Botanical_Classification_of_Coffee - Rodrigo Gonzalez (4)

Prévia do material em texto

-- - 
i, 2 BOTANICAL CLASSIFICATION OF COFFEE 
André Charrier and Julien Berthaud 
--- - _ -- - - 
Introduction 
While the international coffee trade is concerned with only two coffee 
species - Coffea arabica and C. canephora - botanists regard as coffee 
trees all tropical plants of the Rubiaceae family, which produce seed 
resembling coffee beans. During botanical explorations of the tropical 
regions, from the sixteenth century onwards, wild coffees also attracted the 
attention of explorers and botanists. Their specimens are found in the 
herbaria and the names of the most famous explorers have been com- 
memorated in both specific and generic epithets. Hundreds of species have 
been described, but the taxonomic classification of the genus Coffea has 
become very complex and rather confused, 
Even the most authoritative classification system of Chevalier (1947) is 
now due for revision in view of the many new species discovered over the 
past 20 years on Madagascar (Leroy, 1961a, b, c, 1962, 1963, 1972a, b) 
and in East Africa (Bridson, 1982). 
In addition the conventional methods of taxonomic classification, which 
are mainly based on morphological characteristics of specimens deposited 
in herbaria, are inadequate to give full justification to the tremendous vari- 
ability encountered in the allogamous wild coffee populations. 
In their efforts to develop a more exacting system of classification, 
coffee botanists and geneticists (Charrier, 1978; Berthaud, 1985) have 
followed the trend of modern taxonomy, which today makes use of a vari- 
ety of scientific disciplines, to reflect the true genetic-historical relationship 
among plants and animals (Clausen, Keck and Hiesey, 1945; Dobshansky, 
1970; Mayr, 1970; Harlan and de Wet, 1971). 
This chapter presents a review of advances made in the taxonomy of 
coffee, starting with the standard classification based on herbaria specimens 
and living collections. This will then be followed by an analysis of data 
from studies of cytotaxonomy and cytogenetics, ecology and plant geo- 
graphy, as well as on biochemical and serological affinities, which have all 
contributed valuable knowledge needed for a better coffee classification. 
Detailed botanical descriptions of cultivated coffee species have been 
presented earlier by various authors (Wellman, 1961; Haarer, 1962; Coste, 
1968; Carvalho et al., 1969) and therefore need not be repeated here. 
_ _ .- - 
O RSTO M Fonds Doc" mentaire 
,: 3 , /I 5% 2% 4 
Cote i 
.; 
= 14 Botanical Classification of Coffee 
Taxonomy 
The first botanical description of a coffee tree, under the nam Jasminum 
arabicanum, was made in 1713 by A. de Jussieu, who studied a single plant 
originating from the botanic garden of Amsterdam. However, Linnaeus 
(1737) classified it as a separate genus Coffea with the then only one 
known species C. arabica. 
Many new species of Coffeu have been discovered during exploration of 
the tropical forests of Africa since the second half of the nineteenth cen- 
tury. Several botanists have tried to describe these species, but this led 
often to confusion and numerous epithets have proved to be synonymous, 
Special mention should be made of the extensive taxonomic work of 
Chevalier (1947) on the Coffea species of Africa and Madagascar and of 
Lebrun (1941), who paid particular attention to the coffees of Central 
Africa, especially those found in Zaire. 
Of recent date are the detailed taxonomic studies of Leroy (1967,1980) 
on the coffee species of Madagascar and of Bridson (1982) on the coffee 
species found in East Africa. Especially Leroy’s (1980) efforts to indicate 
the relationship between species of the genus Coffea and those of Psilan- 
thus and others are of particular importance to the understanding of the 
whole Coffeu spectrum. His most important criteria for differentiating the cy Axillary flowers 
.- Monopodial development 
o, 
4- 5 * Terminal flowers - Predominantly sympodial 
development 
Figure 2.1 : Coffee Type Placentation 
- Long corolla tube 
- Anthers not exserted 
- Short style 
GENUS: Psilenthus 
p. 
Subgenus Psilenrhus 
P. 
Subgenus Peracoffee 
Botanical Classification of Coffee 15 
Table 2.1 : Classification System for the Genera Cottea and Psilanthus 
According to Leroy (1 980), with Indication of Geographical Distribution 
Family: Rubiaceae, Subfamily: Cinchonoidee 
SubGenus Localisation 
. Psilenthopsis (ChevJLeroy Africa 
. Barecoffee (Leroy) Leroy 
Genus 
Coffee L. . Coffee Africa, Madagascar 
Africa, Madagascar 
Psihnthus (Ho0k.f.) . Parecoffea (Miquel) Leroy Africa, Asia 
New Guinea 
Africa . Psilanthus (Hook. f.) 
1 
Table 2.2: The Two Criteria used by Leroy (1980) to Differentiate Genera 
and Subgenera 
Criterion 1 
- Short corolla tube 
- Anthers exserted 
- Long style 
GENUS: Coffee 
I 
L. 
Subgenus Coffea 
Subgenus Berecoffee 
genus Coffea from all other genera within the family Rubiaceae are the 
type of gynaecium and placenta (Figure 2.1) The classification system pro- 
posed by Leroy (1980) for the genera Coffea and Psilunthus, with indi- 
cation of geographical distribution of the sub-genera, is presented in Table 
2.1. The differentiation between genera and sub-genera is based on two 
main criteria (Table 2.2): 
(1) flower shape, more in particular the length of the corolla tube (long 
or short) associated respectively with exerted anthers and style (Figure 
2.6-2) or inserted anthers and short style (Figure 2.6-1). 
(2) growth habit and type of inflorescence: monopodial with axillary 
flowers or sympodial with terminal flowers. 
However, the second criterion is not unconditional, as both growth habits 
can occasionally be found on the same plant in some species. 
On the other hand, Chevalier (1947) tried to group the species within 
the genus Coffea into the following four sections: 
.. - ^^ 
16 Botunical Classifi'culion o j Cojjee ' . 
s Argocoffea, Paracoffea, Mascarocoffea, Eucoffea. 
. 
According to Leroy's (1967) classification Argocoffea should be 
excluded fom the genus Coffea, because the seeds do not resemble coffee 
beans, and the section Paracoffea should be considered as a sub-genus of 
Psilanrhus (Table 2.1). The section Eucoffea, now more correctly named 
Coff'ea, and Mascarocoffeu include most of the presently known coffee 
species. 
Coffea 
This is divided into five subsections according to very diverse criteria: tree 
height (Nanocoffea), leaf thickness (Pachycoffea), fruit colour (Erylhro- 
coffea, Melanocoffea) and geographical distribution (Mozambicoffea). The 
species grouped under each subsection are shown in Table 2.3. 
Lebrun (1941) proposed the following series for Coffea species found in 
Central Africa, using the increasingly complex structure of the inflor- 
escence as main criterion: 
Abyssinicae, Robustae, Libericae. 
The Libericae series contain a single species, because all the species with 
thick leaves and large fruits recognised by Chevalier were placed in syn- 
onymy with C liberica by Lebrun. Consequently, the subsection Pachy- 
cof'ea of Chevalier becomes equivalent to the series Libericae of Lebrun 
and to C. liberica in a broad sense. Chevalier saw during his taxonomic 
studies only a few herbarium specimens of coffee species from East Africa, 
but nevertheless observed the originality of this material. Bridson (1982) 
carried out detailed studies on coffee species for the Flora of Tropical East 
Africa (not yet published). A list of coffee species and taxa found in this 
part of Africa, many of which are still incompletely known, is given in 
Table 2.4. Many of these coffee species are well adapted to drier climates 
by their xeromorphic characteristics and the short interval between flower- 
ing and mature seed (only three months versus 8-12 months for most other 
coffee species). 
Mascarocoffea 
The coffee species belonging to this section all have one characteristic in 
common: the absence of caffeine, which was first reported by Bertrand( 1902). The tremendous variety in forms has hampered taxonomic classi- 
fication. A first regrouping of species in series was-made by Chevalier 
(1938), but in view of the more than 50 different forms (Portbres, 1962; 
Leroy, 1961a, b, c, 1962, 1965, 1982) found within material collected dur- 
ing the early 1960s Leroy proposed a revision of the series within the 
Muscarocoffeu section (Table 2.5). This classification takes into account 
Botunical Clussijicarion of Cojyee 17 
Table 2.3: The Grouping of Species in the Subsection €ucoffea According 
to Chevalier (1947) 
Species 
C. canephora 
C. congensis 
Subsections 
Erythrocoffea C. arabica 
Pach ycoffea 
Melanoco ffea 
Nanocoffea 
Mozambico ffea 
C. abeokutae 
C. liberica 
C. klainii 
C. o yemensis 
C. dewevrei 
c. slenoph ylla 
C. carissoi 
C. mayombensis 
C. humilis 
C. brevipes 
C. togoensis 
C. schumanniana 
C. eugenioïdes 
C. kivuensis 
C. mu findiensis 
c. zanguebariae 
C. raceinosa 
C. ligustroides 
C. salvatrìx 
Table 2.4: List of Taxa of Coffea from East Africa Adapted from Bridson 
(1 982) 
C. fadenii 
C. mongensis C. mufindiensis 
C. kivuensis - subsp. mufindiensis 
- subsp. lundazlensis C. salvairix 
- subsp. autrah C. pseudozanguebariae 
C. pawekania C. sp. A 
c. sp. 0 C. ligustroides 
c. sp. c C. zanguebariae 
c. sp. D C. racemosa 
C. sp. E 
C. sp. F 
C. sp. G C. paolia 
C. sp. H 
c. sp. I 
C. sp. J 
c. sp. u 
leaf, fruit and seed characteristics. Charrier (1978) gives in his synthesis on 
Mascarocoffea a concise botanical description of all known species with 
details of their geographical distribution on Madagascar. 
