmers assess new ways to achieve those 

goals. The results are quantifiable goals and hypotheses developed 

together (Fig. 1), leading to a shared, common representation 

of the problem to be examined and solved through farmerresearcher 

collaboration. 

The group then considered two questions for broadening 

participation: 

1. How can nonfarmer stakeholders (wholesalers, input 

suppliers, local government, NGOs, etc.) contribute to 

technology development 

Figure 2 shows the advantages for each type of stakeholder. 

Business stakeholders can obtain concrete advantages 

that will improve their business, while participatory 

research can be a social laboratory for government 

to learn about the actual and potential effects of 

policies. Consumers can have a better chance of obtaining 

the type of rice (or other agricultural product) of 

the quality they desire. 

2. How can technology development go beyond a few sites 

of intensive farmer-researcher interaction, without losing 

participation 

Several types of processes were proposed (Fig. 3). One 

process is based on farmers becoming trainers to enable 

other farmers to undertake participatory research 

on their own, and participatory research spreading as a 

method spontaneously among communities. Two especially 

useful tools for this process are (1) study tours by 

communities new to the process to communities that 

already have experience and results through participatory 

research, and (2) farmer field schools to develop 

site-specific field management techniques through joint 

observation by farmers and researchers. The latter 

method is well known for its successes in integrated 

pest management (IPM), but its use should be broad- 


Local 

government 

Farmers and 

participatory 

research 

Input 

suppliers 

l 

l 

Can know effect of a policy 

Useful for democracy 

l Learn farmers’ needs 

- shift business strategy 

l Social responsibility as member 

of community 

- feel more useful 

Wholesalers 

Large 

retailers 

Niche 

retailers 

Manufacturers 

See their clientele 

Have long food chain 

l gain platform of communication 

with distributors and consumers 

Consumers 

Determine market size for investment 

l 

Possibility of new high-quality rice 

Fig. 2. Types of stakeholders (ovals) and advantages (boxes) of participatory research for each. 

Study tour 

Farmers 

trained 

to be trainers 

Farmers field school 

(originally IPM) 

IBM (integrated 

biological management) 

Other communities 

Retail 

Input 

suppliers 

Other 

Pilot community 

Knowledge 

communities: 

workshops, 

information 

b 

technologies* 

Media 

l soap opera 

l “3 reductions, 

3 gains” 

Spontaneous interaction 

Message should be simple 

*GIS, generic model 

tailor each stakeholder can access 

Web-based e-participation and e-learning 

Fig. 3. Scaling-up process. 


ened to integrated biological management (IBM). A second 

process can grow out of the first, in the formation 

of information communities. These can use workshops 

and, where possible, information technology, to develop 

models that can be tailored to the needs of many types 

of stakeholders, both farmers and nonfarmers. These 

can provide a basis for e-participation and e-learning. 

Finally, results from participatory research can be used 

by extension through improved media approaches. These 

should use formats that farmers enjoy, such as soap 

operas (successful in the Philippines) and simple messages, 

such as “3 reductions, 3 gains” (successful in 

Vietnam). 


SESSION 14 

Potential for diversification in rice-based 

systems to enhance rural livelihoods 

CONVENER: M. Hossain (IRRI)

Agricultural diversification in Asia: opportunities 

and constraints 

Prabhu Pingali 

Rapid economic and income growth, urbanization, and globalization 

are leading to a dramatic shift of Asian diets away 

from staples and increasingly toward livestock and dairy products, 

vegetables and fruit, and fats and oils. The tendency for 

per capita rice consumption to decline with income growth 

and with urbanization has been widely documented in the literature 

(Ito et al 1989, Huang and David 1993). FAO projections 

indicate that the per capita consumption of rice will level 

off by 2015 and start to decline by 2030 (FAO 2003). 

The rice sector in Asia is facing the dual challenge of 

sustaining high rates of rice productivity growth while transforming 

itself from a subsistence-oriented monoculture system 

to a diversified market-oriented system. This paper examines 

the scope for the diversification of rice-based farming in 

Asia. Economic, agro-climatic, and technological constraints 

to the commercial transformation of subsistence rice systems 

are identified. Priorities for research, and the primary components 

of a policy agenda, are described. 

The diversification of rice-based farming systems 

A recent FAO/World Bank study on farming systems and poverty 

suggested that diversification is the single most important 

source of poverty reduction for small farmers in South and 

Southeast Asia (FAO and World Bank 2001). The three most 

important systems described in the study are all rice-based 

farming systems (see Table 1): the tropical lowland rice sys- 

tem, the rice-wheat system, and the rainfed uplands. They account 

for about 80% of the agricultural population and some 

50% of the total agricultural area in Asia. The tropical lowland 

and rice-wheat systems are the dominant sources of rice 

supply in Asia. These systems witnessed rapid productivity 

growth during the Green Revolution and their productivity 

continues to be high in the post-Green Revolution period. Yet, 

pressure to diversify out of rice is also the greatest in these 

systems, primarily because of low returns to rice relative to 

high-value alternatives such as vegetables (Pingali et al 1997). 

The feasibility and cost of substituting other crops vary 

across all three farming systems. The flexibility of farmers to 

respond to the changing relative prices and relative profitability 

in their crop choice decision-making can be described in 

terms of the level of investments (both physical and human 

capital) required in switching from rice to nonrice crops and 

vice versa. Flexibility is low during the rainy season for the 

tropical lowlands and the rice-wheat zone because the drainage 

costs for growing nonrice crops can be prohibitive (Pingali 

et al 1997). Upland areas, however, can oscillate between rice 

and nonrice crops with minimum additional investment. 

Access to markets and the relative prices of rice and 

nonrice crops, especially horticulture, are additional determinants 

of diversification. Although roads and market places are 

important, proximity to urban areas expands the range of 

nonrice diversification options, especially for fresh produce. 

Table 1. Main rice-based farming systems in Asia. 

Land area Agricultural 

Farming system Amount (% of region) population Principal livelihoods 

(% of region) 

Tropical lowland rice 11 32 

Agricultural population 604 Irrigated and rainfed rice, vegetables, 

Cultivated area 93 legumes, off-farm activities 

Irrigated area 42 

Rice-wheat 9 22 

Agricultural population 416 Irrigated rice; wheat, vegetables, livestock 

Cultivated area 93 including dairy, off-farm activities 

Irrigated area 158 

Rainfed uplands 30 26 

Agricultural population 636 Cereals, legumes, fodder, livestock, horticul- 

Cultivated area 189 ture, seasonal migration, and off-farm ac- 

Irrigated area 38 tivities 

Source: Tables 5.1 and 6.1 in FAO and World Bank (2001). 

Note: Population figures are in millions, area figures in million hectares. 


During the wet season, rice will continue to be the dominant 

source of income in all but upland environments. In the 

irrigated lowland rice and rice-wheat systems, dry-season rice 

and/or wheat will continue to be the major source of income. 

Areas with good market access and those near urban centers 

will increasingly diversify to nonrice crops and vegetable production. 

Dry-season cropping activities in rainfed areas are 

limited because of technical problems related to timely and 

effective crop establishment, limited moisture (or excessive 

moisture in some cases), and generally modest or high yield 

instability. Off-farm activities are often more dependable 

sources of income, suggesting that dry-season cropping intensities 

would remain low even if technical problems in crop 

production were solved. 

Diversification constraints 

Diversification out of rice is constrained by market availability 

and size, land suitability and rights, irrigation infrastructure, 

and labor supply. Where output demand is relatively elastic, 

the returns from investments in land, technology, and time 

spent in learning about new crops are relatively higher. 

Diversification and risk 

Diversification from a rice-monoculture system to a nonrice 

crop system could lead to increased variability in farm household 

income, which basically comes from yield or price fluctuations. 

Moreover, the diversification “start-up” phenomenon, 

of high prices for several seasons leading to oversupply and a 

consequent collapse of prices, is all too common. This can be 

countered by measures to expand the market by lowering transaction 

costs, improving external linkages, or providing storage 

and processing technologies. Effective rural financial institutions 

will also assist in risk spreading and in the sharing of 

the benefits of commercialization more widely across the community 

and region. 

Land suitability and land rights 

The ability to profitably convert rice lands to nonrice crops is 

constrained by the drainage requirements for lowlands and 

erosion control investments in uplands. It is important to understand 

that not all lands can be converted to nonrice production. 

Even for lands that can be converted, substantial investments 

in land improvement need to be made to sustain the 

long-term productivity and profitability of nonrice crop production. 

Investments in land improvement are likely to be made 

only where secure rights to land exist. 

In upland areas, where market access is good, the profitability 

of diversified field crop production on soils not highly 

susceptible to erosion is high. For soils susceptible to erosion, 

profitability of field crop production is determined by the level 

of erosion control investments. Where such investments are 

high, tree crops may be a more viable option than field crops, 

particularly when field crop production has allowed land degradation 

to occur. In upland areas with poor market access, the 

returns from diversification out of subsistence rice production 

are limited in areas with either type of soil. 

Farmers’ interest in erosion control measures is directly 

related to land values and market access and is conditional on 

the availability of suitable technologies. Secure rights to land 

create the incentives farmers need to invest in land improvements 

that conserve and increase the long-term productivity 

growth that can be induced by the start of commercialization. 

Irrigation infrastructure 

Many observers have argued that existing irrigation systems 

constrain diversification because of the rigid designs of infrastructure 

and inflexible water delivery systems (Schuh and 

Barghouti 1988). It is argued that this inflexibility prevents 

appropriate allocation of water to nonrice crops, thus limiting 

farmers to rice monoculture. Based on these arguments, technology-based 

solutions to diversification within irrigation systems 

are advocated, mainly capital investment in improved 

conveyance, diversion, and drainage systems. 

An alternative argument would be that the failure to diversify 

within irrigation systems is the result of incentive failures 

resulting from the centralized allocation of unpriced irrigation 

water. Policies that establish markets in tradable water 

rights could establish incentives to economize on water and 

choose less water-intensive crops (in the dry season), by inducing 

water users to consider the full opportunity cost of water 

(Rosegrant et al 1995). The establishment of transferable water 

rights can provide maximum flexibility in responding to 

changes in crop prices and water charges as demand patterns 

and comparative advantage change and diversification of cropping 

proceeds (Rosegrant and Binswanger 1994). 

Labor supply 

Does diversified cropping increase labor requirements Yes; 

relative to rice, the per hectare labor requirements for onions, 

vegetables, and other high-value crops are substantially higher. 

Given the higher labor requirements for cropping and drainage, 

nonrice crops on irrigated lands are grown on extremely 

small plots, in general about a fourth of the rice area. 

In addition to crop labor requirements, the supervision 

time required from the farmer is also significantly higher: this 

may be the dominant labor constraint to high-value nonrice 

crop production given the highly inelastic nature of management 

labor available in the farm household compared to hired 

labor augmented by seasonal migrants. 

Implications for research 

Research should focus on providing farmers with the flexibility 

to make crop choice decisions and to move relatively freely 

among crops. 

Both substantial crop-specific research and system-level 

research will be required to provide farmers with flexibility of 

crop choice. Crop-specific research includes increases in yield 

potential, shorter duration cultivars, improved quality charac- 

Session 14: Potential for diversification in rice-based systems to enhance rural livelihoods 421

teristics, and greater tolerance of pest stresses. System-level 

research includes land management and tillage systems that 

allow for shifts of cropping patterns in response to changing 

incentives and farm-level water management systems that can 

accommodate a variety of crops within a season. Also important 

at the system level is research on the carryover effect of 

inputs and management practices across crops, for instance, 

high insecticide applications or the effects of intensification in 

terms of prolonged water saturation, the buildup and carryover 

across crops of pest populations, rapid depletion in soil micronutrients, 

and changes in soil organic matter that could lead 

to reduced productivity of rice monoculture systems over the 

long term. 

Given growing populations and income-induced demand 

for increased cereal consumption, there continues to be a strong 

need to seek higher productivity levels for the staple cereals. 

The need to increase the productivity of cereals is higher the 

greater the diversion of high-potential irrigated lands to 

noncereal pursuits. 

Agenda for a food and agricultural policy 

Commercialization trends require a paradigm shift in agricultural 

policy formulation and research priority setting. The relevant 

development paradigm for the 21st century is one of 

food self-reliance, in which countries import a part of their 

food requirements in exchange for diverting resources out of 

subsistence production. Agricultural policy should emphasize 

maximizing farm household income rather than generating food 

surpluses. 

Governments have a difficult task to perform. On the 

one hand, continued food security needs to be assured for populations 

that are growing in absolute terms; on the other hand, 

research and infrastructural investments need to be made for 

diversification out of the primary staples. As Timmer (1992) 

stated, “Ultimately, the process of rural diversification must 

be consistent with the longer-run patterns of structural transformation.” 

The process of agricultural diversification should not be 

expected to be frictionless. Significant equity and environmental 

consequences can arise in the short to medium term unless 

appropriate policies are followed. Long-term strategies to facilitate 

a smooth transition to commercialization include investment 

in rural markets, transportation and communication 

infrastructure to facilitate integration of the rural economy, crop 

improvement research to increase productivity, and crop management 

and extension to increase farmers’ flexibility and reduce 

possible environmental problems from high input use; 

and establishment of secure rights to land and water to reduce 

risks to farmers and provide incentives for investment in sustaining 

long-term productivity. 

References 

FAO. 2003. World agriculture: towards 2015/2030. London (UK): 

Earthscan Publications Limited. 

FAO and World Bank. 2001. Farming systems and poverty: improving 

farmers’ livelihoods in a changing world. Rome and Washington, 

D.C.: FAO and World Bank. 

Huang J, David CC. 1993. Demand for cereal grains in Asia: the 

effect of urbanization. Agric. Econ. 8:107-124. 

Ito S, Peterson W, Grant W. 1989. Rice in Asia: Is it becoming an 

inferior good Am. J. Agric. Econ. 71:32-42. 

Pingali PL, Hossain M, Gerpacio RV. 1997. Asian rice bowls: the 

returning crisis Wallingford (UK): CAB International. 

Rosegrant MW, Binswanger H. 1994. Markets in tradable water 

rights: potential for efficiency gains in developing country 

irrigation. World Dev. 22:1613-1625. 

Timmer CP. 1992. Agriculture and economic development revisited. 

Agric. Syst. 40:21-58. 

Notes 

Author’s address: Director, Agricultural and Development Economics 

Division, FAO, United Nations, Rome, Italy. 

Consequences of technologies and production 

diversification for the economic and environmental 

performance of rice-based farming systems in East 

and Southeast Asia 

Huib Hengsdijk, Marrit van den Berg, Reimund Roetter, Wang Guanghuo, Joost Wolf, Lu Changhe, and Herman van Keulen 

Rice-based ecosystems in East and Southeast Asia are being 

challenged by the simultaneous requirements for more diversified 

products (Hossain 1998), increased productivity, and 

reduced environmental impact (Dobermann et al 2004). Location-specific 

conditions, such as access to labor and product 

markets, and biophysical conditions determine the potential 

for and constraints to diversification, adoption of technological 

innovations, and a productivity increase in such systems. 

The interrelated issues at stake in rice-based ecosystems are 

complex and require multidisciplinary research methods in 

which knowledge and information from different disciplines 

are integrated and synthesized. Farm household modeling 


(Singh et al 1986) as applied within the project “Integrated 

Resource Management and Land-use Analysis” (IRMLA) is 

one of such methods. 

Here, a farm household model is presented to study the 

performance of two household types differing in off-farm employment 

opportunities. The aim is to identify the scope for 

development for both types of households, taking into account 

multiple economic, social, and environmental objectives (such 

as increasing income and minimizing environmental impacts 

of agricultural activities). Pujiang County in Zhejiang Province, 

China, is used as a case study area for development and 

application of the model. 

Description of Pujiang 

Pujiang County is located in the center of Zhejiang Province, 

about 100 km southwest of the provincial capital Hangzhou, 

which offers ample off-farm employment opportunities. Despite 

the declining rice cultivation area in Pujiang, single and 

double rice-cropping systems are still the predominant agricultural 

land-use activities. About 57% of the 18,000 ha of 

cultivated land consists of rice paddies, while the remainder is 

used for upland crops, including different types of vegetables 

and fruits. 

Methodology 

The methodology applied has two major model components: 

(1) a farm household model based on linear programming techniques 

(Singh et al 1986) and (2) a technical coefficient generator 

called TechnoGIN (Ponsioen et al 2003) for quantifying 

input-output relations of the various production activities. 

The farm household model selects land-use options from a 

range of alternatives generated with TechnoGIN, while maximizing 

farm income subject to certain boundary conditions 

and restrictions. In this study, we identify two types of households 

that differ only in off-farm employment opportunities 

providing extra income to the household. 

TechnoGIN 

TechnoGIN is a generic expert tool for integrating different 

types of information on crop production and generating inputs 

and outputs of cropping systems. Based on soil, crop, and technology 

characteristics, TechnoGIN allows the characterization 

of specified cropping systems in terms of their relevant inputs 

and outputs, such as yield, crop residues, nutrient and biocide 

use, and labor requirements. Quantification of input-output 

relationships is based on the interpretation of survey data for 

representative cropping systems. TechnoGIN can also be applied 

for producing information that is often not available from 

surveys, such as losses of nutrients to the environment and the 

environmental impact of biocides. 

Farm household model 

The model accounts for the main characteristics of farm households 

in Pujiang: (1) a high labor/land ratio, (2) limited availability 

of financial capital, and (3) increasing opportunities 

for off-farm employment. The model maximizes household 

income from crop production and off-farm income, subject to 

the availability of land and family labor, required self-sufficiency 

in rice, access to working capital and off-farm employment, 

available agricultural technology, and market prices. 

Labor and financial balances are computed per 10-day period 

and are key constraints in our simulations. 

A major constraint to the adoption of new crops and 

improved technologies is the limited availability of working 

capital. Increasing employment opportunities in the nonfarm 

sector can alleviate this constraint, as the resulting income can 

be used to purchase inputs for crop production. We illustrate 

this process by distinguishing two farm households in this study 

that differ only with respect to their access to off-farm employment. 

The first household (HH1) has one family member 

(out of three) that is employed full-time in the nonfarm sector 

at a wage rate double the cost of hiring an agricultural laborer, 

that is, 5 versus 2.5 RMB per hour. The other household (HH2) 

uses available labor only on its own farm. Both households 

have a working capital of 1,000 RMB and a farm size of 0.4 

ha. 

Using the farm household model, we analyzed the following 

four scenarios: 

1. Reference. This represents the actual situation in 

Pujiang in which most farmers select cultivation of 

double or single rice systems. 

2. Technological innovations in rice cultivation. In this 

scenario, the same rice production systems are available 

as in the reference scenario, but now using technologies 

aimed at increasing crop productivity and 

reducing fertilizer use, through the use of hybrid rice 

varieties and site-specific nutrient management 

(SSNM) (Guanghuo et al 2004). 

3. Vegetables. In addition to rice cropping systems, vegetable-based 

cropping systems are available in this 

scenario. 

4. Vegetable prices. Here, the price ratio between rice 

and vegetables is changed through a stepwise reduction 

in vegetable prices to a minimum of 10% of the 

prices used in the vegetables scenario. 

Results 

Reference 

Both households select double rice systems that are more profitable 

than single rice. The economic performance of HH1 is 

much better than that of HH2 because of income earned through 

off-farm employment (Table 1). Since off-farm employment is 

more remunerative than farming activities, HH1 hires a limited 

amount of external labor during the peak period around 

harvesting of early rice and transplanting of late rice. Therefore, 

on-farm income of HH1 is slightly lower than that of 

HH2. 


Table 1. Results (in units per year) of the reference, technological innovation, and vegetables scenarios 

for farm households HH1 and HH2. 

Reference Technological Vegetables 

Item Unit innovation 

HH1 HH2 HH1 HH2 HH1 HH2 

On-farm income RMB a 2,996 3,076 3,593 3,712 17,477 10,083 

Off-farm income RMB 14,400 0 14,400 0 14,400 0 

On-farm labor Month 88 92 109 117 208 132 

Off-farm labor Month 360 0 360 0 360 0 

Hired labor Month 4 0 4 0 112 36 

N losses kg N ha –1 285 285 237 242 583 366 

Biocide index ha –1 604 604 604 604 2,115 969 

Rice production Mg farm –1 4.2 4.2 4.7 4.8 1.2 2.1 

a 8.2 RMB = US$1. 

Technological innovations in rice cultivation 

Both households predominantly select double rice systems with 

hybrid rice varieties and SSNM. These innovations increase 

on-farm income by about 20% for both households (Table 1). 

In addition, environmental pollution through N emissions decreases 

by almost 20% compared to the reference situation. At 

the same time, rice production of both farms increases by about 

10%. Off-farm employment is still the major cash earner in 

HH1. 

Vegetables 

Because of the introduction of vegetables, on-farm income of 

HH1 and HH2 increases almost five times and two and half 

times, respectively, compared with that in the previous scenario 

(Table 1). HH1 grows double rice on 25% of its land, 

which is just enough to meet household needs for rice. On the 

remainder of the land, a rotation of celery, leafy vegetables, 

and garden radish is grown. HH2 also selects this vegetable 

rotation but on less than 10% of its land. Almost 75% is cultivated 

to a rotation of leafy vegetables followed by single rice. 

The rest of HH2’s land remains fallow as the household lacks 

working capital to purchase inputs for other cropping activities. 

Environmental pollution by both households increases 

dramatically with the introduction of vegetables (Table 1). 

Vegetable prices 

Up to a reduction in vegetable prices of 40%, the behavior of 

both households is similar, with household income gradually 

decreasing as a consequence of the lower returns for vegetables 

(Fig. 1). Then, the share of rice in the cropping patterns of 

both households increases steadily up to a reduction in vegetable 

prices of 90% for HH1 and 80% for HH2, below which 

only rice is produced. The increase in rice in the cropping pattern 

of HH2 is associated with an expansion of its cultivated 

area from 0.34 ha to the complete landholding of 0.40 ha. At 

the same time, N emissions to the environment decrease for 

both households. 

Household income (RMB) 

35,000 

30,000 

25,000 

20,000 

15,000 

10,000 

5,000 

0 

HH1 

HH2 

Rice production (Mg farm –1 ) 

5.0 

4.5 

4.0 

3.5 

3.0 

2.5 

2.0 

1.5 

1.0 

0.5 

0 

N loss (kg ha –1 ) 

600 

500 

C 

400 

300 

200 

100 

0 

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 

Fraction of original vegetable price 

Fig. 1. Effect of vegetable price (1.0 is the survey-based mean 

vegetable price, decreasing toward the left) in Pujiang for (A) household 

income, (B) rice production, and (C) N losses. 

B 

A 


Conclusions 

The results obtained in this study have several important implications 

for policymakers, farmers, and other stakeholders 

concerned with sustainable development in the rural-urban 

fringes of East and Southeast Asia: 

The economic performance of farm households is 

dominated by their access to working capital through, 

for example, off-farm employment. 

Technical innovations (SSNM) can increase on-farm 

income and rice production and decrease environmental 

pollution. 

The introduction of vegetables in the cropping system 

leads to strong increases in household income, 

but may increase income inequalities among farm 

households and is detrimental to the environment. 

Access to credit and improved extension services are 

a prerequisite for, in particular, poor farmers to shift 

their production to activities with higher added values 

(such as vegetable production) that contribute to 

the improvement of their livelihood. 

References 

Dobermann A, Witt C, Dawe D, editors. 2004. Increasing productivity 


management. Enfield, N.H. (USA): Science Publishers, Inc., 

and Los Baños (Philippines): International Rice Research Institute 

(IRRI). 410 p. 

Guanghuo W, Sun Q, Fu R, Huang X, Ding X, Wu J, He Y, 

Dobermann A, Witt C. 2004. Site-specific nutrient management 

in irrigated rice systems of Zhejiang Province, China. 

In: Dobermann A, Witt C, Dawe D, editors. Increasing productivity 


management. Enfield, N.H. (USA): Science Publishers, 

Inc., and Los Baños (Philippines): International Rice Research 

Institute (IRRI). p 243-263. 

Hossain M. 1998. Sustaining food security in Asia: economic, social, 

and political aspects. In: Dowling NG, Greenfield SM, 

Fischer KS, editors. Sustainability of rice in the global food 

system. Davis, Calif. (USA): Pacific Basin Study Center, and 

Manila (Philippines): International Rice Research Institute. 

p 19-44. 

Ponsioen TC, Laborte AG, Roetter RP, Hengsdijk H, Wolf J. 2003. 

TechnoGIN-3: a technical coefficient generator for cropping 

systems in East and Southeast Asia. Quantitative Approaches 

in Systems Analysis No. 26. Wageningen (The Netherlands). 

69 p. 

Singh I, Squire L, Strauss J. 1986. Agricultural household models: 

extensions, applications, and policy. Baltimore, Md. (USA): 

The Johns Hopkins University Press. 

Notes 

Authors’ addresses: Huib Hengsdijk, Plant Research International, 

Wageningen UR, P.O. Box 16, 6700 AA Wageningen, The 

Netherlands, E-mail: Huib.Hengsdijk@wur.nl; Marrit van den 

Berg, Plant Production Systems, Wageningen University, P.O. 

Box 430, 6700 AK Wageningen, and Development Economics, 

Wageningen University, P.O. Box 8130, 6700 EW 

Wageningen, The Netherlands; Reimund Roetter, Alterra, Soil 

Science Centre, Wageningen UR, P.O. Box 47, 6700 AA 

Wageningen, The Netherlands; Wang Guanghuo, Zhejiang 

University, Hangzhou, China; Joost Wolf, Alterra, Soil Science 

Centre, Wageningen UR, P.O. Box 47, 6700 AA 

Wageningen, The Netherlands; Lu Changhe, Institute of Geographic 

Sciences and Natural Resources Research, Chinese 

Academy of Sciences, Building 917, Datun Road, Beijing 

100101, China; Herman van Keulen, Plant Research International, 

Wageningen UR, P.O. Box 16, 6700 AA Wageningen, 

and Plant Production Systems, Wageningen University, P.O. 

Box 430, 6700 AK Wageningen, The Netherlands. 

Rural poverty and agricultural diversification in Thailand 

Alia Ahmad and Somporn Isvilanonda 

Thailand has experienced steady economic growth and structural 

changes in its economy in the last four decades that enabled 

it to gain a position among the newly industrialized nations. 

Although a great deal of this growth comes from industrial 

development, the agricultural sector has contributed significantly 

to the process through exports, a cheap food supply, 

and the release of labor for industrial development. The structural 

changes associated with economic growth reflect the 

changing role of agriculture in the economy. The share of agriculture 

in GDP declined from 44% in the early 1960s to 10% 

in recent years (Isvilanonda 1998). Its share in employment 

has shown a similar trend, albeit at a slower pace. The share of 

agriculture in total employment fell from 83% in 1957 to 57% 

in 1999. In more recent years, it has hovered around 50% 

(Mundlak et al 2002). 

Although Thailand has been very successful in reducing 

poverty because of rapid and steady economic growth, rural 

poverty, especially in certain regions, is a serious problem. 

Rural-urban disparities also increased after the financial crisis 

because of the inability of the urban sector to absorb rural labor 

at a rapid rate, and the declining importance of agriculture 

in the total value-added. We underscore in this paper that Thai 

agriculture still has a major role to play, and the major challenge 

is to switch to high value-added products, in other words, 

agricultural diversification. 


Background and aims of the study 

Agricultural diversification in Thailand means primarily 

deintensification of rice production. Thailand became a major 

rice-exporting nation in the second half of the 19th century 

because of its abundant suitable land. The favorable land/person 

ratio is one of the reasons behind Thailand’s late adoption 

of land-saving modern rice technology. However, the nation 

started losing its monopoly position in the international market 

in the late 1960s. Traditionally, squeezing rice farmers 

through an export tax, a rice premium, and government control 

of rice sales was the main feature of Thai policies 

(Siamwala 1975). Discrimination against rice farmers has 

weakened since the 1970s partly because of competition in 

the international market. The government supported the adoption 

of new technology through investment in irrigation/infrastructure, 

although this constitutes a small proportion of total 

rice area (Isvilanonda 1998). In spite of this, rice is facing the 

problem of low profitability, mainly because of the declining 

demand in both the international and domestic markets. Ideally, 

the declining profitability of rice should induce farmers 

to switch to other crops. But this has not happened uniformly— 

the pattern of deintensification of rice and diversification into 

other crops and nonfarm activities differs from region to region. 

The overall aim of this paper is to analyze the pattern of 

diversification at the farm level, its effects on farm income, 

and the constraints faced by farmers in different regions and 

under different production environments. 

This paper addresses the following questions: 

Is diversification out of rice more difficult in poorer 

regions than in prosperous regions 

Is diversification into other agricultural activities 

easier in irrigated areas growing modern rice than in 

rainfed areas growing traditional rice 

What factors affect the ability of farmers to switch to 

high-value crops 

How is diversification related to income inequality 

in rural areas and across regions 

Concepts and measures used in the study 

Three types of diversification have been considered: 

1. Crop diversification—diversification out of rice intensification 

measured in terms of both land and value 

of production. 

2. Diversification of agricultural production measured 

in terms of value of production. 

3. Intersectoral diversification into nonagricultural activities 

measured in terms of household income. 

The study areas and data 

The study is based on household-level data from two regions, 

Suphan Buri (SB) in the Central Plain and Khon Kaen (KK) in 

the Northeastern Province, the commercial and traditional ricegrowing 

areas in Thailand, respectively. Three production environments 

are selected in each province. These two regions 

were investigated in surveys at different times—1995, 1998, 

and 2001-02. The number of households investigated was 234 

in 1995, 240 in 1998, and 280 in 2001-02. 

Results 

Land use and cropping patterns 

According to our household-level survey data (2001), total 

sown area is much smaller (only 2.19 ha) in KK than in SB 

(11.7 ha), with a decline since 1995 in KK but an increase in 

SB. The larger farm size coupled with the increase in cropping 

intensity because of irrigation facilities has resulted in a rise in 

average sown area in SB. 

Rice intensification 

In spite of greater cropping intensity, the percentage of sown 

area under rice in 2001 in SB is lower (86.67) than in KK 

(90.67). In SB, rice intensification has declined since 1995 in 

contrast to KK, where it has increased. Differences are considerable 

among the villages in SB, where rice intensification 

is lower in irrigated areas. Rice intensification in the floodprone 

areas is extremely high and has influenced the region’s 

average. In KK, on the other hand, rice intensification is higher 

in irrigated areas than in rainfed or drought-prone areas. Looking 

at the diversification pattern, it appears that, in SB, it is 

cash crops and water chestnuts, rather than fruits, that led to 

diversification out of rice. In KK, the importance of upland 

crops has declined with a corresponding increase in fruit trees. 

Although all households grow rice in both regions, the 

proportion of households growing only rice is lower in SB 

than in KK. In KK, it is increasing, whereas in SB it is decreasing. 

Differences are considerable between SB and KK in 

rice productivity: 3.44 versus 2.47 t ha –1 , the annual average 

for each region as a whole. In SB, rice productivity differs in 

different production environments. It is very high in irrigated 

and flood-prone areas (4.5–5 t ha –l ) compared with rainfed 

areas with less than 1 t ha –l . 

In contrast to SB, there are no significant differences in 

productivity in different production environments in KK since 

farmers prefer to grow local varieties in these villages, even in 

irrigated villages. 

Diversification in terms of gross value 

of agricultural production 

In both Suphan Buri and Khon Kaen, the shares of rice in terms 

of sown area are much greater than the shares in terms of gross 

value of crops. But the trend is moving in the opposite direction. 

Rice is gaining importance in terms of both value and 

sown area in KK, and is declining in importance in SB. There 

are significant differences within each region depending on 

the production environments. 

In Suphan Buri, diversification has increased except in 

flood-prone areas. Both rice and water chestnuts are a top crop, 


and the crop diversification index indicates a high level of diversification. 

In Khon Kaen, on the other hand, because of the increasing 

importance of rice, the crop diversification index in irrigated 

and drought-prone villages shows a declining trend, but 

a rising trend is observed in the rainfed village. On average, it 

has declined since 1995, with a slight increase in 1998. 

Constraints to diversification within agriculture 

There appear to be three categories of problems: basic inputs 

(soil, labor, water), institutional factors (access to land, credit, 

knowledge), and price/marketing factors. Households in SB 

face more constraints with respect to basic facilities and marketing 

risks, whereas those in KK experience more institutional 

problems. A high person-land ratio and a smaller farm size 

may be important factors. The differences between the two 

regions are mainly due to the commercial versus subsistence 

nature of production. In KK, very few households, 8 out of 

142, mentioned marketing problems, and no household considered 

price fluctuation as a constraint. 

Household income from different sources 

While KK is more dependent on rice as a source of income 

within agriculture, its dependence on rice in total income is 

significantly less because of its dependence on nonfarm activities. 

In 2001, the percentage of nonfarm income in KK was 

63% versus 26% in SB. 

Poverty/inequality 

First, we look at inequality between regions, followed by inequality 

within each region. Inequality is expressed in two 

ways: first in terms of the ratio of average per capita income in 

the two regions (SB/KK), and second in terms of the ratio of 

the income of the poorest 20% and the richest 20%. In 1995, 

average per capita income in SB was 2.8 times the income in 

KK. There has been a decline in inequality between 1995 and 

2001. In 1995, the poorest 20% of the population in SB earned 

1.3 times more than its counterpart in KK and inequality among 

the poor between the two regions increased in 2001. Sources 

of inequality between the regions are size of landholding, irrigation 

facilities, size of households, and nearness to metropolitan 

areas. 

Although the level of inequality in SB is higher than in 

KK, inequality is declining in SB at a faster rate. The lower 

level of inequality in KK than in SB is due to more equal distribution 

of the access to production factors. In SB, on the 

other hand, inequality in education may have played a role. 

However, the poor in SB have managed well both in absolute 

terms and in a dynamic sense because of the higher level of 

access to resources and opportunities. 

Determinants of diversification-regression analyses 

An OLS regression was run to find out the determinants of 

crop diversification (Table 1). We have used the crop diversi- 

Table 1. Regression results in different crop years 

for Suphan Buri and Khon Kaen. 

Variable Coefficient t-statistic 

Crop year 1995 

Age –0.002344 –1.365289 

Farmsize 0.007129 1.016620 

Personland –0.008233 –1.296398 

No.debt 0.028868 0.765422 

Rentratio –0.078738 –1.568819 

Riceprice 0.001471 0.068920 

Irrigation 0.372536 4.212958 

Rainfed 0.257906 3.108211 

C 1.099079 6.067455 

R-square 0.113992 

Adjusted R-square 0.083308 


Age –0.006359 –3.917268 

Debt 1.61E–07 0.452216 

Farmsize –0.005485 –0.725704 

Personland –0.011813 –1.482655 

No.debt 0.010956 0.239136 

Rentratio –0.081504 –0.954818 

Riceprice –0.054093 –1.417187 

Irrigation 0.464375 5.070744 

Rainfed 0.115971 1.268601 

C 1.658754 8.398210 

Adjusted R-square 0.257375 0.000000 

R-square 0.286060 


Age 0.000607 0.327060 

Debt 5.73E–07 1.580791 

Farmsize 0.006290 0.995595 

Personland –0.018344 –2.923910 

No.debt –0.033540 –0.989760 

Rentratio 0.160279 2.274592 

Riceprice –0.027360 –0.675996 

Irrigation 0.572512 6.849020 

Rainfed 0.462669 4.854177 

C 0.896186 4.138147 

R-square 0.188981 6.990510 

Adjusted R-square 0.161947 0.000000 

Variable explanation 

Age 

Age of household head (years) 

Debt 

Amount of loan (baht) 

Farmsize 

Total landholding of household 

Personland Ratio of population/total land 

No.debt 

No. source of loan 

Rentratio 

Ratio of rent land/total land 

Riceprice 

Paddy price at farm gate 

Irrigation Dummy variable (irrigation = 1, 

other = 0) 

Rainfed Dummy variable (rainfed = 1, 

other = 0) 

Note: Using flood-prone is base for production environment’s 

dummy variable (irrigation, rainfed, flood-prone). 


fication index for each household as the dependent variable 

explained by access to land (+), price of rice (–), irrigation, 

infrastructure and water control (+), agricultural credit (+), 

proximity to urban areas (+), age of the household head (–), 

demographic pressure (–), and the proportion of rental land in 

total land (–). 

In 1995, the above factors explained 25% of the variation 

in diversification among the sample households. Irrigation 

and the age of the household head have turned out to be 

significant at the 1% level with positive effects and negative 

effects, respectively. The model worked poorly for 1998. It 

explained only 8% of the variation in the dependent variable, 

and irrigation was the only significant factor at the 1% level. 

The model using 2001 data explained 16% of the variation, 

and several factors were found to be significant—irrigation 

and the person/land ratio were significant at the 1% level, 

whereas the rental ratio was significant at the 5% level. The 

person/land ratio had a negative effect as expected. 

Conclusions and policy implications 

Diversification primarily refers to deintensification of rice production 

and a switch to other cash crops, fruits, livestock, and 

aquaculture. It has been found that rice intensification is higher 

and increasing in the northeast compared with the Central Plain. 

Also, diversification in terms of land-use pattern and gross 

value of agricultural production is lower in the former region 

than in the latter. 

The factors that constrain diversification in these two 

regions are quite different because of differences in the nature 

of farming (commercial versus subsistence). The main constraints 

perceived by KK farmers are the lack of access to production 

factors, whereas in SB farmers face marketing problems. 

The differences in the ability to diversify in the two regions 

are reflected in the growing inequality between regions, 

whereas inequality has declined in the more prosperous regions. 

It should be noted that these two regions are at different 

stages of agricultural transformation (Timmer 1988) and the 

policy implications for agricultural development and diversification 

are therefore different. In the Central Plain, policies 

should be directed to removing contraints in the production 

and marketing of high-value crops. On the other hand, in the 

Northeastern Province, the main problem is to increase the 

productivity of rice cultivation as well as cultivation of other 

crops. Policies should be directed to securing property rights, 

irrigation, infrastructure, and credit facilities. 

References 

Isvilanonda S. 1998. Rice production and consumption in Thailand: 

the recent trend and future outlook. Thailand J. Agric. Econ. 

55. 

Mundlak, Larson, Butzer. 2002. Determinants of agricultural productivity 

in Thailand, Indonesia and the Philippines. World 

Bank Web site. 

Siamwalla A. 1975. A history of rice price policies in Thailand. In: 

Puey Ungphakorn et al, editors. Finance, trade and economic 

development in Thailand. Essays in honour of Kunying Suparb 

Yossundara. 

Timmer PC. 1988. The agricultural transformation. In: Chenery H, 

Srinivasan TN, editors. Handbook of development economics. 

Amsterdam (Netherlands): Elsevier Science Publishers 

B.V. 

Notes 

Sustaining higher efficiency in rice production 

Amelia S. delos Reyes, Arelene Julia B. Malabayabas, and Mercedita A. Sombilla 

Authors’ addresses: Alia Ahmad, Department of Economics, Lund 

University, Sweden; Sompom Isvilanonda, Department of 

Agricultural and Resource Economics, Kasetsart University, 

Bangkok, Thailand. 

Poor farmers are less able to cope with shortfalls in crop production 

and as a consequence diversify their activities as a 

precaution. This occurs more frequently in rainfed areas where 

uncertainties are higher because of unpredictable environmental 

elements (IRRI 1995). In Asia, close to 30% of the total productive 

area is categorized as rainfed; the majority of this is 

rice-based, with animal raising as a key component (Thornton 

et al 2002). 

This paper presents part of the results of the socioeconomic 

study on crop-animal interactions under the project Sustainable 

Food-Feed Systems and Improved Livelihoods of the 

Poor in Rainfed Lowland Areas, which is funded by the 

Systemwide Livestock Program. It seeks to explain the crucial 

role of animal production in promoting greater efficiency in 

rice production, and provides recommendations for development 

strategies to progressively improve the crop-animal production 

system. It is hypothesized that animal raising helps 

increase rice income and improves production efficiency. 

Integrated crop-animal systems on small rainfed rice farms 

Crop-animal integration on rainfed farms comes in two broad 

categories: systems combining crops (mainly rice), 

nonruminants, ponds, and fish; and systems combining crops 

(also mainly rice) and ruminants (Devendra 1995). The most 

common is the interaction between ruminants, particularly 

cattle/buffalo and rice. Cattle/buffalo provide animal traction 

during land preparation and manure (directly or indirectly) to 


enhance the soil. Rice straw, other rice by-products, and residues 

from other crops are fed to the animals in return. The 

degree of such interaction varies across countries and across 

areas within countries. In Khon Kaen, Thailand, for example, 

the use of cattle as draft power has been replaced by tractors 

and power tillers because of the combined effect of various 

factors such as the changing demand for crops and other products, 

the growing scarcity of the rural labor market, the decline 

in livestock population, etc. The degree of crop-animal 

interaction in Long An Province in Vietnam, which is close to 

Ho Chi Minh City, is less than that in An Giang, which lies 

close to the border of Cambodia. How does this trend affect 

rice production efficiency, especially among small rainfed farm 

areas 

Methodology 

The study employs the stochastic production frontier model 

popularized by Aigner et al (1977) to compare productive efficiency 

of different rice farm groups. The empirical form is a 

yield function expressed as 

n 

λn Y j = λn a + Σ b i λnX ij + e j 

i=1 

where j = 1 … n total number of samples; Y j represents the 

yield of the jth farm; X i = 1 × k vector of input quantities of the 

jth farm;a, b i = parameters to be estimated; e j = error term 

such that e j = ν j + µ j , where ν j = random error that is beyond 

the control of the production unit. 

Costs and returns in rice production 

Table 1 shows the average costs and returns in rice production 

of sample respondents at the different project sites. At most of 

the project sites, net income (above all costs) is shown to be 

highest among those belonging to Group 1. This is especially 

the case among the respondents in Cambodia and in Vietnam, 

in An Giang in particular. When the noncash costs are excluded, 

the net income of Group 1 in all countries (except in Indonesia 

and Thailand) is highest. Noncash costs account for up to a 

46% share of the total cost in Thailand, mostly the imputed 

cost of family labor. The imputed cost of the use of own animals 

for power or for manure accounts for from 2% to 18% of 

the total cost. It is in Indonesia where a large amount of animal 

manure is used in rice cultivation. 

Stochastic frontier production results 

Table 2 shows the results of the stochastic frontier production 

analysis. The yield function is expressed as a Cobb-Douglas 

function; hence, the coefficients are actually elasticities. A 

majority of the explanatory variables in the yield function have 

the right signs. Important to note are the coefficients of the 

dummy variables representing pest infestation, partial irrigation, 

and the use of high-yielding varieties, which indicate a 

substantial impact on rice yield. The negative sign for the area 

coefficient has been shown in other studies (Husain et al 2001). 

It indicates that farmers strive hard to increase yield to enhance 

production from a small piece of land. The only coefficient 

that is significant, but with the wrong sign, is that for the 

price of seed. The value here is imputed by farmers, who may 

have based it on the price of rough rice. The coefficient of 

animal days is negative but this is insignificant. The sign implies 

that farmers tend to be using more animal days than are 

required, which could be the case for farm households in Group 

1. Sample respondents reported having employed from 10 animal 

days (in Thailand, where tractors are used more extensively) 

to 56 animal days (in Cambodia) per hectare per year 

in rice cultivation. These numbers seem high compared to the 

normal use of animals that averages about 20 days per hectare 

per season or up to 40 days per hectare per year of two seasons 

(personal communication with V. Balasubramanian, IRRI, 

2004). The coefficients of the dummy variables for Groups 2 

and 3, respectively, are negative but insignificant. Nevertheless, 

they seem to confirm the beneficial effect of animal use 

in rice fields, especially in small marginal areas, in terms of 

both productivity and sustainability. 

Among the variables influencing technical efficiency, 

total income, the dummy for education, and the dummy for 

farming experience stand out. The estimated coefficients of 

these variables are significant and the negative sign indicates 

their tendency to reduce inefficiency. The negative sign of the 

coefficient for the small animal ownership dummy, although 

not significant, is related to these animals’ role as a quick source 

of cash/capital for the purchase of inputs. The negative coefficient 

of the Group 2 dummy, although not significant, further 

confirms its lesser tendency to bring about an efficiency improvement. 

The coefficient of the variable representing ownership 

of large animals is significant, but with a positive sign. This 

supports the conclusion derived from the yield regression portion 

regarding the seemingly extensive use of animals beyond 

the level that would further improve production efficiency. The 

training variable is estimated with a positive but insignificant 

coefficient. Only about 33% of the sample respondents at the 

project sites indicated having undergone training on crop production 

practices and technologies, be it on rice or other crops, 

and only 16% on animal production management techniques. 

The highest participation rate is reported in Vietnam. Deficiency 

in extension and training services is typical in remote 

and marginal areas because of a limited budget from the public 

sector. Private companies still do not have enough incentives 

to provide such investments. The coefficient of the credit 

variable in the technical inefficiency portion is significant, but 

with a positive sign. Capital for credit is often scarce in developing 

countries, much more so for small farmers because of 

the risk and cost involved. Farmers have often relied on moneylenders 

or private traders that advance funds until the crop 

is sold. This system of credit has not always been advantageous 

for improving productivity. It has rather led farmers into 

a vicious circle of debt, especially in unfortunate cases when a 


Table 1. Costs and returns in rice production across project sites (in US$ per ha). 

Indonesia Vietnam Philippines 

Item Cambodia Thailand 

East Java West Java All An Giang Long An All N. Ecija Samar All 

Group 1 a 

Yield 3.84 4.25 4.25 4.32 3.48 4.22 3.82 2.05 2.74 1.83 

Total returns 1,485.56 663.2 663.2 858.97 511.21 818.53 1031.3 571.91 751.29 233.76 

Total costs 487.53 615.47 615.47 589.23 415.43 569.02 516.03 321.15 397.25 344.28 

Imputed 203.88 370.71 370.71 223.05 181.51 218.22 68.85 106.73 91.93 229.63 

Cash 283.65 244.76 244.76 366.18 233.92 350.8 447.18 214.42 305.32 114.64 

Net returns above cash cost 1,201.91 418.44 418.44 492.79 277.28 467.73 584.12 357.49 445.97 119.12 

Net returns above total cost 998.03 47.73 47.73 269.74 95.77 249.51 515.27 250.76 354.04 –110.51 

Group 2 a 

Yield 3.76 4.36 5.15 4.25 3.21 3.93 3.78 3.88 2.05 3.66 2.64 

Total returns 1,453.57 626.79 1,080.29 675.82 671.9 566.11 588.15 967.56 359.83 893.89 340.53 

Total costs 746.34 498.94 501.05 499.17 601.81 332.5 388.6 445 174.54 412.21 307.71 

Imputed 109.35 207.53 102.19 196.14 210.22 91.85 116.51 37.3 30.07 36.42 145.97 

Cash 636.99 291.41 398.86 303.02 391.59 240.64 272.09 407.7 144.47 375.79 161.74 

Net returns above cash cost 816.58 335.38 681.43 372.79 280.31 325.46 316.06 559.86 215.36 518.1 178.79 

Net returns above total cost 707.23 127.85 579.24 176.65 70.09 233.61 199.54 522.56 185.29 481.69 32.82 

Group 3 a 

Yield 2.70 4.84 4.39 4.25 3.08 3.08 2.96 1.91 2.35 2.76 

Total returns 976.82 872.53 941.45 938.70 644.23 644.23 616.62 555.38 580.9 316.92 

Total costs 493.85 494.42 648.80 642.63 637.32 637.32 389.29 302.23 338.51 302.77 

Imputed 129.11 91.06 228.13 222.65 284.02 284.02 35.39 57.43 48.25 128.01 

Cash 364.74 403.36 420.67 419.98 353.30 353.30 353.90 244.80 290.26 174.76 

Net returns above cash cost 612.08 469.17 520.78 518.72 290.93 290.93 262.72 310.58 290.64 142.16 

Net returns above total cost 482.97 378.10 292.65 296.07 6.91 6.91 227.33 253.15 242.39 14.15 

Average all 

Yield 3.73 4.36 4.46 4.25 4.18 3.84 4.06 3.78 2.04 2.91 2.65 

Total returns 1,437.32 641.81 955.34 795.69 835.00 554.67 735.32 979.01 559.05 769.03 319.32 

Total costs 508.46 518.43 634.03 575.16 591.23 349.78 505.38 480.12 311.56 395.84 307.33 

Imputed 189.73 230.84 215.54 223.33 222.28 110.53 182.55 54.42 98.04 76.23 141.21 

Cash 318.73 287.59 418.49 351.83 368.95 239.24 322.83 425.7 213.52 319.61 166.12 

Net returns above cash cost 1,118.59 354.23 536.85 443.86 466.05 315.43 412.49 553.31 345.53 449.42 153.20 

Net returns above total cost 928.86 123.38 321.31 220.53 243.77 204.89 229.95 498.89 247.49 373.19 11.99 

a Group 1 includes farm households that own cattle or buffalo and use those animals in their rice cultivation. Group 2 includes farm households that own cattle and/or buffalo but that are not used in rice cultivation. 

Group 3 includes farm households that do not own any cattle or buffalo. 


Table 2. Results of the stochastic production frontier regression, IRRI-ILRI 

Project, 2004. 

Variables Coefficient t-value 

Yield regression component 

Constant 8.26769 146.81 

Farm area (ha) –0.17048 –9.83 

Seed use (kg ha –1 ) 0.00066 1.67 

Family labor (no. of days) 0.00004 0.17 

Hired labor (no. of days) 0.00015 1.45 

Animal power (no. of days) –0.00007 –1.39 

Machine use/rental (no. of days) 0.00011 2.88 

Fertilizer use (kg ha –1 ) 0.00029 7.55 

Dummy for pest infestation –0.10163 –2.56 

Dummy for partial irrigation 0.22159 6.94 

Dummy for use of high-yielding varieties 0.11421 2.28 

Dummy for dry season –0.01945 –0.64 

Dummy for Group 2 –0.05569 –0.97 


Technical inefficiency component 

Constant –15.55516 –3.23 

Total income ($ per household) –0.00822 –4.29 

Dummy for small animal ownership –0.25559 –1.08 

Dummy for large animal ownership 3.77959 3.16 

No. of male laborers to household size 0.00072 1.71 

Dummy for schooling beyond 6 years –2.63059 –3.32 

Dummy for farming experience beyond 20 years –1.49879 –3.41 

Dummy for participation in training 0.16356 0.63 

Dummy for credit availment 0.65659 2.69 


Dummy for Group 3 5.87091 3.14 

Mean efficiency, all groups 0.69 

Group 1 mean efficiency level 0.72 



Source: Food-Feed Systems Survey, 2002-03. 

crop fails. Larger farmers have better access to credit, often 

for reasons of provision of collateral and lower administrative 

costs of servicing larger loans. But their yields are often not 

higher than those of small farms. 

Also shown in Table 2 are the estimated technical efficiency 

coefficients across the three groups of rice farm households. 

The overall technical efficiency estimate for all the respondents 

is 0.69. It is highest among rice farm households 

belonging to Group 1 and lowest among farm households in 

Group 3. The differences, however, are not very large. 

Conclusions 

Crop-animal integration is beneficial from the standpoint of 

the savings on input expenditures, particularly on energy use 

in rice cultivation. The advantage of closer integration in terms 

of promoting production efficiency, however, was not so strong. 

The results seem to indicate the need to stress to farmers the 

proper techniques for the use of animals in land preparation to 

enhance their contribution to promoting production efficiency. 

Additional results gathered from the stochastic production frontier 

analysis also indicate that there is still some room for improving 

yield and efficiency in rainfed rice production. The 

recommendations indicated are nothing new: the provision of 

water and the development of high-yielding and pest-resistant 

varieties should continue to be on the priority list of the research 

and development agenda for these production environments, 

and public budgets for training and extension should 

be enhanced to reach remote and marginal areas. 

References 

Aigner DJ, Lovell K, Schmidt P. 1977. Formulation and estimation 

of stochastic frontier production models. J. Econometrics 

6(1):21-37. 

Devendra C. 1995. Environmental characterization of crop-animal 

systems in rainfed areas. In: Devendra C, Sevilla C, editors. 

Crop-animal interactions. IRRI Discussion Paper Series No. 

6. Manila (Philippines): International Rice Research Institute. 

p 43-64. 


Husain AMM, Hossain M, Janaiah A. 2001. Hybrid rice adoption in 

Bangladesh: a socio-economic assessment of farmers’ experience. 

Research Monograph Series No. 18. Bangladesh Rural 

Advancement Committee, Bangladesh. 

IRRI (International Rice Research Institute). 1995. Fragile lives in 

fragile ecosystems. Proceedings of the International Rice Research 

Conference, 13-17 Feb. 1995. Manila (Philippines): 

IRRI. 976 p. 

Thornton PK, Kruska RL, Henninger N, Kristjanson PM, Reit RS, 

Atieno F, Odero AN, Ndegwa T. 2002. Mapping poverty and 

livestock in the developing world. Nairobi (Kenya): International 

Livestock Research Institute. 

Notes 

Authors’ address: A.S. delos Reyes, assistant scientist II; A.J.B. 

Malabayabas, contractual researcher; M.A. Sombilla, consultant, 

International Rice Research Institute, Los Baños, Philippines. 

Determinants of agricultural diversification in Vietnam: 

changes at the farm level in the Mekong 

and Red River deltas 

Magnus Jirström and Franz-Michael Rundquist 

A central part of current economic development in Vietnam 

concerns the transformation of the agricultural sector and development 

of the rural economy. For a rice economy, like 

Vietnam’s, this process includes the challenge to diversify out 

of rice (cf. Jirström and Rundquist 1999, Taylor 1994). This 

paper focuses on the farm-level processes of agricultural diversification 

in Vietnam’s two rice-bowl areas: the Mekong 

River Delta (MRD) and the Red River Delta (RRD). 

Data, methodology, and organization of this paper 

This paper presents results from a research project 1 within 

which household panel survey data were collected on three 

occasions in the MRD—1996, 1999, and 2002—and on two 

occasions in the RRD—1999 and 2002. The selection of study 

areas and households was carried out through a multistage 

stratified random sampling method. In the MRD, data gathered 

from 180 households cover the time-span 1996-2002. In 

the Red River Delta, 189 farm households were surveyed in 

1999 and re-surveyed in 2002. Our analysis in this paper will 

be limited to what we call rice diversification changes measured 

as changes in the dependence on income from rice in the 

total household economy and agricultural production system. 

Land use and diversification characteristics 

of Vietnam’s delta regions 

Table 1 displays the significant difference in the size of landholdings 

in the two regions, with the RRD farmers operating 

farms only a third of the size of the farms of the Mekong farmers. 

In both regions, however, land use is dominated by rice, 

which is grown intensively, with a cropping index of 172% in 

the north and 219% in the south. In the Mekong, a rice intensification 

process raised the cropping index from 168 in 1996 to 

219 six years later. The Mekong farmers form part of a more 

general development of the 1990s in which farmers in several 

Mekong provinces have switched from double to triple rice 

cropping as a result of improved irrigation and drainage infrastructure 

(Choeng-Hoy Chung 1997). 

In the Red River Delta, a higher proportion of cropped 

land is allocated to upland, or dry, crops, a fact that is explained 

by the climatic conditions in the north not allowing 

rice cultivation in the winter season. In terms of crop diversification, 

Mekong farmers have made efforts to grow new crops 

to increase income. During the 1990s, many diversified into 

fruit production, but a combination of fluctuating and low product 

prices as well as pest- and water management-related problems 

seems to have discouraged farmers to the extent that the 

share of land allocated to fruit plantations dropped to half by 

2002. 

1 “Structural Transformation of Southeast Asia’s Agricultural Systems: Agricultural Diversification in Vietnam,” funded by the Swedish Research Council 

(Vetenskapsrådet). The project is carried out in cooperation with the Cuu Long Delta Rice Research Institute (CLRRI) in Can Tho and the Institute for Environment 

and Sustainable Development, Hanoi. 


Table 1. Agricultural and income characteristics in the Mekong River Delta (MRD) and 

Red River Delta (RRD). a MRD RRD 

Variables 

1996 2002 1999 2002 

Total area (ha) 172.8 183.5 54.9 54.1 

Cropped area (ha) 332.2 440.7 112.5 108.0 

Rice sown area (ha) 289.4 402.4 95.0 93.3 

Rice sown area (%) 87.1 91.3 84.4 86.4 

Upland crop sown area (ha) 18.4 23.6 15.7 12.4 

Upland crop sown area (%) 5.5 5.4 14.0 11.5 

Fruit crop area (ha) 24.4 14.8 1.8 2.3 

Fruit crop area (%) 7.4 3.4 2.0 2.0 

Average land operator per household (ha) 0.96 1.02 0.29 0.29 

Average total household income 14,378 16,925 17,126 21,174 

Average rice production income 7,733 7,749 4,505 3,360 

Average nonrice crop income 1,629 958 1,793 1,613 

Average noncrop agricultural income 221 1,703 4,479 4,669 

Average total off-farm income 4,965 6,515 6,419 12,414 

Average off-farm nonfarm income n.a. 4,045 6,097 11,072 

a All income values are expressed as 000 VND (Vietnamese dong) in constant 2002 values; US$1 = 

15,279 VND (2002). n.a. = not available. T-tests for paired samples show that the difference between the 

two periods for the MRD is significant at the 1% level for cropped area, rice sown area, and fruit crop area. 

Also, for the MRD, significant differences at the 5% level are found for total household income, noncrop 

agricultural income, and total off-farm income. In the case of the RRD, significant differences at the 5% 

level, or higher, are found for all income variables. For land use only, the changes in the upland crop sown 

area are significant, but this time at the 1% level. 

Source: Information is based on sample surveys comprising 180 farm households in the MRD and 189 

households in the RRD from a larger research project by the authors titled “Structural Transformation of 

Southeast Asia’s Agricultural System: Agricultural Diversification in Vietnam.” 

Turning our attention to household income and its composition, 

the aggregate numbers in Table 1 show that Mekong 

farmers, in spite of access to much larger landholdings, on 

average have approximately 20% lower income than farmers 

in the north. This finding is in line with the more general observation 

that rural inhabitants in the Mekong region become 

relatively poorer than in other parts of the country (Government 

of Vietnam 2000). The mentioned increasing intensity of 

land use in the MRD has not been able to reduce the increasing 

income gap between the two regions. In terms of dependence 

on rice income for total income, the regional difference 

is significant. In 2002, 46% of an average Mekong farm 

household’s income was derived from rice production, while 

an average Red River Delta household depended only to some 

16% on rice. The share of rice in the total income was nevertheless 

decreasing in both survey areas (54% in the MRD in 

1996 and 26% in the RRD in 1999). 

In spite of the rather dramatic area intensification of rice 

production in the Mekong, farmers earned approximately the 

same income from rice in 2002 as in 1996. The lower profitability 

of rice production in the MRD is also typical for the 

RRD, where income from rice cultivation from the same size 

of cropped land fell by some 25% from 1999 to 2002. 

Because of the mentioned difficulties in diversifying into 

fruit production, nonrice crop income has decreased by 4% in 

the MRD. On the other hand, noncrop agriculture (animal husbandry 

and aquaculture) has grown in the Mekong and now 

represents some 10% of total income. In the RRD, its share 

has fallen somewhat but remained relatively higher at about 

22% in 2002. 

The process of intersectoral diversification explains most 

of the growth in income in both regions. Off-farm income has 

grown in the MRD at a rate high enough for it to retain its 

relative importance—38% of total income. In the RRD, offfarm 

income has almost doubled, and farm households there 

now depend more on off-farm income—about 52% of total 

income—than on on-farm income. While the proximity to urban 

areas and income sources are more important in the RRD, 

farm families in the Mekong are more dependent on agriculture-related 

off-farm income—38% of total off-farm income 

in the MRD vis-à-vis 11% in the RRD. 

Behind the averages 

The farming and income characteristics discussed so far are 

based on mean sample values. The studied agricultural diversification 

process is not, however, a smooth process including 

all farm households in an equitable manner. In this second section 

of the paper, we try to capture some of the dynamics of 

the diversification process by analyzing and discussing four 

categories of rice farm households. 


Methodologically, changes are described by looking at 

specific categories of farm households and their “movement” 

in rice dependence between the two points in time. Rice dependence 

has been defined as the proportion of income from 

rice in the household total income. These proportions have 

been categorized into four quartiles (0–25%, 25–50%, etc.). 

In the second stage, changes in category affiliation were examined, 

and several new categories were defined. First, those 

who stayed in the first quartile at both times were classified as 

Stable high diversifiers (households displaying a high level of 

diversification, i.e., low dependence [d”25%] on rice income 

in their total income). The opposite group (those who showed 

a high dependence [>75%] on rice income in their total income) 

was classified as Stable low diversifiers. Second, those 

who changed by two or three quartiles were classified as Significant 

ascenders and Significant descenders, respectively, 

depending on the direction of their changes. Finally, those who 

changed only one quartile between the two points in time were 

classified as “Moderate changers.” In the presentation (Table 

2) and discussion below, we focus on only four of these categories—Stable 

high diversifiers, Stable low diversifiers, Significant 

ascenders, and Significant descenders representing 

the more “spectacular” groups of the “stayers” and “movers.” 

Our data confirm a picture of significantly more diversified 

farm households in the RRD. In 2002, only 10 out of 189 

households depended on rice production for more than 50% 

of their total income. In the MRD, the situation is more dynamic, 

with approximately a fifth of the sample being involved 

in rapid diversification out of rice. 

Among the four categories, the Stable high diversifiers 

at both times and in both regions show the highest household 

income. However, the group Significant ascenders is rapidly 

catching up, with a doubling—MRD 104%, RRD 115%—of 

total household income. Significant descenders, on the other 

hand, are characterized by falling total income. Finally, in the 

Mekong, it is possible to discern a group of households—Stable 

low diversifiers—characterized by above average and increasing 

size of landholdings following a rice specialization path. 

This group, which depends almost completely on rice cultivation, 

has greatly increased its total income to a level well above 

the sample average. 

Determinants and patterns of diversification 

Searching for determinants of the observed differences in diversification, 

several variables were selected—access to family 

labor, education, farm size, and sown area. The composition 

of income may in itself not only be an effect of the diversification 

process, but also may determine its direction and 

speed for individual households. This section focuses on the 

MRD region, where the share of farm households experiencing 

significant changes in the composition of income is relatively 

greater than in the already more diversified RRD region. 

Furthermore, only in the RRD is there a group of farmers 

following a nondiversification path. Starting with the households 

experiencing the most dramatic changes in terms of the 

level of income diversification, the following observations can 

be made for the four categories. 

Significant ascenders as a group started out in 1996 at 

an income level well below the average total sample (cf. Table 

1) but were operating slightly above average landholdings. During 

the period, they were able to allocate more of their total 

harvested area to nonrice crops and were successful in raising 

their income from this source. Also, the increasing noncrop 

agricultural income contributed to their overall success, which 

primarily was explained by a dramatic increase in off-farm 

income. 

Significant descenders were, in terms of total income, 

starting from a relatively higher level than the previous group. 

However, their access to land was only two-thirds of that of 

the ascenders and, as their off-farm income dropped, they were, 

in spite of some increase in land, and a significant increase in 

cropping index from 166 to 255, not able to compensate for 

the negative change. The area used for nonrice crops did not 

change, but their net income from this source turned negative. 

This was due to the problems of pest attacks and water management 

related to fruit production. 

Stable high diversifiers enjoy high income from all income 

sources but rice. During the period, these households 

allocated somewhat less land to nonrice crops (from 0.54 to 

0.40 ha) but were instead able to increase their noncrop agricultural 

income. Over the six-year period during which the 

majority of households in the total sample intensified their 

cultivation by growing a third (summer-spring) rice crop, the 

Stable high diversifiers group only marginally increased the 

cropping index of its land. This then seems well in line with a 

strategy of diversifying both agricultural and nonagricultural 

income. 

Stable low diversifiers seem to follow a rice specialization 

strategy. They have significantly increased their landholdings, 

and, with a cropping index of 266, they are the most intensive 

land users. An important question is whether such a 

strategy is sustainable over time. The possibility to further increase 

the intensity of land use is very limited, and only a rising 

price of rice will keep this group on a par with an economy 

experiencing overall economic growth. 

Data in Table 2 do not reveal any strong links between 

the variables education and household labor, on the one hand, 

and the diversification changes, on the other. Apart from the 

striking difference in the level of education between heads of 

household in the RRD and MRD regions, this variable does 

not differ significantly between the diversification categories. 

Conclusions 

A household’s diversification process is complex. Our data 

tally with other data sources (e.g., Government of Vietnam 

1994, 2000) pointing at an ongoing agricultural diversification 

process in Vietnam. The change in the RRD especially 

seems more unidirectional. In the Mekong, where rice dominates 

to a greater extent, different strategies—whether adopted 

by preference or not remains to be analyzed—seem to co-ex- 


Table 2. Diversification changes in the Mekong River Delta (MRD) and Red River Delta (RRD) for rice diversification changes. a 

Stable high diversifiers Significant ascenders Significant descenders Stable low diversifiers 

Item MRD RRD MRD RRD MRD RRD MRD RRD 

1996 2002 1999 2002 1996 2002 1999 2002 1996 2002 1999 2002 1996 2002 1999 2002 

Number of households 16 16 70 70 34 34 12 12 7 7 3 3 20 20 0 0 

Total household income 21,353 24,164 23,332 27882 10,298 20,986 9,713 20,883 14,101 10,040 11,804 4,647 11,928 20,093 – – 

Rice production income 2,431 2,520 3,664 2,655 9,188 5,416 5,737 2,910 3,814 8,939 1,508 3,000 12,040 18,702 – – 

Nonrice crop income 6,496 3,894 2,311 1,594 124 1,133 1,255 1,736 1,551 -699 836 366 404 –176 – – 

Noncrop agricultural 3,117 4,862 6,513 7,674 –730 1,766 1,928 3,719 510 664 1,701 265 –1,208 461 – – 

income 

Off-farm income 9,309 12,888 10,844 15,959 1,716 12,671 793 12,518 8,226 1,136 7,759 1,016 692 1,106 – – 

Education (no. of y) 6.1 5.0 10.1 10.1 4.2 4.9 9.3 9.3 4.1 5.2 8.0 8.0 6.2 5.3 – – 

Labor, family 15–65 y 2.9 3.7 3.5 3.4 3.0 4.4 4.3 4.0 2.6 4.0 3.0 2.7 3.2 4.2 – – 

Total sown area (ha) 1.08 1.38 0.49 0.47 1.86 1.92 0.69 0.60 1.11 1.89 0.46 0.62 2.66 4.79 – – 

Rice sown area (ha) 0.56 0.98 0.42 0.41 1.77 1.76 0.60 0.52 0.79 1.59 0.41 0.58 2.51 4.72 – – 

Land operated (ha) 0.72 0.85 0.26 0.24 1.05 0.86 0.33 0.30 0.67 0.74 0.28 0.31 1.38 1.80 – – 

a All income values are expressed as 000 VND (Vietnamese dong) in constant 2002 values; US$1 = 15,279 VND (2002). T-tests for paired samples show that the difference between the two periods for the MRD is significant at 

the 1% level for rice area for Stable high diversifiers. Similarly, significant differences at the 1% level are found for Significant ascenders in the MRD for total household income, rice production income, noncrop agricultural income, 

and labor. Finally, for the MRD, a 1% level of significance is found for Stable low diversifiers for rice production income, noncrop agricultural income, total sown area, and rice sown area. Significant differences at the 5% level for 

the MRD are found for education (Stable high diversifiers); education (Significant descenders); total household income and labor (Stable low diversifiers). Significant differences at the 10% level in the MRD are found for labor and 

total sown area (Stable high diversifiers); nonrice crop income (Significant ascenders); rice production income, nonrice crop income, and rice sown area (Significant descenders); and land (Stable low diversifiers). In the RRD, 

significant differences at the 1% level are found for total household income, rice production income, and off-farm income for Significant ascenders. Also, significant differences at the 5% level for the RRD are found for off-farm 

income (Stable high diversifiers) and for total sown area and rice sown area (Significant descenders). Finally, significant differences at the 10% level are found for the RRD for total household income, total sown area, and rice sown 

area (Stable high diversifiers); rice area (Significant ascenders); and noncrop agricultural income (Significant descenders). 

Source: Information is based on sample surveys comprising 180 farm households in the MRD and 189 households in the RRD from a larger research project by the authors titled “Structural Transformation of Southeast Asia’s 

Agricultural Systems: Agricultural Diversification in Vietnam.” 


ist. The next step in the search for determinants of diversification 

in this project will be to break down the material further 

in order to analyze whether differences in “geographical capital” 

reveal any spatial diversification patterns at the district 

and village levels (Sen 2003). Differences in physical access 

to product and labor markets as well as varying agroecological 

conditions at the village level will be research issues in this 

next step. 

References 

Choeng-Hoy Chung. 1997. Agricultural growth within a strategy for 

sustainable rural development in Vietnam. Vietnam’s Soc. 

Econ. Dev.: Soc. Sci. Rev. 11:10-29. 

Government of Vietnam. 1994. Vietnam Living Standard Survey 

(VLSS) 1992/1993. Hanoi (Vietnam): General Statistical Office. 

Government of Vietnam. 2000. Vietnam Living Standard Survey 

(VLSS) 1997/1998. Hanoi (Vietnam): General Statistical Office. 

Jirström M, Rundquist F-M. 1999. Diversify or perish! Structural 

transformation of agricultural systems: intensification and 

diversification in Vietnam. In: Cederlund K, Friberg T, Wikhall 

M, editors. Geografi I Lund, Essäer tillägnade Gunnar 

Törnqvist, Institutionen för kulturgeografi och ekonomisk 

geografi, Lunds Universitet. p 74-87. 

Sen B. 2003. Drivers of escape and descent: changing household 

fortunes in rural Bangladesh. World Dev. 31(3):513-534. 

Taylor DC. 1994. Agricultural diversification: an overview and challenges 

in ASEAN in the 1990s. ASEAN Econ. Bull. 10(3):264- 

279. 

Notes 

Authors’ address: Department of Social and Economic Geography, 

Lund University, Sölvegatan 12, S-223 62 Lund, Sweden. 

Growth of the rural nonfarm economy in Bangladesh: 

determinants and impact on poverty reduction 

Mahabub Hossain 

The role of nonfarm activities in promoting growth of the rural 

economy and reducing poverty is well documented (Hymer 

and Resnic 1969, Shand 1986, Ranis and Stewart 1993, 

Rosegrant and Hazell 2000). Termed the rural nonfarm 

economy (RNFE), this sector accounts for a large proportion 

of rural employment and income, and grows faster than agriculture 

with the development of the overall economy. As 

Rosegrant and Hazell (2000; p 81) observe, “From relatively 

a minor sector, often largely part-time and subsistence-oriented 

at the early stages of development, the rural nonfarm economy 

develops to become a major motor of economic growth in its 

own right, not only for the countryside but for the economy as 

a whole. Its growth also has important implications for the 

welfare of women and poor households, sometimes helping to 

offset inequities that can arise within the agricultural sector.” 

Generating productive employment for the growing labor 

force remains a formidable challenge for the Bangladesh 

economy. The capacity to absorb the incremental rural labor 

force in agriculture is extremely limited because of (1) no scope 

for expansion of the land frontier, (2) the intensity of cropping 

has almost reached the limit, and (3) the growth of crop production 

now depends almost entirely on technological progress, 

resulting in low employment elasticity of output. Recent censuses 

and labor-force surveys show a dramatic structural change 

in the composition of the rural labor force in favor of nonfarm 

activities (Hossain et al 2002). Questions arise, however, about 

whether the expansion is due to “push” or “pull” factors that 

determine whether it is a positive development for poverty 

reduction (Islam 1984, Hossain et al 1994). 

This paper uses primary data available from a recent 

survey of a nationally representative sample of rural households 

to (1) study the nature of the RNFE in Bangladesh, (2) 

analyze the determinants of participation in the RNFE, and (3) 

assess the impact of participation on a reduction in rural poverty. 

Data and methodology 

The data for the study are drawn from a repeat survey of a 

nationally representative sample. The benchmark survey was 

implemented in 1987-88 by the Bangladesh Institute of Development 

Studies (BIDS) on 1,245 rural households from 62 

villages in 57 districts, drawn by using a multistage random 

sampling method. In the first stage, 64 unions were randomly 

selected from the list of all unions in the country. In the second 

stage, one village was selected from each of the unions that 

best represented the union with regard to landholding size and 

literacy rate. A census of all households in the selected villages 

was conducted to stratify the households with regard to 

the size of landownership and land tenure. A random sample 

of 20 households was drawn from each village. The International 

Rice Research Institute (IRRI) studied the same villages 

in 2000-01. A sample of 30 households from each village was 

taken using the stratified random sampling method. The stratification 

was based on the wealth ranking technique of the par- 


ticipatory rural appraisal (PRA) method. The 2000-01 sample 

included households and their descendents covered in the 1987- 

88 survey. The author supervised the implementation of both 

surveys. 

This paper uses a narrow definition of the RNFE that 

includes only nonagricultural activities. We distinguish three 

types of such activities: 

Manual labor-based activities, such as self-employment 

in cottage industries, mechanics, wage employment 

in rural business enterprises, transport operations, 

and construction labor (nonfarm labor); 

Human capital-based occupations, such as salaried 

service in public- and private-sector institutions, 

teachers, religious leaders, lawyers, village doctors, 

and various types of personal services (services); and 

 

Physical and human capital-based activities, such as 

agro-processing, shopkeeping, peddling, petty trading, 

medium- and large-scale trading, and contractor 

services (business enterprises). 

We analyzed the factors influencing the participation of 

rural households in the RNFE to examine whether the expansion 

is caused by push or pull factors. The proportion of household 

income derived from the RNFE was used as a measure of 

the intensity of participation, which was related to several explanatory 

variables using multivariate regression. Since the 

value of the dependent variable is truncated at both ends (from 

0 to 100), the TOBIT model was used for estimating the values 

of the parameters. The model was estimated separately for 

the three groups of RNFE activities mentioned above. 

An income determination function was estimated with 

household-level data to assess the contribution of different 

factors to the growth of rural nonfarm income. The incomeearning 

capacity of the household would obviously depend on 

the size of land owned and operated, the number of earning 

members in the family, and the amount of nonland fixed assets 

used in productive activities. The productivity of family workers 

and the choice of economic activities would depend on the 

quality of labor embodied through investment in human resources, 

particularly education and training. The location of 

the village with regard to infrastructure such as roads and electricity 

connections would be an important determinant of the 

productivity of assets and profitability of enterprises, as the 

development of infrastructure opens up opportunities for nonfarm 

employment and processing, storage, and marketing of 

agricultural products (Ahmed and Hossain 1990). The explanatory 

variables of the model were selected in view of the above 

considerations. 

The impact of the growth of the RNFE on poverty was 

assessed by measuring a class of poverty indices (Foster et al 

1984)—the head count index, the poverty gap index, and the 

squared poverty gap index—and comparing them for households 

classified by involvement in the RNFE. 


Characterization of the rural nonfarm economy 

In 2000, 52% of the earning members of the households reported 

nonagricultural activities as the primary occupation and 

another 10% as the secondary occupation. The corresponding 

numbers estimated from the 1987 survey were 34% and 15%, 

respectively. In 2000, 30% of the workers reported these activities 

as a secondary occupation, substantially lower than the 

41% reported in 1987. In 2000, the major nonfarm activities 

were Services, accounting for 22% of all rural employment, 

followed by Nonfarm labor comprising 16%. Business enterprises 

accounted for 14% of employment but 23% of income 

because of higher productivity. 

The level of labor productivity is a good indicator of the 

strength of the RNFE. If labor productivity were lower than 

the agricultural wage rate, it would support the hypothesis of 

the operation of push factors behind the expansion of the RNFE. 

Higher labor productivity, on the other hand, is evidence of 

the existence of pull factors. The estimates of labor productivity 

obtained from the resurvey show that compared to agricultural 

wage rate the productivity is 10% to 40% higher for nonfarm 

labor and two to 3.5 times higher for Services and Business 

enterprises. The average productivity in the RNFE increased 

from US$1.43 per day in 1987 to $2.28 in 2000, an 

increase of 3.6% per year. The evidence thus supports the 

proposition of the existence of pull factors, that higher productivity 

and wage earnings in most nonfarm activities are luring 

labor from relatively low-productive, risky, and backbreaking 

farm activities. 

Factors affecting participation in the RNFE 

The parameters of the TOBIT model estimated to identify factors 

affecting participation in nonfarm activities are reported 

in Table 1. The following points can be noted from the findings. 

Participation in manual labor-based activities appears 

to be poverty-driven. The intensity of participation in nonfarm 

labor is negatively associated with the size of ownership of 

land and nonland fixed assets, and the level of education of 

the workers. The negative coefficient of the technology variable 

indicates that adoption of high-yielding rice varieties reduces 

the pressure on participation in the nonagricultural labor 

market. The negative association with age indicates the 

preference of the younger generation for nonfarm jobs over 

arduous agricultural wage labor. The coefficient of the dependency 

ratio shows a positive association of participation with 

subsistence pressure. 

Education is the most important factor affecting participation 

in Services. The larger size of landownership, access to 

land in the tenancy market, and the intensity of the adoption of 

modern varieties seem to reduce pressure on participation in 

Services (presumably for those at the lower end of the productivity 

scale), as indicated by the statistically significant negative 

coefficients for these variables. The negative coefficients 

of the variable representing number of workers and subsis- 


Table 1. Factors affecting participation in rural nonfarm activities: estimates of a TOBIT 

model. a 

Factors Business Services Nonfarm 

enterprises labor 

Size of land owned (ha) –2.668 –10.697* –50.464* 

(–1.44) (–4.39) (–7.19) 

Area under tenancy (% of holding) –0.095 –0.226* 0.050 

(–1.86) (–3.80) (0.73) 

Age of the household head (y) –0.516* 0.504* –0.966* 

(–3.59) (3.38) (–4.81) 

Household workers (no.) 9.935* –3.002 23.642* 

(5.07) (–1.39) (7.49) 

Dependency ratio (consumers/workers) 4.328* –3.659* 1.833 

(3.67) (–2.82) (1.09) 

Average education of worker (y of schooling) 0.678 5.011* –3.925* 

(1.57) (10.63) (–5.86) 

Value of nonland fixed assets (thousand Taka) 0.071* 0.009 –0.833* 

(7.31) (0.86) (–5.24) 

Coverage of modern rice varieties 0.033 –0.102* –0.173* 

(% of cultivated area) (1.29) (–3.48) (–4.28) 

Status of infrastructure development 11.966* 3.159 7.670 

(villages with developed infrastructure (3.47) (0.83) (1.63) 

= 1) 

Constant term –44.417* –50.154* –10.345 

(–4.90) (–5.02) (–0.85) 

Sigma 61.288* 63.887* 72.213* 

(30.57) (27.85) (25.91) 

Log likelihood function –4,158 –3,548 –3,000 

a 

The dependent variable is measured as the share (%) of the nonfarm activity in total household income. 

Numbers within parentheses are asymptotic “t” values. * denotes statistical significance at the 1% probability 

of error. 

Source: Estimated from household-level data from the BIDS-IRRI surveys. 

tence pressure indicate a smaller family size in households 

engaged in service-sector activities. 

Major determinants for participation in Business enterprises 

are the accumulation of nonland assets, a larger number 

of workers in the household, and access to developed infrastructure. 

Participation is higher in younger households. It is 

interesting to note that the intensity of the adoption of highyielding 

rice varieties and the level of education of the worker 

do not exert a significant influence on participation in trade 

and business activities. Even the less-educated involve themselves 

in business provided they have access to capital. 

Contribution of the RNFE to income 

and reduction of poverty 

The average household income is estimated from the survey at 

$1,232 for 2000, a growth of 2.2% per year over 1987-2000. 

The growth in rural income was almost entirely on account of 

nonagricultural activities. The fastest-growing economic activities 

were business and services, followed by nonrice farm 

labor. The income from rice production and agricultural wage 

labor declined in absolute terms. The share of nonagriculture 

in total household income grew from 42% in 1987 to 54% in 

2000. 

The following regression equation best represents the 

variation in income earned from nonagricultural activities 

across the sample households: 

INCM = 3.56 + 399 WRKR + 0.22 CPTL + 26 EDCN (1) 

(0.90) (20.64) (35.11) (6.30) 

+ 403 INFR + 319 MIGRN 

(7.14) (6.57) 

+ 0.29 MIGRN*CPTL R-square = 0.66, F 

(16.84) 

= 607, N of cases = 1,887 

where INCM is the amount of income from nonfarm sources 

(US dollars), WRKR is the number of family workers engaged 

in nonfarm activities, CPTL is the value (US$) of nonland fixed 

assets, EDCN is the years of schooling for the worker, INFR is 

a dummy variable for villages with access to paved roads and 

electricity, MIGRN is a dummy variable for households with a 

migrant member, and MIGRN*CPTL is an interaction term 

with migration and nonland fixed assets. The numbers within 

parentheses are the estimated t-values for the regression coefficient. 

The value of R-square indicates that about 66% of the 

variation in nonagricultural income across households is explained 

by the variables included in the model. 

Since the equation is estimated in linear form, the value 

of the regression coefficient shows the marginal returns from 

the factor. Thus, a worker employed in nonagriculture earns 

on the margin $399 per year, and the marginal rate of return on 

investment in nonagricultural capital is 22%. A worker earns 

an additional income of $26 for each year of schooling. House- 


Table 2. Estimates of poverty: by primary occupation of the household head and size of land owned, 2000. 

Land owned (ha) Head count index a Poverty gap index a Squared poverty gap index a 

Agriculture. Nonagriculture Agriculture. Nonagriculture Agriculture. Nonagriculture 

With up to 0.2 ha 70.1 46.0 32.2 13.0 18.5 5.5 

0.2 to 1.0 ha 40.5 11.2 14.4 2.9 7.3 1.1 

Over 1.0 ha 11.1 0.0 3.1 0.0 1.2 0.0 

All households 48.3 30.9 20.3 8.7 11.2 3.6 

a For methodology of estimating poverty, see Foster et al (1984). 

Source: Own estimates based on household-level data from IRRI sample survey. 

holds with migrant members earn an additional $319. If the 

village had electricity connections and access to paved roads, 

the household earns an additional $403. The statistically significant 

regression coefficient of the interaction variable of 

migration with capital indicates that remittances help the accumulation 

of nonagricultural capital, and that the capital is 

invested in activities that give higher returns (29%). 

The contribution of different factors to income is estimated 

by dividing marginal returns by the average returns for 

specific factors (elasticity). It is found that 51% of the nonagricultural 

income is on account of nonagricultural labor, 19% 

on account of capital, 15% on account of education, and 8% 

on account of infrastructure. There was a strong association of 

nonagricultural employment with education and infrastructure. 

Thus, investment for education and infrastructure development 

contributes significantly to the growth of nonfarm income. Does 

engagement in the RNFE help reduce poverty To answer this 

question, we estimated various measures of poverty and compared 

them for households in which at least one member is 

employed in nonagricultural activities with those that depend 

entirely on agriculture for livelihoods. Since land is a dominant 

asset in the rural economy, we also made similar comparisons 

for different landownership groups, to dissociate the 

effect of differential endowment of land. Table 2 reports the 

findings. For the entire sample, the households with income 

below the poverty line were estimated as 43% (head count 

index). The poverty ratio was 31% for households engaged in 

nonagricultural activities vis-à-vis 48% for households engaged 

exclusively in agriculture. The difference is more pronounced 

for measures representing intensity and severity of poverty. 

Thus, diversification into nonagriculture contributes to a substantial 

reduction in poverty. The conclusion holds when the 

effect of the differential endowment of land is controlled. 

