Fruit-frugivore interactions in a Malagasy littoral forest - Universiteit ...

Fruit-frugivore interactions in a Malagasy littoral forest: a community-wide approach of seed dispersal An BOLLEN 

Fruit-frugivore interactions 

in a Malagasy littoral forest: 

a community-wide approach 

of seed dispersal 

Proefschrift voorgelegd tot het behalen van 

de graad van doctor in de Wetenschappen 

aan de Universiteit Antwerpen te verdedigen door 

An BOLLEN 

Promotoren: Prof. dr. L. Van Elsacker 

Prof. dr. R. Verheyen 

Prof. dr. J. Ganzhorn 

UNIVERSITEIT ANTWERPEN 

Faculteit Wetenschappen 

Departement Biologie 

ANTWERPEN 2003

Universiteit Antwerpen 

Faculteit Wetenschappen 

Departement Biologie 

Promotoren: Prof. dr. L. Van Elsacker 

Prof dr. R. Verheyen 

Prof dr. J. Ganzhorn 

Antwerpen, 2003 

Fruit-frugivore interactions in a Malagasy 

littoral forest: a community-wide 

approach of seed dispersal 

Interacties tussen vruchten en frugivoren 

in een Madagaskisch littoraal regenwoud: 

een benadering van zaadverspreiding 

op niveau van het ecosysteem 

Proefschrift voorgelegd tot het behalen van de graad van doctor in de Wetenschappen 

aan de Universiteit Antwerpen te verdedigen door 

An BOLLEN

Cover photographs: 

Front: 

Back: 

-‘ambora’ Tambourissa purpurea (Monimiaceae) © An Bollen 

-‘bodono’ Cheirogaleus medius (Cheirogaleidae) © Han Remaut 

-‘hazombato’ Campylospermum obtusifolium (Ochnaceae) © An Bollen 

-‘horovana’ Hypsipetes madagascariensis (Pycnonotidae) © Han Remaut 

-‘tsilantria’ Vaccinium emirnense (Ericaceae) © An Bollen 

-‘varika’ a male Eulemur fulvus collaris (Lemuridae) © An Bollen 

-aerial photograph of the Sainte Luce area (middle) © An Bollen 

-‘sanirambavy’ fruits of Tina thouarsiana (Sapindaceae) with bill traces of 

Coracopsis nigra (Psittacidae) © An Bollen 

-‘fanihy’ Pteropus rufus (Pteropodidae) © An Bollen 

-‘rotry mena’ Syzygium sp.2’ (Myrtaceae) © An Bollen 

-seeds of ‘voapaky vavy’ Uapaca littoralis (Euphorbiaceae) with gnawing marks 

of ‘voalava’ rodents © An Bollen 

-Eulemur fulvus collaris feeding in a fruiting tree © An Bollen 

-‘sjihely’ sardine-like fishes in Dec ‘00 in the village of Ambandriky © An Bollen 

-the people of the Ambandriky at the landing strip of Martin Pêcheur in front of the 

hill of S17 © An Bollen 

Throughout this PhD taxonomic names of the plant species follow the Flora of Madagascar 

(Humbert 1936-1984), the data base of Missouri Botanical Garden (http://mobot.mobot.org/W3T) 

ISBN 90-5728-042-6 

© An Bollen

TABLE OF CONTENTS 

Table of contents 

Acknowledgements ........................................................................................................1 

Preface ...........................................................................................................................5 

I. General Introduction..................................................................................................7 

II. Main Chapters ......................................................................................................... 13 

1. Phenology of the littoral forest of Sainte Luce, south-east Madagascar.............. 15 

Bollen A, Donati G 

Submitted to Biotropica 

2. Tree dispersal strategies in the littoral forest of Sainte Luce .............................. 37 

(south-east Madagascar) 

Bollen A, Van Elsacker L, Ganzhorn JU 

Submitted to Oecologia 

3. Relations between fruits and disperser assemblages in the littoral .................... 59 

forest of south-east Madagascar: a community level approach 

Bollen A, Van Elsacker L, Ganzhorn JU 

Submitted to Journal of Tropical Ecology 

3a. Feeding ecology of Pteropus rufus (Pteropodidae) ................................... 85 

in the littoral forest of Sainte Luce (south-east Madagascar) 

Bollen A, Van Elsacker L 

Acta Chiropterologica 4(1): 33-47 

3b. Feeding ecology of Coracopsis nigra (Psittacidae) in ............................ 101 

the littoral forest (south-east Madagascar). 