Of considerable interest are also the efforts of Lobreau-Callen and 
Leroy (1980) and Chinnappa and Warner (1981) to establish a palyno- 
I 
. - 18 Botanical Classification of Coffee 
2 
, 
Table 2.5: List of the ’Series’ proposed by Chevalier (1938) and Revised by 
Leroy for the Mascarocoffea of Madagascar 
Series: 
- Verae Chev. 
- Multiforae Chev. 
- Scleroph yllae Chev. 
- Brachysiphon Dubard 
- Terminales Chev. 
- Garcinioides 
- Mauritianae 
- Macrocarpae 
Chev. 
Chev. 
Chev. Humblotianae Ler. 
- Unclasssd 
logical basis for coffee taxonomy. It enables differentiation at the generic 
level, as the 3-porate pollen of Coffea species is easily distinguished from 
the 4-5 porate pollen of Psilanthus and Paracoffea species. Unfortunately, 
the eight morphologically different types of pollen described by Chinnappa 
and Warner (1981) bear little relation to the subsections and series in the 
taxonomic classification of the genus Coffea. 
The wealth of forms encountered in wild coffees is”the result of the 
interaction of the genetic variability of coffee populations with an endless 
range of ecological ‘microniches’ (Forster, 1980) of natural habitats. 
Botanists have often failed to take this variation into account by restricting 
themselves mostly to morphological characters, when they tried to establish 
clear-cut distinctions between species in which genetic differentiation is still 
occurring. 
Souces of Botanical Information 
Information on coffee species in relation to habitats, taxonomic charac- 
teristics, genetic diversity and geographical distribution can be found in 
herbaria, travel reports of botanical explorers and in the living collections 
of research stations. 
Major Herbaria 
Because of the historical ties with Africa and South East Asia most Euro- 
pean botanists of earlier times sent their collections to herbaria in their 
home countries, with the result that these herbaria are characterised by a 
high degree of geographical specialisation. This applies also to the herb- 
arium specimens of coffee, which are therefore important sources of infor- 
mation for botanists when preparing new missions to collect and preserve 
coffee genetic resources in centres of genetic diversity ín Africa. Below 
follows a brief description of the herbaria most relevant for coffee: 
Botanical Classification of Coffee 19 
(1) The Royal Botanic Garden at Kew and the Natural History Depart- 
ment of the British Museum in London, UK: comprehensive collections 
from the Sudan, Uganda, Kenya, Tanzania and other eastern and south- 
ern African countries, as well as from English speaking West Africa and 
Angola. 
(2) Jardin Botanique National de Belgique at Meise, Belgium: with 
emphasis on Zaire. 
(3) Museum National d’Histoire Naturelle at Paris, France: with com- 
prehensive collections from Guinea, Ivory Coast, Cameroon, Gabon, 
Congo, the Central African Republic and the Malagasy Republic. 
(4) Botanische Garten und Botanische Museum at Berlin-Dahlen, 
Germany: with collections from Togo, Cameroon and Tanzania, which 
were however almost completely destroyed in 1943. 
(5) Herbarium Vadense of the Department of Plant Taxonomy and 
Geography of the Agricultural University of Wageningen and the 
Rijksherbarium of the University of Leiden, The Netherlands: with 
coffee specimens from respectively Africa (especially Cameroon) and 
Indonesia. 
(6) Erbario Tropicale di Firenze, Florence, Italy: collections from 
Ethiopia and Somalia. 
(7) Botanical Institute of the University of Coimbra and Centro de 
Botanica da Junta de Investigaçoes Cientificas do Ultramar at Lisbon, 
Portugual: with collections from Angola and Mozambique. 
These European herbaria may already give a very good impression of 
the overall distribution of the wild coffees. Nevertheless, the importance of 
national herbaria in Africa should not be underestimated, as they often 
represent part of the duplicate herbarium specimens sent to Europe and 
some have also extended their collections through their own explorations. 
A comprehensive list covering all the coffee specimens present in the 
herbaria of Europe and Africa is lacking. It would be very difficult to 
realise, because determinations made are sometimes inaccurate or only 
approximate descriptions. Many of the collected plants have still to be 
determined taxonomically and to be classified. 
Natural Habitats of the Wild Coflees 
Wild coffee trees are components of the understorey of tropical forests in 
Africa. The observations made by collectors include precise descriptions of 
habitats, which are quite distinct for C. arabica and other coffee species. 
C arabica. All botanists, who have explored the forests on the south- 
western highlands of Ethiopia, agree in their observation that this is the 
centre of diversity of C. arabica, but that it is very difficult to fiid truly wild 
populations (Sylvain, 1955; Von Strenge, 1956; Meyer, 1965; Friis, 1979). 
' 20 Botanical Classification Of coffee 
i This species is indeed very common to the understorey of the forest, but 
most of the trees are regularly harvested by the local people, who usually 
carry out some form of maintenance by clearing the bush to facilitate pick- 
ing of the ripe cherries. Guillaumet and Hallé (1978) distinguished the 
following stages from practically wild to truly cultivated coffee plots: (a) 
natural populations maintained in situ; (b) populations improved by intro- 
duction of young coffee trees from elsewhere; (c) farmers' plots established 
On the other hand, Thomas (1942) found in secondary forest on the 
Boma plateau in south-eastern Sudan truly wild populations of C. arabica, 
which apparently grew there without human interference. 
Semi-wild populations of C. arabica can also be found at an altitude of 
about 1500 m in the upland forest on Mt Marsabit in north Kenya 
(Berthaud, Guillaumet, Le Pierrb and Lourd, 1980). It is not clear, 
whether these trees are truly wild or in earlier times brought from Ethiopia 
by man. 
The Other Coffee Species. Apart from some subspontaneous populations 
of C. liberica and C. canephora found in Ivory Coast and Central Africa, 
most other wild coffee species appear to grow in their natural habitat, 
which consists of the understorey of tropical forests. However, these forests 
are an extremely complex and organised environment. All thecollectors 
have noticed a really specific adaptation of coffee trees to elevation, alti- 
tude, rainfall and soil types. 
In East Africa rainfall is largely controlled by altitude and mountains 
stand out as wet areas covered in dense forest (Lind and Morrison, 1974). 
Typical examples are found in Kenya: Mt Marsabit with C. arabica and the 
Taita Hills with C. fadenii. In contrast, where variations in altitude are 
much less pronounced as in West Africa, the well-drained hill tops are the 
driest zones. For example, in the Ira forest of Ivory Coast C. srenophylla 
grows on the top of the hill, while C canephora and C. liberica occupy the 
lower humid zones (Figure 2.2). This location of C. stenophylla popu- 
lations is common in the western part of the Ivory Coast and is an indi- 
cation of its specific adaptation to dry conditions. In the eastern part of that 
country C. stenophylla is found in the dry lowlands with a semi-deciduous 
type of forest. 
C. humilis represents a good example of the importance of the biotope: 
whatever the altitude, this species grows in the small talwegs and at the 
border of swamps of south-westem Ivory Coast. This species is able to 
, grow id very wet environments, which would explain its limited area of 
distribution. Thomas (1944) reported that in Uganda C. eugenioides is 
always found in the drier zones of the forest and C. canephora in more 
humid environments. 
In the Central African Republic Berthaud and Guillaumet (1978) 
' 
. with plant material derived from wild coffee trees. 
WJ-. 
I 
- 
E 
Y, 
a, o 
m 
.- u 
N 
.- 
O 
Rotanical Classification of Cojjee 21 
O O r 
22 Botanical Classification of Coffee 
- . observed also the presence of C. dewevrei in the periphery of gallery forests 
bordering savannah land, where the soil is well-drained (Figure 2.3-1). In 
the same country C. congensis was found on the eroded banks of the 
Oubangui river and on sandy islands, but always in seasonally flooded 
areas. When the river is high, the coffee roots hold the sandy soil in place, 
with the result that each coffee tree stands on a little hillock (Figure 2.3-2). 
On Madagascar, within the same geographical area where C. resinosa 
and C. richardii are found together, the distribution of the two species is 
closely associated with pedological characteristics: spodosols for C. 
resinosa and oxysols for C. richardii. 
The habitats of the various coffee species correspond closely with spe- 
cific biotopes and these should, therefore, be very well known in order to 
be able to discover wild coffee populations. Such ecological data are also 
essential for selecting appropriate growing conditions for living collections 
and testing of the agronomic qualities for cultivation. 
Figure 2.3: Cross Section of Wild Coffee Habitats. 1. C. dewewrei(in dark) 
at the periphery of a gallery forest (Lihou - Central African Republic). 
2. C. congensis (in dark) on sandy banks and small hillocks (Louma island 
- Central African Republic). (From Berthaud and Guillaumet, 1978.) 
Botunical Classification of Coffee 23 
Collecting Missions for Wild Cojyee Species 
As already mentioned, exploration for wild coffee species started together 
with that of other tropical plants in the sixteenth century and was particu- 
larly intense in Africa towards the end of the last century and during the 
first part of the twentieth century. Evidence for that era is mostly found in 
herbaria and very little in the existing living collections. 