Conclusions 

The capacity of agriculture to generate productive employment 

and to provide a decent standard of living is becoming 

increasingly limited in Bangladesh. Rural households recognize 

this constraint and have been addressing it by using the 

agricultural surplus for investment in education and accumulation 

of nonagricultural capital to facilitate occupational mo- 

bility from farm to nonfarm activities. As a result, the rural 

nonfarm sector has been expanding and has already become a 

major component of the rural economy. The findings in this 

paper demonstrate that the RNFE in Bangladesh is a vibrant 

subsector contributing to productivity growth and reduction 

of poverty. Greater public-sector investment in education and 

expansion of the physical infrastructure will accelerate the process 

of expansion of the nonfarm sector, and contribute to the 

sharing of benefits with resource-poor households. 

References 

Ahmed R, Hossain M. 1990. Development impact of rural infrastructure 

in Bangladesh. IFPRI Research Report No. 83. Washington, 


Foster JE, Greer E, Thorbeck E. 1984. A class of decomposable poverty 

measures. Econometrica 52(3):761-766. 

Hossain M, Rahman M, Bayes A. 1994. Rural non-farm economy in 

Bangladesh: a dynamic sector or a sponge of absorbing surplus 

labor SAAT Working Paper, International Labor Organization, 

New Delhi. 

Hossain M, Bose ML, Chowdhury A. 2002. Changes in agrarian 

relations and livelihoods in rural Bangladesh: insights from 

repeat village studies. In: Ramachandran VK, Swaminathan 

M, editors. Agrarian studies. New Delhi (India): Tulika Books. 

p 369-391. 

Hymer S, Resnick S. 1969. A model of an agrarian economy with 

non-agricultural activities. Am. Econ. Rev. 59(4):493-506. 

Islam R. 1984. Non-farm employment in rural Asia: dynamic growth 

or proletarization J. Contemp. Asia 14:306-324. 

Ranis G, Stewart F. 1993. Rural non-agricultural activities in development: 

theory and application. J. Dev. Econ. 40:175-201. 

Rosegrant MW, Hazell PBR. 2000. Transforming the rural Asian 

economy: the unfinished revolution. Hong Kong: Oxford 

University Press. 

Shand R. 1986. Off-farm employment in the development of rural 

Asia. Australian National University, Canberra. 

Notes 

Author’s address: Economist and head, Social Sciences Division, 

International Rice Research Institute, DAPO Box 7777, Metro 

Manila, Philippines, e-mail: m.hossain@cgiar.org. 


SESSION 15 

Challenges to expanding rice production 

in unfavorable environments 

CONVENER: S. Tobita (JIRCAS) 

CO-CONVENER: R. Lafitte (IRRI)

In vitro selection of somaclonal and gametoclonal variants 

for salt tolerance in rice 

Nguyen Thi Lang, Dang Minh Tam, Hiromi Kobayashi, and Bui Chi Buu 

Tissue-cultured cell lines can be selected in vitro for resistance 

to various stresses. Selection is done by placing a stresscausing 

agent in tissue cultures containing dividing cells. Tissue 

culture techniques have been widely used for breeding 

purposes, especially in selection for stress tolerance. Tissue 

culture is a source of genetic variability that gives rise through 

genetic modifications during the process of in vitro culture to 

a phenomenon called somaclonal variation. The possible causes 

of somaclonal variation include chromosome aberrations, DNA 

amplification, and the occurrence of transposable elements. 

Salt tolerance is an important plant character in areas where 

seedling growth is a problem. This paper reports on the current 

status of tissue culture technology at the Cuu Long Delta 

Rice Research Institute (CLRRI). Our primary objective is to 

demonstrate that useful germplasm can be obtained from tissue 

culture and another way found to produce salt-tolerant lines 

using anther culture techniques in which anthers of F 1 hydrids 

having one or both parents with salt tolerance are cultured. A 

significant feature of this work is to determine what in vitro 

treatment, including concentration of the stressing agent and 

the period of time in culture, will produce the greatest probability 

of field-tolerant plants for a specific environment. This 

study aimed to create genetic variability in rice cell culture to 

obtain salt-tolerant plants through in vitro selection. 


Twelve rice cultivars were used to conduct in vitro selection 

for salt stress. The work on salt stress tolerance was done at 

concentrations of 0.5%, 1.0%, and 1.5% NaCl. 

Mature seeds were dehusked and surface-sterilized with 

70% ethyl alcohol for 30 sec and then for 30 min with 5% 

sodium hypochlorite (commercial bleach is mainly hypochlorite). 

Surface-sterilized seeds were rinsed several times with 

sterile distilled water before inoculation on callus induction 

medium, consisting of MS basal organic and inorganic components 

(Murashige and Skoog 1962) supplemented with 3.0% 

sucrose, 2,4-D at 1.0–2.0 mg L –1 , and kinetin at 0.0–0.5 mg 

L –1 (Nguyen Thi Lang 2002). 

The cultures were incubated under both dark and light 

conditions (16 h day/8 h night). The temperature was maintained 

at 22–25 ºC. The MS culture medium was supplemented 

with 3.0% sucrose, 0.8% agar, and various concentrations of 

growth regulators. The medium pH was adjusted to 5.8 prior 

to autoclaving at 15 psi for 20 min. 

The basal medium for rice tissue culture used is similar 

to that developed by Murashige and Skoog (1962). Unlike 

many other crops, rice requires a unique growth regulator combination, 

sucrose concentration, and light regime for each cultivar. 

Culture vessels are either glass screw vials or jars. The 

vials contain 10 mL of medium and are used for callus initiation. 

The jars containing 20 mL of medium are used after embryogenic 

callus, which is isolated and ready for plant regeneration. 

For callus induction, mature seeds are surface-sterilized 

and placed on the appropriate medium for a cultivar. At the 

end of 28 days—referred to as a “passage”—the original explant 

and all of its associated calli are transferred to a fresh 

medium. After the second passage, embryogenic calli are isolated 

and transferred again to a fresh medium. Embryogenic 

calli are usually yellow or cream-colored and dense in appearance. 

For long-term maintenance, embryo calli must be carefully 

selected and all nonembryogenic and dead calli removed 

at each transfer. 

Three crosses were tested for callus induction and plant 

regeneration. The F 1 seeds were grown in pots in the 

screenhouse, and panicles were collected at the stage when 

the auricle distance of the flag leaf to that of the subtending 

leaf was 5–10 cm. The anthers were plated in various callusinduction 

media when the pollen grains were at the mid-uninucleate 

to early binucleate stage of development. The plated 

anthers were kept in cold at 10 °C for 8 d and were then transferred 

under dim light at 25 °C until calli formed. 

Calli of about 1 mm in diameter were plated in N 6 + 2 

mg L –1 2,4-D + 1 mg L –1 NAA medium and incubated under 

1,000 lux at 25 °C for 4 wk. The calli were then transferred to 

MS + 1 mg L –1 kinetin + 0.5 mg L –1 NAA + 2 mg L –1 BA 

medium, until the plantlets were ready for transfer to normal 

growth conditions. 

The plantlets were grown in nutrient solution for 2 wk. 

The plantlets from each callus were individualized, with each 

one considered as a line. These were then transferred to pots 

and grown in the phytotron until maturity. 

Seeds were then surface-sterilized with 0.1% HgCl 2 for 

2–3 min and rinsed thoroughly with distilled water. The sterilized 

seeds were soaked in water for 24 h and incubated for 48 

h at 30 ºC. The pregerminated seeds were sown, two seeds per 

hole and 10 holes per variety, on a styrofoam sheet with 100 

holes with a nylon net bottom. The sheet was floated in distilled 

water for 4 d in nutrient solution in a plastic tray and 

salinized to EC = 6 dS m –1 by NaCl. After 4 d, the seedlings 

were salinized to EC = 12 dS m –1 by adding NaCl to the nutrient 

solution. The culture solution was renewed weekly and its 

pH was maintained daily at 5.0 by adding either 1 N NaOH or 


HCl. After 2 wk of salinization with EC = 12 dS m –1 , salinity 

symptoms were scored using a modified standard evaluation 

system (SES). 

Results 

Table 1. Plant regeneration of rice calli under NaCl stress. 

Variety 

% callus regeneration 

0.5% NaCl 1% NaCl 1.5% NaCl 

Pokkali 100.0 100.0 51.3 

AS996 100.0 55.6 68.8 

Nep Ao Gia 100.0 100.0 60.0 

Trang Diep 92.9 74.2 55.6 

Trang Thai Lan 100.0 100.0 73.1 

Mong Chim Roi 100.0 100.0 63.6 

KDM 105 75.0 92.9 44.1 

Doc Phung 100.0 100.0 64.0 

Doc Do 100.0 100.0 75.0 

Soc Nau 100.0 100.0 91.8 

IR29 (check) 100.0 78.6 47.6 

Somatic mutation 

Since the first successful plant regeneration by tissue culture, 

many variant plants have been obtained through somaclonal 

variation. 

Time for callus initiation was from 5 to 10 days after 

placing seeds on agar medium. The ability to form calli of 

cultivars on various media showed different results: 

1 Very good I MS + 2,4-D (2 mg L –1 ) 

2 Good II MS + 2,4-D (1 mg L –1 ) 

3 Medium III N 6 + 2,4-D (2 mg L –1 ) 

4 Average IV N 6 + 2,4-D (1 mg L –1 ) 

5 Dead 

In the effect of media on embryo callus formation, the 

medium MS + 2,4-D (2 mg L –1 ) was not significantly different 

from other media. For cultivar effect, Doc Do was considered 

as the best variety, followed by Pokkali, Soc Nau, Nep Ao 

Gia, Mong Chim Roi, and KDM 105, which were also viewed 

as more suitable genotypes for embryogenic callus formation 

ability than others. 

In the effect of media and cultivar interaction, the treatments 

Doc Do × MS + 2,4-D (2 mg L –1 ) and Pokkali × MS + 

2,4-D (2 mg L –1 ) obtained very good results. Trang Thai Lan 

× MS + 2,4-D (2 mg L –1 ) and KDM 105 × MS + 2,4-D (2 mg 

L –1 ) obtained good results on embryogenic callus formation 

ability. 

The effect of various concentrations of 2,4-D medium 

on callus induction showed that callus initiation was observed 

after 3 d of culture in all cultivars, irrespective of the light or 

dark period. 

The result showed that medium MS + 2,4-D (2 mg L –1 ) 

was more suitable to embryogenic callus formation than the 

others. 

The effect of cultivars on embryogenic callus formation 

percentage indicated that Doc Do, Soc Nau, Nep Ao Gia, and 

AS996 were genotypes that exhibited more suitability to develop 

in cultural medium than others. 

For interactions on embryogenic callus formation percentage, 

treatments Doc Do × MS + 2,4-D (1 mg L –1 ) and Nep 

Ao Gia × MS + 2,4-D (2 mg L –1 ) obtained a percentage of 

embryogenic callus formation of 86.7% and 80%, which were 

significantly different at the 0.01 probability level, respectively. 

In conclusion, Doc Do, Nep Ao Gia, and other genotypes 

such as Pokkali and Soc Nau were also suitable for embryogenic 

callus formation. The optimum medium for embryogenic 

callus formation was MS + 2,4-D (2 mg L –1 ). 

Nonembryogenic (NE) calli were not regenerable when 

transferred to the regeneration MS medium supplemented with 

1.0 mg L –1 benzylaminopurine (BAP) and 0.5 mg L –1 of naphthalene 

acetic acid (NAA). On the other hand, when embryogenic 

(E) calli were transferred to the regeneration medium, 

root initiation and callus differentiation were observed within 

1 wk of incubation. Visible shoot formation was noted 4 wk 

later on cultivar Trang Thai Lan (82 shoots per E callus from 

0.5 to 1.0 cm, the average percentage of E calli from shoots 

was 18%) and Trang Diep (3 shoots per E callus, the average 

percentage of E calli from shoots was 18%). The other cultivars 

will continue to be observed. 

For salt-treatment studies, attempts to produce embryogenic 

calli were conducted in saline medium as 0.5%, 1.0%, 

and 1.5% NaCl added to the callusing medium specific for 

each genotype, as per results obtained in the initial experiment 

(Table 1). Induced calli were excised 2 wk after inoculation of 

explants, transferred to fresh medium for proliferation for 21 

d, and then transferred to the regeneration medium. 

NaCl (1.5%) added to the regeneration medium increased 

the percentage of calli showing regeneration by 90.8% and 

75.0% in Soc Nau and Doc Do. In Trang Thai Lan, AS996, 

and Pokkali, 73.1%, 68.8%, and 51.3% reduction in the number 

of regenerating calli was recognized. Salt-stress-sensitive 

genotype IR29 did not clearly respond. 

Anther culture of F 1 

crosses 

Haploid plants are useful for practical and genetic studies. 

Anther and pollen culture are simple methods for obtaining 

haploid plants. Selection efficiency also increases because 

anther culture is based on gametic instead of sporophytic selection. 

The probability of obtaining a desired genotype in 

haploid (1/2 n) is much higher than in diploid (1/4 n) where n 

= number of genes controlling a particular character. The 

crosses from Teqing/Doc Phung, Tequing/Pokkali, and 

Tequing/Soc Nau have been made. Variation in culturability 

in terms of callus induction and plant regeneration among these 

crosses was observed (Table 2). The results indicate that 

Teqing/Doc Phung and Tequing/Pokkali were induced on calliproducing 

green plants (3.1, 18.2, and 0.9 reciprocity) (Table 

2) with the medium N 6 + 2 mg L –1 2,4-D + 1 mg L –1 NAA and 

plant regeneration of Teqing/Doc Phung, Tequing/Pokkali, and 

Session 15: Challenges to expanding rice production in unfavorable environments 443

Table 2. Percentage of embryogenic calli formation (%) from F 1 . 

Anthers Callus Calli- Green Albino Anther 

Cross on plates production Calli producing plant plant culture 

(no.) (%) plated green plants production production plants 

(%) (%) (%) (%) 

Tequing/Doc Phung 480 92.0 64 3.1 17.8 48.4 9.1 

Tequing/Pokkali 678 186.7 192 18.2 14.4 75.0 39.1 

Tequing/Soc Nau 491 47.5 113 0.9 0.9 53.6 0 

Tequing/Soc Nau obtained from the medium MS + 1 mg L –1 

kinetin + 0.5 mg L –1 NAA + 2 mg L –1 BA. In screening of 

anther culture-derived lines in a salt-screening nursery, a total 

of 23 lines were evaluated for salt tolerance. Among 23 lines 

tested, 3 promising lines from Tequing/Doc Phung were identified 

to be better than the resistant check. 

Estimation of salt tolerance in saline soil 

A total of 104 A4 lines, the original variety having been used 

as a control, were tested in normal and saline fields. Some 23 

lines exhibited salt tolerance. Variations were considerable 

among the lines. Nine lines were selected for further observation. 

Grain filling per panicle, 1,000-grain weight, spikelets 

m –2 , and yield were statistically different at the 5% level of 

probability from the original variety. 

Conclusions and suggestions 

MS + 2,4-D (2 mg L –1 ) is considered as a medium suitable for 

cultivating target cultivars. The genotypes Doc Do and Nep 

Ao Gia and others such as Pokkali, Soc Nau, Mong Chim Roi, 

and KDM 105 are also suitable for embryogenic callus formation. 

The regeneration MS medium supplemented with 0.5 

mg L –1 BAP and 1.0 mg L –1 NAA affected Trang Thai Lan (82 

shoots per E callus from 0.5 to 1.0 cm, obtained 18% E calli 

from shoots) and Trang Diep (3 shoots per E callus, obtained 

18% E calli from shoots). The other cultivars have been continuously 

observed. 

In NaCl 0.5% medium, E calli of all cultivars developed 

normally. Both treated and nontreated calli that survived in 

1.0% and 1.5% NaCl were of considerable benefit for further 

studies. 

Variation in culturability in terms of callus induction and 

plant regeneration among these crosses was observed. The results 

indicated that Teqing/Doc Phung and Tequing/Pokkali 

were induced on calli-producing green plants (3.1, 18.2, and 

0.9 reciprocity). Among 23 lines tested, three promising lines 

from Tequing/Doc Phung were identified to be better than the 

resistant check. 

The performance test of the 23 selected lines in the field 

was within 10 dS m –1 . The resulting three lines developed from 

this study have a potential for salt tolerance since their agronomic 

characters are much better than those of the original 

cultivar. 

References 

Chu CC, Wang CC, Sun CS, Yin KC, Chu CY, Bi FY. 1975. Establishment 

of an efficient medium for anther culture of rice 

through comparative experiments on nitrogen sources. Sci. 

Sin. 18:659-668. 

Murashige T, Skoog F. 1962. A revised medium for rapid growth 

and bioassay with tobacco tissue cultures. Physiol. Plant. 

15:473-497. 

Nguyen Thi Lang. 2002. Protocol for basics of biotechnology. Ho 

Chi Minh City (Vietnam): Agricultural Publishing House. 

Notes 

Authors’ addresses: Nguyen Thi Lang, Dang Minh Tam, and Bui 

Chi Buu, Cuu Long Delta Rice Research Institute, Omon, Can 

Tho, Vietnam; Hiromi Kobayashi, National Agrobiological 

Research Institute, Tsukuba, Ibaraki, Japan, e-mail: 

ntlang@hcm.vnn.vn. 


Submergence damage in rice and challenges in expanding 

the crop’s adaptability to submerged conditions in West 

and Central Africa 

Koichi Futakuchi 

In West and Central Africa (WCA), deepwater area accounts 

for 6% of the whole rice-growing area, mainly in the floodplains 

of large rivers such as the Niger River; and rainfed lowland 

and mangrove swamp account for 20% and 7%, respectively 

(Jones 1999). About one-third of all rice cultivation area 

faces danger from waterlogging. In addition, submergence 

could occur even in irrigated lowland at the seedling stage 

because of poor water management. 

In the floodplains, sudden flood can submerge the crop 

and waterlogging could continue during the growth of rice. To 

avoid submergence damage, farmers sometimes dibble rice 

seeds on ridges (Singh et al 1997). In some locations such as 

Mopti in Mali and the Sokoto Rima River floodplains in Nigeria, 

Oryza glaberrima, which is the other cultivated rice species 

besides O. sativa, and domesticated in WAC more than 

3,500 years ago (Jones et al 1997), is still cultivated in spite of 

its unfavorable traits in yield formation, that is, lodging and 

grain shattering, perhaps because of its high adaptability to 

waterlogging. Traditional O. sativa cultivars are preferably 

cultivated by farmers, too. One example is Gambiaka, which 

is considered to be adaptable to medium-deepwater conditions 

(50 to 100 cm deep). This cultivar has more vigorous seedling 

growth than improved cultivars; Gambiaka’s height and dry 

weight of 21-day-old seedlings were 42.3 cm and 123 mg, while 

those of WITA 4, WARDA’s improved cultivar released in 

Nigeria for the rainfed lowland ecology, were 28.9 cm and 75 

mg, respectively. Gambiaka therefore has a higher possibility 

to escape from submergence than WITA 4. 

In savannah and forest zones in WCA, inland valleys 

prevail and rice in valley bottoms experiences waterlogging 

for several days after excessive rainfall. Farmers transplant 

large older-aged seedlings in fields to keep them away from 

possible submergence. However, transplanting of such seedlings 

is associated with lower yield than appropriate-aged seedlings 

because larger seedlings receive more damage at transplanting 

and have a shorter remaining growth duration after 

transplanting, which cannot compensate for the damage. 

Waterlogging and rice plants 

Damage to rice plants under waterlogging 

Rice is an aquatic plant and is able to grow well under flooded 

conditions. In deepwater conditions, rice plants promote the 

elongation of the coleoptile, leaf, or stem, which can reduce 

the danger from submergence. However, rice receives severe, 

sometimes fatal, injury when it is under complete overhead 

submergence conditions for several days, especially for juvenile 

plants. 

Water has slower rates of gas exchange, less capacity to 

hold gases inside, and a higher extinction coefficient for light 

than the air. Floodwater is usually turbid so that severe shading 

occurs in the water. Under complete submergence, therefore, 

photosynthesis and respiration of rice plants are depressed. 

Low rates of gas diffusion cause the accumulation of physiologically 

active gases such as ethylene produced in situ in the 

plant (Jackson and Ram 2003). Visible symptoms of rice plants 

under submergence are faster elongation of a few leaves, yellowing 

of old leaves, slow or negative dry-matter growth, and 

shoot decay. After a reduction in water levels, most leaves or 

the whole shoot may collapse. Some can die and the survivors 

also suffer from poor rates of new leaf emergence and lodging 

(Jackson and Ram 2003). When rice plants under submergence 

succeed in re-contacting with the air because of elongation and 

survive, they are highly likely to lodge after the water level 

recedes. 

Varietal development for tolerance 

of complete submergence 

Physiological mechanisms of submergence tolerance are well 

elucidated in rice compared with other abiotic constraints such 

as drought. Jackson and Ram (2003) have widely reviewed 

mechanisms of injury by and tolerance of complete submergence 

in rice referring to both current and pioneer studies. The 

best source of submergence tolerance in rice was already identified, 

which is Indian cultivar FR13A selected from a local 

landrace. Mackill et al (1993) introduced the tolerance into an 

agronomically useful cultivar by conventional breeding. Although 

the tolerance was considered to be a quantitative trait 

controlled by several genes, studies with molecular markers 

revealed that a single locus, Sub 1, controlled most of the tolerance 

(Xu and Mackill 1996). This locus was mapped on rice 

chromosome 9, and a fine-scale map around Sub 1 was also 

developed (Xu et al 2000). The markers mapped close to Sub 

1 can reduce the need for time-consuming and high-cost field 

screening. 

Plant types adaptable to waterlogging in WCA 

In the floodplains of large rivers, waterlogging could continue 

for a long time. Since existing cultivars with tolerance of complete 

submergence can survive or receive less damage than 

normal cultivars under submergence of several days, their type 

is not suitable. An elongation type that can escape from sub- 


Relative elongation rate to the initial height at day 0 (%) 

80 

TOG 5810 TOG 6283 

60 

40 

20 

0 

80 

CG 14 CG 20 

60 

40 

20 

0 

80 

60 

Norin 30 

IR36 

40 

20 

0 0 5 10 15 20 25 0 5 10 15 20 25 

Days of submergence 

Fig. 1. Changes in the rate of increased height to the initial height with the time of the 

submergence treatment (cited from Futakuchi et al 2001). Vertical bars indicate standard 

deviation of three replications. Closed circles = submergence; open circles = 

nonsubmergence (control). 

mergence is desirable. To tolerate complete submergence 

caused by sudden flooding until plants re-contact with the atmosphere 

by elongation, it will also be desirable for them to 

have submergence tolerance to a certain extent. Vigorous initial 

growth will be important, too, to reduce the danger from 

submergence. Resistance to lodging or recovering ability from 

lodging after a water level reduction will be required for practical 

cultivars acceptable to farmers. 

In inland valley bottoms, submergence-tolerant types will 

be useful. In existing cultivars such as FR13A, however, the 

tolerance does not mean unharmed by submergence (Jackson 

and Ram 2003). Although such cultivars relatively tolerate submergence, 

they also have some damage and collapse more or 

less after submergence for several days. Rice has well-developed 

aerenchyma and it receives much less damage when keep- 

ing in contact with the air by leaves or other organs. Since 

sudden flooding could occur in the bottoms, the same type as 

with the floodplains mentioned above will be highly suitable. 

Setter and Laureles (1996) have reported that there is a 

trade-off between stimulated elongation and tolerance in completely 

submerged conditions, suggesting difficulty in developing 

cultivars with both elongation ability and tolerance. 

Futakuchi et al (2001b) have tested Oryza glaberrima in relation 

to elongation ability and tolerance under submergence. 

Four O. glaberrima lines, out of which two were identified as 

submergence-adaptable types (TOG 5810 and TOG 6283) and 

the others were identified as upland-adaptable types (CG 14 

and CG 20), were cultivated in pots and submerged at the end 

of the stem elongation stage with two O. sativa checks (Norin 

30 and IR36). Figure 1 depicts changes in relative elongation 


Photosynthetic rate (nmol cm –2 h –1 ) 

1,800 

1,600 

1,400 

1,200 

1,000 

800 

600 

400 

200 TOG 5810 

0 

TOG 6283 

1,800 

1,600 

1,400 

1,200 

1,000 

800 

600 

400 

200 

0 

CG 14 

CG 20 

1,800 

1,600 

1,400 

1,200 

1,000 

800 

600 

400 

200 Norin 30 

IR36 

0 

0 5 10 15 20 25 0 5 10 15 20 25 

Days of submergence 

Fig. 2. Changes in the rate of photosynthetic oxygen evolution in a leaf slice with the time of 

the submergence treatment (cited from Futakuchi et al 2001). Vertical bars indicate standard 

deviation of three replications. 

rate—the percentage ratio of increased height to the initial 

height—after the beginning of the submergence treatment. 

Submergence promoted elongation for all lines/varieties tested 

but the extent of elongation was much larger in O. glaberrima 

than in O. sativa. Judging from the elongation ability, O. 

glaberrima should have much less tolerance of submergence 

than the O. sativa checks. However, O. glaberrima maintained 

high photosynthetic rates, which were determined by oxygen 

electrode, compared with Norin 30, and rates similar to those 

of IR36 (Fig. 2). The same result was obtained with chlorophyll 

content. O. glaberrima had better tolerance than expected 

from elongation ability under submergence. 

Out of the O. glaberrima lines tested, CG 14 showed the 

fastest elongation soon after submergence (Fig. 1) though the 

line had been identified as an upland type. Thirty interspecific 

progenies from the cross of CG 14 and an O. sativa cultivar 

(WAB56-104) were tested in terms of elongation ability in 

medium-deepwater: 50 cm deep (Futakuchi et al 2001a). This 

treatment started at maximum tillering and continued for 4 

weeks. CG 14 showed the best elongation at 2 wk after the 

treatment and a few interspecific progenies rivaled CG 14 

though elongation was promoted for all interspecifics during 

the treatment. One week after the end of the treatment, however, 

the actual plant height of the interspecifics in the control 

(always in shallow-water conditions) caught up with that of 

the interspecifics grown in the treatment. Yield and tiller numbers 

were much depressed by the medium-deepwater treatment. 

However, lodging was observed only with CG 14. 

The growth of O. glaberrima at the seedling stage is 

vigorous. For example, plant height and dry weight of CG 14 

at 21 days after seeding were 46.4 cm and 54.3 mg, while those 

of Bouake 189, a leading O. sativa variety for lowlands in 

Côte d’Ivoire, were 30.4 cm and 38.0 mg, respectively. 


Oryza glaberrima could be a source to develop plant 

types adaptable to waterlogging in WCA, providing ability to 

escape from and moderately tolerate submergence. Since the 

two O. glaberrima lines, TOG 5810 and TOG 6283, that had 

been identified as submergence-adaptable types prior to the 

trial did not have better tolerance than IR36 (Fig. 2), it might 

be difficult to identify lines with strong tolerance of submergence 

in O. glaberrima. The problem of O. glaberrima is susceptibility 

to lodging. This could be overcome by crossing with 

O. sativa varieties having strong columns. 

References 

Futakuchi K, Audebert A, Jones MP. 2001a. Stem elongation of interspecific 

O. sativa × O. glaberrima progenies in deep water 

conditions. Jpn. J. Crop Sci. 70(extra issue 2):279-280. 

Futakuchi K, Jones MP, Ishii R. 2001b. Physiological and morphological 

mechanism of submergence resistance in African rice 

(Oryza glaberrima Steud.). Jpn. J. Trop. Agric. 45:8-14. 

Jackson M, Ram PC. 2003. Physiological and molecular basis of 

susceptibility and tolerance of rice plants to complete submergence. 

Ann. Bot. 91:227-241. 

Jones MP, Dingkuhn M, Aluko GK, Mande S. 1997. Interspecific 

Oryza sativa L. × O. glaberrima Steud. progenies in upland 

rice improvement. Euphytica 92:237-246. 

Jones MP. 1999. Food security and major technological challenges: 

the case of rice in sub-Saharan Africa. In: Horie T, Geng S, 

Amano T, Inamura T, Shiraiwa T, editors. World food security 

and crop production technologies for tomorrow. Proceedings 

of the International Symposium, October 1998. Kyoto (Japan): 

Kyoto University. p 57-64. 

Mackill DJ, Amante MM, Vergara BS, Sarkarung S. 1993. Improved 

semi-dwarf rice lines with tolerance to submergence of seedlings. 

Crop Sci. 33:749-753. 

Setter T, Laureles E. 1996. The beneficial effect of reduced elongation 

growth on submergence tolerance in rice. J. Exp. Bot. 

47:1551-1559. 

Singh BN, Fagade S, Ukwungwu MN, Williams C, Jagtap SS, 

Oladimeji O, Efisue A, Okhidievbie O. 1997. Rice growing 

environments and biophysical constraints in different 

agroecological zones of Nigeria. Met. J. 2:35-44. 

Xu K, Mackill DJ. 1996. A major locus for submergence tolerance 

mapped on rice chromosome 9. Mol. Breed. 2:219-224. 

Xu K, Xu X, Ronald PC, Mackill DJ. 2000. A high-resolution linkage 

map of the vicinity of the rice submergence tolerance locus 

Sub 1. Mol. Gen. Genet. 263:681-689. 

Notes 

Author’s address: The Africa Rice Center (WARDA), 01 BP 2031 

Cotonou, Benin, e-mail: k.futakuchi@cgiar.org. 

Ecological, morphological, and physiological 

aspects of drought adaptation of rice in upland 

and rainfed lowland systems 

Shu Fukai and Akihiko Kamoshita 

Many rice-growing areas in the world do not enjoy the availability 

of full irrigation water, and a rice crop often experiences 

drought at some time during its growth. Responses of 

different genotypes to water stress have been studied for a long 

time, and several physiological and morphological characters 

have been suggested to be responsible for drought tolerance in 

rice. Grain yield of some genotypes is affected less than others 

by drought, but genotypic adaptation to drought is not consistent 

across different drought conditions. The single factor determining 

genotypic variation in yield under drought conditions 

is crop phenology, particularly flowering time. However, 

within the same phenology group, no particular physiological 

or morphological characters have been found to be consistently 

responsible for genotypic variation in yield. This lack of universally 

useful traits has resulted in rather slow progress in 

developing drought-tolerant cultivars in rice (Fukai et al 1999). 

Some of the inconsistency in genotypic yield performance 

is caused by different types of drought that the rice crop 

may encounter. Ecological factors such as the presence of flood 

water before drought develops, as is often the case in rainfed 

lowland rice, are now recognized to affect genotypic adapta- 

tion to drought. For example, deep-root characters that are 

useful to minimize the adverse effects of drought in some upland 

conditions may not be expressed fully under rainfed lowland 

conditions where anaerobic soil conditions prevent the 

development of deep root systems. Timing of drought is another 

ecological factor; if drought develops during the vegetative 

stage, the genotype’s ability to recover from drought is 

important, whereas, if drought develops around flowering time, 

the plant’s ability to tolerate drought becomes a key factor in 

producing higher yield than with other genotypes. This paper 

describes physiological and morphological characters that are 

useful under different types of drought (Table 1), and how that 

understanding can assist in the selection of drought-tolerant 

rice cultivars. 

Terminal drought 

Among genotypes of a similar maturity type, genotypes that 

can maintain high leaf water potential are often advantageous 

in producing higher yield under terminal drought conditions 

(Jongdee et al 2002, Pantuwan et al 2002). Higher water po- 


Table 1. Physiological and morphological characters of 

drought tolerance under each drought type. 

Type of drought 

Terminal drought 

Intermittent drought 

Vegetative-stage drought 

Primary/secondary characters 

Early flowering 

Short delay in flowering 

High leaf water potential 

Low leaf death score 

Deep-root system 

Late flowering 

Longer delay in flowering 

Large green leaf area 

Tillering after drought period 

tential and associated turgor maintenance and carbohydrate 

supply to spikelets would assist with fertilization and hence 

reduce spikelet sterility. 

Positive turgor is required for stem extension and exsertion 

of panicles. Delay in flowering will be small with maintenance 

of turgor, and this is related to higher yield under terminal 

drought (Pantuwan et al 2002). 

Under terminal drought, the maintenance of favorable 

water conditions for reproductive organ growth and securing 

adequate grain sink size are a primary drought-tolerance mechanism, 

and a conservation strategy of water use may be considered 

to be effective. Since water runs out by maturity in terminal 

drought, prolonging water availability to the plants is required 

for higher yield. To achieve this, water may be extracted 

rather slowly by a smaller root system or smaller shoot system 

with reduced water demand (Pantuwan et al 2002). 

Intermittent drought 

In contrast to terminal drought, intermittent drought is broken 

by a rainfall event, and hence there is no strong need for conservation 

of water as such. A deep root system with higher 

root density is likely to be useful under these conditions. Association 

between high root length density and the amount of 

water extracted has been well demonstrated (Lilley and Fukai 

1994), and, under upland conditions particularly, these root 

characters contribute to high yield. With intermittent drought, 

rainfall can replenish soil water, and hence cycles of water 

extraction-replenishment would take place. Under these conditions, 

well-developed root systems can take advantage of 

extracting water to a greater extent. This may not apply to terminal 

water conditions because the rate of water extraction 

would be faster and the duration of water extraction shorter 

(Lilley and Fukai 1994), resulting in early exhaustion of a limited 

amount of stored water. 

On the other hand, expression of deep-root characters is 

not always good under rainfed lowland conditions. Conditions 

of anaerobic soil that are likely to develop at some time in the 

growth cycle limit deep-root development (Pantuwan et al 

1997). Partitioning to deep roots is smaller under flooded conditions 

(ca. 0.3–1.2%) than under droughted conditions (ca. 

3–17%) in pots (Azhiri-Sigari et al 2000). Genotypic variation 

in deep-root development has been reported under anaerobic 

conditions, with their ranking generally similar between 

anaerobic and aerobic conditions, although in some cases interactions 

were recognized. 

Vegetative-stage drought 

Drought often develops during the early wet season, particularly 

in bimodal rainfall areas. This causes a major problem of 

failure in nursery establishment and transplanting. Seedling 

vigor is often related to quicker development of the deep-root 

system before drought development, which accelerates water 

extraction and maintains growth during intermittent drought. 

Genotype by initial soil moisture interactions were found for 

deep-root development. 

Genotypes may differ in their recovery growth after vegetative-stage 

drought. This may be related to the amount of 

leaf that remains after drought (Mitchell et al 1998) or ability 

to tiller after drought (Lilley and Fukai 1994). 

In our recent experiments in Cambodia, a short delay in 

flowering was associated with lower yield under early-season 

drought conditions, in contrast to the case of terminal drought, 

in which a short delay was advantageous. In the former case, 

early-flowering varieties flowered before full recovery and 

hence yield decreased, whereas late-flowering varieties had 

more time to recover before flowering took place. 

QTLs for drought-tolerance traits 

QTL links among primary, secondary, and integrated traits were 

examined from the reviews of previous studies. Five mapping 

populations have been extensively examined for QTL identification 

(CO39/Moroberekan, IR64/Azucena, Bala/Azucena, 

CT9993/IR62266, and IR58821/IR52561). Phenotyping environments 

were mostly under upland conditions. The emphasis 

was given to CT9993/IR62266, which has been phenotyped 

under lowland conditions (Kamoshita et al 2002, Lanceras et 

al 2004). 

The mapped traits were (1) primary traits (root traits, 

osmotic adjustment, cell membrane stability, carbon isotope 

discrimination), (2) secondary traits (e.g., leaf rolling, canopy 

temperature, leaf water potential), and (3) integrated traits (e.g., 

grain and biological yields, percent spikelet sterility, delay in 

flowering). Phenology and plant types (e.g., plant height) were 

also mapped. QTL results are more available for root traits 

observed in pots and for easily scored drought-avoidance traits 

(e.g., leaf rolling). 

Thirty-four chromosome regions with multiple QTLs of 

different traits (QTL clusters) were identified. Twenty QTL 

clusters contained both primary and integrated traits (Table 

2). In RG939-RG214 (chromosome 4), QTLs for deep and 

thick root parameters, root penetration ability, and root-pulling 

force were identified, as well as a QTL for grain yield, 

with their positive effects coming from the deeper-root parent, 


Table 2. Analysis of chromosome regions (QTL clusters) that have QTLs for any combination of two or 

more trait groups (primary, secondary, and integrated traits). Number of QTL clusters with the two 

identical trait groups is shown, together with one or two typical examples of QTL clusters for those 

trait group combinations. 

Trait group Trait Marker interval r 2 Allele for positive effect 

Integrated trait—primary trait (20) a 

Cluster 13 

RG939-RG214 (14.3 cM) b (chr. 4) c 

Integrated Grain yield (s) RG939-RG476 15.8 CT9993 

Deep root/tiller RG476-RG214 4.7 CT9993 

Root thickness RG476-RG214 29.9 CT9993 

Root thickness RG939-RG476 11.0 CT9993 

Primary Penetrated root thickness RG939-RG476 31.3 CT9993 

Penetrated root weight RG939-RG476 11.5 CT9993 

Root penetration index RG939-RG476 11.0 CT9993 

Root-pulling force RG214-RG620 19.9 CT9993 

Cluster 9 

RZ313-C63 (16.6 cM) b (chr. 3) a 

Integrated Biological yield (s) d RZ313-EM17_1 15.5 IR62266 

Primary Osmotic adjustment (s) EM17_1-C63 9.9 IR62266 

Primary trait—secondary trait–integrated trait–plant type–phenology (2) a 

Clusters 1, 2 e 

CDO345-RG345 (54.8 cM) b (chr. 1) c 

Root weight RG109-EM11_11 9.3 CT9993 

Root thickness CDO345-ME10_14 8.8 CT9993 

Primary Penetrated root thickness RG957-RG345 9.2 CT9993 

Deep root weight RM212-R2417 9.4 IR62266 

Deep root ratio C813-RG957 8.1 IR62266 

Deep root/tiller RM212-R2417 9.8 IR62266 

Rooting depth RM212-R2417 7.1 IR62266 

Leaf rolling (s) ME10_14-RG109 15.2 CT9993 

Leaf rolling (s) RM212-RG957 – – 

Leaf drying (s) ME10_14-RG109 20.8 CT9993 

Secondary Drought score (s) RM315-RG109 – – 

Drought score (s) RM212-RG957 – – 

Canopy temperature (s) RM315-RG109 – – 

Canopy temperature (s) RG109-EM11_11 17 – 

Relative water content (s) C813-RM212 12.1 CT9993 

Leaf water potential (s) C813-R2417 – – 

Integrated Grain yield RM212-R2417 13 – 

Grain yield RG957-RG345 9 – 

Grain yield (s) (ME2_16-RG811 f ) 7.8 IR62266 

Harvest index (s) RZ909-CDO345 18.6 CT9993 

Spikelet number RG109-EM11_11 15.2 CT9993 

Plant type Plant height (s) ME10_14-RG109 27.8, 46.8 CT9993 

Plant height (s) CDO345-RM102 36.2, 32.6, 23.1 CT9993 

Plant height ME10_14-EM11_11 45.8, 46.5 CT9993 

Plant height RZ909-RM102 46.1, 41.1 CT9993 

Panicle number (s) CDO345-RG109 18.6, 20.5, 20.8 IR62266 

Panicle number CDO345-RG109 21.7, 20.7, 8 IR62266 

Percent light interception C813-R2417 – – 

Phenology Heading/flowering day CDO345-ME10_14 17, 13 – 

Heading day (s) EM11_11-ME4_18 15 – 

a Numbers in parentheses are total number of QTL clusters for the corresponding two trait groups. b Length of marker intervals 

in centi-Morgans. c Chromosome numbers of the corresponding locations of QTL clusters. d The letter “s” in parentheses after 

trait names indicates that the corresponding traits were recorded under severe water stress environments. e Clusters 1 and 2 

are close, at 4.0 cM distance away. f With 24.2 cM apart from RG345 in cluster 2. 


CT9993 (Babu et al 2003). In this marker interval, QTLs for 

deep and thick root traits were found in CO39/Moroberekan 

and IR58821/IR52581. 

While many QTLs for integrated traits are linked with 

those for root traits, QTLs for osmotic adjustment and for biomass 

yield under mild stress in upland are located in RZ313- 

C63 on chromosome 3, with their positive effects from IR62266 

(Table 2). A QTL for osmotic adjustment is located on chromosome 

8 (G187-R1394A), which is homologous with the 

segment on chromosome 7 in wheat with the osmoregulation 

gene, together with other QTLs for drought-avoidance traits. 

In C813-R2417 in clusters 1 and 2 (chromosome 1), QTLs for 

relative water content, leaf water potential, leaf rolling, and 

drought scores were found together with a QTL for percent of 

radiation interception, indicating the effects of canopy size on 

plant water status (Table 2). QTLs for plant height with large 

genetic effects and the sd-1 gene are located in these clusters, 

together with a number of QTLs for primary, secondary, and 

integrated traits. 

In C813-RG811 (chromosome 2), QTLs for primary 

traits (penetrated root thickness through hardpan, deep-root 

traits) and secondary traits (relative water content, canopy temperature 

under drought, leaf drying) were found in more than 

one mapping population, as well as three QTLs for grain yield 

in three environments with different water regimes. The positive 

effects for the allele contribution came from CT9993 for 

both the primary and secondary traits, but grain yield was always 

higher for the shallower-rooting parent, IR62266, with 

the positive effect of the QTL for stress yield from IR62266, 

suggesting the highly integrated manner of yield determination. 

The putative QTLs for these primary and secondary 

drought-tolerance traits were incorporated to elite lines using 

the advanced backcross strategy, and the effects of markerassisted 

selection were evaluated. The putative candidate genes 

for drought tolerance are being clarified by saturation mapping 

(Nguyen et al 2004). 

Conclusions 

Drought-tolerance characters appear to differ under different 

types of drought. Understanding of prevailing drought types 

and associated characters is required for efficient use of these 

characters to increase drought tolerance in rainfed rice. These 

characters will then need to be incorporated into high-yielding 

genotypes with flowering time appropriate to the region of 

concern. 

Several QTL clusters were mapped that contained primary/secondary 

drought-tolerance traits and/or integrated traits, 

together with QTLs for phenology or plant type with large genetic 

effects. Some of them can be used for the advanced backcross 

strategy for developing more drought-tolerant cultivars. 

References 

Azhiri-Sigari T, Yamauchi A, Kamoshita A, Wade LJ. 2000. Genotypic 

variation in response of rainfed lowland rice to drought 

and rewatering. II. Root growth. Plant Prod. Sci. 3:180-188. 

Babu RC, Nguyen BD, Chamarerk V, Shanmugasundaram P, 

Chezhian P, Jeyaprakash P, Ganesh SK, Palchamy A, 

Sadasivam S, Sarkarung S, Wade LJ, Nguyen HT. 2003. Genetic 

analysis of drought resistance in rice by molecular markers: 

association between secondary traits and field performance. 

Crop Sci. 43:1457-1469. 

Fukai S, Pantuwan G, Jongdee B, Cooper M. 1999. Screening for 

drought resistance in rainfed lowland rice. Field Crops Res. 

64:61-74 

Jongdee B, Fukai S, Cooper M. 2002. Leaf water potential and osmotic 

adjustment as physiological traits to improve drought 

tolerance in rice. Field Crops Res. 76:153-163. 

Kamoshita A, Zhang J, Siopongco J, Sarkarung S, Nguyen HT, Wade 

LJ. 2002. Effects of phenotyping environment on identification 

of QTL for rice root morphology under anaerobic conditions. 

Crop Sci. 42:255-265. 

Lanceras JC, Pantuwan G, Jongdee B, Toojinda T. 2004. Quantitative 

trait loci associated with drought tolerance at reproductive 

stage in rice. Plant Physiol. 135:384-399. 

Lilley JM, Fukai S. 1994. Effect of timing and severity of water 

deficit on four diverse rice cultivars. I. Rooting pattern and 

soil water extraction. Field Crops Res. 37:205-213. 

Mitchell JH, Siamhan D, Wamala MH, Risimeri JB, Chinyamakobvu 

E, Henderson SA, Fukai S. 1998. The use of seedling leaf 

death score for evaluation of drought resistance of rice. Field 

Crops Res. 55(1-2):129-139. 

Nguyen TTT, Klueva N, Chamareck V, Aarti A, Magpantay G, Millena 

ACM, Pathan MS, Nguyen HT. 2004. Saturation mapping of 

QTL regions and identification of putative candidate genes 

for drought tolerance in rice. Mol. Genet. Genomics 272:35- 

46. 

Pantuwan G, Fukai S, Cooper M, O’Toole JC, Sarkarung S. 1997. 

Root traits to increase drought resistance in rainfed lowland 

rice. Proceedings of an International Workshop, Ubon 

Ratchathani, Thailand, 1996, Australian Centre for International 

Agricultural Research, Canberra, ACT. p 170-179. 

Pantuwan G, Fukai S, Cooper M, Rajatasereekul S, O’Toole JC. 2002. 

Yield response of rice (Oryza sativa L.) genotypes to drought 

under rainfed lowland. 3. Plant factors contributing to drought 

resistance. Field Crops Res. 73:181-200. 

Notes 

Authors’ addresses: S. Fukai, The University of Queensland, 

Brisbane, Australia, e-mail: s.fukai@mailbox.uq.edu.au; A. 

Kamoshita, The University of Tokyo, Nishi-Tokyo, Japan, e- 

mail: akamoshita@fm.a.u-tokyo.ac.jp. 


Managing iron toxicity in lowland rice: the role 

of tolerant genotypes and plant nutrients 

Kanwar L. Sahrawat 

Iron toxicity is a widespread nutrient disorder affecting the 

growing of wetland rice in the humid tropical regions of Asia, 

Africa, and South America. Large areas of wetlands ideally 

suited for rice production remain underused, especially in West 

and Central Africa, because of iron toxicity as a constraint. 

Iron toxicity has been reported to reduce rice yields by 12– 

100% depending on the intensity of the stress and tolerance of 

the rice cultivars (Sahrawat et al 1996, Sahrawat 2004). Iron 

toxicity of wetland rice is associated with a high concentration 

of ferrous iron in soil solution (Ponnamperuma et al 1955). 

The stress occurs in reduced soils when a toxic amount of ferrous 

iron is mobilized in soil solution in situ or when inflow 

brings in soluble iron from upper slopes (van Breemen and 

Moormann 1978). 

Iron toxicity occurs in soils (mostly Ultisols, Oxisols, 

and acid sulfate soils) high in active iron and potential acidity, 

irrespective of organic matter and texture. But texture, cation 

exchange capacity, and organic matter content influence the 

concentration of ferrous iron in soil solution, in which iron 

toxicity occurs (van Breemen and Moormann 1978). Plantand 

growing-medium-related factors such as plant age, accumulation 

of hydrogen sulfide, organic acids, and other reduction 

products also influence iron toxicity occurrence in rice 

(Sahrawat 2004). 

Iron toxicity symptoms vary with rice cultivars. They 

are characterized by a reddish brown, yellow, or purple-bronzing 

or orange discoloration of the lower leaves of the rice plants. 

Typically, iron toxicity symptoms are manifested as tiny brown 

spots starting from the upper tips and spreading toward the 

bases of the lower leaves. With progress in iron toxicity, the 

brown spots coalesce on the interveins of the leaves. With increased 

iron toxicity stress, the entire affected leaves look purplish 

brown, followed by drying of the leaves, which gives the 

rice plant a scorched appearance. Equally important, the roots 

of rice plants affected by iron toxicity become scanty, coarse, 

short and blunted, and dark brown in color; with the alleviation 

of the stress, the roots may slowly recover to the usual 

white color. Iron toxicity symptoms on rice leaves and changes 

in root color and morphology are useful for diagnosis of the 

stress. Toxicity symptoms commonly develop at the maximum 

tillering and heading growth stage, but may be observed at any 

growth stage of the rice crop. 

Since the first report of its occurrence (Ponnamperuma 

et al 1955), iron toxicity in rice has been reported in several 

countries in Asia, South America, and West and Central Africa 

(van Breemen and Moormann 1978, Yoshida 1981, De Datta 

et al 1994, Sahrawat 2004). 

Iron toxicity is a complex nutrient disorder and the deficiencies 

of other nutrients, especially phosphorus (P), potassium 

(K), calcium (Ca), magnesium (Mg), and zinc (Zn), are 

considered in the occurrence of iron toxicity in rice (Ottow et 

al 1983). Other nutrients may play an important role not only 

in reducing the effect of iron toxicity but also in the expression 

of iron tolerance by various rice cultivars (Sahrawat et al 

1996, Sahrawat 2004). Deficiencies of P, K, Ca, Mg, and manganese 

(Mn) decrease the iron-excluding power of rice roots 

and can thus affect the rice plant’s tolerance of iron toxicity 

(e.g., see Yoshida 1981, Sahrawat 2004). Deficiencies of Ca, 

Mg, and Mn are not commonly observed in lowland rice, except 

probably on acid sulfate soils; deficiencies of P, K, and 

Zn therefore deserve special attention (Yoshida 1981). 

This paper critically reviews recent research on the role 

of tolerant genotypes and plant nutrients in reducing iron toxicity. 

The ultimate goal is to provide information that can be 

used for increasing rice production and productivity on irontoxic 

wetlands on a sustainable basis. 

Tolerant genotypes for reducing iron toxicity 

Rice cultivars differ in their tolerance for iron toxicity and the 

selection of rice cultivars with superior iron tolerance is an 

important component of research for reducing iron toxicity. 

Genetic differences in adaptation to and tolerance for irontoxic 

soil conditions have indeed been exploited for developing 

rice cultivars with tolerance for iron toxicity (Gunawardena 

et al 1982, DeDatta et al 1994). Breeding and screening efforts 

at the International Rice Research Institute in the Philippines 

and at WARDA (West Africa Rice Development Association) 

in Côte d’Ivoire have identified a number of rice cultivars 

for growing in iron-toxic soils (De Datta et al 1994, 

Sahrawat 2004). 

Sahrawat et al (1996) evaluated 20 lowland rice cultivars 

for tolerance of iron toxicity at an iron-toxic site in 

Korhogo, Côte d’Ivoire, under irrigated conditions. The cultivars 

differed in tolerance of iron toxicity. Grain yields varied 

from 0.10 to 5.04 t ha –1 and iron toxicity scores, based on the 

extent of bronzing symptoms on foliage, ranged from 2 to 9 (1 

indicates normal growth and 9 indicates that most plants are 

dead or dying). Further evaluation of rice cultivars during 1992- 

97 showed that, among three promising iron-tolerant cultivars, 

CK 4 was the top yielder (mean grain yield 5.33 t ha –1 ), followed 

by WITA 1 (4.96 t ha –1 ) and WITA 3 (4.46 t ha –1 ), and 

tolerant check Suakoko 8 (3.80 t ha –1 ) (Table 1). These and 

other results suggest that high rice yields and iron toxicity tol- 


Table 1. Grain yields (t ha –1 ) of WITA 1 and WITA 3 rice cultivars during 

1992-97 relative to the performance of iron-tolerant (Suakoko 8 and CK 

4) and iron-susceptible (Bouake 189) check cultivars under irrigated conditions 

in the wet season at an iron-toxic site in Korhogo, Côte d’Ivoire. a 

Year CK4 WITA1 WITA 3 Bouake 189 Suakoko 8 LSD (0.05) 

1992 – 4.33 5.04 2.87 4.85 1.080 

1993 5.87 5.53 5.17 4.08 5.07 0.630 

1994 6.05 6.66 4.30 4.69 3.73 1.100 

1996 3.76 3.24 3.21 2.81 2.57 0.760 

1997 5.63 5.02 4.59 4.99 2.79 1.345 

Mean 5.33 4.96 4.46 3.88 3.80 

a 

Each season, all cultivars received a uniform application of 100 kg N ha –1 , 50 kg P ha –1 , 

and 10 kg Zn ha –1 . 

Source: Sahrawat et al (2000). 

erance are physiologically compatible (Sahrawat et al 2000, 

Audebert and Sahrawat 2000). 

Work done at WARDA in West Africa showed that some 

Oryza glaberrima cultivars, adapted to lowland rice-growing 

conditions, possess a higher tolerance for iron toxicity than 

their O. sativa counterparts. Sahrawat and Sika (2002) conducted 

experiments at an iron-toxic site (Korhogo, Côte 

d’Ivoire) during the 2000 wet and dry seasons to evaluate the 

performance of promising O. sativa (CK 4, tolerant check; 

Bouake 189, susceptible check) and O. glaberrima (CG 14) 

cultivars. While CK 4 and Bouake 189 showed typical iron 

toxicity symptoms in varying degrees, CG 14 plants did not 

show any iron toxicity symptoms at all as measured by iron 

toxicity scores. Although CG 14 did not give high grain yields 

because of its lower harvest index, lodging of the crop, especially 

under the application of nutrients, and shattering of seeds 

at maturity, it showed remarkable tolerance for iron toxicity. 

Research shows that CG 14 has a high tolerance for iron toxicity 

and remains an obvious choice as a donor for iron tolerance 

in breeding programs (Sahrawat and Sika 2002, Sahrawat 

2004). 

Role of other nutrients in reducing iron toxicity 

Table 2. Effects of field applications of nutrients on grain yield of 

iron-tolerant (CK 4) and susceptible (Bouake 189 and TOX 3069- 

66-2-1-6) rice cultivars on an iron-toxic soil at Korhogo, Côte d’Ivoire 

(1995-98). a Grain yield (t ha –1 ) 

Treatment 

CK 4 Bouake 189 TOX 3069-66-2-1-6 

No fertilizer 4.3 (3) b 3.4 (5) 2.9 (7) 

N 4.4 (3) 4.1 (5) 3.3 (7) 

N + P 5.3 (2) 4.3 (4) 4.2 (5) 

N + K 4.8 (2) 4.4 (4) 3.8 (5) 

N + Zn 4.8 (2) 4.6 (4) 4.6 (5) 

N + P + Zn 5.0 (2) 4.4 (4) 4.2 (4) 

N + K + Zn 5.2 (2) 4.6 (3) 4.6 (4) 

N + P + K 5.4 (2) 4.5 (3) 4.5 (3) 

N + P + K + Zn 5.7 (2) 4.7 (3) 4.7 (3) 

LSD (0.05) 1.01 1.02 1.15 

a 

The data are an average of four years (1995-98). All cultivars received a uniform 

application of N (100 kg ha –1 ), P (50 kg ha –1 ), K (80 kg ha –1 ), and Zn (10 kg ha –1 ). 

b 

Iron toxicity scores are given in parentheses on a scale of 1 to 9, where 1 = normal 

growth and 9 = most plants are dead or drying. 

Source: Sahrawat et al (2001). 

A high concentration of iron in soil solution can cause nutrient 

imbalance through antagonistic effects on the uptake of nutrients, 

including K and Zn. The deficiency or lack of availability 

of other nutrients can also affect the rice plant’s ability to 

decrease uptake of iron in the tops through physiological functions 

carried out by roots such as iron oxidation, iron exclusion, 

and iron retention (Yoshida 1981, Audebet and Sahrawat 

2000, Sahrawat 2004). Thus, it is not entirely surprising that 

the application of other nutrients reduces iron toxicity and 

improves yield of rice on iron-toxic soils. Several reports show 

that applications of nutrients such as P, K, and Zn reduce iron 

toxicity, improve growth, and increase rice yield (Sahrawat et 

al 2001, Nayak et al 2004, Sahrawat 2004). Sahrawat et al 

(2001) showed that applications of N, P, K, and Zn in various 

combinations reduced iron toxicity and increased yields of irontolerant 

and -susceptible rice cultivars. The increase in grain 

yields of iron-susceptible cultivars was more than that of irontolerant 

cultivars (Table 2). 

Conclusions 

Iron toxicity can be reduced by using iron-tolerant cultivars 

and by applying other nutrients whose availability is negatively 

affected by a high concentration of iron in soil solution. The 

intensified use of iron-toxic wetlands in the future is inevitable 

for meeting the food needs of the ever-growing population 

in tropical regions, where iron-toxic soils are an important 

natural resource for food production. An integrated use of 

tolerant genotypes and improved soil and nutrient management 

is more practical for sustainable increases in rice productivity. 


References 

Audebert A, Sahrawat KL. 2000. Mechanisms for iron toxicity tolerance 

in lowland rice. J. Plant Nutr. 23:1877-1885. 

De Datta SK, Neue HU, Senadhira D, Quijano C. 1994. Success in 

rice improvement for poor soils. In: Proceedings of the Workshop 

on Adaptation of Plants to Soil Stresses, 1-4 August 1993, 

University of Nebraska, Lincoln, Nebraska. INTSORMIL 

Publication No. 94-2. Lincoln, Nebraska (USA): University 

of Nebraska. p 248-268. 

Gunawardena I, Virmani SS, Sumo FJ. 1982. Breeding rice for tolerance 

to iron toxicity. Oryza 19:5-12. 

Nayak SC, Sahu SK, Mishra GC, Sandha B. 2004. Comparison of 

different amendments for alleviating iron toxicity in rice. Int. 

Rice Res. Notes 29:51-53. 

Ottow JCG, Benckiser G, Watanabe I, Santiago S. 1983. Multiple 

nutritional soil stress as the prerequisite for iron toxicity of 

wetland rice (Oryza sativa L.). Trop. Agric. (Trinidad) 60:102- 

106. 

Ponnamperuma FN, Bradfield R, Peech M. 1955. Physiological disease 

of rice attributable to iron toxicity. Nature 175:265. 

Sahrawat KL. 2004. Iron toxicity in wetland rice and the role of 

other nutrients. J. Plant Nutr. 27:1471-1504. 

Sahrawat KL, Diatta S, Singh BN. 2000. Reducing iron toxicity in 

lowland rice through an integrated use of tolerant genotypes 

and plant nutrient management. Oryza 37:44-47. 

Sahrawat KL, Mulbah CK, Diatta S, DeLaune, RD, Patrick WH Jr, 

Singh BN, Jones MP. 1996. The role of tolerant genotypes 

and plant nutrients in the management of iron toxicity in lowland 

rice. J. Agric. Sci. Cambridge 126:143-146. 

Sahrawat KL, Sika M. 2002. Comparative tolerance of Oryza sativa 

and O. glaberrima rice cultivars for iron toxicity in West Africa. 

Int. Rice Res. Notes 27:30-31. 

van Breemen N, Moormann FR. 1978. Iron-toxic soils. In: Soils and 

rice. Manila (Philippines): International Rice Research Institute. 

p 781-800. 

Yoshida S. 1981. Fundamentals of rice crop science. Manila (Philippines): 


Notes 

Author’s address: International Crops Research Institute for the Semi- 

Arid Tropics, Patancheru 502324, Andhra Pradesh, India, e- 

mail: k.sahrawat@cgiar.org. 

Soil acidity and related problems in upland rice in the tropics 

Kensuke Okada and Matthias Wissuwa 

Upland rice is grown on a total of 19.1 million ha in Asia (60%), 

Latin America (30%), and Africa (10%). The humid to 

subhumid climatic conditions prevailing in upland rice areas 

have typically led to various degrees of soil acidity because of 

deep weathering and leaching of cations. In fact, the major 

soils for upland rice are Ultisols and Alfisols in Asia and West 

Africa, and Oxisols and Ultisols in Latin America. Oxisols and 

Ultisols are especially typical acid soils. 

Soil constraints are less critical for irrigated rice because 

flooded conditions increase the availability of nutrients and 

stabilize soil pH closer to neutral, even in acid soils. On upland 

soils, however, soil acidity can cause yield losses of up to 

50% (Sarkarung 1986). It is therefore necessary to manage 

soil acidity and related soil constraints to increase the productivity 

of upland rice. 

Growth-limiting factors for upland rice in acid soils 

The problems of acid soils are complex and are often regarded 

as an “acid-soil syndrome.” The major growth-limiting factor 

for upland rice differs depending on the degree of soil acidity. 

Roughly speaking, where soil pH (in water) is lower than 

4.3, the Al concentration in the soil solution can be higher than 

100 µM. This is potentially toxic, with negative effects on the 

elongation of rice roots under solution culture conditions. In 

upland rice fields, such soil conditions are not the norm but 

can be induced by inappropriate management practices. For 

example, in Oxisols in the subhumid savannas of Colombia, 

this degree of acidity occurred only in the surface soil layer in 

the middle stage of rice development. This enhanced acidity 

was created by the sequential split application of urea, which 

reduced soil pH because of nitrification and by potassium fertilizers 

that exchange Al at the exchange site of clay minerals 

and increase Al concentration in soil solution in the moderately 

intensive upland rice cultivation system (Okada and 

Fischer 2001). 

Rice plants are moderately tolerant of acid soils in general, 

but showed wide genotypic variation in tolerance of this 

severe soil acidity. Semidwarf indicas were susceptible but 

tropical/temperate japonicas were usually tolerant. The causal 

factor inducing genotypic differences was investigated under 

these conditions and it was found to be low Ca availability 

rather than high Al (Okada and Fischer 2001). The relative 

growth of susceptible varieties was more correlated to exchangeable 

Ca than to exchangeable Al and Al saturation (Fig. 

1). The tolerant genotype had a higher Ca absorbing capacity 

of the roots, probably owing to higher preferential adsorption 

of Ca over Al at the surface of the root cell wall (Okada et al 

2003). 

When soil pH is higher than 4.3, the concentration of Al 

in soil solution is not toxic to the elongation of rice roots. This 

is the case for the commonly found acidic soils in West Africa 

(Ultisols and some Alfisols), where the major chemical soil 

constraint is low P content/availability rather than soil acidity 

itself. 


Relative yield (%) 

110 

A B C D 

r = –0.51 (ns) r = 0.51 (ns) r = 0.58* r = 0.43 (ns) 

100 

90 

80 

70 

4.4 4.6 4.8 1.5 2.0 2.5 3.0 0.2 0.3 0.4 75 80 85 

pH Exch-Al (cmol kg –1 ) Exch-Ca (cmol kg –1 ) Al sat. (%) 

Fig. 1. The response of susceptible varieties of upland rice ( Oryzica 1; Oryzica Llanos 5) to soil pH 

(A), exchangeable Al (B), exchangeable Ca (C), and Al saturation (D) in the topsoil of savanna (0 to 20 

cm). Results of 1993 to 1996 at two sites in savannas of Colombia. 

Genetic improvement of upland rice 

for tolerance of soil acidity 

Strongly acid soils are most predominant in South America. 

Besides the Amazonian rainforest, subhumid tropical savannas 

cover 240 million ha (Cerrados in Brazil and Llanos in 

Colombia and Venezuela). Dominant soils are the highly weathered 

Oxisols and Ultisols. A typical example of the chemical 

characteristics are pH 4.8 (1:1 water), organic matter content 

4.2%, available P (Bray-2) 2.5 ppm, and exchangeable cations 

at Ca 0.18, Mg 0.13, K 0.09, and Al 2.1 cmol kg –1 soil. 

The aluminum saturation of the soil is 84%. Thus, these soils 

are characterized by low pH and low effective CEC (sum of 

exchangeable cations), with most of the CEC occupied by exchangeable 

Al. 

To develop rice varieties with high yield potential and 

acid-soil tolerance for the purpose of introducing upland rice 

in the rice-pasture cropping system of the savanna region was 

one objective at CIAT/Colombia. More than 1,360 cultivars 

of diverse genetic background were evaluated for acid-soil tolerance 

in the typical Oxisols (with 84–86%Al saturation) of 

the Colombian savannas (Sarkarung 1986). The acid-tolerant 

accessions were crossed with blast-resistant lines, further evaluated, 

and finally a commercial cultivar, Oryzica Sabana 6, was 

released in Colombia in 1991. This variety frequently achieved 

outstanding yield (3,500–4,000 kg ha –1 ) under upland acid 

savanna conditions, and has desired characteristics such as long 

grain and high amylose content. In 1996, a succeeding variety, 

Oryzica Sabana 10, was developed. 

Marker-assisted breeding for low-P tolerant varieties 

As mentioned above, phosphorus (P) deficiency is a major 

abiotic stress that limits rice productivity, particularly under 

upland conditions in acid soils such as Ultisols and Alfisols 

(Kirk et al 1998). In addition to areas of low absolute soil-P 

content, P deficiency can also arise in soils where P is strongly 

bound to soil particles. Dobermann et al (1998) estimated that 

more than 90% of added fertilizer P may be rapidly transformed 

to P forms that are not easily available to plants. The development 

of rice cultivars capable of using a higher portion of the 

fixed P already present in soils may be an attractive and costeffective 

approach to increasing rice yields where P deficiency 

is the major constraint. 

One promising step toward developing more P-efficient 

cultivars was the identification of the Pup1 locus in a QTL 

mapping population that was derived from rice cultivars 

Nipponbare (low P uptake) and Kasalath (high P uptake) and 

that had been evaluated on a highly P-fixing volcanic ash soil 

(Wissuwa et al 2002). This locus was transferred from the donor 

variety Kasalath into a Nipponbare genetic background. 

The resulting near-isogenic line, NIL-Pup1, was genetically 

92% identical to Nipponbare, but, when grown on the highly 

P-fixing soil, NIL-Pup1 had a threefold higher biomass than 

Nipponbare (Fig. 2). This improvement in growth was due to 

the ability of the NIL to acquire more P from the soil. Higher 

root growth rates and more efficient P uptake per unit root size 

(uptake efficiency) were the main differences between NIL- 

Pup1 and Nipponbare (Fig. 2). These effects were detected 

only under P deficiency but not when additional P (50 kg ha –1 ) 

was supplied to the deficient soil. This suggested that Pup1 

did not affect root growth potential per se but that it helped 

maintain root growth when P was limiting. Model simulations 

have shown that an increase in P uptake efficiency in the range 

observed for NIL-Pup1 would be sufficient to explain subsequent 

increases in P uptake and plant growth, including root 

growth (Wissuwa 2003). Higher P-uptake efficiency is therefore 

the most likely mechanism affected by the Pup1 locus. 

Currently, efforts at IRRI are directed toward identifying 

the gene(s) at the Pup1 locus. 

Further fine-mapping of Pup1 has advanced considerably 

and Pup1 has now been mapped to a 240-kb interval containing 

31 putative genes (Wissuwa, unpublished data). Most 

of these are hypothetical genes without similarity to known 

genes and those few that were functionally annotated were not 

associated with P metabolism, root growth, or P-uptake mechanisms 

(based on the Nipponbare sequence data). This suggests 

that the Pup1 locus either represents a novel gene or that the 

gene is simply absent or highly distorted in Nipponbare. It 

could also indicate, however, that the Pup1 locus performs a 


Nipponbare NIL-Pup1 t-test 

–P 

P uptake (mg) (–P) 1.89 4.38 ** 

P uptake (mg) (+P) 12.44 13.68 ns 

Root weight (g) (–P) 0.85 1.46 ** 

Root weight (g) (+P) 3.25 3.51 ns 

P-uptake efficiency a (–P) 2.20 3.02 ** 

+ Pup1 – Pup1 

P-uptake efficiency a (+P) 3.82 3.89 ns 

Fig. 2. (Left) Effect of the Pup1 locus on plant growth on a highly P-deficient, highly P-fixing soil 

(Andosol). NIL-Pup1 (+Pup1) has about three times as much P uptake, root dry weight, and total 

biomass as Nipponbare (–Pup1). (Right) Effect of the Pup1 locus on P uptake, root weight, and 

P-uptake efficiency. P-uptake efficiency is estimated as mg P taken up per g of root weight. 

regulatory function, only indirectly affecting the uptake mechanism. 

References 

Dobermann A, Cassman KG, Mamaril CP, Sheehy JE. 1998. Management 

of phosphorus, potassium, and sulfur in intensive, 

irrigated lowland rice. Field Crops Res. 56:113-138. 

Kirk GJD, George T, Courtois B, Senadhira D. 1998. Opportunities 

to improve phosphorus efficiency and soil fertility in rainfed 

lowland and upland rice ecosystems. Field Crops Res. 56:73- 

92. 

Okada K, Fischer AJ. 2001. Adaptation mechanisms of upland rice 

genotypes to highly weathered acid soils of South American 

savannas. In: Ae N et al, editors. Plant nutrient acquisition: 

new perspectives. Tokyo (Japan): Springer-Verlag Tokyo. 

p 185-200. 

Okada K, Fischer AJ, Perez-Salasar FA, Cañon-Romero Y. 2003. 

Difference in the retention of Ca and Al as possible mechanisms 

of Al resistance in upland rice. Soil Sci. Plant Nutr. 

49:889-895. 

Sarkarung S. 1986. Screening upland rice for aluminum tolerance 

and blast resistance. In: Progress in upland rice research. Proceedings 

of the Second International Upland Rice Conference, 

4-8 March 1985. Manila (Philippines): International Rice 


Wissuwa M. 2003. How do plants achieve tolerance to phosphorus 

deficiency: small causes with big effects. Plant Physiol. 

133:1947-1958. 

Wissuwa M, Wegner J, Ae N, Yano M. 2002. Substitution mapping 

of Pup1: a major QTL increasing phosphorus uptake of rice 

from a phosphorus-deficient soil. Theor. Appl. Genet. 105:890- 

897. 

Notes 

The physiological foundation of crop breeding 

for stress environments 

A. Blum 

Authors’ addresses: K. Okada, National Agricultural Research Center, 

3-1-1 Kannondai, Tsukuba, Ibaraki 305-8666, Japan; M. 

Wissuwa, International Rice Research Institute, Los Baños, 

Philippines. 

Plant sciences that address the different physiological, genetic, 

and environmental aspects of crop yield are becoming more 

divergent as time progresses. This is a result of the explosion 

of knowledge in each discipline on the one hand and the human 

limitation in coping with such an amount of knowledge. 

As expertise becomes more specific, the chances for interaction 

between scientific disciplines are diminishing. This is accentuated 

by funding agencies that do not consistently encour- 

age integrated research in plant sciences. Henceforth, while 

the problem of crop production under environmental stress is 

a problem of the whole organism and the whole cropping systems 

in relation to their environment, science is becoming less 

capable of dealing with this whole while it becomes proficient 

in dealing with increasingly smaller components of the whole. 

This is reflected in the young upcoming scientists just out of 

universities who enter the field of abiotic stress resistance with 


the ambition of making a contribution. As they were not always 

trained to understand the whole, they try to define and 

address abiotic stress resistance through the constrained view 

that they have on this trait as seen from the peephole of their 

own narrow discipline. This definition is naturally constrained 

and very often misleading. 

The definition of stress resistance is not just a question 

of scientific formality. It is a major key to understanding what 

stress resistance entails and how plant physiology can make a 

contribution to the genetic improvement of crop-plant stress 

resistance by way of conventional breeding, molecular biology, 

and genomics. This paper will attempt to explain this issue. 

There is no unified abiotic stress resistance at the level 

of the gene or the whole plant. It is very hard to accept the 

often suggested conclusions of functional genomics studies in 

that a certain gene ascribes plant resistance to cold, heat, 

drought, salinity, and probably also to a certain disease. However, 

a unified definition of abiotic stress resistance was put 

forward more than 30 years ago by Levitt (1972) and discussed 

by Blum (1988) in relation to plant breeding. This definition 

provides the basis for applying stress physiology to crop improvement. 

There is no unified protocol for assessing stress 

resistance. However, a unified concept and definition of stress 

resistance can help us move forward in understanding a specific 

stress resistance, its correct measurement, and its manipulation 

in breeding. 

Abiotic stress resistance is controlled by stress avoidance 

and/or stress tolerance. Stress avoidance enables the plant 

to avoid exposure of the plant system(s) to the stressor by excluding 

the stressor or the strain it produces from the system(s). 

Stress tolerance is the capability to sustain plant function in 

the presence of the stressor or the strain in the plant. 

Stress avoidance 

Avoidance mechanisms are diverse and depend on the type of 

stress. Avoidance can take place in the whole plant, the organ, 

or the cellular levels. 

Some examples are warranted. For drought, despite the 

receding supply of water, a plant can maintain hydration by 

either efficient moisture capture from soil or by reduced water 

use at a minimal cost to assimilation. The plant can maintain 

cellular hydration through osmotic adjustment, despite a reduction 

in whole-plant water potential. For salinity stress, plants 

such as halophytes can avoid salinization by excluding sodium 

at the root. Plants can also activate specific vacuolar Na+/H+ 

antiports, which enable vacuolar accumulation and 

compartmentation of toxic sodium, avoiding its presence in 

the cell main sites of life processes. 

Similarly, there are mechanisms of avoidance for various 

mineral toxicity stresses, mineral deficiency stress, and 

even heat stress. Even for freezing stress, the supercooling of 

internal tissue water enables avoiding of cell dehydration, which 

is caused by extracellular ice formation. For heat stress, various 

whole-plant features such as transpiration and reflective 

leaf surfaces allow decreasing the heat load on the plant and 

thus avoiding excessive heating. 

Stress tolerance 

As a general rule, plants can tolerate the presence of the stressor 

and its strain mainly by reverting to a quiescent state, such 

as dormancy. The most notable agronomic case is freezing tolerance 

as seen in winter cereals and hardy trees. Plants generally 

do not grow at freezing temperatures but they can survive 

this temperature in a dormant state and recover when temperature 

returns to normal. Similarly, dehydration tolerance can 

take place in dormant embryos in the dry seed or in a resurrection 

plant that can survive desiccation. Both will return to 

growth when re-hydrated. 

The following not quite hypothetical example will serve 

to accentuate the importance of the physiological dissection 

of abiotic stress resistance phenomena in terms of their definitions. 

A certain research found that two cultivars differed in 

their photosystem activity under drought stress. Both cultivars 

had an activity of 3 empirical units under nonstress conditions. 

Under drought stress, cultivar A had an activity of 1 unit while 

cultivar B had an activity of 2 units. Hence, cultivar B was 

taken as stress tolerant in terms of photosystem activity and 

researchers stated that they will now embark on research to 

investigate what biochemical or structural factors allow the 

resistant B to sustain photosystem activity under drought and 

perhaps this could open a new avenue into breeding droughttolerant 

crops. However, if the investigators would have tested 

their plants for water status (e.g., relative water content, RWC), 

they would have found that cultivar A had 60% RWC while B 

had 85% RWC under stress. This would have indicated that 

photosystem activity was higher in B because B was dehydration-avoidant. 

Its advantage over A in photosystem activity had 

nothing to do with photosystem biochemistry under stress. 

Cultivar B simply was not exposed to dehydration as was cultivar 

A. 

Without citing the hundreds of papers dealing with the 

agronomic and physiological aspects of drought stress resistance 

(e.g., www.plantstress.com), it can be concluded that the 

most common and effective factor of drought resistance in crop 

plants as well as native vegetation (e.g., Chaves et al 2002) is 

dehydration avoidance rather than dehydration tolerance. Dehydration 

tolerance as conditioned by a dormant state is by 

and large less prevalent in the agronomic domain where growth 

and production under stress are prime requirements. Indeed, 

there are limited cases where drought tolerance is required, 

such as cereal seedling survival under post-germination drought 

in the Mediterranean region. It is important, however, to understand 

the difference between survival as a component of 

tolerance and sustained productivity as enabled by avoidance. 

On the other hand, freezing and heat resistance are exceptions 

in that tolerance constitutes a major mechanism. 

This brings the discussion to another example, involving 

ABA. Molecular biology research recognized that one path- 


way of stress signal transduction involves ABA production and 

ABA-responsive genes. The immediate conclusion turned into 

an axiom was that ABA was a “stress hormone” responsible 

for plant resistance to stress. This axiomatic conclusion ignores 

past plant physiology research indicating that ABA causes 

reproductive failure and growth retardation, which oppose productivity. 

On the other hand, ABA induces dormancy. Hence, 

one can generalize that ABA may be a positive factor in stress 

tolerance and survival and a negative factor in stress avoidance 

and productivity under stress. It would therefore be unlikely 

for crop breeders seeking to sustain yield under stress to 

consider ABA as a stress resistance factor in crop plants. Moreover, 

recent research even indicated (e.g., Landi et al 2001) 

that maize lines selected for low leaf ABA content were more 

drought-resistant in terms of yield than high-ABA lines. 

Finally, there is serious concern among leading stress 

physiologists about the quality of testing for stress resistance 

in the domain of molecular biology. Too many studies that attempt 

to prove the effect of a certain gene or a mutation on the 

stress tolerance of experimental transgenic plants are unacceptable 

and unreasonable in terms of stress physiology. Testing 

may be faulty, measurements physiologically unacceptable, 

and the conclusions drawn can be erroneous. By not adhering 

to strict and legitimate stress physiology, such tests defeat their 

own purpose of identifying suitable genes for use in stress 

breeding. What is needed, rather, is a legitimate physiological 

dissection of gene function under stress in terms acceptable to 

physiologists and breeders. 

Conclusions 

This paper underlines the importance of being able to integrate 

the relevant knowledge produced over time in various 

plant science and agronomic disciplines toward breeding for 

stress resistance. It highlights the requirement for dissecting 

apparent abiotic resistance phenomena and resistance gene 

actions in the correct physiological terms of the immediate 

products and whole-plant responses. As plant and environmental 

sciences are becoming more fragmented and divergent, it 

will become very difficult to sustain applied research in plant 

stress resistance without the appropriate integration of plant 

sciences in education. Plant stress perspectives in molecular 

biology, physiology, genetics, breeding, agronomy, and environmental 

sciences should be merged into a new educational 

discipline of plant abiotic stress in institutes of higher education. 

References 

Blum A. 1988. Plant breeding for stress environments. Boca Raton, 

Fla. (USA): CRC Press. 208 p. 

Chaves MM, Pereira JS, Maroco J, Rodrigues ML, Ricardo CPP, 

Oserio ML, Carvalho I, Faria T, Pinheiro C. 2002. How do 

plants cope with water stress in the field Photosynthesis and 

growth. Ann. Bot. 89:907-916. 

Landi P, Sanguineti MC, Conti S, Tuberosa R. 2001. Direct and correlated 

responses to divergent selection for leaf abscisic acid 

concentration in two maize populations. Crop Sci. 41:335- 

344. 

Levitt J. 1972. Responses of plants to environmental stresses. New 

York, NY (USA): Academic Press. 695 p. 

Notes 

Author’s address: www.plantstress.com, PO Box 16246, Tel Aviv, 

Israel, e-mail:ablum@plantstress.com. 

Expression of a serine protease 

during microsporogenesis in rice 

Kentaro Kawaguchi, Naoshi Dohmae, Shuichi Matsuba, Hideyuki Funatsuki, Yutaka Sato, and Masao Ishimoto 

Pollen formation is one of the most important processes in 

plant reproduction and is widely known to be very sensitive to 

various forms of environmental stress, including low temperature. 

The most sensitive stage of pollen was shown to be the 

young microspore stage that was demonstrated using pot-grown 

rice under phytotron conditions (Satake and Nishiyama 1970). 

Male sterility induced by chilling temperature (12 to 18 °C) at 

the booting stage causes a serious loss of grain yield in rice 

and this chilling injury has been one of the most serious problems 

for agriculture in the temperate region (Yoshida 1981). 

To solve this problem, it is important to understand the physiological 

nature of anther development under normal and chill- 

ing temperature. Although the microscopic size of the rice anther 

has been a factor limiting progress in its biochemical and 

physiological analyses, recent advances in analyses of the rice 

genome have allowed us to use new approaches to solve the 

problem, such as small material size. For example, changes in 

the expression of transcripts in rice anthers caused by low temperature 

could be monitored using cDNA microarray analysis 

satisfactorily (Yamaguchi et al 2004). Our research objective 

is to discover the genes that have a key role in chilling injury 

by proteome analysis. In this study, we screened proteins separated 

from normal and chilled anthers, and the selected glycoprotein, 

which was specifically down-regulated by chilling, for 


their identification. We will discuss the possible function of 

the proteins in pollen development and susceptibility to low 

temperature. 

Proteome analysis for development of rice anthers 

Screening anther proteins 

Proteins, which would have important roles in microsporogenesis, 

were expected to accumulate in a stage-specific manner 

in developing anthers. First, we examined anther proteins 

from rice plants grown under normal conditions (25 °C day/ 

19 °C night). Anther proteins prepared from the two distinct 

stages (microspore and heading) were separated by SDS-PAGE 

and visualized with Ponceau S dye or a glycoprotein detection 

kit. However, neither staining method showed remarkable 

changes in patterns of the protein bands between the two stages. 

On the other hand, the combination of SDS-PAGE and concanavalin 

A (Con A) lectin blotting analysis showed that several 

glycoproteins accumulated stage specifically. The 85-, 

82-, and 55-kDa bands gradually increased, and 64- and 63- 

kDa bands appeared transiently at specific stages. The stagespecific 

accumulation suggested that these glycoproteins function 

in the temporal events that occurred during anther development. 

Next, the effects of low temperature on the accumulation 

of glycoproteins were examined. Rice plants were kept at 

12 °C for 4 days at the booting stage for low-temperature treatment. 

Chilling stress modified the accumulation of a particular 

glycoprotein, and the 85-kDa glycoprotein was down-regulated 

in chilled anthers. In contrast, the other glycoproteins 

showed little change by the chilling treatment. So, we selected 

the 85-kDa glycoprotein for further analysis as a candidate 

that could be involved in the mechanisms of chilling injury in 

rice. 

Primary structure of the glycoprotein 

Two-dimensional gel electrophoresis (2DE) showed that the 

85-kDa glycoprotein was apparently separated as multiple spots 

at pI 5–6. We obtained the internal peptides and analyzed the 

amino acid residues by a protein sequencer and MALDI-TOF 

MS. Comparison of the sequence information with SWISS- 

PROT and EMBL using the BLAST program showed higher 

homology with the subtilisin-like serine proteases expressed 

in developing anthers in rice (Yoshida and Kuboyama 2001), 

lily (Kobayashi et al 1994), and tomato (Riggs et al 2001). 

Western analysis using the antisera raised against the lily protease 

showed a cross reaction to the glycoprotein from rice 

anthers that confirmed the result of the structural analysis. 

The primary structure reveals that the glycoprotein possesses 

all the characteristics of the subtilisin-like serine protease 

family. The amino acids that are forming the catalytic 

triad—aspartate (D), histidine (H), and serine (S)—and the 

substrate binding site, asparagine (N), at the active site are 

conserved. The N-terminal region is hydrophobic, suggesting 

that it acts as a signal peptide, and is transported via a secretory 

pathway. 

Possible function of serine protease 

in the development of rice anther 

Subtilisin-like serine proteases are a family of endoproteases 

present in diverse tissues in microorganisms, plants, and animals, 

and they are believed to be present in the extracellular 

spaces, on the plasma membrane, and on the Golgi apparatus. 

The essential functions in proteolysis for cell development were 

reported. The yeast (Saccharomyces cerevisiae) kex2 gene 

encoding a membrane-bound subtilisin-like serine protease was 

required for processing of precursors of α-factor and killer 

toxin (Mizuno et al 1988), and so on. In contrast, limited information 

on biological functions is available in plants, such 

as kex2p-like protease activity in tobacco suspension culture 

cells (Jiang and Rogers 1999). 

Developmentally regulated fluctuation of the protease 

during microsporogenesis suggests that it has important roles 

in the development of pollen in rice anthers. Subtilisin-like 

serine proteases have been studied in detail in unicellular organisms, 

such as Bacillus spp. and yeast, etc. This case is especially 

interesting because the unicellular organisms seem to 

be analogous to microspores. It has been thought that the protease 

is transported via a secretory pathway to the extracellular 

spaces where the protease converts exogenous organic 

sources (proteins or peptides) to amino acids for obtaining 

nutrients. Rice microspores may be supplied with amino acids 

by action of the protease during their development. It is also 

possible that the protease is involved in the processing of precursors 

of important enzymes to enhance pollen development. 

However, these hypotheses are not yet demonstrated experimentally. 

The disappearance of glycoproteins under chilling 

stress might be one cause of the pollen abortion. If the protease 

has an important role in pollen development as discussed 

above, its down-regulation under chilling might cause irreversible 

injury to the microspore. To improve the chilling-susceptible 

nature of rice anthers, it is therefore necessary to understand 

the physiological role of serine protease during anther 

development. 

Future prospects 

In this paper, we described a subtilisin-like serine protease that 

accumulates specifically in developing rice anthers. Although 

structural analysis showed that a glycoprotein was identified 

as a subtilisin-like serine protease, the proteolytic activity has 

not yet been practically examined. To clarify the relationships 

between this glycoprotein and male sterility under low temperature 

in rice, it is important to know whether the glycoprotein 

actually functions in microsporogenesis. What is the natural 

substrate of the protease in anthers Where does the protease 

function The answers to these questions will result in 

important keys to understanding the physiological mechanisms 

of chilling injury in rice anthers. 