Bollen A, Van Elsacker L 

Submitted to Ostrich 

4. Fruit characteristics in a dry deciduous and a humid littoral forest................... 111 

of Madagascar: evidence for selection pressure through abiotic 

constraints rather than through co-evolution with lemurs as seed dispersers 

Bollen A, Donati G, Fietz J, Schwab D, Ramanamanjato J-B, 

Randrihasipara L, Van Elsacker L, Ganzhorn JU 

In: Fruits and frugivores, L Dew and JP Boubli (Eds.) 

In Press. Kluwer Academic Press, Dordrecht 

5. The Malagasy littoral forest: threats and possible solutions ............................ 133 

Bollen A, Donati G

Table of contents 

III. General conclusion.............................................................................................. 147 

Samenvatting.............................................................................................................. 153 

Sommaire ................................................................................................................... 159 

References ................................................................................................................ 165

ACKNOWLEDGEMENTS 

Acknowledgements 

Many people have provided me with assistance, guidance and support during the time it 

has taken me to complete this PhD for which I am very grateful: 

First of all I want to thank my supervisor, Linda Van Elsacker, who has supported me 

from the beginning of this PhD, never losing faith. Thanks for letting me introduce not only 

lemurs but also bats, birds and rats within this primatological group and for always being 

ready to read my manuscripts over and over again, to exchange ideas and to motivate 

me to present research to the larger public. Sincere thanks as well to Rudi Verheyen who 

was ready to act as my co-promoter and who has encouraged me throughout the PhD. 

It was Jörg Ganzhorn who got me to Madagascar in the first place. He helped me 

greatly during field research, in the analyses and written part of this PhD. Many of the 

interpretations presented here were developed in discussion with him. His impressive list 

of publications and unique expertise in tropical ecology have been an ideal guide for me 

during these four years. Jörg, thank you very much for the appreciated help and all the 

time you put into this work while not really having any spare time. 

I would also like to express my gratitude to all my CRC colleagues on whom I could 

count for correcting ‘rough’ drafts and providing useful suggestions. In particular I would 

like to thank Catherine Roden, Jeroen Stevens, Adinda Sannen, Rebekka Deleu, Isra 

Deblauwe, Hilde Vervaecke and Kristin Leus. My special thanks to Adinda who helped 

me greatly in the final phase, guiding me through all the official procedures and lay-out 

options. Many thanks as well to Vera Cuypers and Paule Stevens who made all the 

administrative paperwork a lot easier. 

Language-wise I was very lucky to get help from Eric Arnhem who was ready to 

correct the French summary of this PhD. As for English, it was wonderful to work together 

with Don and Val Sibley who closely scanned my drafts for awkward phrases, 

grammatical errors and other mistakes. Previous versions of the chapters benefited 

greatly from the comments of Steve Goodman and Paul Adrian Racey, for which many 

thanks. 

I am very grateful to Irene Tomaschewski from the Department of Conservation and 

Ecology (University of Hamburg) for carrying out biochemical analyses on the fruits. 

Thanks to the statistical advice of Dries Bonte and Stefaan Van Dongen SAS became 

a little less incomprehensible to me. They put me back on the right track when I was lost 

in the complex world of statistical tests and models. My sincere thanks for your help! 

I would further like to thank the members of the jury who have taken time to read 

through this PhD, I am looking forward to your comments. In this respect I am very 

grateful for the useful suggestions and constructive criticism of Prof Luc Lens on the first 

version of this PhD. 

Obviously this PhD would never have been possible in the first place without the 

funding of the FWO (Funds for Scientific Research, Flanders), for which many thanks. I 

further thank the Centre for Research and Conservation (Antwerp Zoo) and the 

Department of Biology (University of Antwerp) as well as DOCOP (University of Antwerp) 

for additional funding. 

1


A special word of thanks to the whole team of QMM in Fort-Dauphin for welcoming 

me at their study site and for helping me out with transport, radio-contact and food. 

Thanks to Rufen for the ‘architectural’ innovations at the campsite, which greatly 

improved our living standards. Laurent Randrihasipara and Jean-Baptiste 

Ramanamanjato, thanks for putting up with ‘la belge exigeante’ and Manon Vincelette, 

thank you for always taking time to answer my numerous questions. 