Of course, all coffee research centres maintain collections of the culti- 
vated coffee species. These working collections are used for the improve- 
Table 2.6: Coffee Collecting Missions Since 1960 
Germplasm 
Countries Collectors' Coffee maintenance 
Years explored Organisations names species countries 
1964 
1966 
1960 
to 
1974 
1975 
1975 
to 
1980 
1977 
1982 
1983 
Ethiopia FAO 
Ethiopia ORSTOM 
Madagascar Museum 
Mauritius 
Reunion Island ORSTOM 
Comoro Islands 
IFCC 
Central Africa ORSTOM 
Republic IFCC 
Ivory Coast ORSTOM 
Kenya ORSTOM 
IFCC 
Tanzania ORSTOM 
IFCC 
Cameroon ORSTOM 
IFCC 
Meyer C. arabica Ethiopia 
Monaco India 
Narasimhaswamy Tanzania 
Fernie 
Greathead 
Guillaumet 
Hall6 
Leroy 
Porteres 
Vianney-Liand 
Guillaume t 
Charrier 
Guillaumet 
Eerthaud 
Berthaud 
Eerthaud 
Guillaumet 
Lourd 
Eerthaud 
Anthony 
Lourd 
Anthony 
Couturon 
de Namur 
Costa Rica 
Ethiopia 
Madagascar 
Ivory Coast C. arabica 
Cameroon 
Mascarocoffea Madagascar 
(up to 50 taxa) 
C. congensis Ivory Coast 
C. dewevrei Central African 
C. canephora Republic 
Cafbier de la Nana 
C. liberica Ivory Coast 
C. stenoph ylla 
C. canephora 
C. huniilis 
Paracoffea sp. 
Psilanthus sp. 
C. arabica Kenya 
C. eugenioides Ivory coast 
C.zanguebariae 
C. fadenii 
C. zanguebariae Tanzania 
C. mufindiensis Ivory Coast 
c. cmephora Cameroon 
C. liberica Ivory Coast 
C. congensis 
C. brevipes 
c. sp. 
Psilanthus sp. 
c. sp. 
'I 
Botanical Classification of Coffee 25 
* .- 24 Botanical Classification of Coffee 
- 
. 
ment of C. arabica and C. canephora. Such collections have been extended 
regularly by introductions of more or less selected plant material from 
other coffee stations, botanic gardens or local plantations. The other coffee 
species are usually scarcely represented. 
Awareness of the lack of variability in the existing coffee collections 
made the FAO and French organisations (ORSTOM, IRCC, the Museum 
at Paris) intensify their efforts to collect coffee germplasm during the last 
twenty years. The most important collecting missions are listed in Table 
2.6, with indication of the countries covered, names of the collectors, the 
species collected and the countries where the material is maintained 
(Meyer er al., 1968; Charrier, 1980; Berthaud, Guillaumet, Le Pierrès and 
Lourd, 1977; Leroy, 1961a, b, cy 1962, 1963, 1982). Emphasis in the 
collection of coffee germplasm was particularly on C. arabica because of its 
economic importance, but a number of non-cultivated species have also 
been collected, such as those of the section Mascarocoffeu (caffeine free), 
the subsection PachycojTea, C. congensis (progenitor of the Congusta 
hybrids), C. eugenioides (presumed progenitor of the allotetraploid C. ara- 
bica) and of the related genus Psilanthus. 
Living Collections 
A comprehensive botanical study should also include the World living 
collections of coffee. The FAO prepared an inventory of the existing coffee 
collections in 1960 (Krug, 1965) and this was updated in 1978-79 
(unpublished). However, these reports do not fully reflect the actual 
importance and extent of the genetic diversity of the living collections. The 
location of living collections is depicted in Figure 2.4. 
Important collections of C. arabica with material of the Ethiopian centre 
of genetic diversity are present at Jimma (Ethiopia), Turrialba (Costa 
Rica), Campinas (Brazil), Chinchina (Colombia), Lyamungu (Tanzania), 
Ruiru (Kenya), Foumbot (Cameroon), Man (Ivory Coast) and Ilaka-Est 
(Madagascar). There is a unique collection of species of the section 
Mascarocofeu at Kianjavato (Madagascar). 
The main African coffee species are kept in the living collections at Divo 
and Man in Ivory Coast: more than 10 species of the section Coffeu are 
represented by hundreds of thousands of genotypes collected from natural 
populations in different countries since 1965 (Table 2.6). 
At the same time the working collections of C. arabica and C. cane- 
phora maintained in India, Cameroon, Togo, Angola and other countries 
also contain very valuable material. 
In tropical Africa one can find, at least for the time being, wild coffee 
trees in undisturbed forests and subspontaneous coffee in traditional agri- 
cultural areas. However, it has become a matter of urgency to continue 
with the collection of coffee germplasm, where the natural habitats are 
being threatened by human activities. 
2 
o 
2 
C 
u) 
al 
O 
al 
O. 
v) 
al 
.- 
.- 
o 
Y-O 
u) 
C 
O 
o 
a, 
O o 
ul 
C 
> 
o) 
C 
u) 
X w 
.- c. 
- - 
.- 
ï 
.- c. .- 
.. 
L 7 
J 
o) 
h 
26 Botanical Classification of Coffee 
* . Figure 2.5 presents a map of Africa, indicating the areas of origin or 
high genetic diversity for coffee species of the sections Coffea and 
Mascarocoffea. This map is based on accumulated information from herb- 
aria, collecting missions and living collections. 
Cytotaxonomy and Reproductive Systems 
Chromosome Num ber 
Results of studies on chromosome numbers in coffee carried out since the 
1930s have been reviewed by Sybenga (1960). The basic genome of the 
genus, x = 11 chromosomes, is typical for most of the genera of the family 
Rubiaceae. Chromosome counts were made for most species of the genus 
Coffea and for some representatives of the genus Psilanthus. 
In the section Coffea all species are diploid with 2n = 22 chromosomes, 
except for the tetraploid C. arabica which has 2n = 4x = 44 chromosomes. 
In the section Mascarocoffeu the chromosome number of more than 20 
species has been determined (PortBres, 1962; Leroy and Plu, 1966; Fried- 
man, 1970; Louam, 1972). Species belonging to this section are all diploid 
(2n = 22) and the very large variability encountered in this section cannot 
therefore be attributed to variation in chromosome number. 
In the genus Psilanthus the chromosome number 2n = 22 has been con- 
firmed for P. humberrii (Leroy and Plu, 1966), P. horsfieldiana (Bouhar- 
mont, 1959), P. bengalensis (Fagerlind, 1937) and P. mannii (Couturon, 
personal communication). 
Deviations from the normal chromosome number do occur in 
exceptional cases. For example, in C. arabica a whole series of polyploids 
have been found (Sybenga, 1960): triploids (3n = 33), pentaploids (5n = 
55), hexaploids (6n = 66) and octoploids (8n = 88). Haploids, or more 
correctly called di-haploids with n = 2x = 22 chromosomes, occur in low 
frequencies- in seedling offspring as weak plants with narrow leaves and 
have been called monosperma (Mendes and Bacchi, 1940). 
Diploid C. canephora with sectorial tetraploid chimaeras have been 
found and induction of artificial autotetraploidy by colchicine treatment of 
germinating seed (Mendes, 1939) or shoot apices (Berthou, 1975; Noirot, 
1978) is possible. Methods of recovering the naturally occurring haploid 
plants in C. canephora have been developed recently (Couturon and 
Berthaud, 1982). Doubling with colchicine results in homozygous plants of 
C. canephora, which are of great interest to coffee breeding (see Chapter 
The morphology of coffee chromosomes was studied by Mendes (1938) 
and Bouharmont (1959, 1963). The latter author described some 10 
species and prepared an average idiogram for the 11 chromosomes of the 
basic genome for the genus Coffea. Coffee chromosomes are relatively 
3). 
Botanical Classification of Coffee 27 
Figure 2.5: Natural Distribution of Coffee Species in Africa and 
Madagascar 
* C. arabica 
* C. brevipes 
0 C. canephora 
C. congensis * C. eugenioides 
C. humilis 
0 C. liberica 
#t C. mufindiensis 
V C. racemosa 
O C. stenophylla 
c C. zanguebariae 
0 Mascarocoffea 
I 
i .- 28 Botanical Classification of Coffee Botanical Classification of Coffee 29 
small (1 - 3pm), but modem methods of staining chromosomes (banding) 
open up possibilities of studying the chromosome morphology in more 
detail. 
‘i 
Reproductive Systems 
Psilanthus. In this genus flowers have a short style and long corolla tube 
(Figure 2.6-1). System of incompatibility appears to be absent and most 
species, e.g. P. bengalensis, are autogamous. However, occasional cross 
pollination and consequently genetic exchange may occur, as heterozygous 
plants have been found after analysis by electrophoresis. 
Coffa arabica. Flowers of this species are typical of the genus Coffea: 
short corolla tube, long style and exerted stamens (Figure 2.6-2). Such 
morphology would permit natural cross pollination, but nevertheless C. 
arabica is largely autogamous. Fruit set after self pollination is 60 per cent 
or higher (Carvalho et al., 1969). In Ivory Coast, Le Pierrb (personal com- 
munication) effected self pollination on 32 trees of F, progenies of crosses 
between various accessions of C. arabica from Ethiopia. Of the 8,400 
flowers 5,400 set fruit after self pollination. This means a success rate of 65 
fruits to 100 flowers. However, this percentage varied from 3 to 91 per 
cent between trees, which could indicate variation in the degree of self fer- 
tility in wild populations of C. arabica. Meyer (1965) reported 40-60 per 
cent cross pollination in wild populations of C arabica in Ethiopia. 
Most studies on the degree of natural cross pollination were carried out 
on cultivars of C. arabica, which underwent many cycles of selection. By 
using the recessive marker genes Cera (yellow endosperm) and Purpura- 
scens (purple leaves) Carvalho and Krug (1949) in Brazil and Van der 
Vossen (1974) in Kenya found percentages of natural out-pollination rang- 
ing from 7 to 15 per cent. 