References 

Jiang L, Rogers JC. 1999. Functional analysis of a golgi-localized 

kex2p-like protease in tobacco suspension culture cells. Plant 

J. 18:23-32. 

Kobayashi T, Kobayashi E, Sato S, Hotta Y, Miyajima N, Tanaka A, 

Tabata S. 1994. Characterization of cDNAs induced in meiotic 

prophase in lily microspores. DNA Res. 1:15-26. 

Mizuno K, Nakamura T, Ohshima T, Tanaka S, Matsuo H. 1988. 

Yeast kex2 gene encodes an endopeptidase homologous to 

subtilisin-like serine proteases. Biochem. Biophys. Res. 

Comm. 156:254-256. 

Riggs CD, Zeman K, De Guzman R, Rzepczyk A, Taylor AA. 2001. 

Antisense inhibition of a tomato meiotic proteinase suggests 

functional redundancy of proteinases during microsporogenesis. 

Genome 44:644-650. 

Satake T, Nishiyama I. 1970. Male sterility caused by cooling treatment 

at the young microspore stage in rice plants. V. Estimation 

of pollen developmental stage and the most sensitive stage 

to coolness. Proc. Crop Sci. Soc. Jpn. 39:468-473. 

Yamaguchi T, Nakayama K, Hayashi T, Yazaki J, Kishimoto N, 

Kikuchi S, Koike S. 2004. cDNA microarray analysis of rice 

anther genes under chilling stress at the microsporogenesis 

stage revealed two genes with DNA transposon castaway in 

the 5′-flanking region. Biosci. Biotechnol. Biochem. 68:1315- 

1323. 

Yoshida S. 1981. Fundamentals of rice crop science. Los Baños (Philippines): 


Yoshida KT, Kuboyama Y. 2001. A subtilisin-like serine protease 

specifically expressed in reproductive organs in rice. Sex. Plant 

Reprod. 13:139-199. 

Notes 

Authors’ addresses: Kentaro Kawaguchi, Shuichi Matsuba, Hideyuki 

Funatsuki, Yutaka Sato, and Masao Ishimoto, National Agricultural 

Research Center for Hokkaido Region (NARCH), 

Hitsujigaoka, Toyohira, Sapporo, Hokkaido 062-8555, Japan, 

e-mail: kentaro@affrc.go.jp; Naoshi Dohmae, The Institute 

of Physical and Chemical Research (RIKEN), Hirosawa, 

Wako, Saitama 351-0198, Japan. 

Acknowledgment: We thank Professor K. Hiratsuka (Yokohama 

National University) for his kind gift of the antiserum. 

Responses to chilling temperature at the early stage 

of development in rice: geographical clines 

and genetic bases as revealed by QTL analysis 

Kazumitsu Onishi, Noriko Ishigoh-Oka, Mieko Adachi, and Yoshio Sano 

Common wild rice as well as indica-type rice are distributed 

in tropical and subtropical areas, while japonica type rice is 

grown from tropical to temperate areas. Tolerances of chilling 

or low temperature are considered to play a role in the wide 

distribution of cultivated rice toward the northern region. The 

response to chilling temperature (nonfreezing temperature) is 

a complex phenomenon that includes primary and secondary 

injury by exposure to low temperature, depending on the stage 

of development (Kratsch and Wise 2000, Allen and Ort 2001). 

The genetic variation in responses to low temperature has been 

reported in cultivated rice to be associated with its taxonomic 

groups as well as geographic distribution (Oka 1958, Kotaka, 

and Abe 1988, Nagamine and Nakagahra 1990). Among various 

responses to low temperature, acclimation has not yet been 

evaluated substantially in rice. Acclimation, which is a response 

to cold temperature by switching to a more cold-tolerant physiological 

state, is known to play a role in freezing tolerance in 

various plant species, including wheat and barley (Thomashow 

1990). The objectives of this study are to investigate genetic 

complexities in response to low temperature among wild and 

cultivated strains and to evaluate their genetic bases by QTL 

analysis. 


We examined responses to chilling temperature at the three 

stages of early development (germination, plumule, and seedling 

stages), using 40 cultivated and 21 wild strains. These 

were selected from diverse origins distributed from tropical 

areas to Hokkaido (43 o N), Japan. Hokkaido is one of the northernmost 

areas of rice cultivation. Chilling tolerance was evaluated 

as follows: 

1. Germinability under chilling temperature was tested 

with hulled seeds placed on moistened filter papers 

in petri dishes which were incubated in the dark at 13 

°C. The number of germinated seeds was counted 

daily until 20 d after sowing. The germination coefficient 

(percentage of germinated seeds/mean days to 

germination) was used for comparison. 

2. Chilling tolerance at the plumule stage was tested by 

incubating germinated seeds with plumule length of 


1.5 cm, at 0–1 o C for 2 d. After they grew at 26 o C for 

6 d, the degree of injury was compared. 

3. Chilling tolerance at the seedling stage was tested using 

12-d-old seedlings. The seedlings were grown at 

5 °C for 4 d, and the degree of recovery was observed 

5 d after growing at 26 o C. 

4. Acclimation was evaluated using 12-d-old seedlings. 

All the examined strains appeared to die after incubation 

at 5 °C for 8 d; however, some were found to 

be tolerant after acclimation. For acclimation, the 

seedlings were incubated at 12 °C for 3 d before treatment 

at 5 °C for 8 d. The degree of recovery was 

observed 5 d after growing at 26 o C, giving an index 

of acclimation. 

All the degrees of tolerance were scored as an index value 

ranging from 0 (susceptible type) to 4 (tolerant type), making 

two replicates of five seedlings per line. The seedlings were 

grown under 14-h photoperiod. 

To elucidate the genetic bases of chilling tolerance, the 

recombinant inbred lines (RILs) from the cross between A58, 

a japonica type from Hokkaido (43 o N), Japan, and W107, a 

wild strain from Orissa (20 o N), India, were analyzed. After 

crossing A58 with W107, 79 RILs were established from the 

F 5 plants by means of the single-seed-descent method. In total, 

263 markers, including the reported PCR-based markers 

(SSR, STS, and CAPS markers) as well as MITE markers 

(Takagi et al 2003), were used for map construction. The F 5 

and F 6 plants of each RIL were used for QTL analysis of chilling 

tolerance. QTL analysis was conducted by simple interval 

mapping and MQM mapping methods using MapQTL version 

4.0 (Van Ooijen et al 2002). 


Genetic variation in chilling tolerance 

and its geographic distribution 

A wide range of variation was observed in all four index values; 

however, the pattern of genetic variation was different 

between taxa and varietal groups (Table 1). Among the tolerances 

examined, only germinability under chilling temperature 

was distinctly differentiated between cultivated and wild 

forms. No tendency was found that cultivars from Hokkaido 

showed a higher germinability under chilling temperature, suggesting 

that germinability under chilling temperature might 

have been selected for during domestication. At the plumule 

stage, japonica (including javanica) type rice was distinctly 

more tolerant than indica and wild strains. This tolerance is 

considered to reflect changes in the indica-japonica differentiation 

since it is diagnostic between the two types, as reported 

by Oka (1958). The same tendency was also found in the ability 

of acclimation, suggesting that the variation might be related 

to an intrinsic difference between the two types. However, 

another explanation is that the variation might have resulted 

from adaptation to environments within a taxonomic 

group. 

Table 1. Comparisons of the means of index values for chilling 

tolerance among japonica, indica, and wild rice strains. 

Trait Type a No. of Mean b (range) 

strains 

Germinability Jp 27 2.9 a (2.1–4.0) 

In 13 2.6 a (1.7–3.1) 

W 21 0.8 b (0–2.4) 

Plumule Jp 27 2.2 a (0.2–4.0) 

In 13 1.0 b (0–4.0) 

W 21 0.9 b (0–2.5) 

Seedling Jp 27 3.5 a (1.3–4.0) 

In 13 2.3 a (0–4.0) 

W 21 2.7 a (0–4.0) 

Acclimation Jp 27 3.8 a (2.0–4.0) 

In 13 1.9 b (0–4.0) 

W 21 1.6 b (0–4.0) 

a Jp, In, and W indicate japonica (and javanica), indica, and wild strains, respectively. 

b Values followed by different letters are significantly different at the 

1% level (Scheffe’s test). 

To examine micro-evolutionary forces, the relation between 

the index values and latitudes of their origins was investigated. 

Comparisons were made separately in wild and 

cultivated strains. Regarding tolerance in the seedling stage, a 

latitudinal cline was reported in cultivated strains (Nagamine 

and Nakagahra 1990). Our study revealed that tolerance of the 

12-d-old seedling was weakly correlated with latitude in cultivated 

strains but not in wild strains. On the other hand, a significant 

correlation was detected in both plumule tolerance and 

acclimation ability within either wild or cultivated strains (Fig. 

1). The distribution of the indica type is usually restricted to 

the tropics, while that of the japonica type (including javanica) 

is from tropical to temperate areas, which makes it difficult to 

distinguish contemporary and historical processes as mentioned. 

In contrast, the wild progenitor may reflect the ongoing 

processes responding to environments. In wild rice strains, 

higher correlations were detected in plumule tolerance and 

acclimation ability. It should be noted that acclimation has not 

been recognized as an adaptive factor in rice. The highest correlation 

in acclimation suggests that acclimation plays a significant 

role in chilling tolerance under natural conditions. 

The genetic architecture in chilling tolerance 

All the four tolerances examined were intercorrelated with each 

other among 61 rice strains. Among the responses to chilling 

temperature, acclimation was strongly correlated with the other 

tolerances (r = 0.60 to 0.71), showing a complex phenomenon 

of adaptation to chilling temperature. The genetic bases were 

investigated by QTL analysis using 79 RILs derived from A58 

× W107. A58 (a japonica type from Hokkaido, Japan) and 

W107 (annual type of Oryza rufipogon from Orissa, India) 

were tolerant and susceptible for the responses to chilling temperature 

examined, respectively. The map of 1,302 cM in total 

length was constructed with 263 molecular markers. Among 

RILs, no significant correlation was observed between the tol- 


Latitude (°N) 

40 

30 

40 

30 

Japonica (and javanica) 

Indica 

Wild 

20 

20 

10 

0 

10 

r = 0.44** r = 0.50** 

0 

–10 

0 1 2 3 4 

–10 

0 1 2 3 4 

40 

40 

30 

30 

20 

20 

10 

10 

0 

r = 0.45** 

0 

r = 0.70** 

–10 

0 1 2 3 4 

Tolerance 

–10 

0 1 2 3 4 

Tolerance 

Fig. 1. The relations between latitudinal locations and chilling tolerance at the plumule stage (A) or the 

ability of acclimation (B) among 40 cultivated and 21 wild rice strains. ** indicates significance at the 1% 

level. 

erances except between the tolerances of seedlings with and 

without acclimation, suggesting that their genetic bases are 

largely independent. Based on QTL analysis, the three significant 

QTLs for plumule tolerance were detected on chromosome 

1 (LOD = 3.9, near the centromere), chromosome 11 

(LOD = 6.0, long arm), and chromosome 12 (LOD = 6.8, short 

arm). For the other three tolerances, however, only lower LODpeaks 

(suggestive QTLs) were observed on chromosomes 5, 

11, and 12. 

Furthermore, the three chilling tolerances, excluding germinability, 

were significantly correlated with genomic similarity 

index (GSI). GSI was computed as a mean value of 109 

codominant PCR-based markers by assigning the genotypes 

as 1 (A58), –1 (W107), and 0 (heterozygote), expressing the 

genomic likeness of an RIL to the parental strains. The results 

showed that chilling tolerance might be controlled by multiple 

QTLs distributed across the genome as supposed from the QTL 

analysis. In the northernmost areas of cultivation, it is considered 

that rice plants have been improved by adaptive changes 

in response to low temperature. However, our study revealed 

that no major QTLs are involved in the tolerances examined. 

These results demonstrate that the responses to chilling temperature 

are rather complex traits and that the combination of 

various genes acting on different stages of development or 

under different temperature regimes might be responsible for 

their local adaptation. 

References 

Allen DJ, Ort DR. 2001. Impacts of chilling temperatures on photosynthesis 

in warm-climate plants. Trends Plant Sci. 6:36-42. 

Kratsch HA, Wise RR. 2000. The ultrastructure of chilling stress. 

Plant Cell Environ. 23:337-350. 


Kotaka S, Abe N. 1988. The varietal difference of germinability at 

low-temperature in rice varieties and the testing method for 

the percentage establishment of seedlings. J. Agric. Sci. Tokyo 

43:165-168. (In Japanese.) 

Nagamine T, Nakagahra M. 1990. Genetic variation of chilling injury 

at seedling stage in rice, Oryza sativa L. Jpn. J. Breed. 

40:449-455. 

Oka HI. 1958. Intervarietal variation and classification of cultivated 

rice. Ind. J. Genet. Plant Breed. 18:79-89. 

Takagi K, Nagano H, Kishima Y, Sano Y. 2003. MITE-transposon 

display efficiently detects polymorphisms among the Oryza 

AA-genome species. Breed. Sci. 53:125-132. 

Thomashow MF. 1990. Molecular genetics of cold acclimation in 

higher plants. Adv. Genet. 28:99-131. 

Van Ooijen JW, Boer MP, Jansen C, Maliepaard C. 2002. MapQTL 

4.0: software for the calculation of QTL position on genetic 

maps. Wageningen (Netherlands): Plant Research International. 

Notes 

Authors’ address: Graduate School of Agriculture, Hokkaido University, 

Sapporo 060-8589, Japan, e-mail: 

onishi@abs.agr.hokudai.ac.jp. 

QTL analysis on plasticity in lateral root development 

in response to water stress in the rice plant 

Hong Wang, Yoshiaki Inukai, Akihiko Kamoshita, Len Wade, Joel Siopongco, Henry Nguyen, and Akira Yamauchi 

Lateral roots constitute a great proportion of the root system 

of plants in quantity. Two distinct types of lateral roots have 

been identified in rice: the L type, which is generally long and 

thick, and capable of branching into higher-order laterals; and 

the S type, which is short and nonbranching but normally numerous 

(Yamauchi et al 1996). These two lateral roots differ 

in vascular structure in such a way that the vascular system of 

the S type is much less developed (Kono et al 1972). The ability 

of a genotype to alter its phenotypic expression in response 

to the environment is termed “phenotypic plasticity” (Bradshaw 

1965). Yamauchi et al (1996) and Bañoc et al (2000) suggested 

that plastic development of lateral roots under water-deficit 

conditions may be one of the key traits of some genotypes for 

the adaptation to such stress. This study aimed to identify QTLs 

for the plasticity of lateral root development in rice plants in 

response to water stress using a population of doubled-haploid 

lines (DHLs). We also attempted to test a hypothesis that 

developmental plasticity in response to water stress may differ 

among different types of lateral roots. 


Ninety-six DHLs were derived from a cross between CT9993- 

5-10-1-M (japonica, upland adapted) and IR62266-42-6-2 (indica, 

lowland adapted) (henceforth abbreviated as CT9993 and 

IR62266, respectively) (Kamoshita et al 2000). 

Plants were grown hydroponically using a modified 

growth pouch method in an environmentally controlled growth 

chamber for 2 weeks, where the day/night temperature was set 

at 30/25 ºC, relative humidity at 70%, and photoperiod at 16 

hours. A plastic container (25 cm in length, 19 cm in width, 27 

cm in depth) was filled with water of different PEG concentrations 

as described below. Filter papers (18.0 cm in width and 

26.0 cm in height) were hung from plastic pipes (0.8 cm in 

diam) fixed on the side wall of the container at 1-cm intervals, 

and germinated seeds of each of the DHLs were sown between 

the two neighboring sheets of filter paper so that the grown 

roots would be sandwiched between them. When harvesting, 

it was easy to remove the root samples from the filter papers 

without any loss or damage. Water stress was induced by using 

PEG-6000. Preliminary observation showed that, in the 

range of 0 to 120 g L –1 , the concentration of 80 g L –1 resulted 

in the most contrasted root response versus 0 g L –1 (control), 

and thus the concentration was set at 0 and 80 g L –1 , representing 

well-watered and water-stress conditions, respectively. 

In the harvested root systems, seminal, nodal, and lateral 

roots were identified. Lateral roots were classified further 

into three types: L-type laterals, those that were long and thick, 

and branched into a higher order; M types, those that were 

long and thick but without a branch yet; and S types, those that 

were short, slender, and nonbranching. The length of seminal, 

nodal, and L- and M-type lateral roots, number of nodal roots 

per plant, and number of the three types of lateral roots per 

unit length of seminal root axis were measured. Plasticity of a 

given trait was evaluated as the difference between mean values 

of the trait in stressed and control plants for each DHL. 

The marker map with 315 markers, including 145 restriction 

fragment length polymorphism, 153 amplified fragment 

length polymorphism, and 17 microsatellite markers, has 

been constructed at Texas Tech University, Lubbock, Texas 

(Zhang et al 2001). Total map distance is 1,788 cM. Putative 

QTLs (main-effect QTLs assuming no epistasis) for the traits 

were identified based on QTL Mapper (version 1.0) (Wang et 

al 1999a,b). Main-effect QTLs were declared significant at 

the threshold of 0.005 and with the log of odds (LOD) value 

set higher than 2.0. 


Table 1. Mean values and range of root parameters per plant of the two parents and DHL population under wellwatered 

conditions and water (PEG) stress. 

IR62266 CT9993 DHLs 

Root traits Well- PEG Well- PEG 

watered watered Well- watered PEG 

Mean Mean Mean Mean Mean Range Mean Range 

Shoot height (cm) 10.6 10.2 8.6 6.8 12.0 6.0–21.0 9.0 3.0–15.8 

Seminal root length (cm) 20.2 8.6 23.8 24.7 26.1 6.7–53.0 19.6 5.7–40.5 

Nodal root number 6.0 7.3 4.0 4.8 6.4 2.0–11.0 5.2 1.0–10.0 

Nodal root length (cm) 23.3 48.2 40.8 40.9 62.7 13.6–167.1 40.8 3.0–94.8 

L-type lateral root number a 2.0 1.8 0.1 0.6 0.7 0–2.7 0.7 0–3.2 

M-type lateral root number a 1.2 0.6 0 0.6 0.5 0–2.2 0.6 0–2.7 

S-type lateral root number a 10.2 8.7 9.9 9.1 9.4 2.8–19.6 9.6 1.6–23.0 

L-type lateral root length b 39.3 17.0 2.7 3.0 17.0 0–79.4 16.5 0–74.6 

M-type lateral root length b 21.3 1.7 0 2.0 7.3 0–40.7 5.5 0–39.3 

a On the basis of cm –1 seminal root axis. b On the basis of 0–10-cm seminal root axis base from shoot. 

Results 

The mean values and range of root parameters per plant of the 

two parents and DHL population under well-watered and water-stress 

(PEG) conditions are shown in Table 1. The parameters 

showed normal distributions, and transgressive segregations 

were observed among the DHL population. 

There was a significant difference in the response of the 

three types of lateral root development to water stress between 

the two parents. CT9993 had a shorter shoot and longer seminal 

root, but fewer lateral roots than IR62266. The shoot height 

of CT9993 was suppressed by water stress, while that of 

IR62266 was not affected. The seminal root length in IR62266 

decreased significantly under water stress, while that of CT9993 

showed little change. The parents and the DHLs showed different 

degrees of plasticity in lateral root development and 

nodal root growth in response to water stress. 

Table 2 presents the results of putative QTLs for the plasticity 

of four lateral root parameters. Two QTLs near ME97- 

K706 and K706-ME24 for the plasticity of L-type lateral root 

number per cm seminal root axis were found on chromosome 

2. Three QTLs near ME216-RG811 and RG811-ME83 on 

chromosome 1, and near G188-CDO202 on chromosome 5, 

respectively, were identified for the plasticity of S-type lateral 

root number per cm seminal root axis in response to water 

stress. The QTL for the plasticity of the longest L-type lateral 

root was detected on chromosome 7 between EM173 and 

ME215. Two QTLs contributing to the plasticity of the longest 

M-type lateral root were identified on chromosome 8 between 

ME54-ME55 and ME104-ME54, respectively. 

Table 2. Chromosome (Chr.) and marker intervals likely containing 

QTLs (P

Bradshaw AD. 1965. Evolutionary significance of phenotypic plasticity 

in plants. Adv. Genet. 13:115-155. 

Kamoshita A, Wade LJ, Yamauchi, A. 2000. Genotypic variation in 

response of rainfed lowland rice to drought and rewatering. 

III. Water extraction during the drought period. Plant Prod. 

Sci. 3(2):189-196. 

Kono Y, Igata M, Yamada N. 1972. Studies on the developmental 

physiology of the lateral roots in the rice seminal roots. Proc. 

Crop Sci. Soc. Jpn. 41:192-204. 

Wang DL, Zhu J, Li ZK, Paterson AH. 1999a. Mapping QTLs with 

epistatic effects and QTL by environment interactions by mixed 

linear model approaches. Theor. Appl. Genet. 99:1255-1264. 

Wang DL, Zhu J, Li, ZK, Paterson AH. 1999b. User manual for 

QTL Mapper version 1.0. Texas A&M Univ., College Station, 

TX. 

Yamauchi A, Pardales JR Jr, Kono Y. 1996. Root system structure 

and its relation to stress tolerance. In: Ito O, Johansen C, Adu- 

Gyamfi JJ, Katayama K, Kumar Rao JVDK, Rego TJ, editors. 

Dynamics of roots and nitrogen in cropping systems of the 

semi-arid tropics. Tsukuba (Japan): Japan International Research 

Center for Agricultural Sciences. p 211-233. 

Zhang J, Zheng HG, Aarti A, Pantuwan G, Nguyen TT, Tripathy JN, 

Sarial AK, Robin S, Babu RC, Nguyen BD, Sarkarung S, Blum 

A, Nguyen HT. 2001. Locating genomic regions associated 

with components of drought resistance in rice: comparative 

mapping within and across species. Theor. Appl. Genet. 

103:19-29. 

Notes 

Authors’ addresses: Hong Wang, Soil and Fertilizer Institute, Chinese 

Academy of Agricultural Sciences; Hong Wang, Yoshiaki 

Inukai, and Akira Yamauchi, Graduate School of 

Bioagricultural Sciences, Nagoya University; Akihiko 

Kamoshita, Field Production Science Center, University of 

Tokyo; Len Wade, School of Plant Biology, University of 

Western Australia; Joel Siopongco, International Rice Research 

Institute; Henry Nguyen, Department of Agronomy, 

Plant Sciences Unit, University of Missouri, e- 

mail:ayama@agr.nagoya-u.ac.jp. 

Acknowledgment: This research was partially supported by the Japan 

Society of the Promotion of Science with a Grant in Aid 

for Scientific Research (No. 16380015). 


The session on abiotic stresses, formally titled “Challenges to 

expanding rice production in unfavorable environments,” was 

convened by S. Tobita (JIRCAS) and R. Lafitte (IRRI). The conveners 

recognized that the session should deal with the linkages 

across different disciplines and across different levels of institutions 

and organizations for the challenge, as well as understanding 

the kinds of abiotic stresses in rice ecologies. So, we invited 

five speakers with a background in rice physiology or soil science, 

with each having specialized experience in salinity, drought, 

submergence, iron toxicity, and soil acidity, respectively. The 

molecular biology of stress responses in the rice crop were to be 

reviewed, taking into account some of the recent progress in this 

area. It was also good to have a lecture on the principles of 

stress physiology; herein the uniqueness of the rice crop would 

automatically stand out. Because the Crop Science Society of 

Japan concurrently organized a session on cold/heat tolerance in 

rice, we did not examine temperature stress in the session, but 

called for poster presentations on this subject. 

Our session began with more than 100 participants. Because 

R. Lafitte could not attend the symposium, the session 

was chaired by G.V. Subbarao (JIRCAS) and me. I briefly explained 

the aims and outlined the structure of the session, followed by 

the first speaker, Nguyen Thi Lang from the Cuu Long Delta Rice 

Research Institute (Vietnam). She discussed tissue culture and 

in vitro approaches for selecting salt-tolerant somaclones in rice, 

for which successful examples were reported. The second lecturer 

was Prof. S. Fukai of the University of Queensland, Australia 

(whose paper is co-authored by A. Kamoshita, the University of 

Tokyo), who talked about experiences with drought studies at 

several field sites in Southeast Asia, mostly in the rainfed low- 

land ecology. He emphasized the importance of interactions between 

genotype and environment in crop performance under 

drought stress. Characterization of the type of drought at a proper 

site is sometimes not regarded as important but it does affect 

the selection criteria for better drought tolerance, he explained. 

K. Futakuchi, a Japanese scientist currently working for the Africa 

Rice Center, discussed his research on submergence tolerance 

in rice, including African rice (Oryza glaberrima) and Nericas 

(interspecific hybrids of Asian and African rice). His achievements 

covered the physiological elucidation of responses to submergence 

stress and genotypic variation in submergence tolerance. 

The audience understood that O. glaberrima rice would be a good 

genetic resource for achieving rice with higher tolerance of submergence 

stress in this region through wide hybridization. 

K.L. Sahrawat, working for the International Crops Research 

Institute for the Semi-Arid Tropics (ICRISAT) in India, discussed 

another problem also prevalent in West Africa, iron toxicity, especially 

in the irrigated and rainfed lowland ecologies. With his background 

in soil chemistry, he explained a detailed process of chemical 

development of toxic iron ions in the soil. He also showed us 

the importance of nutrient management and the use of varieties 

tolerant of iron toxicity. 

K. Okada (National Agriculture Research Organization, Japan), 

in a paper co-authored by M. Wissuwa (IRRI), spoke about 

soil acidity problems in tropical upland areas. An investigation of 

typical Oxisols in South America suggested that the real cause of 

the genotypic difference did not result from tolerance of toxic Al 

but from higher Ca absorbing capacity of the roots, probably because 

of the higher preferential adsorption of Ca over Al at the 

cell wall. He also mentioned recent progress in future applica- 


tions of the marker-assisted breeding approach to improve adaptation 

to soil acidity and related nutrient deficiencies in upland 

rice. The next speaker, J. Bennett from IRRI, the only molecular 

biologist in this session, reviewed old and new techniques for the 

study of molecular mechanisms of tolerance for many kinds of 

abiotic stresses, with an example involving drought tolerance. 

The session had a general discussion after the six presentations. 

The chairman emphasized the linkages across various 

disciplines for expanding rice production in unfavorable environments. 

What and how should the linkages be among agronomy, 

physiology, soil science, genetics, breeding, molecular biology, 

and so on The speakers especially discussed drought and soil 

acidity. We recognize that a multidisciplinary characterization of 

the targeted stresses could be the best strategy for our goals. A. 

Blum, founder of plantstress.com from Israel, synthesized the 

discussion based on his long and distinguished career and handson 

experience in stress physiology. He emphasized the physiological 

foundation of crop breeding for abiotic stress tolerance, 

with a good example of drought avoidance and tolerance. Many 

in the audience were convinced that, without a physiological understanding 

of crop stress, research could go the wrong way and 

selection and breeding programs could be useless. 


SESSION 16 

Pest management with minimal 

environmental stress 

CONVENER: C. Vera Cruz (IRRI) 

CO-CONVENER: Y. Suzuki (NARO)

Integrated biodiversity management (IBM) in rice fields 

Keizi Kiritani 

Kiritani (1975) stated that the central issue for agriculture in 

the future would be how to manage and optimize biodiversity, 

stability, and productivity within agroecosystems. The paddy 

ecosystem is an integrated, water-dependent system, which can 

contain many kinds of living organisms: birds, fish, reptiles, 

amphibia, arthropods, and plants. Paddy fields were originally 

wetlands that are artificially constructed devices for rice production. 

Nowadays, very few natural wetlands remain, and 

many aquatic organisms now depend partly or fully on paddy 

fields. 

To raise both land and labor productivity, the Japanese 

government has promoted the conversion of poorly drained 

wet paddy fields into well-drained ones in association with a 

policy to consolidate fragmented farmlands. U-shaped concrete 

ditches have replaced traditional earth ditches and irrigation-supply 

canals have been separated from drainage canals, 

which effectively reduced the variety of habitats for 

aquatic organisms. 

Biodiversity of arthropod fauna in paddy fields 

Infestations by rice borers, Chilo suppressalis and Scirpophaga 

incertulas, and the intermittent and sudden occurrences of outbreaks 

of Nilaparvata lugens were the major causes of losses 

in rice yield in temperate Asia. These were the major pests 

until around 1965. Thereafter, leaf- and planthoppers and the 

viral diseases RSLV and RDV transmitted by them were predominant 

for about 30 years. Meanwhile, the rice water weevil, 

Lissorhoptrus oryzophilus, invaded Japan in 1976 from 

the United States, inflicting serious damage to rice in the late 

1980s. Since 1995, the damage caused by various kinds of 

stink bugs and mirids has become the most serious problem. 

Many species of arthropods with diverse types of life 

cycles occupy different habitats within the paddy 

agroecosystem. Sympetrum dragonflies emerge from paddy 

fields and stay in coppiced woodlots to mature sexually before 

returning to paddy fields to oviposit. The eggs hatch in 

the following spring when irrigation water becomes available. 

Newly emerged adults of the water scorpion, Ranatra chinensis, 

move from paddy fields to irrigation ponds for overwintering. 

Oviposition takes place in paddy fields in the next spring. The 

migratory planthopper pests, N. lugens and Sogatella furcifera, 

are annually replenished by a long-range immigration from 

tropical endemic habitats. 

The biodiversity of the paddy agroecosystem therefore 

depends not only on the paddy fields themselves but also on 

water channels, irrigation ponds, levees, surrounding fallow 

fields, neighboring farmlands, secondary forests, wetlands, rivers, 

and remote hibernating areas (Kiritani 2000). 

IPM perspective in relation to biodiversity 

In the past, most studies on paddy ecosystems have focused 

on productivity and its stability in terms of rice yields. 

Arthropods in paddy ecosystems can be classified into three 

main groups according to their ecological requirements: (1) 

resident species adapted to the continuous cropping of rice in 

the same field, (2) migratory species adapted to exploit rice as 

an annual crop, and (3) aquatic species originating from stillwater 

habitats in wetlands. Concerning groups 1 and 2, integrated 

pest management (IPM) programs, which have a primary 

objective of maximizing economic profit on the farm, 

have been implemented with various degrees of success. Although 

IPM is becoming widespread, those insects (Tada-nomushi 

= species of unknown or uncertain function that routinely 

occur in the habitat) that have no direct economic impact 

on rice production have been mostly ignored as an important 

element in the rice ecosystem. Consequently, some aquatic 

insects are in danger of extinction, thus requiring conservation 

(Kiritani 1979, 2000, Kiritani and Naba 1994). 

The green rice leafhopper, Nephotettix cincticeps, is 80% 

of the diet of a lycosid spider, Pardosa pseudoannulata, in 

paddy fields. No lycosid spiders, however, developed to adults 

when fed only N. cincticeps. When lycosid females were allowed 

to prey upon mixed species of prey, their fecundity 

greatly increased (Suzuki and Kiritani 1974). Those species 

such as chironomids and collembola, for example, that are 

neither pests nor natural enemies, and yet are useful as alternative 

food of generalist predators, can be referred to as minor, 

yet important, components of the community (Kiritani 

2000). 

Immigration of spiders to paddy fields occurs after the 

appearance of chironomids. Early insecticide applications to 

control rice stem borers often result in the resurgence of 

planthoppers and leafhoppers 1 month later because insecticide 

treatments simultaneously kill spiders and chironomids 

(Kobayashi 1961). In the tropics, prevention of outbreaks of 

planthoppers and leafhoppers depends on protection of earlyacting 

natural enemies by avoiding early insecticide spraying 

(Way and Heong 1994, Settle et al 1996). 

Levees are likely to act as refuges for various kinds of 

natural enemies of arthropod pests that occur in upland crops 

grown close to paddy fields. A dwarf spider, Ummeliata 

insecticeps, common in paddy fields dispersed from levees by 

ballooning in late May to uplands remains there until the end 

of the rainy season. It behaved like a specific predator attacking 

a newly hatched colony of larvae of Spodoptera litura in 

taro fields (Nakasuji et al 1973). Another example is the 

anthocorid bugs, Orius spp., that are effective natural enemies 

of Thrips palmi, a serious invasive alien pest of eggplant. O. 


IBM 

Intensity of management 

100 

IPM 

Conservation 

IPM 

Density 

IPM 

Economic injury level 

IBM 

50 

Conservation 

Conservation 

Extinction threshold 

Fig. 1. Illustration of the concepts of IPM, conservation, and IBM. 

nagaii and O. sauteri occur on rice and on white clover grown 

on levees, respectively, before invading eggplant fields in early 

June (Ohno and Takemoto 1997). 

Integrated biodiversity management (IBM) 

0 

Paddy 

field 

Levee Ditch Irrigation 

pond 

Secondary 

forest 

Fig. 2. Difference in the intensity of management in accordance 

with habitats. (Adapted from Kiritani 2000.) 

A new concept, integrated biodiversity management (IBM), 

has been proposed under which IPM and conservation are reconciled 

and made compatible with each other (Fig. 1). IPM 

requires that densities of each pest species be kept below their 

specific economic injury level. In conservation, target species 

have to be managed to remain above a specific extinction 

threshold. Since some presently rare carnivorous aquatic 

arthropods, such as Lethocerus deyrollei, and some large-sized 

dytiscid beetles have been recorded as pests of fish culture 

when they were abundant, these species also should be managed 

to keep their populations below defined economic injury 

levels (Kiritani 2000). 

The status of a pest species could be changed by IPM 

into a Tada-no-mushi (minor or nontarget insect), which can 

function as potential food for generalist predators. S. incertulas 

is currently almost extinct in Japan. From the viewpoint solely 

of an economically oriented IPM, however, this is of little consequence 

because S. incertulas was an important rice pest to 

be controlled. But, in view of IBM, such relatively rare species, 

such as S. incertulas and some aquatic insects, can be 

considered a target for conservation. 

The arthropods inhabiting paddies require various habitats 

for the completion of their life cycles. The relative importance 

of IPM and conservation changes along a continuum away 

from the paddy field through the levee, ditch, irrigation pond, 

and coppiced woodlots (Fig. 2). The two lines cross at a point 

most appropriate for a specific location as well as for the target 

species concerned (Kiritani 2000). 

The concept of IBM is not limited to the paddy ecosystem, 

but is also applicable to all types of agricultural systems. 

Crops range from those that require intense IPM intervention 

with little consideration of species conservation, for example, 

greenhouse crops, to those for which high levels of pest control 

and biodiversity preservation can be attained, for example, 

a complex home garden or backyard in the tropics. 

Conclusions 

It is necessary that a set of elemental habitats be available for 

completion of the life cycle of many species and that these 

multiple habitats be within the range of dispersal of these species. 

A set of habitats, including host plants (prey), shelter, 

hibernacula, mating places, etc., is essential to ensure the persistence 

of diverse species. For aquatic insects, irrigation ponds, 

coppiced woodlots, and poorly drained wetlands are necessary 

habitats in addition to paddy fields. Because cleaning an 

irrigation pond could result in the complete destruction of the 

aquatic fauna in the pond, neighboring ponds that supply the 

newly cleaned pond with aquatic species should exist within 

an appropriate distance, for example, 1 km for dragonflies 

(Moriyama 1997). 

It is recommended to adopt IPM strategies and tactics 

that are compatible with conservation. 

Preventing alien species from invading the paddy ecosystem 

is very important. Alien species often jeopardize the 

conservation of endangered species by competition and inducing 

additional chemical control applications. 

Special consideration should be given to avoid lethal 

effects on species that are vulnerable to pesticides, such as 

aquatic, univoltine, and carnivorous or monophagous species. 

Bioaccumulation of persistent biotoxins is far greater in aquatic 

systems than in terrestrial systems. Contamination of irrigation 

water with pesticides must therefore be avoided as much 

as possible. We can keep the amount of pesticides to a minimum 

by applying knowledge of the behavioral ecology of the 

pests. 

Farm management techniques that make the difference 

as great as possible between the population levels for an EIL 

Session 16: Pest management with minimal environmental stress 469

(economic injury level) and an extinction threshold should be 

introduced in the IBM system. As an alternative to the EIL, we 

could use another EIL in which the “E” refers to “ecological 

or environmental.” This new EIL, however, has yet to be established. 

In general, global warming may work in favor of natural 

enemies (except for spiders) by increasing the number of generations 

more than for their host species (Kiritani 1999). Biological 

control is expected to become a more important control 

tactic in the future. Uncertainty remains, however, regarding 

the extent to which host-parasitoid phenology will be synchronized 

after an increase in the number of generations. Parasitism 

and predation, similar to those in paddy fields in the 

tropics, would be expected to increase through this numerical 

response and enhance the natural control. It is inevitable that 

implementation of IBM involves some trial and error. We need 

an adaptive approach toward IBM. We should not only invite 

active involvement of persons interested in evaluation and 

improvement, but should also adopt a modest attitude toward 

developing and improving an ongoing IBM design. 

References 

Kiritani K. 1975. Pesticides and ecosystems. J. Pesticide Sci. 

Commem. Issue:65-75 (In Japanese.) 

Kiritani K. 1979. Pest management in rice. Ann. Rev. Entomol. 

24:279-312. 

Kiritani K. 1999. Shift of IPM strategy for rice under global warming 

in temperate areas. In: Zhang R, Gu D, Zhang W, Zhou C, 

Pang Y, editors. Integrated pest management in rice-based 

ecosystem. Guangzhou (China): Editorial Department of Journal 

of Zhongshan University. p 235-244. 

Kiritani K. 2000. Integrated biodiversity management in paddy fields: 

shift of paradigm from IPM toward IBM. Integ. Pest Manage. 

Rev. 5:175-183. 

Kiritani K, Naba K. 1994. Development and implementation of rice 

IPM in Japan. In: Heinrichs EA, editor. Biology and management 

of rice insects. New Delhi (India): Wiley Eastern Limited. 

p 713-731. 

Kobayashi T. 1961. The effect of insecticidal applications to the rice 

stem borer on the leafhopper populations. Special report on 

the forecasting of agricultural pests and diseases no. 6. 126 p. 

(In Japanese with English abstract.) 

Moriyama H. 1997. What is protecting paddy fields Tokyo (Japan): 

Nobunkyo. 205 p. (In Japanese.) 

Nakasuji F, Yamanaka H, Kiritani K. 1973. The disturbing effect of 

micryphantid spiders on the larval aggregation of the tobacco 

cutworm, Spodoptera litura (Lepidoptera: Noctuidae). Kontyu 

41:220-227. 

Ohno K, Takemoto H. 1997. Species composition and seasonal occurrence 

of Orius spp. (Heteroptera: Anthocoridae), predacious 

natural enemies of Thrips palmi (Thysanoptera: 

Thripidae), in eggplant fields and surrounding habitats. Appl. 

Entomol. Zool. 32:27-35. 

Settle WH, Ariawan H, Astuti ET, Cahyana W, Hakim AL, Hindayana 

D, Lestari AS, Pajarningsih. 1996. Managing tropical rice pests 

through conservation of generalist natural enemies and alternative 

prey. Ecology 77:1975-1988. 

Suzuki Y, Kiritani K. 1974. Reproduction of Lycosa pseudoannulata 

under different feeding conditions. Jpn. J. Appl. Entomol. Zool. 

18:166-170. (In Japanese with English abstract.) 

Way MJ, Heong KL. 1994. The role of biodiversity in the dynamics 

and management of insect pests of tropical irrigated rice: a 

review. Bull. Entomol. Res. 84:567-587. 

Notes 

Author’s address: Ito City, Shizuoka, e-mail: kiritanik@ybb.ne.jp. 

Habitat manipulation in sustainable pest management 

in the rice ecosystem of the Yangtze River Delta 

Xiaoping Yu, Jianming Chen, Zhongxian Lu, Xusong Zheng, Hongxin Xu, and Juefeng Zhang 

Rice planthoppers, Nilaparvata lugens Stål and Sogatella 

furcifera Horvath, are major insect pests in the Yangtze River 

Delta rice ecosystem. About 50–100 million tons of rice were 

lost per year during 1973-2000 in China because of the incidence 

of rice planthoppers (Yu et al 1999). Usually, the injudicious 

use of insecticides in the rice ecosystem caused serious 

problems, such as resurgence and resistance of insect pests, 

residue, and adverse effects on the environment. Predators and 

parasitoids were found to be an effective bio-agent to suppress 

insect pests, but natural enemies per se usually could not efficiently 

suppress insect pests below the damage threshold because 

of the low parasitism and predation rate of natural en- 

emies caused by an unstable rice habitat (made unstable by 

harvesting, pesticide application, and other practices) and winter 

interruption (Way and Heong 1994). Therefore, neighboring 

habitats are very important in conserving and enhancing 

the population of natural enemies in rice fields (Yu et al 1998). 

To evaluate the role of habitat diversity in promoting natural 

enemies in rice habitats, the movement of spiders and Anagrus 

spp. between rice and neighboring habitats was monitored. 

Meanwhile, spraying experiments to determine the optimal 

dosage of chemicals were conducted on a demonstration farm 

with various habitat manipulations. 



Test insects 

Adults of delphacids, N. lugens and Sacchasydne procerus in 

rice fields and S. procerus in neighboring Zizania caduciflora 

L. habitats, were collected, and were reared in culturing mylar 

cages separately. The egg parasitoid A. nilaparavate was collected 

by exposing potted rice and Zizania plants bearing hopper 

eggs. The spider Pirata subpiraticus was collected using a 

Blower-vacuum suction machine and sorted in the laboratory 

for further experiments (Xu et al 1987). 

Dispersal patterns of egg parasitoids 

between habitats using sticky traps 

Yellow sticky traps made with a wooden plate (50 cm 2 ) were 

used for monitoring the dispersal of Anagrus spp. on the field 

bunds between habitats. The yellow sticky papers were replaced 

every 6 hours, and Anagrus spp. trapped on sticky paper were 

identified and counted. 

Movement of the spider P. subpiraticus between rice 

and Zizania habitat measured by the capture 

and recapture test 

In the field capture and recapture test, all spiders marked with 

rubidium were released in the releasing plot of a Zizania habitat. 

Spiders in neighboring rice fields were sampled by a 

Blower-vacuum suction machine at 5, 10, 20, and 40 m from 

the releasing plot on the 1st, 3rd, 7th, and 15th day after releasing. 

All specimens were digested and analyzed for Rb content 

by an Atomic Absorption Spectrometer (Perkin Elmer 

Model 3300) at the wavelength of 780.0 nm (Berry et al 1972, 

Perfect et al 1985). 

Effects of various spraying times and dosages 

in rice fields with habitat diversity 

A field of 0.45 ha planted with hybrid rice Shanyou 63 as singleseason 

rice was selected for testing. The treatments of insecticide 

spraying in rice fields were arranged as follows: (1) 

monocrotophos was sprayed on 20 June for leaffolder, (2) 

Fipronil was sprayed on 6 July for stem borer, (3) buprofezin 

was used on 1 August and 15 August for brown planthopper, 

and (4) spraying water only was the check. All treatments were 

replicated three times. Habitat manipulation was done surrounding 

rice fields, such as planting soybean on the field bund, 

digging a canal in the field at 5-m intervals, and reconstructing 

the field bund. The arthropods in each plot were collected 

using Blower-vacuum suction machines and yield loss was 

calculated after rice harvesting. 

Results 

The results showed that many egg parasitoids (Anagrus spp.) 

moved into rice fields from neighboring Zizania habitats: about 

60 per sticky trap per day were caught during July and August. 

A high frequency of movement occurred from early June to 

early August, especially in July and August. More Anagrus 

Table 1. The number of marked and unmarked lycosid spiders recaptured 

at various distances from a releasing plot. 

Days after Distance from a Individuals Individuals 

releasing releasing site (m) captured marked 

1 

1st day 5 35 2 

10 19 0 

20 19 1 

40 28 0 

3 

3rd day 5 32 2 

10 26 1 

20 16 1 

40 18 0 

7 

7th day 5 45 0 

10 28 0 

20 15 0 

40 57 0 

15 

15th day 5 24 1 

10 25 0 

20 17 0 

40 21 1 

spp. moved into rice fields from neighboring Zizania habitats 

than from other cultivated habitats. After October, numerous 

Anagrus spp. moved back to Zizania habitats from rice habitats. 

The proper arrangement of rice fields and Zizania habitats 

could significantly enhance the number of egg parasitoids 

in rice fields. 

The capture and recapture test with Rb marking demonstrated 

that the lycosid spider could disperse up to 40 m into 

rice fields from a releasing plot on the 15th day after release. 

However, more marked spiders were captured at the 5-m site 

during the first to third day (Table 1). The results showed that 

the spider P. subpiraticus could rapidly move into rice fields 

after rice transplanting, suggesting that spiders in neighboring 

Zizania fields could efficiently promote the spider population 

in rice fields. 

With the farmers’ conventional spraying (4 times per 

season of rice), most spiders and Anagrus spp. were killed; 

however, the number of brown planthoppers was not significantly 

lower than that in the other treatments with 0–3 

sprayings. The number of leaffolders (Cnaphalocrocis 

medinalis) and stem borers (Chilo suppressalis) in 3- and 4- 

spraying treatments was significantly higher than in treatments 

with 0, 1, and 2 sprayings. More rice yield was lost in 0–2- 

spraying treatments than in treatments with 3 and 4 sprayings. 

There was no difference in rice yield between 3- and 4-spraying 

treatments (Table 2). It is suggested that 1–2 insecticide 

sprayings could be saved, especially sprayings after August. 

Discussion 

The results indicated that a high number of spiders and Anagrus 

spp. dispersed into rice fields after rice transplanting through- 


Table 2. Test on optimizing spraying of pesticides in paddy fields. a 

Application Spiders Anagrus spp. Brown Leaffolder Stem borer Yield loss 

No. of sprayings time (date) (no. per m 2 ) (no. per m 2 ) planthopper (no. per hill) (deadhead %) (%) 

(no. per m 2 ) 

No spraying 22.7 a 12.2 a 18.3 a 2.9 a 0.24 a 36.01 a 

One spraying 6 Jul 23.4 a 10.3 a 27.6 a 4.5 a 0.06 b 14.33 b 

Two sprayings 6 Jul, 15 Aug 17.5 a 4.5 b 33.1 a 4.9 a 0.03 b 9.04 b 

Three sprayings 6 Jul, 1 Aug, 15 Aug 12.3 a 4.2 b 25.3 a 0.9 b 0 b 0.51 c 

Four sprayings 20 Jun, 6 Jul, 1 Aug, 15 Aug 3.4 b 1.2 c 17.2 a 1.1 b 0 b 0 c 

a The transplanting date was 1 June. Data were analyzed with Duncan’s multiple range test between treatments. Means followed by a different letter in the same column are 

significantly different at P

Evaluating augmentative releases of the mirid bug 

Cyrtorhinus lividipennis to suppress brown planthopper 

Nilaparvata lugens in open paddy fields 

Masaya Matsumura, Satoru Urano, and Yoshito Suzuki 

Augmentative releases of natural enemies are a widely used 

method of controlling many pests, mainly in greenhouse vegetables 

(Van Driesche and Bellows 1996). In open rice fields, 

however, few attempts have been made to manage insect pests 

by releasing natural enemies. The mirid bug Cyrtorhinus 

lividipennis Reuter is the egg predator of the brown planthopper 

Nilaparvata lugens (Stål) and the whitebacked planthopper 

Sogatella furcifera (Horváth), the two major pests of rice 

throughout Asia (Chiu 1979). Although the potential growth 

rate of C. lividipennis is very high (Suzuki and Tanaka 1996, 

Matsumura and Suzuki 1999), the initial population density of 

C. lividipennis in Japan is usually low (Teramoto et al 1996). 

This is because C. lividipennis is not able to overwinter successfully 

in Japan, and colonization occurs annually following 

long-distance migration from southern China. Thus, we evaluated 

the effectiveness of augmentative releases of C. 

lividipennis to suppress N. lugens in open paddy fields. 


Mass rearing of C. lividipennis 

As no mass-rearing technique for C. lividipennis using an artificial 

diet has been established, we developed a simple method 

for mass rearing of C. lividipennis on rice seedlings using N. 

lugens eggs as food. To calculate the reproductive efficiency 

of C. lividipennis, 25 or 50 pairs of adult C. lividipennis were 

introduced into plastic cages (30 × 25 × 28 cm) containing 5– 

6-day-old rice seedlings infested with 50 gravid females of N. 

lugens. After 2 weeks, new rice seedlings were introduced into 

the plastic cages and the old rice seedlings were removed after 

all the insects had moved to the new ones. Thereafter, the rice 

seedlings were renewed at 7-day intervals. After 5 weeks, the 

number of adult offspring was counted. The experiments were 

conducted in the laboratory at a temperature of 25 ± 1 °C and 

a photoperiod of 16L:8D. 

Augmentative release experiment 

In 1999-2001, C. lividipennis adults (1-week-old adults produced 

in the laboratory) were augmentatively released at 

predator:prey (C. lividipennis adult:N. lugens adult) ratios of 

0:1 (nonrelease control), 1:1, and 1:2 (only in 1999) into replicated 

experimental blocks (7 × 7 m) in open paddy fields. 

The release was done at the start of the immigrant generation 

of N. lugens (early July). Because the density of N. lugens 

immigrants was quite low in those three years, N. lugens adults 

were also released at a rate of 0.2 individual per hill prior to 

the release of C. lividipennis. In 2001, the effect of an additional 

release of C. lividipennis at the start of the second generation 

(early August) was also evaluated. 

A routine population census for the two planthoppers 

(released N. lugens and naturally-occurring S. furcifera) and 

C. lividipennis was conducted using a FARMCOP suction sampler 

(Cariño et al 1979). A plastic cylinder (30 cm in diameter 

and 70 cm in height) was used to cover a rice hill. All arthropods 

inside the cylinder were collected with the sampler, stored in 

70% alcohol, and counted in the laboratory under a binocular 

microscope. The number of rice plants sampled was 15 per 

experimental block per sampling date. Censuses were made at 

intervals of 5–10 days from late July to late September (in 

1999) or early October (in 2000 and 2001). 

Generation boundaries and mean population density of 

the two planthoppers and C. lividipennis for each successive 

generation were calculated based on Kuno’s (1968) method. 

All analyses were conducted using a 3-year data set of mean 

population densities. The effect of the natural enemy release 

on population dynamics of N. lugens was determined by 

Yamamura’s (1999) key-factor/key-stage analysis. All statistical 

analyses were performed with the JMP version 5 Statistical 

Package. 


Mass rearing of C. lividipennis 

Five weeks after the introduction of C. lividipennis adults into 

the plastic cage, 286.8 ± 15.7 and 436.4 ± 30.3 individual adult 

offspring were obtained from 25 and 50 C. lividipennis adults, 

respectively. The population growth rate of C. lividipennis was 

5.7 and 4.4 per generation for the 25- and 50-adult treatments, 

respectively. 

The present method of mass rearing of C. lividipennis is 

high enough for stock maintenance and use in release experiments. 

For practical and commercial uses of C. lividipennis as 

a biological control agent, however, there is a need to establish 

a mass-rearing technique using an artificial diet. 

Augmentative release experiment 

The population density of the first-generation N. lugens (late 

July to early August) was significantly suppressed when C. 

lividipennis was released at predator:prey ratios of 1:1 (highratio 

release) (ANOVA, F = 8.38, P

Population density per hill log (n + 1) 

3.0 

High-ratio release 

Low-ratio release 

Nonrelease 

2.0 

1.0 

G1 

G2 

G3 

A 


Nonrelease 

G1 

G2 

G3 

B 


Low-ratio release 

× 2 releases 

Nonrelease 

G1 

G2 

G3 C 

0.0 

Rep. 1 

Rep. 1 

Rep. 1 

3.0 

2.0 

1.0 

0.0 

Rep. 2 

Jul Aug Sep Oct 

Rep. 2 


Month 

Rep. 2 


Fig. 1. Population trends of Nilaparvata lugens in each replicated block (replication 1 and 2) following augmentative releases of Cyrtorhinus 

lividipennis in open paddy fields. (A) C. lividipennis adults were released in early July at predator:prey (C. lividipennis adult:N. lugens adult) 

ratios of 0:1 (nonrelease), 1:1 (high-ratio release), and 1:2 (low-ratio release) in 1999, (B) at ratios of 0:1 and 1:1 in 2000, and (C) at ratios 

of 0:1 and 1:1, and an additional release in early August (high ratio × 2 releases) in 2001. G1, G2, and G3 indicate the first, second, and 

third generation, respectively. 

Table 1. Key-factor/key-stage table for the population density of the third generation in 

Nilaparvata lugens. 

Generation 

Factor df F P 

G1/G0 G2/G1 G3/G2 Total 

Natural enemy release 1 3.129 –0.199 0.798 3.728 11.0 0.002 

Replication 1 0.295 –0.055 –0.026 0.215 1.9 0.201 

Year 2 0.872 3.491 –0.407 3.957 17.6 0.001 

Unknown variability 7 0.445 0.339 0.229 1.014 

Total 11 4.742 3.576 0.595 8.913 

was the most significant factor suppressing the density of the 

third-generation N. lugens (Table 1). One-time release of C. 

lividipennis was insufficient to suppress the density of thirdgeneration 

N. lugens (Fig. 1C). However, an additional release 

of C. lividipennis at the second generation (early August) successfully 

regulated the mean population density of the thirdgeneration 

N. lugens to 13.6–36.1 individuals per hill (1.13– 

1.56 in log scale), which is lower than the density at which 

“hopper-burn” damage occurs (Fig. 1C). 

Stepwise multiregression analysis revealed that the most 

important factor decreasing the population growth rate from 

the initial to the first generation of N. lugens was the density 

of released C. lividipennis (P

population increase of C. lividipennis depends on S. furcifera 

early in the season and on N. lugens later in the season. 

Our field experiment suggests that the augmentative release 

of C. lividipennis is highly promising for regulating N. 

lugens density to levels low enough to prevent “hopper-burn” 

damage in open paddy fields. To further enhance the effect of 

the released natural enemy, additional techniques such as a 

banker plant system (Van Lenteren 1995) should be worthwhile 

to evaluate. For open rice fields, eggs of the delphacid 

planthopper Sogatella vibix (Haupt) (a nonpest species of rice 

plants) found on the Japanese barnyard millet Echinochloa 

utilis Ohwi et Yabuno could serve as a noncrop plant reservoir 

for food and reproduction of C. lividipennis. The evaluation 

of this system is now in progress (Matsumura M and Urano S, 

unpublished data). 

References 

Cariño FO, Kenmore PE, Dyck VA. 1979. The FARMCOP suction 

sampler for hoppers and predators in flooded rice fields. Int. 

Rice Res. Newsl. 4(5):21-22. 

Chiu SC. 1979. Biological control of the brown planthopper. In: 

Brown planthopper: threat to rice production in Asia. Los 

Baños (Philippines): International Rice Research Institute. 

p 335-355. 

Kuno E. 1968. Studies on the population dynamics of rice leafhoppers 

in a paddy field. Bull. Kyushu Agric. Exp. Stn. 14:131- 

246. 

Matsumura M, Suzuki Y. 1999. Effects of prey species and honeydew 

feeding on development and reproduction of the mirid 

bug, Cyrtorhinus lividipennis Reuter (Heteroptera: Miridae). 

Proc. Assoc. Plant Prot. Kyushu 45:63-67. 

Matsumura M, Suzuki Y. 2003. Direct and feeding-induced interactions 

between two rice planthoppers, Sogatella furcifera and 

Nilaparvata lugens: effects on dispersal capability and performance. 

Ecol. Entomol. 28:174-182. 

Suzuki Y, Tanaka K. 1996. Reproductive traits of Cyrtorhinus 

lividipennis Reuter. Proc. Assoc. Plant Prot. Kyushu 42:69- 

72. 

Teramoto T, Nakasuga T, Yokomizo K. 1996. Seasonal prevalence 

of occurrence of the brown planthopper, Nilaparvata lugens, 

and predacious mirid bug, Cyrtorhinus lividipennis, in 

Nagasaki, Japan. In: Hokyo N, Norton GA, editors. Pest management 

strategies in Asian monsoon agroecosystems. 

Kumamoto (Japan): Kyushu National Agricultural Experiment 

Station. p 55-62. 

Van Driesche RG, Bellows TS Jr. 1996. Biological control. New York 

(USA): Chapman & Hall. 539 p. 

Van Lenteren JC. 1995. Integrated pest management in protected 

crops. In: Dent D, editor. Integrated peat management. London 

(UK): Chapman & Hall. p 533-537. 

Yamamura K. 1999. Key-factor/key-stage analysis for life table data. 

Ecology 80:533-537. 

Notes 

Authors’ addresses: Masaya Matsumura and Satoru Urano, National 

Agricultural Research Center for Kyushu Okinawa Region, 

Nishigoshi, Kumamoto 861-1102, Japan; Yoshito Suzuki, 

National Agricultural Research Center, Tsukuba, Ibaraki 3005- 

8666, Japan. 

Developing a rice production system 

for sustainable pest management 

T.W. Mew 

The development and cultivation of shorter-duration and agronomically 

well-adapted varieties has fed half of the world’s 

total population of about six billion people, mostly in Asia. 

This not only allowed intensification of the rice crop in time 

and space, which reduced the demand for land to grow it to 

meet this population pressure, but it also accounts for 30% to 

50% of agricultural production and 50–90% of the calories 

consumed by these people (Hossain and Fischer 1995). While 

there is a need to achieve greater rice productivity, intensification 

also makes the rice crop more vulnerable to attack by rice 

pests and diseases (Mew et al 2004). 

Pests and diseases are a moving target. Genetic elasticity 

allows them to readjust their population to adapt to new 

environments and on new resistant varieties bred by plant 

breeders. It is not uncommon that the scenario of pest and disease 

profiles shifts as production systems change. Accompa- 

nying this intensification is the use of genetically uniform varieties 

that reduce the buffering capacity in the rice-cropping 

system. This situation creates some concerns regarding modern 

rice production in relation to pest outbreaks. Below, we 

outline some of these concerns and propose that scientists 

working on pest management should consider sustainable rice 

production systems as a means to manage pests and diseases 

as an important option for another doubly green revolution in 

agriculture. 

The first concern is the deployment of a few high-yielding 

varieties over large areas, leading to a decline in cultivar 

diversity. The general landscape of rice production in intensive 

rice ecosystems is characterized as a monoculture system. 

A reduction in cultivar diversity is a concern of rice intensification 

because maintaining adequate diversity and resilience 

is important in the humid tropics of Asia, where pressure 


of diseases and pests is unusually high. Current disease and 

pest management systems rely primarily on new resistant varieties 

and/or the application of pesticides. Dependence on varietal 

resistance often results in a “boom-and-bust” cycle because 

of pathogen adaptation on the widely grown few resistant 

varieties. The continuous cultivation of these modern varieties 

also replaces some of the popular traditional cultivars. 

While this is a true concern, still, the reality is to achieve the 

current level of rice production. Thus, it is not feasible and 

practical to plant low-yielding but diverse traditional varieties. 

The challenge that rice scientists, especially those involved 

in pest management, face is how to increase the rice 

supply or retain the high level of productivity of modern cultivars 

to feed a growing population with reduced arable land 

area and increased scarcity of natural resources on the one 

hand, and maintain cultivar diversity so as to minimize losses 

from disease and pest outbreaks on the other. In rice production, 

however, host-plant resistance is a key to sustaining rice 

productivity in the long term. To sustain the useful life of the 

resistance, resistance deployment as part of gene management 

should be considered an important option (Wolfe 1985). Although 

resistance deployment has its own limitations, it is an 

attractive option to reduce pest outbreaks and disease epidemics 

frequently caused by blast, tungro, and other diseases in 

the tropics. Genetic diversity would provide complementary 

disease resistance and other positive agronomic functions, such 

as resistance to lodging, good grain quality, and others. It is 

also efficient in R-gene use and management. It would reduce 

the selection pressure on pathogens and slow down pathogen 

evolution. The deployment of modern high-yielding and traditional 

high-quality varieties together, that is, if the match is 

feasible, would increase the productivity of rice land and farmers’ 

income, reduce pesticide use, and thus protect rice production 

environments. Such a system would also harmonize 

the production of modern and traditional varieties, and offer a 

means of in situ landrace conservation. This has been demonstrated 

in Yunnan, China, using interplanting of modern highyielding 

hybrids with traditional glutinous rice with high grain 

quality (Zhu et al 2000). 

The second concern is the increase in input use to match 

the intensification. Farmers use more fertilizer and also more 

pesticides. The consequence of this practice is related to environmental 

deterioration and an increase in production cost. 

This situation suggests that rice production is subject to increasing 

environmental constraints. Many issues are related to 

natural resource management leading to potential sustainable 

crop production practices. In tropical Asia, however, the health 

base of the poor-quality farmer-saved seed for planting has 

been neglected despite the fact that seed base technology is 

the key to increasing crop production. Recent surveys indicated 

that there have been extensive losses of rice yield and 

grain quality because of the low-quality seed used for planting 

by farmers. 

Farmers in general lack the knowledge and awareness 

of planting high-quality seed for crop production, which could 

be an important means of pest management for sustainable rice 

production. The widely popularized integrated pest management 

for years has appeared to focus on the crop in fields. 

Nowhere else in the practice of farmers’ seed health management 

has it been taken into account. Yet, poor-quality seed for 

planting is responsible for at least a 10% crop loss based on 

the work done in Bangladesh in a project on “Seed Health 

Improvement” supported by DFID (Department for International 

Development, UK) (T.W. Mew, Taher Mia, and Mahabub 

Hossain, unpublished results). The research suggested that 

farmers could gain 10% in crop yield if high-quality seed were 

used for planting (Mew and Cottyn 2001). We need to realize 

that poor-quality seed for planting also reduces the genetic 

potential of modern rice varieties. Planting high-quality seed 

is also a means of disease management because seed carries 

more than 80% of the rice pathogens. In any given seed lot, 

rice seeds carry several pathogens in various degrees of intensity. 

It is difficult to demonstrate the impact on crop production 

by managing individual pathogens carried by seed. By 

applying the seed health concept, that is, if healthy seed is 

used for planting to manage seedborne pathogens, the diseases 

derived from the seedborne inoculum could be greatly minimized. 

Therefore, this is a system of sustainable crop production 

and pest management. It reduces the production cost by 

lowering the seeding rate and cost of weed and insect pest 

management in the field. 

The third concern relates to soil health, an area that has 

been widely studied. In this paper, we shall not discuss issues 

of soil health. It is likely that, if we achieve the above, we shall 

achieve health of the ecosystem. A rice production system that 

takes into account genetic or cultivar diversity and seed health 

for planting, then coupled with soil health, would sustain rice 

production and pest management. 

References 

Hossain M, Fischer K. 1995. Rice research for food security and 

sustainable agricultural development in Asia: achievements 

and future challenges. GeoJournal 35:286-298. 

Mew TW, Leung H, Savary S, Vera Cruz CM, Leach JE. 2004. Looking 

ahead in rice disease research and management. Crit. Rev. 

Plant Sci. 23(2):103-127. 

Wolfe MS. 1985. The current status and prospects of multiline cultivars 

and variety mixtures for disease resistance. Annu. Rev. 

Phytopathol. 23:251-273. 

Zhu Y, Chen H, Fan J, Wang Y, Li Y, Chen J, Fan JX, Yang S, Hu L, 

Leung H, Mew TW, Teng PS, Wang Z, Mundt CC. 2000. Genetic 

diversity and disease control in rice. Nature 406:718- 

722. 

Notes 

Author’s address: IRRI, DAPO Box 7777, Metro Manila, Philippines, 

e-mail: t.mew@cgiar.org. 


Evaluation of a leaf blast simulation model (BLASTMUL) 

for rice multilines in different locations and cultivars, 

and effective blast control using the model 

T. Ashizawa, M. Sasahara, A. Ohba, T. Hori, K. Ishikawa, Y. Sasaki, T. Kuroda, R. Harasawa, K.S. Zenbayashi, and S. Koizumi 

Blast, caused by Pyricularia grisea (Cooke) Sacc., is one of 

the most devastating diseases of rice in Japan. However, levels 

of field resistance to the disease in most Japanese rice cultivars 

are relatively low because the cultivars have been developed 

to mainly improve eating quality in compensation for 

loss of the resistance. For this reason, blasticides have been 

used frequently to control the disease, but Japanese farmers 

meet the demand for fewer applications of chemicals because 

of food safety and environmental consciousness. To solve this 

dilemma, multilines, which are mixtures of near-isogenic lines 

(NILs) with different complete resistance to blast, have been 

released in Japan. Sasanishiki multiline (named Sasanishiki 

BL, currently mixed with Pik, Pik-m, Piz, and Piz-t lines in the 

proportion of 1:1:4:4) has been cultivated in Miyagi Prefecture 

since 1995. In Toyama Prefecture, Koshihikari multiline 

(Koshihikari Toyama BL), consisting of Piz-t, Pib, and Pik-p 

lines in the proportion of 4:4:2, has been cultivated since 2002. 

Moreover, Niigata Prefecture is planning to cultivate another 

Koshihikari multiline (Koshihikari Niigata BL) on about 90,000 

ha in the prefecture beginning in 2005. For effective blast control 

with multilines, we need to estimate the extent of blast 

suppression using them. However, it is difficult to overcome 

the limitations of field trials. The simulation approach can provide 

an opportunity to clarify the extent of disease suppression 

using multilines. 

A simulator of plant disease foci and epidemics in cultivar 

or NIL mixtures, EPIMUL, was developed by Kampmeijer 

and Zadoks (1977). EPIMUL can simulate spatial disease 

progress in mixtures, and was used for forecasting rust epidemics 

in mixtures at early stages of disease epidemics. However, 

no simulation model has been developed for rice 

multilines. Among simulation models for rice blast epidemics, 

BLASTL developed by Hashimoto et al (1986) is one of the 

most reliable simulators. Hence, we modified BLASTL and 

developed a multiline simulation model called BLASTMUL. 

BLASTMUL was developed based on the results of field trials 

using rice cultivar Sasanishiki and its NILs at the National 

Agricultural Research Center for Tohoku Region (NARCT) 

from 1998 to 2001. However, its usefulness has not been validated 

yet in different locations and multilines. 

The objective of this study is to evaluate the applicability 

of the model in different locations and multilines for effective 

control of leaf blast using the multiline approach. 

The model 

BLASTMUL is developed by modifying BLASTL. In 

BLASTMUL, a new parameter of spore dispersion and deposition 

between two hills of NILs is set (Ashizawa et al 2001). 

Our model calculates the number of lesions per hill in a mixture 

of susceptible and resistant NILs in a given proportion 

under various weather conditions. 

Field trials in Miyagi Prefecture 

To evaluate the appropriateness of BLASTMUL in different 

locations, we conducted paddy field trials at the Furukawa Agricultural 

Experiment Station in Miyagi Prefecture from 2000 

to 2001. The seeds of Sasanishiki and Sasanishiki BL No. 4 

were mixed uniformly in the proportions of 1:0 and 1:1. The 

mixed seeds were sown in upland nurseries and then transplanted 

by machine in 10 × 10-m paddy-field plots at a distance 

of 30 × 15 cm on 19 May 2000 and 15 May 2001. These 

experiments were arranged in a randomized block design with 

two replications. As inoculum sources, diseased seedlings of 

Sasanishiki inoculated with the blast isolate Ina86-137 (Japanese 

race 007.0, virulent to Sasanishiki, avirulent to Sasanishiki 

BL No. 4) were placed in each plot on 23 June 2000 and 20 

June 2001. Lesions more than 5 mm in size per hill were 

counted at 6- to 7-day intervals from the start till the end of 

leaf blast epidemics. Meteorological data were recorded beside 

the fields using a weather sensor system. 

Field trials in Niigata Prefecture 

We also conducted paddy-field trials under natural blast epidemic 

conditions at Ojiya in Niigata Prefecture from 2002 to 

2003. Mixed seeds of rice cultivar Koshihikari and its Pita 

line, whose mixture proportions were 1:0, 1:1, and 1:3, were 

sown and then four seedlings per hill were transplanted by 

hand at a distance of 30 × 18 cm in 4.2 × 5.3-m paddy-field 

plots on 20 May 2002 and 22 May 2003. These trials were 

arranged in a randomized block design with three replications. 

Koshihikari and the Pita line are susceptible and resistant, respectively, 

to the predominant Japanese races distributed in 

the fields. The disease assessment was as described earlier. 

Meteorological data were recorded using a dew sensor and a 

weather sensor system (AMeDAS) located at Tokamachi, close 

to Ojiya. 


Number of lesions per hill 

35 

30 

25 

20 

Sim. S:R (1:0) in 2000 

Sim. S:R (1:1) in 2000 

Obs. S:R (1:0) in 2000 

Obs. S:R (1:1) in 2000 

Sim. S:R (1:0) in 2001 

Sim. S:R (1:1) in 2001 

Obs. S:R (1:0) in 2001 

Obs. S:R (1:1) in 2001 

Number of lesions per hill 

30 

25 

20 

15 

Sim. S:R (1:0) in 2002 

Sim. S:R (1:1) in 2002 

Sim. S:R (1:3) in 2002 

Obs. S:R (1:0) in 2002 

Obs. S:R (1:1) in 2002 

Obs. S:R (1:3) in 2002 

A 

15 

10 

10 

5 

5 

0 

0 

25 30 35 40 45 50 55 

Days from 1 June 

Fig. 1. Observed (Obs.) and simulated (Sim.) development of rice 

leaf blast in a mixture of susceptible (S) cv. Sasanishiki and its 

resistant (R) near-isogenic line, Sasanishiki BL No. 4, at proportions 

of 1:0 and 1:1 from 2000 to 2001. The field experiments 

were performed at the Furukawa Agricultural Experiment Station 

in Miyagi Prefecture. 

10 

8 

6 

Sim. S:R (1:0) in 2003 

Sim. S:R (1:1) in 2003 

Sim. S:R (1:3) in 2003 

Obs. S:R (1:0) in 2003 

Obs. S:R (1:1) in 2003 

Obs. S:R (1:3) in 2003 

Results 

4 

In Miyagi Prefecture, leaf blast lesions were first observed on 

11 July 2000 and 7 July 2001. The numbers of lesions per hill 

on 27 July 2000 and 23 July 2001 were 0.29 and 32 in the pure 

stands of Sasanishiki and 0.02 and 13 in the 1:1 mixture of 

Sasanishiki and Sasanishiki BL No. 4, respectively. In contrast, 

simulated numbers of lesions with BLASTMUL on 27 

July 2000 and 23 July 2001 were 0.4 and 32 in the pure stands 

of Sasanishiki and 0.05 and 7 in the 1:1 mixture of Sasanishiki 

and Sasanishiki BL No. 4 (Fig. 1). 

In Niigata Prefecture, leaf blast lesions were first observed 

on 3 July 2002 and 3 July 2003. The numbers of lesions 

on 24 July 2002 and 24 July 2003 were 27 and 8 in the 

pure stands of Koshihikari, 6 and 3 in the 1:1 mixture of 

Koshihikari and its Pita line, and 2 and 0.6 in the 1:3 mixture, 

respectively. Meanwhile, the simulated numbers of lesions on 

24 July 2002 and 24 July 2003 were 24 and 8 in the pure stands 

of Koshihikari, 5 and 2 in the 1:1 mixture of Koshihikari and 

its Pita line, and 0.8 and 0.6 in the 1:3 mixture of them (Figs. 

2A and B). 

Discussion 

Visual comparisons between the observed and simulated values 

were made subjectively in this study. Although the numbers 

of observed lesions in the mixture trials in Miyagi Prefecture 

were slightly larger than those of the ones observed in 

2000, the simulated numbers of lesions were similar to the 

2 

0 

25 30 35 40 45 50 

Days from 1 June 

Fig. 2. Observed (Obs.) and simulated (Sim.) development of rice 

leaf blast in mixtures of susceptible (S) cv. Koshihikari and its 

resistant (R) near-isogenic line, the Pita line, at proportions of 

1:0, 1:1, and 1:3 from 2002 (A) to 2003 (B). The field experiments 

were performed at Ojiya in Niigata Prefecture. 

observed ones. In 2001, the simulated numbers of lesions were 

slightly smaller than those observed in the 1:1 mixture. However, 

the simulation results in the 1:0 mixture were as expected 

for the year. Although the simulated numbers of lesions in our 

trial at Niigata were slightly smaller than the data observed in 

the 1:0 mixture, the simulation results in the 1:1 and 1:3 mixtures 

agreed well with the observed ones in 2002. These results 

suggest that the model is appropriate for evaluating rice 

mixtures for blast control in different locations and cultivars. 

Recently, we modified the model again to simulate leaf 

blast lesion development in a mixture of four NILs for four 

blast races. The modified BLASTMUL will contribute to evaluating 

the efficacy of the increased number of mixtures under 

various conditions, including an increase in disease pressure. 

In addition, the weather data have been recorded at NARCT 


for four years, in Miyagi Prefecture for two years, and in Niigata 

Prefecture for two years. These data will also contribute to 

developing a more accurate model. 

The 11 genes with complete resistance to blast (Pik-s, 

Pia, Pii, Pik, Pik-m, Piz, Pita, Pita-2, Piz-t, Pik-p, and Pib) 

mainly exist in Japanese rice cultivars. However, most of the 

Japanese cultivars possess Pik-s, Pia, or Pii, which are ineffective 

against the predominant Japanese blast races 001.0, 

003.0, and 007.0. For this reason, there are now eight complete 

resistance genes available for blast control. 

We are currently clarifying the effect of increasing the 

number of NILs in multilines on the efficacy of rice blast control 

using BLASTMUL. However, as noted earlier, available 

complete resistance genes are limited. Hence, we have to incorporate 

new resistance genes from foreign countries into the 

NILs for blast control using the multiline approach. 

References 

Ashizawa T, Zenbayashi K, Koizumi S. 2001. Development of a 

simulation model for forecasting rice leaf blast epidemics in 

multilines Jpn. J. Phytopathol. 67:194. (Abstr., in Japanese.) 

Hashimoto A, Hirano K, Matsumoto K. 1984. Studies on the forecasting 

of rice leaf blast development by application of the 

computer simulation SP. Bull. Fukushima Pref. Exp. Stn. 2:1- 

104. (In Japanese, English summary.) 

Kampmeijer P, Zadoks JC. 1977. EPIMUL, a simulator of foci and 

epidemics in mixtures of resistant and susceptible plants, 

mosaics, and multilines. Wageningen (Netherlands): Centre 

for Agricultural Publishing and Documentation. 50 p. 

Notes 

Authors’ addresses: T. Ashizawa, K.S. Zenbayashi, and S. Koizumi, 

National Agricultural Research Center for Tohoku Region, 3 

Yotsuya, Omagari, Akita 014-0102, Japan; M. Sasahara and 

A. Ohba, Miyagi Prefectural Furukawa Agricultural Experiment 

Station, 88 Fukoku, Osaki, Furukawa, Miyagi 989-6227, 

Japan; T. Hori, K. Ishikawa, Y. Sasaki, T. Kuroda, and R. 

Harasawa, Niigata Agricultural Research Institute, 857 

Nagakura, Nagaoka, Niigata 940-0826, Japan, e-mail: 

toketa@affrc.go.jp. 

Association of candidate defense genes 

with quantitative resistance to rice blast 

and in silico analysis of their characteristics 

G. Carrillo, J. Wu, B. Liu, N. Sugiyama, I. Oña, M. Variar, B. Courtois, J.E. Leach, P.H. Goodwin, H. Leung, and C.M. Vera Cruz 

Many traits of significance in agriculture exhibit quantitative 

inheritance. To characterize these quantitative traits, the candidate 

gene approach has been applied in several cereal crops 

such as maize, wheat, and rice because of the detection of 

multiple genes defining a complex trait, their partial effects on 

phenotypic variation, and their less precise localization on 

genetic maps in comparison with traits conferring qualitative 

inheritance (Pflieger et al 2001, Wilson et al 2004). 

The availability of the whole rice genome sequence from 

public and private sequencing efforts has provided opportunities 

to predict putative functions of a gene based on sequence 

information, thus allowing the identification of candidate genes. 

Candidate genes are DNA sequences similar to known genes 

or conserved motifs that make it possible to infer their biological 

functions. Through their association with disease resistance, 

they become candidate defense-response (DR) genes 

for conferring particular phenotypes. Using the candidate gene 

approach, we studied the associations between candidate gene 

markers with disease and insect resistance in a doubled-haploid 

rice mapping population (Ramalingam et al 2002). This 

frame map has been a useful reference for selecting candidate 

genes involved in both pathogen recognition and general plant 

defense for analysis of mapping populations for improving 

resistance to rice blast. 

With the sequencing of many genomes in recent years 

along with the genetic and bioinformatic resources for rice, 

our ability to comprehensively characterize multigene families 

has also been greatly enhanced. These resources include 

the physical, linkage, and expressed sequence tag (EST) maps, 

a collection of >130,000 ESTs, gene prediction algorithms and 

automatic genome annotation systems, and online resources 

for comparative grass genomics. Through these resources, we 

examined DR genes retrieved from the Rice Genome Program 

database using in silico analysis. 

Identifying candidate genes by association 

with blast quantitative resistance 

A japonica cultivar, Moroberekan from Africa, known to exhibit 

durable resistance to rice blast in Asia was crossed to a 

popular upland cultivar, Vandana, following the advanced backcross 

QTL approach. This approach allows QTL analysis of 

the mapping population while simultaneously developing elite 

germplasm for commercial use (Tanksley and Nelson 1996). 

Eighty BC 3 F 3 and BC 3 F 4 lines derived from advanced back- 


cross populations of Vandana/Moroberekan were analyzed for 

seedling (BC 3 F 3 ) and neck blast resistance in the greenhouse 

and blast nursery at IRRI and screening sites in India. A collection 

of 25 microsatellite markers, resistance gene analog 

(RGA) primers, and 51 candidate resistance genes were used 

for genotypic analysis of the lines. Six candidate gene markers, 

NBS-LRR (Pic19), Rp1 (a maize rust resistance gene), 

NBS (7-4F), PR5 (thaumatin), RGA 1-8 (resistance gene analog, 

LRR), and RGA6-7 (resistance gene analog, Kinase), and 

one simple sequence repeat (SSR) marker (RM21) were significantly 

associated with quantitative blast resistance in rice 

(P = 0.01). When tested against a single isolate, PO6-6, four 

candidate genes (oxalate oxidase, 14-3-3, RGA8-4, and RGA1- 

10) and four SSR markers (RM21, RM168, RM215, and 

RM250) were associated with resistance to this single isolate. 

Of these markers, two were associated with resistance involved 

in a reduction in diseased leaf area (DLA), five with reduced 

lesion number, and one with reduced lesion size. These markers 

accounted for observed phenotypic variation ranging from 

9.1% to 28.7%. Seedling reactions in the blast nursery and 

greenhouse tests in the Philippines were highly correlated. Field 

tests of BC 3 F 3 and BC 3 F 4 lines in India across screening sites 

identified six lines showing < 5% DLA and 13 lines showing 

quantitative resistance; however, only two (V4M-6-1-B and 

V4M-5-3-B) of these quantitatively resistant lines have good 

agronomic acceptability with a varying level of neck blast resistance. 

Other than Vandana/Moroberekan mapping populations, 

others such as TXZ/SHZ2 (Liu et al 2004), Oryzica Llanos 

5/Way Rarem (M. Bustamam, pers. comm.), and 

Ilpumbyeo/Moroberekan (S.S. Han, pers. comm.) are used to 

determine the association of candidate defense genes with quantitative 

resistance to blast. 

Accumulating different mechanisms 

of quantitative blast resistance 

To accumulate different mechanisms involved in quantitative 

resistance to blast, 15 BC 3 F 3 or BC 3 F 4 parental lines derived 

from Vandana/Moroberekan showing partial resistance and 

carrying positive candidate alleles were selected and crossed 

in all pairwise combinations. Cluster analysis of DNA profiles 

showed that the BC 3 population was genetically similar (>85%) 

to the recurrent parent Vandana (Wu et al 2004). Seeds of 10 

F 2 families were selected based on partial resistance of the 

BC 3 F 3 or BC 3 F 4 parental lines to seedling blast and neck blast, 

association of the parental lines with positive alleles, and phenotypic 

similarity to Vandana. Among the 10 F 2 families, 14 

F 2 plants showed a good to moderate level of agronomic acceptability 

and high level of seedling blast resistance (0.75– 

3.2% DLA). Resistant lines from selected families were advanced 

to F 3 until F 5 . At F 4 , the top 10% of the lines (60 out of 

>600 lines) derived from the progenies of VM5/VM14, VM6/ 

VM14, and VM82/VM14 had acceptable agronomic traits, and 

served as a basis for further selection and advancement to F 5 . 

The field performance of these advanced BC lines indicated 

that the major QTLs have been captured in the BC lines. In 

this study, only a small proportion of alleles were from 

Moroberekan. It is possible that the loci contributing to disease 

resistance may not be polymorphic, thus precluding their 

use as informative markers. 

To identify informative markers for analysis of F 4 

intermated lines of BC 3 F 5 Vandana/Moroberekan populations, 

sequence tagged site (STS) markers derived from rice candidate 

gene sequences and SSR markers located in the region of 

each candidate gene were identified using TIGR and Gramene 

databases (www.gramene.org). Thirty SSRs gave polymorphic 

fragments between Vandana and Moroberekan. Of 11 DR genes 

(oxalate oxidase, oxalate oxidase-like proteins, aspartyl protease, 

14-3-3, PR-1, probenazole-induced protein, peroxidases, 

HSP90, thaumatin-like proteins, adenosylhomocysteinase, and 

aldose reductase) analyzed, polymorphic markers were obtained 

for oxalate oxidases, PR-1, probenazole-induced protein, 

and HSP90. Using 30 random SSRs dispersed in the rice 

genome and polymorphic candidate gene-associated markers 

in Vandana and Moroberekan, an initial set of F 4 resistant elite 

lines (7) in Cavinti, Philippines, was scanned for introgression 

of Moroberekan regions into Vandana and co-localization of 

the consensus candidate defense genes (Fig. 1A and B). Interestingly, 

primers derived from 1,000 bp upstream of two of 

the four oxalate oxidases in chromosome 3, SSRs for aspartyl 

protease in chromosome 7, and oxalate oxidase-like proteins 

in chromosome 8 co-localized with particular candidate genes 

in target chromosomal regions (Fig. 1B). However, only two 

lines, IR78221-19-6-56 and IR78222-20-7-148, which co-localized 

with oxalate oxidases in chromosome 3 and oxalate 

oxidase-like proteins in chromosome 8, were resistant to seedling 

and neck blast in Almora, a blast hotspot in the Indian 

Himalaya. Based on these results, we hypothesized that (1) 

highly resistant lines IR78221-19-6-56 and IR78222-20-7-148 

may carry combinations of several effective candidate genes 

under high disease pressure in Almora, and (2) the contribution 

of individual candidate genes conferring different mechanisms 

of quantitative resistance to blast has been accumulated 

in these lines. We are currently designing PCR primers for 

other consensus DR genes identified from multiple mapping, 

linkage, and microarray experiments to determine the combination 

of effective candidate defense genes in 27 intermated 

F 5 lines resistant in Almora. 

In silico analysis of selected candidate defense genes 

Initially, we focused on consensus DR genes for in silico analysis. 

Sequences were retrieved from the Rice Genome Program 

database (www.tigr.org/tdb/e2k1/osa1/). For genes occurring 

in gene families such as the germin-like proteins (GLP), we 

derived phylogenetic trees using the retrieved sequences. We 

also checked for conserved promoter motifs and identified ciselements 

in the 1,000-bp upstream regions using MEME (http:/ 

/meme.sdsc.edu/) and PLACE (Plant Cis-Acting Elements, 

www.dna.affrc.go.jp/PLACE/). 