Having a degree in Zoology, it was great to be surrounded by excellent botanists 

familiar with the Malagasy flora. Petra De Block was the first to introduce me to the 

existing literature on taxonomy of Malagasy plants. Johny Rabenantoandro and Faly 

Mbolatiana from the Missouri Botanical Garden managed to replace most ‘vernacular’ 

names into scientific ones, which is not at all simple! Thanks for your patience in trying to 

guide me through the herbaria at Tana. 

Prof. Berthe Rakotosamimanana (Dept. Anthropologie) and Prof Daniel 

Rakotondravony (Dept Biologie Animale) of the University of Antananarivo helped me to 

obtain research permits and saved me from weeks of administrative hassle, for which my 

sincere thanks. I further acknowledge ANGAP, the Ministère des Eaux et Forêts and the 

Commission Tripartite of the Malagasy government for their collaboration and permission 

to work in Madagascar. 

The great help of my field assistants was essential in my field research. Thanks to 

them I managed to recognize many different tree species and find my way in the forest. In 

particular I would like to acknowledge Sambo Givet and ‘Ramanga’ (Honoré Aly) with 

whom I spend many hours in the forest. It was an enriching experience and unforgettable 

cooperation in ‘équipe fruit’. I am very grateful as well to Josette, our cook, for spoiling us 

daily with fresh sisika 1 and rano apango 2 and to Kadoffe for his work in phenology and his 

companionship. Ramisi Edmond, our local plant specialist, helped me tremendously in 

identifying and recognizing plant species. Many other Antanosy helped me in the field 

(Julson, Bruno, Claude, Rivo, …). Misaotra! 3 I would further like to express my 

appreciation for the assistance and goodwill of the people of Ambandrika, 

Ampanosatomboky et Manafiafy. They always welcomed me and I have great respect for 

the Présidents des villages (Olaf Abel Isaia, Leonnaise, Sambo Pierre, Milson) who were 

always ready to discuss about conservation issues. 

During my field research numerous Malagasy and foreign students joined me at the 

campsite for shorter or longer periods. They contributed a lot to a great and stimulating 

working environment. Thanks for sharing all those wonderful moments Jack, Tolona, 

Andreas, Nicoletta, Valentina, Rick, Petra, Kaï, Odon and Jose. 

Closer to home I would like to acknowledge my friends and family who supported me 

throughout this PhD in different ways, for which my sincere thanks. The risk of forgetting 

someone is too big, which is why I have not tried to list everyone here but I am sure they 

know I deeply appreciate their support! 

My parents have been a great support to me for as long as I can remember. They 

always have been very open-minded and accepted my wish to get out into the world and 

explore far away places. Thanks for being there for me unconditionally and for 

encouraging me to follow my ‘idealistic’ dreams. My brother always supports me too and I 

am happy to see how close we still are even though both of us spend a lot of time at 

different places on the globe. 

2


Finally there is someone special to whom I am most grateful. Giuseppe Donati has 

been of crucial importance to me during this PhD. How could I ever forget the year we 

spent together in the forest, my neighbour from the tent next-door; the long nights talking 

about our exciting findings in the forest; endless discussions trying hard to find solutions 

to the conservation issues; Sanzinia 4 at breakfast; Italian salami for Belgian chocolates. 

We shared both great and tough moments. Too much to describe here but I am sure he 

knows what I am talking about! Misaotra betsaka zoky be an ala!!! 3 

Then of course there is the forest with all its mysteries and wonders that I was able to 

explore a little more each day. It is a precious place and I feel very fortunate to have been 

able to work out there. I hope this unique littoral forest of Sainte Luce will survive for 

many more generations to come… 

Misaotra betsaka! 3 Antwerp, 25 October ‘03 

1 shrimps, 2 rice water, 3 Thanks!, 4 Sanzinia madagascariensis : boa 

3

PREFACE 

Preface 

This work brings together the main data collected during a 14-month field research in the 

littoral forest of Sainte Luce, south-east Madagascar. Nearly all chapters have been 

written to be suitable for publication and can as such be read independently from each 

other. Overlap in the introduction, materials and methods sections of the different papers 

was therefore unavoidable. In the different chapters distinct sub-samples from the fruit 

database and phenological data were used depending on the research question under 

consideration. 