Figure 2.6: 1. Flowers of Psilanthus, on the Right: Flower with Removed 
Corolla; 2. Flowers of Coffea 
Diploid Species of the Genus CojTea. Most diploid coffee species have 
proved to be highly self incompatible including 24 tested species of the 
section Mascarocoffea. Devreux, Vallayes, Pochet and Gilles (1959) 
describe how, after self pollination, pollen tube growth on the stigma of C. 
cunephoru becomes distorted and further penetration into the style is 
blocked. Berthaud (1980) produced further evidence for a gametophytic 
system of incompatibility in C. cunephora, which is controlled by one gene 
with multiple alleles. A similar mechanism appears to operate also in Con- 
gusta coffee (a hybrid between C. corrgensis and C. canephora). On the 
other hand, observations on C. liberica indicate that the incompatibility 
reaction can be delayed until the day of anthesis. Penetration of the pollen 
tubes in the stigma and style could take place with ‘bud pollination’ 
(Hamon, personal communication). 
One notable exception appears to be an accession in the living coffee 
collections in Ivory Coast of unknown origin but resembling C. brevipes 
and self-compatible (Le Pierrb and Louarn, personnal communication). 
We observed that some pollen tubes grow through the style while others 
were blocked at the papillary zone of the stigma, as in incompatible com- 
binations. The offspring of such trees was very homogeneous. Could this 
mean mutation of an S allele making it inoperative? 
Because of the high degree of self-incompatability in most diploid 
species, a high level of heterozygosity will be maintained in populations and 
this has great consequences to breeding (see Chapter 3). 
Wild Populations of Coffea 
The variability existing in natural coffee populations has been studied to a 
limited extent so far. Porteres (1937) was one of the first to describe the 
variability of populations of C. cunephoru, C. liberica and C. stenophyllu 
and their offspring in Ivory Coast. Considerable work was also carried out 
on Java on numerous progenies of introduced coffee species (Cramer, 
1957). 
Botanists and geneticists have in their recent efforts to explore and pre- 
serve coffee genetic resources applied various methods to describe the vari- 
ability present in wild coffee populations, including: (a) morphological 
observations and numerical taxonomy; (b) analysis of electrophoretic varí- 
ants; (c) studies of the frequency distribution of incompatibility alleles 
within and between populations; and (d) genetic analyses with progenies of 
controlled crosses. 
A few examples of natural coffee populations may serve as an illustra- 
tion. 
30 Botanical Clussification of Coffee 
I 
Coffea arabica 
The great variability in natural populations of C. arabica has been apparent 
to most botanists and geneticists, who visited and explored the south- 
western highlands of Ethiopia. 
An electrophoretic analysis ofsix enzyme systems produced similar 
homogeneous patterns, both for accessions from Ethiopia and from Mt 
Marsabit in Kenya (Berthou and Trouslot, 1977, and unpublished). This 
made it possible to describe the electrophoretic type for C. arabica. It is of 
inter st to note here, that this analysis generally pointed to an expression of 
two alleles at each locus, which would support earlier conclusions from 
classical genetic analyses (Carvalho et al., 1969) that the allotetraploid C. 
arabica is a functional diploid for most gene expressions. 
This uniformity of the species is lost if the morphological characteristics 
differences) are examined. Actually, hierarchical variance analysis carried 
out by Reynier, Pemes and Chaume (1978) and Louarn (1978) showed 
that differences between origins and between families are both significant, 
but with the interorigin variance component being several times higher. 
The caffeine content of C. arabica germplasm from Ethiopia varies from 
0.8 to 1.9 per cent (of dry matter), a variation not correlated with geo- 
graphical distribution. 
It is difficult to differentiate ‘varieties’ within this (semi-)wild material 
because of the considerable variation existing between trees of the same sub- 
population and such trees produce also heterogeneous progenies. The term 
variety (or cultivar) should therefore be reserved for the cultivated forms of 
this species, which are indeed homogeneous due to several generations of 
line selection. 
C. canephora: the ‘Nana’ Population 
This population is located in a gallery forest of the savanna zone in the 
Central African Republic and it represents the northern limits for the dis- 
tribution of C. canephora. This material is well adapted to the climatic and 
pedological conditions of that area and it has been cultivated locally since 
1923. The original population was rediscovered in 1975 (Berthaud and 
Guillaumet, 1978) and about 100 plants raised from seed and cuttings 
were added to the coffee collection in Ivory Coast. 
Observations on morphological characteristics and floral biology indi- 
cate that there is considerable variation, but no sub-populations could be 
detected. Some characters like colouration of mature fruits appear to be 
fixed, while others are heterozygous. This was also confirmed by electro- 
phoretic analysis. An analysis of incompatibility points to a considerable 
number of S alleles being present in that population. 
Taxonomically, there is now sufficient evidence to consider the Nana 
7 
of germplasm collections (origin differences) or their progenies (family 1 
I 
ßotunical Classification of Coffee 31 
population as part of the species C. canephora. The diversity found within 
this material appears similar to that encountered in other populations of C. 
canephora. 
C. stenophylla 
Wild populations of this species exist in west and east Ivory Coast. Pheno- 
typically these populations are easily distinguishable as is shown in Table 
2.7. Electrophoretic analysis indicates that these populations are very 
homogeneous but the alleles of the western populations differ from those 
fixed in the eastern ones. In a population of 1000 trees of western origin, 
the total number of S alleles was less than 10, which would suggest that the 
population originates from five parent trees at the most. Notwithstanding 
the expected relatively low genetic diversity, caffeine content deter- 
minations gave a range from 0.9 to 1.9 per cent (on dry matter), a variation 
similar to C. arabica. 
The geographical isolation apparently produced considerable genetic 
diversity between populations, as interpopulation crosses gave distinct 
hybrid vigour. 
C. zanguebariae 
When examining herbarium specimens one realises that the specific name 
applies to two taxa, one without and one with fruit stalks. These forms are 
indicated here as A and B, which is synonymous with the taxa Cojfea sp. A 
and C. pseudozanguebariue described by Bridson (1982). Various popu- 
lations, collected in Kenya by Berthaud et al, (1980), are now under obser- 
vation in Ivory Coast. The main characters of these two taxa are compared 
in Table 2.8. Analysis by electrophoresis has also indicated differences 
between the two taxa. Three of the studied populations were found to 
belong to form B, while one included both forms as well as intermediary 
types. Plants of the A form have all flowers open on the same day, while 
trees of form B flower over three consecutive days (Hamon et al., 1984) 
Hybridisation between the two forms is possible, although the chances of 
natural cross pollination between trees of the B form only should be higher. 
Table 2.7: Morphological Characteristics of the Western and Eastern 
Forms of C. stenophylla in Ivory Coast 
Observed character Eastern form Western form 
Secondary branching pattern very numerous branches numerous branches 
very small Leaves 
1 2 Flowers per fascicle 
Flower shape globulous 
black black Fruit colour 
small 
oblong 
1 . 
32 Botanical Classification of Coffee Botanical Classification of Coffee 33 
Table 2.8: Morphological Characteristics of the A and B Forms of 
C. zanguebariae. (After Hamon and Anthony, to be published.) 
A form E form 
thick thin 
long short 
small big 
6 6 
1 3 
5 6-7 
very short long 
green brown 
Leaf-thickness 
Stipule length 
Domatia 
Day range between shower and first flowering 
Number of flowering days 
Corolla lobes number 
Fruit stalk length 
Fruit colour before ripening 
However, Anthony (personal communication) and Louarn (1982) carried 
out crosses between trees of the A and B forms as well as between the A 
form and several other species. Crosses between A and B forms gave a 
considerably lower fruit set than crosses between the A form and for 
instance C. eugenioides and C. racemosa. It appears that a genetic barrier 
has evolved between the two forms but not due to geographic separation. 
Such a genetic barrier had not evolved, however, between the earlier 
mentioned subpopulations of C. stenophylla, regardless of the considerable 
morphological differences. 
C. liberica 
This species also shows distinctly different forms: C. liberica var. liberica 
from West Africa (Guinea, Liberia, Ivory Coast) and C. liberica var. de- 
wevrei from Central Africa (usually called Excelsa coffee). Forms of C. lib- 
erica found in Cameroon appear to be intermediate between these two 
extremes, but this still requires verification. 
Mascarocoffea 
A large number of different species (or taxa) have been distinguished 
already (Charrier, 1978). However, the relatively high success of inter- 
specific hybridisation within this section indicates that only weak genetic 
barriers have evolved. Figure 2.7 presents a hierarchical classification, 
mainly based on morphological characteristics. 
Multispecific Populations 
The existence of multispecific wild coffee populations is supported by 
observations made during various collecting missions carried out in the past 
20 years. 
Apart from the coexistence of the A and B forms of C zanguebariae in 
one population in Kenya, the following associations have been encountered 
in Ivory Coast and in the Central African Republic: 
Figure 2.7: Differentiation of the Taxa of Mascarocoffea by a Hierarchical 
Classification 
A206 
t A315 
u) 
A8 
A403 
A31 7 
A18 
A20 
A204 
A307 I r 
A205 
A230 
A l 38 
A213 
A743 
A320 
1 
Index of Similarity 
L 
a s O ? O p' '9 O O 9 m O O 
' 34 Botanical Classification of Coffee Botanical Classification of Coffee 35 
- C. liberica + C. canephora (Gbapleu, Ivory Coast) 
- C. liberica + C. humilis (Tai, Ivory Coast) 
- C. libericu + C. stenophylla + C. canephora (Ira, Ivory Coast) 
- C. liberica + C. congensis + C. canephora (Bangui, Central 
African Rep.) 