Scanning the rice genome identified more than 40 GLPs, 

four of which are putative oxalate oxidases in chromosome 3. 


A 

0 

25 

50 

75 

100 

rm61a 9.3 

rm569 11.0 

rm545 24.7 

rm157a 44.1 

rm7 64.0 

Putative 

oxalate 

oxidases 

0 

Eukaryotic 

aspartyl 

25 protease 

rm2 36.1 

rm214 49.7 

50 

rm336 61.0 

rm10 63.5 

rm6011 73.2 

75 

rm1279 75.6 

0 

rm38 16.4 

25 

rm310 38.5 

rm6208 42.9 

50 

rm3215 49.5 

rm547 58.1 

rm137 69.0 

75 

Oxalate 

oxidase-like 

proteins 

125 

0 

rm473d 113.2 

rm426 122.3 

rm168 122.8 

rm186 127.4 

rm411 127.9 

rm148 140.7 

1 3 5 

2 4 6 

7 

8 

100 

9 

rm5720 115.2.6 

rm248 118.0 

1 3 5 

2 4 6 

7 9 

8 

100 rm80 103.7 

rm281 119.9 

1 3 5 

2 4 6 

7 9 

8 

Chromosome 8 

Chromosome 8 

Chromosome 8 

B 

100 

rm473 113.2 

rm426 122.3 

125rm168 122.8 

rm186 127.4 

rm411 127.9 

rm148 140.7 

1 2 3 4 5 6 7 8 9 

rm426 

osglp10up 

osglp10 

osglp11up 

osglp11 

osglp12 

osglp13up 

osglp13 

rm168 

1 2 3 4 5 6 7 8 9 

Fig. 1. (A) Genome scan of Vandana/Moroberekan intercross progenies using SSRs, 1: Moroberekan; 2: Vandana; 

3–7: more resistant lines; 8–9: less resistant lines. Examples of candidate genes associated with each of the 

critical areas marked: putative oxalate oxidases in chr. 3, putative aspartyl protease in chr. 7, and oxalate oxidaselike 

proteins in chr. 8. Heterozygous loci are colored green. Because of selection in BC 3 F 3 prior to intermating, our 

susceptible lines in fact carry some degree of quantitative resistance. Therefore, Vandana is still the susceptible 

control in this analysis. (B) Detailed marker analysis of chr. 3 loci identified with 4 putative oxalate oxidases on 1: 

Moroberekan; 2: Vandana; 3–7: more resistant lines; 8–9: less resistant lines. PCR primers used were designed 

from the coding region of each gene. Primers were also designed in the 1,000-bp upstream region of each gene 

(OsGLP10UP, OsGLP11UP, and OsGLP13UP). Heterozygous loci are colored green. Monomorphic markers are colored 

light blue. 

These oxalate oxidase gene sequences are located next to each 

other, with similarities ranging from 90% to 98%. For each 

gene, there was variation in the copy number of cis-elements 

related to biotic stress responses, such as W box, WNPR1, 

and WRKY, indicating that these genes have potential associations 

with the response of rice to pathogen infection, such 

as the blast fungus. 

Phylogenetic analyses clustered chromosome 3 oxalate 

oxidases with barley oxalate oxidase Y14203. Several GLP 

sequences in chromosome 8 were similar in conserved domain 

structure to barley oxalate oxidase-like protein X93171. These 

genes are thought to play a role in defense response by degradation 

of oxalate produced by fungi, the production of active 

oxygen species for oxidative cross-linking of cell wall components, 

and stress signaling by increasing extracellular 

[Ca + ion] (Dunwell et al 2000). 

Another class of DR genes is the thaumatin-like proteins 

or PR5 proteins. At IRRI, overexpression of PR5 delays the 

onset of disease symptoms of sheath blight in transgenic rice. 

A scan of the rice genome revealed at least 20 thaumatin-like 

proteins dispersed in all chromosomes except chromosome 5. 

Of these, thaumatin-like sequences in chromosomes 6 and 7 

have been mapped to IR64/Azucena doubled-haploid lines 

(Ramalingam et al 2003). Analysis of the 1,000-bp upstream 

of each gene reveals that the number of copies of cis-elements 

such as WRKY, WBOX, and GCC core varies from one gene 

to the next. PR5 proteins in barley were found to bind to 

β-1,3-glucans from yeasts and fungi (Trudel et al 1998). In 

Vandana/Moroberekan, thaumatin was significantly associated 

with partial blast resistance (Wu et al 2004). 


The availability of rice genetic and bioinformatic resources 

has clearly overcome the limitations for relating QTLs to candidate 

genes and metabolic pathways. To improve efficiency 

in the selection and pyramiding of desirable plant traits, markerassisted 

selection may now allow candidate defense genes with 


quantitative effects to be combined efficiently. The level of 

quantitative resistance can be estimated based on the presence 

of markers associated with candidate defense genes or QTLs 

in a plant/line. To contribute to the practicality of this approach, 

we continue to investigate the genetic behavior of the QTLs in 

different genetic backgrounds and find consensus candidate 

genes. We also continue to develop markers derived from consensus 

candidate defense genes conferring partial resistance 

instead of just close linkage to resistance genes, and develop 

efficient and economical breeding procedures using molecular 

genotyping. Our research has therefore focused on these 

areas aimed at facilitating the selection of cultivars with broadspectrum 

resistance against blast pathogen populations in diverse 

rice-growing environments. 

References 

Dunwell JM, Khuri S, Gane PJ. 2000. Microbial relatives of the 

seed storage proteins of higher plants: conservation of structure 

and diversification of function during evolution of the 

cupin superfamily. Microbiol. Mol. Biol. Rev. 64:153-179. 

Liu B, Zhang S, Zhu X, Yang Q, Wu S, Mei M, Mauleon R, Leach J, 

Mew T, Leung H. 2004. Candidate defense genes as predictors 

of quantitative blast resistance in rice. Mol. Plant-Microbe 

Interact. 17:1146-1152. 

Pflieger S, Lefebvre V, Causse M. 2001. The candidate gene approach 

in plant genetics: a review. Mol. Breed. 7:275-291. 

Ramalingam J, Vera Cruz CM, Kukrejah K, Chittoor JM, Wu J, Lee 

SW, Baraoidan M, George ML, Cohen MB, Hulbert S, Leach 

JE, Leung H. 2002. Candidate resistance genes from rice, 

barley, and maize and their association with qualitative and 

quantitative resistance in rice. Mol. Plant-Microbe Interact. 

16(1):14-24. 

Tanksley SD, Nelson RC. 1996. Advanced backcross QTL analysis: 

a method for the simultaneous discovery and transfer of valuable 

QTLs from unadapted germplasm into elite breeding lines. 


Trudel J, Grenier J, Potvin C, Asselin A. 1998. Several thaumatinlike 

proteins bind to -1,3-glucans. Plant Physiol. 118:1431- 

1438. 

Wilson LM, Whitt SR, Ibañez AM, Rocheford TR, Goodman MM, 

Buckler ES IV. 2004. Dissection of maize kernel composition 

and starch production by candidate gene association. Plant 

Cell Preview, www.aspb.org. 

Wu JL, Sinha PK, Variar M, Zheng KL, Leach JE, Courtois B, Leung 

H. 2004. Association between molecular markers and blast 

resistance in an advanced backcross population of rice. Theor. 

Appl. Genet. 108:1024-1032. 

Notes 

Authors’ addresses: G. Carrillo, J. Wu, B. Liu, N. Sugiyama, I. Oña, 

H. Leung, and C.M. Vera Cruz, IRRI, DAPO Box 7777, Metro 

Manila, Philippines; M. Variar, Central Rainfed Upland Rice 

Research Station, Hazaribag 825 301, Jharkand, India; B. 

Courtois, CIRAD-CA, Avenue Agropolis, Montpellier Cedex 

5, France; J.E. Leach, Colorado State University, Fort Collins, 

Colo. 80523-1177, USA; P.H. Goodwin, University of Guelph, 

Guelph N1G 2W1, Canada. 

Transferring resistance genes 

among different cereal species 

Bingyu Zhao, Shavannor Smith, Jan Leach, and Scot Hulbert 

The transfer of genes into cereal crops from related grasses 

has been an important tool for improving the disease and pest 

resistance of some cereals. The approach to date has been limited 

to cereals and their close relatives because conventional 

genetic approaches have been used. If resistance genes could 

be effectively transferred among distantly related grasses in 

different tribes (e.g., rice, maize, and wheat), this could have a 

large effect on resistance breeding. Of particular relevance is 

the observation that a given cereal species is typically resistant 

to most of the pathogens that cause disease in other cereal 

species. If the basis of some of these nonhost resistances is 

due to the same types of genes that cause host resistance, we 

may be able to identify and transfer these genes to transfer the 

resistance. The contribution of typical host resistance genes to 

nonhost resistance is not clear, but similarities between the 

two types of resistance (Thordal-Christensen 2003) suggest 

that R genes probably play a role. One limitation of effective 

transfer of resistance genes is the fact that many R genes do 

not function in the heterologous genetic backgrounds of distantly 

related species (Hulbert et al 2001). 

Transfer of a maize resistance gene to wheat and barley 

The first transfer of a disease resistance gene among distantly 

related cereals was the maize Rp1-D gene, which was transferred 

to wheat and barley (Ayliffe et al 2004). Rp1-D confers 

resistance to common rust (Puccinia sorghi) in maize, but no 

rust resistance was observed in wheat or barley lines expressing 

the gene after inoculation with rust isolates from three different 

rust species. Although the Rp1-D transformation con- 


struct worked well in maize, most of the transcript appeared 

truncated in the other cereals. It is therefore possible that the 

lack of function was due to improper expression of the gene. It 

is also possible that the lack of phenotype was due to the absence 

of the corresponding Avr gene in the other Puccinia species 

that attack these cereals. Experiments with other Rp1 genes, 

however, also failed to associate phenotypes with these genes. 

These experiments included one variant Rp1 gene that conditioned 

cell death spontaneously, without pathogen inoculation 

(Smith and Hulbert, unpublished). Thus, it seems that the Rp1 

genes will not function in these other cereals, either due to 

improper expression or the lack of a functional signal transduction 

pathway. 

A maize gene controls resistance to a rice pathogen 

It is possible to assay resistance reactions to nonhost bacterial 

species by infiltrating leaves of nonhost species and assessing 

whether a hypersensitive reaction (HR) occurs. A gene controlling 

a hypersensitive reaction to Xanthomonas oryzae pv. 

oryzicola, the rice bacterial streak pathogen, was identified in 

this manner in maize (Zhao et al 2004a). Although infiltration 

of the bacterium into maize leaves does not cause disease regardless 

of whether they carry the Rxo1 gene, only maize lines 

that carry the gene show the strong HR response. A molecular 

genetic analysis of the pathogen verified that the HR is under 

control of a classic gene-for-gene interaction. All strains of X. 

o. pv. oryzicola tested elicited the resistance reaction in maize 

lines carrying Rxo1, and a single gene was isolated from the 

bacterium that was responsible for the HR (Zhao et al 2004b). 

The avrRxo1 gene was identified by transferring a cosmid library 

made from X. o. pv. oryzicola into a strain of X. o. pv. 

oryzae that did not cause the HR when infiltrated into maize. 

A single cosmid clone was identified that conferred the ability 

to induce the HR. After testing individual genes from the cosmid 

in X. o. pv. oryzae, the ability to induce HR in an Rxo1-dependent 

manner was associated with a single gene. The avrRxo1 

gene encodes a novel protein that shares no homology with 

any known protein in public databases. 

Different maize lines were polymorphic for the Rxo1 

gene, allowing the gene to be mapped to the short arm of maize 

chromosome 6. An anonymous R gene-like sequence was then 

used as a probe to identify a gene family that cosegregated 

with the Rxo1 gene in several large mapping populations. The 

family consisted of five genes in the maize line B73 (Rxo1). 

Three of the five sequences were predicted to code for resistance 

genes, while the other two appeared truncated by 

retroelement-like sequences. RT-PCR experiments with primers 

designed to amplify all of the genes indicated that one family 

member was the most highly transcribed because all of 20 

cloned amplification products corresponded to this gene. Sequence 

analysis of RT-PCR products amplified with primers 

designed to be gene-specific verified that the other two family 

members predicted to code for resistance gene proteins were 

also transcribed, presumably at lower levels. 

Identification of the Rxo1 gene 

Several approaches were used to associate the putative resistance 

gene family with the Rxo1 gene and determine which, if 

any, of the family members confers the resistance phenotype. 

An RNAi silencing approach was first used to silence the family 

and determine whether this eliminated the resistance reaction. 

Transgenic HiII (Rxo1) maize lines were made that express 

both RNA strands of a fragment of the most highly transcribed 

family member. Eight of 15 transgenics carrying the 

RNAi construct showed little or no HR after inoculation with 

X. o. pv. oryzicola, indicating that silencing of the gene family 

silenced the Rxo1 phenotype. In addition, the lack of HR 

cosegregated perfectly with the transgene in progeny of one of 

the transgenics. The results indicated that one or more of the 

gene family members confers the Rxo1 phenotype. 

A transient transformation assay was used to examine 

the ability of the specific family members to interact with 

avrRxo1 and initiate HR. In these experiments, plasmid constructs 

carrying each of the potentially functional family members 

were separately cobombarded into Mo17 (rxo1) maize 

leaf cells with a plasmid expressing the avrRxo1 gene. The 

avrRxo1 gene was engineered for expression in plant cells by 

placing it between the HBT35S promoter (Sheen 1993) and 

the NOS terminator. The maize genes were used with their 

native promoters and also with the HBT35s promoter. A third 

plasmid, which was also included in each of the bombardment 

experiments, expressed the GUS gene as a reporter for cell 

viability. Previous experiments had indicated that transient 

delivery of the avrRxo1 gene into B73 maize cells expressing 

the native Rxo1 gene caused cell death and inhibited reporter 

gene expression. Delivery of the Rxo1 gene and avrRxo1 gene 

together would also be expected to reduce reporter gene expression 

if they conditioned cell death before noticeable 

amounts of the GUS protein were made. Only one of the gene 

family members noticeably reduced reporter gene expression 

in these experiments: the gene that was previously determined 

to account for most of the transcript in cDNA cloning experiments. 

The inhibition of reporter gene expression was nearly 

complete when a strong (HBT35s) promoter was used to drive 

the gene and only partial suppression was observed when the 

native promoter was used. These transient assay results indicated 

that the most highly transcribed family member is the 

functional Rxo1 gene. 

To verify the identity of the Rxo1 gene, stably transformed 

maize lines were constructed expressing the putative 

gene. A genomic copy of the gene with its native regulatory 

elements was transformed into a modified version of the maize 

line HiII that does not carry the Rxo1 gene. T 0 and T 1 plants 

expressing this gene gave HRs following inoculation with 

X. o. pv. oryzicola. 

Rxo1 functions in rice to control bacterial streak 

A transient assay was used to test whether Rxo1 would function 

to interact with avrRxo1 in rice. When Rxo1 and avrRxo1 


were cobombarded into rice, they reduced cell viability, as they 

did in maize, indicating that they interacted to induce HR. 

Stable rice transformants were then made in the rice cultivar 

Kitaake by Agrobacterium-mediated transformation. Rice cultivar 

Kitaake is highly susceptible to all strains of X. o. pv. 

oryzicola. In the T 0 generation of the transgenic rice lines, 36 

out of 82 lines carrying the native Rxo1 gene showed a strong 

HR when infiltrated with X. o. pv. oryzicola. The HR was confined 

to the infiltrated region, and the bacterial infection did 

not spread through the leaf. In untransformed control plants, 

the infiltrated region became water-soaked within 48 hours 

and the bacterial infection spread up the leaves, eventually 

killing the leaf. Segregating progeny were examined from two 

T 0 families that showed HR and the resistance response segregated 

as a single gene, as expected. The results therefore demonstrated 

that Rxo1 functions as a nonhost resistance gene in 

rice and conditions resistance reactions to X. o. pv. oryzicola. 

Transferring the Rxo1 gene to hybrid rice would therefore be 

expected to significantly improve its resistance to X. o. pv. 

oryzicola. 


The function of the Rxo1 gene in controlling bacterial streak 

resistance after transfer from maize to rice provides hope that 

resistance gene transfer among distantly related cereal species 

will be a valuable method of expanding our arsenal of genes 

for crop improvement. The existence of roughly 10,000 species 

of grasses provides a nearly unlimited source of genes 

and cereal genomes carry large numbers of resistance gene 

sequences (Monosi et al 2004). The challenge in using these 

genes is to develop efficient methods to find the useful ones 

and predict whether they will function in the backgrounds of 

heterologous species. 

References 

Ayliffe MA, Steinau M, Park RF, Rooke L, Pacheco MG, Hulbert 

SH, Trick HN, Pryor AJ. 2004. Aberrant mRNA processing of 

the maize Rp1-D rust resistance gene in wheat and barley. 

Mol. Plant-Microbe Interact. 17(8):853-864. 

Hulbert SH, Webb CA. 2001. Resistance gene complexes: evolution 

and utilization. Ann. Rev. Phytopathol. 39:285-312. 

Monosi B, Wisser RJ, Pennill L, Hulbert SH. 2004. Full-genome 

analysis of resistance gene homologs in rice. Theor. Appl. 

Genet. (In press.) 

Sheen J. 1993. Protein phosphatase activity is required for lightinducible 

gene expression in maize. EMBO J. 12(9):3497- 

3505. 

Thordal-Christensen H. 2003. Fresh insights into processes of 

nonhost resistance. Curr. Opin. Plant Biol. 6:1-7. 

Zhao B, Ardales EY, Brasset E, Claflin LE, Leach JE, Hulbert SH. 

2004a. The Rxo1/Rba1 locus of maize controls resistance reactions 

to pathogenic and nonhost bacteria. Theor. Appl. Genet. 

109:71-79. 

Zhao B, Ardales EY, Raymundo A, Trick H, Leach JE, Hulbert SH. 

2004b. The avrRxo1 gene from the rice pathogen Xanthomonas 

oryzae pv. oryzicola confers a nonhost defense reaction on 

maize with resistance gene Rxo1. Mol. Plant-Microbe Interact. 

17:771-779. 

Notes 

Authors’ addresses: Bingyu Zhao, Shavannor Smith, and Scot 

Hulbert, Department of Plant Pathology, Kansas State University, 

Manhattan, Kansas, USA; Jan Leach, Bioagricultural 

Sciences and Pest Management, Colorado State University, 

Ft. Collins, Colorado, USA, e-mail: 

shulbrt@plantpath.ksu.edu. 

Screening of allelopathic activity from rice cultivars 

by bioassay and field test 

Yoshiharu Fujii, Hiroshi Araya, Syuntarou Hiradate, and Kaoru Ebana 

The objectives of this study are to discriminate allelopathy 

from other competition in rice by a specific bioassay (plant 

box method), and to check weed suppression activity in paddy 

fields. The final goal is weed control without herbicides, or 

less use of herbicides. 

The main target of breeding rice cultivars in Japan has 

been good taste and high productivity. Recently, cultivar 

Koshihikari, famous for its good taste, has been sweeping over 

all of Japanese rice cultivation except for Hokkaido, a northern 

district of Japan, and a need remains for still more good 

rice cultivars with good taste. Now, many other new candidates 

are being prepared for the market. On the other hand, 

sustainable agriculture, which aims at less impact on the envi- 

ronment and makes sure human beings have a sustainable existence, 

is needed all over the world. In paddy fields, problems 

of weed control, which had been tedious and laborious 

work for a long time, have been successfully conquered by the 

introduction of specific and powerful herbicides, and few problems 

remain except for some perennial weeds in Japan. We 

thus need to be able to use allelopathy to control paddy weeds 

without using much or any herbicide and to accomplish safe 

and labor-saving rice cultivation. Our group therefore started 

research on allelopathy in rice. Our target is to search for potent 

allelopathic rice varieties and to use these strains for weed 

control in paddy fields without herbicides or with little use of 

them. 


Weed dry weight (g/30 × 60 cm) 

25 

Others 

20 

Barnyardgrass 

15 

Konagi 

10 

5 

Inuhatarui 

0 

Awa-Akamai Houman jinjya 

Akamai 

Midorimai 

Soujya 

Akamai 

Fig. 1. Weed suppression by allelopathic rice in paddy fields. 

Nebarimochi 

Nihonbare 

For allelopathy of rice, little research has been done so 

far. Though effects of rice residues on the following crop were 

reported (Chou 1986), little research has been done on the effect 

of root exudates of rice. Our group has been engaged in 

work to search for allelopathic plants and to examine allelopathy 

and ascertain its mechanism. In this research, we have developed 

a new system that can prove the possibility of allelopathy, 

called the “plant box method” (Fujii 1999). This method 

is a sort of mixed planting using agar (no nutrient) as a medium. 

By using this method, we can visualize allelopathy by 

root exudates, and we can observe differences between varieties, 

even in the same species. 

Some farmers suggested from their experience that red 

rice of Japanese old-type rice varieties has few accompanying 

weeds. Some rice cultivars used before the introduction of 

herbicides are possibly resistant to paddy weeds. And, some 

rice cultivars or ancestors of rice growing in tropical countries 

where weed growth in paddy fields is serious may be resistant 

to weeds. So, we started screening allelopathic activity of rice 

cultivars by the plant box method. Our work is quite preliminary, 

and our findings might not apply directly to actual weed 

control. But, we hope that these approaches to the biological 

control of paddy fields will awaken Asians to weed management 

using allelopathy from their original rice cultivars. 

A total of 500 rice cultivars, varieties, and related plants 

were tested. They were supplied from the Japanese Gene Bank 

(MAFF, Tsukuba), the core collection by Dr. M. Okuno 

(HNAES), the Malaysian Gene Bank (MARDI), the red rice 

collection by Dr. T. Itani (Hiroshima Prefectural University), 

and the wild and traditional rice collection in Bhutan by Dr. Y. 

Sato (Shizuoka University). 

The plant box method was followed according to Fujii 

(1999). The receiver plant was lettuce; Great Lakes 366 is 

highly susceptible to bioactive substances. The plant box 

method was used for assaying the activity of diffusible phytotoxic 

metabolites (potentially allelopathic compounds) in agar 

in the laboratory and is summarized briefly as follows. Rice 

plants were grown in sand culture in pots for 4 to 8 weeks. 

Roots were washed with tap water and then with deionized 

water to remove sand. Roots of three rice plants were placed 

gently into a nylon-mesh-bounded cylinder (3.2 cm diameter 

by 6.6 cm tall), which was placed upright in one corner of a 70 

× 70-cm-wide by 100-cm-deep clear plastic box. Plant stems 

were taped to the corner of the box to stabilize the plants in the 

box. Hot liquid agar (0.75%; 250 mL per box) was cooled to 

40 °C and poured into the boxes, covering all root material, 

but not covering stems. Boxes containing agar and rice plants 

were cooled on ice until the agar had solidified (about 30 min). 

Thirty-one lettuce seeds were inserted partially into the surface 

of the agar in an equidistant grid pattern. Seeds were about 

halfway into the agar with the germ (pointed) end down using 

porcelain forceps. Boxes were sealed around the rice stems 

and over the box surface with plastic wrap to minimize water 

loss, and placed in a fluorescent-lighted incubator for 5 days 

(25 °C, 14-h days; 20 °C, 10-h nights). Lettuce seedlings were 

then removed and root and shoot lengths determined to the 

nearest mm. Regression analysis was used to estimate lettuce 

root length at zero distance from rice roots (y-intercept), and 

slope of the inhibition curve. 

Experiments were conducted with rice using an agar diffusion 

plant box method. Apparent allelopathic activity was 

observed in the second year of a 5-year field trial in transplanted 

water-seeded rice. In this study, the red Shrine rice, 

called Awa-Akamai, reduced total weed biomass by about 80– 

90% compared with the commercial Japanese cultivar 

Nipponbare (Fig. 1). The mechanisms responsible for weed 

suppression in Awa-Akamai may be valuable for improving 

weed suppression in commercial rice cultivars. Suppression 

of barnyardgrass and broadleaf paddy weeds has been shown 

in the field in Tsukuba with Awa-Akamai and other red rice 

cultivars. 

In plant box tests, roots of several rice lines, including 

Awa-Akamai, were highly inhibitory to the elongation of lettuce 

roots. Nipponbare (a nonsuppressive cultivar in field tests 


Table 1. Evaluation of allelopathic activity of rice by the plant box method (selected data). Data are shown in 

inhibitory % to the control by the plant box method. 

Name of cultivar R a (%) Name of cultivar R a (%) Name of cultivar R a (%) 

Awa-Akamai R 96 Canaboug bong R 78 Mitsuyou 23 41 

Melor 94 Sathi R 78 Suigen 258 41 

Sisumuchimai 92 Kahei 78 Nankin 11-gou 39 

Tsushima-Akamai R 93 Kouketsu-mochi R 77 Koganemasari 38 

Soujya-Akamai R 91 Hokuriku 127 77 Kameno-o 37 

Deng Mack Tek 89 Li Zi Hong R 77 Nihonbare 36 

Keichou 2 gou 88 Mack Kheua R 76 IR30 32 

Yamasaka 85 Dam Ngo R 75 IR24 28 

Asominori 85 Nago-akaho R 73 Shinsyuukaneko 28 

Kurumi-wase 84 Taducan 72 Reihou 26 

Dular 84 Nepal No. 1 72 Sasanishiki 24 

Takane-nishiki 84 Chuuseishinsenbon 68 Zinriki 21 

Moroberekan 83 Khauk Yoe 64 Toyonishiki 21 

Tou-Boshi 83 PI312777 62 Rexmont 20 

Padi Keinikir Puti R 83 Shikoku-mai R 62 Si Li Gu 20 

Sekiyama 83 Fujisaka 5 gou 58 Todorokiwase 20 

Jyukkoku 83 Dourado Precoce 55 Basmati 19 

Mock Kheua 81 Karh R 54 Asahi 19 

Vista 81 Nankin-kaoriine 54 Masria 19 

Nepal No. 8 81 Koshihikari 54 Sari Queen 18 

Daw Dam 80 Taichuu 65 49 Kantou 146 13 

Houman-jinjyamai R 80 Huao Ba Wang 46 Kouryowai 4 13 

Mushashi-kogane 80 Sabarmati 45 Nourin 8 gou 10 

a 

R in this column means red rice. 

in Tsukuba) was usually less inhibitory to lettuce root elongation 

than was Awa-Akamai. In plant box experiments in which 

lettuce seed planting was delayed 2 or 3 days, allowing 

allelochemicals to diffuse greater distances, apparent allelopathic 

activity was greater following longer delays. 

The results of the effect of rice cultivars are shown in 

Table 1, in which % means the inhibitory percentage of lettuce 

radicle length in the root surface of rice by extrapolation of 

the radicle length of each point apart from the root zone of 

rice to the control. These data show the activity and quantity 

of root exudates. Results of screening 500 rice varieties showed 

a wide variation in activity. Some varieties showed no activity, 

some showed strong activity, and, in some cases, the top of the 

lettuce radicle became brown to black. The strongest activity 

was of Awa-Akamai, an old red rice reserved in the old shrine 

in Shikoku. Its inhibitory activity was equivalent to the activity 

of Mucuna and Avena species that showed the strongest 

activity we ever tested. In Table 1, the letter R shows red rice 

and, generally speaking, the allelopathic activity of red rice is 

stronger than that of other rice cultivars. 

Though our research aimed to search for potent allelopathic 

rice varieties that inhibit the growth of accompanying 

weeds, the work reported here is only the first step because we 

used lettuce as a model plant and did not yet estimate the allelopathic 

effect in general paddy fields. But it became clear 

that there were distinct differences among rice varieties. 

We would like to assess allelopathic activity by using 

paddy weeds as the acceptors, to identify allelopathic substances 

and specific genes in relation to allelopathy, and to 

breed new varieties that have allelopathic ability to control 

weeds. This project may take another 10 years or more, and it 

might be impossible to suppress paddy weeds completely by 

only allelopathy, but, in combination with other biological 

control, allelopathic rice strains will help us to achieve sustainable 

and safe rice production. 

References 

Azmi M, Abdullah MZ, Fujii Y.2000. Exploratory study on allelopathic 

effect of selected Malaysian rice varieties and rice field 

weed species. J. Trop. Agric. Food Sci. 28(1):39-54. 

Baker P. 1992. Weed resistance hidden in rice genes. Rice J. 8-9 

March. 

Chou C-H. 1986. The role of allelopathy in subtropical 

agroecosystems in Taiwan. In: Putnam AR, Tang C-S, editors. 

The science of allelopathy. New York (USA): John Wiley 

& Sons. p 57-73. 

Fujii Y, Shibuya T, Yasuda T. 1991. Discrimination of allelopathy of 

upland rice, taro, and oat by substitutive experiment and its 

modified experiments. Jpn. J. Soil Sci. Plant Nutr. 62:357- 

362. 

Fujii Y. 1994. The allelopathic effect of some rice varieties. In: Integrated 

management of paddy and aquatic weeds in Asia. FFTC 

Book Series, No. 45. p 160-165. 

Fujii Y. 1999. Discrimination and proof methods for allelopathy by 

bioassay, greenhouse and field tests. In: Macias FA, Galindo 

JCG, Molinillo JMG, Cutler HG, editors. Recent advances in 

allelopathy. Vol. 1. p 25-28. 


Fujii Y, Gealy DR. 2000. Screening of allelopathic activity from rice 

cultivars. In: Oono K, Komatsuda T, Kadowaki K, Vaughan 

D, editors. Integration of biodiversity and genomic technology 

for crop improvement. Tokyo (Japan): Ministry of Agriculture, 

Forestry, and Fisheries. p 135-136. 

Fujii Y, Shibuya T, Nakatani K, Itani T, Hiradate S, Parvez MM. 

2004. Assessment method for allelopathic effect from leaf litter 

leachates. Weed Biol. Manage. 4(1):19-23. 

Notes 

Authors’ addresses: Yoshiharu Fujii, Hiroshi Araya, and Syuntarou 

Hiradate, National Institute for Agro-Environmental Sciences, 

Tsukuba, e-mail: yfujii@affrc.go.jp; Kaoru Ebana, National 

Institute of Agrobiological Sciences, Tsukuba, Japan. 

Evaluating near-isogenic lines with QTLs 

for field resistance to rice blast from upland rice cultivar 

Sensho through marker-aided selection 

Norikuni Saka and Shuichi Fukuoka 

Rice blast caused by Pyricularia oryzae is the most serious 

disease of rice in the world, and cultivation of disease-resistant 

cultivars is highly desirable for stable rice production. In 

Japan, rice cultivars with a complete rice blast-resistance gene, 

such as Kusabue, suffered from a massive outbreak of blast 

disease in the 1960s. Thereafter, the importance of a program 

for breeding varieties with field resistance to rice blast was 

strongly recognized (Takita and Solis 2002). However, the 

mechanism of field resistance has not been fully analyzed because 

of the complicated hereditary mode and difficulty of 

bioassay; thus, breeding did not proceed satisfactorily. In this 

study, we produced recombinant inbred lines using DNA markers 

for high field resistance of upland rice to simplify complicated 

phenomena, and to develop a method of breeding resistant 

cultivars. 

History of the breeding of rice blast resistance at AARC 

At the Aichi Agricultural Research Center (AARC), Mountainous 

Region Agricultural Research Institute, we have been 

breeding blast-resistant lines of rice since 1933 in an area where 

an outbreak of rice blast occurred, by introducing the field 

resistance to blast of upland rice cultivar Sensho into paddy 

rice. Until now, scores of rice-blast-resistant cultivars have been 

bred. However, it has been difficult to eliminate the inferior 

characteristics of upland rice such as burning up and poor quality, 

and only paddy rice with a part of the upland rice genes 

introduced has been released (Saka et al, in preparation). Breeding 

for excellent eating quality satisfying the needs of Japanese 

consumers along with field resistance to rice blast has 

not succeeded because these traits are controlled by polygenes. 

Therefore, we tried to resolve the long-pending problems using 

genomic techniques. 

Mapping and breeding NILs having each QTL region 

Kato et al (2002) crossed upland rice cultivar Sensho with blastsusceptible 

japonica paddy rice variety Norin 29 and analyzed 

the resistance to leaf blast of the progeny by QTL analysis. 

They found four QTL regions, which were detected in the vicinity 

of RFLP markers G271 (LOD = 24.8) and G177 (5.7) 

of chromosome 4 (Targets 1 and 4), in the vicinity of C1172 

(3.6) of chromosome 11 (Target 2), and in the vicinity of S826 

(3.4) of chromosome 12 (Target 3). 

Based on this finding, we backcrossed (B 4 F 4 –B 3 F 6 ) 

Mine-asahi, a susceptible variety with good eating quality, with 

Sensho to breed lines having each of the four QTL regions 

independently, using DNA markers for selection. We bred NIL 

(near-isogenic line)-1, which has the Mine-asahi genome except 

for the RFLP marker G271 region of Sensho (smaller than 

22.3 cM), the Target 1 region; NIL-2, which has the Mineasahi 

genome except for the SSR marker RM229–RM287 region 

(smaller than 16.9 cM) of chromosome 11 of Sensho, the 

Target 2 region; and NIL-3, which has the Sensho genome 

except for the SSR marker region RM313–S13752 (smaller 

than 14.7 cM) of chromosome 12 of Sensho and a part of chromosomes 

1 and 9, where a QTL was not detected, the Target 3 

region. In addition, NILs that showed resistance, although the 

chromosome with the resistance gene was not specified, were 

bred (NIL-5). From the results of a race test, the complete 

resistance genes of these NILs were Pia and Pii, like their backcrossed 

parent Mine-asahi. 

Effect of each QTL region on blast severity 

The leaf and panicle blast tests gave consistent results in the 

two years of testing, and the difference between the criterion 

varieties having Pia and Pii and other cultivars was clear. Under 

the same test conditions, we performed a leaf blast test for 

each NIL. The degree of resistance varied with the NIL, and 

was in the order of NIL-1 > NIL-2 >NIL-5 > NIL-3 (Fig. 1). 


1 2 3 4 

4 5 6 7 8 9 

Fig. 1. Leaf blast resistance in NILs of Mine-asahi backcrossed to Sensho. 1 = Chubu 105 (r), 2 = 

Mine-asahi (ms), 3 = Mine-hibiki (rm), 4 = Hitomebore (s), 5 = Sensho (rr), 6 = NIL-1 (Target 1: Chr. 4- 

1), 7 = NIL-5 (the region undecided), 8 = NIL-2 (Target 2: Chr. 11), 9 = NIL-3 (Target 3: Chr. 12), 1–5 = 

check varieties. 

This result is in agreement with the results of QTL analysis 

reported by Kato et al (2002). The result of the panicle blast 

test was similar to that of the leaf blast test, confirming that 

the QTL region for resistance to leaf blast also confers resistance 

to panicle blast. 

The field resistance to rice blast of the four NILs was 

compared with that of the criterion varieties (Saka 1996) (Fig. 

2). NIL-1, which had the highest resistance, developed severer 

symptoms of leaf blast than Sensho, but was more resistant 

than Chubu 105, which is the criterion variety with high resistance, 

and was judged to have very high resistance. The resistance 

to panicle blast of NIL-1 was slightly lower than that of 

Sensho but was higher than that of Tiyonishiki, which had 

slightly high resistance. Thus, resistance of NIL-1 to panicle 

blast was judged to be high to slightly high. These results suggest 

that NIL-1, having the Target 1 region, has very high resistance 

to rice blast. In particular, resistance to leaf blast of 

NIL-1 was very high compared with that of the criterion varieties, 

and we consider that this line could be cultivated without 

the use of agrochemicals for control of this disease in areas 

where rice blast occurs commonly in Japan. The resistance 

to leaf blast of NIL-2 having the Target 2 region was 

similar to or slightly higher than that of Mine-hibiki, a criterion 

variety, and the resistance to panicle blast of NIL-2 was 

similar to that of Tiyonishiki, which has slightly high resistance. 

Mine-hibiki and Tiyonishiki are the varieties having the 

highest resistance to rice blast among the recommended varieties 

in Japan, and they show high resistance in the area where 

rice blast commonly occurs. Therefore, Target 2 was considered 

to contribute by itself to stable rice production. NIL-5, in 

which the region for the resistance gene was not specified, 

had medium resistance to leaf and panicle blast. NIL-3, having 

Target 3, had slightly higher resistance to leaf and panicle 

blast than Mine-asahi, which has slightly low resistance, and 

was judged to have medium to slightly low resistance. 

Elite NILs with a narrow Sensho region 

The above-described four NILs retained the inferior traits of 

Sensho such as poor eating quality and coarseness of stems 

and leaves compared with repeatedly backcrossed parent cultivar 

Mine-asahi. To narrow the Sensho region, we selected a 

line with a narrower region for the resistance from 672 individuals 

of Mine-asahi*3/Sensho (B 2 F 2 ) having the Target 1 

region, using eight PCR markers included in RFLP marker 

G271–G89B (14.6 cM). As a result, we succeeded in selecting 

a NIL that has the Mine-asahi genome except for the chromosomal 

segment smaller than 0.5 cM having Target 1. This region 

coincided with the estimated locus of rice blast-resistance 

gene pi21 (Fukuoka and Okuno 2001), and this NIL had field 

resistance similar to that of the NIL having Target 1. Thus, the 

NIL is considered to be a useful breeding material. In the future, 

we will perform linkage drag associated with introgressed 

Sensho DNA to eliminate undesirable characteristics by 

marker-aided selection, using various NILs having different 

sizes of Sensho region that were selected in this study. 

The next target 

The level of field resistance to blast of NILs having QTL regions 

from Sensho was clarified as mentioned above. Among 

them, the NIL with Target 1 had high resistance, which was 

not observed in known paddy rice varieties in Japan, and we 

may be able to grow varieties with this region introduced without 

using agrochemicals for control of rice blast. Hereafter, 

not only the simple use of this region but also examination of 


Panicle blast severity 

(poor) 

9 

r 

mr 

s 

8 

6 

7 

3 

4 

m 

6 

2 

mr 

5 

1 

4 

5 

3 

2 

1 

0 

(excellent) 0 1 2 3 4 5 6 7 8 9 (poor) 

Leaf blast severity 

Fig. 2. Blast resistance in NILs of Mine-asahi backcrossed to Sensho. 1 = NIL-1 

(Target 1: Chr. 4-1), 2 = NIL-2 (Target 2: Chr. 11), 3 = NIL-3 (Target 3: Chr. 12), 4 

= NIL-5 (the region undecided), 5 = Sensho (rr), 6 = Mine-asahi (ms); blue dotted 

line: leaf blast check varieties; Chubu 105 (r), Mine-hibiki (mr), Hitomebore 

(s); green dotted line: panicle blast check varieties; Tiyonishiki (mr), Nipponbare 

(m). 

the cumulative effects and confirmation of the accompanying 

inferior characteristics may be necessary. 

Recently, we started a national project to introduce these 

QTL regions into the leading Japanese variety of rice, 

Koshihikari, which occupies 32% of the paddy fields in Japan. 

If field resistance to rice blast can be introduced into this 

variety while retaining its excellent eating quality, we will be 

able to cultivate rice stably without disturbing the environment. 

Along with the above project, we are now trying to clone 

the pi 21 gene (Fukuoka and Okuno 2001). By clarifying the 

function of this gene, we should be able to understand easily 

the mechanism of complicated field resistance to rice blast. 

References 

Fukuoka S, Okuno K. 2001. QTL analysis and mapping of pi21, a 

recessive gene for field resistance to rice blast in Japanese 

upland rice. Theor. Appl. Genet. 103:185-190. 

Kato T, Endo I, Yano M, Sasaki T, Inoue M, Kudo S. 2002. Mapping 

of quantitative trait loci for field resistance to rice blast in 

upland rice, Sensho. Breed. Res. 4:119-124. 

Saka N. 1996. Panicle blast resistance: test method using the area 

where outbreak of rice blast occurred. In: Yamamoto R, 

Horisue N, Ikeda R, editors. Rice breeding manual. Tokyo 

(Japan): National Agricultural Research Center: Yokendo 

Printing Co., Ltd. p 15-19. 

Saka N, Kudo S, Sugiura K, Okuda T, Terashima T. History for 70 

years, present situation and the future of the breeding of rice 

blast resistance in the Mountainous Region Agricultural Research 

Institute, AARC. Plant Prod. Sci. (In preparation.) 

Takita T, Solis O. 2002. Rice breeding at the National Agricultural 

Research Center for the Tohoku Region (NARCT) and rice 

varietal recommendation process in Japan. Bull. Natl. Agric. 

Res. Cent. Tohoku Reg. 100:93-117. 

Notes 

Authors’ addresses: Norikuni Saka, Aichi Agricultural Research 

Center (AARC), Mountainous Region Agricultural Research 

Institute. Inabu, Aichi 441-2513, Japan, e-mail: 

norikuni_saka@pref.aichi.lg.jp; Shuichi Fukuoka, National 

Institute of Agrobiological Sciences (NIAS), Tsukuba, Ibaraki 

305-8602, Japan, e-mail: fukusan@affrc.go.jp. 



Pest management relies on the use of appropriate technologies 

that are based on location-specific pest composition, environment, 

and farmers’ socioeconomic constraints. Integrated pest 

management (IPM) attempts to prevent pathogens, insects, and 

weeds from causing economic crop losses by using a variety of 

management techniques that are cost-effective and cause the 

least damage to the environment. The objective of this session 

was to provide participants with new ideas and a better understanding 

of key issues in pest management to develop environment-friendly 

technologies, rather than to draw any conclusions 

from the studies presented. 

The oral session consisted of two subsessions: the first 

dealt with insect pest management and the second with disease 

management. Subsession 1 featured three presentations and 

four presentations were delivered in subsession 2. K. Kiritani and 

T.W. Mew presented their philosophy and critical overview on insect 

pest management and disease management for sustainable 

rice production, respectively. K. Kiritani focused on integrated 

biodiversity management (IBM), a new concept on how to merge 

economic optimization and conservation in pest management. 

T.W. Mew focused on three issues and research agendas: (1) the 

need for increased genetic diversity, (2) the health of farmersaved 

seeds for increased storage and shelf-life, and (3) soil 

health. To ensure sustainability in rice production systems and 

achieve the next Green Revolution, he advocates using appropriate 

varieties, emphasizing seed health, and using cropping practices 

that ensure a healthy ecosystem. The other two presentations 

made by X. Yu and M. Matsumura in the first half of Session 

16 discussed the use of natural enemies, that is, how to 

make use of domestic and migratory natural enemies as the most 

promising control measures in IPM. The three presentations made 

by T. Ashizawa, C.M. Vera Cruz, and S.H. Hulbert in the second 

half of Session 16 were all related to the use of host-plant resistance 

against pathogens. The first two presentations covered blast 

management using simulation modeling with BLASTMUL for rice 

multilines, and the association of candidate defense genes with 

quantitative disease resistance and in silico analysis of their characteristics. 

According to T. Ashizawa, BLASTMUL simulations for 

leaf blast epidemics usually agreed with those observed in different 

locations and multiline series, suggesting that this is useful 

for forecasting leaf blast epidemics in multilines and is a powerful 

tool in determining effective mixture proportions of nearisogenic 

lines in multilines. C. Vera Cruz and colleagues employed 

the candidate gene approach aimed at accumulating different 

mechanisms of quantitative resistance to blast. This approach 

integrates the molecular analysis of host-pathogen interactions, 

gene mapping and expression, and selection for blast resistance 

in elite rice lines. S.H. Hulbert discussed the transfer of resistance 

genes between cereal species. He explained how the function 

of different resistance genes was evaluated after transfer 

between distantly related cereal species such as rice and maize. 

R. Wang summarized the oral session He remarked that 

the studies presented showed the evolution of the science of 

pest management in relation to understanding pest systems, that 

is, the scientific understanding has gone beyond the use of component 

technologies into the production system. The current trend 

makes use of converging strategies for pest management as depicted 

in the areas of enhancing natural control, ecosystem diversity, 

and genetic diversity, all contributing toward increasing 

resilience of the system. Modern tools for studying host-plant 

resistance are being employed in understanding and managing 

the system. The potential for protecting the environment through 

reductions in pesticide application has also been recognized. 

Applying the concepts and tools in the context of modern science, 

R. Wang saw a great potential for collaboration among 

advanced research institutions and national agricultural research 

and extension systems. 

A total of 36 posters were presented in Session 16, covering 

insect pest, disease, and weed management. Two of the presentations 

were selected to represent the diversity of topics in 

this session. These posters can be read in this proceedings. These 

are “Evaluating near-isogenic lines with QTLs for field resistance 

to blast from upland rice cultivar Sensho through marker-aided 

selection” by N. Saka and S. Fukuoka and “Screening of allelopathic 

activity from rice cultivars by bioassay and field test” by Y. 

Fujii, H. Araya, S. Hiradate, and K. Ebana. 


SESSION 17 

Rice supply and demand 

CONVENER: O. Koyama (JIRCAS) 

CO-CONVENER: D. Dawe (IRRI)

International trade in rice: recent developments 

and prospects 

Concepcion Calpe 

In the past several decades, the international rice market has 

undergone major changes, in particular a shift in the general 

policy setting, a strong expansion in the volumes of trade, and 

a lingering tendency for world prices to decline in real terms 

and relative to the other two most-traded cereals, wheat and 

maize. Nonetheless, the world rice market continues to be regarded 

as distorted, thin, segmented, and volatile. This paper 

discusses whether these attributes still portray the market in 

the light of the trade liberalization drives prevailing in the 

1990s. 

On the policy front, interventions have diminished in the 

wake of the market liberalization launched by several countries 

since the late 1980s. The WTO agreement, in 1994, also 

disciplined government policies and helped improve market 

access. Nonetheless, rice continues to be one of the most protected 

commodities in both developing and developed countries, 

subject to high tariff and nontariff barriers, export restrictions 

or aids, state trading, and domestic market interventions. 

Since the early 1960s, trade in rice has expanded at about 

3% per annum, not much different from the pace of growth in 

wheat or maize trade (Fig. 1). However, growth has been far 

from steady. The liberalization thrust of the 1990s coincided 

with a period of dynamic expansion in the volume of rice trade, 

which succeeded a decade of relatively lackluster growth in 

the 1980s. The volume of rice exchanged rose from less than 7 

million tons in 1961 to 24 million t in 2000 and has continued 

to expand further in the early 2000s, surpassing 28 million t in 

2001 and 2002. Nevertheless, the international rice market is 

still small relative to the other major cereals, with an average 

of 27 million t in 2000-03, about one-quarter of the volume 

traded in wheat and a little over one-third of the trade in maize. 

Rising import demand by countries in Asia and Africa 

was the main force underpinning trade in rice in the 1990s and 

early 2000s. The increases in imports were often a reflection 

of more open trade policies but were also prompted by several 

production setbacks, for instance, in 1997 in the wake of an El 

Niño weather anomaly. Despite the consolidation of countries 

in Africa and the Near East as important and stable destinations 

of rice trade, the demand side of the rice international 

market remains highly dispersed geographically, with the top 

ten importers accounting for only 40% of the total. 

Most of the trade expansion witnessed in the past two 

decades was met by traditional exporters. Thailand, in particular, 

has maintained its leadership as the top rice exporter 

since 1980. Major inroads were made by Vietnam, which became 

the second most important source of trade supplies in 

the 1990s, a position from which it was eclipsed in the early 

2000s, when India started granting export subsidies. Despite 

Million t 

160 

140 

120 

100 

80 

60 

40 

20 

0 

1961 

1964 

1967 

1970 

1973 

1976 

Wheat: Y = 48.147e 0.0269t 

1979 

Year 

Fig. 1. Trends in global cereal trade, 1961-2003. 

changes in the relative positions of the major exporters, the 

supply side of the rice international market is still highly concentrated, 

with the top four exporting countries (Thailand, 

India, Vietnam, and the United States) supplying 66% of trade 

and the top ten more than 90% of the total. 

Rice is not a homogeneous commodity and there are now 

more than 50 different published international price quotations 

for rice. In fact, there are distinct submarkets featured according 

to several criteria, the most important of which are variety, 

quality (defined mainly by the percentage of brokens), and the 

degree of processing. 

The expansion of trade witnessed in the 1990s was accompanied 

by small but significant changes in the structure of 

the world rice market and in the relative importance of each 

segment (Table 1). The bulk of global trade continues to be in 

the form of milled, indica, and higher-quality rice (defined as 

containing less than 20% of brokens). However, aromatic rice 

varieties, lower-quality rice, parboiled, and paddy have made 

large in-roads and have increased market shares. Those gains 

were mainly at the expense of trade in japonica, higher-quality, 

and milled rice. 

Most of the highlighted changes can be associated with 

shifts in the geographical pattern of trade. The increasing importance 

of Africa and several Asian countries as destinations 

of rice flows, in particular, has sustained a large increase in the 

trade of lower-quality rice. The growing importance of aromatic 

rice varieties in global trade reflects brisk imports to the 

European Union (mainly of Basmati rice, imported under preferential 

access conditions), the United States, Canada, and 

Australia. However, it can also be associated with large deliveries 

of Hom Mali rice (a fragrant variety from Thailand) to 

1982 

1985 

Maize: Y = 24.223e 0.0351t 

Rice: Y = 6.0645e 0.0321t 

1988 

1991 

1994 

1997 

2000 

2003 


Table 1. World trade in rice. 

1992-94 2001-03 

000 t Share (%) 000 t Share (%) 

Variety 

Indica 11,663 76 20,068 75 

Japonica 2,132 14 3,186 12 

Aromatic 1,353 9 3,322 12 

Glutinous 115 1 242 1 

US$ t –1 

1,400 

1,200 

1,000 

800 

600 

400 

200 

Rice price nominal 

Rice price real 

y = 814.52e –0.031x 

R 2 = 0.7354 

Quality 

High 11,781 77 20,226 75 

Low 3,482 23 6,592 25 

Degree of processing (forms) 

Paddy 263 2 1,122 4 

Husked 508 3 1,077 4 

Milled 12,559 82 20,639 77 

Parboiled 1,934 13 3,980 15 

Total trade 15,263 100 26,818 100 

countries in Africa, particularly Côte d’Ivoire, Ghana, and 

Senegal, albeit with a high percentage of brokens. On the other 

hand, high degrees of protection in japonica rice markets such 

as Japan, the Republic of Korea, the European Union, and 

Turkey have constrained the opportunities to expand japonica 

rice trade. Tariff escalation, whereby the more processed forms 

of a commodity are assigned higher tariff rates, has favored a 

strong expansion of trade in paddy, principally to Latin America 

and the Caribbean. 

An international commodity market is considered “thin” 

when it represents a relatively small proportion of global production. 

The international rice market represented only 3% to 

5% of global production in the 1980s, but a strong expansion 

of world trade since the mid-1990s has made it “deepen,” as it 

has come to represent 7% of global production in recent years. 

Nonetheless, the international rice market remains thin compared 

with wheat or maize, whose trade now accounts for some 

18% and 13% of global production, respectively. 

Thin commodity markets are often subject to large swings 

in volumes since relatively small changes in supply or use in 

important producing countries may give rise to large increases 

or contractions in their exports or imports. In general, however, 

such countries have preferred to balance their domestic 

markets by building up or drawing supplies from stocks, with 

trade considered only a “residual” option. 

Global trade in rice has fluctuated widely over the past 

two decades, from a minimum of 10.6 million to 28.3 million 

t, and variability measured by the coefficient of variation (CV) 

was high at 37% compared with a 12% variability of global 

rice production. It was also much higher than the variability of 

wheat and maize trade, which had a CV of 6% and 8%, respectively. 

Variability of rice trade measured decade by decade 

pointed to a greater stability. In the 1980s, volumes exchanged 

on the international rice market fluctuated within a 

relatively small range of 11 million to 14 million t, resulting in 

0 

1960 

1963 

1966 

1969 

1972 

1975 

Fig. 2. International rice prices. 

1978 

1981 

1984 

1987 

1990 

1993 

Year 

1996 

1999 

2002 

a measure of variability on the order of 8%, not very different 

from that prevailing on the wheat and maize markets. Trade 

fluctuated within a much broader band of 12 million to 28 

million t in the 1990s, which gave rise to a much higher CV of 

26%. Thus, the strong tendency for international trade to grow 

in the 1990s was associated with much greater volatility in 

volumes. 

The international rice market has also been characterized 

by a long-term tendency for world rice prices (represented 

by the Thai 5% broken rice, f.o.b. Bangkok) to fall in real 

terms (deflated by the index of unit value of manufactured 

goods) from 1961 to 2003 (Fig. 2). The decline in constant 

US$ has been 3% per annum and, in 2003, rice was worth less 

than 40%, in real terms, of its 1961 value. Rice prices also 

declined relative to wheat and maize. If 1 t of rice could be 

exchanged for 2.5 t of wheat in 1961, it could be bartered for 

only 1.3 t in 2003. A similar loss in value was evidenced relative 

to world maize prices. 

Although variability in the volume of rice trade rose in 

the 1990s compared with the 1980s, this did not cause the variability 

of world prices to follow suit. On the contrary, rice prices 

have become more stable over time, to the point of achieving 

levels of volatility similar to those exhibited by wheat and maize 

prices. Actually, on an annual frequency basis, prices in the 

1990s were more stable for rice than for wheat or maize, in 

sharp contrast to the pattern prevailing in the 1960s and ’80s. 

Stabilization of world rice quotations was also evidenced on a 

monthly frequency basis. 

Thus, the rising variability of trade flows was not associated 

with more volatile world prices, which have instead stabilized. 

Several explanations can be offered to explain this 

paradox. First, the deepening of the international market has 

meant a greater dependability of supplies. The existence of 

large buffer stocks, the improved flows of information on markets 

and prices, and the introduction of disciplines (restrictions 

on policies because of trade agreements such as WTO) 

on national and international policies are also believed to have 

fostered price stability on the market, in spite of the wider 

fluctuations in the volumes of trade. 

Session 17: Rice supply and demand 493

In summary, it can be said that international trade in rice 

has become less distorted, less thin, more unstable volumewise, 

but more stable price-wise and more dependable. This 

could have important implications for policymakers by encouraging 

them to lower domestic protection to the rice sector and 

increase their countries’ reliance on trade. However, there is 

still much uncertainty on whether the tendencies observed in 

the 1990s will linger into the rest of the 2000s and in the de- 

cades to come. Against this backdrop, the outcome of the ongoing 

multilateral trade negotiations will be of particular importance 

in shaping the future of the international rice market. 

Notes 

Author’s address: Food and Agriculture Organization. 

Rice in the world at stake 

Shoichi Ito 

Rice has been consumed as a staple food more than any other 

grain in the world. However, in Asia, where 90% of global 

rice is consumed, per capita rice consumption (PCRC) has been 

declining over time. In Japan, PCRC decreased by almost half 

from 125 to 65 kg during the last 45 years. In Taiwan (China), 

it decreased by more than two-thirds from 160 to 50 kg during 

the same period. As a result, domestic rice production in both 

countries has decreased substantially. The declining trend of 

PCRC seems likely to continue even though the governments 

of these nations have tried hard to reverse the trend (Fig. 1). 

The same type of situation has been occurring in other 

Asian countries such as South Korea, China, and India. The 

current situation in South Korea is similar to what happened in 

Japan in the 1960s when the decline began. In China, the pace 

of the decline currently appears to be small at 0.5 kg per year, 

decreasing from 110 kg in 1991. However, the diet in large 

areas of China is quite similar to the Taiwanese diet. Therefore, 

it would not be surprising even if the PCRC in China 

were to decline as fast as what has happened in Taiwan (China). 

The declining trend of PCRC in the Asian countries was first 

reported by Ito et al (1989) but it has not hit bottom yet. It was 

big news in Japan in May 2004 that PCRC in Japan dropped 

by as much as 2.4% during the preceding 12 months (Asahi 

Newspaper, June 2004). The countries where PCRC is decreasing 

are not only economically developed countries. This has 

spread all over Asia. 

Even in China, which is the world’s largest rice-producing/-consuming 

country (holding 1.3 billion people), its PCRC 

has been decreasing steadily during recent years. PCRC in 

China was increasing until the early 1990s; however, it peaked 

at 110 kg in 1991 and then started decreasing. It was as low as 

105 kg in 2004, about a 4% drop during the last decade. If the 

PCRC dropped by 1 kg in China, this would mean a surplus of 

1.3 million tons of rice generated in the nation. Population in 

China is estimated to peak in the 2030s; therefore, total rice 

consumption in the country could decrease anytime soon if 

the current situation continues. 

A continuous decrease in PCRC in a country eventually 

leads to a decrease in total consumption in the country. And 

this can lead to a decrease in domestic production. This causes 

a political problem just like the one that prevailed in Japan 

and Taiwan (China) during the last three decades. It is still a 

huge problem involving governments, producers, and consumers. 

In Taiwan, as total consumption decreased to a third, production 

decreased to a third as well. The situation in Japan is 

the same. This type of decrease in rice production, which is 

the core of Asian agriculture, is a tremendous setback. The 

drastic plunge in PCRC in Taiwan and Japan is not unique, but 

any Asian country would have to face the same fate if it did 

not take any effective measures. 

Pressure on market prices 

Rice production is quite sensitive to changes in market prices. 

The rice of Thai 5% broken was as low as US$237 per metric 

ton (mt), on a milled basis, in 1993. It increased since then and 

reached $338 in 1996 and remained relatively high until 1998. 

Meanwhile, rice production in Asia increased continuously, 

breaking records every year through 1999, reaching 371 million 

mt. The world total production, reflecting the situation in 

Asia, reached a new record at slightly over 400 million mt in 

1999, the first time in history to reach the 400 million mt level. 

However, the market prices continued to plunge and dropped 

way below $200 at $173 per mt by 2001. The price in 2001 

was the lowest since 1972. Although market prices recovered 

some since then, they are still drifting around $250. 

Accounting for about 90% of world rice production, 

Asian rice producers are quite sensitive to changes in market 

prices, and changes in global production are a reflection of the 

changes in Asian production. Asian rice production decreased 

three years in a row to as little as 344 million mt in 2002, a 

7.3% decrease from the record level in 1999. Asian production 

recovered some in 2003 and 2004. However, the historically 

largest amount recorded in 1999 has not been achieved 

even after five years. During the last five decades, each historical 

production record has been renewed mostly every year 

or after three years as the longest period after the preceding 

record. Nowadays, however, the preceding record has not been 

surpassed even at the fifth year. The very low prices around 

2000 have affected rice growers. In real terms, taking the consumer 

price index (CPI) into consideration, the low prices in 

2001 were practically the lowest in history. The market prices 


kg 

200 

160 

Milled basis 

Japan 

Taiwan (China) 

South Korea 

China 

120 

80 

40 

0 

1961 

1963 

1965 

1967 

1969 

1971 

1973 

1975 

1977 

1979 

1981 

1983 

1985 

1987 

1989 

1991 

1993 

1995 

1997 

1999 

2001 

2003 

2005 

2007 

Year 

Source: S. Ito, World Food Statistics & Graphics (http://worldfood.muses.tottori-u.ac.jp). 

Original data sources: USDA: PS&D View, November 2004; USBC: International Data Base, July 2003. 

Fig. 1. Per capita consumption of rice for Japan, Taiwan (China), South Korea, and China during 1961-2004. 

in 2003 and 2004 recovered some; however, the rise in prices 

is not enough to boost production and set a new record. Since 

world consumption has been greater than production during 

the past few years, the world rice stock has been diminishing 

dramatically. Therefore, market prices should be rising, but 

they have appeared to stop rising since early 2004. 

Consumption-derived supply and forecasts for rice 

A decrease in PCRC indicates weaker demand. The demand 

curve for rice in Asia may be shifting leftward on a per capita 

basis, meaning that rice market prices may not rise enough to 

increase production. Instead, rice production in Asian countries 

may decrease in the future as revealed in Japan and Taiwan 

(China) as well as recently in China. 

Because the power of declining PCRC is so strong, even 

a modeling technique should change from a market-oriented 

equilibrium to a consumption-derived supply model. In fact, 

there are strong indications that rice consumption is not influenced 

much by lower rice prices, although higher rice prices 

tend to affect rice consumption like the example in Japan reported 

by Asahi Newspaper. For any estimate of the rice equilibrium 

model for Asian nations, it is important to develop a 

reasonable rice consumption model. 

Chinese PCRC is currently at 105 kg and is declining at 

approximately 0.5% annually, while the declining rates for Japan 

and Taiwan for the last 45 years are approximately 1.8% 

and 3.5%, respectively. While the FAPRI (2004) estimates rice 

demand and supply for the next decade or so, some researchers 

forecast for China with three scenarios for the next five 

decades using declining rates of 0.33%, a current rate of 1.77% 

applying the Japanese rate, and 3.45% applying the Taiwanese 

rate. According to the results, by 2050 Chinese PCRC would 

become 90.0, 46.0, and 20.8 kg with the three rates, respectively. 

The total rice consumption of China in 2050 would be 

estimated at 127 million, 65 million, and 30 million mt, respectively, 

with the three decreasing rate scenarios vis-à-vis 

the current 135 million mt. Together with other nations’ declining 

PCRC, total rice consumption in the world in 2050 

would increase to only 525 million, 463 million, and 427 million 

mt, respectively, in comparison with the current level of 

412 million mt, using the population estimated by the United 

Nations (2003) to increase to 9 billion by 2050 (Fig. 2). 

Conclusions 

Rice in Asia is the basis for food, culture, and the economy. 

Declining per capita rice consumption in Asia would eventually 

lead to a decline in rice production. Since rice is the major 

crop in Asia, downsizing in rice production would mean a loss 

of agricultural competitiveness as well as less value of assets 

and resources. Asian people heavily depend on agriculture for 

household income. Weak agriculture would create more unemployed 

people and more poverty. The Asian food culture, 

which is based on rice, would also lose ground inside and outside 

Asia. All these losses would be tremendous in Asia. 


1,000 mt 

600,000 

500,000 

World (China rate –0.331% 



400,000 

300,000 

200,000 

100,000 

0 

1960 

1963 

1966 

1969 

1972 

1975 

1978 

1981 

1984 

1987 

1990 

1993 

1996 

1999 

2002 

2005 

2008 

2011 

2014 

2017 

2020 

2023 

2026 

2029 

2032 

2035 

2038 

2041 

2044 

2047 

2050 

Year 

Fig. 2. World rice total consumption, 1960-2050. 

While rice is the major agricultural crop and the major 

source of income and calories for most of the Asian people, 

less production of rice with weak demand would mean that 

Asian people would have a less competitive agriculture relative 

to that of other regions in the world and more serious environmental 

and poverty problems in Asian rural areas. This is 

a critical problem not only for Asian agriculture but also for 

any rice producer in the world. 

It is therefore important for Asian nations to cooperate 

among themselves and cope with the declining per capita rice 

consumption. While more human consumption as table food 

should be promoted, more diversified use of rice should be 

explored even more. 

Note: All the data for rice are on the milled basis provided originally 

from the U.S. Department of Agriculture. The World Food Statistics 

& Graphics (by Ito, http://worldfood.muses.tottori-u.ac.jp) is designed 

using the USDA’s PS&D Online data set. 

References 

Asahi Newspaper: Rice consumption declining consecutively for 9 

months, 9 June 2004. 

FAPRI (Food and Agriculture Policy Research Institute). 2004. 

FAPRI world agricultural outlook 2004: world rice, 

www.fapri.iastate.edu/Outlook2004/PageMker/ 

9_WldRice.pdf. 

Ito S, Wesley E, Peterson F, Grant W. 1989. Rice in Asia: is it becoming 

an inferior good Am. J. Agric. Econ. 71:32-42. 

Ito S. World food statistics and graphics, September 2004: http:// 

worldfood.muses.tottori-u.ac.jp. 

United Nations. 2003. United Nations Expert Meeting on World 

Population in 2003, Department of Economic and Social Affairs, 

Population Division New York, 9 December 2003, 

www.un.org/esa/population/publications/longrange2/ 

longrange2.htm. 

USDA (U.S. Department of Agriculture). 2004. PS&D Online, September 

2004: www.fas.usda.gov/psd/complete_files/ 

default.asp. 

Notes 

Author’s address: Tottori University, Japan, e-mail: 

sito@muses.tottori-u.ac.jp. 


Rice consumption in China: Can China change 

rice consumption from quantity to quality 

Chien Hsiaoping 

Rice is the most important crop in China, with the highest level 

of production. China’s rice production accounts for 28% of 

the world’s total production. China produced 161 million tons 

of rice in 2003, but this figure is relatively low compared with 

past years. The rice grown in China consists of indica and 

japonica varieties. In general, indica varieties are produced 

mainly in the south. Landraces of japonica are produced in the 

north on the lower reaches of the Chang Jiang River, and improved 

varieties of japonica that originated from Japanese rice 

are cultivated mainly in the northeast. In recent years, japonica 

varieties have increased rapidly, accounting for 26% of the 

total production, whereas indica varieties account for nearly 

70%. 

In China, rice has been consumed since ancient times. 

The main staple food in China has been either rice or wheat, 

depending on the suitability of local environments for cultivation 

and the diversity of races and eating habits. Because of 

migration between the southern and northern regions, rice consumption 

is now seen all over the country. It is currently considered 

that 840 million people (65% of the total population) 

eat rice as a staple food. The nationwide per capita rice supply 

has dropped slightly since 1984, when it peaked at 99 kg, to 

91 kg in 2000. 

Trends of rice consumption 

Although per capita rice consumption in China is a statistical 

matter and precise figures have not been obtained, recent documents 

(based on the food balance sheet) issued by the Chinese 

Ministry of Agriculture indicate that polished rice consumption 

is in the 90-kg range. This figure is considerably higher 

than that in Japan (65 kg in 1999). Based on a recent food 

balance sheet for rice (by crop years), 85% of total demand 

was for food consumption (Fig. 1). In addition, yearly seed 

demand has remained at 1.5 to 1.6 million tons, whereas demand 

for rice for industrial processing has increased slightly 

to 1.8 million tons (2001-02). There is a significant loss during 

distribution and storage, but the total loss has decreased to 

some extent for four years (from about 8 million to about 7 

million tons). Feed demand has increased for four years. Its 

share of total demand has increased from 4% to 6%, and consumption 

of rice as animal feed has reached nearly 8 million 

tons. Under the current situation, future rice demand will be 

determined by the rate of decrease in rice consumption and 

the trend of population increase. 

Direct consumption accounted for more than 90% of total 

rice demand at the beginning of the 1970s. Since the latter half 

of the 1970s, it has started to decrease because of the increase 

% 

100 

80 

60 

40 

20 

0 

1998-99 

1999-2000 

Years 

2000-01 

2000-02 

Export 

Loss 

Processing 

Feed 

Seed 

Food 

Fig. 1. Share of rice demand. Source: www.agri.org.cn/ 

analysis/include/4/dmjs.htm, 2002 China agricultural 

development report. 

in choice of other dietary options and increased consumption 

of meat and other foods. According to statistics on rice purchases 

in urban areas, direct rice consumption peaked in the 

mid-1980s, and per capita rice consumption (not including consumption 

at restaurants) has decreased ever since. In 2000, it 

was 46 kg, a decrease of 12% from 1995 figures. In rural areas, 

as estimated from the total consumption of rice and wheat, 

direct rice consumption has decreased slightly (90 kg in 1995 

to 89 kg in 2000). However, the quantitative difference in consumption 

between urban and rural areas is still significant. 

Increased income and rice consumption 

In general, changes in rice consumption are influenced by consumer 

income and price. However, price data on rice are insufficient, 

and rice prices have been under the control of the 

government for a long time because of the importance of rice 

as a staple. In this section, we analyze the relationship between 

income and consumption of rice and other foods. 

We obtained details of income elasticities in the 1980s 

and 1990s from data published by the State Statistics Bureau. 

Income elasticity of rice in urban areas has had negative values 

since the 1980s, indicating that rice is now seen as an inferior 

commodity in these areas (Table 1). On the other hand, in 

the 1980s, the income elasticity of rice in rural areas was 0.15 

(meaning that when income increases by 10%, rice consumption 

rises by 1.5%), demonstrating that rice was still a valued 


Table 1. Income elasticity in staple foods. a 

Food 

Urban area 

Rural area 

1981-91 1992-2000 1981-91 1992-2000 

Rice –0.27 –0.35 0.15 –0.16 

Vegetable oil 0.42 0.33 0.97 0.59 

Pork 0.17 –0.18 0.44 0.55 

Beef/mutton 1.39 (0.03) 0.43 1.23 

Poultry 1.43 0.40 0.97 1.59 

Eggs 0.53 0.58 1.07 1.36 

Marine products (0.06) 0.51 0.79 1.14 

a Numbers in parentheses indicate poor statistical fit. 

commodity. However, in the 1990s, income elasticity of rice 

in rural areas developed negative values, and rice consumption, 

as in urban areas, showed a downward tendency despite 

increased income. 

Food diversification was a factor in this decrease in rice 

consumption, and increased consumption of livestock products 

is considered to be an influencing factor. For the income 

elasticities of all stock farm products and vegetable oil (except 

pork in the 1990s), positive values were observed and 

there was an upward tendency with increasing income, although 

the extent of each fluctuation was different. In urban areas, the 

income elasticities of livestock products were high in the 1980s. 

However, the elasticities declined in the 1990s, and pork in 

particular showed negative values, indicating decreasing consumption. 

We consider that these trends reflect the recent emergence 

of health-conscious attitudes toward lifestyle in urban 

China. In contrast, unlike the trend in urban areas, elasticities 

in rural areas showed positive values. A comparison of values 

in the 1980s and 1990s revealed that the elasticities of livestock 

products (except vegetable oil) in the 1990s were higher 

than those in the 1980s. Values for beef and mutton, poultry, 

eggs, and marine products were greater than 1, indicating that 

these products were regarded as superior goods. Income elasticity 

of pork in rural areas also increased, and a further increase 

in consumption because of rising income is expected. 

From our analysis of the fluctuations in income elasticities in 

urban and rural areas between the two periods, we strongly 

expect that consumption of livestock products will increase in 

the future owing to rises in the income of the rural population 

of 800 million. 

Analysis of the patterns of consumption of staple foods 

by income group revealed that, in urban areas, rice was mainly 

consumed as a staple food. The higher the income was, the 

lower the consumption of rice became. Compared with rice, 

wheat was more likely to be consumed as a processed food. 

We focused on unfilled buns and bread, which were the processed 

wheat products consumed most often. Consumption of 

processed foods such as buns and bread increased as income 

rose, whereas consumption of strong and plain flour decreased 

as income rose. We tried to obtain a simple regression model 

that related income and consumption of various foods. Income 

elasticities of rice by low-, middle-, and high-income groups 

were –0.85, –0.60, and –0.64, respectively; rice was an inferior 

commodity for all groups. For bread, income elasticity 

values were elastic in the middle- and high-income groups (1.15 

and 0.87, respectively). For traditional buns, high income elasticities 

were demonstrated in the low-, middle-, and high-income 

groups (2.16, 1.26, and 0.90, respectively). Because of 

their lower prices compared with those of other processed 

foods, traditional processed foods enjoy a high demand and 

are thus regarded as increasingly superior as income becomes 

lower. In the middle- and high-income groups, increased consumption 

of Western-style foods such as bread was caused by 

a rise in income. On the basis of trends in the purchase of main 

foods in restaurants and office dining rooms, we obtained significant 

elastic values for processed foods in the middle- and 

high-income groups. Therefore, it was evident that preferences 

for processed foods that could save consumers the trouble of 

cooking increased as income rose. 

Also, consumption of livestock products increased as 

income increased. Breakfast for households in urban areas 

tended to become a meal consumed out. In addition, many 

two-income families were characterized by their tendency to 

eat out for lunch. However, with the exception of rice, there 

were no precise statistical data on the consumption of different 

styles of food eaten out. Therefore, we considered that the 

actual amounts consumed in urban areas might be greater than 

those shown in the statistical data obtained. 

Conclusions 

We have reviewed the influence of eating habits and income 

on rice consumption, but there is another influencing factor: 

population. In China, which currently has about 1.3 billion 

people, increased rice consumption because of population 

growth is expected because the population has been increasing 

by nearly 10 million a year, even though the growth rate is 

less than 1%. In addition, there has been a change from quantity 

to quality in terms of consumer demand for rice. The main 

factors driving this change in demand are an improvement in 

the quality of indica rice by establishing a market for brand 

rice, a shift in consumption from indica to japonica rice, and 

the migration of people from rural to urban areas because of 

urbanization. We consider that these trends will not change 

significantly in the future because rice is a highly self-sufficient 

crop in China. In recent years, there have been trends 

toward decreased rice consumption because of increased diversity 

of eating habits, increased demand for high-quality rice, 

and increased sophistication of foods. We have also confirmed 

that demand has increased for japonica rice. The supply of 

rice with improved quality and the production of rice crops 

with a high degree of self-sufficiency are goals for rice consumption 

in China. 


References 

Chien HP. 2003. Rice consumption in China. Agric. Econ. 69:5-12. 

Ministry of Agriculture. 2004. 2003 China agricultural statistical 

material. Beijing (China): China Agriculture Press. 

Sawada M. 1986. The effects of prices, income and household characteristics 

on food demand: a demand system approach. J. 

Rural Econ. 57(4):229-239. 

Wang ML. 2004. On the economy of japonica rice in China. Issues 

Agric. Econ. 2004(4):35-39. 

Notes 

Author’s address: Research Information Division, Japan International 

Research Center for Agricultural Sciences (JIRCAS), e-mail: 

chienp@jircas.affrc.go.jp. 

Surplus rice supply in Asia 

Vichai Sriprasert 

Asia has long been dominant in rice production, consumption, 

and trade. More than 92% of world rice production, 90% of 

rice consumption, and 75% of rice surplus for exports are in 

Asia. The abundant rice supply in Asia explains why 60% of 

the world’s six billion people are concentrated in Asia. Will 

this picture change in the future 

Determinants 

Rice supply is determined by both natural endowments and 

human factors. Water, fertile soil, sunlight, and suitable temperature 

are prerequisites for growing rice. Human factors 

could include particular government policies on free trade versus 

government control, investments in infrastructure in irrigation 

and transportation, and emphasis on technology, for 

example, genetic engineering or traditional biotechnology. Both 

natural and human factors can have a tremendous effect on 

producing rice surpluses or deficits. Population density or human-land 

ratios can also become an overriding determinant. 

Origins 

Even though the definitive origins of rice have yet to be established, 

it is quite reasonable to believe that rice originated in 

Asia. The shear sizes of the Asian population and rice production 

today are good testimonies that rice must have started 

somewhere in Asia. Remains of rice plants from 10,000 B.C. 

were discovered in Spirit Cave on the Thailand-Myanmar border. 

Rice and farming implements dating back at least 8,000 

years have been found along the Yangtze River in China. Rice 

production outside of Asia constitutes only one-ninth of that in 

Asia. Because of the crop’s origins in Asia, rice production, 

consumption, and trade are all concentrated in Asia. 

Trade 

International trade in rice, in the modern era, started to develop 

about 150 years ago in Southeast Asia. The West, particularly 

the British, extended its colonial campaign with trade 

and threats in the East around the mid-1800s. China lost Hong 

Kong to the British after a humiliating defeat in the Opium 

War. Hong Kong was supposedly on a long-term lease, but the 

British never did pay any rent to China over 155 years (from 

1842 to 1997). India also lost its independence in 1857 after 

conflicts with East India Company. The U.S. naval officer 

Matthew C. Perry forced Japan to trade with the West in 1854. 

Next, the British annexed Burma (now Myanmar) in the 

west, and forced Thailand to open up for trade, particularly in 

rice. The French controlled Vietnam and Cambodia to the east. 

The British invested heavily in irrigation and transportation in 

Burma, while the French did the same in Indochina. Thailand 

also invested a moderate amount in similar infrastructure. Indians 

poured into Burma, while the Chinese immigrated into 

Vietnam and Thailand. Rice production in excess of the domestic 

requirement was stimulated by new demand for international 

trade. Immigrant entrepreneurs brought in their trading 

and managerial skills together with their initial capital. The 

British and the Indians helped Burma to export rice to Europe, 

India, and Sri Lanka, while the Chinese helped Thailand and 

Vietnam to export to Hong Kong and Singapore, and the Vietnamese 

were expanding exports to Indonesia and the Philippines. 

The rice economy of Southeast Asia flourished despite 

fears that foreigners would exploit their countries. Prior to the 

Second World War, Burma managed to export more than 3 

million tons annually. Thailand and Vietnam took turns being 

the second- or third-largest exporters at around 1 million tons. 

Foreign exploitation did indeed turn out to be a blessing in 

disguise. 

Economic systems 

Nationalism and independence gained after the Second World 

War did not help Burma to continue producing huge rice surpluses 

for export as when the country was operating under a 

market economy during the long period of British rule. 

Rice exports reached 2.1 million tons as early as 1900, 

and peaked at 3.4 million tons in 1934. In 1948, the year of the 


country’s complete independence, rice exports dropped to 1.2 

million tons. After 1965, rice exports stayed below 1 million 

tons for most years until today. 

The ravages of the Second World War, and the adoption 

of socialism, that is, nationalizing everything, including the 

rice sector, together with a policy to keep rice prices very low, 

destroyed farmers’ incentives to produce surplus rice for export. 

Similar changes from a market economy to a centrally 

planned economy occurred in Vietnam after the Second World 

War, and the country suffered the same consequences: rice 

exports dissipated and the country became a large rice importer 

of 1 million tons instead in 1969. However, Vietnam 

was quick to change toward a more market economy in 1986 

under the Doi Moi (renovation policy). In a few years, Vietnam 

became the second-largest rice exporter after Thailand. 

Unlike Burma and Vietnam, Thailand started out with a 

free market economy and keeps the same system today. Thailand 

surpassed both Vietnam and Burma and emerged as the 

world’s largest rice exporter for more than two decades in a 

row. From 1966 to 1982, the U.S. and Thailand took turns 

being the largest rice exporter, but China on some occasions 

also occupied the top position. 

Pressure from population density 

Rice exporters are countries that are generally endowed with 

favorable human-land ratios; they also must have a proper 

economic system that provides enough incentives for rice producers. 

Population density per km 2 for the major rice exporters 

discussed here are Myanmar (62), Thailand (121), and Vietnam 

(246). In contrast, traditional rice importers prior to the 

Second World War with their respective population density 

were Hong Kong (6,688), Singapore (6,430), Bangladesh 

(926), Taiwan (627), South Korea (491), Japan (336), India 

(318), Sri Lanka (298), the Philippines (282), China (134), 

and Indonesia (121). 

The importers that have limited land to support a large 

population must overcome their scarce natural endowments 

with technology and management. Japan, China, and India are 

good examples of success stories. 

Japan 

Rice imports into Japan reached 1.4 million tons in 1919 and 

2.3 million tons in 1938, and then declined to 1.1 million tons 

in 1943 and less than 1 million tons since 1956 onward. Up 

until 1945, Japan promoted rice production and imported 

mostly from its two colonies, Korea and Taiwan. Rice consumption 

in Japan peaked in 1962 at 118 kg per capita, but is 

only 61 kg per capita now. Japan has always emphasized high 

technology and high rice prices to their producers and at the 

same time tightly controlled rice production, imports, and distribution. 

This self-sufficiency policy has been so successful 

that Japan today is in the process of deregulating the rice industry 

and using about US$3 billion a year to reduce rice growing 

and allow more market forces to determine rice prices and 

distribution. Nonetheless, Japan is among the pioneers in rice 

genomic research together with China and the U.S. Field yields 

and milling yields for rice are among the highest in the world. 

But the policy to keep high rice prices for farmers precludes 

Japan from becoming a major rice exporter because Japanese 

rice was 13 times higher in price than U.S. rice in 2000. 

China 

China has a population of 1.3 billion to feed, the largest in the 

world. The famous Yellow River has run dry nearly every year 

since 1972. Because of population pressure, the government 

has managed to stay at the frontier of knowledge in rice production. 

China had a parallel development of its own during 

the Green Revolution. It led the world in hybrid rice, and now 

keeps abreast of rice genomics and genetic engineering. Its 

average field yield of 6.2 t ha –1 is way above the world’s average 

yield of 3.9 t ha –1 . 

In 1923, China’s rice imports reached 1.4 million tons; 

by 1956, it could export 1.1 million tons. In 1973, it exported 

2.6 million tons, and by 2000 exported 3.1 million tons. A few 

times in the 1960s and 1970s, China was the largest rice exporter 

in the world. 

India 

India used to import a lot of rice: 2.5 million tons in 1915, 3.7 

million tons in 1919, a peak of 4.4 million tons in 1934, and a 

gradual drop to 1 million tons in 1941. India became self-sufficient 

in 1977. Exports increased from 0.09 million tons in 

1994 to 4.9 million tons in 1995 and 6.7 million tons in 2002. 

Technology and the market economy are the key 

The world was not comfortable with food security in the early 

1960s as rice production could not catch up with population 

growth. Like a miracle, IRRI introduced high-yielding varieties, 

which were of short maturity, and two to three crops could 

be planted in a year. These varieties were also responsive to 

fertilizer. That scientific breakthrough was called the Green 

Revolution. A food crisis has been averted up until today. However, 

the growth of rice production seems to have leveled off 

and again there is concern that population may outgrow rice 

production in the decades ahead. In the last few years, production 

has been running behind consumption. This situation must 

be corrected because it cannot be sustained. 

The Green Revolution helped farmers to grow more rice 

and real prices therefore came down. Roughly, from 1953 to 

1999, rice production increased threefold, but prices dropped 

to one-fourth of those of the beginning period. 

Under such circumstances, a few observations are worth 

pointing out. Asia and Oceania’s shares of rice imports decreased 

from 69% in 1961 to 38% in 2002. This implies that 

those Asian importers became more self-sufficient and needed 

to import less. 

On the other hand, Asian exporters increased their shares 

from 70% to 75% during the same period. However, poor 


Burmese farmers, even with limited alternatives, did not produce 

any more rice for export. In contrast, its neighbor farmers 

in Thailand and Vietnam, under a market economy, were 

motivated to grow more and more. 

Will Asia continue to provide rice surpluses 

If we look at Thailand as a case study, we can come up with 

some interesting insights. First of all, Thailand enjoys the most 

favorable human-land ratio. Second, the farmers are very skillful 

in growing rice; they can do the best with what little they 

have, and survive. Third, the traders, millers, and exporters 

who are serving the farmers are also very skillful and competitive. 

They are willing to buy and sell rice even without a margin. 

There are plenty of such traders around, but they do not 

get any recognition for such a good deed. Society seems to 

prefer traders that charge high fees for their services! Fourth, 

Thai bankers are willing to finance such a competitive rice 

business, which has a good track record on paying interest. 

The principal has seldom, if ever, been recalled. 

Since all the major players (farmers, millers, and exporters) 

are all very lean, the rice industry just grows faster and 

faster. Rice has been freely exported from Thailand for about 

150 years. It took the first 125 years to reach an annual export 

level of only 2 million tons. The level of 5 million tons took 

only 25 years. The export pace for each additional 1 million 

tons became shorter and shorter, from over 60 years to 5 and 

now moving toward about 1 year! Our annual rice exports are 

likely to jump from 7 to 9 million tons in just four years at the 

end of 2004! This is a miracle of the free market. 

This outstanding export performance was achieved under 

some very unfavorable circumstances. Thai field yields 

are the lowest among all our competitors. Our government is 

intervening in the market to raise paddy prices, while India is 

poised to provide a rice export subsidy again. Japan, the EU, 

and the U.S. continue to provide heavy domestic support to 

their rice farmers. 

If Thailand were to have the same percentage of irrigation 

area as China or Japan, our field yields would be twoand-a-half 

times more than what they are today. Thailand alone 

can meet global rice imports! Do we have enough water to get 

the job done Yes, if we can obtain some water from the 

Salween River, which simply flows to the sea without being 

used much by anyone. 