The work can be subdivided into three major sections: the introduction, the main section 

containing five chapters and the conclusion. The introduction situates the main topic of 

this work, describes the study site and the goal of this study. It further gives the outline of 

the thesis. At the time of compiling this thesis one chapter has been published (Chapter 

3a), one is in course of publication (Chapter 4) and four are being reviewed (Chapter 1, 2, 

3, 3b). Chapter 5 gives an overview of the conservation issues in the littoral forest and 

outlines possible applications, which will be used for conservation suggestions in a later 

publication. Finally the conclusion links all chapters together, summarizing the most 

relevant findings of this study. 

Unfortunately the four year period was too limited to integrate all collected data into this 

work. It will therefore not finish with this doctoral study. Other parts of my work such as 

the more experimental preliminary work on seed predation and germination have been 

presented during conferences in the form of oral and poster contributions: the British 

Ecological Society (BES) meeting in Reading in 2001 (Bollen and Van Elsacker 2001), 

the Association for Tropical Biology and Conservation (ATBC) meeting in Panama 2002 

(Bollen et al. 2002; Bollen and Van Elsacker 2002b), the ATBC meeting in Aberdeen 

2003 (Bollen and Van Elsacker 2003) and the Benelux Congress of Zoology in Antwerp 

(Bollen and Van Elsacker 2002c). A lecture given at the University of professional 

education Larenstein in the Netherlands as part of a course on Tropical Ecology 

completes this list. As I work with the Centre for Research and Conservation, which is 

based at the Antwerp Zoo, there were several occasions on which written and oral media 

on this seed dispersal project were presented to a larger audience (Zoo Magazine, 

Course Primatology, Op de Koffie, Save the Bonobos, Africamatters). 

5

GENERAL INTRODUCTION 

General introduction 

ZOOCHORY AND FRUGIVORY 

Scientists and evolutionary biologists as early as Darwin (1859), Wallace (1879) and 

Kerner (1898) acknowledged the importance of seed dispersal. The number of studies on 

dispersal ecology has only increased substantially during the last three decades. There 

was a great interest in understanding the role played by frugivory and seed dispersal in 

the dynamics of forests, particularly those in tropical forests. Van der Pijl (1969) was the 

first to give an elaborate survey of the modes of seed dispersal. This was based on the 

classification of Ridley (1930), who defined the agent of transport as the criterion for the 

main dispersal classes. Seeds are not mobile themselves, so their movement must be 

effected by dispersal vectors, whether abiotic or biotic. The main classes of seed 

dispersal are autochory, anemochory, hydrochory and zoochory, which respectively 

means seed dispersal by the plant itself, by wind, water or animals (Van der Pijl 1969). 

Each of these categories can be further subdivided into several subclasses, but I will only 

elaborate on zoochory, since this is the focus of my research. The unit of dispersal is a 

diaspore, which in zoochorous plants is nearly always the seed. Three different types of 

zoochory can be distinguished; endo-, syn- and epi- (or exo-) zoochory. Endozoochory 

occurs when diaspores are transported within the animal, either intentionally or 

accidentally. Synzoochory takes place when diaspores are intentionally carried in the 

mouth and epizoochory when diaspores are accidentally carried on the outside of the 

animal (Van der Pijl 1969). In tropical rainforest zoochory, in particular endo- and 

synzoochory, is the most common way of seed dispersal (Charles-Dominique 2001). 

About 75% or more of all plant species depend on vertebrates for the dispersal of their 

seeds (Howe & Smallwood 1982). In Mediterranean scrubland and tropical dry woodland 

only 50-70% of the plants are zoochorous and in temperate forests the percentages are 

even lower (30-40%). Zoochorous plant species are even virtually absent in alpine and 

desert vegetation (Jordano 2000). 

The general principle of zoochory is fairly simple. Frugivores rely on fruits as their 

essential food source for survival, while at the same time, as seed dispersers, they 

represent the dynamic link between the fruiting plant and the seedlings. Fruits facilitate 

the dispersal of seeds by providing benefits to seed dispersers (Van der Pijl 1969; Witmer 

and Van Soest 1998; Jordano 2000). The seed is associated with soft and fleshy edible 

fruit pulp with attractive signals (colours, smell) on which the frugivores orient themselves 

to locate ripe fruit. The rewards offered by the plants include nutritious fruit pulp, while 

through directed dispersal by frugivores plants can colonize new vacant areas and avoid 

disproportionate mortality near the parent plant (Janzen 1970; Connell 1971; Chapman 

and Chapman 1996; Wenny 2000). Of course other animals also exploit this mutualism. 