Multispecific populations have also been found on Madagascar with 
species of the section Mascarocoffeu (Charrier, 1978), while in Uganda 
associations of C. cunephora with C. liberica or C. eugenioideshave been 
reported by Thomas (1944). 
These observations made on wild coffee populations strongly suggest 
that: 
(1) hybridisation between geographically adjoining species in wild 
populations occurs only sporadically. On Madagascar hybrids between 
two species were found on two occasions, while in Ivory Coast occa- 
sional interspecific hybrids were noticed in the seed progeny of wild 
coffee trees; 
(2) interspecific hybrids represent a transitory state, which may explain 
why they are rarely encountered in natural conditions. 
Various factors appear to restrict chances of gene exchange between 
species in multispecific populations: 
(1) species coexist more side by side within a limited area rather than in 
a real mixture: their separation can be due to microclimatic (different 
sides of a hill), pedologic (different soils) or topographic differences; 
(2) flowering patterns differ between species: C. canephora usually 
flowers early in the year, while C. liberica and C. stenophylla tend to 
flower later: the time interval between a flowering inducing rainshower 
and anthesis is also often one day longer for C. canephora than for C. 
liberica. 
Clearly, studies of wild coffee populations have contributed considerably 
to a better understanding of the relation between species of multispecific 
associations. Genetic exchange between coffee species through interspecific 
hybridisation should be taken into account when formulating theories on 
the evolution of coffee species. 
Evolution of the Coffee Gene Pool 
Most studies on the genetic relationships of species in the genus Coffea 
before 1960 were almost entirely restricted to the subsection Erythro- 
v) 
al 
o 
al 
.- 
n a 
K3 
a 
n 
i3 
O 
C 
al 
2" al 
m 
2 
u) 
al 
lh 
lh 
o 
o 
G= 
o 
al 
.- 
n 
t 
al 
C 
r 
I 
*- 
O 
o 
+.I 
al 
t 
al 
.- 
5: u 
al s 
3 
a 
*- 
O 
al 
> 
al 
.- 
.. 
2 
- 
(1) N 
O .- O 
t5 
c 
o E m 
. 36 Botanical Classification of Coffee 
d 
coffea. More recently, however, the crossability and chromosomal homo- 
logy among other coffee species have received much attention, particularly 
in Brazil (Carvalho and Monaco, 1968), India (Vishveshwara, 1975), 
Madagascar (Charrier, 1978) and in Ivory Coast (Capot, 1972; Louarn, 
1982). Coffee taxonomists have also started to establish the biochemical 
and serological affinities of coffee taxa through research on enzyme poly- 
morphism, cytoplasmic DNA and serological reactions. 
Considerable progress has thus been made in the understanding of the 
genetic-historical relationship among coffee species. 
The Diploid Coffee Species 
From the extensive cytogenetic studies on interspecific hybrids - a sum- 
mary of the most relevant data together with major references is presented 
in Table 2.9 - the following main conclusions can be drawn as regards the 
genetic differentiation of the diploid coffee species 
* . 
. 
Botanical Classification of' Coffee 37 
(1) Absolute crossing barriers appear to be absent within the genus 
Coffea, although considerable variation in the degree of successful 
hybridisation exists between species of different taxa. 
(2) Exact quantitative information on the genetic relation of species is 
difficult to obtain due to the influence of genotype, crossing techniques 
and environment on the rate of success of interspecific hybridisation. 
(3) Maternal effects can be very significant in crosses between species, 
which differ in the time interval between flowering and ripe fruits: the 
best results are usually obtained when the species with the quickest for- 
mation of the albumen is used as female parent. 
(4) In general, crosses between species within one group present the 
least difficulty: for instance C. canephora and C. congensis, which both 
belong to the subsection Erythrocoflea, or species of the section 
Mascarocoflea. However, considerable success can sometimes be 
obtained as well with taxa of different taxonomic groups: for instance C. 
Iiberica and C. eugenioides or C. dewevrei and C. stenophylla. The 
species Psilanthopsis kapakata (erroneously classified in another genus 
by Chevalier) can easily be crossed with C. canephoru and C. 
eugenioides. 
(5) C. eugenioides performs in general much better in crosses with all 
other diploid taxa than C. canephora. This difference is particularly evi- 
dent in crosses with species of the section Mascarocofea (Figure 2.8) 
(6) Intergeneric hybridisation between Coffea and Psilanrhus has been 
unsuccessful so far. 
All diploid species of the genus Coflea have maintained a similar 
chromosomal structure, which would arise from the same basic A genome 
(monophyletic origin). Charrier (1978) observed a variable rate of PMCs 
Figure 2.8: Genetic Relationships Between Species of Coffea (according to 
number of hybrids per 100 flowers). From Louarn, 1982 (1,2,3); Charrier, 
1978 (4); Carvalho and Monaco, 1968 (5) 
(1) More than 19 hybrids per 100 flowers 
, 
ERlIrUIOCOFFl 
(2) 6-18 hybrids per 100 flowers 
(3) Less than 5 hybrids per 100 flowers 
Corqm.ir 
Llberico 
* - 38 Botunical Classification of Coffee 
* 
’ 
(pollen mother cells) with 11 bivalents in FI hybrids between diploid 
species and found this to be highly correlated with fertility. It appears that 
all diploid coffee species have preserved the identity of their common 
origin during the evolutionary history notwithstanding the geographical iso- 
lation. This lack of chromosomal differentiation was also observed by 
Bouharmont (1959; 1963) in the caryotypes of ten African coffee species. 
An exception is the chromosomal aberrations found in hybrids between C. 
kianjavatensis and C. canephora (Lanaud, 1979). The genetic diversifi- 
cation of the diploid coffee species would therefore result essentially from 
genetic differentiation, which causes various levels of interspecific incon- 
gruity. 
A good example of species which are morphologically and ecologically 
clearly distinct, but show almost normal chromosome pairing and good fer- 
tility in hybrids, are C. canephora (distribution in West and Central Africa) 
and C. congensis (limited to the banks and flooded islands of the Congo 
and Oubangui rivers). The latter species occupies a specific ecotype well 
isolated from the former by a very strict ecological barrier (Berthaud and 
Guillaumet, 1978). In the subsection Pachycoffea the exact position of C. 
fiberica sensu stricto (berries with thick mesocarp) in relation to C. 
dewevrei (berries with thin mesocarp) is not yet clear. On the other hand, 
the abundant number of taxa in the section Mascarocojyea described by 
Leroy often correspond to allopatric populations with distinct morpho- 
logical characteristics but nevertheless without strongly developed crossing 
barriers. Apparently, effective geographical isolation has produced geneti- 
cally divergent populations through genetic drift and natural selection 
pressure. A similar situation could also be found in East African coffee 
species (Leroy, 1982; Hamon, Anthony and Le Pierrks, 1984). From a 
taxonomic point of view some taxa found in East Africa should not be 
classified as different species. 
On the other hand, genetic divergence is in species like C. canephora, C. 
liberica and C. eugenioides exhibited as partial incongruity. For example, 
Louarn (1982) obtained fairly fertile FI hybrids between these three 
species, but their meiotic behaviour was characterised by a reduced pairing 
of the chromosomes (only 40-50 per cent of the PMCs had 11 bivalents). 
In addition there is a genetic effect on the chromosome pairing and the 
hybrid fertility within the interspecific combination C. liberica X C. 
canephora (buarn , 1980). Similar genetic divergence appears to exist also 
between some series of Mascarocoffeu. 
There is a particularly sharp differentiation between C. canephora and 
the section Mascarocofleu: the rare FI hybrids obtained are weak and 
almost completely sterile. This genetic divergence is less pronounced for 
C. eugenioides,since crosses between this species and Mascarocoffea pro- 
duces a higher proportion of fairly fertile FI hybrids. 
It is in fact astonishing that a geographic isolation, started in the Upper 
Botanical Classification of Coffee 39 
Cretaceous, when Madagascar was separated from the main African con- 
tinent, did not have a more profound effect on the chromosomal differen- 
tiation between coffee species. In this respect it is interesting to note the 
affinity of the Malagasy coffee species with East African coffee species 
(Charrier, 1978), in particular the similarity between C. grevei and C. 
rhamnifolia (ex C. paolia) (Leroy, 1980). 
Little work has so far been done on the intergeneric relationship 
between Coffea and Psilanthus. 
Cytogenetic Evidence for the Origin of Coffea arabica 
C. arabica, the only tetraploid species in the genus Coffea, is indigenous to 
the highlands of south-western Ethiopia and south-eastern Sudan. The 
diploid meiotic behaviour and the fact that its centre of genetic diversity is 
situated outside the area of distribution of the diploid coffee species, indi- 
cate an allotetraploid origin (Carvalho, 1952). According to Grassias and 
Kammacher (1975), C. arabica has to be considered as a segmental allo- 
tetraploid. 
C. eugenioides and C. canephora (or C. liberica or C. congensis) have 
often been assumed to be the ancestral parents of C. arabica (Carvalho, 
1952; Cramer, 1957; Narasimhaswamy, 1962). However, meiotic pairing 
of chromosomes of the two genomes in interspecific hybrids of C. 
eugenioides and C. canephora was observed 10 be better (Louarn, 1976) 
than in dihaploid plants of C. arabica (Mendes and Bacchi, 1940; 
Berthaud, 1976). Besides, duplication of such interspecific hybrids (allo- 
diploids) produces tetraploids with autotetraploid meiotic behaviour, such 
as the formation of quadrivalents. Louarn (personal communication) has 
observed this in different interspecific combinations. 