The Green Revolution got the world out of the food crisis of 

the 1960s. Biotechnology, particularly genetic engineering, may 

change the landscape that we can see today. It is entirely possible 

that genetically engineered rice seeds will thrive in Africa, 

where the population and rice imports have been growing 

faster than in any other regions in the world. Africa’s share 

of the world population jumped from 9% in 1950 to 13% in 

2000 and is projected to reach 20% in 2050. Africa’s share of 

world rice imports was only 8% in 1961; it became 25% in 

2002. 

Just take Nigeria as an example. Its population is 120 

million. Rice consumption per capita is only 23 kg versus 58 

kg for the average world consumption. Annual rice imports 

are about the same as the domestic production of 2 million 

tons of milled rice. The f.o.b. per metric ton of Thai 100% 

parboiled is about US$265, freight to Lagos is $85, and the 

import duty is in excess of 100% of cost and freight. That means 

that any rice produced locally will have an advantage or price 

protection against imported rice of ($265 + $85) + $85 = $435 

per metric ton! Under normal circumstances as in Thailand, 

rice production and the rice industry in Nigeria ought to explode! 

Asia, and particularly Thailand, will continue to generate 

rice exports to Africa only because Africa does not take 

advantage of its potential to substitute rice imports. 

Myanmar can regain her prewar prominence in rice similar 

to what Vietnam did. It is the miracle of the free market 

economy that Myanmar must turn to eventually to prevail. 

The warming of Earth may hurt rice yields and reduce 

Asian rice area in the vast lowland delta areas because of higher 

sea levels. However, advances in biotechnology, particularly 

in genomics, may overcome many if not all of the constraints 

we are aware of today. Together with freer trade in the world, 

if it ever really comes to fruit, rice production and trade will 

also change much beyond our imagination. It is quite certain 

that under a recurring of the oil crisis, more crops will be grown 

for the production of gasohol in competition with grain grown 

for food. 

Notes 

Author’s address: President of Rice Exporters Association (Thailand), 

e-mail: Riceland@Ricesources.com. 

Acknowledgments: Various facts and statistics are from the following 

sources: U.S. Bureau of the Census, International Data 

Base: FAOSTAT; Rice almanac, IRRI 2002; The economy of 

Asia: R. Barker, R.W. Herdt and B. Rose, 1985; The rice 

economy of monsoon Asia: V.C. Wickizer and M.K. Bennett, 

1921. 

Any surprises 


Review of existing global rice market models 

Eric J. Wailes 

Numerous models of the global rice market have been developed 

and are used by university and governmental organizations 

and agencies. This paper discusses several of the key 

model frameworks for which documentation and projections 

exist. An overview of the various models in terms of type and 

scope is presented. Comparisons are made in terms of model 

structure and differences in the baseline development process, 

model output, and projections. Differences in country/regional 

coverage, product disaggregation, time period of projections, 

and the purpose of the various models are noted. The challenges 

and prospects for improving quantitative modeling of 

the global rice economy are discussed. 

The need for quantitative models 

Quantitative models of the world rice market are needed for 

three alternative analytical needs (Liu and Seeley 1987). First, 

we need models to use for short-term impact analysis. These 

short-term models must be relatively flexible to address issuespecific 

policy and market questions, since we never know in 

advance what the particular problem, policy, or market issue 

will be. Second, we need models that are capable of intermediate-term 

production, consumption, trade, and price forecasting. 

Ideally, these models should be multicommodity and capable 

of accounting for substitution effects. It is also extremely 

useful if there is a spatial dimension that can account for trade 

flows. The third use we make of market trade models is for 

longer-term policy analysis. The longer-term models typically 

are used to generate baseline projections against which policy 

alternatives can be compared. Typically, the longer-term models 

are multicommodity and have a time horizon of ten years 

or longer for the projection period. They tend to be more complex 

and useful for generating analyses that evaluate changes 

in market structure and policy regimes. Most of these models 

are partial equilibrium that focus on a broad set of commodities 

and countries but take as exogenous variables income, 

exchange rates, interest rates, employment, and other factor 

markets. Computable general equilibrium (CGE) models such 

as the Global Trade Analysis Project (GTAP) have been used 

extensively by the World Bank and the WTO Secretariat for 

analyses of the Uruguay Round and more recently of the Millennium 

(Doha) Round of WTO negotiations (Hertel 1997, 

World Bank 2003). The macro variables and factor markets 

are endogenized in the CGE model framework. However, because 

of the large data and parameter requirements of the GTAP 

model and other CGE frameworks, there is considerable sacrifice 

in terms of reflecting current market and policy conditions. 

At present, the GTAP model is based on 1997 data and 

will be updated to 2001 data only by spring 2005. 

Overview of quantitative models 

This paper will discuss primarily the third longer-term category 

of partial equilibrium quantitative models. Models of 

the third type, which are widely recognized, include the World 

Food Model (WFM) of FAO, the IMPACT model of the International 

Food Policy Research Institute (IFPRI), the AGLINK 

model of the Organization for Economic Cooperation and 

Development (OECD), the Country-Commodity Linked System 

(CCLS) of the United States Department of Agriculture 

(USDA), and the Food and Agricultural Policy Research Institute 

(FAPRI) model of a consortium of U.S. universities. 

The FAO WFM is no longer maintained. The WFM was 

a large multicommodity and multicountry model used to examine 

resource, policy, and market issues. Currently, within 

FAO, two global modeling efforts are in a developmental process, 

including the @2030 model, which is more aggregated 

by country/region than the previous World Food Model (FAO 

2003). The second modeling effort within FAO is known as 

COSIMO, which will be a multicommodity, multicountry model 

based on the OECD AGLINK framework but with much greater 

disaggregation by countries. The IMPACT model maintained 

by IFPRI is an analytical framework used to examine alternative 

scenarios for global food demand, supply, and trade. IM- 

PACT covers 16 commodities, including rice, and covers 36 

countries and regions, accounting for essentially all of the 

world’s food production and consumption. It is structured as a 

set of country or regional submodels that are linked through 

trade and world reference prices that clear international markets. 

Unlike the @2030 and IMPACT models, the AGLINK, 

CCLS, and FAPRI models are updated at least annually through 

a process known as a reference scenario or baseline exercise, 

which is then published annually by each of the respective institutions 

as a 10-year global agricultural outlook. In this paper, 

comparisons of the AGLINK, CCLS, and FAPRI rice 

models are made first by providing a brief description of each 

model with respect to model type, scope, model structure, and 

reference scenario development, including the use of extramodel 

information. This is followed by some comparisons of 

output and projections. Finally, challenges and prospects for 

improving quantitative global rice modeling frameworks will 

be discussed. 

The AGLINK model maintained by the OECD is a set 

of country modules of OECD member nations and selected 

nonmember nations. As such, the model links the country modules 

through trade and international reference prices. The rice 

component of each country model therefore constitutes the basis 

for representation of the global rice model. Country models 

include major rice importers and exporters among the OECD 


members. Non-OECD member models for rice are maintained 

for other major rice importers and exporters. The international 

reference rice price is the Thai 100%B long-grain price. The 

model aggregates rice trade across quality and types and is 

nonspatial. Domestic rice prices are specified through price 

transmission equations. The model is dynamic and policy-specific 

with respect to domestic support and trade interventions. 

The model structure is based on an elasticities framework. The 

reference scenario (baseline) projection process is initiated 

annually with a questionnaire sent to member and nonmember 

nations to obtain updated data and policy information. This 

provides an empirical basis to evaluate an initial baseline developed 

by the Secretariat. Revisions are made and then submitted 

for review to the Working Party on Cereals. This baseline 

is next evaluated for sensitivity to key macro and policy assumptions 

and finally is presented as a preliminary outlook 

draft for consideration by the Working Party on Agricultural 

Policies and Markets of the Committee for Agriculture. This 

iterative review process implies that the final baseline projections 

are heavily conditioned by representatives from member 

and participating nonmember nations. The review process ensures 

that the final estimates reflect the input of the country 

and commodity experts; however, the Secretariat takes full 

responsibility for the final published baseline projections 

(OECD 2004). 

The USDA Country-Commodity Linked System (CCLS) 

is maintained at the Economic Research Service. This framework 

includes a set of 42 foreign country or regional models 

and the Food and Agriculture Policy Simulator (FAPSIM) 

model of U.S. agriculture. The country models estimate production, 

consumption, stocks, and prices. Country models are 

integrated through trade and international reference prices 

through the LINKER module. The country models are dynamic 

and reflect policies and institutional behavior such as tariffs, 

subsidies, and trade restrictions. A rest-of-the-world model 

handles any excluded country-commodity coverage. 

The country models are structured to account, where 

appropriate, for both imports and exports of a given commodity. 

The international reference price is the Thai 100%B longgrain 

price. The trade framework aggregates rice across quality 

and types and is nonspatial. The baseline is generated 

through a Delphi approach, similar to the AGLINK process. 

As such, the projections reflect the composite of model results 

and judgment-based analysis by country and commodity specialists 

(USDA 2004). 

The Arkansas Global Rice Model (AGRM) is the international 

rice model framework used in the FAPRI baseline 

and is maintained at the University of Arkansas. The AGRM is 

based on a multicountry econometric framework. The country 

models consist of a supply sector, a demand sector, and a set 

of trade and price linkage equations. All these equations are 

estimated using econometric techniques. The model is dynamic 

and nonspatial in trade. Individual country models are then 

linked through net trade—a specification that highlights the 

interdependence of countries in the world rice economy. 

Simulation is conducted for the purpose of the baseline 

projection and policy analysis. The Thai long-grain price is 

used to clear the international long-grain rice market and the 

California ex-mill medium-grain price is used to clear the international 

medium-/short-grain rice market. Projections include 

national levels of production (area harvested and yields), 

consumption, net trade, stocks, and prices. 

Government interventions to the extent possible are explicitly 

reflected in the model’s structure. These policies are 

incorporated in the model’s supply, demand, export (or import), 

stocks, and price transmission equations, and are thus 

implicitly reflected in the model solution. Major importing and 

exporting countries or regions are explicitly included in the 

AGRM model. All other countries are included in the rest-ofthe-world 

(ROW) region (FAPRI 2004). 

Unlike the AGLINK and CCLS frameworks, which are 

country models with commodity submodels, the FAPRI framework 

is a set of commodity models with country submodels. 

Country consistency and coherence are maintained through the 

FAPRI baseline meetings and review. The baseline process 

begins with an initial “meltdown” where macroeconomic and 

policy assumptions are shared across the various commodity 

models. The meltdown process involves an iterative process 

of commodity model simulations and calibrations based on 

sharing cross-commodity estimates of prices, production, consumption, 

stocks, and trade. Following the meltdown, a preliminary 

set of baseline projections is reviewed by commodity 

and country experts drawn from national and international 

government agencies and industry organizations. This review 

is then used to revise and recalibrate the models to generate a 

final baseline publication. 

In addition to the AGRM framework, the University of 

Arkansas maintains a spatial equilibrium model framework of 

the global rice economy known as RICEFLOW. Unlike the 

previous three frameworks, RICEFLOW is distinctively different 

in several important respects. First, it is a static model 

that solves to optimize welfare quasi-rents in international rice 

trade subject to transaction costs, including transportation, 

national policy interventions, and bilateral, regional, and multilateral 

trade agreements. A second important difference is 

that RICEFLOW generates spatial rice trade estimates. Finally, 

and perhaps most importantly, RICEFLOW is disaggregated 

by rice type, degree of processing, and quality. This framework 

is flexible in terms of country aggregation, with as many 

as 49 rice importers and 13 rice-exporting countries or regions. 

The baseline is generated to approximate actual trade flows 

for a given base year. Adjustments to calibrate the model to 

the base year are made primarily through modifications in the 

transaction cost constraints. Policy and market analysis is conducted 

in a comparative static (equilibrium displacement) 

framework in which policies and other constraints are relaxed. 

The strength of this model is the country and commodity space 

disaggregation relative to the AGLINK, CCLS, and AGRM 

models. It is particularly useful to examine trade policy issues 

such as regional trade agreements, tariff escalations, preferences, 

and nontrade barriers (Wailes 2004). The major limita- 


tion is its inability to track dynamic adjustments to policy and 

market changes. 

Table 1 provides a summary of the key differences and 

similarities of the four global rice models discussed. 

Model projection comparisons 

As noted above, both model structure and the process of developing 

a baseline for the AGLINK, USDA, and AGRM/ 

FAPRI models are similar. All three frameworks are used to 

generate annual 10-year outlook projections. Table 2 presents 

the 2004 projections for global rice production, consumption, 

and the international long-grain reference price, Thai 100%B 

long grain. 

The models reflect some significant differences in their 

projection of global rice consumption. The USDA model 

projects an increase of 31 million mt over the next 10 years 

compared with the AGLINK, with a projected increase of 33 

million mt, and the AGRM model, with a projected increase 

of 36 million mt. Increases in production are projected in all 

three frameworks to exceed the consumption growth, suggesting 

an increase in global rice stocks. The AGLINK model 

projects a 39 million mt increase in production, 6 million mt 

greater than consumption growth. The USDA baseline production 

increases by 41 million mt over the next 10 years, 10 

million mt greater than the projected consumption increase. 

Table 1. Summary of model characteristics for global rice quantitative 

models. 

Model characteristics AGLINK CCLS AGRM RICEFLOW 

Dynamic Yes Yes Yes No 

Countries/regions 13 31 31 62 

Product disaggregation No No L/M High 

Extra-model info High High High Low 

Spatial trade flow No No No Yes 

The AGRM framework shows increases in production of 41 

million mt also, but only 5 million mt greater than the projected 

consumption increase. 

More striking differences are reflected in the price projections. 

The USDA model clearly projects a significantly lower 

price path than the AGLINK and AGRM price projections. 

While consistent with the relatively lower consumption and 

higher surplus production projections, the USDA projected 

price by 2013 is 27% lower than the AGRM projection and 

17% lower than the AGLINK estimate. 


Quantitative models of the global rice market are typically part 

of a multicommodity framework. This characteristic is critically 

important as the dynamics of demand and supply are reflected 

in a framework that includes income and relative price 

effects. An important challenge in quantitative modeling of 

the global rice economy is the need to be able to capture the 

heterogeneous and creative nature of government intervention 

in this market. It is difficult to monitor changes in policy regimes 

as countries are slow to notify the WTO market interventions. 

As policy regimes change, for example, in domestic 

support from coupled to decoupled payments, it is difficult to 

obtain statistical estimates of program effects given the unavailability 

of time-series data. 

Disaggregated commodity space and inclusion of key 

exporting and importing countries/regions are highly desirable. 

While comparative static frameworks are useful, it is also valuable 

to understand how markets adjust from one period to the 

next. Finally, it is especially desirable for models to simulate 

the rice economy in a general equilibrium framework, especially 

for countries in which the rice economy is dominant in 

agricultural production and food consumption. Unfortunately, 

all of these desirable characteristics are difficult to capture in 

any one model. Data constraints, costs of developing and main- 

Table 2. Global rice production, consumption, and price projections. 

Item 

Year 

2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 

OECD 

Production (million mt) 393 409 417 422 425 429 434 438 441 444 448 

Consumption (million mt) 415 415 418 421 425 428 434 438 441 444 448 

Prices (Thai milled 100% B, US$) 204 232 252 264 275 286 291 299 303 310 316 

USDA 




AGRM/FAPRI 




Sources: OECD (2004), USDA (2004), and FAPRI (2004). 


taining models, and access to extra-model information ultimately 

lead to trade-offs in the scope and design of quantitative 

models. Critical to understanding model results is knowing 

not only what the model includes but also what the model 

excludes in its analytical framework. Fortunately, through the 

development of information technology, modelers and users 

have access to alternative quantitative model results. This means 

that decisions based on model output can be evaluated and incorporated 

into decision-making without unrealistic reliance 

upon any single model. Internal and external model validation 

must be an important dimension of quantitative models. Models 

are rarely correct in projections but achieving success in 

direction and magnitude of market and policy effects is the 

standard against which all models must be judged. 

References 

FAO. 2003. World agriculture: towards 2015/2030: an FAO perspective. 

Bruinsma J, editor. London (UK): Earthscan. 

FAPRI. 2004. FAPRI 2004 U.S. and world agricultural outlook. Staff 

Report 1-04. Ames, Iowa (USA): Iowa State University. 

Hertel TW, editor. 1997. Global trade analysis: modeling and applications. 

Cambridge (UK): Cambridge University Press. 

Liu K, Seeley R, editors. 1987. The International Agricultural Research 

Consortium: agricultural trade modeling, the state of 

practice and research issues. Staff Report No. AGES861215. 

International Economics Division, Economic Research Service, 

U.S. Department of Agriculture. 

OECD. 2004. OECD agricultural outlook 2004-2013. Paris. 

USDA. 2004. USDA agricultural baseline projections to 2013. Staff 

Report WAOB-2004-1. Office of the Chief Economist, USDA. 

Prepared by the Interagency Agricultural Projections Committee. 

Washington, D.C. 

Wailes EJ. 2004. Implications of the WTO Doha Round for the rice 

sector. Proceedings, FAO Rice Conference, Rome, Italy. 

World Bank. 2003. Global economic prospects 2004. In: Realizing 

the development promise of the Doha Agenda. Washington, 

D.C. (USA): World Bank. 

Notes 

Author’s address: 217 Agriculture Building, Department of Agricultural 

Economics and Agribusiness, Division of Agriculture, 

University of Arkansas, Fayetteville, Arkansas 72701, 

USA, e-mail: ewailes@uark.edu. 

Household rice consumption in Japan: quantity and price by 

income while controlling for household types 

Kimiko Ishibashi, Yoshinobu Kono, and Yuuji Ooura 

As food consumption has become diversified, rice consumption 

has gradually declined in Japan. Individual consumption 

of food products varies significantly by age (Ishibashi 1997), 

indicating the importance of the age factor in food demand 

analysis. 

Household income is highly associated with the age of 

the household head under the seniority system in the Japanese 

labor market. Real income-consumption profiles are likely obscured 

by the fact that Japanese older people (old in age and 

generation as well) consume more rice on average than (currently) 

younger ones, and that older children generally eat more 

rice than smaller ones. 

Based on the analysis of the panel data of household 

expenditure surveys, distinct differences are detected in quantity 

and price paid for rice among major household types. We 

then tried to estimate the age factor-free income elasticities of 

demand for rice cross-sectionally by major household types. 

Data 

The Japanese government Bureau of Statistics (JBC) conducts 

the Family Income and Expenditure Survey (FIES) monthly 

with approximately 8,000 households across the nation on a 

continuous basis, with one-sixth of them renewed every month. 

The age composition of family members, annual income, and 

monthly purchases of various products and services of each 

household are recorded in the data. We analyzed the panel data 

of approximately 96,000 households each year for 8 selected 

years from 1987 to 2001 (Courtesy of the JBC). 

Methodology 

We first classified the panel by the age of the household head 

and household size, with the age of family members taken into 

consideration. 

First, the amount of rice purchased and average rice price 

paid were examined by each household type listed below, then 

household purchases were regressed against the household income 

as follows, using the aggregate data grouped by every 

0.5 million yen in household annual income. 

ln Q = a + b ln Y 

where Q = household monthly purchases (= consumption), Y 

= household annual income, a = intercept, and b = coefficients 

for income. 


Amount of rice purchased (kg mo –1 ) 

16 

14 

A couple in their 30s with two children under 10 years of age 

A couple in their 40s with two children aged 10–20 

A couple in their 50s with a child in his/her 20s 

A couple in their 60s without dependents 

12 

10 

8 

6 

4 

3,500 4,000 4,500 5,000 5,500 6,000 

Yen 

Fig. 1. Changes in consumption of rice with respect to quantity and price paid by selected 

household types. 

Results 

Major household types in Japan 

According to the JBC, the sample household is selected in proportion 

to the size of the household types in Japan. The major 

household types selected for our analysis were 

Households in which a couple are in their 30s with 

two children under 10 years of age, 5,663 in 1987 

and 3,979 in 2001 in total number, for example. 

Households in which a couple are in their 40s with 

two children age 10 to 20, 4,584 in 1987 and 2,880 in 

2001, respectively. 

Households in which a couple are in their 50s with a 

child in his/her 20s, 2,840 in 1987 and 3,099 in 2001, 

respectively. 

Households in which a couple are in their 60s and 

have no dependents, 3,761 in 1987 and 7,449 in 2001, 

respectively. 

Rice consumption by household types 

Figure 1 clearly shows that changes in consumption of rice 

with respect to quantity and price varied greatly among the 

selected household types from 1987 to 2001. The average price 

paid declined consistently from 1987 to 2001 (downward slope 

along the X-axis) in every household type, with the younger 

households showing greater declines. At the same time, a decrease 

in quantity purchased occurred for all household types 

(downward slope along the Y-axis). The households of a couple 

in their 40s with two children from age 10 to 20, in particular, 

demonstrated the largest drop from approximately 15 kg per 

month in 1987 to approximately 9 kg per month in 2001 (the 

topmost line in Fig. 1). 

Figure 1 also shows that the households in which a couple 

were in their 50s with a child in his/her 20s and those of a 

couple in their 60s without dependents share strikingly similar 

tendencies in consumption changes, with the two lines overlapping 

over time, despite big differences in income between 

the two categories, that is, the former register more than 10 

million yen per year for most years, whereas the latter category 

has 2.0 to 4.0 million yen. Regarding the average prices 

paid, those in their 50s paid higher prices (rightward horizontally) 

than those in their 60s in every year. 

Within the same household type, that is, when the age of 

family members is controlled, clear tendencies are observed 

that the greater the household income, the smaller the consumption 

and the higher the price paid. For households of a 

couple in their 40s with two children from 10 to 20, for example, 

those with an annual income of 2.0 to 4.0 million yen 

purchased 16.8 kg of rice per month on average at the average 

price of 5,257 yen per 10 kg in 1987, whereas those with income 

over 10 million yen purchased 12.3 kg at 5,897 yen in 

the same year. The corresponding numbers for 1996, for example, 

are 12.7 kg at 4,815 yen versus 8.9 kg at 5,099 yen, 

and for 2001 are 8.9 kg at 3,743 yen versus 7.5 kg at 4,430 

yen, respectively. 

Income elasticities for rice 

Most previous studies, using cross-sectional data, suggest that 

income elasticities for rice should be positive in sign (Appendix 

table 4 in FIES annual report, various issues). To determine 

the pure income effects on rice consumption, we ran a 

regression by household types. 

Table 1 shows estimates of income elasticities for consumption 

of rice by household type, using panel data aggregated 

by income groups in 0.5 million yen. Both the households 

in which a couple are in their 30s with two children under 

10 years of age and those of a couple in their 40s with two 

children from 10 to 20 are estimated to carry negative income 

elasticities, –0.1 to –0.2, with meaningful t-values for most of 

our survey years. When these two types are grouped together, 

however, we get positive income elasticities, 0.2 to 0.3, with 

significant t-values and much greater R 2 s. This might suggest 

that household purchases of rice tend to increase with the growing 

up of children from under age 10 to teens, whereas the 

household income also tends to increase in accordance with 

the aging of household heads from their 30s to their 40s. The 


Table 1. Estimates of income elasticities for rice by household types, selected years from 1984 to 2001, using panel data 

aggregated by income groups (in 0.5 million yen). a 

Household types 

A couple in their 30s with two A couple in their 40s with two Two types aggregated 

children under 10 years of age children from 10 to 20 

b-estimated Adj. R 2 t-value b-estimated Adj. R 2 t-value b-estimated Adj. R 2 t-value 

–0.21 0.44 –3.89 –0.08 0.11 –1.94 0.15 0.19 2.63 

–0.14 0.11 –1.91 –0.21 0.46 –4.68 0.15 0.27 3.20 

–0.13 0.12 –1.93 –0.04 –0.03 –0.59 0.32 0.52 5.31 

–0.08 0.02 –1.21 –0.16 0.15 –2.28 0.24 0.32 3.60 

–0.17 0.06 –1.55 –0.18 0.13 –2.17 0.22 0.21 2.73 

0.15 0.13 2.00 –0.16 0.07 –1.58 0.29 0.32 3.57 

–0.11 0.02 –1.24 0.19 0.14 2.18 0.34 0.48 4.74 

0.02 –0.04 0.22 0.04 –0.04 0.30 0.29 0.26 3.11 

–0.18 0.11 –2.00 –0.18 0.07 –1.67 0.06 –0.01 0.91 

a 

Double log model used: ln Q = a + b ln Y, where Q = household monthly purchases and Y = household annual income. Glutinous and nonglutinous rice for 

2001, nonglutinous rice for other years. 

above results may strongly suggest a need for classifying households 

by the age of family members if we wish to determine 

the pure income effects on rice consumption in Japan using 

cross-sectional data in particular. 

Conclusions 

Tentative conclusions from our investigations are as follows: 

1. The household quantities of rice purchased and prices 

paid differ, and the patterns of change in consumption 

over time also vary significantly by types of 

household. 

2. When the age of family members and the household 

type are controlled, the households with a larger income 

are found to purchase a smaller amount of rice, 

and to pay higher prices. 

3. The cross-sectional income elasticities for rice are estimated 

to be negative in sign and significantly different 

from zero in Japan, opposed to most previous 


studies, including the government FIES annual reports. 

Our findings, as shown in Figure 1, suggest that age effects 

should overwhelm income effects in rice consumption 

and prices paid are found to differ significantly by age of household 

head and household income. This should be taken to imply 

the importance of incorporating the age factor into analysis 

to better comprehend Japanese rice consumption, including 

future projections. 

References 

Ishibashi K. 1997. Changes in Japanese dietary patterns by age. Is 

the “Japanese style” disappearing Natl. Agric. Res. Ctr. Rural 

Econ. Res. 36:17-31. 

Notes 

Authors’ address: National Agricultural Research Center, Kannondai 

3-1-1, Tsukuba, Ibaraki, Japan. 

This session dealt with the international rice market situation, 

with emphasis on Asian countries. The main objectives were to 

obtain common understanding on past trends and future prospects 

in rice supply and demand among researchers by exchanging 

views, as well as to identify research priorities in this field. 

These include the identification of emerging factors influencing 

supply and demand of rice and the improvement of analytical 

tools such as econometric models. 

In addition to day-to-day market fluctuations, the supply 

and demand situation structurally and fundamentally shifts according 

to economic and social environmental changes. Agricul- 

tural research, especially the development of technologies, should 

be consistent with these structural shifts, so that research results 

will be relevant. This session therefore attracted a large 

audience consisting of both economists and natural scientists. 

Regarding the demand side, we are encountering epochmarking 

phenomena such as diversification of the rice-based dietary 

pattern in Asian countries, the increase in processed and 

nonfood use for rice, and quality differentiation in the rice market. 

For the supply side, skepticism on yield potential prevails, as 

water use and chemical inputs are no longer unlimited. The cost 

of production tends to increase in many countries. In addition, 


environmental concerns and recognition of the external economics 

of paddy farming lead to the introduction of new types of 

support policies. 

The market itself also confronts drastic changes. Information 

technology accelerates the flows of market information worldwide. 

New market players appeared locally and globally through 

both vertical market integration and global trade liberalization. 

Market volatility seems to have become more evident, and food 

security issues are often mentioned in the context of price fluctuation 

as well as in the lack of market access. Our session tried 

to elucidate these changes and future prospects for rice. 

The session was co-chaired by O. Koyama (JIRCAS) and D. 

Dawe (IRRI). Five speakers made presentations. C. Calpe (FAO) 

covered the recent changes in international rice trade and pointed 

out that the current rice market is not as thin, segregated, or 

protected as it used to be. The disaggregated trade statistics 

shown by quality and variety classes were examined. Some arguments 

were made whether rice was becoming a normal trade 

good, though the general feeling was that it still keeps its special 

characteristics to a lesser extent. S. Ito (Tottori University) discussed 

future rice consumption in Asia, showing statistical evidence 

from some countries. He forecast that Asian agriculture, 

and consequently Asian rural society, would lose ground if rice 

consumption declined at the pace experienced in Japan, Taiwan, 

and Korea. He emphasized the need of investment to generate 

new demand in the region. Some argued that the potential requirement 

for additional rice consumption by the low-income 

population should not be underestimated. 

H. Chien (JIRCAS) gave details on the dynamic shift in demand 

from quantity to quality observed in China, and concluded 

that japonica rice would be welcomed by expanded markets. It 

was noted that the quality differences among japonica rice would 

determine the future geographical distribution of japonica production 

in China. She also mentioned that national food security 

was still a major concern with respect to grain policy in China. V. 

Sriprasert (Thai Rice Exporters Association) stressed the massive 

supply potential in Asia based on the historical lessons and the 

current level of investment in some exporting countries in Asia. 

His view was that there would be plenty of business opportunities, 

including in Africa, if some adverse institutional restrictions 

were removed. 

E. Wailes (University of Arkansas) overviewed various activities 

currently conducted in the field of quantitative models on 

the international rice market. He stressed that model types were 

subject to the purpose of the analyses; thus, the evaluation of 

model results should also be done accordingly. It was extremely 

useful for the audience to grasp the cutting edge of modeling 

methodologies for the rice market. 

In general discussion, the future demand for rice, whether 

it will grow or not, was again a controversial issue. The importance 

of nonfood uses, such as for energy, bio-materials, and 

feed, was suggested. Areas will be emerging for both research 

and business opportunities. It was also suggested that research 

on the supply and demand situation requires information exchange 

among researchers; thus, wider cooperation should be promoted 

across business, academics, and international organizations. The 

session successfully provided information and instructions for 

participants from a wide range of research subjects. 


SESSION 18 

Impact of globalization on rice farmers 

CONVENER: M. Umemoto (NARO) 

CO-CONVENER: D. Dawe (IRRI)

Impacts of distorted trade policies on rice productivity 

Manitra A. Rakotoarisoa 

Studies (e.g., Wailes, 2003) showed that, although the removal 

of tariff and export subsidies in the highly distorted rice market 

would benefit the world, developing countries that produce 

rice but are net rice importers (e.g., Bangladesh, Brazil, 

Indonesia, Madagascar, Nigeria, and the Philippines) would 

lose. For these countries, with mostly poor rice farmers and 

low productivity, the increase in the producer’s gain would be 

too little to overcome the consumer’s loss as the market price 

rises. Seemingly, liberalization yields little benefit unless productivity 

rises first. But, did the lack of price competitiveness, 

caused in part by distorted rice policies that depressed revenues 

and profits, prevent poor farmers from adopting available 

technology 

The objective of this paper is to determine how distorted 

policies in the international rice market affect productivity and 

competitiveness in the rice sector. Specific objectives are to 

(1) measure the level and growth of total factor productivity 

(TFP) in the main rice-producing countries, (2) explain the 

sources of productivity gaps among rice producers, and (3) 

determine the interactions among trade liberalization, productivity, 

and competitiveness. The focus is on 33 rice-producing 

countries during the period 1961-2002. This study provides a 

new framework for measuring and analyzing sources of the 

productivity gap. 

Rice markets and policies 

Rice trade is thin and represented only 6.5% of global rice 

consumption during 1997-2002 (Childs 2002). Wailes (2003) 

reviewed current policies in the rice-producing countries in 

the OECD (Organization for Economic Cooperation and Development), 

where production and exports are highly subsidized 

and protected. These policies have caused distortions in 

prices and resource allocation for developing countries. During 

1961-2002, the price-adjusted inflation of rice in the world 

market has tended downward despite the growing global rice 

demand, but the price of imported farm inputs has increased 

sharply. 

Framework 

Measuring productivity 

We assume that the rice production function is Y it = F (X 1 it ,..., 

X k it , ev it), where index i represents country, t is the time dimension, 

and superscripts 1 …k represent input type. Y represents 

output, X’s are inputs, and X k is labor. The exponent of the 

residual ν it represents the level of technology. Using Cermeño 

et al (2003) henceforth as CMT, the residual in a dynamic form 

is ν it = µ i + λ t + Φν it–1 + ε it , where ε it˜ iid (0,σ ε 2 ), µ, and λ 

represent individual and time-specific effects, and µ, λ, and ε 

are uncorrelated to each other. CMT considered two other alternatives 

of the expression of ν it in replacing λ t by (1) θ i *t 

indicating that the growth rate of TFP (coefficient on time trend) 

is different for each country and (2) θ*t indicating that the 

TFP growth rate is the same for all countries. Taking the logarithms 

of the production function and taking into account these 

three expressions of the residual, production per worker is 

k–1 k–1 

(1a) y it = γy it–1 + Σ α j x j it + Σ β j xj it–1 + δ 0 Χk it 

j=1 j=1 

+ δ 1 Χ k it–1 + θ i *t + µ i + ε it 

k–1 k–1 

(1b) y it = γy it–1 + Σ α j x j it + Σ β j x j it–1 + δ 0 Χ k it 

j=1 j=1 

+ δ 1 Χ k it–1 + θ i *t + µ i + ε it 

k–1 k–1 

(1c) y it = γy it–1 + Σ α j x j it + Σ β j xj it–1 + δ 0Χ k it 

j=1 j=1 

+ δ 1 Χ k it–1 + λ t + µ i + ε it 

Variables y and x represent the log of output and input per 

worker, and ε’s are error terms. 

Competitiveness and total factor productivity 

We define a competitiveness index C as the ratio between value 

added estimated at world prices and input cost estimated at 

shadow prices: C = pV/(ωL+rK) where ω and r are shadow 

prices of labor L and capital K. Variable K combines all 

nonlabor inputs such as human capital, land, fertilizer, and tractor 

uses. p is the world price of the value added V; V = f (K, L, 

A) where A is the level of technology. As in Nishimizu and 

Page (1986), price p can be called “terms of trade,” reflecting 

differences between output price and intermediate input costs. 

The change in competitiveness index is 

dC dp dω dr dL 

(2) = – s L – s K – (s L – a L ) 

C p ω r 

L 

dK 

– (s K – a K ) – aA .dA 

K 


where s L = ω.L/(ω.L + r.K) and s K = r.K/(ω.L + r.K) are the cost 

shares of labor and capital at shadow prices, and a K and a L are 

input elasticities of substitution. 

The first three terms in (2) represent price competitiveness, 

PC, which increases with C as the world price increases. 

The fourth and fifth terms indicate production efficiency: as 

cost shares decline toward input elasticities, C grows. The last 

term is technological progress, affecting the growth of C. 

TFP gap and price competitiveness 

PC, the first three terms in (2), also represents the (negative) 

value of the world’s TFP (or TFP price) rate of change. The 

fourth and fifth terms vanish in the long run as input elasticities 

approach input cost shares to indicate price efficiency. 

Without trade distortions and under production efficiency, C 

depends only on the TFP gap between the domestic country 

and the rest of the world, that is, 

(3) 

From Coe et al (1997), we add control variables such as 

openness (OPEN), human capital (HK), and access to agricultural 

machinery—a function of equipment price and PC—and 

propose the following model: 

(4) 

where P w and q are the prices of rice and inputs and farm 

equipment. Subscripts w and i designate the world’s and individual 

country’s variables. 

Data 

Growth of 

TFP gap + Disturbances = 

(home/world) 

Output 

price 

Intermediate 

– – 

good’s price 

Other 

factor 

costs 

Growth 

of 

Change in 

+ production 

efficiency 

TFPw 

log ( ) t = α 0 + α 1 log(Pw) t + α 2 log(q) t 

TFPi 

+ α 3 log (HK w / HK i ) t + α 4 log(OPEN w / OPEN i ) t + η it 

We divide the 33 countries into three groups. Group 1 includes 

OECD countries, rice producers: Australia, EU (Italy), Japan, 

Korea, and the United States. Group 1 has high rice yields, 

subsidizes heavily its rice production and exports, and has high 

import protection. Group 2 includes developing countries, rice 

exporters, with little or no subsidies and relatively high yields: 

Argentina, China, Colombia, Egypt, Guyana, India, Myanmar, 

Pakistan, Suriname, Thailand, Vietnam, and Uruguay. Group 3 

includes mostly low-income countries that produce and import 

rice from the other two groups: Bangladesh, Brazil, Cambodia, 

Guinea, Indonesia, Iran, Laos, Madagascar, Mali, Malaysia, 

Nepal, Nigeria, Peru, the Philippines, Sri Lanka, and Tanzania. 

Data are yearly from 1961 to 2002. Appendix A explains 

the data sources. 

Table 1. Explaining the productivity gap among rice-producing countries 

(1971-2002). 

Group 2 a (low- and Group 3 b (low- 

Variables middle-income income rice 

rice exporters) importers) 

Dependent variable: log (TFP1/TFPi) 

Independent variables in log term: 

Output price –0.053*** –0.050*** 

(–4.34) c (–6.09) 

Farm equipment price ratio 0.112*** 0.149*** 

(8.11) (14.63) 

Openness ratio 0.300* 0.258 

(1.86) (1.47) 

Human capital ratio –1.017** –0.512*** 

(–2.22) (–3.16) 

Lag-dependent variable 0.145 0.409** 

(0.78) (2.33) 

Durbin Watson 2.01 2.09 

a Group 2 includes Argentina, China, Colombia, Egypt, Guyana, India, Myanmar, Pakistan, 

Suriname, Thailand, Vietnam, and Uruguay. b Group 3 includes Bangladesh, 

Brazil, Cambodia, Guinea, Indonesia, Iran, Laos, Madagascar, Mali, Malaysia, Nepal, 

Nigeria, Peru, Philippines, Sri Lanka, and Tanzania. c Numbers in parentheses and 

below the coefficients are t-values. The symbols *** , **, and * are levels of significance 

at 0.01, 0.05, and 0.1, respectively. 

Results 

Measures of TFP level and growth rate 

Between-groups estimation. In all three models, (1a), (1b), and 

(1c), µ i ’s, the group-specific effects, are statistically significant 

for group 1 and group 2 (group 3 is the basis for comparison). 

As expected, the high-income countries in group 1 have 

the highest level of TFP versus group 2 and group 3. Model 

(1b) produces the best fit and indicates, through θ, that the 

trend in world rice productivity grew at 0.4% on average per 

year during 1961-2002. 

Within-group estimation. For group 1, model (1b) yields 

the best fit. The coefficient θ is significant, indicating that TFP 

in group 1 grew at 1.6% per year on average. For group 2, 

model (1a) gives the best fit. Country-specific effects on TFP 

are significant in eight out of 11 countries. Model (1b) provides 

similar results and shows that TFP in group 2 grew at 

1% per year during 1961-2002. 

For group 3, model (1b) best represents the data. All 

country-specific effects of TFP are significant. Also, TFP 

growth is about 0.8% per year, meaning that the “catching-up” 

hypothesis does not hold; countries in group 3 started at lower 

TFP levels and their TFPs have grown slower than in group 2 

and group 1. 

Explaining productivity gaps 

From model (1a), we build a time series on TFP levels and 

growth for each group. Estimates of TFP levels are plotted in 

Figure 1 showing the widening TFP gap, especially between 

group 3 and group 1. We employ TFP measures to estimate the 

parameters of (4). 

Results are in Table 1. Coefficients have the expected 

signs and are significant, except for the openness variable. The 

Session 18: Impact of globalization on rice farmers 511

Index 

3.0 

2.5 

Group 1 (high-income producers) 

Group 2 (low-income exporters) 

Group 3 (low-income importers) 

2.0 

1.5 

1.0 

0.5 

0.0 

1961 1965 1969 1973 1977 1981 1985 1989 1993 1997 2001 

1963 1967 1971 1975 1979 1983 1987 1991 1995 1999 

Year 

Fig. 1. Estimated TFP index levels for the main rice-producing countries (1961-2002). 

Group 1 includes Australia, EU (Italy), Japan, Korea, and the United States. Group 2 

includes Argentina, China, Colombia, Egypt, Guyana, India, Myanmar, Pakistan, 

Suriname, Thailand, Vietnam, and Uruguay. Group 3 includes Bangladesh, Brazil, 

Cambodia, Guinea, Indonesia, Iran, Laos, Madagascar, Mali, Malaysia, Nepal, Nigeria, 

Peru, Philippines, Sri Lanka, and Tanzania. 

TFP gap has grown partly because of the lack of price competitiveness. 

A 10% decrease in the world price below its free 

trade level has increased the TFP gap ratios between group 1 

and the other two groups by 0.5%. Coefficients on q are statistically 

significant. Each percent increase in equipment price 

would widen the TFP ratios by about 0.11% and 0.15% for 

groups 2 and 3. The rising equipment price reduced poor farmers’ 

access to the inputs necessary to adopt available technology. 

As a result, their TFP levels lag behind TFP in rich countries 

and the TFP gap widens. 

The coefficients on OPEN are positive, and are significant 

at the 0.1 level for group 2, but not significant for group 

3. This supports early criticism (Rodrik 1992) that openness 

does not necessarily guarantee a reduction in TFP gaps. The 

coefficients on HK ratios are significant, indicating that rapid 

growth of human capital in groups 2 and 3 could help reduce 

TFP gaps. 

Conclusions and implications 

This paper focused on how distorted trade policies affect productivity 

and competitiveness in the rice sector. We estimated 

TFP levels and growth rates and explained the sources of the 

TFP gap among rice-producing countries. We found that the 

TFP gap in the rice sector has widened. The catching-up hypothesis 

does not hold: TFP levels and growth rates have been 

higher in high-income countries than in developing countries. 

The lack of price competitiveness, partly because of heavy 

production and trade subsidies in the OECD, was one of the 

sources of this productivity gap. Limited access to equipment 

has increased the TFP gap. Also, trade openness has a weak 

positive impact on the TFP gap; this may partly reflect difficulties 

in choosing proxies (Edwards 1997). 

Results imply that removing distortions on output price 

and input access can boost revenue and profit for rice producers 

in developing countries. That will allow investments for 

the use of available technology and reduce the TFP gap. 

Appendix A: Variables and sources 

Inputs: X f , fertilizer use, proxyed as the total fertilizer use 

times the share of harvested rice fields over arable land (UN 

FAO); X k , amount of labor in rice production proxyed as the 

rural labor force, that is, active workers in rural areas (World 

Bank); X tr , physical capital, proxyed as the average number of 

tractors used per unit of agricultural land times rice-harvested 

area; X l , rice-harvested area (UN FAO). 

HK: human capital, is net secondary enrollment ratio (World 

Bank). 

OPEN: openness, is trade value over GDP (World Bank). 

Pw: world price of rice; proxy is the f.o.b. price of 5% broken 

milled rice from Bangkok, Thailand; prices are in 1982 US 

dollars constant (World Bank). 

q: price of agricultural and farm equipment; proxy is the index 

price of agricultural and farm inputs and equipment times the 

exchange rate (UN FAO, U.S. Department of Labor, World 

Bank). 


: rental price of capital, or lending interest rate (World Bank). 

Y: milled rice production in metric tons (UN FAO). 

References 

Childs N. 2002. Rice outlook. Economic Research Service, United 

States Department of Agriculture, Washington, D.C., USA. 

Cermeño R, Maddala GS, Trueblood M. 2003. Modeling technology 

as a dynamic error components process: the case of the 

inter-country agricultural production function. Econ. Rev. 

22:289-306. 

Coe D, Helpman E, Hoffmaister A. 1997. North-south R&D spillover. 

Econ. J. 107(440):134-149. 

Edwards S. 1997. Openness, productivity and growth: what do we 

really know National Bureau of Economic Research Working 

Paper No. 5978. 

Nishimizu M, Page J. 1986. Productivity change and dynamic comparative 

advantage. Rev. Econ. Stat. 68:241-247. 

Rodrik D. 1992. Closing the technology gap: Does trade liberalization 

really help In: Helleiner G, editor. Trade policy, industrialization, 

and development: new perspectives. Oxford (UK): 

Clarendon Press. 

United Nations Food and Agriculture Organization. Various years. 

Agriculture database. 

United States Department of Labor. Bureau of Labor Statistics. Price 

indexes. 

Wailes E. 2003. Global trade and protection regime in rice trade. 

Draft paper. University of Arkansas. 

World Bank. 2003. World development indicators. 

Notes 

A new strategy for group farming in Japan 

Kyoichi Miyatake 

Authors’ address: Markets, Trade, and Institutions Division, International 

Food Policy Research Institute, 2033 K Street N.W., 

Washington, D.C. 20006, USA, e-mail: 

m.rakotoarisoa@cgiar.org and a.gulati@cgiar.org. 

Japanese rice farming was influenced by the liberalization of 

the domestic market and minimum access to rice trading. The 

price of rice decreased by 27% from 1992 to 2002. Because of 

this problem, the Japanese government promotes the improvement 

of the productivity of rice production. This policy considers 

large-scale individual farmers (10–20 ha) and cooperative 

management farms (35–50 ha) to be model farms. These 

are estimated to be only 45,000 of the total of 3 million Japanese 

farmers. So, in regions where large-scale farms aren’t 

found, “community-wide group farming” based on the rural 

community is expected as a model of farming. 

In Japanese agricultural areas, traditionally, rural communities 

of 20 to 30 farmers have been the basic units not only 

of social life but also of rice-farming work, such as the management 

of irrigation water and joint farm work. Group farming 

is an activity based on the rural community, which aims to 

modernize and restructure rice production through joint use of 

farm machines, joint farm work, and community land-use arrangements. 

Therefore, about 10,000 farming groups were established 

in Japan. 

The Noguchi farming group of Johana town, Toyama 

Prefecture, is an example of community-wide group farming. 

The Noguchi farming group is incorporated and is now working 

on the organic cultivation of rice, too. This report analyzes 

the group’s farm management activities for over three decades 

and discusses its improvements. 

Outline of the area 

Traditionally, most farmers in Johana have been engaged in 

rice growing on their land of 1 ha or so. But, in the 1970s, 

factories of metals and electronics were constructed in Toyama 

Prefecture, and farmers obtained new jobs in nonagricultural 

industries. Under those conditions, most farmers in the town 

became part-time farmers and took part in community-wide 

group farming to maintain their rice production through the 

joint use of farm machines, joint farm work, and community 

land-use arrangements because their farm size was too small 

to modernize rice production and make enough money. 

As a result, 21 of the 32 rural communities in Johana 

now have group-farming units. The farmland managed by these 

units is two-thirds of the paddy field in Johana (1,367 ha). On 

the other hand, only 33 of the total of 649 town farmers are 

full-time farmers engaged in beef-cattle raising, fruit production, 

or vegetable cultivation in addition to rice growing. 

Foundation of the Noguchi farming group 

In Noguchi community, 16 farmers started the joint use of ricefarming 

machines in 1975. In 1983, when restrictions on rice 

growing had become serious, Noguchi community started the 

joint production of soybean and barley as an alternative to rice. 

Then, in 1985, as a project financed by Toyama Prefecture, 

the community introduced rice-farming machines and constructed 

a joint rice drying and processing plant (Table 1), 

thereby starting community-wide agriculture with the participation 

of all 23 farmers in the community. In general, community-wide 

farming groups adopted the joint use of machines in 

individual fields, and members got the harvest from individual 

paddy fields. But, in the group farming in Noguchi reorganized 

in 1985, a new system was adopted: the members worked 

together and the harvest was processed and sold jointly in the 

community. And the earnings were distributed to the member 

farmers as land rent and wages for farm work. 


Table 1. Outline of the Noguchi farming group. 

Item 

Characteristics 

Members 

23 farm households (all households in Noguchi 

community): 47 participate in joint farm work 

(28 males, 19 females), of whom 16 are farm 

machine operators 

Farm area 29.4 ha (paddy fields; all plots enlarged to 0.3 

ha) 

Crops 

Rice: 21.4 ha (entirely organically cultivated); 

Koshihikari 60%, Gohyaku Angoku (rice for 

sake brewing) 40%; soybean 8.0 ha; barley 

4.3 ha (double cropping with soybean) 

Yield 

Rice 4.6–5.2 t ha –1 (brown rice basis); soybean 

1.3–2.5 t ha –1 ; barley 3.0–4.0 t ha –1 

Machines and facilities Two 43-hp, one 34-hp, and one 29-hp tractor; 

two 8-row transplanters; two 5-row combines; 

drying and processing building (436 m 2 ); a 

set of seedling culture equipment and a forklift; 

six 36-Koku dryers for rice (1 Koku = 180 

L); a soybean combine 

Working hours for rice 213 h ha –1 of paddy fields; of which daily management 

work (100 h) is entrusted to member 

farmers 

Distribution of earnings Net returns $6,646 composed of land rent 

$2,000 ha –1 , daily management fee $2,727 

ha –1 , and wage for joint work $1,918 ha –1 ; 

wages $16.36 h –1 for males and $14.55 

h –1 for females 

In those years, most farmers in the community experienced 

a generation shift from the first generation, which had 

established the organization for joint machine use in 1975, to 

their sons as the second generation. The younger farmers had 

served as full-time workers in nearby cities, so it became more 

important to increase the efficiency of group farming. This 

new system of profit distribution was introduced by the younger 

generation as its members overcame the opposition of the older 

generation, who wanted to continue getting the harvest from 

each paddy field. As a result, working time for rice decreased 

from 297 h ha –1 (local average) to 213 h ha –1 , although that 

included 100 h for daily management done by individual farmers. 

In addition, the farming group was incorporated in 1988, 

which allowed it to conclude land-lease agreements with farmers 

in the community and to stock earnings for future renewal 

of machines and facilities. Since then, the 43-hp tractor, 8-row 

transplanter, and 5-row combine harvester have been replaced 

without financial support. 

Securing the labor force 

The Noguchi farming group does, as its joint work, the main 

rice cultivation work, such as seedling production, preparation 

of paddy fields, transplanting, harvesting, and drying and 

processing. The farming group entrusts daily management 

tasks, including water management and hand weeding, to the 

owners of land for wages (seven retiring farmers asked their 

neighbors). The growing of soybean and barley, which is highly 

mechanized, is totally carried out by the joint efforts of the 

farming group. In this group, 16 members practiced the main 

joint work as operators, of which 11 are young farmers who 

work full-time in nonfarm jobs during the day, and so joint 

work is performed mainly on the weekends. On weekdays, five 

other farmers, who retired from nonagricultural work, take 

charge of farm work. Their agricultural know-how is useful 

for seedling management and drying control. On the other hand, 

seedling operation and transplanting require intensive work 

and are carried out on the weekends by all 47 inhabitants of 

working age, including elderly people and high school students 

in Noguchi community. To promote the entrustment of 

rice cultivation management and participation in joint farm 

work, the farming group has improved the distribution method 

of earnings by reducing the amount of land rent to farmers in 

the community and raising the fund for the entrustment fee 

and wages. The daily management fee increased from 

US$2,091 to $2,727 ha –1 and wages rose from $10.91 to $16.36 

h –1 for males and from $8.64 to $14.55 h –1 for females. Because 

these are very attractive rates for farmers, it was easy to 

secure the necessary labor force. 

Organic farming activities 

Recently, organic cultivation of high-quality rice has been encouraged 

in an attempt to create profitable rice farming. It is 

generally practiced by large, full-time farmers because organic 

farming requires farmers to prepare the soil by applying sufficient 

compost every year (20 t ha –1 in Johana), to restrict chemical 

application very carefully according to an agreement with 

buyers, and to conduct intensive nursing by continuous field 

observations. 

But farmers in Johana are mostly part-timers, and the 

group-farming systems play a leading role in the promotion of 

organic farming of rice; farmers cooperate with each other in 

the application of manure. Farmers also inspect paddy fields 

in turns to jointly make decisions about fertilizing and disease 

and insect control. These groups’ daily operation records were 

used as evidence for inspection of chemical use. On the other 

hand, the agricultural cooperative association in Johana did 

marketing promotion and made a contract of dependable selling 

with consumers to encourage organic farming for farming 

groups. By those efforts, farmers sold organic rice at a premium 

of $0.23 kg –1 . Now, organic farming has been introduced 

to 21% of the total paddy fields of the town, especially the 

Noguchi farming group, which has switched all of its rice production 

to organic cultivation. 

Closing remarks 

With this history since 1975, the Noguchi farming group has 

become an example of community-wide group farming in Japan. 

The Noguchi farming group has maintained and expanded 

its organization by the following four strategies (Table 2). First, 

the farming group was able to raise the efficiency of farm work 

by changing joint machine use on individual fields into joint 


Table 2. Strategies of the Noguchi farming group. 

Strategies of the Noguchi farming group 

Outcomes 

Adoption of joint production and selling Working hours for rice 

of rice in the community instead of 

decreased to 213 h ha –1 from 

joint machine use in individual paddy fields. 297 h ha –1 of local average 

Change of distribution method of earning High wage rates: 

by reducing land rent and raising wage $16.36 h –1 for males, 

to secure necessary labor force. 

$14.55 h –1 for females 

Incorporation of group farming 

Renewal of tractor, 

allowed financing for future renewal 

transplanter, and combine 

of machines for rice cultivation. 

without financial support 

Organic rice cultivation succeeded in 

Farmers sold organic rice 

profitable farming. at a premium of $0.23 kg –1 

production and selling. Second, by reducing land rent payments 

and raising wages for farmers, farmers are encouraged 

to take part in joint farm work. Third, the incorporation enabled 

the group to have stocked earnings, which made it easier 

to renew farm machines. Finally, by introducing organic rice 

cultivation, profitable farming was obtained. With these strategies, 

group farming has overcome the overinvestment in machines 

and ineffective management, the situation that Japanese 

part-time farmers usually experienced, and has obtained 

productivity as high as that of large, full-time rice growers. 

That is the reason the Japanese government considers community-wide 

group farming to be an exemplary model of rice 

farming and promotes that instead of further globalization. 

Notes 

Author’s address: NARO, e-mail: kyoichi@affrc.go.jp. 

The role of the rice economy after the implementation 

of agricultural policy reform and trade liberalization 

from the perspective of farmers: the case of rice farmers 

in Java, Indonesia 

Roosiana 

Rice has traditionally been the main source of livelihood among 

Indonesian farmers. Rice has also remained a unique and strategic 

commodity, closely related to food security and economic, 

social, and political concerns. Rice has been the staple food 

for the majority of Indonesian people and still holds the highest 

share among consumption commodities. The rice sector 

employs more than 18 million farming households, or about 

70.8% of the total farming households in 2003, of which the 

farming household itself constitutes more than 48% of the total 

households. 

Because rice has an important role in the Indonesian 

economy, the government needs to protect both rice farmers 

and consumers, and assure enough rice supply to meet demand. 

Under the New Order administration (1969-98), the primary 

effort in agricultural development was devoted to increasing 

rice production to the point of achieving self-sufficiency. 

Through intensification programs, the so-called BIMAS (literally 

meaning mass guidance) and INMAS (literally meaning 

mass intensification) programs, Indonesia managed to boost 

rice production and achieve rice self-sufficiency in 1985. Dur- 

ing those periods, rice farmers mostly enjoyed subsidies, irrigation 

establishment, and market protection from the government. 

Because of the economic crisis, beginning in 1998, Indonesia 

introduced changes in its rice policies to include rice 

trade liberalization. The government opened the rice economy 

more to the global market. Prior New Order strategies for the 

rice economy were abolished gradually, including some subsidies. 

The rice sector has lost some protection since rice trade 

was liberalized in 1998, and local farmers now have to face 

global competition, whereas the government strategy has 

shifted to food security rather than self-sufficiency. 

This development has necessitated more research to examine 

the performance of the rice economy as affected by food 

security strategies, mostly from a macroeconomic point of view. 

There are only a few studies on farmers whose lives revolve 

mainly around rice as their main livelihood. Using the case 

study approach, this paper attempts to investigate the role of 

the rice economy given some rice policy reforms made, from 

the perspective of Indonesian rice farmers in Java that histori- 


cally were the country’s main rice producers. Likewise, the 

study tries to determine some factors that influence the performance 

of the rice economy at the farm level, such as equity 

aspects. 

The study area and sample and data collection 

With growing attention to data quality and cost-effectiveness, 

small-sample surveys were designed to obtain comparative data 

for relatively developed areas of rice farming. These villages 

were in Central Java and Yogyakarta provinces, which were 

purposely selected to represent current and common conditions 

of rice farmers in Java. The selected villages, Gelaran I 

hamlet of Bejiharjo and Karangtengah hamlet of Karangtengah 

village in Gunungkidul Regency, Sedayu hamlet of Awonggo 

village in Magelang Regency, and Ceporan and Towangsan in 

Klaten Regency, represented relatively developed rice-farming 

areas in Java. 

Respondents were selected purposely from a number of 

rice farmers who were available during the surveys. All respondents 

were farmers working in their own fields. Of the 

total of 101 respondents, 21 were from Bejiharjo village, 25 

from Karangtengah village, 25 from Awonggo village, and 30 

from Ceporan and Towangsan villages. A coded questionnaire 

containing two major parts was formulated. The first part had 

questions dealing with general information on the farming 

household, its activities, and general social and economic conditions. 

Questions in the second part asked farmers to recall 

their conditions before 1998 in comparison with recent conditions. 

Role of the rice economy for farmers 

As has been implied in prior sections, the main role of rice for 

paddy farmers has been closely related to the household’s income. 

In the past, farmers could have a larger rice field than 

now; therefore, farmers could rely on rice as the main source 

of livelihood. As generations have passed and land has been 

shared with more people, especially by inheritance, farmers 

who hold land have increased in number. However, the size of 

landholding for each individual farmer has decreased over 

generations. 

Field surveys indicated that 84 out of 101 respondents 

(83.17%) were farmers with a landholding of less than 0.5 ha. 

The rest of the respondents had less than 1 ha of rice field. 

Most respondents were rice farmers who always spent their 

lives in the field everyday, and who have been rice farmers for 

more than 20 years. Only 16 farmers of the 101 respondents 

had other nonrice agricultural activities besides rice cultivation 

as their main source of income. 

The surveys tried to find out what farmers thought about 

the role of the rice economy in their livelihood, particularly 

after the policy reform, and most of them still considered rice 

as their main source of livelihood. However, since their fields 

were small, they used some portions from the harvest for their 

own consumption and for social and community purposes. 

Farmers must have other assets such as livestock for extra expenses. 

As a matter of fact, those farmers still keep growing 

rice more as a tradition than to make a profit. 

Rice performance and equity 

Economic development today gives more attention to the relation 

between equity and efficiency (Baldassarri and Piga 1996). 

While development in the rice economy achieved efficiency 

goals with increased rice production, equity objectives seemed 

to be left behind, particularly for rice farmers. Studies on equity 

in the Indonesian rice economy are still few. Prior studies 

suggested that factors affecting equity in the Indonesian rice 

economy mostly came from technological changes and rice 

strategies. Jatileksono (1987) in his study about equity achievement 

concluded that technology changes worsened income 

distribution between favorable and unfavorable areas unless 

modern varieties along with irrigation expanded to the unfavorable 

areas. Another study by Naylor (1991) stated that primarily 

rice producers, a high-income group, benefited from 

the transfer effects of government strategies in rice prices and 

irrigation investment. In contrast to national-level studies, our 

study focuses on the performance of the rice economy regarding 

equity at the farm level. Based on survey findings, our 

study presents three factors affecting equity of the rice economy 

after the policy reform of 1998. These factors were landholding, 

the rice selling price and prices of consumption goods, 

and an assessment of modern technologies and new information. 

Landholding. Landholding has become crucial for traditional 

rice farmers. There were no changes in the 

size of land ownership after the reform. Even though 

productivity was quite high, on average 4–5 t ha –1 of 

unhusked dry rice, with a small piece of land, it was 

difficult for farmers to improve their level of welfare. 

The rice selling price and prices of consumption 

goods. Except for farmers of Ceporan and Towangsan 

villages, farmers had a rice selling price that was 

slightly higher than the floor price. However, the 

prices of consumption goods have moved higher than 

the profit gained from rice cultivation. As a result, 

income from rice was not sufficient to pay for public 

services such as education for children and health 

services. Rice policy reform thus did not bring significant 

changes to intervening prices for better farmers’ 

welfare. 

Technology changes and new information. Based on 

survey results, it was found that rice production after 

the reform did not change much. Some farmers increased 

their rice production per ha, but most of them 

did not. The cropping pattern during the year also 

remained the same, and was not affected by the reform, 

which eliminated subsidies and incentives given 

to farmers. However, with the higher cost of inputs, 

when production was low and the rice selling price 

was not high, farmers had almost no ability to make 


savings to improve their economy. When the opportunity 

to expand their fields was small, rice farmers 

could optimize field cultivation by applying technology 

changes and new information in rice farming and 

income generation. However, limited assessment was 

made of this. 

Concluding remarks 

Agricultural policy reform and trade liberalization since 1998 

have brought about some changes in rice strategies that led to 

less protection and fewer incentives for rice farmers. Results 

of the field surveys in five villages in Central Java and 

Yogyakarta provinces showed that rice farmers still considered 

rice as their main source of income despite insufficiency 

in income generation from rice cultivation. The present situation 

of rice at the farm level suggested three factors that affected 

equity achievement: landholding, the rice selling price 

and prices of consumption goods, and technology changes and 

new information. An assessment of rice farmers regarding these 

three factors is indeed necessary, particularly to improve their 

level of welfare. 

References 

Baldassarri M, Piga G. 1996. Distributive equity and economic efficiency: 

trade-off and synergy. In: Baldassarri M, Paganetto L, 

Phelps ES, editors. Equity, efficiency and growth: the future 

of the welfare state. Hampshire and London (UK): MacMillan 

Press Ltd. p 257-275. 

Jatileksono T. 1987. Equity achievement in the Indonesian rice 

economy. Yogyakarta (Indonesia): Gadjah Mada University 

Press. 120 p. 

Naylor R. 1991. Equity effects of rice strategies. In: Pearson S, Falcon 

W, Heytens P, Monke E, Naylor R, editors. Rice policy in 

Indonesia. Ithaca, N.Y. (USA): Cornell University Press. 

p 138-161. 

Notes 

Author’s address: Graduate School of International Development, 

Nagoya University Furo-cho, Chikusa-ku, Nagoya 464-8601, 

Japan, e-mail: m020109d@mbox.nagoya-u.ac.jp. 

El Niño sensitivity, resource endowment, and socioeconomic 

characteristics: the case of wetland rice in Java, Indonesia 

Shigeki Yokoyama and Bambang Irawan 

Objectives and method 

Regional factors influencing climatic vulnerability were analyzed 

at the subdistrict (kecamatan) level on Java island of 

Indonesia in the case of the 1997-98 El Niño. Out of 1,454 

subdistricts where wetland rice grows, area loss was observed 

in 58%, with an average rate of –6.9% per subdistrict. Most of 

them were located in the highland regions of the southern parts 

of West and Central Java provinces. 

The principal component analysis was applied to clarify 

the local characteristics that determine El Niño vulnerability. 

The analysis incorporated 16 variables of four categories: (1) 

natural conditions, (2) level of infrastructure, (3) household 

characteristics, and (4) farming performance. The data were 

obtained from the Central Bureau of Statistics and came from 

four publication series: (1) Land use in Java, (2) Agricultural 

census 1993, (3) Potency of villages, and (4) Annual reports 

from the agricultural offices of Kabupaten in Java. The description 

of variables is as follows. 

Natural conditions 

X1: Percentage of villages with altitude

Table 1. Coefficient of principal factors and their correlations with variables. a 

Item 

Principal factor 

F1 F2 F3 F4 F5 F6 F7 

Factor coefficient 4.353 2.441 1.713 1.020 0.912 0.851 0.818 

Factor contribution (%) 27.2 15.3 10.7 6.4 5.7 5.3 4.6 

Cumulative contribution (%) 27.2 42.5 53.2 59.5 65.2 70.6 75.2 

Coefficient of correlation of principal factors and variables 

X1 –0.497 0.004 –0.126 0.033 0.004 –0.018 0.002 

X2 0.088 0.043 0.007 0.050 0.494 –0.034 0.022 

IN1 –0.549 0.010 0.185 0.006 0.037 –0.001 –0.012 

IN2 –0.630 –0.077 –0.035 0.001 0.001 0.014 –0.114 

IN3 0.575 0.013 0.087 –0.009 0.002 0.002 0.152 

IN4 –0.012 0.558 0.082 –0.147 0.013 0.051 0.001 

IN5 0.094 –0.541 –0.040 0.149 –0.004 0.059 0.016 

IN6 0.159 –0.070 0.146 0.159 0.073 0.004 –0.211 

IN7 –0.001 0.183 0.597 –0.037 –0.115 –0.081 –0.019 

IN8 –0.211 –0.002 0.072 0.105 –0.012 0.474 0.010 

HH1 –0.596 0.076 0.184 0.021 0.002 0.017 0.003 

HH2 –0.518 0.125 0.139 0.011 0.004 0.027 0.002 

HH3 0.003 –0.681 0.011 –0.039 0.024 0.015 0.001 

HH4 –0.174 –0.506 0.009 –0.564 0.005 0.007 0.083 

R1 0.095 –0.074 0.462 0.003 –0.188 –0.042 0.001 

R2 –0.368 –0.035 0.006 –0.059 –0.016 0.003 0.171 

a Boldface = high coefficient of correlation. 

Wetland rice performance 

R1: Crop intensity per year (%) 

R2: Yield (kg ha –1 ) 

Results of principal component analysis and grouping 

Results of principal component analysis show that 75% of the 

total variance could be explained by seven principal factors 

(Table 1). The first factor, highly correlated to variables X1, 

IN1, IN2, IN3, HH1, and HH2, represents production conditions, 

contributing 27% to the total variance. The second factor, 

related to IN4, IN5, HH3, and HH4, stands for social infrastructure. 

The third factor, related to IN7 and R1, indicates 

technological level and productivity. The fourth factor, related 

to HH4, could be interpreted as social structure. The fifth factor, 

related to X2, indicates rainfall. Those five factors, being 

able to explain about 65% of the total variance, are used for 

kecamatan grouping. 

Table 2 shows the grouping of kecamatans based on five 

principal factors. The comparison of mean values of variables 

between groups shows that several variables were highly correlated 

to one another. For instance, kecamatans with the most 

villages located in regions with altitude of 500 m or in upland regions 

(G2 and G3). Wetland rice yield in lowland kecamatans was 

also higher than in upland ones because the former generally 

had better irrigation networks. This was clarified by the higher 

proportion of technical irrigated land in lowland regions (IN2). 

G2 and G3 were both located in upland regions and had 

high annual rainfall, though several other variables showed 

differences between the two. G2 was relatively high in IN4 (% 

of villages with paved main road) but relatively low in IN5 

(number of households to number of cars), IN7, IN8, HH3, 

and R1. This means that G2 and G3 had a different level of 

infrastructure, household landholding, and crop intensity for 

wetland rice, even though they had similar natural resource 

endowments. 

A similarity of natural resource endowments is also observed 

for G1 and G4. Both groups were located in lowland 

regions and had relatively low annual rainfall. However, G1 

had a better transportation infrastructure (IN4) and larger average 

farm size (HH1 and HH2) than G4. In contrast, G4 had 

a higher crop intensity for wetland rice than G1. 

Rainfall decrease, area decrease, and vulnerability 

The rainfall decrease caused by El Niño varies by locality because 

of the difference in natural resource endowments. However, 

at a homogeneous level of rainfall decrease, production 

loss might differ, depending on farmers’ capability to cope with 

the effects of a rainfall decrease on their agricultural production. 

The farmers’ ability to adjust to a rainfall decrease is a 

function of the characteristics of farmers’ households, infra- 


Table 2. Rainfall decrease, area decrease, and vulnerability index by group of kecamatans. 

Group of kecamatans 

Variable Description of variable Lowland area Upland area 

G1 G4 G2 G3 

Characteristics of kecamatans 

IN4 Percentage of villages with paved main road (%) High Low High Low 

IN5 Ratio of number of households to number of cars Low Low Low High 

IN6 Ratio of wetland area to number of water pumps Low Low Moderate High 

IN7 Ratio of wetland area to number of agicultural input kiosks High High Low High 

IN8 Percentage of villages accessible to mass transportation High High Low High 

HH1 Ratio of wetland area to farmland High Moderate Low Low 

HH2 Mean wetland size (ha per farm household) High Moderate Low Low 

HH3 Percentage of farm households with wetland size 

Comparison between groups G1 and G4, both lowland 

regions, shows similar results. The higher vulnerability index 

of group G4 than group G1 was accompanied by a lower transportation 

infrastructure (IN4) and farm size (HH1 and HH2) 

and higher crop intensity (R1). 


The subdistricts of Java island, Indonesia, were grouped into 

four based on the combination of location and crop intensity: 

G1 (low altitude [well irrigated], low crop intensity), G2 (high 

altitude [less irrigated], low crop intensity), G3 (high altitude 

[less irrigated], high crop intensity), and G4 (low altitude [well 

irrigated], high crop intensity). Because of the higher rainfall 

decrease and poor irrigation, the impact of the area decrease is 

higher on highlands (G2, G3). While El Niño sensitivity, defined 

as A/R (A: percentage of area loss, R: percentage of rainfall 

reduction), is higher in G3 and G4, where more rice is 

cultivated in the dry season compared to G1 and G2. This finding 

is consistent with the fact that El Niño-induced drought 

was more prominent in the dry season of 1997. Other characteristics 

of an area highly sensitive to El Niño are (1) remoteness 

or poor transportation infrastructure, (2) lower availability 

of water pumps, (3) lower availability of input kiosks, and 

(4) high proportion of small farms (less than 0.5 ha). As expected, 

areas with good infrastructure are less vulnerable to 

climatic shocks, while smaller farms are more vulnerable. 

Notes 

Authors’ addresses: Shigeki Yokoyama, National Agriculture & Biooriented 

Research Organization, Japan, Kannondai 3-1-1, 

Tsukuba, Ibaraki, Japan; Bambang Irawan, Indonesian Center 

for Agro Socio-Economic Research and Development, Jalan 

A. Yani No. 70, Bogor, Indonesia, e-mail: 

syokoyam@affrc.go.jp. 

Global competitiveness of medium-quality Indian rice: 

a PAM analysis 

B.V. Chinnappa Reddy, M.S. Raghavendra, and Lalith Achoth 

During the last five decades, rice area in India has increased 

from 20 million to about 45 million hectares and rice production 

from 22 million to 90 million tons. Productivity increased 

from 700 to 2,000 kg ha –1 . Rice is a major cereal for more 

than 70% of the world’s poor in Asia, where more than 90% of 

rice production and consumption take place. Despite the rapid 

expansion of rice trade globally, as argued by Gulati and 

Narayanan (2003), the world rice market is characterized by 

thinness, volatility, and segmentation. Rice markets are affected 

by distortions caused by a plethora of controls imposed by 

both developed and developing countries because of various 

political, economic, and food security compulsions. In the long 

run, trade liberalization is likely to benefit both consumers and 

producers in developing countries, in addition to the availability 

of cheaper grain in the importing countries (Gulati and 

Naranayan 2003). 

India is a traditional exporter of high-quality basmati 

rice. But the export of nonbasmati rice was rather small in 

earlier years, as it was not export-competitive (Chand 2002). 

However, the devaluation of the Indian rupee drastically altered 

the export scenario of the country. This was followed by 

a gradual relaxation of restrictions on the export of rice, which 

resulted in a gradual increase in exports of medium-quality 

rice. To assess the export potential and consequent welfare 

gain from rice, it is mandatory to assess the export competitiveness 

of medium-quality rice. As Karnataka symbolizes 

medium-quality rice production in the country, we have chosen 

this state for assessing the export competitiveness and concomitant 

welfare gain due to the export trade of rice. 

Methodology 

Policy matrix analysis 

The trade competitiveness of Indian medium-quality rice was 

analyzed using the framework of the policy analysis matrix 

(PAM) as proposed by Monke and Pearson and presented by 

Yao (1997). To construct the PAM, world prices (free on board, 

FOB, for exports, and cost, insurance, and freight, CIF, for 

imports) were used. For domestic factors, which are not traded 

internationally, social costs (Pxi) were calculated using the 

value of marginal product approach that uses factor shares (Si) 

of various inputs (Xi) together with the mean value of inputs 

and outputs (Y) and price (Py). Social cost was obtained as 

Pxi = { (Si/Xi) × Y} × Py. The nontradable inputs considered 

were labor (human and bullock), farmyard manure, and rental 

value of land. The tradable inputs were seed, fertilizers, and 

pesticides. 

Nominal protection coefficients (NPCs), effective protection 

coefficients (EPCs), and domestic resource cost (DRC) 

were computed to reveal trade competitiveness. Trade competitiveness 

was estimated using the above three measures for 

rice in India using the data from Karnataka on the importable 

hypothesis for the two periods, preliberalization (1985-86 to 

1991-92) and postliberalization (1996-97 to 2000-01). 

NPC was calculated as a ratio between the domestic price 

and the international price of a comparable grade of rice, adjusted 

for all the transfer costs such as freight, insurance, handling 

costs, margins, losses, etc. 


EPC for rice was defined as EPCi = (VA i 

d 

/VA i b ) where 

VA i 

d 

is the value added of output i at domestic prices and VA i 

b 

is the value added of output i at border prices. 

DRC was obtained as DRC = a ij v j /(P i 

b 

– a ij P i b ) where a ij 

(j = K + 1 to n) is the technical coefficient (input use per unit 

of output) for domestic resource (nontraded intermediary input), 

j is the production of output i, and v j is the shadow price 

of such an input. 

The cost of cultivation data needed for computing the 

PAM were collected from the Directorate of Agriculture, Government 

of Karnataka, Bangalore. The international reference 

price of the crop was collected from various issues of the FAO 

production yearbook and from the FAO Web site. 

Partial equilibrium analysis 

The extent of price distortions created by the policies in the 

face of free trade was computed and likely gain or loss to producers 

and consumers was also estimated. A partial equilibrium 

approach proposed by Lutz and Scandizzo (1980) was 

used to evaluate the impact of the price changes on demand, 

supply, and welfare based on the following measures: 

1. Net social loss in production (NSLp) 

NSLp = ½ (Qw – Q) (Pw – Pp) 

2. Net social loss in consumption (NSLc) 

NSLc = ½ (Cw – C) (Pw – Pc) 

3. Total net social loss (NSL) 

NSL = NSLp + NSLc 

4. Welfare gain of producers (Gp) 

Gp = Q (Pp – Pw) – NSLp 

5. Welfare gain of consumers (Gc) 

Gc = Q (Pw – Pc) – NSLc 

6. Net effect of liberalization on welfare in the state 

Q (Pp – Pw) – Q (Pw – Pc) 

7. Change in government revenue 

dG = (NSLp + NSLc) – Gp – Gc 

where Qw = production at world prices, Q = production at 

domestic prices, Pw = border prices, Pp = price faced by domestic 

producers, Pc = price faced by domestic consumers, 

Cw = consumption at world prices, and C = consumption at 

domestic prices. 

The basic parameters of supply and demand elasiticies 

needed for this evaluation were drawn from a study by Reddy 

(1997). To compute production values, the wholesale price of 

rice was used. For consumption values, the retail price was 

used. The world reference prices were derived from the international 

price and adjusted for transportation, marketing, and 

trading margins. 


Trade competitiveness of Indian rice 

Trade competitiveness of a commodity reveals whether a country 

has an opportunity to engage in export trade. It is interesting 

to observe that rice, which is the major crop in Karnataka 

State, had been largely competitive on an importable basis with 

its NPC values being below unity during the reference period. 

Table 1. Results of a policy analysis matrix (PAM) of rice. 

Year NPC a EPC DRC SRP 

1985-86 0.81 1.19 0.95 0.62 

1986-87 0.89 1.20 0.99 0.52 

1987-88 0.81 1.12 0.91 0.42 

1988-89 0.68 0.91 0.76 0.25 

1989-90 0.63 0.76 0.63 0.10 

1990-91 0.65 0.82 0.70 0.17 

1991-92 0.66 0.79 0.68 0.11 

Average b 0.74 1.00 0.82 0.35 

1992-93 0.77 0.97 0.82 0.35 

1993-94 0.65 0.88 0.73 0.20 

1994-95 0.69 1.00 0.79 0.32 

1995-96 0.64 1.00 0.78 0.27 

1996-97 0.67 0.93 0.77 0.27 

1997-98 0.55 0.68 0.53 0.04 

1998-99 0.52 0.61 0.49 –0.05 

1999-2000 0.53 0.59 0.49 –0.07 

2000-01 0.64 0.84 0.59 0.15 

Average c 0.58 0.73 0.58 0.07 

a 

NPC = nominal protection coefficients, EPC = effective protection coefficients, 

DRC = domestice resource cost, SRP = subsidy ratio to producer 

coefficient. b Average of 1985-86 to 1991-92. c Average of 1992- 

93 to 2000-01. 

EPC estimates showed that, in only five years during the 17- 

year reference period, it was more than 1, indicating that the 

state had protected the crop only in those years. However, for 

the reference period, the average EPC revealed that Karnataka 

is an efficient producer of rice (Table 1). Over the years, EPC 

had been declining, which implies an increasing rate of competitiveness 

of rice. This could be due to the emergence of 

efficient production technology and the impact of economic 

reforms in the country. 

The estimates of DRC revealed that the state had a comparative 

advantage in rice production (DRC was below 1). The 

level of DRC shows that the value of domestic resources used 

in producing 1 ha of rice in Karnataka was less than the cost of 

its import. DRC level decreased in the postliberalization period, 

which reveals an improvement in the comparative advantage 

of rice production in recent years. The subsidy ratio 

to producer coefficient (SRP), which was computed to analyze 

the degree of state protection for the rice crop, was 0.07 

for the state for the period 1996-97 to 2000-01. This implies 

moderate state protection for rice production. However, the 

levels of incentives provided to farmers are very meager compared 

to the magnitude of protection in developed countries. 

Welfare gains and welfare losses because 

of liberalization 

The results of an analysis of the impact of trade liberalization 

are presented in Table 2. These are estimated based on derived 

elasticities and on the estimated NPCs. The supply elasticity 

used for the analysis was 0.10 for rice. The loss to society 

because of liberalization of rice trade was estimated at 

around Rs. 5,800. The loss to society because of inefficiency 

in production resulting from a rise in the price was Rs. 4,200. 


Table 2. Welfare impact of rice trade liberalization. a 

Net monetary effects of price distortions on rice in Karnataka 

NSLp NSLc NSL Estimated welfare gain Estimated welfare loss Net effect of 

(Rs.) (Rs. 000) (Rs. 000) of producers, WGp of consumers, WGc Liberalization on 

(Rs. million) (Rs. million) welfare in the state 

4,200 1,600 5,800 7,718.55 2,864.35 4,854.20 

Gains and losses because of projected changes in prices resulting from 

Value of production Percentage of WGp Value of consumption Percentage of WGc to value 

at Pp (v) (Rs. million) to value of production at Pc (w) (Rs. million) of consumption 

36,219.80 21.31 41,074.00 6.97 

Production and consumption effects of price changes of selected commodities in Karnataka 

Production Consumption Estimated change Estimated change 

(million t) (million t) in production (million t) in consumption 

4.19 2.94 0.453 0.799 

a US$1 = Rs. 46.00. 

The effects of price distortion on production and consumption 

are additive with respect to trade effects. For Karnataka, the 

liberalization of agriculture would result in a change in production 

because of the change in prices. International prices 

adjusted for transportation costs were higher by 21% than domestic 

prices during the postliberalization period (2001-02). 

These higher world prices would result in increased domestic 

production of the crop to the extent of 0.453 million t of rice. 

But, higher international prices would result in a decrease in 

rice consumption of 0.799 million t. 

Welfare gains were much larger than the respective welfare 

losses. Results revealed that welfare gains to producers 

were 21.31% (Rs. 7,718.55 million) of the total value of production. 

Conversely, consumers in the state and region incurred 

substantial welfare loss because of the rise in rice prices (Rs. 

2,864.35 million) in 2001-02. 

Conclusions 

The study shows that liberalization will benefit the rice sector 

in terms of giving farmers a better deal. Consumers may have 

to pay a higher price because of the limited domestic supply 

and increase in prices. The positive impact on the farming community 

may lead to more efficient rice production and in the 

process increase the export prospects of rice. 

References 

Chand R. 2002. Trade liberalisation, WTO and Indian agriculture. 

New Delhi (India): Mittal Publications. 

Gulati A, Narayanan S. 2003. Rice trade liberalization and poverty. 

In: Mew TW, Brar DS, Peng S, Dawe D, Hardy B, editors. 

Rice science: innovations and impact for livelihood. Manila 

(Philliphines): International Rice Research Institute. p 939- 

956. 

Lutz E, Scandizzo P. 1980. Price distortions in developing countries: 

a bias against agriculture. Eur. Rev. Agric. Econ. 7:5- 

27. 

Reddy BDR. 1997. Dynamics of cropping pattern in dry zones of 

Karnataka and implications of global trade liberalisation. PhD 

thesis. University of Agricultural Sciences, Bangalore, India. 

Yao S. 1997. Comparative advantages and crop diversification: a 

policy analysis matrix for Thai agriculture. J. Agric. Econ. 

48:211-222. 

Notes 

Authors’ address: Department of Agricultural Economics, University 

of Agricultural Sciences, GKVK, Bangalore 560 065, 

India, e-mail: batlahalli@yahoo.com. 


Behavior and strategies of Japanese rice producers 

under globalization 

Masaki Umemoto 

In 2002, the domestic supply of Japanese rice consisted of 

8.876 million tons of production and 0.77 million t of minimum 

access import. Rice was cultivated on 1.688 million ha. 

To maintain the crop’s taste, farmers have not been pursuing 

high yield, but, thanks to good weather, the average rice yield 

was 5.27 t ha –1 and almost constant since 1994. On the other 

hand, domestic demand for rice decreased from 10.26 million 

t (1992) to 8.95 million t (2002). Japanese per capita rice consumption 

decreased drastically from 1960 to 1980, and this 

considerable decrease in rice consumption still continues. It is 

also important to notice that, in that period, the producers’ price 

of rice fell, corresponding to the decrease in demand (Fig. 1). 

This shows that the consumption of rice itself is diminishing. 

The decreasing demand together with the high average 

yield necessitate a reinforcement of rice supply adjustment. In 

2002, the set-aside area of rice production reached 39% of 

paddy fields. Therefore, to convert crops from rice, a lot of 

subsidy was paid to farmers. The government expended 

US$2.467 billion for this rice supply adjustment program in 

2002, and it occupied one-half of financial expenditures related 

to rice policy. Consequently, a decrease in these expenditures 

became an urgent issue. 

On the other hand, in 1992, the government designed a 

structural change in arable farming, targeting 80,000 farms cultivating 

20 ha each, with a target of sharing 60% of paddy 

fields by 2020. Nevertheless, in 2000, the share of the rice 

crop area cultivated by large-scale farms (more than 5 ha of 

rice crop area) was only 14%. This means that the accumulation 

of farmland for core farmers is stagnant and structural 

adjustment has not progressed as the government expected. 

These facts show that nowadays there are three main issues 

in Japanese rice production. First, market-oriented rice 

production is required. It is especially essential to grasp market 

needs and seek out new demand for rice. Second, the conversion 

from rice to other crops in paddy fields should be reinforced 

to achieve a rice supply adjustment. It is also important 

to increase the rate of food self-sufficiency. Third, the government 

needs to tackle reform of the production structure by encouraging 

large-scale farmers. 

Japanese rice policy reform and objectives of this paper 

To solve these problems, the government launched a Rice 

Policy Reform in April 2004. By implementing the policy, the 

government expected to cope with market globalization through 

a remarkable change from the existing rice supply adjustment 

program. The government abolished the set-aside program to 

achieve efficient land resource allocation for rice production 

Rice crop area (000 ha) 

Producers’ rice price (US$ t –1 ) 

3,000 

2,750 

2,500 

2,250 

2.000 

1,750 

Rice crop area 

Rice yield 

Set-aside area (000 ha) 

Demand for rice (10,000 t) 

Rice yield (kg ha –1 ) 

1,100 

Set-aside area 

Demand 

Rice price 

1,000 

900 

800 

700 

600 

1,500 

500 

1994 1996 1998 2000 2002 

1995 1997 1999 2001 

Year 

Fig. 1. Trend of Japanese rice supply, demand, and price. Source: 

References for rice production Production Bureau, Ministry of Agriculture, 

Forestry, and Fisheries. 

under market mechanisms by 2010. The government also aims 

to encourage large-scale farms. Therefore, the government is 

now considering a conversion from a price-support policy to 

direct payment for targeted farmers. 