For example, some invertebrate and vertebrate frugivores capitalize on fleshy fruits 

without dispersing the seeds or, even worse, by destroying them. In this respect, 

frugivore species can be subdivided based on their role in the ecosystem (Gautier-Hion et 

al. 1985; Debussche and Isenmann 1992; Jordano 2000). First of all legitimate seed 

dispersers swallow fruits entirely, digest the pulp and defecate or regurgitate intact seeds. 

Secondly, fruit pulp specialists or seed droppers tear off the pulp and drop the seeds. 

7


Finally, seed predators discard the pulp, extract and digest or crack the seed. The latter 

can be considered as granivores. 

Fruit-frugivore interactions represent an important aspect in tropical forest dynamics. 

Fruit resources are thought to be crucial in sustaining certain vertebrate populations in 

some tropical areas (Terborgh 1986a; Gautier-Hion and Michaloud 1989; Julliot 1997). 

While zoochorous fruits are very abundant in the tropics, frugivores make up the bulk of 

vertebrate biomass in tropical forests (Fleming et al. 1987; Gautier-Hion et al. 1985). In 

general, birds and mammals (mainly primates and bats) are the most important 

vertebrate frugivores, which swallow and defecate, regurgitate or spit seeds away from 

the parent plant. However some records of frugivory by reptiles, fish and invertebrates 

exist as well (Van der Pijl 1969; Corlett 1998). Obligate and occasional frugivory are the 

extremes along a gradient of fruit-eating. Most frugivores supplement their fruit diet to a 

greater or lesser extent with animal prey, flowers, leaves, nectar, and seeds (Fleming et 

al. 1987; Corlett 1998). According to Terborgh’s definition (1986a) are frugivores only 

those animal species whose diet is composed of at least 50% fleshy fruits. Worldwide, 17 

bird families can be considered strictly frugivorous (Snow 1981). Among mammals, 

obligate frugivores are rare with the exception of Pteropodids (Old World bats, Marshall 

1983). As for primates, fruit is found in the diet of 91% of the species studied to date 

(Jordano 2000). 

Seed dispersal is a complex multi-step process that links the end of the reproductive 

cycle of adult plants with the establishment of their offspring (Wang and Smith 2002). The 

pre-dispersal, dispersal and post-dispersal phases make up the intermediate processes 

between seed production and recruitment of adult trees (Fig. 1). The pre-dispersal phase 

is actually the fleshy fruit-frugivore interface, which includes the probability of the fruit to 

be selected, eaten and dispersed by a certain frugivore. The major focus in this phase is 

the fruit as food source with its morphological display and nutritional reward, including its 

spatial and temporal availability both seasonally and annually. Fruit production, fruiting 

period and fruit crop size are other traits influencing this phase (Garber and Lambert 

1998). The actual dispersal phase indicates mainly the stage at which the consumers 

forage and feed on the fruits. Here, fruit and seed handling and processing determine the 

roles different frugivores may fulfil in this ecosystem. Finally in the post-dispersal phase 

the seed fate is concentrated on, which includes seed shadow, germination and growth of 

seedlings. This phase further includes secondary seed dispersal and seed predation. 

Since seed dispersal takes place at the final stage of a plant’s reproductive life cycle, it 

has the potential to wipe out previous effects of pollination and fruit growth phases 

(Jordano 2000) and thus also to alter vegetation recruitment (Wang & Smith 2002). 

SEED DISPERSAL HYPOTHESES 

The study of frugivory and zoochory has led to early conceptual contributions. Several 

hypotheses were developed trying to explain the evolution within zoochorous fruits. First 

of all during the 1980s the field of frugivory and seed dispersal focused on a central 

paradigm: co-evolution (Janzen 1980; Levey and Benkman 1999). In this process 

organisms interact closely, influencing each other’s evolution, which leads to reciprocal 

adaptations. With respect to seed dispersal, certain plant species may evolve specific 

fruit and seed traits to facilitate dispersal while the behaviour and diet of the frugivore 

responds to these changes. Tight co-evolutionary fruit-frugivore interactions have been 

proposed by several authors (Howe 1977; Tutin et al. 1991; Chapman et al. 1992b). 