Triploid hybrids, as a result of crosses between C. arabica and diploid 
species, show vigorous growth, but they are almost completely sterile as 
would be expected. Table 2.10 summarises the observations on meiotic 
behaviour of interspecific hybrids between C. arabica and different diploid 
coffee species made by various coffee geneticists. The number of bivalents 
plus trivalents formed during meiosis is, with few exceptions, close to 11. 
This would suggest that one genome of C. arabica has close affinity to the 
genome present in the diploid species, therefore that in the genus Coffea all 
species share the same basic genome and have a monophyletic origin. 
C. arabica could have arisen from natural hybridisation between two 
ancestral diploid coffee species followed by unreduced gamete formation 
(see Demarly, 1975, for a description of the rare events that can lead to the 
occurrence of unreduced gametes). The degree of homology of the two 
genomes could have been high as a consequence of the monophyletic 
origin of the participating species, but various mechanisms (preferential 
pairing, genetic regulation of the synapsis) are likely to have played an 
3 - 40 Botanical Classification of Coffee 
rg 
2 
í i 
O 
II 
U 
C 
m 
m 
Q 
m 
- 
*2 
E! 
G 
C 
al 
d al 
m 
2 
ln 
al 
v) 
v) 
o 
o 
Y- o 
a, 
L1 
al 
C 
O 
3 
O 
> 
m x 
al 
.- 
t? 
4- - 
.+- 
I 
.- 
m 
- e m 1 2 8 1 8 1 I I 1 I I I 
Botanical Classification of Coffee 41 
important role in the progressive diploidisation from the archetype tetra- 
ploid to the present amphidiploid C. arabica. 
Lobreau-Callen and Leroy (1980) observed in their palynological 
studies that C. arabica produces two types of pollen: one type related to C. 
canephora and one closely resembling pollen of C. rhamnifolia, a xero- 
phytic species indigenous to the coastal regions of Somalia and Kenya. This 
would not necessarily contradict the foregoing cytogenetic evidence, but a 
definite conclusion on the origin of C. arabica will have to await further 
studies, including the intergeneric level between Coffea and Psilanthus. 
Biochemical and Serological A ffiniries in Coffea 
Enzyme Polymorphism. Studies on enzyme polymorphism in Coffea 
started with Payne and Fair-Brothers ( 1 9 7 9 who compared total crude 
proteins and malate dehydrogenase in seeds of C. arabica and C. cane- 
phora, and Guedes and Rodrigues (1974), who studied the phenoloxidase 
variability after polyacrylamide electrophoresis of extracts taken from 
genotypes of C. arabica, used as differentials for the identification of 
physiological races of leaf rust (Hemileia vasfafrix). 
Methods of analysing enzyme polymorphism have been further 
developed to study the genetic affinities between various coffee popu- 
lations by Berthou and Trouslot (1977) and Berthou et al. (1980) in Ivory 
Coast. The horizontal starch-gel electrophoresis technique was applied 
initially to three and later to eight enzyme systems and this led to the iden- 
tification of a number of loci. Allozymatic frequencies were estimated for 
natural coffee populations. Estimates of genetic distances between species, 
whereby an index developed by Nei (1972) for 3-8 enzyme systems was 
used, gave the following information: 
(1) C. canephora and C. congensis have the same allozymes with differ- 
ent frequencies; the genetic distance between the species is larger than 
between populations of the same species: the Nana taxon is therefore 
clearly to be considered as part of C. canephora. 
(2) The genetic distance between C. liberica from Ivory Coast and C. 
dewevrei from the Central African Republic is of the same order as that 
between C. liberica and C. humilis from Ivory Coast. This evidence 
justifies a clear distinction between C. liberica and C. dewevrei, which 
both belong to the subsection Pachycoffea. 
(3) The genetic distances between the species C. canephora, C. liberica 
and C. eugenioides are considerable. 4 
(4) The enzymic affinities of C. arabica with all different diploid coffee 
species are similar. According to Berthou and Trouslot (1977) C. 
arabica would result from the complementary electrophoretic bands of 
acid phosphatases and esterases of C. eugenioides and C. canephora or 
C. congensis. 
? . 
= 42 Botanical Classification of Coffee Botanical Classification of Coffee 43 
(5) Lower enzymic affinities have been found between species of the 
genus Coffea and the genus Psilanthus such as Paracoffea ebracteolata. 
.. 
Cytoplasmic DNA. Berthou et al. (1980, 1983) have applied methods of 
identifying DNA of cytoplasmic organelles (chloroplasts and mito- 
chondria) to coffee. Such methods include separation by electrophoresis on 
agarose slab gels of fragments of DNA obtained by bacterial restriction 
enzymes. 
Electrophoresis of fragments of chloroplast DNA, obtained by the Hpa 
II enzyme (from Haemophilus parainfluenzae) suggest that the following 
species have a similar origin: C. arabica and C. eugenioides; C. canephora, 
the Nana taxon and C. congensis 
Electrophoresis of fragments of mitochondrial DNA, obtained by action 
of the Sal I enzyme (from Streptomyces albus), suggested the following 
affinities: (a) great similarity between C. canephora and the Nana taxon; 
(b) great similarity of C. arabica with C. eugenioides and C. congensis; (c) 
considerable divergence between C. canephora and C. arabica or C. 
eugenioides; (d) wide genetic divergence between C. dewevrei and C. 
liberica and Paracoffea ebracteolata 
Evidently, electrophoresis of mitochondrial DNA confirms the distinc- 
tion between the subsections Abyssinicae (C. arabica, C. eugenioides and 
C. congensis) and Robustae ( (C. canephora) earlier proposed by Lebrun 
(1941). The affinity of C. canephora with C. congensis, as indicated by the 
similarity of chloroplast DNA is clear, but these two species show differ- 
ences in their mitochondrial DNA. This type of analysis has also indicated 
that the Nana taxon should be considered as an ecotype of C. canephora. 
SerologicalAffinities. Antisera obtained for C. arabica and C. canephora 
were used in serological reactions with soluble antigens from seed of 
various coffee species (Hofling and Oliveira, 1981). It was possible to 
establish the following relationship: C. arabica has more affinity with C. 
congensis and C. eugenioides than with C. canephora. 
Other Chemical Affinities. Chromatographic analysis of the flavonoid 
component (Longo, 1975) indicated that Psilanthopsis kapakata belongs 
to the genus Coffea rather than Psilanthus as earlier proposed by 
Chevalier. 
Additional data, based on simple chemical tests applied by Ram, 
Sreenivason and Vishveshwara (1982), suggest that C. eugenioides, C. 
rucemosa and P. kapakata are more closely related to each other than to C. 
salvatrix. Nevertheless, all four species are rightly classified under the sub- 
section Mozambicoffea (see also Chapter 13: 334). 
Conclusions 
From this review of advances in the taxonomy and botany of coffee the fol- 
lowing main conclusions can be drawn. 
(1) A general revision of coffee taxonomy should take into account the 
new insights into the relationships of coffee species, acquired by the 
extensive exploration of natural habitats and centres of high genetic 
diversity, as well as by the application of modern concepts of the bio- 
logical series and of biosystematic classification systems. 
(2) During the evolutionary organisation of the coffee gene pool differ- 
entiation into species was not accompanied by the development of 
strong crossing barriers. This offers considerable prospects for intro- 
gressing desirable characters into cultivated species or the development 
of new cultivars by interspecific hybridisation. Of particular interest are 
the taxonomic affinity of C. congensis and C. canephora (Congusta 
hybrids) and the crossability of C. eugenioides with many other diploid 
species. There are also indications that C. eugenioides, C. congensis, C. 
canephora and the allotetraploid C. arabica have originated from 
common ancestral forms. 
(3) There is a great urgency for intensifying studies of the still existent 
wild coffee populations, especially in those geographic areas where such 
populations are most threatened by extinction. The preservation of this 
coffee germplasm should be secured, either by conservation of the 
natural forest habitats or by establishing living collections. The following 
regions require special attention in this respect: 
- The highlands of south-western Ethiopia and south-eastern Sudan 
(Boma Plateau): the centre of high genetic diversity for C. arabica; 
- Uganda and Central Africa (Gabon, Congo, Zaire): the major areas 
of distribution for many diploid coffee species; 
- The Malagasy Republic: the centre of genetic diversity for species of 
the section Mascarocoffeu. 
References 
Berthaud, J. (1976) ‘Etude cytogdnttique d’un haploide de Coffea arabica L‘ Café Cacao 
Berthaud, J. (1980) ‘L’incompatibilitk chez Coffeu canephora: mtthode de test et deter- 
Berthaud, J. (1985) ‘Les Ressources Gtnttiques Chez les Caftiers Africains: Populations 
Thé, 2 4 91-6 
minisme genttique’, Café Cacao Thé, 24, 267-74 
Sylvestres, Bchanges Gknttiques et Mise en Culture’ Doctoral Thesis, University of Paris 
XI, Orsay (in press). 
Berthaud, J. , Guillaumet, J.L , Le Pieds , D. and Lourd, M. (1977) ‘Les prospections des 
caftiers sauvages et leur mise en collection’ in 8th International Colloquium on the 
Chemistry of Coffee, ASIC, Paris, pp.365-72 
44 Botanical Classification of Coffee 
’ Berthaud, J. and Guillaumet, J.L., (1978) ‘Les caféiers sauvages en Centrafrique’ Cafe Cacao 
Berthaud, J., Guillaumet, J.L., Le Pierrès, D. and Lourd, M. (1980) ‘Les caféiers sauvages du 
Kenya: prospection et mise en culture’, Cafe Cacao Thé, 24, 101-12 
Benhou, F. (1975) ‘Méthode d’obtention de polyploides dans le genre Coffea par traitements 
localisés de bourgeons à la colchicine’, Cafe Cacao Thé, 19, 197-202 
Berthou, F. and Trouslot, P. (1977) ‘L‘analyse du polymorphisme enzymatique dans le genre 
Coffea: adaptation d’une méthode d’électrophorèse en skrie: premiers résultats’, in 8th 
International Colloquium on the Chemistry of Coffee, ASIC, Paris, pp. 373-83. 