This paper discusses the reactions and strategies of Japanese 

rice producers responding to various economic and political 

changes. Whether the structure of Japanese rice farming 

will change as the government hopes depends highly on farmers’ 

responses. Therefore, it is essential to understand farmers’ 

strategies in their specific socioeconomic context. Reactions 

to policy and market conditions differ according to farming 

types. Therefore, first I describe the major farming types of 

rice producers. Next, I discuss their socioeconomic characteristics 

and behavior. Finally, I conclude with the future perspectives 

of Japanese rice production and the necessary conditions 

for changes in agricultural structure. 

Major farming types of rice producers and their behavior 

Japan has various types of rice producers such as the part-time 

family farm, large-scale family farm, agricultural corporate 

firm, group-farming organization based on the rural community, 

farm contractor, rural agricultural public corporation, and 

agricultural cooperative. However, the following four types 


are the major rice producers: part-time family farm, groupfarming 

organization based on the rural community, large-scale 

family farm, and agricultural corporate firm. 

The part-time family farm is still a major rice producer 

in Japan, and its number was 2.76 million in 2000. It works on 

rice production during days off and earns its main income from 

nonagricultural jobs. The part-time family farm has a priority 

to maintain paddy fields as the family estate inherited from 

predecessors and not to consider profits from rice production. 

Therefore, this farm might increase rice crop areas to correspond 

to the abolishment of the set-aside program, and, even 

if the rice price falls, some of them will continue rice production. 

But, the number of part-time family farms is gradually 

decreasing according to the aging of farmers. The young generation 

especially hesitates to do work with the rice crop. 

Group-farming organizations based on the rural community 

mainly consist of part-time family farms, but their activities 

can be regarded as an improvement of productivity and 

reconstruction of regional farming. The common features of 

group farming are the collective ownership of machinery and 

facilities, joint work on rice cultivation, and regional adjustment 

of land use. The purposes of this group farming are to 

reduce production costs and maintain paddy fields. The number 

of group-farming organizations was only 9,961 in 2000. 

Recently, however, the leading actors of rural areas such as 

the personnel of agricultural cooperatives or agricultural extension 

services are actively encouraging farmers to organize 

group farming. Meanwhile, the government is promoting groupfarming 

organizations to formalize as a corporate body because 

the government assumes that the corporate body is a 

more suitable economic entity than the group-farming organization. 

It is expected that the number of group-farming organizations 

will increase, resulting in a rationalization of rice 

production, while the conversion of group-farming organizations 

to corporate bodies that pursue profit seems difficult. 

Large-scale family farms and agricultural corporate firms 

are keenly interested in increasing farm size and diversifying 

their business vertically and/or horizontally. Both types are 

expected to be leading rice producers. The large-scale family 

farm is mainly managed and worked by family labor. Meanwhile, 

agricultural corporate firms employ nonfamily members 

and assign them some management work. The farm sizes 

of large-scale family farms are 10 to 20 ha, and two or three 

family members usually cultivate rice, wheat, soybean, and 

feed crops. On the other hand, agricultural corporate firms are 

50 to 100 ha, and their investment in machinery and facilities 

is three to five times as much as that of the family farm. 

The large-scale family farm mainly aims at expanding 

rice, wheat, and soybean crop area to reduce production costs 

and increase farm income. Meanwhile, agricultural corporate 

firms are pursuing both diversified agribusiness and expansion 

of farm size. Through these strategies, they manage to 

keep enough working time for full-time employees. 

These two types have several different characteristics, 

but both equally reinforce marketing. First, they usually sell 

polished rice to consumers or retail stores directly. Second, 

some of them address the production of organic rice and/or 

chemicals used in rice. Japanese consumers have recently paid 

attention to food safety and to labeling (production areas, varieties), 

taste, and price level. So, to meet these consumers’ 

needs, producers are trying to supply branded products. Third, 

they are introducing a traceability system for their products 

because many Japanese consumers require information on the 

quality of rice and its cultivation methods. Finally, they invite 

their customers (consumers, retailers) to their farms for interpersonal 

contact to keep clients in good standing. These Japanese 

rice producers’ strategies are effective for advantageous 

marketing, resulting in higher profitability of rice production. 

The future of Japanese rice production 

From the review of major rice producers’ strategies, we summarize 

our conclusions. To resolve the first three issues, such 

as the decrease in rice demand, conversion from rice crops in 

paddy fields, and structural change, it is important to encourage 

the large-scale family farms and agricultural corporate firms 

by supporting their strategies. Therefore, whether Japanese rice 

farming copes with globalization and the changes in farming 

circumstances depends on how much and how fast those viable 

producers dominate production. 

But the Rice Policy Reform, which abolished the compulsory 

set-aside program, might lead to an increase in supply, 

resulting in a drop in price. Core farmers are most susceptible 

to a drop in price. Figure 2 shows the effects of a price drop 

for large-scale farms. The Rice Policy Reform prepared a compensation 

program for rice income in the case of a price drop, 

but the amount that farmers receive from that policy is limited, 

within about 10% of existing rice income. Therefore, even for 

a 20-ha farm, if the rice price declines to under $2.70 per kg 

(the price level in September 2004), it will be difficult to gain 

comparable income with nonagricultural work. Especially in 

the case of smaller farms (15 ha, 10 ha), at each price they 

couldn’t get enough income. Additionally, if the government 

abolishes the financial support for farmers who participate in 

the set-aside program, farm income will decline according to 

the price decline (Fig. 2). 

Therefore, it is important to accelerate the accumulation 

of farmland for farmers who pursue expanded farm size, and, 

to avoid a decline in farm income corresponding to a drop in 

price, the government needs to prepare an effective direct income 

compensation policy for core farmers immediately. 

Notes 

Author’s address: National Agricultural Research Center, Kannondai 

3-1-1, Tsukuba, Ibaraki, Japan, e-mail: umemoto@affrc.go.jp. 


Income of full-time labor (US$ thousand year –1 ) 

72.7 

Farm land size (20 ha) 

63.6 

Income of nonagricultural worker 

per year ($54,500) 

54.5 

45.5 

36.4 


27.3 

18.2 


9.1 

Farm land size (30 ha but no subsidy) 

2.7 2.6 2.4 2.3 2.1 2.0 1.8 1.7 1.5 

Rice price (US$ kg –1 ) 

Fig. 2. Effects of a decline in price. Note: the numbers in this figure were calculated by using linear 

programming methods. The farm model is as follows: two family laborers. Crops are rice, wheat, and 

soybean. The farmer receives a subsidy from the government for the wheat or soybean crop in the 

paddy field (US$4,545 ha –1 ) and rice income compensation within 10% of existing income, but the 

nonsubsidized farmer receives nothing. 

The future perspective of upland rice farmers in Indonesia 

in the era of globalization 

Yusman Syaukat and Sushil Pandey 

Rice is the dominant staple food for about 220 million people 

in Indonesia. Indonesia achieved self-sufficiency in rice in 1984 

following the intensive effort of the government through various 

public assistance programs for increasing rice production. 

Unfortunately, rice self-sufficiency could not be maintained 

for long and the import volume slowly crept up over time. In 

2003, rice imports reached 2 million metric tons, which made 

Indonesia the biggest rice-importing country in the region 

(Sidik 2004). 

Land is the basic resource for agricultural activities. The 

upland and rainfed lowland play an important role in the agricultural 

sector of Indonesia, since a vast area of cultivated land 

(about 65.8%) is under rainfed conditions (BPS 2003). The 

existence of upland is dominant on all the main islands of Indonesia 

(Table 1). In Java, the proportion of upland area is 

about 33.5%. However, the high incidence of rural poverty, 

instability of production, and fragility of the environment are 

among the important characteristics of rainfed upland areas 

(Partohardjono 1993). 

Upland rice is one important product of upland agriculture. 

In 2002, the total production of upland rice reached 2.6 

million tons, whereas lowland rice produced 45.2 million t. 

Sumatra and Java islands account for the major share of upland 

rice area and production in Indonesia. Total upland rice 

area of Indonesia is 1 million ha and upland rice accounts for 

less than 10% of the total rice area of Indonesia. Government 

efforts to achieve self-sufficiency in rice during the 1980s focused 

on lowland irrigated rice. Upland rice has remained 

somewhat neglected in the public-sector assistance programs 

of Indonesia. 

In upland areas, where the opportunity costs of land and 

labor are relatively low, Indonesia may still have a comparative 

advantage in upland rice production. The objectives of 

this paper are (1) to evaluate the importance of upland rice 

and its relation with poverty level and (2) to discuss the likely 

impact of trade liberalization on upland rice production and 

the livelihoods of upland farmers. 

Characteristics of upland rice production 

Upland rice production is characterized by tiny percentages of 

marketed output, low yields, and relatively low levels of use 

of modern inputs. Upland rice production activities involve a 

large number of Indonesians. Although the contribution of 

upland rice to the total national rice production is relatively 

low, it has an important role in supporting farmers’ livelihoods. 

Most upland farmers cultivate upland rice mainly to meet their 

family needs. Because of the small area of land ownership, 

low yield, and lack of government support, most upland rice 

farmers are poorer than their wetland counterparts. 


Table 1. Total area of agricultural land in Indonesia by islands, 2003 (excluding Maluku 

and Papua). 

Island 

Total area (ha) 

Irrigated land Rainfed land Upland Total 

Sumatera 1,691,544 710,997 7,097,329 9,499,870 

17.81% 7.48% 74.71% 100.00% 

Java 4,663,175 879,417 2,792,861 8,335,453 

55.94% 10.55% 33.51% 100.00% 

Bali and Nusa Tenggara 708,886 89,283 1,471,482 2,269,651 

31.23% 3.93% 64.83% 100.00% 

Kalimantan 321,534 413,829 6,687,296 7,422,659 

4.33% 5.58% 90.09% 100.00% 

Sulawesi 1,116,706 327,196 2,953,964 4,397,866 

25.39% 7.44% 67.17% 100.00% 

Indonesia 8,501,845 2,420,722 21,002,932 31,925,499 

26.63% 7.58% 65.79% 100.00% 

Source: BPS (2003). 

Following trade liberalization agreements under the Uruguay 

Round (signed in 1994), the government of Indonesia 

deregulated trade in food crops, including rice. In line with 

these agreements, Indonesia has abolished input subsidies, 

output price protection, and reduced tariffs for rice. The resulting 

price disparity between the international and domestic 

retail price has increased the imports of cheap rice into Indonesia 

(Sidik 2004). Although imports of cheap rice are expected 

to have a positive effect on the poverty of poor consumers, 

commercial rice farmers will have to adjust production to the 

prevailing price regime. 

Upland rice, farmer livelihoods, and poverty 

Upland rice is still an important food in the hilly region where 

farmers have limited access to lowlands, irrigation facilities 

are limited, and market access is poorly developed. Under these 

conditions, the opportunity costs of land and labor become 

relatively low. Subsistence production of upland rice is observed 

under such conditions. Though upland is considered to 

be “marginal land,” it plays an important role since farmers 

use it to cultivate rice and secondary crops. 

Upland farmers tend to have a different perspective in 

growing rice than wetland farmers. For wetland farmers, rice 

is a commercial commodity. Wetland farmers grow rice to 

obtain profits by selling their products to the market. In contrast, 

upland farmers mostly grow rice to make sure that they 

obtain enough production to meet their food requirements. They 

sell rice only if there is a surplus of production over consumption. 

Poverty is a significant and persistent problem in the 

agricultural sector of Indonesia. About 23.7 out of 37.1 million 

poor people of Indonesia (63.9%) are those who work in 

agriculture (BPS 2002). The relationship between the percentage 

of upland rice area (UpArea, i.e., ratio between total area 

of upland rice and the total area of both upland and wetland 

Percent 

70 

Upland rice 

60 Wetland rice 

50 

40 

30 

20 

10 

0 

Kept as seeds Consumed Paid as land rent Sold 

Fig. 1. Comparision of upland and wetland rice disposal. 

rice) and the incidence of poverty (Poor) using district-level 

data is estimated as 

Poor = 17.5336 + 0.0456 UpArea 

(30.721) (1.895) R 2 = 0.102 

The results indicate that there is a positive and significant (at 

the 5% level) relationship between the percentage of poverty 

and percentage of upland area in each district, meaning that 

every 1% increase in the area of upland rice is associated with 

an increase in poverty incidence of 0.05%. 

A survey of upland rice farmers in West Java was conducted 

in 2003 to generate household-level data regarding 

upland rice systems. The main reasons given by farmers for 

growing upland rice were (1) to meet their household needs 

(80%) and (2) because of “custom”—they just know (upland) 

rice activities (9%). Allocation of upland rice production is 

mainly to meet household needs (66%) (Fig. 1). This indicates 

the importance of upland rice in farmers’ livelihoods. They 

grow upland rice to make full use of the available family labor 

and land. Even if lowland rice is available in the market, its 


purchase is limited because of the lack of income and the subsistence-oriented 

livelihood strategy. 

Future prospects for upland rice 

As upland rice is a subsistence crop, its production is unlikely 

to be affected directly by the availability of cheap imported 

rice following trade liberalization. Survey results indicate that 

farmers who don’t have sufficient purchasing power will continue 

to grow upland rice even when cheap rice is available in 

the market at a low price. In years when the production of 

upland rice is low, farmers simply substitute other crops such 

as cassava and maize for upland rice instead of purchasing 

rice in the market, mainly because of a lack of income. So, the 

effect of trade liberalization on upland rice is likely to depend 

more on how it will affect overall farm income. As trade in 

agricultural commodities is liberalized, other commercial crops 

such as rubber, coconut, and cloves that substitute for upland 

rice may become more profitable. A shift toward these crops 

may gradually lead to less reliance on upland rice as farmers’ 

income and purchasing capacity improve. 

Only the poorest of the poor who are dependent on upland 

rice for their livelihoods and those who lack other means 

of enhancing their income are likely to continue to grow upland 

rice. Hence, although there is likely to be some shrinkage 

in upland rice area in the wake of trade liberalization, the reduction 

in upland rice area is likely to be proportionately less 

than that for lowland rice area. However, in the long run, as 

adjustments in the prices of crops that are substitutes for upland 

rice occur in response to trade liberalization, upland rice 

may lose its comparative advantage altogether. 

References 

BPS (Central Statistics Agency of Indonesia). 2002. Data dan 

informasi kemiskinan tahun 2002: buku 2 - kabupaten. Jakarta, 

Indonesia. 

BPS. 2003. Land area by utilization in Indonesia: agricultural survey. 

Jakarta, Indonesia. 

Partohardjono S. 1993. Upland agriculture in Indonesia: recent trends 

and issues. In: Bottema JWT, Stoltz DR, editors. Upland agriculture 

in Asia. Proceedings of a workshop held in Bogor, 

Indonesia, 6-8 April 1993. 

Sidik M. 2004. Indonesia rice policy in view of trade liberalization. 

A paper presented at the FAO Rice Conference held in Rome, 

Italy, 12-13 February 2004. 

Notes 

Authors’ addresses: Yusman Syaukat, Department of Agricultural 

Socio-Economics, Bogor Agricultural University, Bogor, Indonesia, 

e-mail: ysyaukat@indo.net.id; Sushil Pandey, International 

Rice Research Institute, Philippines, e-mail: 

sushil.pandey@cgiar.org. 

Impact of globalization on rice farmers in Thailand 

Rangsan Pitipunya 

Globalization makes the world smaller and affects people’s 

lives in many aspects. With the fast development of information 

technology (IT), our way of life has changed drastically. 

IT has markedly changed our way of life, economics, society, 

and policy worldwide. Every country has to prepare itself to 

cope with this change. Besides IT development, the World 

Trade Organization (WTO), which aims to create efficient and 

fair international trade, has been established. To promote agricultural 

trade, three measures, market access and tariff reduction, 

domestic support reduction, and export subsidy reduction, 

have been implemented since 1995. In general, it is believed 

that the WTO and IT will boost the volume of agricultural 

trade, which will create benefits for many people, especially 

farmers. However, in Thailand, this belief is questioned; 

therefore, it is crucial to discuss the impact of the WTO and IT 

on farmers. Because the number of rice farmers in Thailand is 

the most compared to other farmers in the country, this paper 

will focus only on the impact of the WTO and IT on rice farmers. 

Thailand, the WTO, and IT 

Thailand and the WTO 

As a member of the WTO, Thailand committed to the following: 

1. Reduce average tariffs on 740 tariff lines by 24% in 

10 years (1995-2004), with a minimum reduction of 

10% on all tariff lines. The average tariff rate will 

decrease from 49% in 1995 to 27–40% (37% on average) 

in 2004 (DTN 2004). 

Thailand opened its domestic rice market in two 

ways. The first one was to set import tariff rate quotas 

at 30%. The quota amounts were set at 237,863 

and 249,757 tons in 1995 and 2004, respectively. The 

second was to reduce tariffs for imported rice (more 

than quota) from 58% in 1995 to 52% in 2004. For 

processed rice products, there is no tariff quota, but 

there will be a tariff reduction and commitment to an 

import tariff rate. 


2. Reduce domestic support by 13% in 10 years. In 1999, 

rice production support was US$333.03, dropping 

from $354.53 in 1995 (DTN 2004). 

Development of information technology 

Thailand has been using IT for about 20 years. However, this 

technology was not widely used until 1995, when the Thai government 

established its IT 2000 policy framework for IT development 

from 1995 to 2000, with three fundamental prerequisites 

that must necessarily be in place and function together: 

(1) a national information infrastructure, (2) a well-educated 

populace and adequate IT labor, and (3) good governance with 

a dare to dream and a resolve to act. The termination of IT 

2000 was followed by a second, called IT 2010, which was 

approved by the cabinet in March 2002. The IT 2010 policy 

framework, implemented from 2001 to 2010, set a higher goal 

of transforming the Thai economy into a knowledge-based 

economy (Krutkaew 2004). The Ministry of Information Technology 

and Communication has been established to take these 

responsibilities. An IT master plan has also been set up to promote 

widespread IT use. Consequently, IT has been used widely 

in the last 2–3 years. 

The impact of the WTO and IT on Thai rice farmers 

It was expected that Thailand should have been able to increase 

its amount of rice exports because its climate is suitable 

for rice cultivation. Many millions of tons of rice are left over 

that can be exported to the world market every year. According 

to the WTO agreement, its members have to reduce domestic 

rice production support, export subsidies, and other trade 

barriers. This is a good chance for Thailand to increase its 

amount of rice exports. Consequently, the domestic rice market 

had to adapt itself to new conditions that certainly affect 

Thai rice farmers. In addition, at present, IT has been widely 

used and is also another factor affecting farmers’ life. 

For rice exports, it was found that the amount of exported 

rice decreased in 1996. It dropped from 6,198 million 

tons in 1995 to 5,460 million tons in 1996. After that, the 

amount of rice exports has obviously increased. Even though 

it fell in some years, it climbed to over 7 million tons since 

2001 and in 2003 the amount of exported rice was 7.343 million 

tons. It can be seen that income from rice exports has 

expanded since 1995. Income was $1,948.19 million, which 

rose from $1,555.05 million in 1994. Then, it gradually declined 

to about $1,600 million and again climbed to $1,846.84 

million in 2003 (Table 1). 

The variation in export income depended not only on 

the export amount but also on the export price, which had a 

close relationship with the farm price and profit. Table 1 showed 

that farmers could obtain profit from rice production almost 

every year. In 1995, the average profit from rice was $17.10 

per ton. After that, profit rose to more than $40 per ton in 1996 

and remained at a high level until 1999. Since then, profit declined 

to $6.20 per ton in 2000. Because the product price 

dropped drastically, farmers lost $4.72 per ton in 2001. After 

Table 1. Important information on rice production in Thailand. a Year 

Items 

1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 

Milled rice export 

Amount (million t) 4.86 6.20 5.46 5.57 6.54 6.84 6.14 7.69 7.33 7.35 

Value (million US$) 1,555.05 1,948.19 1,998.22 2,074.85 2,098.29 1,950.59 1,631.38 1,577.46 1,629.41 1,846.84 

Price ($ t –1 ) 320.06 314.33 365.96 372.68 320.83 285.22 265.64 205.10 222.16 251.41 

Farm price ($ t –1 ) 149.62 161.74 197.05 166.32 159.63 143.33 114.99 97.40 108.98 117.68 

Net profit ($ t –1 ) 15.54 17.10 46.59 40.87 61.12 27.16 6.20 -4.27 5.88 6.65 

Total product of paddy (t) 

Wet season 16.48 18.16 17.73 17.78 18.79 18.66 19.02 19.79 20.90 19.63 

Dry season 1.97 2.95 4.29 4.55 4.79 4.34 5.16 6.06 5.62 6.43 

Average product of paddy (t ha –1 ) 

Wet season 1.83 2.01 1.93 1.94 2.06 2.07 2.10 2.14 2.26 2.16 

Dry season 3.96 4.28 4.51 4.42 4.14 4.20 4.10 4.34 4.17 4.21 

Cultivated area (million ha) 

Wet season 8.98 9.02 9.19 9.17 9.11 9.00 9.05 9.24 9.25 9.11 

Dry season 0.50 0.69 0.95 1.03 1.16 1.03 1.26 1.39 1.35 1.53 

Sources: Office of Agricultural Economics, own calculations. 


that, farmers gained profit from rice production again, but it is 

rather low. 

Because of earning profit, Thai rice farmers expanded 

their production by improving productivity and expanding 

planting area. For productivity, Table 1 showed that since 1994 

productivity has not been increasing steadily. In the rainy season, 

yield fluctuated from 1.83 to 2.26 t ha –1 , whereas, in the 

dry season, it was 3.96 to 4.50 t ha –1 . The main factor affecting 

rice productivity was overall environmental conditions, 

particularly rainfall. Production technique, particularly planting 

method, compared between before and after being a WTO 

member, is similar. However, because of the health hazard resulting 

from chemical substances, a few farmers have begun 

nonchemical practices in rice cultivation lately. 

Even though the expansion of rice production by increasing 

productivity was not distinct but in the case of increasing 

planting area was obvious, particularly in the dry season, planting 

area increased about 0.093 million ha per year. It increased 

from 0.50 million ha in 1994 to 1.53 million ha in 2003. As a 

result, rice production in the dry season in the same period 

rose from 1.97 million to 6.426 million t. Because of land limitations, 

the planting area in the rainy season increased slightly 

from 8.98 million ha in 1994 to 9.11 million ha in 2003. Therefore, 

the total product also increased a bit from 18.161 million 

t to 19.631 million t. 

As mentioned above, it seems that being a WTO member 

provides a good opportunity for Thai rice exports and enables 

farmers to expand their production. However, there is a 

question whether the increase in rice exports since 1995 resulted 

from being a WTO member or not. Poapongsakorn and 

Corman (2004) indicated that the increasing rate of rice exports 

in 1995-2000 (after being a WTO member) was 3.88% 

per year lower than that during 1985-1994, when exports increased 

4.15% per year. Furthermore, it was noted that the 

market share of Thai rice in developed countries after 1995 

seemed to decrease, particularly in Japan and the European 

Union. In addition, statistical analysis cannot confirm the significant 

impact of agreements under the WTO framework on 

the expansion of Thai rice exports. Besides, some additional 

observations found that the expansion of Thai rice exports may 

be due to the depreciation of the Thai currency (baht) because 

of the economic crisis. Pouthipong (2003) indicated that the 

foreign exchange rate was a significant factor affecting demand 

for Thai rice exports. 

Regarding IT, it has been found that the technology affected 

people living in big cities who have a high education 

and young age. But its impact on rice farmers who are poor, 

have a low education, and are old was not clear. However, in 

the near future, when IT infrastructure has been completely 

set up throughout the country, it is believed that the technology 

will become an important tool of Thai rice-farming development. 

Conclusions 

It still cannot be concluded that the WTO is a factor affecting 

the expansion of the Thai rice market. However, after being a 

member of this organization, the trend of Thai rice exports has 

been increasing. Some research results revealed that Thai competitiveness 

in the world rice market has increased (Tawhon 

2003). Moreover, the WTO is a forum for its members to negotiate 

and reach agreement. Therefore, Thailand should cooperate 

with this organization closely. Creating subgroup members 

having the same or relevant problems should be carried 

out continuously to strengthen bargaining power. 

Though the impact of IT on Thai rice farmers has not 

been seen clearly, in the near future it is expected to be an 

important tool in the development of rice farmers and the agricultural 

sector. Therefore, Thai rice farmers should prepare 

themselves to cope with this kind of technology. 

References 

DTN (Department of Trade Negotiations). 2004. Negotiation on 

agricultural product under WTO and FTA. Bangkok (Thailand): 

DTN. 205 p. 

Krutkaew C. 2004. Strategy and direction of information and communication 

technology development in Thailand. Exec. J. 

24:26-30. 

Poapongsakorn N, Corman S. 2004. Multilateral negotiation on 

agricultural product: unbalance, failure, and future of Doha 

Round. In: Thai economy and multilateral negotiation. Proceedings 

of Symposium 27, 17 August 2004. Bangkok (Thailand): 

Faculty of Economics, Thammasart University. p 5-1– 

5-100. 

Pouthipong P. 2003. A study of export demand for Thai rice before 

and after devaluation. Bangkok (Thailand): Kasetsart University. 

68 p. 

Tawhon K. 2003. A composition of factors affecting rice export 

growth of Thailand. Bangkok (Thailand): Kasetsart University. 

174 p. 

Notes 

Author’s address: Kasetsart University, Bangkok, Thailand, e-mail: 

fecorsp@ku.ac.th. 


The rice economy and rice policy in China 

Li Ninghui 

China is the largest country that produces rice in the world and 

also the largest country that consumes rice. China’s rice production 

plays a very important role not only in China’s grain 

economy but also in the world rice economy. 

Table 1 shows that China produces more than 30% of 

the rice with around 20% of the world’s paddy area. This implies 

that the yield of rice in China is higher than the average 

in the world. However, these shares have been going down. 

This downward trend mainly results from the Chinese 

government’s policy of structural adjustment of rice production 

to enhance the production of high-quality rice while reducing 

low-quality rice in response to the change in consumers’ 

behavior in rice consumption. Chinese demand for rice 

consumption is shifting from quantity to quality as economic 

development occurs. Owing to the progress of science and technology 

for agricultural production, China’s grain supply versus 

demand has changed from chronic shortage to surplus in a 

good harvest year, keeping a balance in a normal year and a 

temporary shortage in a bad harvest year, since the end of the 

1990s. Thanks to the good rice harvest in the second half of 

the 1990s, the surplus grain supply gave the Chinese government 

a good chance to smoothly implement policies of adjusting 

the agricultural production structure, including the rice 

production structure, to meet consumers’ diversified demand. 

Figure 1 tells us that we can categorize rice production 

and consumption into three stages. The first stage was the period 

before the mid-1980s, when production and consumption 

rose rapidly and in the same phase. During that period, production 

and consumption were almost balanced. The second 

stage was the period from the mid-1980s to the end of the 

1990s, when production rose with a large fluctuation and consumption 

went up smoothly. During that period, the rice surplus 

and deficit switched several times, with a large surplus in 

the end. The third stage was the period starting at the beginning 

of the 2000s, when both production and consumption went 

down, but production at a faster speed. Fortunately, the trend 

Table 1. China’s share of the rice economy. 

China’s share of rice 1970 1980 1990 1999 2000 2001 

In the domestic grain economy (%) 

Cultivated area 27.1 28.9 29.1 27.6 27.6 27.2 

Output 45.8 43.6 42.4 39.0 40.7 39.2 

In the world rice economy (%) 

Cultivated area 24 23 23 20 19.4 19.0 

Output 36 38 38 34 31.3 30.0 

Data source: National Bureau of Statistics of China. 

(000 t) 

150,000 

140,000 

Production 

Consumption 

130,000 

120,000 

110,000 

100,000 

90,000 

1980 

1982 

1984 

1986 

1988 

1990 

1992 

1994 

1996 

1998 

2000 

2002 

Year 

Fig. 1. Milled rice production and consumption from 1980 to 2002. 


of a sharp decrease in rice production in recent years has been 

halted and recovered, according to a report from the National 

Bureau of Statistics. 

The Chinese government’s rice policy targeted high output 

for a long time in order to feed China’s huge population, 

with limited arable land. In the 1960s, a new variety of rice, 

called semi-short-stem rice, was cultivated and extended in 

China, whose output relies heavily on fertilizer use, while the 

Green Revolution was being experienced in other countries 

during the same period. In the early 1980s, the cultivated area 

of upgraded varieties of rice, which were mainly conventional 

high-yielding varieties and hybrid varieties, accounted for more 

than 98% of the total cultivated area of rice. In the late 1970s 

and 1980s, varieties with disease resistance were cultivated 

and extended in China. 

However, rice production aimed at high yield has faced 

a growing challenge since the early 1990s. After the opening 

of the domestic rice market to consumers starting in 1993, the 

cultivated area of hybrid rice gradually stopped going up, and 

even shrank because the quality of hybrid rice was relatively 

low. The demand for high-quality rice significantly affected 

regional and varietal distribution (i.e., indica and japonica) of 

rice production, which resulted in the rapid expansion of cultivated 

area in northern China, especially northeast China, 

where japonica rice is produced. The share of rice area in northern 

China increased from less than 6% in the early 1980s to 

Table 2. Structural change in rice production of China in 1980-2000. a 

Cultivated area share Cultivated area share Share of 

Year by variety (%) by region (%) hybrid rice 

area (%) 

Indica Japonica South North 

1980 89 11 94 6 14 

1985 88 12 93 7 26 

1990 84 16 90 10 49 

1995 79 21 89 11 52 

2000 73 27 86 14 50 

a Cultivation of hybrid rice reached a peak in 1991-92. Cultivated area of hybrid rice 

accounted for 40% of the total cultivated area of rice. 

10% in 1990 and 14% in 2000. In some provinces along the 

middle reaches and downstream of the Yangtze River, such as 

Jiangsu, Zhejiang, Shanghai, and Anhui, where indica rice is 

the traditional product, farmers produce more and more 

japonica, even surpassing indica. This resulted in the share of 

japonica area increasing from 11% in 1980 to 16% in 1990 

and 27% in 2000. 

By varieties, we see that in the 1960-70s, a group of 

precocious early rice and cold-resistant late rice was cultivated 

in China, which resulted in the extension of double-cropped 

rice and increased total output greatly. 

Double-cropped rice yielded 7.04 t ha –1 in 1975, 56% 

higher than the yield of single-cropped rice (4.43 t ha –1 ). This 

difference was 63% in 1990-94, when the yield of singlecropped 

rice increased to 6.63 t ha –1 and the yield of doublecropped 

rice increased to 10.63 t ha –1 . This big difference was 

the main reason that the cultivated area of double-cropped rice 

expanded under the guidance of the policy for output. 

The output growth of double-cropped rice, however, was 

obtained at the cost of two times more labor input and fertilizer 

input per unit of paddy field than those in single-cropped 

rice. 

The rapid development of the rural economy and growth 

of rural household income enhanced the opportunity cost of 

rural labor. In response to the change in production environments 

and conditions, the Chinese government adjusted the 

policy that encouraged multiple-cropped rice. As a result, the 

cultivated area of double-cropped rice decreased from 66% of 

the total area in 1980 to 58–60% in 1990. 

Rice policy in China 

At present, the new policies that have been implemented for 

rice production are mainly the following: 

Protection of necessary paddy fields. With this policy, 

if a field is determined to be necessary as a paddy 

field, it will be forbidden to be used in nonagricultural 

activities, but it is allowed to be changed to plant 

another crop so that it can be reused as a paddy field 

in case it is needed in the future. 

Table 3. Cultivated area and yield of early rice, late rice, and single-harvest rice. 

Cultivated area distribution (%) Yield (t ha –1 ) 

Double-cropped Single-cropped Double-cropped Single-cropped 

Year rice rice rice rice 

Early Late Middle rice and Early Late Middle rice and 

rice rice one-season late rice rice rice one-season late rice 

1975 36 35 29 3.80 3.24 4.43 

1980 33 33 34 4.43 3.29 4.64 

1985 30 30 40 5.10 4.65 5.96 

1990 28 30 42 5.49 5.13 6.50 

1994 27 33 40 5.11 5.53 6.55 

Data source: China Rice Research Institute, 1993. Agricultural Yearbook of China, 1990-1995. 


Construction of a rice production base. The unit of a 

base is a county. There are about 200 bases so far. 

The central government provides 2 billion yuan of 

favorable loans to support rice production in these 

bases. 

Support of natural rice production. The standard of 

natural rice production has been established, and natural 

rice production bases have been determined and 

are being constructed in Heilongjiang, Jilin, Zhejiang, 

and Jiangsu. The trend is that these bases will be expanded 

soon. 

Support of research and extension of rice production 

science and technology. The Chinese government 

strengthened its financial input into research on “super 

rice” and rice genes. 

The new policies that have been implemented 

for rice procurement are mainly the following: 

Cancellation of the policy that procured early indica 

rice at a protected price in 2000. In the early 1990s, 

the cultivated area of early indica rice accounted for 


Notes 

 

25% of the total cultivated area of rice, and this output 

accounted for 10% of the total output of rice. 

However, on the demand side, most consumers didn’t 

like early indica rice because its taste was not so good. 

Therefore, the government cancelled the protected 

price for procuring this product in order to get farmers 

to reduce its production. 

High quality and high price. The Chinese government 

increased the procurement price of high-quality 

rice to encourage its production. To carry out this 

policy and enforce the management of rice quality, 

the Chinese government implemented a new standard 

of rice quality in 2000 to replace the old one implemented 

in 1986. 

Author’s address: Institute of Agricultural Economics, Chinese Academy 

of Agricultural Sciences, Beijing, China, e-mail: 

linh@mail.cass.net.cn. 

The theme of session 18 was the “Impact of globalization on rice 

farmers.” We focused on farmers’ response to the progression of 

globalization. Globalization is usually treated as a matter of WTO 

agricultural negotiations on agricultural economics. Globalization 

is the liberalization and deregulation of the market system, but it 

also fosters the interactions among policy, the economy, information, 

and culture of each country. Especially, the information 

technology (IT) revolution will strongly affect agricultural production. 

Therefore, it is important to consider globalization from a 

wider perspective. 

The session focused on three major topics: 

1. What are the major farming types of rice farmers (rice 

producers) and what are their socioeconomic characteristics 

in each country 

2. How do rice farmers face changes in economic conditions 

brought about by globalization, and how do they 

react to those changes 

3. What are the future perspectives of rice farmers in each 

country 

From these viewpoints, we explored the conditions of rice 

farmers among Asian countries. 

With this concept, we held two symposia. Symposium 1 

had four presentations: a comparative analysis for total factor 

productivity in 33 rice-producing countries from 1961 to 2002 

(“Impacts of trade liberalization on competitiveness and productivity 

in the rice sector,” M. Rakotoarisoa, USA), an empirical 

study on Japanese group faming (“New strategy of group farming 

in Japan,” K. Miyatake, Japan), research on Indonesian rice farmers 

under globalization (“The role of the rice economy after the 

implementation of agricultural policy reform and trade liberaliza- 

tion from the perspective of farmers,” Roosiana, Indonesia), and 

the classification of areas against climatic shocks (“El Niño sensitivity, 

resource endowment, and socioeconomic characteristics: 

the case of wetland rice in Java, Indonesia,” S. Yokoyama, Japan). 

These analyses indicated that current problems in rice production 

and the impacts of globalization on rice farmers really 

differed among countries and areas. They also indicated that farmers 

were undertaking regionally specific behavior regarding their 

socioeconomic circumstances. Questions and discussion for these 

presentations mainly concentrated on how researchers grasped 

that diversity. A statistical approach and field research were applied 

but a detailed farming survey will be indispensable for an 

empirical study on rice production and rice producers’ behavior. 

Symposium 2 had four presentations by speakers from Japan, 

Indonesia, Thailand, and China. They reported their countries’ 

main issues on rice production and rice farmers. 

M. Umemoto (“Behaviors and strategies of Japanese rice 

producers under globalization”) showed the main types of Japanese 

rice producers and their behaviors. Next, he analyzed the 

impact of the rice price decline on large-scale family farms and 

agricultural cooperative firms. Finally, he pointed out the need 

for an effective income compensation policy in Japan. 

Y. Syaukat (“Future perspectives of upland rice farmers in 

Indonesia in the globalization era”) analyzed the impact of globalization 

on Indonesian upland rice farmers. He explained that 

upland rice is an important product of upland agriculture, but it 

has remained somewhat neglected in public-sector assistance 

programs. Then, he concluded that only the poorest of the poor 

who are dependent on upland rice for their livelihoods are likely 


to continue to grow upland rice. In the long run, with adjustments 

in crop prices in response to trade liberalization, upland 

rice could lose its comparative advantage. 

R. Pitipunya (“Impact of globalization on rice farmers in 

Thailand”) discussed the impacts of trade liberalization on Thai 

rice farmers. He paid special attention to the impacts of the IT 

revolution. His statistical analysis had shown that the significant 

impact of the WTO on rice exports was not confirmed, and the 

expansion of Thai rice exports might have resulted from the depreciation 

of the Thai currency. He then mentioned that IT was 

expected to be an important tool in agricultural development in 

the near future; therefore, farmers have to prepare to cope with 

this technology. 

Li Ninghui (“Rice economy and rice policy in China”) analyzed 

the trend of Chinese rice supply and demand. He stated 

that, in China, rice production and consumption have decreased. 

Especially, the amount of rice production had dropped drastically. 

He also pointed out that, after the opening of the domestic 

rice market in 1993, the cultivated area of hybrid rice gradually 

declined because its quality was relatively low. The demand for 

high-quality rice significantly affects the regional layout and variety 

layout of rice production. He then introduced the new rice 

policy in China: protection of necessary paddy fields, construction 

of a rice production base, support of natural rice production, 

and cancellation of the policy that procured early indica-type rice. 

All of this would encourage the production of high-quality rice. 

These reports discussed the various problems related to 

rice production and rice farmers in each country, especially referring 

to socioeconomic conditions. With these presentations, questions 

mainly focused on three issues. First was the evaluation of 

government policy regarding globalization. Next were the prospects 

for the trend of international rice supply and demand. It 

was recognized that whether China, with its huge population, 

would continue to be self-sufficient in rice or not could become a 

serious problem for the international rice market. Finally were 

the prospects for rice export competitiveness. Because R. 

Pitipunya reported low productivity of Thai rice, the possibilities 

for it were discussed. 

It was concluded that it was difficult to summarize briefly 

the theme of this session, the impact of globalization on rice 

farmers, because rice production possesses specific and varied 

socioeconomic characteristics in each country. The theme of this 

symposium overrides other important issues in each country’s 

agriculture; therefore, it should be treated as a long-term subject. 

To clarify those specific problems, it is important to tackle 

with comparative analyses various countries’ rice production aspects. 

Therefore, cooperative and continuous research would be 

effective. 


SESSION 19 

Climate change and rice production 

CONVENER: T. Imagawa (NIAES) 

CO-CONVENERS: T. Hasegawa (NIAES), Y. Hayashi (NIAES), 

and J. Sheehy (IRRI)

Effects of elevated atmospheric CO 2 concentration 

and increased temperature on rice: implications 

for Asian rice production 

T. Horie, H. Yoshida, S. Kawatsu, K. Katsura, K. Homma, and T. Shiraiwa 

It is predicted that the carbon dioxide concentration (CO 2 ) in 

the atmosphere will double within this century the concentration 

at the beginning of the last century and that the increases 

in [CO 2 ] and other greenhouse gases will cause global warming 

of 1.5 to 5.8 °C at the end of this century (IPCC 2001). 

Such global environmental change is considered to have an 

enormous impact on Asian rice production, so that prediction 

of its impact by region is important not only for future food 

security but also for the sustainable development of rice-farming 

societies in Asia. Many experimental studies (Baker 2004, 

Horie et al 2000) and simulation studies (Matthews et al 1995, 

Horie et al 1995) have been made to clarify elevated [CO 2 ] 

and increased temperature effects on rice. On the basis of these 

previous studies and a multilocational field experiment (Horie 

et al 2003) covering wide-ranging rice genotypes and environments 

in Asia, we developed a new rice model for simulating 

growth and yield of different genotypes under variously 

different environments (Yoshida et al 2004, Yoshida and Horie 

2005, Horie 2005). 

This paper briefly reviews important findings from the 

previous studies on [CO 2 ] and temperature effects on rice, then 

explains the structure of our rice model and results of model 

simulation on rice yield at different locations in Asia under 

increased [CO 2 ] and temperature conditions, and discusses the 

implications of the simulation results for future Asian rice production. 

Rice responses to [CO 2 

] and temperature: a review 

To understand rice responses to [CO 2 ] and temperature, extensive 

research has been done using experimental facilities 

of closed chambers (Allen et al 1995), open-top chambers 

(Ziska et al 1997), and temperature gradient chambers (Horie 

et al 2000) since 1980. Since results of these studies have been 

fully reviewed in Horie et al (2000), only important findings 

from these studies for model synthesis are summarized here. 

Those previous studies showed that elevated [CO 2 ] accelerated 

rice development; nearly doubled [CO 2 ] increased leaf 

photosynthesis by 30–70%, canopy photosynthesis by 30–40%, 

and crop biomass yield by 15–30%, depending on genotype 

and environment; and that elevated [CO 2 ] had a minor effect 

on rice nitrogen (N) uptake, which appeared to be associated 

with the relatively insensitive response of leaf area growth to 

[CO 2 ]. Those rice responses to [CO 2 ] resulted in a substantial 

grain yield increase under elevated [CO 2 ] and nearly optimum 

temperature conditions. Analysis of reported data on rice yield 

response (R) to [CO 2 ] under optimum temperatures revealed 

that the relative response with yield at 353 µmol mol –1 with 

unity can be well approximated by the following equation: 

R = a (Ca – Co)/{K + (Ca – Co)} (1) 

where Ca is ambient [CO 2 ], Co is [CO 2 ] at which R becomes 

zero, and a and K are empirical constants. Indica rice genotypes 

showed a higher yield response to [CO 2 ] than japonica 

genotypes (Fig. 1). The values of a, K (µmol mol –1 ), and Co 

(µmol mol –1 ) for indica genotypes were 1.89, 288, and 30.6, 

respectively, and those for japonicas were 1.39, 124, and 30.6, 

respectively. Thus, relative yield increases at 700 µmol mol –1 

are estimated to be 32% for indicas and 17% for japonicas. 

With the increase in daily maximum temperature averaged 

over flowering period above about 36 ºC, rice yield generally 

declined because of spikelet sterility induced by high 

temperatures. Importantly, elevated [CO 2 ] increased spikelet 

susceptibility to high-temperature damage (Kim et al 1996). 

Nearly doubled [CO 2 ] decreased the threshold temperature for 

high-temperature damage of spikelets by 1–2 ºC more than 

the ambient [CO 2 ]. 

Simulation of rice growth and yield under elevated [CO 2 

] 

and increased temperatures 

The model used for the simulation 

Many attempts have been made to predict global climate change 

effects on rice by rice growth simulation models (Matthews et 

al 1995, Horie et al 1995). However, the models employed for 

prediction did not consider the important interacting effect of 

[CO 2 ] and temperature on high-temperature-induced spikelet 

sterility, nor genotypic differences in the responses to [CO 2 ] 

and temperature that are described above. Considering these 

insufficiencies of existing rice models, we synthesized a physiologically 

based rice growth and yield simulation model named 

GEMRICE (Genotype by Environment simulation Model for 

RICE). Since important components of this model have been 

reported in Yoshida et al (2004 and 2005) and Horie et al 

(1995), we describe only briefly the structure of GEMRICE. 

GEMRICE consists of six model components for simulating 

processes of phenological development, photosynthesis, 

respiration, biomass growth, spikelet number determination, 

and grain filling. The phenological development process 

component is the same as the previous model SIMRIW (Horie 

et al 1995). The photosynthesis component simulates leaf pho- 


R (relative yield) 

2.0 

1.5 

Ueda et al (2000) 

Kim et al (1996) 

Kim et al (2003) 

Huang et al (2002) 

Allen et al (1995) 

Baker (2004) 

Ziska et al (1997) 

Ueda et al (2000) 

Indica 

R 2 = 0.85 

1.0 

Japonica 

R 2 = 0.77 

0.5 

Ca 

0 

0 200 400 600 800 1,000 

Ca=CO 2 concentration (mmol mol –1 ) 

Fig. 1. [CO 2 ]-yield response curves of indica (open symbols) and japonica (closed symbols) 

rice genotypes grown under nearly optimum temperatures. The figure was drawn with experimental 

data from various reports shown in the figure. 

tosynthetic rate as a function of radiation flux density, temperature, 

and [CO 2 ], in which stomatal conductance and leaf 

N content are given as genotype-dependent plant parameters. 

Canopy photosynthetic rate is given by integrating leaf photosynthetic 

rate with respect to leaf area index (LAI), taking into 

account solar radiation flux density in the canopy as a function 

of LAI. Respiration consists of growth and maintenance 

respirations, both of whose rates are given as functions of plant 

N content. The maintenance respiration rate is also a function 

of temperature. The photosynthetic and respiration rates are 

both calculated on an hourly basis by inputting hourly temperature 

and solar radiation, which were estimated from their 

daily values. 

Crop biomass growth rate (CGR) is calculated on a daily 

basis by subtracting daily respiration from daily photosynthesis. 

The number of spikelets per unit area is calculated as the 

difference between the spikelet number differentiated and that 

degenerated. The differentiated spikelet number is determined 

by multiplying plant N content at 2 wk before full heading by 

a genotype-dependent coefficient, and the degenerated spikelet 

number by a function of CGR during the 2-wk period before 

full heading. Spikelet sterilities caused by both cool temperature 

at the meiosis stage and high temperature during flowering 

are simulated by the respective functions of temperature. 

The [CO 2 ] effect on high-temperature-induced spikelet 

sterility is also considered. Grain-filling rate is calculated as a 

function of sink and source size and temperature. Sink size is 

given by a logistic function of the current grain yield and the 

maximum sink size, which in turn is a function of effective 

spikelet number per unit area (= spikelet number produced – 

spikelet number sterilized). Source size is the sum of the current 

daily photosynthesis and amount of nonstructural carbohydrates 

stored in the plant. 

In the current version of GEMRICE, LAI and plant N 

content are given as external parameters by assuming that effects 

of [CO 2 ] and temperature on those parameters are relatively 

small. 

Estimation of parameter values and model validation 

The model GEMRICE contains 49 plant parameters to describe 

rice growth and yield formations, of which 37 are genotypeindependent 

and 12 genotype-dependent. To estimate values 

of these parameters and validate the model, a cross-locational 

field experiment named the Asian Rice Network Experiment 

(ARICENET, Horie et al 2003) was conducted on growth and 

yield of nine common genotypes. The experimental sites were 

eight locations: Iwate, Nagano, Kyoto, and Shimane in Japan; 

Nanjing and Yunnan in China; and Chiang Mai and Ubon 

Ratchathani in Thailand, covering 39N (Kitakami, Iwate) to 

15N (Ubon Ratchathani). The nine genotypes used included 

representative cultivars from indica, temperate and tropical 

japonicas, and a glaberrima-sativa hybrid (NERICA). The 

experiment was conducted in 2001 and 2002 at the respective 

sites. 

Approximately half of the ARICENET data were used 

for parameter estimation and the remainder for model validation. 

GEMRICE with parameter values thus estimated could 

simulate spikelet number per unit area, biomass, and grain yield 

of ARICENET data with R 2 values of 0.86, 0.80, and 0.87, 

respectively (Yoshida et al 2003). 

Session 19: Climate change and rice production 537

Yield change from current [CO 2 ] and temperature (%) 

40 

20 

0 

60 

40 

20 

–40 

60 

40 

20 

0 

0 

–20 

–20 

–40 

–60 

–80 

[CO 2 ] 700 ppm + 0 °C 

[CO 2 ] 700 ppm + 2 °C 

[CO 2 ] 700 ppm + 4 °C 

IR72 

Nipponbare 

Iwate Kyoto Nanjing Yunnan Ubon 

Fig. 2. Percentage yield change from the current under 700 µmol 

mol –1 [CO 2 ] with 0, 2, and 4 ºC temperature rises, simulated for 

cultivars IR72 and Nipponbare at Iwate and Kyoto in Japan, Nanjing 

and Yunnan in China, and Ubon Ratchathani in Thailand. Climate 

data in 2001 and 2002 at each location were used as base climate. 

Simulated effects of elevated [CO 2 

] 

and increased temperature on rice yield 

Growth and yield simulations were made for different rice 

genotypes grown under 700 µmol mol –1 [CO 2 ] and different 

degrees of temperature rise at the respective ARICENET sites. 

The [CO 2 ] of 360 µmol mol –1 and weather conditions in 2001 

and 2002 at the respective sites were used as base [CO 2 ] and 

climate conditions for the simulation. Figure 2 represents the 

simulation results on cultivars of IR72 (indica) and Nipponbare 

(japonica) at Iwate and Kyoto in Japan, Nanjing and Yunnan 

in China, and Ubon in Thailand as percentage yield change 

from the base conditions at each site. Simulation results on 

IR72 at Iwate and Nipponbare at Ubon are not shown in Figure 

2 because these cultivars showed very poor adaptability to 

the climates of those sites. 

The model predicted higher yield responses to elevated 

[CO 2 ] in IR72 (indica) than in Nipponbare (japonica) simulated 

for all the locations under the current temperature condi- 

tions, which agrees well with the observation shown in Figure 

1. Since the model assumed the same leaf photosynthetic response 

to [CO 2 ] for both genotypes, the higher yield response 

of IR72 to [CO 2 ] was due to its larger sink formation ability 

and, hence, more source limitation under ambient [CO 2 ] than 

Nipponbare. The model predicted that doubled [CO 2 ] alone 

will increase yield of the two genotypes by about 20–30% at 

most locations simulated. This effect of [CO 2 ] was drastically 

reduced by a 2 ºC temperature increase at all locations except 

Iwate in northern Japan. At Kyoto, where summer temperature 

is high, doubled [CO 2 ] with a 2 ºC temperature rise will 

significantly reduce the yield of Nipponbare, reflecting the 

increased spikelet susceptibility to high temperature damage 

by elevated [CO 2 ]. It was predicted that doubled [CO 2 ] with a 

4 ºC temperature rise would have severe negative effects on 

rice yield at most Asian locations except for northern areas 

such as Iwate. The negative effects were more pronounced in 

warm-temperature regions such as Kyoto and Nanjing than in 

tropical areas such as Ubon. However, the results at Ubon in 

Figure 2 were on wet-season rice, and severe negative effects 

on dry-season rice yield were predicted there under conditions 

of doubled [CO 2 ] with a temperature rise of more than 2 ºC. 

The review of the previous studies and the model simulation 

indicated that the anticipated global climate change associated 

with increased greenhouse gases will have large and 

different effects on rice production in Asia, depending on locations 

and genotypes. The model predicts that, while doubling 

[CO 2 ] with a temperature rise of more than 2 ºC will 

significantly increase rice yield in cool temperate areas, it will 

drastically reduce the yield in warm temperate areas and for 

dry-season rice in the tropics. This is an important subject for 

further studies to identify adaptive technologies for mitigating 

the negative effects of global warming on rice production in 

each region in Asia. 

References 

Allen LH Jr, Baker JT, Albrecht SL, Boote KJ, Pan D, Vu JCV. 1995. 

Carbon dioxide and temperature effect on rice. In: Peng S, 

Ingram KT, Neue H-U, Ziska LH, editors. Climate change 

and rice. Berlin (Germany): Springer-Verlag. p 258-277. 

Baker JT. 2004. Yield responses of southern US rice cultivars to 

CO 2 and temperature. Agric. Forest Meteor. 122:129-137. 

Horie T, Nakagawa H, Ohnishi M, Nakano J. 1995. Rice production 

in Japan under current and future climates. In: Matthews PB, 

Kropff MJ, Bachelet D, van Laar HH, editors. Modeling the 

impact of climate change on rice production in Asia. 

Wallingford (UK): CAB International. p 143-164. 

Horie T, Baker JT, Nakagawa H, Matsui T, Kim H-Y. 2000. Crop 

ecosystem responses to climatic change: rice. In: Reddy KR, 

Hodges HF, editors. Climate change and global crop productivity. 

Wallingford (UK): CAB International. p 81-106. 

Horie T, Yoshida H, Shiraiwa T, Nakagawa H, Kuroda E, Sasaki T, 

Hagiwara H, Kobata T, Ohsumi M, Kobayashi K. 2003. Analysis 

of genotype by environment interaction in yield formation 

processes of rice grown under a wide environmental range in 

Asia. 1. Asian Rice Network (ARICENET) research and preliminary 

results. Jpn. J. Crop Sci. 72(extra issue 2):88-89. 


Horie T. 2005. Determination of the yield potential and associated 

traits in rice. Gamma Field Symposia 43. Institute of Radiation 

Breeding, NIAS. (In press.) 

Huang J, Yang H, Dong G, Wang Y, Zhu J, Yang L, Shan Y. 2002. 

Effects of free-air CO 2 enrichment (FACE) on yield formation 

in rice (Oryza sativa). Chinese J. Appl. Ecol. 13:1210- 

1214. 

IPCC (Inter-Governmental Panel on Climate Change). 2001. Climate 

change 2001: the scientific basis. Contribution of working 

group I to the third assessment report of IPCC. London 

(UK): Cambridge University Press. 

Kim H-Y, Horie T, Nakagawa H, Wada K. 1996. Effects of elevated 

CO 2 concentration and high temperature on growth and yield 

of rice: the effect on yield and its components on Akihikari 

rice. Jpn. J. Crop Sci. 65:644-651. 

Kim H-Y, Lieffering M, Kobayashi K, Okada M, Matthew WM, 

Gumpertz M. 2003. Effects of free-air CO 2 enrichment and 

nitrogen supply on the yield of temperate paddy rice crops. 


Matthews RB, Horie T, Kropff MJ, Bachelet D, Centeno HG, Shin 

JC, Mohandas S, Singh S, Defeng Z, Lee MH. 1995. A regional 

evaluation of the effect of future climate change on 

rice production in Asia. In: Matthews RB, Kropff MJ, Bachelet 

D, van Laar HH, editors. Modeling the impact of climate 

change on rice production in Asia. Wallingford (UK): CAB 

International, p 95-139. 

Ueda T, Nakagawa H, Okada K, Horie T. 2000. Varietal differences 

in yield and spikelet fertility of rice in response to elevated 

CO 2 concentration and high temperature. Jpn. J. Crop Sci. 

69(extra issue 1):112-113. 

Yoshida H, Horie T, Ohnishi M, Katsura K. 2003. Analysis of genotype 

by environment interaction in yield formation processes 

of rice grown under a wide environmental range in Asia. 2. A 

model explaining genotypic and environmental variation in 

spikelet number per unit area. Jpn. J. Crop Sci. 72(extra issue 

2):90-91. 

Yoshida H, Horie T. 2005. Analysis of genotype by environment 

interaction in yield formation processes of rice grown under a 

wide environmental range in Asia. 5. Jpn. J. Crop Sci. 73(extra 

issue 2). (In preparation.) 

Ziska LH, Namuco O, Moya T, Quilang J. 1997. Growth and yield 

response of field-grown tropical rice to increasing carbon dioxide 

and air temperature. Agron. J. 89:45-53. 

Notes 

Authors’ address: Graduate School of Agriculture, Kyoto University, 

Sakyo-ku, Kyoto 606-8502, Japan, e-mail: 

horiet@adm.kais.kyoto-u.ac.jp. 

Monitoring rice growth and development using 

a crop model: the case of northern Japan 

Masaharu Yajima 

Rice yield has increased dramatically with the development of 

improved nursery techniques, new varieties resistant to lodging 

and diseases, and new yield-enhancing agrochemicals. 

Although rice farmers commonly attain high yield, yield fluctuations 

sometimes occur because of unfavorable weather conditions 

during the rice-cropping season in particular areas. A 

recent study shows that summer weather in northern Japan since 

1982 appears to exhibit a distinct five-year cycle, with a pressure 

difference between Wakkanai in Hokkaido and Sendai in 

Tohoku. The temperature also follows similar five-year cycles 

and the second year of each cycle is typical of the Yamase, a 

cold northeasterly wind, and is consistent with the cool summer 

in northern Japan (Kanno 2004). For instance, severe spikelet 

sterility because of cool weather that occurred during meiosis 

of the pollen mother cell to the flowering stage of the rice 

crop in some areas of Tohoku Region in 1980, 1988, 1993, 

and 2003 greatly reduced yield. 

This paper intends to describe the research approach and 

initial results of wide-area monitoring of crop growth and development 

in 1993 and 2003 using a rice crop model (Yajima 

1996) to construct the initial version of a climatic early warning 

system against cool summer damage of rice under changing 

climate. 

Monitoring crop growth and development 

using a rice crop model 

The approach involved using the development stage and spikelet 

sterility and yield models, and crop and geo-climatic data 

sets to develop a prediction system. 

Based on the development stage (DVS) model designed 

by Horie (1987), values of 0, 1, and 2 are assigned to emergence 

(seed germination), heading, and physiological maturity, 

respectively. The value of DVS at any point of crop development 

is calculated by integrating the developmental rate 

(DVR), expressed as a function of daylength and daily mean 

temperature, with time. 

Generally, dry matter accumulation at any point in rice 

growth is proportional to the accumulated solar radiation intercepted 

by the canopy, except at the final stage of rice growth, 

when DVS is close to 2. The model used to estimate daily dry 

matter increase was adopted from Monteith (1977). 

The spikelet sterility model on the relationship between 

spikelet sterility and cool-temperature sensitivity of the rice 

plant at the panicle development stage was proposed by Yajima 

et al (1989), with the following equation: 


G = Σ (To – Ti) W(DVS) 

for Ti < 18.9 and 1.20 > DVS > 0.72 (1) 

where W(DVS) is the panicle sensitivity factor for cool temperature 

as a function of DVS. Rice shows two sensitivity peaks 

to cool temperature. The first is observed at DVS = 0.88, which 

corresponds to meiosis of the pollen mother cell, and the second 

is at DVS = 1.07, which corresponds to the mid-flowering 

stage. 

Grain yield is proportional to dry matter accumulation 

at physiological maturity (W m ) and is expressed as follows: 

Y = HI × (1 – UF/100) × W m (2) 

where Y is the grain yield (brown rice) and HI is the harvest 

index, with a value of 0.4. 

Crop data such as variety planted, transplanting date, 

heading date, and actual paddy yield were provided by the 

Department of Statistics, Ministry of Agriculture, Forestry, and 

Fisheries (MAFF). The Automated Meteorological Data Acquisition 

System (AMeDAS) provided the real-time weather 

information all over Japan. Digital national land information 

in Japan is also provided by the Ministry of Land, Infrastructure, 

and Transport for the information on land elevation, land 

use, and other geographical data on each rectangular grid-point 

1 km × 1 km in dimension. The mesh climatic data system 

provided an opportunity to estimate the daily mean temperature 

and solar radiation at each grid-point so that real-time 

weather information would be available at every mesh (1 km 2 ) 

all over Japan (Seino 1993). 

The investigation focused on Tohoku Region involving 

the monitoring of Aomori, Iwate, Miyagi, Akita, Yamagata, 

and Fukushima prefectures in 1993 and 2003. Each prefecture 

was divided into three to five subdivisions based on geographic 

and climatic conditions. Using the two abovementioned models, 

crop development and percentage of sterility were estimated 

for Tohoku Region in 1993 and 2003. The real-time 

mesh data on daily air temperature and solar radiation at 1 km 

× 1 km grid-points were estimated by using Seino’s method 

with daily AMeDAS data. Crop data were provided by MAFF. 

Results from simulation 

Spikelet sterility and yield index 

The percentage of sterility caused by cool temperature was 

estimated for Tohoku Region in 1993 and 2003. Spikelet sterility 

estimates of 50–100% were obtained in the Pacific 

Ocean–side prefectures, particularly in areas with large negative 

temperature deviations (3–5 o C) from climatic normals. 

Spikelet sterility estimates for the prefectures on the Japan Sea– 

side, which ranged from 10% to 30%, may be due to the negative 

deviation of air temperature from climatic normals by 1– 

3 o C, particularly during meiosis of mother pollen cells until 

heading in these areas. 

When the estimated spikelet sterility in Tohoku Region 

was plotted against that observed by the Department of Statistics 

of MAFF, a highly significant linear relationship (R = 

–0.932***) was obtained in 2003. This suggests that the spikelet 

sterility model is suitable for prediction purposes. 

Rice growth and yield 

As described in the methodology, the combined use of DVS, 

crop growth, and spikelet sterility models leads to the estimation 

of yield. In this case, daily mean temperature and solar 

radiation from the mesh data and the crop growth data were 

used. Yields were estimated using the rice crop model for 

Tohoku Region in 1993 and 2003. Estimated yields based on 

the actual daily mean temperature and solar radiation were 

generally high (4–5 t ha –1 ) on the Japan Sea–side, particularly 

in some parts of Aomori, and most parts of Akita and Yamagata 

prefectures. Low yield estimates were obtained on the Pacific 

Ocean–side of Aomori and Iwate prefectures. On the inland 

and southern portions of Iwate, and about 40% of Fukushima, 

yield ranged from 2 to 4 t ha –1 . 

With the climatic normal as the basis of the yield estimate, 

it is apparent that the estimated yield in the Japan Sea– 

side prefectures surpassed 5.5 t ha –1 . These areas consistently 

yielded high even when the yield estimates were based on actual 

temperature and solar radiation. Estimated potential yields 

in Aomori (Pacific Ocean–side) could be as high as 4.5 t ha –1 

under the climatic normal, but yield estimates were below 1 t 

ha –1 under the adversely cool temperature in 1993. Similarly, 

the estimated potential yield in Iwate and Miyagi prefectures 

decreased by 1–3.5 t ha –1 under the adverse cool weather. 

As a check for the reliability of the yield estimates, actual 

yields under climatic normals and cool weather conditions 

were plotted against estimated yields under two climatic 

conditions. The good fitness of the actual and estimated yields 

in the unity slope proved the reliability of the rice crop model 

in the monitoring and prediction of yield (Fig. 1). 

Importance of shifting heading date 

by changing transplanting days 

It is known that cool damage, which is sometimes called the 

accident element, differs greatly according to heading date. 

Using the crop model, the effects of shifting transplanting date 

on sterility and yield were examined under weather conditions 

in 2003. It was proved that sterility decreased considerably 

only by delaying transplanting slightly in week 1 on the Pacific 

Ocean–side in Aomari, Iwate, and Miyagi prefectures. It 

seems reasonable to expect a reduction in cool summer damage 

by shifting the transplanting date or by changing rice varieties 

to minimize the risk from unfavorable weather under 

changing climate (Table 1). 

Conclusions 

A crop model was used to develop a monitoring and forecasting 

system for rice development, spikelet sterility, and yield in 

Tohoku Region, Japan, in 1993 and 2003. With the use of meshweather 

data, rice models were developed and varietal characteristics 

of rice planted in the monitored area, spatial distribu- 


Estimated yield (t ha –1 ) 

7 

6 

Normal year 

2003 

5 

4 

3 

(1:1) 

2 

r = 0.944 

1 

0 1 2 3 4 5 6 7 

0 1 2 3 4 5 6 

Observed yield (t ha –1 ) 

Fig. 1. Spatial distribution of estimated yield (left) and its comparison with observed yield in 

2003 (right), Tohoku Region. 

Table 1. Estimated sterility and yield resulting from shifting transplanting date in 2003. 

Prefecture 

Sterility (%) Yield (t ha –1 ) 

Agroecological 

zone 2003 1 wk 2 wk 2003 1 wk 2 wk 

later later later later 

Aomori Aomori 27.2 6.5 7.3 3.77 4.88 4.47 

Tsugaru 14.3 3.8 6.3 4.87 5.42 4.79 

Nanbu 47.4 6.6 9.5 2.26 4.05 3.56 

Shimokita 46.5 11.9 11.2 2.23 3.76 3.63 

Iwate Upper Kitakami 23.2 11.4 14.4 3.55 3.83 3.17 

Lower Kitakami 13.5 5.2 6.9 4.06 4.32 3.96 

Southeast 25.4 7.9 8.5 3.31 4.10 3.80 

Shimohei 57.1 23.4 14.2 1.66 2.92 2.87 

North 58.0 26.9 16.9 1.70 2.83 2.78 

Miyagi South 15.4 4.9 5.4 3.74 4.15 3.95 

Central 14.6 3.3 3.9 3.71 4.19 3.97 

North 12.5 3.7 4.9 3.90 4.22 3.97 

East 15.0 3.2 3.1 3.70 4.31 4.15 

tion of mean air temperature, and its deviation from climatic 

normals, as well as spikelet sterility caused by cool temperature, 

were determined. A significant linear correlation was 

obtained between the estimated spikelet sterility and yield. The 

results suggest the importance of using a crop model to monitor 

and forecast rice development stages and spikelet sterility 

at the regional level or in areas affected by cool temperature 

damage. With this method, extension staff could easily provide 

information on the possible occurrence of spikelet sterility 

in particular areas, which could enable rice growers to take 

necessary countermeasures to minimize the yield reduction 

from cool temperature. 

References 

Horie T. 1987. A model for evaluating climatic productivity and water 

balance of irrigated rice and its application to Southwest Asia. 

Southwest Asian Studies, Kyoto Univ. 25:62-71. 

Kanno H. 2004. Five-year cycle of north-south pressure difference 

as an index of summer weather in northern Japan from 1982 

onwards. J. Meteorol. Soc. Jpn. 82:711-724. 

Seino S. 1993. An estimation of distribution of meteorological elements 

using GIS and AMeDAS data. Jpn. J. Agric. Meteorol. 

48:379-383. 

Yajima M, Nitto A, Seino H. 1989. Estimation of percent sterility of 

rice spikelets with DVS model. Abstract for 1989 Annual 

Meeting of the Agricultural Meteorological Society, Japan. 

p 58-59. 

Yajima M. 1966. Monitoring regional rice development and coolsummer 

damage. JARQ 30:139-143. 


Yajima M. 1996. Monitoring and forecasting of rice growth and development 

using crop-weather model. Proceedings of the 2nd 

Asia Crop Science Conference, Crop Science Society of Japan. 

p 280-285. 

Yajima M. 2003. Early warning system against cool summer damage: 

case of Northern Japan. In: Coping against El Niño for 

stabilizing rainfed agriculture: lessons from Asia and the Pacific. 

CGPR Center Monograph 43. p 57-70. 

Notes 

Author’s address: Department of Biology and Environmental Sciences, 

National Agricultural Research Center for Tohoku Region, 

National Agricultural Research Organization, Morioka, 

Iwate 020-0198, Japan, e-mail: yajima@affrc.go.jp. 

Coping with climate variability and change in rice 

production systems in the Philippines 

Felino P. Lansigan 

Climate variability is a major source of risk in rice production 

since it greatly affects yield variation and often leads to yield 

losses (Pantastico and Cardenas 1980, Lansigan et al 2000). 

Climate change is also expected to affect crop growth and development, 

resulting in yield reduction (Peng et al 1995, 

Matthews et al 1996). The stability of rice production systems 

in tropical agroenvironments depends on adaptation strategies 

and mitigation measures applied to cope with these events. 

This paper analyzes the effects of climate variability and 

change on rice production in the Philippines. Climatic events 

that are important to rice growth and production are analyzed. 

The paper distinguishes the effects associated with risks from 

long-term weather aberrations and short-term weather fluctuations. 

Coping strategies to adapt to or mitigate the effects of 

climate variability and change are discussed. 

Climate variability and change 

The Philippines has a tropical climate that is conducive to rice 

production year-round. The Philippine climate is generally characterized 

by four climatic types in terms of the relative duration 

and intensity of the wet and dry periods in different parts 

of the country. Type 1 climate has a pronounced wet period 

from May to November, and a dry period from December to 

April. Type 2 climate is characterized by no clear dry season, 

and maximum rainfall is experienced from November to January. 

Type 3 climate is characterized by no distinct wet and dry 

seasons but is relatively dry from November to April. Type 4 

climate has rainfall more or less evenly distributed throughout 

the year. 

Extreme climate variability conditions such as during the 

occurrence of El Niño and La Niña events have greatly affected 

the spatial and temporal distribution of rainfall in tropical 

rice-growing areas such as the Philippines. Figure 1 shows 

the percentile of the 12-month total rainfall in the Philippines 

for the period April to March during El Niño years in the last 

four decades. Significant changes in rainfall patterns occurred 

during different periods within the year. These extreme events, 

with a return period of 7–10 years, have tended to occur more 

frequently in recent years. Historically, El Niño events are often 

accompanied by the occurrence of temperature extremes 

(i.e., high temperatures during unusually prolonged periods 

that induce changes in potential evapotranspiration and soil 

moisture availability), resulting in severe droughts that cause 

reduced crop yields. Different areas across the archipelago experience 

varying degrees of drought. 

The Philippines is often visited by an average of 20 typhoons 

a year, which mostly occur in the eastern and northern 

portions of the archipelago. Typhoons, which are often accompanied 

by intense rainfall and strong winds, usually occur from 

June to October. These cause significant yield losses during 

the wet season, the growing period for most crops. 

Current climate change studies (IPCC 2001) anticipate 

that future climate will experience an appreciable increase in 

temperature, changes in the spatio-temporal distribution of precipitation, 

and increased frequency and intensity of extreme 

weather events. The frequencies of occurrence of successive 

wet and dry periods such as El Niño and La Niña and tropical 

cyclones and typhoons with strong winds are expected to increase. 

In recent years, typhoons have begun to be experienced 

during periods when they were less expected (e.g., November 

or December). Unfortunately, these events coincide with the 

period when rice and other crops are about to be harvested, or 

when the second crop has been planted. 

Effects of climate variability and change 

on rice growth and productivity 

Climate variability affects rice growth and the various crop 

processes and activities in rice production. It determines planting 

date, duration of crop growth, crop yield, and also management 

practices in rice production. The occurrence of extreme 

climate variability such as El Niño or La Niña events 

characterized by a prolonged dry period or heavy rainfall spell 

coinciding with the critical stages of crop growth and development 

may lead to significantly reduced crop yields and extensive 

crop losses. Statistics on rice production and productivity 

in the Philippines (PhilRice-BAS 2000) show that the 


1951-52 1953-54 1957-58 1968-69 1972-73 1976-77 

1997-98 

1982-83 1986-87 1991-92 1992-93 1993-94 1994-95 

Percentile* 

≤10 Severe drought effects 41–60 Near normal to above normal >90 Severe flood damage 

11–20 Drought effects 61–80 Way above normal conditions *Percentile is a way to present 

variability with respect to time 

21–40 Moderate drought effects 81–90 Potential flood damage 

Fig. 1. Spatial and temporal variability of 12-month (April-March) rainfall during El Niño years in the Philippines. (Source: PAGASA 2000.) 

observed declines in production and yield coincide with the 

occurrence of El Niño events. Rice production has declined 

during these extremely dry periods since crop productivity or 

yield level has decreased, and the area planted to rice has been 

reduced to adapt to the anticipated drought period. The risk 

associated with climate variability of rice production in general 

depends mainly on the growth stage of the rice crop when 

the weather aberration occurs. 

Several studies using temperature gradient tunnels, crop 

simulation models, and field experiments have been conducted 

to evaluate the effects of climate change on rice crop growth 

and development, particularly temperature increase and increased 

or double CO 2 concentration (e.g., Horie 1993, Horie 

et al 1996, Matthews et al 1996). In general, simulation analyses 

and field experiments have shown a reduction in rice yields 

because of the sterility of spikelets, which are very sensitive to 

a temperature increase. The same studies have also shown that 

rice varieties respond nonlinearly with increased CO 2 concentration 

(Peng et al 1995). 

Coping strategies for climate variability and change 

Crop management to cope with climate variability 

Inasmuch as climate variability and change are inevitable, rice 

production systems should be able to adapt to weather fluctuations 

and climatic aberrations to minimize their negative 

effects. Coping with or managing climate variability and change 

in rice production systems requires a combination of adapta- 

tion and mitigation measures that involve the choice of rice 

crop variety, adaptation of cultural management practices, and 

understanding of climate science. Crop management measures 

include a range of possible strategies such as the following 

(Lansigan 2003): (1) adjusting the cropping calendar to synchronize 

crop planting and the growing period with soil moisture 

availability, (2) changing the rice variety to plant (i.e., 

planting a drought-tolerant or early-maturing variety), (3) varying 

planting density, and (4) optimizing water-use efficiency 

by improving irrigation facilities and introducing water-saving 

techniques. 

Crop improvement strategies 

Biological measures being pursued to cope with climate variability 

and change include the breeding of drought-tolerant 

varieties and lodging-resistant cultivars in national and international 

rice breeding programs. Drought-tolerant rice varieties 

can be planted that are more adapted to warmer or drier 

conditions. This involves the screening and testing of rice 

germplasm collections that are sources of tolerance of water 

stress or that require less water for crop growth. Lodging-resistant 

cultivars will be useful in areas generally experiencing 

more frequent and intense typhoons with strong winds and more 

precipitation. 

Crop insurance for and risk from climate variability 

Risks associated with reduced rice production and yield losses 

caused by climate variability can be partly alleviated by pro- 


Estimating 

crop area 

Downscaling 

climate forecasts 

Cropping strategy 

Simulating 

seasonal crop 

yields 

Fig. 2. Schematic of a knowledge-based crop forecasting system using seasonal climate prediction, estimation of crop area, and simulation 

of rice yields. 

viding appropriate crop insurance coverage to rice farmers. 

Current crop insurance practices in the Philippines cover only 

the cost of land preparation and rice crop establishment. The 

crop insurance policy at the moment often does not consider 

the varying extent of vulnerability of rice-growing areas in 

different locations with different climatic types as well as the 

differences in risk from climate variability, which are site-specific 

and time-dependent. Although crop insurance has been 

recognized as an effective coping strategy, it is less popular 

among rice farmers because of the usually unaffordable insurance 

premium. Promotion of crop insurance to rice farmers 

requires not only the provision of a more reasonable insurance 

premium but also attractive crop insurance coverage schemes 

or options such as insuring not only the cost of crop establishment 

but also (1) the expected crop yield or (2) the expected 

revenue based on harvestable crop. 

Seasonal climate prediction and rice production 

In recent years, advances and developments in climate science, 

meteorology, and global data observation and monitoring 

networks have enabled the reasonably accurate prediction 

or real-time forecast of seasonal climate up to a 3-month lead 

time. Reliable forecasts of seasonal climate serve as important 

and crucial information to an early warning system to support 

crop production to recommend appropriate management actions 

or measures to take considering the anticipated climatic 

conditions during the cropping period. More specifically, 

knowledge-based crop forecasting that uses these seasonal climate 

forecasts will be useful in guiding rice farmers and concerned 

government agencies and other stakeholders to act accordingly 

to minimize the negative or detrimental consequences 

of climate variability. Such a crop forecasting system currently 

being piloted in the Philippines (Fig. 2) involves using seasonal 

climate forecasts and a validated process-based rice crop 

simulation model to predict rice crop performance under the 

predicted seasonal climate outlook during the growing period. 

Seasonal climate forecasts are often made at a regional 

level, and have to be downscaled to a specific location (e.g., 

province of a country) to produce forecasts of rice yield performance 

in a given area. Crop forecasts provide useful information 

to stakeholders (e.g., farmers, extension workers, traders, 

and planners) in rice production systems at different levels. 

The crop forecasting system also includes efficient and 

effective tools and mechanisms to deliver the forecast information 

to the intended users in the area. 

Conclusions 

The vulnerability of rice production systems to climate variability 

in tropical regions depends on its time of occurrence 

relative to the growth stage of the crop. Advances in crop science, 

meteorology, and other sciences offer potential for knowl- 


edge-based approaches and strategies to manage, adapt to, and 

cope with climate variability and change. Effective and efficient 

adaptation and mitigation measures should be promoted 

to prepare stakeholders in rice production systems to enhance 

their resilience and flexibility when facing inevitable climatic 

events. 

References 

Horie T. 1993. Predicting the effects of climatic variation and effect 

of CO 2 on rice yield in Japan. J. Agric. Meteorol. 48:567- 

574. 

Horie T, Nakagawa H, Ohnishi M, Nakano J. 1996. Rice production 

in Japan under current and future climate. In: Matthews RB, 

Kropff M, Bachelet D, editors. Modeling the impacts of climate 

change on rice production in Asia. Wallingford (UK): 

CAB International. p 143-164. 

IPCC (Intergovernmental Panel on Climate Change). 2001. Climate 

change 2001: impacts, adaptation, and vulnerability. Contribution 

of Working Group II to the Third Assessment Report 

of the IPCC. Washington D.C., USA. 

Lansigan FP, delos Santos WL, Coladilla JLO. 2000. Agronomic 

impacts of climate variability on rice production in the Philippines. 

Agric. Ecosyst. Environ. 82:129-137. 

Lansigan FP. 2003. Assessing the impacts of climate variability on 

crop production, and developing coping strategies in rainfed 

agriculture. UN CGPRT Monograph No. 43, Bogor, Indonesia. 

p 21-36. 

Matthews RB, Kropff M, Bachelet D, editors. 1996. Modeling the 

impacts of climate change on rice production in Asia. 

Wallingford (UK): CAB International and IRRI. 

PAGASA (Philippine Atmospheric, Geophysical, and Astronomical 

Services Administration). 2000. El Niño years in the Philippines. 

Quezon City. 

Pantastico EB, Cardenas AC. 1980. Climatic constraints to rice production 

in the Philippines. In: Agrometeorology of the rice 

crop. Los Baños (Philippines): International Rice Research 


Peng S, Ingram KT, Neue H-U, Ziska LH, editors. 1995. Climate 

change and rice. Berlin (Germany) and Los Baños (Philippines): 

Springer-Verlag and International Rice Research Institute. 

374 p. 

PhilRice-BAS (Philippine Rice Research Institute and Bureau of 

Agricultural Statistics). 2000. Rice statistics handbook (1970- 

1997). Muñoz (Philippines): PhilRice-BAS. 

Notes 

Author’s address: Institute of Statistics (INSTAT), University of the 

Philippines Los Baños, College, Laguna, 4031 Philippines, 

e-mail: fpl@instat.uplb.edu.ph. 

Effect of elevated CO 2 on nutrient uptake 

and nutritional conditions of rice 

Jianguo Zhu 

The China rice/wheat FACE (free-air carbon-dioxide enrichment) 

has been running since 2001. It is the first FACE approach 

to determine the effects of elevated CO 2 on intensive 

agroecosystems with a rice/wheat rotation under field conditions. 

The FACE system followed the Japanese design with 

three FACE rings, 12.5 m in diameter, and three ambient rings. 

The experimental setup was a split-plot design with CO 2 treatments 

(two CO 2 concentrations) as the main plots and N fertilization 

as subplot treatments (two N levels). Different from 

others, the China FACE system is run continuously year-round, 

24 h a day. The target CO 2 concentration ([CO 2 ]) at the ring 

center was 200 µmol mol –1 above ambient. 

Results showed that elevated CO 2 affects nutrient uptake 

and nutritional conditions of rice. Nitrogen content (%) 

in the rice plants sampled at different growth stages decreased 

significantly under FACE treatment, but N accumulation increased 

slightly because of the significant enhancement in dry 

matter production. Ear N concentration increased at the heading 

stage but decreased at the ripening stage. No significant 

effect was found on root N concentration at tillering, but root 

N concentration at jointing, heading, and ripening decreased. 

The FACE treatment resulted in a significant increase in N- 

use efficiency for biomass production (NUEp), which was 

measured at the 28th day after transplanting, at heading, and 

at maturity. A significant increase in N-use efficiency for grain 

output (NUEg) and nitrogen harvest index (NHI) under the 

FACE treatment was also observed. Nitrogen content (%) and 

N accumulation in rice plants increased significantly under 

high-N treatment, but N-use efficiency in rice plants decreased. 

Elevated CO 2 significantly increased P and K uptake in 

aboveground tissues. P fertilization had no significant effect 

on various tissue dry biomass. Leaf P concentration at jointing, 

heading, and ripening increased, but no significant effect 

was found on P concentration in the stem, ear, and root. P and 

K concentration in rice was not influenced in the same way as 

N concentration by elevated CO 2 . The dilution effect is always 

used to explain the decrease in N content led by elevated 

CO 2 ; however, it could not be used to explain the increase in P 

content. The average soil exchangeable K in the FACE ring 

increased more under the NN treatment than under the LN 

treatment in the rice season. The possible increase in soil or- 


ganic acid or HCO 3 

– 

because of elevated CO 2 may increase 

the release of soil mineral K. 

C content in various tissues changed unremarkably and 

the ratio of C over N (C/N) increased. Elevated CO 2 increased 

the C/N ratio in the leaf, stem, ear, and root by 1.8%, 11.1%, 

10.1%, and 18.7%, respectively, under LN, and by 7.6%, 

54.7%, 12.0%, and 33.9% under NN at the rice ripening stage. 

CO 2 enrichment led to a striking increase, particularly 

in the jointing and heading stages, in the concentrations of 

Fe 2+ , Mn 2+ , Ca 2+ , and Mg 2+ in the soil solution, whereas K + 

and Zn 2+ contents decreased. In addition, elevated CO 2 increased 

the concentrations of Si, P, Cl – , and HCO 3 – , but decreased 

S and SO 4 

2– 

in the soil solution. No N-supply effects 

were found on soil solution chemistry. Dissolved organic C, 

soil redox processes, and other biological processes played a 

key role in the dynamics of soil solution. During the rice-growing 

season, elevated CO 2 decreased soil redox potential. However, 

no CO 2 enrichment effects were found on soil-solution 

pH and soil exchangeable cations. 

Elevated CO 2 tended to increase the bioavailability of 

Fe, Mn, Cu, and Zn extracted with DTPA solutions. After the 

2002 and 2003 rice harvests, the concentration of DTPA extractable 

Fe increased by 15% (P ≤0.008) and 13% (P ≤0.001) 

in the soil layer at 0–5 cm, respectively. Enriched CO 2 led to a 

reduction in Cu concentration in all kinds of rice organs, as Fe 

did in roots. However, CO 2 enrichment increased Fe concentration 

in stems, senescent leaves, and ears. Except for roots, 

Zn and Mn followed the general pattern of nutrient reductions 

in rice shoots, but most of them were not statistically significant. 

Based on the increase in biomass under elevated CO 2 , 

total content per hill of Fe, Mn, and Zn in rice roots and shoots 

was enhanced by CO 2 enrichment at four growing stages. No 

nitrogen treatment effect was found on plant-available micronutrients. 

Our results demonstrate that CO 2 -driven changes in soil 

microbial and redox processes in a paddy field could enhance 

soil CO 2 , CH 4 , and cation concentrations in the soil solution. 

Our findings also indicate that elevated CO 2 may stimulate 

soil chemical weathering because of higher P CO2 in the soil 

and more organic acids secreted by plant roots. This stimulation 

may increase soil nutrient mineralization and supply more 

available nutrients with growing plants. In the short term, we 

suggest that increased micronutrient demand in rice plants 

could be met by stimulating the availability of nutrients in the 

soil. However, it must be kept in mind that we may 

overgeneralize these results with only three years’ work. We 

cannot obviate the possibility that belowground responses to 

elevated atmospheric CO 2 are specific to particular soil types, 

cultivation systems, plant species, communities, and ecosystems. 

Therefore, more attention should be paid to the longterm 

effects of increasing atmospheric CO 2 on belowground 

processes in different intact ecosystems in the coming decade. 

The China Rice/Wheat FACE project is one research 

program within the China-Japan Science and Technology Cooperation 

Agreement. The main instruments and apparatus of 

the FACE system were supplied by the Japanese Rice FACE 

project. The project was supported by the Chinese Academy 

of Sciences, National Natural Science Foundation, and the 

Ministry of Science and Technology of China. The project also 

had financial support from the National Institute for Agro-Environmental 

Science of Japan. 

Notes 

Author’s address: State Key Laboratory of Soil and Sustainable Agriculture, 

Institute of Soil Science, Chinese Academy of Sciences, 

Nanjing 210008, China. 