8

PRE-DISPERSAL PHASE 

phenology, pollination 

fruit production 

fruit crop size, fruiting length 

Adult 

tree 

POST-DISPERSAL 

PHASE 

growth of seedlings 

and saplings 

regeneration 

Seedling 

Fruit 

availability 

Secondary 

dispersal 

POST-DISPERSAL PHASE 

seed shadow 

germination & establishment 

seed mortality 

Seed 

rain 

Fig 1. The seed dispersal cycle (based on Wang and Smith 2002) 

PRE-DISPERSAL PHASE 

morphological display 

nutrient content 

feeding selection 

Seed 

uptake 


DISPERSAL PHASE 

primary seed dispersal 

fruit & seed handling 

feeding & dispersal behaviour 

by frugivores 

Seed 

predation 

Secondly based on the idea of co-evolution, a dichotomy of low and high investment 

plants has been put forward (Snow 1971; McKey 1975; Howe 1977; Howe and Estabrook 

1977; Howe 1993). Low investment plants or generalists were described to invest little in 

the nutritional value of their fruits, having watery and sugary pulp. They produce large 

fruit crops during short periods, attracting as many opportunist frugivores as possible. On 

the contrary, high investment plants or specialists produce highly nutritious fruits, during 

elongated periods of time. They have a smaller fruit crop and often large one-seeded 

fruits attracting few specialist frugivores. This dichotomy has further contributed to the 

more differentiated concept of dispersal syndromes. Syndromes represent diffuse coevolution 

among taxa and include mainly broad morphological co-adaptations of fruit and 

seed traits that attract certain taxonomic groups of seed dispersers. Each disperser type 

corresponds to a more or less diversified group of frugivorous animals whose size, 

anatomy and behaviour are compatible with the fruit features. This concept has been 

widely used in literature (Van der Pijl 1969; Howe and Smallwood 1982; Janson 1983; 

Knight and Siegfried 1983; Gautier-Hion et al.; 1985; Willson et al. 1989; Julliot 1996; see 

also Fischer and Chapman 1993; Jordano 1995; Voigt 2001). These three hypotheses 

highlight the importance, mechanisms and consequences of seed dispersal. 

STUDY SITE: LITTORAL FOREST OF SAINTE LUCE (SE-MADAGASCAR) 

The littoral forest of Madagascar represents a particular study site to carry out seed 

dispersal research. First of all, Madagascar is a hotspot of biodiversity and endemism 

(Mittermeier et al. 1998) with high priority for conservation. The majority of all plant 

species is endemic to Madagascar (Dumetz 1999; Schatz 2001), with percentages as 

9


high as 98% for the littoral forest (Rabevohitra et al. 1996; Razafimizanilala 1996). So 

studying fruit-frugivore interactions contributes to a better understanding of the dynamics 

within this ecosystem, that differs greatly in flora as well as fauna from those previously 

studied. Secondly most studies in Madagascar regarding seed dispersal focused only on 

the role of the larger lemur species (Hemingway 1996; Dew and Wright 1998; Overdorff 

and Strait 1998; Birkinshaw 1999, 2001; Ganzhorn et al. 1999a). Frugivory and zoochory 

by bats, birds and the smaller lemurs have only poorly been studied in Madagascar and 

not at all in the littoral forest. There is thus a great need for a community-wide approach 

of zoochory in Malagasy ecosystems, which is carried out for the littoral forest in this 

study. Thirdly, the Malagasy frugivore guild can be considered atypical in its composition. 

There are very few frugivorous bird and bat species here compared to other tropical sites, 

while at the same time large mammals are missing. Furthermore, since humans arrived 

on the island one third of its lemur species has disappeared (Godfrey et al. 1997). This 

relatively species-poor frugivore guild allows us to sample data on all involved species. 

Besides, this particular composition stresses the importance of the remaining frugivores 

in seed dispersal, which will provide crucial data that can be implemented in conservation 

management plans. At present the highly degraded and fragmented littoral forest is under 

severe threat and in urgent need of conservation. Finally, Sainte Luce lies at the 

southernmost position of a rainforest, being south of Capricorn, which may have led to a 

differential impact of abiotic factors here. The high degree of endemism and biodiversity, 

together with the awkward frugivore composition, the southernmost position and the 

urgent need for data on plant-animal interactions in this threatened forest type have all 

contributed to the choice of the littoral forest to focus on community-wide seed dispersal. 