Berthou, F., Trouslot, P., Hamon, S., Vedel, F. and Quetier, F. (1980) ‘Analyse en électro- 
phorèse du polymorphisme biochimique des caféiers: variation enzymatique dans dix-huit 
populations sauvages: C. canephora, C. eugenioides et C. arabica‘ Cafe Cacao Thé, 24, 
Berthou, F., Mathieu, C. and Vedel, F. (1983) ‘Chloroplast and mitochondrial DNA vari- 
ation as indicator of phylogenetic relationships in the genus Coffea L‘ Theoretical and 
Applied Genetics, 65, 77-84 
Bullerin de la Societe de Pharmacie, 5, 283-5 
Publication INEA C, Série Scientifique, 77 
- > The, 22, 171-86 
313-26 
Bertrand, G . (1902) ‘Recherche et dosage de la caféine dans plusieurs espèces de Cafe 
Bouharmont, J. (1959) ‘Recherches sur les affinités chromosomiques dans le genre Coffea’, 
Bouharmont, J. (1963) ‘Somatic chromosomes of some Coffea species’, Euphytica, 12, 254. 
Bridson, D. (1982) ‘Studies in Coffea and Psilanihm (Rubiaceae sub fam. Cinchonoideae\ 
for part 2 of the ’Flora of Tropka1 East Africa: Ribiaceae‘, Kew Bulletin (Royal Botkic 
Garden), 36, 817-59 
Capot, J. (i972) ‘L‘amélioration du Caféier en Côte d‘Ivoire. Les hybrides ‘arabusta’, Café 
Carvalho, A. (1952) ‘Taxonomia de Coffei arabica L. Caractères morphologicos dos 
Carvalho, A. and Krug, C.A. (1949) ‘Genetica de Coffea XII. Hereditariedade da côr 
Cuca0 Thé, 16,3-18 
’ haploides’, Bragantia, 12, 201-12 
amarela da semente’. Erapantia. 9. 193-202 , , ~ ~~~ 
Carvalho, A. and Monaco, LYC. (1968) ‘Relaciones geneticas de especies selectionadas de 
Coffee’, Café, 4,3-19 
Carvalho, A., Fenverda, F.P., Frahm-Leliveld, J.A., Medina, D.M., Mendes, A.J.T. and 
Monaco, L.C. (1969) ‘Coffee (Coffea arabica L. and C. canephora Pierre ex Froehner)’ in 
F.P. Ferwerda and F. Wit (eds). Outlines of Perennial Crop Breeding in the Tropics, 
Veenman & Zonen, Wageningen, pp. 189-241 
Charrier, A. (1978) ‘La srructure génétique des caféiers spontanés de la région Malgache 
(Maxarocoffea). Leurs relations avec les caféiers d’origine africaine (Eucoffea)’, 
Mèmoires ORSTOM (Paris), no. 87 
Charrier, A. (1980) ‘La conservation des ressources génétiques du genre Coffea’, Cafe Cacao 
Charrier, A. (1982) ‘L‘amélioration génétique des cafés’, La Recherche, 13 (136), 1006-16 
Chevalier, A. (1938) ‘Essai d’un groupement systkmatique des caféiers sauvages de 
Chevalier, A. (1947) ‘Les caféiers du globe 111. Systématique des caféiers et faux caféiers. 
The, 24, 249-57 
Madagascar et des îles Mascareignes’, Revue Botanique Appliquee et d’Agriculture Tropi- 
rule, 18, 825-43 
Maladies et insectes nuisibles’, Encyclopedie biologique 28, Fascicule 111, P. Lechevalier, 
Paris 
Chinnappa, C.C. (1968) ‘Interspecific hybrids of Coffea canephora and C. arabica‘, Current 
Science. 37. 676-7 
Chinnappa; C.C. (1970) ‘Interspecific hybrids of C. canephora and C. liberica’, Genetica, 41, 
Chinnappa, C.C. and Warner, B.G. (1981) ‘Pollen morphology in the genus Coffea 
141-6 
(Rubiaceae) and its taxonomic significance’, Botanical Journal of the Linnean Society, 83, 
221-36 
Clausen, D., Keck, D. and Hiesey, W.M. (1945) ‘Experimental studies on the nature of 
species II. Plant evolution through amphiploidy and autoploidy, with examples from 
Maidinae’, Carnegie Institute, Washington, Publ. no. 564 
Botanical Classification of Coffee 45 
Coste, R. (1968) Le Cafeier, G.P. Maisonneuve & Larose, Paris 
Couturon, E. and Berthaud, J. (1982) ‘Presentation d’une methode de récupération 
d’haploïdes spontanes et d’obtention de plantes diploides homozygotes chez C. cane- 
phora’, in 10th International Colloquium on the Chemistry of Coffee, ASIC, Paris, 
pp. 385-91 
Cramer, P.J.S. (1957) in F.L. Wellman, Review of Literature of Coffee Research in Indonesia. 
SIC Editorial lnternational American Institute of AgriculturalSciences, Turrialba, Costa 
Rica 
Demarly, Y. (1975) ‘Amélioration du caféier liée aux progrès génétique. in 7th International 
Devreux, M., Vallayes, G., Pochet, P. and Gilles, A. (1959) ‘Recherches sur I’autostérilité du 
Dobshansky, T. (1970) Genetics of Evolutionary Process, Colombia University Press, New 
Colloquium on the Chemistry of Coffee, ASIC, Pans, pp. 423-35 k 
caféier robusta ( C . canephora Pierre)’, Publication INEAC Serie Scientifique, 78 
York 
in der Familie Rubiaceae nebst Bemerkungen uber einige Polyploiditat problem. Acta 
Horticulturae, I l , 195-470 
Forster, R.B. (1980) ‘Heterogeneity and disturbance in tropical vegetation’ in M.E. Soule and 
B.A. Wilcox (eds.), Conservation Biology, Sinauer Associates Inc., Sunderland, pp. 75-92 
Friedman, F. (1970) Etude biogéographique de Coffea buxifolia Chev.’, Cafe Cacao The, 14, 
Friis, I.B. (1979) ‘The wild populations of Coffea arabica L., and cultivated coffee’, in G . 
,d#, ,., Fagerlind, F.O. (1937) ‘Embryologische, zytologixhe und bestaubungsexperinientelle Studien 
i+, 
3-12 
Kunkel (ed.). Taxonomic Aspects of African Economic Botany, Les Palmas de Gran 
Canaria. Ayuntamiento de las Palmas. pp. 63-8 
Grassias, M. and Kammacher, P. (1975) ‘Observations sur la conjugaison chromosomique de 
Coffea arabica L.’, Cafe Cacao Th4 19, 177-90 
Guedes, M.E.M. and Rodrigues, C.J. Jr. (1974) Disc electrophoretic patterns of 
phenoloxidase from leaves of coffee cultivars’, Portugaliae Acta Biologica, 13, 169-78 
Guillaumet, J.L. and Hallé, F. (1978) ‘Echantillonnage du materiel Coffea arabica recolté en 
Ethiopie’ in A. Charrier (ed.) Etude de la Structure et de la Variabilitè Genetique des 
Cafeiers, IFCC(Paris) Bulletin no 14, pp. 13-18 
Haarer, A.E. (1962) Modern Coffee Production. Leonard Hill, London 
Hamon, S., Anthony, F. and Le Pieds, D. (1984) ‘Etude de la variabilité génétique des 
caféiers spontanés de la section des Mozambicoffea Coffea zanguebariae Bridson et C. sp. 