Effect of soil properties, substrate addition, microbial 

inoculation, and sonication on methane production 

from three rice topsoils and subsoils in the Philippines 

Sudip Mitra and Deepanjan Majumdar 

Production and emission of CH 4 from rice fields, a gas about 

23 times more radiatively active than CO 2 , are an important 

feature of the global carbon cycle (Ramanathan et al 1985). 

Rice fields were identified as a major source of CH 4 and the 

intensity of CH 4 release to the atmosphere depends on several 

factors, including soil parameters (Denier van der Gon et al 

1992, Yao et al 1999, Mitra et al 1999). According to Yagi and 

Minami (1990), emission rates (flux) of CH 4 vary according 

to soil type and the organic matter applied; Bachelet and Neue 

(1993) illustrated the potential importance of differentiating 

soil types when estimating CH 4 emissions from rice fields. 

This study contains a series of experiments to characterize 

CH 4 production in three different rice topsoil and subsoil 

samples and their blends in different ratios created to produce 

gradients in soil properties for assessing CH 4 production potentials 

influenced by these properties. Blending was done with 

the assumption that, during long-term farming, topsoil and subsoil 

get mixed with each other, creating a similar situation. To 

estimate production potential, a laboratory incubation study 

was conducted estimating both entrapped and emitted CH 4 . 

Production potential was tested under special circumstances 

such as sterilization followed by inoculation, straw incorpora- 


tion and sonication to understand the influence of microbial 

addition, substrate addition, and the impact of the possible release 

of organo-minerals from the soil matrix on methane production. 


Soil collection and preparation 

Three rice soils (e.g., Luisiana, Maahas, and Pila) with a wide 

range of soil properties (Table 1) were collected from Central 

Luzon, Philippines, by a core sampler of 8 cm in diameter from 

0–20-cm (topsoil) and 30–50-cm (subsoil) layers. Soil from 

20–30-cm depth was not taken to avoid the possible mixing of 

the topsoil and subsoil. The entire volume of soil layer was 

mixed thoroughly and one subsample was taken for physicochemical 

analysis. After air-drying, soil samples were ground 

and passed through an 80-mesh sieve. Subsamples of air-dried 

topsoil and subsoil were blended in a 1:1, 3:1, and 1:3 ratio 

and mixed thoroughly, air-dried again, and stored in darkness 

at 25 °C temperature until the incubation experiment. 

Incubation study to estimate CH 4 

production 

The experiment was conducted by taking 20 g of soil and 40 

mL of de-ionized water in a 100-mL spoutless glass beaker 

with three replicates, which were then incubated. Gas samples 

were collected and analyzed periodically at 7-day intervals up 

to 91 days of incubation (DOI). The incubation procedure, gas 

sampling, and analysis have been described in detail elsewhere 

by Mitra et al (2002). 

Sterilization. In this set of soils, the beakers containing 

the soil suspensions and rubber corks were heat-sterilized separately 

at 121 °C for 40 min before incubation. After sterilization, 

the beakers were sealed with rubber corks immediately 

under a laminar flow and were then kept in the incubator at 30 

°C for 86 days. Nonsterilized natural soil was also incubated 

as a control. 

Sonication. In a separate set of soils, the beakers containing 

the soil suspension were treated in a sonicator-ultrasonic 

processor at 100 W output for 5 min. The beakers were 

kept in a water bath during sonication to reduce heating of the 

samples. 

Addition of rice straw. Rice straw was air-dried, pulverized, 

and passed through a 40-mesh sieve. Subsamples of 0.4 

g of rice straw powder were added to the soil suspension in the 

beakers. 

Statistical analysis 

Duncan’s multiple range test (DMRT) was done by MSTAT 

(version 1.41), developed by the Crop and Soil Science Division, 

Michigan State University, USA. All other analyses were 

done by MS Excel software developed by Microsoft Corporation, 

USA. 


Methane production and emission 

from topsoil, subsoil, and their blends 

The physicochemical properties of topsoils, subsoils, and 

blends in different topsoil/subsoil ratios are given in Table 1 

jointly with total CH 4 production. Some soil properties, such 

as pH and active Fe content, remained fairly stable in both 

layers and also in the different blends. In contrast, nitrogen 

and carbon content showed a steep decrease with decreasing 

topsoil share; only the nitrogen for Pila soil represented an 

exception from this trend for unknown reasons. In all other 

cases, the five other soil parameters correspond to a gradient 

N and C content of the soil, while other parameters are fluctuating 

without a consistent trend. 

Methane production began relatively early during the 

incubation; the peak appeared within 3 weeks after flooding 

of soil and then came down sharply. This pattern was typical 

for most rice soils in the absence of external substrates (Mitra 

et al 2002). As soon as the inherent C was exhausted by the 

methanogens, the CH 4 production rate dropped sharply. Maahas 

topsoil had a much lower CH 4 production than the other two 

topsoils (Table 1). Production rates were consistently lower 

than 9 µg g –1 soil d –1 . In Pila topsoil, CH 4 production was very 

intense within the first week, resulting in an early maximum, 

and it decreased gradually over the remaining incubation period. 

For all three soils, total CH 4 production showed a decreasing 

trend in the order topsoil > 3:1 blend > 1:1 blend 

>1:3 blend > subsoil. However, while the 1:1 blend yielded 

very low production rates for Luisiana and Maahas (corresponding 

to 1.7% and 0.3% of the production by the respective 

topsoil; see Table 1), this blend resulted in an intermediate 

value for Pila (26% of the production by topsoil). 

For all three topsoils, soil Eh dropped to low values (data 

not shown). However, Eh levels varied in a relatively broad 

range during the later stages. Soil pH stabilized near neutral 

after flooding in all the soils and did not have wide fluctuations. 

Effect of different soil treatments on CH 4 

production 

Sterilization. After sterilization, no CH 4 production was observed 

in the first phase of incubation and a rapid increase was 

seen in production rates after inoculating the topsoils. All three 

topsoils showed maximum production rates at the sampling 

date directly after inoculation, followed by a sudden drop (Fig. 

1). Methane production peaks were much higher in the sterilized 

topsoils than in the control topsoils. When the topsoils 

were re-inoculated, apparently a large amount of unused organic 

C became readily available to the newly introduced 

methanogens. In the subsoils, production levels were generally 

low even after inoculation. Only Pila subsoil showed some 

discernible CH 4 production, but production rates increased with 

a considerable delay compared to the control. 

Different CH 4 production patterns were previously attributed 

to different abundance of viable methanogenic bacte- 


Table 1. Physicochemical properties of topsoil, subsoil, and blends of different topsoil/subsoil ratios and their total methane production. 

pH Active Fe Active Mn Available P Total N Total C Clay Silt Sand CH 4 production a (mean ± SD)(mg kg –1 soil) 

Soils (1:1 H 2 O) (%) (%) (mg kg –1 ) (g kg –1 ) (g kg –1 ) (%) (%) (%) 

Control Sonication Straw app. Sterilization 

Topsoil 

Luisiana 4.3 4.33 0.09 4.1 1.84 18.13 29 66 5 433.9 ± 52.9 g 813.2 ± 101.1 c 2,534.7 ± 638.4 e 609.7 ± 28.5 c 

Maahas 6.4 2.05 0.07 15.0 1.55 16.40 55 36 9 169.3 ± 26.9 d 252.4 ± 51.4 b 2,297.5 ± 286.7 d 44.4 ± 10.1 b 

Pila 7.5 0.50 0.10 62.0 3.64 45.50 21 58 21 837.1 ± 43.7 i 1,070.9 ± 94.9 d 3,067.8 ± 604.5 f 836.2 ± 90.3 d 

3:1 (topsoil + subsoil) blend 

Luisiana 5.4 4.21 0.14 4.3 1.40 11.90 67 30 3 253.3 ± 33.2 f – – – 

Maahas 6.9 2.86 0.19 15 0.90 9.80 58 39 3 39.7 ± 9.1 b – – – 

Pila 8.1 0.98 0.12 97 3.70 31.60 33 54 13 551.3 ± 53.2 h – – – 


Luisiana 5.9 4.11 0.18 5.1 0.90 9.01 68 30 2 7.3 ± 0.9 a – – – 

Maahas 7.2 2.97 0.22 13 0.70 7.52 52 44 4 0.5 ± 0.04 a – – – 

Pila 7.9 0.94 0.09 67 3 26 35 54 11 216.6 ± 15.4 e – – – 


Luisiana 6.2 4.07 0.23 4.3 0.80 7.02 68 30 2 1.3 ± 0.4 a – – – 

Maahas 7.0 3.19 0.24 11 0.60 5.59 50 47 3 1.2 ± 0.2 a – – – 

Pila 8.0 0.69 0.05 44 0.24 21.10 41 49 10 112.7 ± 11.4 c – – – 

Subsoil 

Luisiana 5.3 4.31 0.29 3.1 0.58 0.57 51 43 6 0.1 ± 0.1 a 0.2 ± 0.07 a 1,996.7 ± 282.8 c 0.3 ± 0.1 a 

Maahas 6.8 3.13 0.29 3.5 0.28 2.60 43 54 3 0.3 ± 0.1 a 0.1 ± 0.06 a 1,363.3 ± 291.9 b 0.9 ± 0.7 a 

Pila 7.1 0.50 0.10 19 1.85 1.94 44 47 9 0.9 ± 0.2 a 0.4 ± 0.3 a 734.5 ± 58.2 a 4.0 ± 0.6 a 

a Values followed by the same letters are not significantly different from each other at the 5% level of significance according to Duncan’s multiple range test (DMRT). DMRT was performed separately for CH4 methane production in 

different treatments. 


CH 4 (mg g –1 soil d –1 ) 

25 

20 

15 

Control 

A 

0.06 

0.04 

Control 

E 

10 

5 

0 

0 20 40 0 80 

0.02 

0.00 

0 

20 40 60 80 

Straw incorporation 

120 

B 

100 

80 

60 

40 

20 

0 

0 20 40 6 0 80 

Straw incorporation 

100 

F 

80 

60 

40 

20 

0 

0 20 40 6 0 80 

6 

Sonication 

50 

C 

40 

30 

20 

10 

0 

0 20 40 0 80 

Sonication 

0.16 

G 

0.12 

0.08 

0.04 

0.00 

0 20 40 6 0 80 

80 

60 

40 

20 

PI 

LU 

MA 

Inoculated 

Sterilization 

D 

0.16 

0.12 

0.08 

0.04 

Inoculated 

Sterilization 

H 

0 

0 20 40 6 0 80 

DOI 

0.00 

0 20 40 6 0 80 

Fig. 1. Methane production in topsoils (A,B,C,D) and subsoils (E,F,G,H) as influenced by 

sterilization and subsequent inoculation, sonication, and straw incorporation. DOI = days 

of incubation; PI = Pila, LU = Luisiana, and MA = Maahas. Please note different scaling on 

Y axis. 


ia in the soil samples, which became active when water and 

substrates were provided (Roy et al 1997). However, the results 

of our study with sterilized and inoculated soil samples 

did not give any evidence of an effect by the abundance of 

methanogenic bacteria. 

Methane production is a very complicated biochemical 

process driven by the microbial degradation of organic matter 

under anoxic conditions by methanogenic bacteria and is the 

final step in a series of sequential reduction processes that are 

initiated when soil is flooded (Patrick and Reddy 1978, 

Ponnamperuma 1981). In topsoils, sterilization followed by 

inoculation resulted in an increase in CH 4 production immediately 

following inoculation. Sterilization has not only interrupted 

the decay of inherent soil organic carbon, it has also 

prompted CH 4 production by supplying readily degradable biomass 

C stemming from the destroyed microbial community. 

Apparently, the entire suite of bacteria involved in substrate 

supply and CH 4 production was going through very rapid 

growth after sterilization. This pulse triggered by inoculation 

had resulted in a much higher total CH 4 production than the 

control. Enhancement of CH 4 production on exogenous carbonaceous 

substrate addition has been previously observed (Lu 

et al 2000), but the effect of sterilization and subsequent inoculation 

on methane emissions is rarely studied. Ironically 

enough, this treatment did not enhance CH 4 production in 

Maahas and Pila subsoils versus their respective controls, which 

might be due to less dead microbial biomass after sterilization 

or poor microbial community establishment after inoculation. 

Straw incorporation. The stimulating effect of rice straw 

on CH 4 production was observed for topsoils and subsoils. All 

topsoils had shown high CH 4 peaks within the initial 3 weeks 

of incubation, resulting in similar increments triggered by straw 

application, that is, approximately 2,150 mg CH 4 kg –1 soil more 

production than the control (Fig. 1). Subsoils were also very 

responsive to straw amendment; CH 4 production from the 

amended subsoils exceeded that from the untreated topsoils. 

However, while the straw-borne increment in Luisiana subsoil 

was high, CH 4 production was apparently impeded in Maahas 

and Pila subsoils. But, in all the subsoils, the percent increase 

in CH 4 production compared with the control was much higher 

than what was found in topsoils. 

Effect of sonication of CH 4 production. Sonication is 

often used to separate organo-minerals from the soil matrix, 

namely, organic matter associated with clay (Watson 1971). 

Sonication of soil resulted in diverging effects on CH 4 production 

in topsoils (Fig. 1). Luisiana topsoil showed higher 

production rates after sonication and total production increased 

by approximately 30% vis-à-vis the control. Pila topsoil showed 

an earlier and higher peak, but overall production was almost 

at an identical level. In contrast, production rates generally 

decreased in Maahas topsoil, resulting in a 70% reduction in 

overall production. For the subsoils, CH 4 production increased 

upon sonication but only Pila showed an appreciable (78%) 

stimulation through sonication. 

In Luisiana topsoil, this process seems to have supplied 

substrate that was not degradable without sonication enhancing 

CH 4 production. In Pila topsoil, sonication provided substrate 

that would have been otherwise degraded at a later stage, 

so that there was no change in total production. In Maahas 

soil, however, sonication has reduced CH 4 production. Although 

we have no clue on the mechanisms of this effect, we 

assume an effect of the high clay content (55%) of Maahas 

topsoil. 

Conclusions 

The enhanced rate of CH 4 production after inoculation in sterilized 

soils demonstrated the importance of dead microbial 

biomass as substrate and the external inoculation of 

methanogens, which created a perfect methanogenic condition. 

The establishment of microbial communities was found to be 

a critical factor for regulating CH 4 production after inoculation 

in sterilized soils. Intrinsically, although subsoils had a 

very low CH 4 production potential, the addition of rice straw 

significantly increased their potential. Sonication of soils was 

done to see the effect of the physical disturbance on CH 4 production 

potential. Sonication apparently has certain effects on 

the soil matrix and thus on the association of organic matter 

and clay, which might have regulated CH 4 production in different 

soils. However, further studies are needed to find a 

clearer answer to this observation. 

References 

Bachelet D, Neue HU. 1993. Methane emissions from wetland rice 

areas of Asia. Chemosphere 26:219-237. 

Denier van der Gon H, Neue HU, Lantin RS, Wassmann R, Alberto 

MCR, Aduna JB, Tan MJ. 1992. Controlling factors of methane 

emission from rice fields. In: Batjes NH, Bridges EM, 

editors. World inventory of soil emission potentials. WISE 

Rep 2, ISRIC. Wageningen, Netherlands. p 81-92. 

Lu Y, Wassmann R, Neue HU, Huang C, Bueno CS. 2000. 

Methanogenic responses to exogenous substrates in anaerobic 

rice soils. Soil Biol. Biochem. 32(11-12):1683-1690. 

Mitra S, Jain MC, Kumar S, Bandyopadhyay SK, Kalra N. 1999. 

Effect of rice cultivars on methane emission. Agric. Ecosyst. 

Environ. 73:177-183. 

Mitra S, Wassmann R, Jain MC, Pathak H. 2002. Properties of rice 

soils affecting methane production potentials. 1. Temporal 

patterns and diagnostic procedures. Nutr. Cycl. Agroecosyst. 

64(1-2):169-182. 

Patrick WH Jr, Reddy CN. 1978. Chemical changes in rice soils. In: 

Soils and rice. Manila (Philippines): International Rice Research 


Ponnamperuma FN. 1981. Some aspects of the physical chemistry 

of paddy soil. In: Proceedings of the symposium on paddy 

soils, Beijing, Institute of Soil Science. Beijing (China): 

Academia Sinica, Science Press. p 59-94. 

Ramanathan V, Cicerone RJ, Singh HB, Kiehl JT. 1985. Trace gas 

trends and their potential role in climate change. J. Geophys. 

Res. 90:5547-5566. 

Roy R, Klüber HD, Conrad RF. 1997. Early initiation of methane 

production in anoxic soil despite the presence of oxidants. 

FEMS Microbiol. Ecol. 12:311-320. 


Watson JR. 1971. Ultrasonic vibration as a method of soil dispersion. 

Soils Fert. 37:127-134. 

Yagi K, Minami K. 1990. Effects of organic matter application on 

methane emission from some Japanese paddy fields. Soil Sci. 

Plant Nutr. 36:599-610. 

Yao H, Conrad R, Wassmann R, Neue HU. 1999. Effect of soil characteristics 

on sequential reduction and methane production in 

sixteen rice paddy soils from China, the Philippines and Italy. 

Biogeochemistry 47:269-295. 

Notes 

Modeling the effects of farming management 

alternatives on greenhouse gas emissions: 

a case study for rice agriculture in China 

Authors’ addresses: Sudip Mitra, Center for Global Environment 

Research, The Energy and Resources Institute (TERI), Darbari 

Seth Block, Habitat Place, Lodhi Road, New Delhi 110003, 

India, e-mail: sudip@teri.res.in, sudipmitra@yahoo.com, previously 

with Crop, Soil, and Water Sciences Division, International 

Rice Research Institute, Los Baños, Philippines; 

Deepanjan Majumdar, Department of Environmental Sciences, 

Institute of Science and Technology for Advanced Studies and 

Research (ISTAR), Vallabh Vidyanagar 388120 Gujarat, India. 

Acknowledgments: Support provided by the International Rice Research 

Institute (IRRI), Philippines, and The Energy and Resources 

Institute (TERI), India, is kindly acknowledged. 

Changsheng Li, Steve Frolking, Xiangming Xiao, Berrien Moore III, Steve Boles, Yu Zhang, Jianjun Qiu, Yao Huang, William Salas, and Ronald Sass 

Since the early 1980s, water management of rice paddies in 

China has changed substantially, with midseason drainage 

gradually replacing continuous flooding (Shen 1998). This has 

provided an opportunity to estimate how management alternatives 

affect greenhouse gas emissions at a large regional scale. 

Both methane (CH 4 ) and nitrous oxide (N 2 O) fluxes from 

agroecosystems are highly variable in space and time, affected 

by ecological drivers (e.g., climate, vegetation, and anthropogenic 

activity), soil environmental factors (e.g., temperature, 

moisture, pH, redox potential, and substrate concentration gradients), 

and biochemical or geochemical reactions (Li 2000, 

Li et al, 2004). Process-based models are used to quantify trace 

gas fluxes driven by the local climate, soil, vegetation, and 

management conditions at the site scale. Geographic information 

systems (GIS) databases provide spatially differentiated 

information on climate, soil, vegetation, and management to 

drive the model runs across the region. To quantify the effects 

of water management change on carbon (C) sequestration and 

CH 4 and N 2 O emissions from approximately 30 million hectares 

of rice paddies in China, we integrated a process-based 

model, DNDC, with a GIS database of paddy area, soil properties, 

paddy management (fertilizer use, water management, 

crop residue management, planting and harvest dates), and daily 

weather data. 

DNDC was originally developed for predicting C sequestration 

and trace gas emissions for nonflooded agricultural 

lands, simulating the fundamental processes controlling 

the interactions among ecological drivers, soil environmental 

factors, and relevant biochemical or geochemical reactions, 

which collectively determine the rates of trace gas production 

and consumption in agricultural ecosystems (Li et al 1992, 

1994). Details of management (e.g., crop rotation, tillage, fer- 

tilization, manure amendment, irrigation, weeding, and grazing) 

have been parameterized and linked to the various biogeochemical 

processes (e.g., crop growth, litter production, 

soil water infiltration, decomposition, nitrification, denitrification, 

etc.) embedded in DNDC (Fig. 1). To enable DNDC to 

simulate C and N biogeochemical cycling in paddy rice ecosystems, 

we modified the model by adding a series of anaerobic 

processes. The paddy-rice version of DNDC has been described 

and tested in recent manuscripts (Cai et al 2003, Li et 

al 2004), and is summarized briefly here. 

The Nernst equation, a basic thermodynamic formula 

defining soil redox potential (i.e., Eh) based on concentrations 

of the oxidants and reductants existing in the soil liquid phase 

(Stumm and Morgan 1981), and the Michaelis-Menten equation, 

a widely applied formula describing the kinetics of microbial 

growth with dual nutrients (Paul and Clark 1989), were 

linked to each other in DNDC to simulate variation of soil Eh 

and its impact on rates of the redox reactions, which produce 

or consume CH 4 or N 2 O during the frequent changes between 

saturated and unsaturated conditions driven by water management. 

CH 4 or N 2 O is produced or consumed under certain Eh 

conditions (–300 to –150 mV for CH 4 , and 200 to 500 mV for 

N 2 O production), so the two gases are produced during different 

stages of the varying soil redox potential. The major substrates 

controlling production of CH 4 or N 2 O are dissolved 

organic carbon (DOC), ammonium, and nitrate. DNDC allocates 

the substrates to drive nitrification, denitrification, and 

fermentation based on a simple kinetic scheme, the so-called 

“anaerobic balloon.” An anaerobic balloon is defined as the 

effective anaerobic volumetric fraction of a soil, which is quantified 

by the Eh value as calculated with the Nernst equation. 

When the substrates are allocated within the balloon, they will 


Ecological 

drivers 

Climate Soil Vegetation Human activity 

Annual 

average 

temp. 

Potential 

evapotrans. 

Water demand 

Water uptake 

N demand 

Daily growth 

C0 2 

Litter 

Very labile Labile Resistant 

LAI-regulated 

albedo 

Soil temp. 

profile 

Soil moist. 

profile 

Soil climate 

Evap. 

0 2 

diffusion 

Trans. 

Soil Eh 

profile 

Vertical 

water 

flow 

0 2 

use 

Water stress 

N uptake 

Grain 

Stems 

Root respiration 

Roots 

Plant growth 

Effect of temperature and moisture on decomposition 

NH 4 

+ 

DOC 

Decomposition 

Microbes 

Labile Resistant 

Humads 

Labile Resistant 

Passive humus 

Soil 

environmental 

factors 

Temperature Moisture pH Eh Substrates: NH 4 + , NO 3 – , DOC 

NO – 

2 

NO 

N 2 O 

N 2 

Denitrification 

Nitrate 

denitrifier 

Nitrate 

denitrifier 

N 2 O 

denitrifier 

DOC 

NO 3 

– 

DOC 

NO – 

3 

N 2 O 

Nitrification 

Nitrifiers 

NO 

NH 4 

+ 

NH 3 

NH 3 

Clay- 

NH 4 

+ 

Soil Eh 

Aerenchyma 

DOC 

Fermentation 

CH 4 production 

CH 4 oxidation 

CH 4 transport 

CH 4 

Fig. 1. Structure of the DNDC model. 

participate in the reductive reactions (e.g., denitrification, 

methanogenesis); when the substrates are allocated outside of 

the balloon, they will be involved in oxidative reactions (e.g., 

nitrification, methanotrophy). When a soil is flooded, oxygen 

diffusion into the soil is severely restricted, which reduces the 

soil Eh and causes swelling of the anaerobic balloon. The enlarged 

balloon will hold more DOC or nitrate to stimulate 

methanogenesis or denitrification. This mechanism has been 

embedded in DNDC to link the soil water regime to trace gas 

emissions for rice paddy ecosystems. The modified DNDC 

model has been tested against several methane flux data sets 

from wetland rice sites in the United States, Italy, China, Thailand, 

and Japan (Li et al 2002, Cai et al 2003). The test results 

indicate that DNDC is capable of estimating the seasonal patterns 

and magnitudes of CH 4 and N 2 O fluxes from the sites 

although discrepancies exist for some of the tested cases. 

A GIS database was constructed to hold the 30 million 

ha of rice paddies with different rotation systems at the county 

scale in China. The area occupied by each rotation in each 

county was quantified by combining a county-scale statistical 

database of crop-sown areas with a Landsat TM-derived landcover 

map for all of mainland China (Frolking et al 2002). 

The database also includes soil texture, pH, bulk density, and 

organic carbon content, and daily weather data for 1990 from 

610 weather stations across China. General data on fertilizer 

use, tillage, planting and harvest dates, crop residue management, 

and crop varieties were taken from various sources. Two 

water management scenarios, continuous flooding (CF) and 

midseason draining (MSD), were used in the model simulations. 

The modeled results indicated that the ranges in total 

CH 4 emissions with CF and MSD management were 6,400– 

12,000 and 1,700–7,800 Gg CH 4 -C y –1 , respectively. Taking 

the mean values, changing water management from continuous 

flooding to midseason draining caused a reduction in aggregate 

methane emissions from rice paddies of about 40%, or 

6 Tg CH 4 y -1 . This is 5–10% of total global rice paddy methane 

emissions (Prather et al 2001). The mitigating effect of 

midseason draining on CH 4 flux was highly uneven across the 

country. The highest flux reductions (>200 kg CH 4 -C ha –1 y –1 ) 


were in Hainan, Sichuan, Hubei, and Guangdong provinces, 

with warmer weather and multiple-cropping rice systems. The 

smallest flux reductions (8.0 kg N ha –1 ) in 

Jilin, Liaoning, Heilongjiang, and Xinjiang provinces, where 

the paddy soils contained relatively high organic matter. The 

change in water management had very different effects on net 

greenhouse gas mitigation when implemented across climatic 

zones, soil types, or cropping systems. Maximum CH 4 reductions 

and minimum N 2 O increases were obtained when the 

midseason draining was applied to rice paddies with warm 

weather, high soil clay content, and low soil organic matter 

content, for example, Sichuan, Hubei, Hunan, Guangdong, 

Guangxi, Anhui, and Jiangsu provinces, which have 60% of 

China’s rice paddies and produce 65% of China’s rice harvest 

(see details in Table 1). 

References 

Cai Z, Sawamoto S, Li C, Kang G, Boonjawat J, Mosier A, Wassmann 

R. 2003. Field validation of the DNDC model for greenhouse 

gas emissions in East Asian cropping systems. Global 

Biogeochem. Cycles 17(4):doi:10.1029/2003GB002046. 

Frolking S, Qiu J, Boles S, Xiao X, Liu J, Zhuang Y, Li C, Qin X. 

2002. Combining remote sensing and ground census data to 

develop new maps of the distribution of rice agriculture in 

China. Global Biogeochem. Cycles 16(4):doi:10.1029/ 

2001GB001425. 

Li C, Frolking S, Harriss RC. 1994. Modeling carbon biogeochemistry 

in agricultural soils. Global Biogeochem. Cycles 8:237- 

254. 

Li C, Frolking S, Frolking TA. 1992. A model of nitrous oxide evolution 

from soil driven by rainfall events. 1. Model structure 

and sensitivity. J. Geophys. Res. 97:9759-9776. 

Li C. 2000. Modeling trace gas emissions from agricultural ecosystems. 

Nutr. Cycl. Agroecosyst. 58:259-276. 

Li C, Mosier A, Wassmann R, Cai Z, Zheng X, Huang Y, Tsuruta J, 

Boonjawat J, Lantin R. 2004. Modeling greenhouse gas emissions 

from rice-based production systems: sensitivity and 

upscaling. Global Biogeochem. Cycles 18(1):doi:10.1029/ 

2003GB002045. 

Paul EA, Clark FE. 1989. Soil microbiology and biochemistry. Second 

Edition. San Diego, Calif. (USA): Academic Press. 273 

p. 

Prather M, Ehhalt D, Dentener F, Derwent R, Dlugokencky E, Holland 

E, Isaksen I, Katima J, Kirchhoff V, Matson P, Midgley 

P, Wang M. 2001. Atmospheric chemistry and greenhouse 

gases. In: Houghton JT, Ding Y, Griggs DJ, Noguer M, van 

der Linden PJ, Dai X, Maskell K, Johnson CA, editors. Climate 

change 2001: the scientific basis. Contribution of Working 

Group I to the Third Assessment Report of the Intergovernmental 

Panel on Climate Change. Cambridge (UK): Cambridge 

University Press. p 239-287. 

Shen S. 1998. Soil fertility in China. Beijing (China): Chinese Agricultural 

Press. (In Chinese.) 

Stumm W, Morgan JJ. 1981. Aquatic chemistry: an introduction 

emphasizing chemical equilibria in natural waters. 2nd Edition. 

New York, N.Y. (USA): John Wiley & Sons. 780 p. 

Notes 

Authors’ addresses: Changsheng Li, Steve Frolking, Xiangming 

Xiao, Berrien Moore III, Steve Boles, Yu Zhang, Institute for 

the Study of Earth, Oceans, and Space, University of New 

Hampshire, Durham, NH, 03824, USA; Jianjun Qiu, Institute 

of Agricultural Resources and Regional Planning, Chinese 

Academy of Agricultural Science, Beijing, China; Yao Huang, 

College of Resource and Environmental Sciences, Nanjing 

Agricultural University, Nanjing, China; William Salas, Applied 

Geosolutions, LLC, Durham NH 03824 USA; Ronald 

Sass, Department of Ecology and Evolutionary Biology, Rice 

University, Houston TX 77005, USA, e-mail: 

changsheng.li@unh.edu. 

Acknowledgments: This research has been supported by NASA’s 

Terrestrial Ecology Program (NAG5-12838 and NAG5-7631), 

the NASA EOS-IDS program (NAG5-10135), and the Asian- 

Pacific Global Change Network (Land Use/Management 

Change and Trace Gas Emissions in East Asia—APN 2001- 

16). Development of the anaerobic biogeochemistry in DNDC 

was supported by the USDA Forest Service’s Southern Global 

Change Program. 


Table 1. Provincial paddy area and simulated CO 2 , CH 4 , and N 2 O maximum and minimum fluxes for continuous flooding (CF) and midseason draining (MSD) water 

management. CO 2 flux equals negative change in soil organic carbon storage. Sign convention for all fluxes is positive flux equals net emission from soil. All areas 

and all flux values greater than 10 Gg were rounded to two significant figures. 

Production b CO 2 Gg C y –1 CH 4 Gg C y –1 N 2 O Gg N y –1 

Province a (Gg C y –1 ) 

CF min CF max MSD min MSD max CF min CF max MSD min MSD max CF min CF max MSD min MSD max 

Sichuan c 11,000 1,500 240 1,600 –380 1,800 1,300 1,300 200 49 34 72 47 

Hunan 12,000 370 –1,700 280 –2,000 1,100 550 580 110 38 32 54 41 

Jiangsu 9,000 1,800 1,400 1,900 1,300 1,200 770 1000 480 25 18 36 25 

Jiangxi 7,600 74 –1,000 –3 –1,200 1,100 560 710 72 33 22 54 31 

Anhui 6,300 600 –710 780 –920 1,100 460 870 77 29 14 38 19 

Hubei 8,300 510 –1,200 490 –1,300 1,100 410 650 46 24 13 33 16 

Guangdong 7,600 310 –700 57 –870 860 470 400 150 26 22 40 29 

Guangxi 6,200 1,800 15 1,700 –13 670 260 260 13 38 26 56 33 

Yunnan 2,600 180 –290 130 –320 160 81 18 3 12 12 17 14 

Guizhou 2,200 450 –560 530 –550 200 96 74 9 14 10 21 13 

Fujian 3,600 –1 –180 –39 –250 260 170 140 29 14 10 21 13 

Henan 1,600 340 88 300 26 340 140 230 27 6 3 9 4 

Liaoning 1,800 490 390 660 450 1,800 140 230 61 10 8 19 17 

Jilin 1,700 650 370 680 500 150 100 120 30 12 8 22 24 

Shaanxi 450 130 180 120 160 110 64 77 7 4 3 7 4 

Shanghai 800 160 140 160 120 150 95 120 56 2 2 3 2 

Inner Mongolia n.a. d 120 160 120 180 34 21 27 5 3 2 4 3 

Shandong 520 –47 20 –44 27 61 50 50 34 1 1 2 2 

Hebei 450 –22 22 –5 37 28 26 34 27 1 1 1 1 

Zhejiang 6,100 340 –390 340 –480 730 390 550 140 23 17 34 23 

Heilongjiang 3,200 1,800 1,400 1,800 1,600 270 180 230 78 36 26 44 44 

Tianjin 190 –14 11 7 27 17 16 32 26 0.7 0.6 1 1 

Ningxia 270 –47 –38 –29 –36 2 3 10 2 0.7 0.5 1 1 

Gansu n.a. d 7 11 7 12 2 2 2 0.6 0.1 0.1 0.2 0.2 

Hainan 790 610 210 610 200 200 100 73 19 8 4 12 5 

Beijing n.a. d –1 7 3 7 5 4 6 2 0.1 0.1 0.2 0.1 

Shanxi n.a. d 4 7 6 8 3 3 3 2 0.1 0.1 0.2 0.2 

Xinjiang n.a. d 51 56 57 63 51 27 42 7 2 1 2 

China 94,000 12,000 –2,000 12,000 –3,500 12,000 6,400 7,800 1,700 410 290 610 420 

a Provinces sorted by paddy area. Tibet and Qinghai provinces excluded because of small paddy areas. 

b Mean annual rough rice production for 1991-2000 (IRRI 2004); carbon fraction of dry weight 

assumed to be 50%. c Includes Chongqing Province, which was established from land in Sichuan Province in 1997. d Production data (see note 2 above) not reported individually for some provinces; total 

production for all nonreported provinces = 370 Gg C y –1 ; China total equals sum of all reported and nonreported provinces. 


Impact of rising atmospheric CO 2 on CH 4 emissions 

from rice paddies 

Weiguo Cheng, Kazuyuki Yagi, Kazuyuki Inubushi, Kazuhiko Kobayashi, Hidemitsu Sakai, Han-Yong Kim, and Masumi Okada 

Recent anthropogenic emissions of key atmospheric trace gases 

(e.g., CO 2 , CH 4 , N 2 O, and CFC) that absorb infrared radiation 

may lead to an increase in global mean surface temperature, 

and thus have the potential to bring about changes in climate 

(IPCC 2001). Depending on population growth and energyuse 

scenarios, atmospheric CO 2 concentration is expected to 

rise from 370 ppm currently to about 485 to 1,000 ppm by 

2100 (IPCC 2001). CO 2 and CH 4 are two important greenhouse 

gases and they are responsible for approximately 55% 

and 20%, respectively, of the anthropogenic global warming 

effect. Most atmospheric CH 4 is produced by microbial activities 

in extremely anaerobic ecosystems, such as natural and 

cultivated wetlands, sediments, sewage, landfills, and the rumens 

of herbivorous animals. The methane from rice paddies 

accounts for about 17% of the anthropogenic sources. Also, 

plant-mediated transport is a very important pathway for CH 4 

emissions from cultivated rice soils. Many studies have demonstrated 

a positive effect of elevated CO 2 on rice biomass 

(above- and belowground) production and grain yield. The 

direct effect of elevated CO 2 on rice root biomass and tiller 

number would be expected to indirectly affect CH 4 production 

and emissions from rice fields, although such an effect 

has not been identified previously. Therefore, we carried out a 

controlled-environment chamber experiment at Tsukuba and a 

free-air CO 2 enrichment (FACE) experiment at Shizukuishi, 

Iwate, Japan, for a series of rice growth seasons, in order to 

understand the influence of elevated CO 2 on CH 4 emissions 

from rice paddies. 


Controlled-environment chamber experiment 

This research was conducted at the National Institute for Agro- 

Environmental Sciences (NIAES), Tsukuba, Japan. Six controlled-environment 

chambers (Climatron, Shimadzu, Kyoto) 

were used. The size of the chambers was 4 × 3 × 2 m (L × W × 

H). CO 2 concentration was maintained at 350 µmol mol –1 in 

three chambers (ambient CO 2 ) or at 650 µmol mol –1 in three 

other chambers (elevated CO 2 ). During the day, the CO 2 concentration 

was maintained by a computer-controlled system 

by injecting pure CO 2 to compensate for CO 2 uptake by the 

rice plants (Oryza sativa, cultivar Nipponbare). During the 

night, CO 2 levels increased because of plant respiration, but 

were not allowed to increase to >100 µmol mol –1 higher than 

the daytime levels by means of a computer-controlled air ventilation 

system, which introduced ambient air when necessary 

to reduce the CO 2 concentration. In this experiment, air temperature 

inside all controlled-environment chambers was con- 

trolled to follow ambient air temperature. In the 1998 and 1999 

rice growth seasons, rice straw was applied to the soil at 300 

and 150 g m –2 , respectively. The CH 4 emitted from the chambers 

was determined by an automated monitoring system during 

all rice growth seasons. 

To understand how elevated CO 2 affects CH 4 emissions 

by plant-mediated transport or ebullition diffusion through 

flooding water, and how drainage affects CH 4 emissions in 

elevated CO 2 conditions, a pot experiment in the Climatron 

was carried out in the 2002 season. No rice straw was applied. 

The flooding water was maintained at a 5-cm depth until 2 

September (3 weeks before harvest); then, all pots were drained, 

and the water level was maintained at 5 cm below the soil 

surface. A cylinder and a doughnut chamber were used to measure 

CH 4 flux with and without plants. 

FACE experiment 

The FACE experiments with rice plants were conducted at 

Shizukuishi, Iwate, Japan. The soil was an Andosol paddy soil. 

A randomized complete block design with two levels of target 

atmospheric CO 2 concentration (ambient CO 2 and FACE = 

ambient + 200 µmol mol –1 ) and four replicate blocks was used. 

The rice (Oryza sativa) test cultivar was Akita Komachi. CH 4 

flux was determined with a cylindrical, transparent, acrylic, 

closed-top chamber. The samples were collected and measured 

once every 2 weeks. The total seasonal emissions from each 

plot for both the 1999 and 2000 seasons were calculated and 

analyzed statistically. 


In rice soils, acetate and H 2 -CO 2 are the major substrates for 

CH 4 production. The carbon sources for CH 4 production in 

rice soils are from soil carbon, applying organic material, rice 

root exudates, and root autolysis products. The seasonal CH 4 

emission rates increased by elevated CO 2 (+300 µmol mol –1 ) 

in the controlled-environment chambers were 23% in 1998 and 

145% in the 1999 season with different rice straw application 

rates at 300 and 150 g m –2 , respectively, indicating that the 

effect of elevated CO 2 on CH 4 emissions was influenced by 

rice straw application. A large amount of rice straw applied 

before rice transplanting led to large CH 4 emissions during 

the early rice growth period, whereas there was no difference 

between ambient and elevated CO 2 treatments. Elevated CO 2 

increased CH 4 emissions in rice paddies during middle-later 

rice growth periods because root exudates and root autolysis 

products under elevated CO 2 conditions increased the carbon 

source for CH 4 production in anaerobic soil. Therefore, the 


Tiller number 

and stem size 

+ 

CH 4 

transport 

(by micropores) 

+ 

+ 

+ 

O 2 

diffused to 

CO rhizosphere 

2 

+ CH 

(by aerenchyma) 

4 

oxidation 

– 

CH 4 

emissions 

+ 

+ 

+ 

Root biomass 

+ 

Root 

exudation 

+ 

CH 4 

production 

+ positive effect, – 

negative effect. 

The size of the arrows in the last steps indicated potential changes during all the rice growth period. 

Fig. 1. A conceptual model depicting how elevated CO 2 affected CH 4 emissions from rice 

paddies. 

effect of elevated CO 2 on increasing total CH 4 emissions in 

the 1999 rice growth season was much greater than that in 

1998 because of lesser amounts of rice straw applied in 1999 

(Yagi et al 2000). 

The results from a pot experiment under controlled-environment 

chambers during the 2002 rice growth season 

showed that the increase in CH 4 emissions caused by elevated 

CO 2 was significant after the grain-filling stage of rice; the 

total emissions were enhanced 58% by elevated CO 2 . The CH 4 

emitted by ebullition-diffusion accounted for 11.3% and 11.9% 

of total emissions under ambient and elevated CO 2 conditions, 

respectively. In contrast, no CH 4 was emitted from plant-free 

pots, suggesting that the CH 4 emitted from rice-plant pots was 

most contributed by the rice plant through root exudates and 

root autolysis products. The CH 4 flux decreased when the 

flooding water was drained under both ambient- and elevated- 

CO 2 treatments, implying that drainage management will have 

an important role in mitigating future CH 4 emissions from 

paddy fields, when more CH 4 will likely be emitted from 

flooded rice paddy soils under increasing atmospheric CO 2 

concentration (Cheng et al 2003). 

When free-air CO 2 was enriched to 550 µmol mol –1 , the 

CH 4 emissions from the rice paddy field increased significantly, 

by 38% in 1999 and 51% in the 2000 season. Net CH 4 emissions 

are determined by the balance between methane production 

and methane oxidation in the rice plant and paddy soil 

ecosystems. The CH 4 production potential in the FACE was 

significantly larger than that in the ambient-CO 2 treatments at 

the flowering stage in the middle rice growth period because 

elevated CO 2 increased CH 4 production by root exudates and 

root autolysis products. However, elevated CO 2 had no net 

effect on CH 4 oxidation activity during the same period. Also, 

the elevated CO 2 increased rice tiller numbers and CH 4 transport 

through the rice plant in FACE treatments (Inubushi et al 

2003). 

In conclusion, elevated CO 2 affects CH 4 emissions from 

rice paddies both by increasing the net balance between CH 4 

production and oxidation and by increasing the pass-way 

through the rice plant (Fig. 1). Several studies were made on 

the effect of elevated atmospheric CO 2 concentration on CH 4 

emissions from wetland and paddy soils (Table 1). CH 4 from 

rice paddies and natural wetlands accounts for 32.7% of the 

total global sources to the atmosphere. If we consider that CH 4 

emissions from natural wetlands and rice paddies would be 

enhanced by elevated CO 2 , the global warming potential of 

CH 4 could become more important as atmospheric CO 2 increases, 

as was recently predicted (Lelieveld et al 1998). 

References 

Cheng W, Yagi K, Sakai H, Kobayashi K. 2003. Effects of elevated 

CO 2 on CH 4 and N 2 O emissions from submerged rice soil: a 

pot experiment. 2003 ASA-CSSA-SSSA Annual Meetings, 2- 

6 November 2003, Denver, Colorado, USA. CD-ROM. 

Dacey JWH, Drake BG, Klug MJ. 1994. Stimulation of methane 

emission by carbon dioxide enrichment of marsh vegetation. 

Nature 370:47-49. 

Hutchin PR, Press MC, Lee JA, Ashenden TW. 1995. Elevated concentrations 

of CO 2 may double methane emissions from mires. 

Global Change Biol. 1:125-128. 

Inubushi K, Cheng W, Aonuma S, Hoque M, Kobayashi K, Miura S, 

Kim H, Okada M. 2003 Effects of free-air CO 2 enrichment 

(FACE) on CH 4 emissions from a rice paddy field. Global 

Change Biol. 9:1456-1464. 

IPCC. 2001. Climate change 2001: the scientific basis. Cambridge 

(UK): Cambridge University Press. 

Lelieveld J, Crutzen PJ, Dentener FJ. 1998. Changing concentration, 

lifetime and climate forcing of atmospheric methane. 

Tellus 50B:128-150. 


Table 1. A review of the effects of elevated CO 2 on CH 4 emissions from wetland paddy soils. 

Study description CH 4 emissions Reference 

increasing (%) 

Wetlands 

Tidal marsh, open-top chambers 80 Dacey et al (1994) 

(690 ppm) 

Ombrotrophic mire, open-top chambers 145 Hutchin et al (1995) 

(550 ppm) 

O. aquaticum grew in wetland, glasshouse 136 Megonigal and Schlesinger (1997) 

(700 ppm) 

Paddy soils 

Rice (IR72), IRRI open-top chambers 49–60 Ziska et al (1998) 

(650 ppm) 

Rice (Nipponbare), Climatron (650 ppm) 23–145 This research (Yagi et al 2000) 

Rice (Akitakomachi), FACE (550 ppm) 38–51 This research (Inubushi et al 2003) 

Rice (Nipponbare), pot in Climatron 58 This research (Cheng et al 2003) 

(650 ppm) 

Megonigal JP, Schlesinger WH. 1997. Enhanced CH 4 emissions from 

a wetland soil exposed elevated CO 2 . Biogeochemistry 37:77- 

88. 

Yagi K, Li Z, Sakai H, Kobayashi K. 2000. Effect of elevated CO 2 

on methane emission from a Japanese rice paddy. FACE 2000 

Conference, 27-30 June 2000, Tsukuba, Japan. p 40. 

Ziska LH, Moya TB, Wassmann R, et al. 1998. Long-term growth at 

elevated carbon dioxide stimulates methane emission in tropical 

paddy rice. Global Change Biol. 4:657-665. 

Notes 

Authors’ addresses: Weiguo Cheng, Kazuyuki Yagi, and Hidemitsu 

Sakai, National Institute for Agro-Environmental Sciences, 

Tsukuba 305-8604, Japan; Kazuyuki Inubushi, Faculty of 

Horticulture, Chiba University, Matsudo, Chiba 271-8510, 

Japan; Kazuhiko Kobayashi, Graduate School of Agricultural 

and Life Sciences, the University of Tokyo, Tokyo 113-8657, 

Japan; Han-Yong Kim, Department of Agronomy, Chonnam 

National University, Gwangju 500-757, Korea; Masumi 

Okada, National Agricultural Research Center for Tohoku 

Region, Morioka 020-0198, Japan, e-mail: 

cheng@niaes.affrc.go.jp. 

The detrimental effect of tropospheric O 3 

on lowland rice is ameliorated by elevated CO 2 

T. Ishioh, K. Kobori, and K. Imai 

The atmospheric carbon dioxide (CO 2 ) concentration is increasing 

year by year because of human activities such as fossil 

fuel combustion and deforestation. The elevated levels of 

CO 2 promote photosynthesis of C 3 plants. For example, in lowland 

rice, photosynthesis and biomass production are promoted 

and plant development is accelerated by elevated CO 2 . More 

than a 20% yield increment of rice is often reported under 

doubled CO 2 conditions (Imai et al 1985, Kim et al 1996). On 

the other hand, atmospheric ozone (O 3 ), a major component 

of photochemical air pollutants, is often generated around urban 

areas during the sunny summer season in Japan. Ozone 

causes injury, with visible symptoms, growth inhibition, and 

yield decline in a range of field crops. From now on, under 

further industrialization and global warming, the frequency of 

occurrence of O 3 above the environmental quality standard 

(0.06 ppm in Japan) will increase and detrimentally affect crop 

plants. Thus, O 3 and CO 2 have opposite effects on production 

processes of crop plants. However, investigations on the interaction 

of these two gases in lowland rice are limited (Ishioh 

and Imai 2003, 2004, Kobori and Imai 2003, Olszyk and Wise 

1997). 

In our study, we exposed a japonica lowland rice cultivar 

to O 3 and CO 2 at the vegetative growth stage and at the 

heading to early-maturity stage and examined plant responses 

over short (3 h) and long (2 wk) terms to clarify the interaction 

of these two gases. 


A lowland rice cultivar (Oryza sativa L. cv. Koshihikari) was 

grown outdoors in small plastic pots containing 2.5 kg of soil 

and 4 g of compound chemical fertilizer (NPK=15:15:15, %) 


Experiment 1 (vegetative growth stage) 

In short-term 

1-0 

(3 h) (4 h) 

Experiment 1 (flowering to early-maturing stage) 

Flowering 

In long-term 

1-1 

1-2 

1-3 

1-4 

(1 wk) (1 wk) (1 wk) 

Outdoors 

Flowering to early maturing 

Early maturing 

(1 wk) (1 wk) 

Outdoors 

Fig. 1. Duration of treatments. 

: Gas exposure : O 3 -free air (CO 2 : 400 ppm) 

(experiment 1), or in large plastic pots containing 10 kg of soil 

and 8 g of chemical fertilizer (experiment 2). 

In experiment 1, at the vegetative growth stage (shortterm 

exposure, a half of the 7th leaf on main stem appeared; 

long-term exposure, a half of the 9th leaf on main stem appeared), 

nine plants (one plant per pot) were transferred into 

four naturally-lit phytotrons and were exposed to combinations 

of two levels of O 3 (0 ppm vs 0.1 ppm, 0630 to 1830) 

and two levels of CO 2 (400 ppm vs 800 ppm, 0000 to 2400) 

atmosphere (control plot, 0 ppm O 3 + 400 ppm CO 2 ; 0.1 ppm 

O 3 + 400 ppm CO 2 plot; 0.1 ppm O 3 + 800 ppm CO 2 plot; and 

0 ppm O 3 + 800 ppm CO 2 plot) for up to 3 h (short-term exposure) 

or 2 wk (long-term exposure) under 28/23 °C (day/night) 

and 60% relative humidity conditions (Fig. 1). The concentrations 

of O 3 and CO 2 , respectively, were assumed to be clean 

air or polluted air with O 3 and air of current or future doubled 

CO 2 . Gas exchange rates of fully expanded 6th leaves (shortterm 

exposure) or 8th leaves (long-term exposure) were examined 

for each of six plants with a portable measurement 

system (LI-6400, LI-COR, Inc.) at 0, 1, 2, 3, 4, 5, and 6 h 

(short-term exposure) and 0, 4, 11, 14, 18, and 21 d (longterm 

exposure) of treatment. Environmental conditions were 

CO 2 corresponding to treatment, 1,500 µmol m –2 s –1 PPFD, 

28 °C leaf temperature, and 1.5 kPa saturation deficit. After 

the treatments, plants were measured for their leaf areas and 

separated into leaf blade, leaf sheath + stem, and root fractions 

and oven-dried at 80 °C for 48 h and weighed. Based on 

the leaf area and dry weight values, growth parameters such as 

relative growth rate (RGR), net assimilation rate (NAR), leaf 

area ratio (LAR), leaf weight ratio (LWR), and specific leaf 

area (SLA) were calculated. 

In experiment 2, from flowering to early-maturity stage, 

plants were exposed to combinations of two levels of O 3 and 

two levels of CO 2 concentrations as described in experiment 1 

(Fig. 1). Gas exchange rates were measured at 0, 7, and 14 d 

of treatment. After the treatments, plants were transferred outdoors 

and cultivated ordinarily until full maturity. At maturity, 

plants were harvested and the dry weight of each organ, yield, 

and yield components were determined. 

Results 

At the vegetative growth stage (experiment 1), visible injury 

on the leaf surface appeared rapidly after the O 3 treatment 

(within 1–2 d) and leaf function such as gas exchange declined 

severely. In both the short- and long-term exposures, rapid 

declines in net photosynthetic rate and stomatal conductance 

occurred just after the O 3 treatment and these were not recovered 

when the plants were transferred into O 3 -free air (Fig. 2; 

experiment 1). Therefore, plant growth rate was suppressed 

and both leaf area and dry-matter production decreased in the 

0.1 ppm O 3 + 400 ppm CO 2 plot. Accordingly, the RGR decreased 

after the O 3 treatment because of decreased NAR. The 

LWR in the 0.1 ppm O 3 + 400 ppm CO 2 plot was higher than 

in the other plots in short- and long-term exposures and this 

indicated a decline in dry-matter partitioning to the root fraction. 

On the other hand, in a plot with 0.1 ppm O 3 + 800 ppm 

CO 2 , stomatal conductance decreased under doubling CO 2 and 

this could avoid injury to leaf function. Both net photosynthetic 

rate and growth rate were maintained as high as those of 

the control plot (0 ppm O 3 + 400 ppm CO 2 ). 

From flowering to early-maturity stage (experiment 2), 

the net photosynthetic rate was affected differently by the duration 

of treatments. A single exposure to O 3 decreased the net 

photosynthetic rate but the overall effect did not reflect on the 

whole-plant dry matter at maturity. The relevant factors might 

be the aging of the plant and/or meteorological conditions after 

the treatment, but a clear explanation was not obtained. 

However, once suffering O 3 fumigation, plants showed a tendency 

to have sterile and less fertile caryopses at the lower 

positions of the rachis branch of the panicle. This suggested 

that O 3 also suppressed the translocation of photoassimilate. 

Conclusions 

A high level of O 3 (0.1 ppm) inhibited photosynthesis, drymatter 

production, and yield of lowland rice. However, under 

elevated CO 2 (800 ppm), such effects of O 3 (0.1 ppm) were 

ameliorated. Especially in the exposure experiments at the 


Net photosynthetic rate (mmol m –2 s –1 ) 

40 

Exposure to O 3 

and/or CO 2 

O 3 

-free air 

35 

30 

25 

20 

15 

10 

5 

0 

Stomatal conductance (mmol m –2 s –1 ) 

0.8 

0.7 

0.6 

0.5 

0.4 

0.3 

0.2 

0.1 

0 

0 4 7 11 14 18 21 

Days of treatment 

Fig. 2. Effects of O 3 and CO 2 concentrations on net photosynthetic rate and stomatal conductance 

at vegetative growth stage (experiment 1, long-term; 1-4). 

= 0 ppm O 3 + 400 ppm CO 2 plot, = 0.1 ppm O 3 + 400 ppm CO 2 plot, 

= 0.1 ppm O 3 + 800 ppm CO 2 plot, = 0 ppm O 3 + 800 ppm CO 2 plot. 

vegetative growth stage, the main reason could be a decline in 

stomatal conductance (in other words, stomatal closure) induced 

by elevated CO 2 , which suppressed the invasion of O 3 

inside the leaf cavity as a physical barrier. The O 3 effect on 

yield could be attributed to an increase in infertile caryopses 

at lower positions of the rachis branch of the panicle because 

of a decline in photoassimilate accumulation. Further investigations 

should be done for the light environment, plant age, 

and biochemical reactions in treatments to understand the interaction 

of O 3 and CO 2 in lowland rice. 

References 

Imai K, Coleman DF, Yanagisawa T. 1985. Increase in atmospheric 

partial pressure of carbon dioxide and growth and yield of 

rice (Oryza sativa L.). Jpn. J. Crop Sci. 54:413-418. 

Ishioh T, Imai K. 2003. Effects of atmospheric ozone and carbon 

dioxide concentrations on growth and yield of lowland rice. 

Jpn. J. Crop Sci. 72(Suppl. 2):286-287. (In Japanese.) 

Ishioh T, Imai K. 2004. Interaction between ozone and carbon dioxide 

on gas exchanges in leaves of lowland rice. Jpn. J. Crop 

Sci. 73(Suppl. 2):240-241. (In Japanese.) 


Kim HY, Horie T, Nakagawa H, Wada K. 1996. Effects of elevated 

CO 2 concentration and high temperature on growth and yield 

of rice. 2. The effect on yield and its components of Akihikari 

rice. Jpn. J. Crop Sci. 65:644-651. (In Japanese.) 

Kobori K, Imai K. 2003. Effects of atmospheric ozone and carbon 

dioxide concentrations on photosynthesis of lowland rice. Jpn. 

J. Crop Sci. 72(Suppl. 2):284-285. (In Japanese.) 

Olszyk DM, Wise K. 1997. Interactive effects of elevated CO 2 and 

O 3 on rice and flacca tomato. Agric. Ecosyst. Environ. 66:1- 

10. 

Notes 

Authors’ address: School of Agriculture, Meiji University, Kawasaki, 

Kanagawa 214-8571, Japan, e-mail: imai@isc.meiji.ac.jp. 

Effects of topdressing on grain shape and grain damage 

under high temperature during ripening of rice 

Satoshi Morita, Osamu Kusuda, Jun-ichi Yonemaru, Akira Fukushima, and Hiroshi Nakano 

High temperatures during the rice-ripening period significantly 

diminish grain dry weight and increase grain damage (Nagato 

and Ebata 1960, Yoshida and Hara 1977, Tashiro and Wardlaw 

1991, Morita et al 2004). It is thought that unusually high temperatures 

in recent years have caused a reduction in first-grade 

quality rice in Japan (Terashima et al 2001). The major factors 

in grade reduction were an increase in immature grains with a 

white portion, deep ditch on the grain surface, and thin shape. 

Recently, it has been suggested that these adverse effects of 

high temperature on rice ripening have been accelerated by 

the reduced topdressing during the panicle formation stage 

designed to improve edibility (Terashima et al 2001). 

The objective of this study was to demonstrate the effect 

of topdressing at the panicle-formation stage on grain dry 

weight and the occurrence of immature grains with a white 

portion, deep ditch, and thin shape under high temperature. 

Recently, immature grains with a white portion have been identified 

by grain quality inspectors during the inspection system. 

However, immature grains with a deep ditch and thin shape 

were not detected by machines. To enable a numerical comparison, 

we developed a new method of determining grain shape 

using digitized images. 


Experiments were performed in fields of the National Agricultural 

Research Center for Kyushu Okinawa Region at Chikugo 

(33°12′N, 130°30′E), Japan. The seedlings of rice (Oryza 

sativa L. cv. Hinohikari) growing 4 weeks after sowing were 

transplanted to paddy fields (gray lowland soil) on 18 June 

2003. Planting density was 20.8 hills m –2 (30 cm × 16 cm), 

with 3 seedlings per hill. An experiment was established in a 

randomized block design with three replications (11 m 2 per 

replication). A basal dressing of 6 g N m –2 (commercial NPK 

(16:16:16) fertilizer) was applied in all plots. Topdressing at 

18 and 8 d before heading (DBH) with ammonium sulfate was 

in six patterns: 3–3 (g N m –2 ), 3–1.5, 3–0, 1.5–1.5, 1.5–0, and 

0–0. The average temperature during the 20 d after heading 

was 27.7 ºC, which was 5–6 ºC higher than the optimum temperature. 

To determine yield, yield components, and grain quality, 

48 hills (2.3-m 2 area) from each replication were harvested 

at maturity (2 October). Yield and 1,000-grain weight were 

recorded as filled grain (with a thickness of 1.7 mm or more) 

weights of brown rice adjusted to 15% grain water content. 

Grain-filling percentages were calculated as 100 × filled grain/ 

total spikelet number. Filled grains were graded by the inspection 

system, and grain quality was measured by a grain quality 

inspector (RGQI 10A, Satake, Japan). Nitrogen contents of 

filled grains were measured by the Japan Grain Inspection 

Association (Tokyo, Japan). 

For grain shape analysis, eight grains (3rd, 4th, 5th, and 

6th grains from the top on the 3rd and 4th rachis-branch in 

each panicle) of one panicle per plot were sampled. The 

samples were dehulled with a forceps and cut in half along the 

cross section. The grain images, viewing from the grain top or 

bottom, were digitized through a digital microscope at 50 magnifications. 

After transforming the digitized 24-bit full color 

images into binary ones, an image analysis was performed to 

measure grain contour with rθ polar barycentric coordinates (r 

indicates the distance, %, from the center point, half of X and 

Y radius, respectively, in the grain image, and θ denotes the 

angles of direction toward the vascular bundle). θ angles were 

within 0–180 degrees (i.e., numbers and areas of the cells at 

90 degrees were the averages of those at 270 degrees and 90 

degrees), since the top and bottom of a grain image showed no 

distinction from one another. 


Yield, yield components, and grain quality 

Grain yields were higher at higher levels of N topdressing because 

of higher grain numbers per unit area and higher 1,000- 

ripened-grain weights (Table 1). Grades of grain quality were 

higher, and the percentages of the immature grain number were 

lower, at higher N levels of topdressing, especially at 8 DBH 

(Table 1). There were no significant differences between the 


Table 1. Effects of topdressing at panicle formation stage on yield and grain quality. 

Treatments a Ripened b Grain no. Ripened 1,000- Grade of Sum of Milky Basal Grain N 

(g N m –2 ) grain yield per m 2 grain ripened-grain quality c immature white white content e 

(g m –2 ) (×100) (%) weight (g) (1–3) grains (%) 

3–3 620 304 90.0 22.7 1.7 39.4 8.4 6.2 1.24 * 

3–1.5 617 ns 306 ns 90.1 ns 22.5 ns 1.8 ns 42.6 ns 9.4 ns 9.4 ns 1.18 * 

3–0 593 ns 300 ns 89.4 ns 22.1 * 2.0 ns 46.7 * 10.0 ns 10.0 * 1.11 * 

1.5–1.5 578 ns 282 ns 91.7 ns 22.3 * 1.8 ns 44.5 * 7.8 ns 11.4 * 1.13 * 

1.5–0 561 * f 286 ns 90.7 ns 21.6 * 2.0 ns 51.5 * 8.2 ns 14.0 * 1.07 * 

0–0 544 * 281 ns 90.3 ns 21.5 * 2.3 * 53.5 * 8.8 ns 13.4 * 1.06 * 

a Topdressing at 18 and 8 days before heading with ammonium sulfate. b Grains with a thickness of 1.7 mm or more were regarded as ripened grains. c By inspection system. 1st 

high = 1, 1st low = 1.5, 2nd high = 2, 2nd low = 2.5, 3rd high = 3, 3rd low = 3.5. d Measured by RGQI 10A (Satake). e Measured by the Japan Grain Inspection Association. 

f *Indicates a significant difference from the 3–3 treatment at the 5% level by Dunnett’s multiple range test. 

Ratio between the distance from the center point to the contour 

of grain cross section of 3-3 treatment and that of other treatments 

1.02 

1.00 

3-3 treatment = 1 

0.98 

0.96 

3-3 

3-0 

1.5-0 

0.0 

0.94 

0 30 60 90 120 150 180 

Angle of direction toward vascular bundle (º) 

Fig. 1. Comparison between the grain contour of 3-3 treatment and that of other 

treatments. Standard errors of the ratios between the distance from the center 

point to the grain contour of the 3-3 treatment and that of other treatments 

ranged from 0.007 (1.5-0 treatment) to 0.016 (3-0 treatment), on average, from 

0 to 180º. 

percentages of milky white grains of the 3–3 treatment and 

those of other treatments. However, the percentages of basal 

white grains of the 3–3 treatment were significantly lower than 

those of other treatments, except for the 3–1.5 treatment. These 

results coincide with the report that immature grains with a 

white portion increased at a higher grain number per unit area 

(Terashima et al 2001, Kobata et al 2004) and at a lower N 

level (Terashima et al 2001) under high temperature at ripening. 

In this study, it was clear that the reactions to the increase 

in the nitrogen application change with the type of immature 

grain (i.e., milky white or basal white) under high temperature. 

But, the reason for this has been unclear. Further research 

into the physiological and morphological factors of white portions 

in the endosperm will be required. 

Grain shape 

Image analysis of grain projection from the grain top or bottom 

revealed that lower N levels in the reproductive stage under 

high temperature caused inferior grain growth, especially 

at 30 to 50 degree angles from the vascular bundle direction 

from the center point of the cross section (Fig. 1), where the 

photoassimilates were thought to be accumulated at a later stage 

of ripening. These areas of angles contain a deep ditch made 

from the joint line of the lemma and palea. 

Image analysis of the longitudinal projection from the 

grain lateral side also revealed that lower N levels caused inferior 

growth at ventral shoulder parts (from –120 to –150 

degrees) and at the dorsal bottom square parts (from 30 to 60 

degrees) of grains (Fig. 1). The latter parts corresponded to 

the parts of the white portion in the basal white grain. 

Akiyama et al (1997) have classified rice cultivars by 

image analysis of the size of white cores in the grain. However, 

image analysis of rice grain shape had not been previously 

reported. In this study, the inferior growth position in 

the grains, revealed by image analysis of grain shape, may indicate 

the time and/or position in which a lower nitrogen application 

under high temperature has an immediate adverse 

effect. This suggests that image analysis of grain shape will be 


of value in understanding the adverse effect of high temperature 

and in improving rice ripening under such an environment. 

References 

Akiyama Y, Yamada H, Takahara Y, Yamamoto K. 1997. Classification 

of rice cultivars for sake brewery based on white-core 

characters. Breed. Sci. 47:267-270. 

Kobata T, Uemuki N, Inamura T, Kagata H. 2004. Shortage of assimilate 

supply to grain increases the proportion of milky white 

rice kernels under high temperatures. Jpn. J. Crop Sci. 73:315- 

322. (In Japanese with English summary.) 

Morita S, Shiratsuchi H, Takanashi J, Fujita K. 2004. Effect of high 

temperature on ripening in rice plant: analysis of the effects 

of high night and high day temperatures applied to the panicle 

and other parts of the plant. Jpn. J. Crop Sci. 73:77-83. (In 

Japanese with English summary.) 

Nagato K, Ebata M. 1960. Effects of temperature in the ripening 

periods upon the development and qualities of lowland rice 

kernels. Proc. Crop Sci. Soc. Jpn. 28:275-278. (In Japanese 

with English summary.) 

Tashiro T, Wardlaw IF. 1991. The effect of high temperature on kernel 

dimensions and the type of occurrence of kernel damage 

in rice (Oryza sativa L.). Aust. J. Agric. Res. 42:485-496. 

Terashima K, Saito Y, Sakai N, Watanabe T, Ogata T, Akita S. 2001. 

Effects of high air temperature in summer of 1999 on ripening 

and grain quality of rice. Jpn. J. Crop Sci. 70:449-458. (In 

Japanese with English summary.) 

Yoshida S, Hara T. 1977. Influence of air temperature and light on 

grain filling of an indica and japonica rice (Oryza sativa L.) 

in a controlled environment. Soil Sci. Plant Nutr. 23:93-107. 

Notes 

Authors’ addresses: Satoshi Morita, Akira Fukushima, and Hiroshi 

Nakano, NARO, National Agricultural Research Center for 

Kyushu Okinawa Region, Fukuoka 833-0041, Japan; Osamu 

Kusuda, NARO, National Institute of Vegetable and Tea Science, 

Mie 514-2392, Japan; Jun-ichi Yonemaru, NARO, National 

Agricultural Research Center for Tohoku Region, 

Morioka 020-0198, Japan, e-mail: moritas@affrc.go.jp, 

kusuda@affrc.go.jp, yonemaru@affrc.go.jp, 

afuku@affrc.go.jp, nakanohr@affrc.go.jp. 


The conclusions of the Third Assessment Report (TAR) of the 

Intergovernmental Panel on Climate Change (IPCC) leave no doubt 

that Earth’s climate is changing in a manner unprecedented in 

the past 400,000 years. The report forecasts that by 2100 mean 

planet-wide surface temperatures will rise by 1.4 to 5.8 °C, precipitation 

will decrease in the subtropics, and extreme events will 

become more frequent. However, changes in climate are already 

being observed—the last 60 years were the warmest in the last 

1,000 years and changes in precipitation patterns have brought 

greater incidence of floods or drought globally. The TAR also concluded 

that natural phenomena do not explain these observed 

changes—that there is now clear evidence of human influence. 

Climate change is a development problem as well as an 

environmental problem. It is a development problem because 

those engaged in agriculture must learn how to adapt to the 

consequences of an altered atmosphere, higher temperatures, 

and greater variability and more extremes in weather. Coping simultaneously 

with the other adverse environmental consequences 

of anthropogenic activity, such as air quality, will be no easy matter. 

The combination of changes in air quality and composition, 

acid rain, and climate will produce a new bioclimate for food 

production systems. The potentially beneficial effects of increases 

in CO 2 may be offset by the increases in ozone at ground level 

(tropospheric concentrations) as well as high-temperature stress. 

It is arguable that most developing countries are not well informed 

about the probable impact of bioclimatic change and are therefore 

most vulnerable to its consequences. 

To understand the quantitative effects of climate change 

on rice production, it is necessary to have rice simulation models 

that can predict the outcome of the new bioclimates on rice. The 

models have to have parameters measured in “controlled environments.” 

The first paper of the session by T. Horie and H. Yoshida 

described the development of such a model, with particular reference 

to high-temperature stress and its use to predict future 

yields in China, Thailand, and Japan. 

One curious consequence of climate change is that in some 

regions temperatures could get cooler because of changes in 

currents and atmospheric pressure differentials. The second paper 

of the session by M. Yajima focused on the effects of cool 

weather in northern Japan on spikelet sterility and methods of 

protecting rice against cold damage. 

Coping with the effects on rice of weather variability associated 

with climate change in the form of increased frequency of 

wet and dry periods, tropical cyclones, and typhoons in the Philippines 

was the topic of the third paper by F. Lansigan. 

Climate change will have direct effects on rice but may 

also alter soil nutrient availability for rice. The fourth paper by J. 

Zhu described how free-air CO 2 enrichments can affect plant 

nutrients in the soil solution. 

Rice ecosystems will be affected by climate change, but 

there has also been great concern about the contribution of rice 

to climate change. Understanding the properties of rice soils in 

relation to methane emissions is crucial to efforts to minimize 

those emissions and this topic was explored by S. Mitra and D. 

Majumdar. 

Professor Li and his colleagues investigated the alternatives 

for farmers in managing their crops so as to reduce methane 

emissions. They developed a tool by integrating a processbased 

model with a GIS database to scale up their experimental 

observations from the field to a regional level. They investigated 


the effects of mid-season drainage in different cropping seasons 

and on different soils. Their results highlighted the benefits of 

their management systems in reducing emissions of methane, 

but revealed the associated problem of stimulating the emission 

of nitrous oxide. 

Climate change studies need to cover a range of issues at 

widely different time and space scales. Many uncertainties still 

exist in our predictions of future climates, rice production, and 

impacts of rice cultivation on the environment at each scale. To 

reduce uncertainties and to develop mitigation and adaptation 

strategies, three key components should work together: robust 

modeling, large-scale databasing, and continuous experimentation. 

The great challenge is to achieve a balance between rice 

cultivation and environmental conservation as addressed by T. 

Horie and K. Minami in the general discussion. 


SESSION 20 

Improving rice productivity through IT 

CONVENER: S. Ninomiya (NARO) 

CO-CONVENER: M. Bell (IRRI)

Data mining using combined yield 

and quality maps of paddy fields 

Tadashi Chosa 

Research on precision agriculture has been attracting attention 

with regard to the production of rice in Japan (Umeda 

1999) and other Asian countries (Sung et al 2002). It is thought 

that precision farming could solve the problem of variability 

within and between fields, leading to economic and environmental 

benefits, as well as stable production. The methods of 

precision agriculture are based on information technology. In 

precision farming, farmers make optimal decisions depending 

on site-specific information, such as on fertility, growth, and 

yield. 

In recent times, progress has been made in developing 

information acquisition technology (Shibata et al 2002, Chosa 

et al 2004); progress has also been made with respect to applications 

for site-specific management (Chosa et al 2003). However, 

it is difficult to make recommendations for management 

or applications based on the information acquired. Although 

some successful case studies have been reported (Toriyama et 

al 2003), no general method for producing recommendations 

has been determined. 

Theoretically, if the fertility and variations in fertility of 

an area can be determined, the added nitrogen required to reach 

a targeted nitrogen absorption level, or to decrease variability, 

can be calculated. Similarly, if the absorption of nitrogen is 

determined through yield, a revised amount of nitrogen can be 

calculated for application the following year. However, nitrogen 

absorption is not the same in each area, for each year, 

because of differences in soil conditions, climate, fertilizer 

application method, and type of cultivation. For this and other 

reasons, it is difficult to generalize the formulation of a recommendation. 

This paper presents guidelines for making recommendations 

for the next year’s crop using yield and grain quality 

maps of paddy fields. The information provided by mining the 

data in these maps should be more useful than the information 

provided by the individual maps. Combining the two types of 

maps for the purpose of site-specific management can reveal 

useful information. 

Yield and quality acquisition systems 

To acquire yield information, a hybrid yield-monitoring method 

(Chosa et al 2004a) was used. This method uses a sequencemonitoring 

sensor simultaneously with a batch-monitoring sensor. 

The sequence yield monitor, which is used to measure the 

grain flow rate, consists of an emitting unit and a receiving 

unit fixed to the top of the clean grain auger of the combine 

harvester. The output signal is an analog signal that varies according 

to the particle flow between the emitting and receiving 

units. A batch yield monitor is used to measure the mass of 

grain inside the combine tank. This monitor is a load cell unit 

that is fixed under the grain tank of the harvester and responds 

to particles filling the tank. The output signal of this sensor is 

also analog. As position information is acquired by GPS at the 

same time as yield information, the yield map is drawn after 

the harvest. 

A gathering unit for analysis of the quality of samples 

(Chosa et al 2004b) that can be attached to a combine harvester 

was used to acquire quality information. This system 

consists of a leading tube in the form of a cylinder, receiving 

cups that gather the grain from the leading tube, and a turning 

table that supplies the receiving cups. The leading tube is inserted 

into the combine tank at an angle and leads the grain 

sample from inside the combine tank to the outside. As the end 

of the tube inside the tank is cut obliquely, the angle of the 

tube controls the sampling weight and can be used to stop the 

sampling. When the sampling is stopped, normal operations 

can be conducted. Grain dropped through the leading tube is 

gathered in sampling cups that are supplied by the turning table. 

Grain samples corresponding to random harvest positions can 

be collected through appropriate operation of the turning table. 

After the harvest, the quality parameters of the samples collected 

can be analyzed, such as the moisture content, protein 

content, and outward quality. Since position information is 

acquired through the GPS at the time of sampling, the quality 

map is drawn after the analysis. 

Farm work for data acquisition 

To examine variability in yield and quality, a field survey was 

conducted during combine harvesting. 

The combine harvester used was a modified commercial 

four-row operating combine harvester (head-feeding type, 

25.7 kW, harvesting width 1.2 m). The combine was equipped 

with the yield and quality acquisition systems described above. 

Commercial combine harvesters can easily be equipped with 

both systems. Although the person who runs the combine must 

operate the information acquisition systems, the harvesting operation 

is otherwise the same as is usual for this farm work. 

After the harvest and analysis of the yield, quality, and 

position information, the yield map and quality map were generated. 

Protein content was used as a measure of rice quality. 

The protein content of each sample was measured using nearinfrared 

light. 

The yield and quality variability maps were overlapped 

to determine a recommendation for cropping and fertilizing 

strategies for the next year. 


Table 1. Field categories on the basis of yield and protein content. 

Category Yield Protein Analysis and prescription 

1 High Low Proper management was conducted. 

The same management 

is recommended for the next 

year. 

2 Low Low If the priority is yield, more nitrogen 

is required. 

3 High High If the priority is quality, less nitrogen 

is required. 

4 Low High This problem cannot be controlled 

by fertilizing. Other causes such 

as soil structure, disease, and 

insects should be discussed. 


Variations in yield and quality were observed. The cropping 

and fertilizing strategies for the next year, determined by these 

maps, indicate that areas with lower yields will require more 

nitrogen, and areas that produce rice with excessive protein 

content will require less nitrogen. However, the situation with 

nitrogen absorption is further complicated because of differences 

in soil conditions, climate, fertilizer application, and type 

of cultivation. Occasionally, contradictions arise; for example, 

areas that produce rice with low protein content sometimes 

have high yield. Basing the fertilizing strategy on calculation 

of the nitrogen absorption may not result in an optimal recommendation. 

Therefore, another data-mining method for determining 

fertilizing strategies that use yield and quality maps is 

proposed here. 

Initially, each area is divided into two categories, based 

on the yield map: the high-yield area and low-yield area. Next, 

each area is also divided into two categories based on the quality 

map of the same time period, the low-protein-content area and 

high-protein-content area. Finally, each part of the field is classified 

into four categories by combining all of the above categories: 

category 1 has high yield and low protein content, 

category 2 has low yield and low protein content, category 3 

has high yield and high protein content, and category 4 has 

low yield and high protein content (Table 1). Ideally, every 

area should be managed to have the qualities of category 1. 

Some of the categories may require the same cropping and 

fertilizing strategies in the subsequent year. If the priority is 

yield, the category 2 areas require more nitrogen. If the priority 

is quality, the category 3 areas require less nitrogen. It is 

natural to assume that other problems that cannot be controlled 

by fertilizing, such as soil structure, disease, and insects, occur 

in category 4 areas. Therefore, more surveys may be required 

for category 4 areas to determine an appropriate recommendation. 

Figure 1 shows a category map based on yield 

and quality maps. Maps of this type will constitute useful tools 

for determining recommendations for cropping and fertilizing 

strategies for the following year. A further refinement to the 

method would be to incorporate logical thresholds, and actual 

proof with respect to the recommendations would also be informative. 

(m) 

30 

20 

10 

0 

0 10 20 30 40 50 60 70 80 90 100 

(m) 

Conclusions 

This study has presented guidelines for making recommendations 

for site-specific crop management. Information on yield 

and quality can be obtained using this newly developed system. 

A new data-mining method for determining fertilizing strategies 

using yield and quality maps has been presented here. 

The category maps resulting from this method suggest qualitative 

strategies for the crop of the following year. 

References 

Chosa T, Shibata Y, Omine M, Kobayashi K, Toriyama K, Sasaki R. 

2003. Map-based variable control system for granule applicator. 

J. Jpn. Soc. Agric. Machine. 65(3):128-135. 

Chosa T, Shibata Y, Omine M, Toriyama K, Araki K. 2004a. A study 

on a yield monitoring system for head-feeding combines. J. 

Jpn. Soc. Agric. Machine. 66(2):137-144. 

Chosa T, Shibata Y, Omine M, Nozaki I, Sekiguchi M, Hosokawa H. 

2004b. Development of a gathering unit for quality analyzing 

sample to be put on a combine harvester. Proceedings of the 

40th Annual Meeting of JSAM Kanto branch. p 42-43. 

Shibata Y, Sasaki R, Toriyama K, Araki K, Asano O, Hirokawa M. 

2002. Development of image mapping techniques for site-specific 

paddy rice management. J. Jpn. Soc. Agric. Machine. 

64(1):127-135. 

Sung J, Lee C, Park W, Jung I, Kim S. 2002. The spatial variability 

of yield, growth status, soil properties, elevation and quality 

in Korean rice. ASAE paper: 023085. 

Toriyama K, Sasaki R, Shibata Y, Sugimoto M, Chosa T, Omine M, 

Saito J. 2003. Development of a site-specific nitrogen management 

system for paddy rice. Jpn. Agric. Res. Q. 37(4):231- 

218. 

Umeda M. 1999. Precision Agriculture Project in Kyoto University. 

J. Jpn. Soc. Agric. Machine. 61(4):4-11. 

Notes 

Catergory 1: high yield and low protein content 

Catergory 2: low yield and low protein content 

Catergory 3: high yield and high protein content 

Catergory 4: low yield and high protein content 

Fig. 1. Category map based on yield and quality. 

Author’s address: NARC, Hokuriku, 1-2-1 Inada, Joetsu, Niigata 

943-0193, Japan, tchosa@affrc.go.jp. 

Session 20: Improving rice productivity through IT 567

A wireless sensor network with Field-Monitoring Servers 

and MetBroker in paddy fields 

Masayuki Hirafuji, Tokihiro Fukatsu, Hu Haoming, Takuji Kiura, Tominari Watanabe, and Seishi Ninomiya 

Environmental data such as weather data and crop data such 

as rice growth in paddy fields are necessary for crop management 

and scientific studies. In addition, production history 

systems and traceability systems for food security are becoming 

indispensable. A wireless sensor network could be one of 

the best solutions for this need. 

So far, many kinds of sensor-network solutions, such as 

Mote (Khan et al 1999) and TINI (www.ibutton.com/TINI/), 

have been proposed. However, specification of the sensors is 

poor for monitoring crops. Wireless broadband communication, 

high-resolution image-monitoring technology, and various 

sensors are needed for monitoring rice in real-time in paddy 

fields. For example, a rice blast prediction system, 

MetBLASTAM (http://cse.naro.affrc.go.jp/ketanaka/model/ 

applet/Blastam.html), requires information on air temperature, 

humidity, and leaf wetness. Specific data such as images of 

emerging rice blast are indispensable to revise the prediction 

system. To estimate the photosynthetic rate of rice, the spatial 

distribution of CO 2 concentration should be measured in realtime 

in the paddy field. 

We developed Field-Monitoring Servers for those requirements, 

and constructed a wireless sensor network in paddy 

fields. Their cost is extremely low and their functions are much 

advanced compared with conventional sensor networks or 

weather stations. The observed data are freely available on 

our Web site. Any users can use the data directory, and applications 

such as the rice blast prediction system, through 

MetBroker, which provides data in standard format for applications 

by linking these applications to conventional weather 

databases such as AMeDAS and NOAA. 

The Field-Monitoring Server 

A field server is a Web server installed in fields, and the Field- 

Monitoring Server (FMS) is a kind of field server for realtime 

monitoring. The FMS can monitor fields in real-time and 

use wireless local area network (LAN) hotspots around the 

FMS (Hirafuji 2000). The FMS consists of a field server engine 

(micro Web server for control and data acquisition), a 

wireless LAN access-point, a network camera, and sensors 

(Hirafuji and Fukatsu 2002, Fukatsu and Hirafuji 2003). 

The field server engine is a main panel that functions 

like a Web server, analog I/O, digital I/O, DDS (Direct Digital 

Synthesizer), FPAA (Field Programmable Analog Array), and 

analog multipliers. The DDS generates an alternative signal 

up to 70 MHz. The FPAA serves to construct dynamic analog 

circuits, such as a low-pass filter and super-heterodyne radio 

receiver, using its embedded 20 CAB (Configurable Analog 

Block) and analog multipliers. By combining these elements, 

the FMS can be connected to various sensors and probes directly. 

A typical specification of FMSs follows: 

1. Casing: heat-resistant, waterproof, dustproof, and suitable 

to the landscape. 

2. Sensors: air temperature, humidity, PPFD (photosynthesis 

photon flux density), soil temperature, soil 

moisture, CO 2 concentration, leaf wetness, ultraviolet 

light, infrared light, etc. 

3. Image data: 0.3–8 M pixels depending on cameras. 

4. Wireless communication range: 300 m–10 km depending 

on external antennas. 

5. Hotspot service: within 100 m–1.5 km depending on 

antennas and circumstances. 

6. Communication speed: 54/11 Mbps. 

7. Power supply: AC adapter, embedded solar cell, or 

external solar cells. 

8. Size: diameter is 20 cm and length is about 30 cm. 

9. Cost: less than $500 to $2,000 per unit depending on 

embedded devices. 

The full-wireless Field-Monitoring Server for paddy fields 

Conventional FMSs require cables for the power supply, since 

power from the solar cell embedded on the rooftop of the FMSs 

is too low (0.5 W) to drive an FMS continuously, as it requires 

more than 50 W for nights and the rainy season. The hardest 

job for the installation of FMSs was the construction of wiring. 

This will be much harder in paddy fields. So, the fullwireless 

FMS was developed, especially for paddy fields with 

several energy-saving technologies (Hirafuji et al 2004). Client 

connection FMSs can rest periodically to save power. The 

full-wireless FMSs connect other conventional standard FMSs 

as clients at hotspots provided by the standard FMSs, which 

are driven by an AC adapter or larger solar cells and consume 

more energy to maintain the backbone and the hotspot service. 

Fieldserver-Agents and MetBroker 

Field servers can be controlled by users using a Web browser, 

also simultaneously controlled by a Fieldserver-Agent, which 

is a kind of artificial intelligence (Fukatsu and Hirafuji 2004). 

The collected data are stored in public database servers such 

as a PC-cluster at NARC or a computer center. A Fieldserver- 

Agent controls almost all FMSs in the world and collects data 

automatically, such as the Google Web crawler. The agent is 

accessing FMSs every 1–10 min. Farmers and consumers can 


Fig. 1. Field Monitoring Servers in the paddy field. 

view the data. The sorted images are shown as time-lapse 

movies to monitor plant growth and farms. This function can 

be a production history system and a farm security system. 

Applications can use the data by using MetBroker 

(Laurenson et al 2002), which has transformed conventional 

weather databases and data of FMSs into standardized data 

automatically. MetBroker is also a network service. Applications 

determined to be used by MetBroker can access an enormous 

amount of data through MetBroker. If a new FMS is 

installed in a field, the measured data instantly become available 

applications. 