THESIS OUTLINE 

In this PhD the results of a 14-month field research (November ’99 - February ‘01) are 

presented. Throughout the research, a close collaboration was established with the 

University of Hamburg (Germany), the University of Antananarivo, the Ministry of Water 

and Forestry, Qit Madagascar Minerals and Missouri Botanical Garden in Antananarivo 

(Madagascar). 

This study aims at understanding community-wide seed dispersal in the Malagasy 

littoral forest. Research on both pre-dispersal and dispersal phase was carried out. The 

central focus of the doctoral study is to unravel how phenological, morphological and 

biochemical fruit traits determine and interact in fruit-frugivore dynamics. Seed dispersal 

is approached both ways, from the perspective of trees and frugivores. In the first part, 

fruits are studied with respect to the trees’ investment in morphological display and 

nutritional reward along with their temporal availability. To get insight into the overall fruit 

availability in the littoral forest together with its intra- and inter-annual fluctuation, 

phenological transects and fruit trails were carried out (Chapter 1). Dispersal strategies 

are tested based on existing hypotheses such as co-evolution, low-high investment 

model and dispersal syndromes (Chapter 2). In the next part, the influence of the 

morphological and biochemical traits on the frugivores’ fruit choice, feeding selection and 

fruit and seed processing is investigated. The role of the complete frugivore guild in this 

ecosystem as seed dispersers or predators is determined (Chapter 3). The feeding 

ecology and disperser role of Pteropus rufus (Chapter 3a) and predator role of 

Coracopsis nigra (Chapter 3b) are elaborated on. Subsequently, an inter-site comparison 

between fruit traits and feeding ecology of Eulemur fulvus and Cheirogaleus medius is 

10


made to test the validity of my results at other forest types. Both sites have similar 

frugivore guilds, but involve completely different forest types: the dry deciduous forest of 

Kirindy (west Madagascar) and the humid littoral forest of Sainte Luce (south-east 

Madagascar)(Chapter 4). Finally, the present degradation and fragmentation of the littoral 

forest as well as the disruption of fruit-frugivore mutualisms in Sainte Luce is discussed. 

Based on my understandings of seed dispersal dynamics, certain findings can be 

integrated within existing conservation measures (Chapter 5). 

11

’Hazo tokana tsy mba ala’ 

One tree does not make a forest 

Malagasy proverb 

Drawing of flower and fruit of Ravenala madagascariensis (Strelitziaceae) © An Bollen

Phenology of the littoral forest 

of Sainte Luce, 

south-east Madagascar 

AN BOLLEN, GIUSEPPE DONATI 

BIOTROPICA (SUBMITTED) 

Phenology 

ABSTRACT 

From January 2000 through December 2002 phenological transects were carried out to 

assess monthly leaf, flower and (ripe) fruit presence for 423 individual plants (95 plant 

species, 43 families) within the littoral forest of Sainte Luce. Fruit-on-trail-counts were 

conducted additionally in the year 2000 to allow comparison between both methods. 

Despite low climatic seasonality and absence of a dry season in the littoral forest, interannual 

phenological patterns were seasonal at our site. Within-year variability was 

present with clear periods of abundance and scarcity. All phenophases were highly intercorrelated 

and peak from November through February. This was found in other humid 

Malagasy forests as well, while in dry Malagasy forests phenophases seemed to be more 

spaced in time due to the more seasonal climate. Day length seems to have the strongest 

impact on all phenophases, while rainfall is associated with flushing only and temperature 

with fruiting and ‘ripe fruiting’. Differences in presence of ripe fruits when comparing 

between both sampling methods can be explained by the differential contribution of 

several life forms in both methods, which influence overall fruiting patterns. 

INTRODUCTION 

Phenological studies address the timing of reproductive events in plants such as bud 

formation, flowering and fruiting along with vegetative processes like leaf flushing and 

shedding. Tropical trees display an enormous variety in temporal patterns of flowering 

and fruiting. Both flowers and fruits are patchily distributed in time and space and are 

relatively scarce food items compared to leaves and insects (Howe 1984). Therefore it is 

critical to study food availability and distribution in order to understand the behavioural 

ecology of tropical wildlife. In the tropics, many animal species are frugivorous to a lesser 

or greater degree and in terms of biomass, frugivores are the dominant trophic group in 

most tropical forest mammalian communities (Terborgh 1983). This is further reflected by 

the dominance (60-90%) of zoochorous plant species producing fleshy fruits (Howe and 

Smallwood 1982). 