novo Bridson’, Adamsonia, in press 
Taxon, 20,509-17 
with emphasis on the origin of C. arabica’, Ciencia e Cultura, 33, 66-72 
C. canephora’, Cafe Cacao The, 16, 289-94 
Harlan, J.R. and de Wet M.J. (1971) ‘Towards a rational classification of cultivated plants’, 
Hofling, J.F. and Oliveira, A.R. (1981) ‘A serological investigation of some Coffea species 
Kammacher, P. and Capot, J. (1972) ‘Sur les relations caryologique entre Coffea arabica and 
Krug, C.A. (1965) ‘Ensayo mundial del cafe’ FAO, Rome 
Krug, C.A. and Mendes, A.J.T. (1940) ‘Cytological observations in Coffea‘, Journal of 
Genetics, 39, 189-203 
Krug, C.A. and Carvalho, A. (195 1) ‘The genetics of coffee’, Advances in Genetics, 4, 127- 
68 
Lanaud, C. (1979) ‘Etude de Problemes gtnétiques posés chez le caféier par I’introgression 
de caractères d’une espèce sauvage ( C. kianjavatensis: Mascarocoffea) dans l‘espèce 
cultivée C. canephora: Eucoffea’, Cafe Cacao Thé, 23, 3-28 
Lebrun, J. (1941) Recherches morphologiques et systematiques sur les caféiers du Congo, 
Memoires TXI (3), Institute Royal Colonial Belge, Bruxelles 
Leliveld, J.A. (1940). ‘Nieuwe gezichtspunten voor het selectie ondermek bij 
soortskruisingen naar aanleiding van cytologische gegevens’, Bergcultures, 14, 370-86 
Leroy, J.F. (19614 ‘Coffea novae madagascariensis’, Journal d’Agriculture Tropicale et de 
BotaniqueAppliquee, 18, (1, 2, 3), 1-20 
Leroy, J.F. (1961b) ‘Sur les trois caféiers endemiques de l’archipel des Mascareignes’, Journal 
d’Agriculture Tropicale et de Botanique Appliquee, 8, ( 1 , 2, 3), 21-9 
Leroy, J.F. ( 1 9 6 1 ~ ) ‘Sur deux caféiers remarquables de la forêt sèche du Sud-Ouest de 
~ 
” Z 
rr&i I 
d i 
46 Botanical Classification of Coffee 
1 
._. Madagascar (C. humbertii J.F. Ler, et C. capuronii J.F. Ler.)’, Comptes Rendus de 
I’Academie des Sciences, Paris, 252,2285-7 
Leroy, J.F. (1962) ‘Coffeae novae madagascariensis et mauritianae’, Journal d’Agriculture 
Tropicale er de Botanique Appliquee, 9 (11, 12), 525-30 
Leroy, J.F. (1963) ‘Sur les cafkiers sauvages des îles Mascareignes’, Comptes Rendus de 
I’Academie des Sciences, Paris, 256, 2897-9 
sur l’origine et l’aire du genre Coffed, Comptes Rendus de I’Academie des Sciences, Paris, 
Leroy, J.F. (1967) ‘Recherches sur les cafkiers. Sur la classification biologique des caftiers et I 
265, 1043-5 
Leroy, J.F. (1968) Notice sur les titres et travaux scientifiques de J.F. Leroy Monnoyer, Le 
Leroy, J.F. (1972a) ‘Prospection des cafkiers sauvages de Madagascar: deux espèces 
Leroy, J.F. (1972b) ‘Prospection des cafkiers sauvages de Madagascar: sur deux esphes 
Leroy, J.F. (1980) ‘Evolution et taxogenèse chez les cafkiers. Hypothese sur leur origine’, 
Leroy, J.F. (1982) ‘L‘origine kenyane du genre Coffea et la radiation des espkces a 
Leroy, J.F. and Plu, A. (1966) ‘Sur les nombres chromosomiques des Cogea malgaches’, 
Mans fascicule 11 
remarquables (Coffea tsirananaen. sp., C. kianjavatensisn. sp.)’, Adamsonia, 2 (12), 317- 
28 
sympatriques du Nord’, Adamsonia, 2 (12), 345-58 . 
Comptes Rendus de I’Academie des Sciences, Paris, 291,593-6 
Madagascar’, in 10th International Colloquium on the Chemistry of Coffee, ASIC, Paris, 
pp.413-20 
Cafe Cacao The, I O (3), 216-18 
Lind, E.M. and Morrison, M.E.S. (1974) East African Vegetation. Longman, London i 
I Linnaeus, C. (1737) Hortus Cliffortianus. Amsterdam Lobreau-Callen, D. and Leroy, J.F. (1980) ‘Quelques donntes palynologiques sur le genre 
Coffeu et autres genres du cercle des cafeiers’, in 9th International Colloquium on the 
chemistry of Coffee, ASIC, Paris, pp. 479-506 
Longo, C.R.L. (1975) ‘The use of phenolic pigments in phylogenetic studies on Coged, Ba- 
letin de Divulgação, Escola Superior de Agricultura “Luis de Queiroz”, Sao Paulo, 20, 
dtterminations de leurs nombres chromosomiques’, Cafe Cacao The, 16,312-15 
Louam, J. (1976) ‘Hybrides interspecifiques entre Coffea canephora Pierre et C. eugenioides 
Moore’, Cafe Cacao The, 20,433-52 
Louam, J. (1978) ‘Diversitd compark des descendances de C. arabica obtenues en 
autoftcondation et en fhndat ion libre au Tonkoui’, in A. Charrier (ed.), Etude de la 
Structure et de la Variabilité Genetique des Cafèiers. IFCC Bulletin no 14, pp. 75-8 
Louarn, J. (1980) ‘Hybrides interspkcifiques entre Coffea canephora et C. libericu Rtsultats 
préliminaires sur les hybrides F,, Cufé Cacao Thé, 24, 297-304 
Louam, J. (1982) Bilan des hybridations interspkcifiques entre caftiers africains diploides en 
collection en Côte d’Ivoire’, in 10th International Colloquium on the Chemistry of Coffee, 
Mayr, E. (1970) Populations, Species and Evolution. Belknap Press of Harvard University 
Press, Cambridge, Mass 
Medina, D.M. (1963) ‘Microsporogenest em um hibrido triploide de C. racemosa X C. 
arabica’, Bragantia, 22, 299-318 
Mendes, A.J.T. (1938) ‘Morfologia dos cromosomios de Coffea excelsd, Instituto 
Agronomico de Campinas, Boletin Tecnico, 56 
Mendes, A.J.T. (1939) ‘Duplicapo do numero de cromosomios em cafe, algodao e fumo, 
pela açao da colchicina’, Instituto Agronomico de Campinas, Bobin Tecnico, 57 
Mendes, A.J.T. and Bacchi, O. (1940) ‘Observaçoes citologicas em Coffea V. Uma variedade 
haploide (di-haploide) de C arabica L.’ lmtituto Agronomico de Campinus, Boletin 
Tecnico, 7 7 
Meyer, F.G. (1965) ‘Notes on wild Coffea arabica from southwestem Ethiopia, with some 
historical considerations’, Economic Botany, 19, 136-51 
Meyer, F.G., Fernie, L.M., Narasimhaswamy, R.L., Monaco, L.C. and Greathead, D.J. 
(1 968) FA O Coffee Mission to Ethiopia 1964-65. FAO, Rdme 
~ 171-4, 392-5. 
Louam, J. (1972) ‘Introduction à l’ktude gtnétique des Mriscarocoffea Nouvelles I 
ASIC. Paris, pp. 375-84 
Botanical Classification of Coffee 47 
Monaco, L.C. and Medina, D.M. (1965) ‘Hibridaçoes entre Coffea arabicu e C. kapokata. 
Narasimhaswamy, R.L. (1962) ‘Some thoughts on the origin of C. urabica L. Coffee 
Nei, M. (1972) Genetic disrance between populations. Americun Naturulist, 106, 283-92 
Noirot,M. (1978) Polyploidisation de cafeiers par la colchicine. Adaptation de la technique 
sur bourgeons axillaires aux conditions de Madagascar. Mise en tvidence de chimères 
périclines stables. Cufe, Cacao, The, 22, 187-94 
Payne, R.C. and Fair-Brothers, D.E. (1976) ‘Disc electrophoresic evidence for heterozygosity 
and phenotypic plasticity in selected lines of Coffea arabica L.’ Botanical Gazette, 137, 
1-6 
Portères, R. (1937) ‘Etude sur les caféiers spontanés de la section Eucoffea’, Annales 
agricoles de IXfrique occidentale et Etrangères. lkre partie: Repartition, habitat, I (I), 
68-91: 2ème partie: Espkces, varietes, formes, I (2), 219-63 
Portères, R. ( 1962) ‘Sur quelques cafkiers sauvages de Madagascar’, Journal d’Agriculture 
Tropicale et de la Botuniqice Appliquee, 9 (3-6), 201-10 
Ram, AS., Sreenivasan, M.S. and Vishveshwara, S. (1982) ‘Chemotaxonomy of 
Mozambicoffea’, Journal of Coffee Research (Chikmagalur), 12, 42-6 
Reynier, J.F., Pernes, J. and Chaume, R. (1978) ‘Diversitt observte sur les descendances 
issues de pollinisation libre au Tonkoui’, in A. Charrier (ed.). Etude de la structure et de 
la variabilitt genttique des cafdiers. IFCC Bulletin no. 14, pp. 69-74 
Sybenga, J. (1960) ‘Genetics and cytology of coffee: a literature review’, Bibliograbhia 
Genetica, 19, 217-316 
Sylvain, P.G. (1955). ‘Some observations on Coffea urabica L., in Ethiopia’, Turrialba (Costa 
Rica), 5,37-53 
Thomas, A.S. (1942) ‘The wild Arabica coffee on the Boma plateau, Anglo-Egyptian Sudan’, 
Empire Journul of Experimenrul Agriculre, I 0, 207- 12 
Thomas, AS. (1944) ‘The wild coffees of Uganda’, Empire Journal of Experimental 
Agriculture, 12, 1- 12 
Van der Vossen, H.A.M. (1974) ‘Plant Breeding’, Coffee Reseurch Foundation of Kenya, 
Annual Report 1973-74, pp.40-51 
Analise citologica de um hibrido triploide’, Bragantiu, 24, 19 1-201 
(Turrialba)’, 4, 1-5 
Vishveshwara,-S. (1975) ‘Coifie: commercial qualities of some newer selections’, Indian 
Coffee, 39. 363-74 
Von Siienke, H. (1956) ‘Wild coffee in Kaifa province of Ethiopia’, Tropical Agriculture, 38, 
Wellman, F.L. (1961) Coffee: Botuny, Cultivation and Utilization. Leonard Hill Books, 
297-301 
London 
Botany, Biochemistry and 
Production of Beans and Beverage 
EDITED BY M.N. CLIFFORD AND K.C. WILLSON 
CROÖM HELM 
London & Sydney 
avr 
oßsnm ~ o n d s Documentaire 
cote I: a'&, 
AMERICAN EDITION 
Published by 
THE AVI PUBLISHING COMPANY, INC. g 3 2 , /l--c?n,.a ) !~d 
Westport, Connecticut 1985 > -