Experiments 

We installed five FMSs; three FMSs are in a paddy field and 

one FMS is on a rooftop at NARC’s Yawara experiment station 

in Yawara village. The first FMS (FMS-1) on the rooftop 

is connected to the Internet by the Asymmetric Digital Subscriber 

Line (ADSL), since the Yawara station is not connected 

to the optical fiber network of MAFFIN (the Ministry of Agriculture, 

Forestry, and Fisheries Network). The second FMS 

(FMS-2) is connected to FMS-1 by a repeating mode of wireless 

AP (a kind of ad hoc network), and the third FMS (FMS- 

3) is also connected to the FMS-2 by repeating. The fourth 

FMS (FMS-4) and fifth FMS (FMS-5) are connected by an 

AP-Client mode as a client around FMS-2, which provides a 

hotspot service. FMS-4 and FMS-5 were installed in the paddy 

field (Fig. 1). The power for FMS-4 was supplied by a cable, 

which was hard to use in a such a dirty field. So, we employed 

a full-wireless FMS for the fifth FMS. Power for the FMS-5 

was supplied by an embedded small solar cell. Sensors for the 

FMSs are 

FMS-1: air temperature, humidity, intensity of solar radiation 

(power of solar cell), and inside temperature. 

FMS-2: air temperature (2 types), humidity, intensity of 

solar radiation, inside temperature, and CCD 

camera (0.3 Megapixels). 

FMS-3: air temperature, CO 2 concentration, leaf wetness, 

intensity of UV, and inside temperature. 

FMS-4: air temperature, PPFD (2 layers), and inside temperature. 

FMS-5: air temperature, humidity, recharging current (intensity 

of solar radiation), battery voltage, and 

inside temperature. 

In addition, we tested a 3-Megapixel high-resolution 

digital camera with a magnifying lens to FMS-2. We could 

observe the surface of a rice leaf; the resolution is up to about 

10 micrometers (Fig. 2). This microscope camera can be installed 

in the paddy field and also controlled by the Fieldserver- 

Agent. In total, 23 sensors and two cameras were installed in 

the paddy field. All of the measured data were collected by the 

Fieldserver-Agent automatically. 

We isolated the FMSs from the global IP network using 

VPN for security using FieldServerGateway (Kiura et al 2002). 

However, real-time data and collected data are always available 

on the Field Server Web site (http://model.job.affrc.go.jp/ 

FieldServer/monitor/). 


We needed four visits to the Yawara station for installation of 

the five FMSs and ADSL router. That is troublesome for farmers 

and researchers, so we improved the FMSs and now we 

can finish such installation within a half day. The function of 

the hotspot and ad hoc network could be connected among the 

FMSs successfully, and FMSs now provide a ubiquitous network 

service at the station. The CO 2 concentration in the paddy 

field can be measured continuously and the relationship with 

other monitored data such as UV light is clear. 


Fig. 2. Surface of a rice leaf: live high-resolution image recorded by the Field- 

Monitoring Servers with other environmental data in the paddy field. 

So far, the FMSs were installed in both the paddy field 

and at many experimental sites in several countries such as 

Japan, the United States, Thailand, Denmark, China, Korea, 

Canada, and Taiwan (China). The FMSs in paddy fields became 

extremely dirty after one year, and the sensors were damaged. 

So, we are improving the FMSs; for example, an air filter 

was appended to the air intake. 

Conclusions 

A wireless sensor network is constructed in paddy fields using 

FMSs. We can access the Internet in the area ubiquitously by a 

Wi-Fi wireless-LAN. The sensor network has collected data 

in real-time since July 2003. The collected data were used to 

develop new sensors such as a leaf-wetness sensor. Relationships 

hidden in enormous amounts of measured data and 

weather databases can be found by MetBroker and data mining. 

So far, we lack detailed data in paddy fields where we 

could not install measurement equipment, and only some data 

and numerical estimations have been employed. At present, 

we can now collect an enormous amount of data automatically 

and monitor them in real-time by live cameras with the FMSs. 

Any people, such as researchers and consumers, can inspect 

paddy fields anytime at any place using the collected data. 

This can make rice production more scientific, more reliable, 

and safer. 

References 

Fukatsu T, Hirafuji M. 2003. Development of Field Servers for a 

field monitoring system. Agric. Info. Res. 12:1-12. 

Fukastu T, Hirafuji M. 2004. The agent system for Field Monitoring 

Servers to construct smart sensor-network. Fifth International 

Workshop on Artificial Intelligence in Agriculture, 8-10 March 

2004. 

Hirafuji M. 2000. Creating comfortable, amazing, exciting and diverse 

lives with CYFARS (CYber FARmerS), an agricultural 

virtual corporation. Proceedings of the Second Asian Conference 

for Information Technology in Agriculture. p 424-431. 

Hirafuji M, Fukatsu T. 2002. Architecture of Field Monitoring Servers. 

Proceedings of the Third Asian Conference for Information 

Technology in Agriculture. p 405-409. 

Hirafuji M, Fukatsu T, Hu Haoming. 2004. Full-wireless Field Monitoring 

Server for advanced sensor network. Proceedings of 

AFITA/WCCA, 2004. p 686-691. 

Kahn JM, Katz RH, Pister KSJ. 1999. Mobile networking for smart 

dust. ACM/IEEE International Conference on Mobile Computing 

and Networking (MobiCom 99), Seattle, Wash., USA. 

Kiura T, Fukatsu T, Hirafuji M. 2002. Field server gateway: gateway 

box for Field Monitoring Servers. Proceedings of the Third 

Asian Conference for Information Technology in Agriculture. 

p 410-413. 

Laurenson M, Kiura T, Ninomiya S. 2002. Providing agricultural 

models with mediated access to heterogeneous weather databases. 

Appl. Engin. Agric. 18:617-625. 

Notes 

Authors’ addresses: National Agricultural Research Center, e-mail: 

hirafuji@affrc.go.jp, fukatsu@affrc.go.jp, hhaoming@ 

affrc.go.jp, kiura@affrc.go.jp, tomonari.watanabe@ 

affrc.go.jp, snino@affrc.go.jp. 


Prediction of airborne immigration of rice insect pests 

Akira Otuka, Tomonari Watanabe, Yoshito Suzuki, Masaya Matsumura, Akiko Furuno, and Masamichi Chino 

The whitebacked planthopper Sogatella furcifera and the 

brown planthopper Nilaparvata lugens are major pests of rice 

in eastern Asia. They migrate from southern China into Japan 

mainly in the Bai-u rainy season in June to July every year 

(Seino et al 1987). Generally, the prediction of planthopper 

migration provides farmers and plant protection advisers with 

information about the migration source as well as area of migration. 

It is especially important to know the migration source, 

which helps to understand the immigrant’s characteristics such 

as biotype, pesticide resistance, etc. Conventional prediction 

methods used 6- or 12-hour two-dimensional wind data 

(Rosenberg and Magor 1983, Watanabe et al 1990), but their 

prediction quality was limited. A three-dimensional simulation 

method using a boundary layer model was developed 

(Turner et al 1999). However, a prediction system based on 

this model has not yet been developed. 

To achieve high-precision migration prediction, a realtime 

prediction system was developed using three-dimensional 

simulation models. This paper presents the simulation method 

and its evaluation results. 


Planthopper behavior 

Sogatella furcifera and N. lugens are tiny insects about 3–4 

mm in size and 1–3 mg in weight (Ohkubo 1973). Many field 

observations have shown that the planthoppers take off at either 

dusk or dawn (e.g., Ohkubo and Kisimoto 1971). A radar 

observation showed that S. furcifera and N. lugens flew to an 

altitude of several hundred to 1,000 m above the ground at an 

estimated upward speed of 0.2 m s –1 (Riley et al 1991). The 

species fly at about 1 m s –1 , whereas the wind speed when 

migration occurs is typically more than 10 m s –1 (Seino et al 

1987). Laboratory experiments of tethered N. lugens adults 

indicated that they have the ability to fly for up to 23 hours in 

air of high humidity (Ohkubo 1973). N. lugens ceases flying 

at cooler temperatures; half of them stop beating their wings 

when the air temperature is below 16.5 °C (Ohkubo 1973). 

Their postmigration landing process is not yet fully understood. 

Planthopper migration simulation model 

Takeoff area. Because information on the planthopper’s density 

in source regions was not available, it was assumed that 

the population density in June to July was high enough for the 

planthopper to take off from source regions. In the simulation, 

several takeoff areas of 50–100 km 2 were set up in southern 

China and Taiwan. These takeoff areas were located in paddy 

fields, and covered the major source regions. 

Model. The migration simulation model originates from 

a particle dispersion model, GEARN, developed by the Japan 

Atomic Energy Research Institute (JAERI), which calculates 

atmospheric dispersion of radioactive particles in case of a 

nuclear accident (Ishikawa and Chino 1991). GEARN was 

modified to model the planthopper’s migration behavior. Figure 

1 shows a schematic model. This model calculates the position 

of many planthoppers, and doesn’t discriminate between 

N. lugens and S. furcifera. Nearly 2,000 planthoppers for each 

takeoff area took off randomly within the area at 10 or 21 UTC 

(Universal Time Coordinate), which locally corresponds to 

dusk or dawn, respectively. They then flew up at a speed of 

0.2 m s –1 for 1 hour. Taking off ceased 1 hour later, at 11 or 22 

UTC. During their flight over the sea, the planthoppers moved 

at the same speed as the wind because they are slow fliers. 

Vertical diffusion was taken into account by a random walk 

model, but horizontal diffusion was not because preliminary 

results showed too much horizontal diffusion over Japan. Since 

the planthoppers don’t like cooler air, a temperature ceiling of 

16.5 ºC was set up, and they could not go beyond that ceiling. 

Wind, temperature, and vertical diffusion coefficient data were 

given by numerical weather forecasts. Simulation duration was 

48 h. The model calculated the relative aerial density of 

planthoppers based on their position on the model grid. The 

horizontal resolution was 33 km. Areas of relative density larger 

than zero in the lowest level, which was less than 100 m above 

the ground, were used for prediction. These areas were called 

“migration clouds” (Fig. 2). 

Prediction system 

First, the latest meteorological data were supplied online to an 

advanced numerical weather prediction model, MM5 (Grell 

et al 1994). The model forecast atmospheric fields for the next 

72 hours at 1-hour intervals. Second, these forecast fields were 

supplied to the planthopper migration simulation model, 

GEARN, which calculated displacement of a number of modeled 

planthoppers and predicted relative aerial density. The 

data were converted to PDF files at 3-hour intervals and sent 

to the project’s Web site. These processes were conducted automatically. 

The system predicted migrations over the next two 

days. The maps of relative aerial density provided information 

on the timing and area of arrivals. The system also gave 

possible migration sources. 


Evaluation 

Figure 2 shows an example of predicted migration clouds. An 

evaluation was conducted using daily catch data obtained at 

three sites in Kyushu, western Japan, in 2003 and 2004. The 

sites were Saga (33.17°N, 130.33°E), Kumamoto (32.95°N, 

130.78°E), and Kagoshima (31.52°N, 130.50°E). The system 


Takeoff Flight Over Japan 

Temp. ceiling of 16.5 ºC 

Wind + V-diffusion 

0.2 m s –1 for 1 h 

33 km 

100 m 

50 km 

Sim. grid 

China and Taiwan East China Sea Japan 

Fig. 1. Schematic of the migration simulation model. Black circles represent modeled planthoppers. As 

many as 1,852 planthoppers per one takeoff area were randomly generated at 10 or 21 UTC. The 

simulation duration was 48 h. 

E 

N 

40º 

110º 120º 130º 

40º 

tions were evaluated by a hitting ratio defined as: 

Hitting ratio (%) = Number of days when the migration 

was correctly predicted/Total number of days × 100 

30º 

30º 

The results in 2003 showed a hitting ratio of 79%, and with 

84% the result in 2004. The average hitting ratio of rainfall 

forecasts for the Kyushu region in June in the two years was 

80%. This number was comparable to that of migration prediction. 

20º 

110º 120º 130º 

20º 

Fig. 2. An example of migration clouds. This figure shows the predicted 

distribution of relative density of planthoppers at 6 UTC, 

25 June 2004, which started at 21 UTC from all the takeoff areas. 

predicted daily migrations by seeing whether a migration cloud 

would cover the evaluation site or not. To judge predictions, 

daily catches of S. furcifera captured at the sites in June to 

mid-July were used. Catches of N. lugens were too small to be 

used for the evaluation. Traps used were net traps and a 

Johnson-Taylor-type suction trap. The net trap was a 1-m tow 

net mounted at the top of a pole 10 m high. More than one 

catch was interpreted as a migration event. The daily predic- 

Discussion 

This is the first real-time prediction system to use three-dimensional 

wind fields given numerically by weather forecasts. 

The system enabled migration prediction that presented information 

about the timing and area of migrations. Nonetheless, 

several issues need to be discussed. 

The takeoff assumption. The strong assumption made 

for the prediction was that planthoppers were supposed to take 

off from every predefined takeoff area at every takeoff time. 

The regions where planthopper density is high in June and 

July are located in southern China at latitudes less than 25°N 

(Zhou et al 1995). The takeoff areas in Fujian, Guangdong, 

Guangxi, and Hainan provinces qualify as such regions. In some 

years, however, northern regions around 30°N have been invaded 

by the beginning of June. Therefore, the other northern 

takeoff areas were included. Since the theoretically designated 

takeoff areas widely cover possible source regions, this assumption 

seems reasonable. 


The evaluation period. This was limited to June to mid- 

July for several reasons. The first was that the evaluation period 

must match the takeoff assumption. Conditions in source 

regions in earlier months, such as in April and May, would not 

match the assumption very well since planthopper density 

would still be low (Zhou et al 1995). Therefore, if the evaluation 

period included the earlier months, the evaluation would 

result in too many false positives, and hence an underestimation 

of the hitting ratio. The second reason is that most migrations 

from overseas are concentrated in the Bai-u rainy season, 

making it the period of greatest concern. Third, from an 

evaluation point of view, since preceding migrations in earlier 

months were limited in number, the contamination of local 

planthoppers in the trap data in June and July was expected to 

be low, which was an important condition in conducting an 

accurate evaluation. A local planthopper is defined as a 

planthopper captured by the trap because of taking off shortly 

after landing in Japan, or as a “second-generation” planthopper, 

born to immigrants. After mid-July, evaluation becomes rather 

difficult because of the contamination of locally reproduced 

planthoppers. Therefore, the limited evaluation period of June 

to mid-July was selected. 

Migration source. If a migration cloud from each takeoff 

area is traced and its timing and area are compared with 

actual catch data, an accurate source estimation can be made. 

We have already conducted hourly catch investigations in 2003 

and 2004, which clearly showed sharp migration peaks of several 

hours. Using these data, we could estimate accurate migration 

sources of an order of a few hundred km. 

International cooperation. The current version of the 

prediction system can predict migrations for Korea, and can 

be used to analyze migrations within China. Establishing a 

prediction system for East Asia is one of the challenging next 

steps. 

References 

Grell G, Dudhia J, Stauffer D. 1994. A description of the fifth-generation 

Penn State/NCAR Mesoscale Model (MM5). NCAR 

Technical Note NCAR/TN-398+STR. Boulder, Colo. (USA): 

NCAR. 122 p. 

Ishikawa H, Chino M. 1991. Development of regionally extended/ 

worldwide version of system for prediction of environmental 

emergency dose information: WSPEEDI, (II) long-range transport 

model and its application to dispersion of Cesium-137 

from Chernobyl. J. Nucl. Sci. Technol. 28:642-655. 

Ohkubo N, Kisimoto R. 1971. Diurnal periodicity of flight behaviour 

of the brown planthopper, Nilaparvata lugens Stål, in the 4th 

and 5th emergence periods. Jpn. J. Appl. Entomol. Zool. 15:8- 

16. (In Japanese with English summary.) 

Ohkubo N. 1973. Experimental studies on the flight of planthoppers 

by the tethered flight technique. Jpn. J. Appl. Entomol. Zool. 

17:10-18. (In Japanese with English summary.) 

Riley JR, Cheng X-N, Zhang, X-X, Reynolds DR, Xu G-M, Smith 

AD, Cheng J-Y, Bao A-D, Zhai B-P. 1991. The long-distance 

migration of Nilaparvata lugens (Stål) (Delphacidae) in China: 

radar observations of mass return flight in the autumn. Ecol. 

Entomol. 16:471-489. 

Rosenberg LJ, Magor JI.. 1983. Flight duration of the brown 

planthopper, Nilaparvata lugens (Homoptera: Delphacidae). 

Ecol. Entomol. 8:341-350. 

Seino H, Shiotsuki Y, Oya S, Hirai Y. 1987. Prediction of long-distance 

migration of rice planthoppers to northern Kyushu considering 

low-level jet stream. J. Agric. Meteorol. 43:203-208. 

Turner R, Song Y-H, Uhm K-B. 1999. Numerical model simulations 

of brown planthopper Nilaparvata lugens and white-backed 

planthopper Sogatella furcifera (Hemiptera: Delphacidae) 

migration. Bull. Entomol. Res. 89:557-568. 

Watanabe T, Seino H, Kitamura C, Hirai Y. 1990. A computer program, 

LLJET, utilizing an 850 mb weather chart to forecast 

long-distance rice planthopper migration. Bull. Kyushu Nat. 

Agric. Exp. Stn. 26:233-260. (In Japanese with English summary.) 

Zhou B-H, Wang H-K, Cheng X-N. 1995. Forecasting systems for 

migrant pests. I. The brown planthopper Nilaparvata lugens 

in China. In: Drake VA, Gatehouse AG, editors. Insect migration. 

Cambridge (UK): Cambridge University Press. p 353- 

364. 

Notes 

Authors’ addresses: Akira Otuka, Tomonari Watanabe, Yoshito 

Suzuki, and Masaya Matsumura, National Agriculture and Biooriented 

Research Organization (NARO), 3-1-1, Kannondai, 

Tsukuba, Japan; Akiko Furuno and Masamichi Chino, Japan 

Atomic Energy Research Institute (JEARI), Tokai, Ibaraki, 

Japan. 


Distance education and eLearning 

for sustainable agriculture: lessons learned 

Robert T. Raab and Buenafe R. Abdon 

There is no doubt that the agricultural sector, no less than any 

other, is facing a range of old and new challenges as a result of 

today’s economic and environmental pressures. Key among 

these are population growth, increased market complexity, 

continuing economic inequality, and the need to raise productivity 

without adversely endangering the natural resource base 

(McCalla 2001). 

A growing global population means that agriculture will 

need to produce enough food to feed an expected two billion 

additional people by 2025 and this additional production must 

be achieved with less natural resources. Compounding these 

problems is the changing economic nature of agriculture, with 

increased commercialization, sophistication, and globalization. 

There is a growing consensus that learning will be a major 

factor to help agriculture and agricultural producers successfully 

deal with these challenges. “Knowledge—and related 

information, skills, technologies, and attitudes—will play a key 

role in the sustainable intensification of agriculture and success 

of rural development investments,” stated Alex et al 

(2002). 

Getting the essential knowledge to those who need it 

most remains difficult and expensive, but much optimism has 

been generated as a result of the increased growth and sophistication 

of new electronic information services—even in remote 

rural areas. Information and communication technologies 

(ICTs), and such specialized ICT applications as 

eLearning, are offering new options to deliver knowledge and 

information to farmers directly and indirectly through knowledge 

intermediaries. 

eLearning is one form of distance learning, a type of 

educational situation in which the instructor and students are 

separated by time, location, or both. eLearning typically involves 

the use of the Internet to access learning materials; interact 

with the content, instructor, and other learners; and obtain 

support during the learning process in order to acquire 

knowledge, construct personal meaning, and grow from the 

learning experience. 

Proponents make several convincing arguments about 

the power and potential of eLearning. eLearning provides learning 

opportunities in subjects not offered locally or where local 

offerings lack quality. It is ideally suited for individuals who 

lack time for classroom courses. Perhaps most importantly, 

participating in an online class gives students the skills required 

for lifelong learning. 

Although eLearning is still in its infancy, particularly in 

developing countries, some experience has been gained. Below, 

we will look briefly at where and how eLearning is being 

applied in agriculture and continue on to examine some of the 

lessons learned. 

Applications of digital technologies to agricultural learning 

eLearning and digital technologies are increasingly influencing 

and enriching all forms of agricultural education. This is 

most apparent in informal and formal education, but with tentative 

pilot applications in nonformal education as well. 

Traditional means of informal education and learning in 

agriculture, based on knowledge and skills being passed between 

generations and between community members, are becoming 

much less effective. This is a result of several factors, 

such as fewer experts in rural areas, agricultural innovations 

are increasingly coming from outside the community, fewer 

traditional courses are being offered, and much information is 

time-sensitive and/or needed quickly (Agriculture and Agri- 

Food Canada 2003). Comprehensive and growing repositories 

of online agricultural information provide a powerful means 

of overcoming these constraints and allow the independent 

learner to delve deeply into a myriad of subjects with the click 

of a mouse. 

Formal education has long been limited by geography, 

high cost, and lack of access by particular groups; as a result, 

eLearning and associated technologies are increasingly being 

used to overcome these obstacles. This trend is most prominent 

in higher education in developed countries and recent figures 

show that 80% of the top U.S. and European universities 

will offer global courses in 2004, with many of these offerings 

related to agriculture. 

While informal and formal education are certainly important 

for the future of agriculture, nonformal education is 

arguably the most critical. There is now almost too much information 

available online for informal learning and taking 

full advantage of these resources requires special skills to locate 

and evaluate. Agricultural knowledge acquired through 

formal education is soon outdated and obsolete. Properly conceived 

and developed, nonformal eLearning can substantially 

complement formal and informal efforts and provide up-todate 

and relevant agricultural knowledge. 

Lessons 

Although schools and other providers of education first began 

experimenting with online education only just over a decade 

ago, much has already been learned. Below, some key lessons 

are listed and discussed. 


eLearning works 

Recent studies have shown that some 50% of North American 

farmers take advantage of the informal educational opportunities 

available through the Internet through such services as e- 

mail (85%), searching information on agricultural products and 

services (78%), and news on agriculture (77%). Although these 

opportunities are not yet as prevalent in developing countries, 

evidence is clear that developing-country farmers and agricultural 

professionals are equally eager to learn in this way. The 

success of www.agriwatch.com and the ITC company’s 

eChoupal project (www.digitaldividend.org/case/ 

case_echoupal.htm) provides convincing evidence of this. 

The best eLearning courses are characterized by incorporating 

substantial interaction, are student-centered and 

constructivist, provide learner support, and use an integrated 

technology environment. Several studies have shown that learning 

outcomes associated with formal and nonformal eLearning 

are equal to the outcomes of traditional training courses given 

similar content and good instructional design. 

Experience also shows that, even in developing countries, 

properly designed and delivered online courses are effective 

and in demand. For example, over a period of three 

years, the agLe@rn program of the Asia Pacific Regional Technology 

Centre (APRTC) provided some 900 learning opportunities 

for agricultural professionals in 20 Asian and 17 African 

countries. A survey showed that alumni appreciated the 

knowledge they gained, were excited about their new abilities 

to network with peers around the world, and were better prepared 

to take advantage of digital learning resources. Perhaps 

most importantly, they were actively sharing their newly gained 

knowledge with students, colleagues, and farmers (Raab and 

Abdon 2003). 

Rate of application and adoption 

of eLearning approaches 

The majority of eLearning-related initiatives to date in both 

developed and developing countries have been in the area of 

online publishing of information as resources for informal learning. 

Almost all research and development organizations now 

routinely publish their findings and studies in electronic form 

and make these documents available through the Internet. 

Although aggregated data are not available on formal agriculture-related 

courses, it is clear that more and more universities 

and schools in developed countries are offering online 

learning opportunities. For example, the number of students 

taking online courses in the United States grew at more than 

25% from 1999 to 2002 and some sectors of higher education 

expect an annual growth rate exceeding 25% in online learners 

over the next few years (Allen and Seaman 2003). Developing 

countries lag seriously behind in this area as a result of high 

initial costs and lack of access to information, training, infrastructure, 

and resources (UNDP 2001). 

Nonformal educational opportunities for agriculturalists 

are virtually nonexistent in developing countries and, although 

some are available, they are not widely taken advantage of in 

developed countries. In developed countries, the main factors 

behind this slow adoption have been identified as a lack of 

opinion leaders who can provide expert, trusted advice about 

online learning and perceptions that 

eLearning has uncertain or unproven benefits 

Internet and computer access are not sufficient 

Online interactions are insecure and not confidential 

 

 

Advanced computer skills are required 

The cost is high in relation to benefits (Agriculture 

and Agri-Food Canada 2003) 

These concerns are no doubt shared by farmers and agricultural 

professionals in developing countries, who also have 

to deal with more serious cost, computer literacy, and connectivity 

limitations, in addition to language and literacy constraints. 

The most appropriate targets for eLearning 

in support of agriculture 

While no one questions that the priority target for knowledge 

development efforts must ultimately be agricultural producers, 

reaching this group remains problematic because of a range 

of cultural and technological constraints. It appears, however, 

that these constraints are much less serious among rural knowledge 

intermediaries, the many individuals employed by government 

extension systems, nongovernment organizations, 

academia, and the private sector, who have the responsibility 

to provide information and educational opportunities for farmers. 

It is also apparent that these individuals are effective in 

knowledge transfer and are at least as much in need of new 

knowledge and information as the clients they serve. 

Sustainability of eLearning efforts 

is a major problem 

Experience is showing that only a limited number of eLearningrelated 

initiatives are economically sustainable, with the exception 

of some informal learning resources. Farmers, in developed 

and developing countries, with sufficient financial 

resources are willing to pay for various kinds of market information 

if they feel it can be used to improve profits. Students 

(or their parents) will pay for online learning leading to a formal 

degree or certification if convinced it will result in better 

employment opportunities or a bigger paycheck. It is much 

more difficult to convince a learner to pay for a nonformal 

course for which the personal financial benefits are not clear 

or for which the major beneficiaries may well be others. 

Progress will depend on long-term, public-sector, 

and/or donor support 

The majority of eLearning initiatives, in both developed and 

developing countries, are now supported through government 

grants or research and development money and are generally 

not independently sustainable once the funding runs out. Given 

the current situation and the newness of this approach, it is 

probably not realistic to expect most initiatives to survive without 

continued public funding except in a very few informal 

learning niche markets. 


However, the continued involvement of the public sector 

may well be in society’s best interest. If learning is available 

only to the elite few who can afford it, there is considerable 

danger that the divide between the rich and poor will not 

only remain but grow. Winrock (2003) cautions that while, in 

general, reliance on the private sector is good, “information 

and access to it closely resemble a public good threatened with 

undersupply by market failures.” In cash-strapped developing 

countries, donor support as well will be critical. 

References 

Agriculture and Agri-Food Canada. 2003. An overview of e-Learning 

in Canadian agriculture and Agri-business. www.agr.gc.ca/ 

ren/serv/elearn_e.cfm. 

Alex G, Zijp W, Byerlee D, et al. 2002. Rural extension and advisory 

services: new directions. Rural Development Strategy Background 

Paper #9. Washington, D.C. (USA): Agriculture and 

Rural Development Department, World Bank. 

www.worldbank.org/wbi/sdruralpoverty/ag_extension1/Materials/additional/Rural_extension.pdf. 

Allen IE, Seaman J. 2003. Sizing the opportunity: the quality and 

extent of online education in the United States, 2002 and 2003. 

Sloan-C, The Sloan Consortium. www.sloan-c.org/resources/ 

sizing_opportunity.pdf. 

McCalla AF. 2001. Challenges to world agriculture in the 21st century. 

Agricultural and Resource Economics Update. University 

of California, Davis. www.agecon.ucdavis.edu/outreach/ 

areupdatepdfs/UpdateV4N3/spring2001.pdf. 

Raab R, Abdon BR. 2003. Assessment and use of agLe@rn Knowledge 

and course materials. Bangkok (Thailand): APRTC. 

www.aprtc.org/for_alumni/multiplier.htm. 

UNDP (United Nations Development Programme). 2001. Human 

development report 2001: making new technologies work for 

human development. Published by Oxford University Press, 

Inc., New York City. Retrieved 1 October 2001 from 

www.undp.org/hdr2001/. 

Winrock. 2003. Future directions in agriculture and information and 

communication technologies (ICTs) at USAID. Prepared for 

USAID/Economic Growth, Agriculture, and Trade/Agriculture 

and Food Security. Retrieved 30 August 2003 from 

www.dot-com-alliance.org/documents/AG_ICT_USAID.pdf. 

Notes 

Authors’ address: Sustainable Development eLearning Network, e- 

mail: rraab@sdlearn.net. 

The Rice Knowledge Bank 

Mark Bell and David Shires 

The fast and effective transfer of research findings to farmers 

has always been one of the biggest challenges facing those in 

agricultural development. All too often, new knowledge is successfully 

developed and validated, only to fail in reaching those 

who need it most—the farmers. Into this gap between research 

and impact has stepped the International Rice Research 

Institute’s (IRRI) Rice Knowledge Bank (RKB). Not only is it 

one of the world’s first digital extension services for those who 

provide information and support for farmers, it is also the first 

comprehensive, digital rice-production library containing a 

wealth of information for rice-related training and extension. 

More importantly, it provides this service using a format that 

sets a new standard for knowledge access within the agricultural 

development community. Containing the most up-to-date 

and validated knowledge, the RKB is providing government 

extension services, nongovernment organizations (NGOs), and 

universities with unprecedented access to training and support 

knowledge. 

First and foremost, IRRI is a rice science research institute. 

However, IRRI’s founders knew that, to get research findings 

out of the laboratory and into farmers’ fields, there had to 

be an education component. Over the past 40 years, IRRI has 

used a variety of instructional methodologies and technologies 

to satisfy its education mandate. Printed materials, photo- 

graphic slide–audio tape modules, instructional video, early 

attempts at computer-aided instruction, video-conferencing, 

and information presented on CD-ROM have all supported 

face-to-face classroom instruction both at IRRI headquarters 

and in country training programs. Now, IRRI is developing 

and adding another set of tools to better serve its education 

mandate: those that are collectively known as information and 

communication technologies (ICT), which include distancelearning 

methodologies and use of the Internet. 

IRRI has moved to harness the power of ICT by carefully 

matching the new media’s capabilities to the needs of 

rice-related training to bring relevant knowledge, in the most 

useful form, to field officers when and where they need that 

knowledge. The RKB has received critical acclaim (for example, 

BBC Earth Report, September 2004—the “Further readings” 

section provides more reviews of the RKB) as a tool to 

distill, store, and provide access to the vast array of IRRI’s 

training and support knowledge for rice science and extension. 

Rationale for the RKB 

Ask any farmers what they perceive as their major needs and 

you will probably be told three things: access to credit, a good 


and reliable price for their harvest, and relevant, up-to-date 

knowledge. The knowledge that farmers require covers all aspects 

of their farming, from the efficient management of farms, 

through improved production technologies, to the best 

postharvest practices. Knowledge of markets and marketing is 

also usually included in their list of knowledge needs. IRRI, 

through its extensive 40-year research agenda and associated 

work in social science and extension, has accumulated much 

of the knowledge that will meet these farmer needs. 

However, IRRI recognized early that an ICT solution 

was not going to be successful in achieving impact from its 

research findings and improving the livelihoods of farmers if 

the approach was only to collate a large amount of knowledge 

and make it available through the Internet. To be successful, 

other, more important factors had to be included in the design 

of an ICT-based system. 

The first question IRRI tackled was, “Who would make 

up the primary audience for the IRRI Internet presence and its 

array of rice-related knowledge” Clearly, to achieve the goal 

of obtaining impact and improving livelihoods from IRRI research, 

farmers had to be the ultimate beneficiaries of the 

project, but they were probably not going to be the primary 

users of the RKB. A lack of ready access to ICT infrastructure, 

literacy and language capability, and information skills suggested 

that, at least initially, a direct farmer target audience 

was not appropriate. How then to improve the flow of information 

to farmers IRRI’s work across the region suggested 

that increased capacity of existing farmer intermediaries to 

provide better, more timely knowledge in a more usable form 

was necessary. Thus, IRRI decided that the primary audience 

for the RKB would be intermediaries who work directly with 

farmers—intermediaries within government extension services, 

NGOs, and universities. 

The second issue IRRI examined was, “Which information 

should be included in the RKB” After 40 years of intensive 

research covering all aspects of production, 

postproduction, and marketing across the region, there was no 

shortage of readily available knowledge presenting itself for 

inclusion. Investigations within the target audience suggested 

that a huge bank of information was not the solution but rather 

a careful distillation of the most relevant knowledge would be 

more useful. 

After these decisions were made, IRRI designed a Web 

presence, the RKB, whose primary audience is farmer intermediaries. 

The RKB contains content that meets the audience’s 

needs in improving the delivery of extension services to the 

farming community, and makes relevant, up-to-date knowledge 

available in a form appropriate to its intended use by farmers. 

Assembling the RKB 

The RKB was assembled from the knowledge base of IRRI by 

selecting those knowledge parts that were applicable to the 

intended audience, validating the selected knowledge, and assembling 

it into a dynamic, easy-to-use form. All the knowledge 

was assembled such that it was readily available to the 

target audience in the form it required—online, on CD-ROM, 

or in print. The RKB was structured as a single-source publishing 

resource. 

Thus, the RKB was based on three fundamental characteristics 

that have proved to be vital in making ICT effective 

in training and extension activity: 

the RKB has a definite focus: a clear audience and 

purpose; 

all of its content has the specific characteristics of 

being focused, credible, value-added, and demanddriven; 

and 

single-source publishing provides the form of output 

best suited to users. 

Characteristics of RKB content 

To ensure that the knowledge in the RKB is best suited to the 

audience, four characteristics were established and are applied 

to all content in the RKB: focused, credible, value-added, and 

demand-driven. 

1. Focused. A great temptation when using the immense 

power of ICT to store, search, and deliver information 

is to include as much information as possible— 

the principle that “more is better.” IRRI, of course, is 

well equipped to pursue this strategy because of its 

accumulated wealth of research and farming knowledge 

accumulated over 40 years of activity across the 

region. Therefore, much effort is needed to ensure 

that the RKB remains focused, containing only the 

knowledge that is relevant to extension and research 

training and support. Much other valuable knowledge 

is omitted and is published through other channels by 

IRRI. In this way, the target audience is able to quickly 

identify knowledge that is directly relevant to it without 

having to search through large amounts of what 

is, to it, irrelevant knowledge. 

2. Credible. Because most members of the target audience 

are not trained in many aspects of rice production 

and processing, it is not easy for them to judge 

the accuracy or relevance of knowledge. Therefore, a 

basic characteristic of the RKB has to be that all of 

the content is credible. IRRI instituted and maintains 

a rigorous quality-control process that ensures that 

all content is accurate, up-to-date, and validated. In 

addition, the quality-control process also ensures that 

all content possesses these RKB characteristics. The 

quality-control process involves review of and signoff 

on content by the relevant IRRI scientists and review 

by RKB staff. With these mechanisms in place, 

users can approach the RKB with confidence in the 

quality and relevance of the knowledge they locate. 

3. Value-added. The knowledge captured in the RKB is 

not merely presented in its original scientific form 

but is enhanced by presentation in differing forms that 

make the knowledge more accessible and more directly 

usable. 


Reference guides contain the field-related information 

required by extension officers and scientists. 

They are concise, easy-to-read manuals, not 

scientific papers. They are all presented in an attractive, 

easy-to-use electronic book format. The 

information in the RKB is not stored as Word 

files or PDF files, but in a fully-indexed book 

form for users’ convenience. 

Fact sheets are the distilled essence of topics that, 

on one page, provide all of the vital points about 

that topic. They summarize the essence of information 

in the reference manuals. 

Decision support tools are computer programs 

that lead users easily through a process to solve 

their particular problem. By questioning users, 

the decision support tools lead users to the knowledge, 

reference manual sections, or fact sheets 

that cover their problem. 

eLearning courses have been created to capture 

and present some key training in a form that users 

can access when and where they want and at 

a time of their own choosing. These courses can 

be supported by remote tutoring and electronic 

discussion groups. 

4. Demand-driven. The RKB ensures that its knowledge 

base and the access mechanisms are driven by user 

needs. It ensures this through IRRI’s extensive incountry 

networks that continually identify country 

needs. In addition, the RKB itself maintains extensive 

statistics of usage that help refine the in-country 

needs analyses gathered through traditional means. 

One particular demand-driven innovation is the 

inclusion of country sites, accessible from the home 

page by clicking on the country’s flag. These sites 

contain the best, most relevant local knowledge that 

complements the IRRI knowledge. Much of this 

knowledge is in the local language and is particularly 

directed to local needs and conditions. 

Future of the RKB 

The RKB will continue to develop over the coming years in 

two areas: further content development and the RKB’s application 

in countries across the region. The aim of all future development 

will be to continue to ensure that the focus of the 

RKB remains strong, its content is always up-to-date and accurate, 

and its effective use increases in countries. 

Four activities are planned in the area of content development: 

1. Knowledge will be related to field application. More 

emphasis will be placed on ensuring that the content 

is, as much as possible, linked to field practice so 

that its transfer to farmers is as easy as possible. 

2. Up-to-date content will be maintained through links 

to IRRI’s research agenda and major development 

projects. To ensure that the content is the best and 

most up-to-date available, more emphasis will be 

placed on the use of the RKB by IRRI scientists as a 

tool to increase the impact of their research. Major 

IRRI development projects such as the Irrigated Rice 

Research Consortium (IRRC) and the Consortium for 

Unfavorable Rice Environments (CURE) will be involved 

in both the dissemination of RKB knowledge 

and contributions to the RKB. 

3. Content of the RKB will be supplemented by in-country 

knowledge. To enhance the scope and relevance 

of the content, closer links with country partner 

projects will be encouraged, resulting in more goodquality 

local knowledge being captured in the RKB. 

4. An effective feedback loop to research agendas will 

make the RKB more relevant to the overall development 

of rice-related activity. Links with IRRI’s and 

national research agendas will be strengthened, ensuring 

the continued health of the knowledge base 

that is the RKB. 

To improve the scope and quality of application of the 

RKB in countries across the region, three areas of activity will 

be pursued: 

1. Country use through the identification, training, and 

support of committed partners will be expanded. 

2. The quality and scope of country sites, and the country 

information behind the country flags on the home 

page, will be expanded through the capture of the best 

of the local knowledge available. 

3. The expertise to develop independent knowledge 

banks will be transferred to national systems and they 

will be supported through their development process. 

A vision for the RKB 

The vision that IRRI has for the RKB is much the same now as 

it was when it was first conceived. The RKB has always aimed 

to provide needed, accurate information in the most appropriate 

form for the target audience. The current vision states, 

“IRRI’s Rice Knowledge Bank is the world’s leading repository 

of rice-related extension & research training and support 

materials, being effectively used by intermediaries and farmers 

to improve livelihoods. The Rice Knowledge Bank is the 

primary focal point for networking and sharing information 

between rice communities.” 

Further readings about the RKB 

Atkinson AD, Bell MA. 2003. Organized free for all. Rice Today 

2(1):22-25. 

Atkinson AD, Bell MA. 2003. The Rice Knowledge Bank: strengthening 

capacity. Int. Rice Res. Notes 27(2):5-10. 


Amerasinghe N. 2003. Rice Knowledge Bank: a review. ADB Institute. 

www.adbi.org/articles/29.Rice.Knowledge.Bank/ 

default.php. 

Bell MA, Fredenberg P, Atkinson A, Shires D. 2004. How rice farmers 

benefit from ICT. New Agriculturalist on-line. www.newagri.co.uk/04-4/focuson/focuson3.html. 

Noronha F. 2003. Rice could get a bounty crop, thanks to the Net. 

Express Computer, India’s No. 1 IT Business Weekly. 

www.expresscomputeronline.com/20030106/indcomp2.shtml. 

Notes 

Authors’ address: International Rice Research Institute, DAPO Box 

7777, Metro Manila, Philippines, e-mail: m.bell@cgiar.org, 

d.shires@cgiar.org. 

Integrated information systems 

for crop research and improvement 

Graham McLaren, Arllet Portugal, Alexander Cosico, William Eusebio, Teri Ulat, May Ann Sallan, Victor Jun Ulat, and Richard Bruskiewich 

International germplasm exchange was the engine of the Green 

Revolution. In the past, however, much of the important information 

generated from this exchange was accessible only locally, 

in field books or researchers’ files. Although major international 

initiatives for germplasm collection and conservation 

followed the Green Revolution, much collected material 

is still not used because it is difficult to access. As a result, the 

potential impact on agriculture has not yet been realized. However, 

the free exchange of information, through international 

crop information systems, should provide the foundation for a 

second revolution that adds value to germplasm by seamlessly 

uniting its conservation, evaluation, use, and exchange. 

Furthermore, new technologies in molecular biology and 

genomics mean that traditional phenotypic information must 

be linked to large quantities of sequence and genetic information 

so that functional genomics and allele-mining activities 

can speed up germplasm enhancement. 

The International Maize and Wheat Improvement Center 

(CIMMYT) devised an information strategy and developed 

software on a main-frame computer during the 1980s to facilitate 

unambiguous identification of wheat germplasm, thereby 

establishing links among information from different sources 

(Fox et al 1996a). In 1995, CIMMYT and IRRI canvassed 

other CGIAR centers to establish a project to develop an International 

Crop Information System applicable to a wide range 

of crops. 

The International Crop Information System (ICIS) 

Extensive communication among CGIAR centers highlighted 

the economies to be gained by collaborating on the development 

of an information system that could be used for many 

crops. At a meeting held at CIMMYT in September 1994, a 

prototype CD-ROM was demonstrated and Open Database 

Connectivity (ODBC) was chosen as a programming standard 

for international collaboration. This ensures that the choice of 

database system for a particular crop is independent of the 

decision to use ICIS. 

Several CGIAR centers, national agricultural research 

and extension systems, and advanced laboratories are collaborating 

to develop ICIS as a generic system that will accommodate 

all data sources for any crop. The vision of ICIS is to 

integrate different data types into a single information system 

and provide specialist views and applications that operate on 

a single integrated data platform. ICIS must also seamlessly 

integrate private local data with public central information to 

give local researchers access to global crop information. After 

all phases of development are complete, ICIS will seamlessly 

support a range of activities including germplasm conservation, 

functional genomics, allele mining, breeding, cultivar testing, 

and release. Data will be accessible from CD-ROM or the 

Internet, and users could either adopt the complete system or 

individual components complementary to their own systems. 

The driving force behind ICIS is accessing and sharing data 

rather than providing analytical and statistical tools. This is 

because the major bottleneck to intelligent data integration and 

use was not statistical software, but rather the drudgery of finding, 

extracting, preparing, and managing the data. 

Functionality 

The ICIS system is fast, user-friendly, PC-based, and distributable 

on CD-ROM or via the Internet. It contains 

A genealogy management component to capture and 

process historical genealogies as well as to maintain 

evolving pedigrees, and to provide the basis for unique 

identification and internationally accepted nomenclature 

conventions for each crop. 

A data management system for genetic, phenotypic, 

and environmental data generated by evaluation and 

testing, as well as for providing links to genomic maps. 

Links to geographic information systems that can manipulate 

all data associated with latitude and longitude 

(e.g., international, regional, and national testing 

programs). 


Applications for maintaining, updating, and correcting 

genealogy records and tracking changes and updates. 

Applications for producing field books and managing 

sets of breeding material, and for diagnostics such 

as COPs (coefficients of parentage) and genetic profiles 

for planning crosses. 

Tools to add new breeding methods, new data fields, 

and new traits. 

Tools for submitting data to crop curators and for distributing 

data updates via CD-ROM and electronic 

networking. 

Remote users 

One of the innovative features of ICIS is that it permits independent 

users to integrate their own local data with public central 

data. ICIS does this by allowing read-only access to the 

central database for a particular crop and supporting a local 

copy of the central data model where the local data are stored. 

Apart from providing user-friendly access to the data, so that 

crop scientists can make informed decisions, the system provides 

a local data management system for users and captures 

relevant data for the crop. Periodic updates to the central database 

by users make their data available to all other users as 

well as browsers of the central database. 

The data model 

The data model and database system of ICIS are designed for 

maximum flexibility to cater to as wide a range of crops as 

possible. The model must be independently adopted for a specific 

crop and data entered to create an independent system 

for that crop. 

The Genealogy Management System. The core of ICIS 

is a common genealogical data model, called the GMS, which 

must be sufficiently well designed and universal to accommodate 

a wide range of crops. The functions of the GMS are to 

(1) assign and maintain unique germplasm identification, (2) 

retain and manage information on genealogy, and (3) manage 

the nomenclature and chronology of germplasm development. 

Each germplasm entity is identified by a controlled 

CropID (the domain), a UserID (the authority), and a locally 

assigned GERMPLASM_ID (GID). The logical connection between 

a GID and a packet of seeds or other propagating material 

is that different packets that germplasm specialists would 

not like to mix receive different GIDs. Figure 1 shows information 

on method, location, date of genesis, and other attributes 

managed through the data model. Each germplasm record is 

linked to its progenitors through their GIDs. Germplasm is 

divided into two categories, generative and derivative. Generative 

germplasm is produced by generative methods, which 

tend to increase and combine genetic variability such as crosses 

or mutations. Derivative germplasm is produced by derivative 

methods, which tend to refine and target genetic variability, 

such as selection methods, or which aim to maintain genetic 

status through management methods such as seed increase or 

conservation. Germplasm produced by generative methods may 

have any number of progenitors. Derivative germplasm is derived 

from a single germplasm source. 

Germplasm collects a multitude of labels during the development 

and release process. These are all tracked as NAMES 

in the GMS. One name must be identified as the preferred name. 

NAMES may contain imbedded information, and this can be 

made accessible to application programs for specific name 

types by specifying a format for the name. 

Attributes are text variables used to store information 

about the genesis, genealogy, nomenclature, or chronology of 

germplasm. Attribute types are defined and described as 

USER_DEFINED_FIELDS. Like names, attributes may contain 

imbedded information in the form of subfields or variables 

within the attribute text. 

The Data Management System. The functions of the 

DMS are to (1) store and manage documented and structured 

data from a genetic resource, varietal evaluation, and crop 

improvement studies; (2) link data to specialized data sources 

such as GMS, and soil and climate databases; and (3) facilitate 

inquiries, searches, and data extraction across studies according 

to structured criteria for data selection. 

All types of data can be accommodated in DMS, including 

raw data, observed data, derived data, and summary statistics. 

Data may have continuous or discrete numeric values, or 

text or categorical character values. For example, observations 

on disease resistance or nutrient efficiency of a genotype can 

be numerical measurements, scored or calculated indices, or 

text data. More complex forms of data, such as pictures or 

documents, will also be considered. Figure 2 shows the basic 

data model of the DMS, and shows the linkages between the 

entities described below. 

A STUDY is the basic, reportable unit of research; it is 

synonymous with the notions of experiment, nursery, or survey. 

Since the DMS must deal with any of these, we use the 

term “STUDY.” A STUDY is characterized by a set of scientific 

objectives and testable hypotheses and results in the collection 

of one or more data sets. FACTORS are classifying variables 

in a STUDY, which take values from finite sets of discrete 

LEVELS. Factors are named and described in each STUDY. They 

have three main attributes: the PROPERTY of the experimental 

material or survey units being manipulated or stratified, the 

METHOD or procedure by which the levels are applied, and the 

SCALE or measurement units in which the levels are expressed. 

All levels of a particular factor are expressed in the same scale. 

The names of factors are consistent within studies and equivalent 

factors are linked across studies through common 

PROPERTYs. PROPERTY is subject to a controlled vocabulary to 

facilitate this linkage. 

Data sources such as field objects, sampling units, or 

data classes such as treatment are identified by combinations 

of levels of design or sampling factors. These are referred to 

as OBSERVATION UNITS, which may be grouped into subsets or 

EFFECTS. DATUM values are recorded for one or more VARI- 

ATES. Each VARIATE has the same three attributes as a FACTOR, 

the PROPERTY or trait being measured, the METHOD or proce- 


Germplasm 

Generative germplasm 

Has progenitors 

Derivative germplasm 

(Recursive links) 

Has group 

Has source 

Developed by 

Methods 

Called 

Names 

Developed at 

Locations 

Named at 

Named by 

Developed by 

Users 

Assigned at 

With value of 

Attributes 

Assigned by 

Defined property 

User-defined fields 

Fig. 1. Data model for the ICIS Genealogy Management System. 

dure by which the value is observed or derived, and the SCALE 

or measurement units in which the value is expressed. 

Correcting data 

Corrections and changes will inevitably be made in any database. 

Only authorized users can make these changes, and all 

changes are logged so that the sequence of changes can be 

traced and can be “undone” if required. 

One common occurrence of changes is when new information 

about an existing germplasm record is entered into a 

local database. If the existing record is also in the local database, 

then the local user can complete the changes. But, if it is 

in the central database, changes cannot be completed until the 

central database is updated. Verifying and completing requested 

changes are part of the update process and sufficient information 

and justification need to be recorded in the changes table 

to allow the process to be completed. 

Verification and completion of changes in the central 

database may take some time but local users would like to see 

their changes reflected immediately. This is achieved by the 

DLL, which always checks the local CHANGES table for central 

changes and applies them at run time for the specific installation 

where they are recorded. 

Stand-alone software modules 

Components of ICIS include a Genealogy Management System 

(GMS); Set Generation Module (SETGEN), including the 

External Pedigree Input Tool (EPIT); Field Book Module 


Study 

Applied by 

Property management 

module 

Observed by 

Factor 

Application of 

Expressed in 

Applied by 

Property 

Measured of 

Expressed by 

Variate 

Scale 

Value of 

Indexed by 

Effect 

Method 

Measured by 

Vector 

Element 

of 

Measurement 

of 

Representation of 

Level 

Indexed by 

Observation 

unit 

Data 

Fig. 2. Data model for the ICIS Data Management System. 

(FLDBK); Trait Management System (TMS); Data Management 

System (DMS); a Work Book (WRKBK) for data input 

and query; and the Data Retriever for cross-study data queries 

(RTV). The first four modules focus on germplasm and management 

of genealogy and nomenclature. The next three handle 

the management of evaluation and characterization data, and 

the Retriever provides access to both raw and processed data. 

User-defined data 

Users will be able to define new relationships among 

germplasm by specifying new breeding methods. They can 

specify attributes of the germplasm to be stored, types of names, 

location descriptors, and traits and variates for characterization 

and evaluation data. Data with any form of factorial structure 

can be managed. 

Links to global bioinformatics resources 

Other databases of agricultural information have close links 

with ICIS. The System-wide Information Network for Genetic 

Resources (SINGER) aims to provide global access to genetic 

resources data across all CGIAR-mandated crops and commodities. 

Germplasm records in ICIS that relate to accessions 

in the CGIAR collections are linked with SINGER so that the 

ICIS can provide information on the use and deployment of 

those genetic resources. 

Each individual ICIS database manages data for a particular 

crop and generally does not share data with other crops. 

This integrating role of genetic resources information across 

crops is played by SINGER (Fig. 3). 

ICIS has also enjoyed collaboration with the USDAsponsored 

GrainGenes and Gramene database initiatives. Linking 

these implementations to ICIS will provide access to sequence 

information and molecular maps, which will facilitate 

functional genomics and allele mining and integrate information 

across crops at the genomic end of the spectrum. 

The International Rice Information System (IRIS) 

IRIS is the rice implementation of ICIS. The GMS of IRIS 

stores information on about one and a half million varieties, 

breeding lines, and accessions of rice. This allows pedigree 

analysis to trace germplasm flows and relationships between 

lines that can be used to improve evaluation estimates or plan 

improvement programs. The DMS of IRIS contains 5 million 

data values from over 500 studies from breeding, screening, 

and international testing trials. This allows integrative analysis 

over different environments. These data are available 

through a web interface and are linked to rice data sources 

throughout the world (Bruskiewich et al 2003). 

Conclusions 

The technical challenge for plant scientists and software developers 

is to implement the type of system outlined here. The 

challenge for administrators is perhaps more difficult—to facilitate 

the continued free exchange of information and to nurture 

a scientific culture in which users take full advantage of 

the data of others and, in turn, contribute to shared databases. 


SINGER 

Access to basic genetic resources data: 

accessions, cooperators, standards, characterization, transfer, collecting 

Data sharing across crops 

Central 

databases 

User 

databases 

Wheat Rice Sorghum Others 

Data types: 

l Genetic resources 

data management 

l Characterization 

l Evaluation 

l Genealogy 

l Nomenclature 

l Unique identification 

of germplasm 

ICIS 

data model and information system 

Fig. 3. Relationship between SINGER and ICIS. 

References 

Bruskiewich RM, Cosico AB, Eusebio W, Portugal AM, Ramos LM, 

Reyes T, Sallan MA, Ulat VJ, Wang X, McNalley K, Sackville- 

Hamilton R, McLaren CG. 2003. Linking genotype to phenotype: 

the International Rice Information System. 

Bioinformatics 19(Supp. 1):i63-i65. 

Fox PN, Lopez C, Skovmand B, Sanchez H, Herrera R, White JW, 

Duveiller E, van Ginkel M. 1996. International Wheat Information 

System (IWIS), Version 1. Mexico, D.F.: Centro 

Internacional de Mejoramiento de Maíz y Trigo. On compact 

disk. 

Fox PN, Skovmand B. 1996. The International Crop Information 

System (ICIS)—connects genebank to breeder to farmer’s 

field. In: Cooper M, Hammer GL, editors. Plant adaptation 

and crop improvement. Wallingford (UK): CAB International. 

Notes 

Authors’ address: International Rice Research Institute (IRRI), 

DAPO Box 7777, Metro Manila, Philippines. 

Using APAN for content delivery: possibilities for the CGIAR 

Paul O’Nolan 

In the mid-1990s, the Consultative Group on International 

Agricultural Research (CGIAR) positioned itself well to take 

advantage of new opportunities provided by information technology, 

with projects such as the Integrated Voice and Data 

Network (IVDN) and the System-Wide Information Network 

for Genetic Resources (SINGER)—an application that de- 

pended on new levels of connectivity. The IVDN brought the 

first digital Internet connections to several of the countries in 

which the CGIAR operates. Since then, the world has changed 

a great deal, as evidenced by disconnections from the IVDN 

and connections to local Internet service providers that didn’t 

exist seven or eight years ago. Unfortunately, the CGIAR is no 


longer at the forefront in bringing the fruits of new information 

and communication technologies to bear on scientific and 

development problems. To date, only a single CGIAR center 

is connected to second-generation Internet networks. As a system, 

the CGIAR has fallen years behind where it should be. 

Researchers in many national agricultural research and extension 

systems in Asia and elsewhere now enjoy better connectivity 

to advanced research networks than CGIAR scientists, 

so now is a good time for the CGIAR to do some catching up. 

Recently, the Information and Communication Technology 

and Knowledge Management (ICTKM) program of the 

CGIAR has, with funding from the World Bank, committed 

funds (approx. US$100,000) to kick-start advanced research 

networking (ARN) in the CGIAR. This paper presents some 

ways in which advanced research network connectivity will 

have an impact on research and education in the CGIAR. 

Background 

The CGIAR first joined the Asia Pacific Advanced Network 

(APAN; www.apan.net) when IRRI joined on behalf of the 

system in 1999 under the leadership of the then acting head of 

the IRRI Training Center, Dr. Robert Raab. Using a single PC 

with a nondedicated 128 kbps APAN connection, videoconferencing 

technology was used to have IRRI scientists give 

seminars at the Ministry of Agriculture in Bangkok and attend 

a seminar at IRRI on bioinformatics presented from Singapore. 

Under Raab’s leadership, IRRI funded APAN connectivity for 

the Thai national agricultural research system (NARS) for two 

years, after which the NARS took it over, retained the connection, 

and began to use a national network to disseminate adapted 

and translated versions of IRRI-originated training materials. 

In subsequent years, IRRI’s APAN connection improved, 

eventually becoming a dedicated 2 Mbps link; IRRI went on 

to create the Rice Knowledge Bank 

(www.knowledgebank.irri.org), an electronic, Internet-accessible 

digital repository of knowledge about rice; the first national 

“mirror” copy of the IRRI Web site was created in 

Beijing, hosted by the Institute of Microbiology (at 

www.irri.cn). More recently, in 2004, IRRI signed a memorandum 

with the Agriculture, Forestry, and Fisheries Research 

Council of the Japanese Ministry of Agriculture, Forestry, and 

Fisheries (MAFF/AFFRC) agreeing that IRRI and MAFF/ 

AFFRC shall jointly develop, distribute, and promote mirroring 

systems for improving access to and sharing of Internetbased 

information in Japan and in other countries in the Asian 

region—using the APAN network, the hub of which is in 

Tsukuba, Japan. 

A copy of the International Rice Information System 

(IRIS), a version of the International Crop Information System 

(ICIS) that IRRI is developing in cooperation with partner 

organizations from several countries (home page: http:// 

iris.irri.cgiar.org), is now hosted online at MAFF’s Information 

Network (MAFFIN) headquarters in Tsukuba. Data are 

replicated between MAFFIN and IRRI using APAN and it is 

planned that the MAFFIN network operating center (NOC) 

will be the hub for future data replication to networking centers 

in other Asian countries connected to APAN. Mirrored data 

will also be available via the Internet. IRRI will provide content 

and MAFFIN will provide network resources, including 

storage and bandwidth to this end. 

In principle, a vehicle now exists for sharing a wide variety 

of different kinds of information relevant to agricultural 

research and development among APAN-connected organizations 

in the region. Sharing can be done on a bilateral basis 

between pairs of nodes or via the NOC hub on a one-to-many 

basis. Shareable content could originate in the region, or externally, 

and be made available via partners, such as CGIAR 

institutes. 

There are many reasons to mirror data. In the case of the 

China mirror of IRRI’s Web site, the copy helps students and 

scientists avoid paying for international communications 

charges using a commercial Internet service provider. It has 

been IRRI’s experience in India, Thailand, and elsewhere that 

content stored locally is more likely to be translated and 

adapted. When used with suitable file-sharing software, mirror 

sites may provide faster and more reliable access to large 

data sets by distributing the burden of data transfers. In addition, 

mirror sites, which need not be public, can help in disaster 

mitigation. 

Even with the rapid decline in the cost of bandwidth in 

recent years and its growing availability, it is clear that mirroring 

will remain a useful technology for a long time, and that 

much of the utility will derive from added-value features such 

as data annotation, search capabilities, use of standard formats, 

etc. Even simply sharing metadata about what data are potentially 

available could be very valuable. It should be possible to 

co-opt a peer-to-peer file-sharing tool of the kind typically used 

to share music (often illegally) to share metadata among communities. 

Today, a great deal of valuable scientific information 

is hidden away inside organizations and even its existence 

may be hard to discover. 

APAN and the CGIAR today 

Meanwhile, Internet technology has marched on. Most, but not 

all, of the CGIAR institutes now have local rather than international 

Internet connectivity as they did five years ago. One 

of the CGIAR’s ICTKM projects is a so-called “2nd-Level 

Connectivity” enhancement project, designed to improve connectivity 

at nonheadquarters locations, such as country offices 

and regional offices, with a special focus on Africa. 

However, IRRI is still the only center connected to an 

advanced research network (APAN); we hope that it will be 

joined in this by several sister CGIAR institutes in 2005 (the 

International Potato Center, CIP, in Peru; the International 

Center for Tropical Agriculture, CIAT, in Colombia; and International 

Maize and Wheat Improvement Center, CIMMYT, in 

Mexico are currently planning to connect to Internet2 in 2005, 

with funding from the CGIAR ICTKM advanced research networks 

project, led by IRRI). Today, every networked computer 

in IRRI has a broadband connection to APAN. The APAN link 


Before 

Los Baños 

PHNET, Makati 


2 Mbps 

(E1) 

1.5 Mbps 

APAN 

Commercial 

ISP 

Internet 

Fig. 1. IRRI’s APAN connectivity in 2003: a shared local loop from IRRI to a co-located 

router at the premises of the Philippine Network Foundation (www.ph.net) in Manila for 

a 1.5 Mpbs APAN link and a 1 Mbps commercial Internet connection (one of two in 

operation). ISP = Internet service provider. 

After 

Los Baños 

U.P. 

Los Baños 

PHNET, Makati 

U.P. 

Open 

University 


2 Mbps 

(E1) 


IPB 


6 Mbps 

(3 × E1) 

2 Mbps 

(E1) 

Ai3 

APAN 

ASTI, Quezon City 

PREGINET 

Fig. 2. IRRI’s APAN connectivity in 2004: the 6 Mbps local loop to the Advanced Science 

and Technology Institute (ASTI) at the University of the Philippines Diliman 

(www.upd.edu.ph) will be operational in 2005. Here, a CGIAR institute serves as a bridge 

between both national and international partners. 

to the Philippines increased from 1.5 Mb to 6 Mbps in February 

2004. In the Philippines, the Department of Science and 

Technology has sponsored the creation of a national network, 

PREGINET, the Philippine Research, Education, and Government 

Information Network (http://preginet.asti.dost.gov.ph). 

This network now connects all 19 regions of the Philippines 

with broadband links and more than 70 institutions are connected 

to it. Video-conferencing between nodes and broadcasting 

of meetings using video-streaming technologies are routine. 

Recent changes in IRRI’s APAN connectivity are summarized 

in Figures 1 and 2. 

With the change in the national termination point of 

APAN from PHNET to ASTI, as shown in Figure 2, IRRI opted 

to install a faster local loop that would enable better exploitation 

of the extra APAN bandwidth. The original circuit was 

retained as a backup and made available to the local scientific 

community. Locally, the University of the Philippines Los 

Baños (UPLB), the University of the Philippines Open University 

(UPOU), and the Institute of Plant Breeding (IPB) were 

interconnected with fiber optic cable that converged in an IRRI 

building. They were not connected to APAN, nor, more importantly, 

to PREGINET. 

Apart from local loop communication costs and infrastructure 

costs, advanced research network bandwidth is free 

to the end user. When CGIAR scientists are connected to 

Internet2 or to APAN or similar networks, including connected 

national networks in a growing number of countries, they can 

make use, at no marginal cost, of video-conferencing, videostreaming, 

and other communication-intensive technologies 

(e.g., for large data transfers, and for hosting and synchronizing 

mirror sites of genomic, climatic, and other data). 


Fig. 3. July 2004. A planned Access Grid node begins to take shape at IRRI. The 

facility has been used by IRRI staff for videoconferences with scientists on PREGINET. 

IRRI has begun construction of an Access Grid node 

(www.accessgrid.org). Once additional CGIAR centers connect 

to advanced research networks, the CGIAR network can 

play a bridging role in facilitating face-to-face interactions 

between national partners at some locations and external partners. 

With the Access Grid’s suitability for informal group 

meetings, we hope that the technology “catches on” and is used 

for informal meetings by communities of interest. 

In 2004, four CGIAR centers acquired grid cluster computer 

systems for bioinformatics and genomics research. It is 

planned that these resources will be shareable on the CGIARwide 

area network and improvements in ARN connectivity to 

address this should have spillover benefits for bandwidth for 

content sharing and delivery. 

APAN and the CGIAR tomorrow 

Despite the explosive growth in the number of cell phones in 

the region in recent years, access to communication technology 

is out of reach for many Asians, in particular the rural 

poor. In Bangladesh, for example, it is estimated that only 5% 

of the population has a cell phone despite annual growth rates 

of 85%. 1 This will surely change dramatically in the next 5–10 

years as cellular and wireless Internet devices become ever 

cheaper, more capable, and increasingly convergent. The empowering 

effects of access to communication by the poor are 

well documented. What the world needs, and will surely have 

soon, is an affordable multifunctional device for voice and data 

communications. A new state-of-the-art device, such as an HP 

iPAQ 6300, costs about $500. For half as much, one eminent 

researcher, Raj Reddy, now claims it is possible to make and 

sell a device that connects the world’s poor. 2 

It seems likely that we are, at most, only a few generations 

of cellular and wireless technology away from a $100 

device. This is still not affordable by all individuals but should 

be accessible to most communities in populated areas. 

What will the connection be between researchers using 

networks like APAN and farmers and extension workers with 

Internet access via a prepaid subscription to a wireless Internet 

service The “CGIAR Knowledge Bank in your pocket” will 

be attainable. Whether it will feed anyone, of course, will be 

another challenge. Just delivering content will not be enough. 

Clearly, we are only at the beginning of an exciting era. 

Notes 

Author’s address: Head, Information Technology Services, IRRI, 

DAPO Box 7777, Metro Manila, Philippines, 

p.onolan@cgiar.org. 

1 www.asiamedia.ucla.edu/article.aspparentid=10184. Bangladesh cellphone 

market holds huge potential, Asia Media, 13 Apr. 2004. 

2 www.iht.com/articles/534041.html. At $250, a PC that aims to connect 

world’s poor. John Markov. International Herald Tribune, 16 Aug. 2004. 


A decision support system for site-specific nitrogen 

management of paddy rice 

Ryouji Sasaki, Kazunobu Toriyama, and Yoichi Shibata 

Rice yield and quality depend highly on soil nitrogen fertility 

(Murayama 1979). Yearly fluctuations of temperature and sunshine 

also affect rice growth and yield considerably. Thus, a 

decision support system formed by inputting those physical 

variables has been anticipated. A decision support system for 

rice was developed in Australia to calculate the amount of 

topdressing nitrogen fertilizer based on leaf nitrogen analysis 

(Angus et al 1996). 

Recently, the number of farmers with more than 10 ha 

has increased in Japan, accompanied by land consolidation, in 

which several small fields (0.1 to 0.3 ha) are joined to form a 

large field (1 ha), aiming at increasing production efficiency. 

However, this has often engendered variability in soil fertility 

within a large field, and thus uniform application of fertilizer 

results in the lodging of rice plants and considerable spatial 

variability in grain yield and quality. More information-intensive 

fertilization strategies, by which nitrogen fertilizers are 

precisely applied according to site-specific information about 

soil fertility, are needed to match fertilizer inputs to the seasonal 

pattern of rice crop nutrient demand and soil nitrogen 

supply. Although there was a trial for site-specific nutrient management 

in several Asian countries in cooperation with IRRI 

(Witt and Dobermann 2002), the required precision to overcome 

the variability in the reclaimed large field would be much 

higher. Thus, we have developed a computer-based decision 

support system based on the idea of site-specific nutrient management 

to solve the above problems. 

Simulation model and structure of the system 

Outline of the simulation model 

The simulation model employed in the system contains several 

simple submodels that simulate the development and 

growth of rice plants in daily time steps (Table 1). Most of the 

submodels were modified or evolved from the models by precursors. 

The developmental stage of rice plants is estimated 

by maximum and minimum air temperature, and is expressed 

as a continuous variable. The influence of daylength on panicle 

initiation is not considered in this model. The parameters of 

the model for soil nitrogen mineralization, such as mineralization 

rate constant, mineralization potential, and apparent 

activation energy, should be derived from the experimental 

data. 

Structure of the system 

This decision support system is named “RiceNiSMo” (Rice 

Nitrogen Simulation Model). It is written in Visual Basic and 

runs on PCs with Windows-OS. The target rice plant of the 

system is transplanted Koshihikari, a leading variety in Japan 

with good taste. This system has roughly four functions, as 

follows. 

Registration of nitrogen fertilizer. Not only ammonium 

sulfate or ammonium chloride but also the compound fertilizer, 

high-analysis mixed fertilizer, coated urea (controlledrelease 

urea N) could be registered in this system. Two or three 

different types of nitrogen fertilizers, which are mixed for basal 

dressing by users, could also be registered as original N fertil- 

Table 1. Relationships between simulated items and factors for the simulation. 

Simulated items 

Developmental stage 

Plant age in leaf number 

Soil nitrogen mineralization 

Nitrogen release from coated urea 

Nitrogen uptake 

Growth 

Canopy development 

Intercepted solar radiation 

Elongation of culm 

Lodging index 

Grain yield 

Factors 

Temperature 

Temperature, developmental stage 

Temperature 

Temperature 

Soil N mineralization, effective soil depth, amount of 

N fertilizer, temperature, developmental stage 

Intercepted solar radiation, temperature, plant N status 

Nitrogen uptake, plant N status, developmental stage, 

temperature 

Global radiation, leaf area 

Developmental stage, canopy development at panicle 

initiation stage, amount of N fertilizer as topdressing 

Culm length 

Dry weight, N uptake at harvest stage 


izers. However, the mixed N fertilizer for topdressing is not 

available at present. The temporal nitrogen release from the 

coated urea is simulated by the model reported by Hara (2000). 

Simplified estimation of soil fertility. To estimate soil 

fertility, users are requested to input each date and value in a 

normal year into the system as follows: (1) date of puddling, 

transplantation of seedlings, and harvest; (2) name of nitrogen 

fertilizers and the amount of basal and topdressing; and (3) 

grain yield as brown rice. 

We have collected a large number of data sets for brown 

rice yield and N uptake at harvest from field trials at several 

prefectural experiment stations in the Hokuriku region. These 

data sets were from 1994 to 1999. Grain yield ranged from 2.7 

to 7.0 t ha –1 . There was a close relationship between grain 

yield and N uptake at harvest; therefore, N uptake at harvest 

(Nup) could be calculated by using the yield data. The amount 

of N uptake from fertilizer (Nup_fertilizer) can be estimated 

assuming a reported recovery efficiency of fertilizer N. Thus, 

the N uptake from soil (Nup_soil) could be calculated by subtracting 

Nup_fertilizer from Nup. On the other hand, a temporary 

value for effective soil depth is preset to simulate the 

amount of mineralized soil N per unit land area. Soil N fertility 

depends on the amount of organic nitrogen in the soil profile. 

The importance of nitrogen fertility in subsoil was also 

suggested (Sekiya and Shiga 1977). Therefore, tuning the simulated 

Nup_soil with the Nup_soil calculated from the yield data 

would give a plausible estimate of effective soil depth. Although 

measurement of organic nitrogen along the soil profile 

by near-infrared spectroscopy (Toriyama et al 2003) is convenient, 

it still takes time for soil sampling and preparation for 

measurement. Recently, a yield-monitoring combine with a 

GPS has been developed in Japan (Chosa et al 2002); thus, the 

variability of soil fertility within a large field could be better 

estimated from the monitored yield data by employing the 

abovementioned scheme. 

Calculating the recommended amount of N fertilizer. The 

amount of N fertilizer for basal application and topdressing is 

calculated as follows. At first, agronomic variables, such as 

planned date of fertilization, puddling, transplantation of rice 

seedlings, and the like, and parameters reflecting soil N fertility 

are input into the system. Then, the growth and N uptake of 

rice plants is simulated by input variables with weather data of 

a normal year, and the application rate of basal N and 

topdressing N is computed to close the gap between the estimated 

and required N uptake at both stages. 

Rice variety Koshihikari tends to lodge at harvest. N 

uptake at panicle initiation was closely related to culm length 

at the ripening stage (Sasaki et al 2001), which is a good indicator 

for forecasting lodging at harvest. The optimum amount 

of N topdressing based on growth and N uptake at panicle 

initiation to avoid or reduce lodging could be recommended. 

Users of this system will be able to decide on optimum 

N fertilization based on the calculated values for the respective 

field. If the field is large and has considerable variability 

in soil fertility, then the procedures as above are followed likewise 

for each site so that variability in yield and quality among 

sites can be reduced. 

Application rate and timing of topdressing N. The input 

variables for cultivation and/or amount of fertilizer can be altered 

if necessary, and weather data for the respective year can 

be input daily or timely. The results of simulation for growth, 

canopy development, and the like could be referred to by users. 

The growth stage of both normal and ongoing years could 

be calculated. As fluctuations occur yearly in the soil N supply 

by temperature, the preset application rate of topdressing N 

can be reconsidered. It is recommended to measure the N uptake 

of the rice canopy at panicle initiation and revise the calculated 

value because the simulation model is simplified. Nitrogen 

uptake could be measured by a ground-based image 

mapping system equipped with a GPS and a CCD camera with 

band pass filter for near-infrared images, and a CCD camera 

to take images of the rice canopy vertically from zenith to the 

ground (Shibata et al 2002). 

The optimum date for harvesting could also be predicted 

by this simulation model based on the accumulated temperature. 

Thus, users can prepare harvesting work based on the 

prediction. 

Problems for future research 

Although the initial purpose to develop a decision support system 

for site-specific management of paddy rice was achieved, 

continuous improvement of the simulation model is necessary, 

especially when rice quality becomes a concern. The regulation 

of N content in rice grain based on the simulation model 

like the one above is strongly anticipated because N content is 

the index of rice palatability in Japan. Thus, the system will 

require a model with the relationship between the N status of 

the rice plant and N content of rice grains. The present model 

estimates the amount of N mineralized under specific conditions, 

and thus further study is needed to increase the validity 

of the model. Finally, variable rate applications are essential 

for site-specific management of rice in large-sized fields. For 

that purpose, a yield-monitoring combine is also important. 

Therefore, it is very important to communicate the information 

between these machines and this system on a computer. 

The system also needs a revision to enable the calculation of 

profit based on a price of grain and a cost of fertilizer N and 

the like. 

References 

Angus JF, Williams RL, Durkin CO. 1996. MANAGE RICE: decision 

support for tactical crop management. In: Ishii R, Horie 

T, editors. Crop research in Asia: achievements and perspectives. 

Proceedings of the 2nd Asian Crop Science Conference, 

21-23 August 1995. Fukui City (Japan): Asian Crop Science 

Association. p 274-279. 

Chosa T, Kobayashi K, Daikoku M, Shibata Y, Omine M. 2002. A 

study on yield monitoring system for head-feeding combines. 

I. Adoption of an optical sensor and a load cell as a yield 

monitor. J. Jpn. Soc. Agric. Mach. 64:145-153. 


Hara Y. 2000. Estimation of nitrogen release from coated urea using 

the Richards function and investigation of the release parameters 

using simulation models. Soil Sci. Plant Nutr. 46:693- 

701. 

Murayama N. 1979. The importance of nitrogen for rice production. 

In: Nitrogen and rice. Manila (Philippines): International Rice 


Sasaki R, Horie T, Toriyma K, Shibata Y. 2001. Factors affecting the 

yearly fluctuation in culm length of rice. Jpn. J. Crop Sci. 

70:489-498. 

Sekiya S, Shiga H. 1977. A role of subsoil of paddy field in nitrogen 

supply to rice plants. JARQ 11:95-100. 

Shibata Y, Sasaki R, Toriyama K, Araki K, Asano O, Hirokawa M. 

2002. Development of image mapping techniques for site-specific 

paddy rice management. J. Jpn. Soc. Agric. Mach. 64:127- 

135. 

Toriyama K, Sasaki R, Shibata Y, Sugimoto M, Chosa T, Omine M, 

Saito J. 2003. Development of a site-specific nitrogen management 

system for paddy rice. JARQ 37:213-218. 

Witt C, Dobermann A. 2002. A site-specific nutrient management 

approach for irrigated, lowland rice in Asia. Better Crops Int. 

16:20-24. 

Notes 

Authors’ addresses: Ryouji Sasaki, National Agricultural Research 

Center for Western Region, e-mail: ryouji@affrc.go.jp; 

Kazunobu Toriyama, Japan International Research Center for 

Agricultural Sciences, e-mail: toriyama@affrc.go.jp; Yoichi 

Shibata, Obihiro University of Agriculture and Veterinary 

Medicine, e-mail: yshiba@obihiro.ac.jp. 


The importance of information technology (IT) is increasing rapidly 

in agricultural production and research, as demonstrated by 

several examples of IT-related activities with rice. IT, unlike many 

other technologies, is horizontally applicable in that it has applications 

throughout almost all aspects of agriculture. This session 

discussed the present status of IT-based activities and provided 

an opportunity to discuss the future of IT applications in rice research 

and production. 

During this session, seven oral papers and nine posters 

were presented. Seven of the oral papers and one poster have 

been recommended for inclusion as full papers in the conference 

proceedings. The papers cover the application of IT in precision 

farming, a low-cost wireless-based field-monitoring robot, 

an integrated database system to efficiently transfer knowledge, 

lessons learned in an eLearning activity, an integrated system to 

accelerate rice breeding, and a high-performance Internet infrastructure 

to harness IT applications. The following is the summary 

of each paper and a discussion from each presentation. 

T. Chosa, in “Data mining using combined yield and quality 

maps of paddy fields,” discussed developments in precision farming 

in recent years, particularly in Japan and some other Asian 

countries. Precision farming involves making optimal decisions 

based on site-specific information, such as fertility, growth, and 

previous crop yields. In spite of the progress in information acquisition 

technology and site-specific management, there are still 

difficulties in providing optimal recommendations for crop management 

and no general method for producing recommendations 

has been determined. This paper suggested the need to 

develop a categorical map as a guideline for making recommendations 

for crop management in succeeding years by combining 

yield and quality maps of paddy fields. During discussion, the 

benefits of this approach to farmers were pointed out as one of 

the most important criteria for further development in precision 

farming. 

M. Hirafuji and others, in “A wireless sensor network with 

Field-Monitoring Servers and MetBroker in paddy fields,” discussed 

a low-cost, wireless-based field-monitoring system, Field Server. 

Field Server is a Web server installed in fields. When used with a 

Field-Monitoring Server (FMS), it allows real-time monitoring to 

identify wireless LAN hotspots around the FMS. The FMS consists 

of a Fieldserver-Engine (a micro Web server for control and 

data acquisition), wireless LAN access point, network camera, 

and sensors. By combining these elements, the FMS can be connected 

to various sensors and probes directly. A group of Field 

Servers can be linked to develop a linked sensor network. Data 

acquired from Field Servers by an agent program are stored in 

the FMS archive in an XML format and the stored data are linked 

to a weather database, accessible through a weather data mediator, 

MetBroker, to provide a consistent interface to several 

agricultural software applications. A group of Field Servers has 

been used in paddy fields for a long-term experiment without any 

trouble. Several questions arose during the discussion period— 

particularly about the availability and price of the FMS. 

A. Otuka and others, in “Prediction of airborne immigration 

of rice pest insects,” reported on a new simulation model to predict 

the airborne immigration of rice insect pests to Japan. The 

whitebacked planthopper and the brown planthopper are major 

pests of rice in eastern Asia, and migrate from southern China to 

Japan, primarily in June to July each year. Generally, the prediction 

of planthopper migration provides farmers and plant protection 

advisers with information about the source of the migration 

as well as where the migration is occurring. Knowing the migration 

source is important as it helps to understand pest characteristics 

such as biotype, pesticide resistance, etc. The quality of 

conventional prediction methods that use 6- or 12-h two-dimensional 

wind data is limited. The authors adopted a three-dimensional 

simulation method using a boundary layer model to develop 

a new method to predict pest immigration. Field observations 

of pest immigration over two seasons showed that the new 

method could predict immigration of the pest 2 days before its 

arrival with a precision as high as 80%. Compared with the precision 

of the former method, this precision is fairly high, indicating 


that the new method is useful. The new method can also be a 

tool to determine the take-off areas of pests. 

R. Raab and B. Abdon, in “Distance education and 

eLearning for sustainable agriculture: lessons learned,” described 

the experiments and lessons learned by the authors, who have 

been providing an eLearning service as an NGO activity in the 

Asian region. Specialized information and communication technology 

(ICT) applications such as eLearning are offering new options 

to deliver knowledge and information to farmers directly 

and indirectly through knowledge intermediaries. eLearning provides 

learning opportunities in subjects not offered locally or where 

local offerings lack quality. Agricultural knowledge acquired through 

formal education can soon become outdated and obsolete. Properly 

conceived and developed, nonformal eLearning can substantially 

complement formal and informal efforts and provide up-todate 

and relevant agricultural knowledge. Although not yet as 

prevalent in developing countries, there is clear evidence that 

developing-country farmers and agricultural professionals are 

equally eager to learn in this way. The majority of eLearning-related 

initiatives to date in both developed and developing countries 

have been in the area of online publishing of information as 

resources for informal learning. Although aggregated data are 

not available on formal agriculture-related courses, more and 

more universities and schools in developed countries are offering 

online learning opportunities. The most appropriate targets for 

eLearning in support of agriculture are currently agricultural professionals 

and knowledge intermediaries. 

M. Bell and D. Shires, in “The Rice Knowledge Bank,” reported 

the concept and purpose of the International Rice Research 

Institute’s Rice Knowledge Bank (RKB). Access to information 

is a major limitation to improving the livelihoods of farmers 

in Asia. IRRI therefore developed RKB as the world’s leading 

ICT repository of rice-based extension and training knowledge. 

Containing the most up-to-date and validated knowledge, the 

RKB is providing government extension services, NGOs, and universities 

with unprecedented access to training and support knowledge. 

The RKB was assembled from the knowledge base of IRRI 

by selecting those knowledge parts that were applicable to the 

intended audience, validating the selected knowledge, and assembling 

it into a dynamic, easy-to-use form. All RKB content is 

focused, credible, value-added, and demand-driven. Further, drawing 

on best industry practice, the RKB is developed around the 

concept of “single-source publishing,” which allows target audiences 

to access information in the format best suited to their 

circumstances: through the Internet, on CD-ROM, or on paper. 

Since its launch in September 2002, the RKB has received more 

than 4 million hits. 

G. McLaren and others, in “Integrated information systems 

for crop research and improvement,” reported on the International 

Rice Information System (IRIS) of IRRI. IRIS is the rice portion 

of the International Crop Information System (ICIS), a database 

system for the management and integration of global information 

on genetic resources and crop improvement for any crop. 

Major constraints to developing knowledge-intensive crop improvement 

programs include ambiguous germplasm identification, difficulty 

in tracing pedigree information, and lack of integration 

among genetic resources, characterization, breeding, evaluation, 

and use data. ICIS is being developed to overcome these constraints. 

The Genealogy Management System (GMS) is the core 

database, which ensures unique identification of germplasm, 

management of nomenclature, and retention of all germplasm 

development information. The GMS links germplasm with characterization 

and evaluation data, and a trait ontology with a controlled 

vocabulary ensures integration of information across different 

studies. New components are being added to the database 

to handle the diversity of rice functional genomics data, 

including genomic sequence data, molecular genetic data, expression 

data, and proteomic information. Integration with international 

bioinformatics resources is facilitated by adopting Web 

services technology using both XML Schema and BioMoby standards. 

P. O’Nolan, in “Using APAN for content delivery: possibilities 

for the CGIAR,” explained that the Consultative Group on 

International Agricultural Research (CGIAR) first joined the Asia 

Pacific Advanced Network (APAN) when IRRI joined on behalf of 

the system in 1999. A copy of the IRIS, a version of the ICIS that 

IRRI is developing in cooperation with partner organizations from 

several countries, is now hosted online at MAFF’s Information 

Network (MAFFIN), headquartered in Tsukuba. Data are replicated 

between MAFFIN and IRRI using APAN and it is planned 

that the MAFFIN network operating center (NOC) will be the hub 

for future data replication to networking centers in other Asian 

countries connected to APAN. Mirrored data will also be available 

via the Internet. Recently, IRRI’s APAN connectivity was totally 

changed, and today every networked computer in IRRI has a broadband 

connection to APAN. IRRI has also begun construction of 

an Access Grid node. Once additional CGIAR centers connect to 

advanced research networks, the CGIAR network can play a bridging 

role in facilitating face-to-face interactions between national 

partners at some locations and external partners. 

R. Sasaki and others, in “A decision support system for 

site-specific nitrogen management of paddy rice,” discussed a 

decision support system for site-specific nutrient management. 

As the scale of paddy fields is increasing in Japan, the in-field 

soil fertility variability in these larger fields becomes an issue because 

uniform application of fertilizer results in lodging of rice 

plants and considerable spatial variability of grain yield and quality. 

Thus, more information-intensive fertilization strategies, by 

which nitrogen fertilizers are precisely applied based on site-specific 

information about soil fertility, are needed to match fertilizer 

inputs to the seasonal pattern of rice crop nutrient demand and 

soil nitrogen supply. The authors developed a computer-based 

decision support system to solve this problem. The simulation 

model employed in the system contains several simple submodels 

that simulate development and growth of rice plants in daily time 

steps. This decision support system is named “RiceNiSMo” (Rice 

Nitrogen Simulation Model).

here - IRRI books - International Rice Research Institute

Create successful ePaper yourself

Delete template?

Save as template?