Quantifying fruit availability has been a primary objective in many studies, which 

focus on the ecology of tropical fruiting trees and their consumers (Chapman et al. 1992a; 

Janson and Chapman 1999). During the alternation of seasons and years in rain forests, 

the availability of vegetative and reproductive plant parts is irregular and induces periods 

of abundance and scarcity of food for consumers (Brugiere et al. 2002). These temporal 

15

Chapter 1 

changes in resource availability are both affected by abiotic or climatic variables as well 

as by biotic factors through herbivory, pollination and seed dispersal. 

Up to present, several short-term phenological studies have been published for 

Madagascar, where 96% of the tree species are endemic (Schatz 2001). Phenological 

studies have been carried out in dry deciduous and semi-deciduous forests in the West 

(Meyers and Wright 1993; Sorg and Rohner 1996; Curtis and Zaramody 1998; 

Rasmussen 1999; Yamashita 2002) as well as in lowland and mid-altitude rainforest in 

the East (Andrianisa 1989; Overdorff 1993a, 1993b; Rigamonti 1993; Freed 1996; 

Hemingway 1996, 1998; Andrews and Birkinshaw 1998). Most of these studies represent 

data from one year and often include only a limited number of species. At present, no 

data are available for the littoral forest. This paper presents the first findings from a 

detailed phenology of plant guilds in the littoral forest of Sainte Luce (south-east 

Madagascar) in the interest of revealing broad community-wide patterns of leafing, 

flowering, fruiting in the course of three years. This study generates indices to the food 

supply available to animal consumers, but also addresses the impact of abiotic factors. 

Furthermore, we compare phenological patterns of different Malagasy forest types. 

In this study several aspects will be looked at closely. First, we describe the temporal 

fluctuations of flushing, flowering and fruiting inter- and intra-annually based on data from 

phenological transects and fruit-on-trail-counts, which are considered to be the most 

common phenological methodologies (Chapman et al. 1992a; Chapman and Wrangham 

1994). Secondly, we look for correlations between these phenological patterns and 

different abiotic factors such as rainfall, temperature and day length, that may trigger 

phenological events. Finally, we compare our data with the results of other Malagasy 

study sites. 

METHODS 

Study site 

The littoral forest of Sainte Luce (24º45'S 47º11'E) is located in south-eastern 

Madagascar, 50km up north of Fort-Dauphin. Research was carried out in a 377-ha large 

forest fragment, called S9 (Fig. 1). Littoral forest is characterised by a relatively open or 

non-continuous canopy, which is 6 to 12m in height with emergents up to 20m. The 

diameter at breast height (DBH) of trees rarely exceeds 30-40cm. Littoral forest grows on 

sandy soils and occurs within 2-3km of the coast at an altitude of 0-20m (Lewis 

Environmental Consultants 1992a). 

Abiotic factors 

Daily rainfall was measured with a plastic rain gauge (TruCheck) during the research 

period. Rainfall data from February till September 2001 were not available for Sainte 

Luce. A thermo-hygrograph (Box Pro) was placed in primary forest at one meter height to 

record the daily march of temperature and relative humidity. Variation in day length was 

calculated using Stephen Moshier’s Ephemeris Program v5.1 (Moshier 1991) for the 

Sainte Luce latitude. Mean monthly values were calculated for all abiotic factors. 

16

20°S 

N 

MADAGASCAR 

50°E 

Sainte Luce 

Fort Dauphin 

SAINTE LUCE 

S9 

fruit trail 

phenological 

transect 

Phenology 

Fig. 1. Map of Madagascar with indication of the study site together with an enlargement of the 

forest fragment ‘S9’ where the study was carried out. A detail of the grid with corresponding 

phenology transects and fruit trails is shown as well. 

Phenological transect 

A systematic floristic and phenology inventory was conducted as part of two doctoral 

research projects in the littoral forest: one on the behavioural ecology of Eulemur fulvus 

collaris (Donati 2002) and the other on seed dispersal and predation by the frugivorous 

guild (Bollen and Van Elsacker 2002, Chapter 3a; Bollen et al. in press, Chapter 4). All 

plant species with DBH greater than 5cm (conform to other Malagasy studies, see 

Meyers and Wright 1993; Overdorff 1993a; Hemingway 1996) within 5m of each side of 

the transects were tagged with fluorescent flags provided by an individual code. Ideally, 

five individuals per species were marked. Additional tree species were added to this list. 

These involved important known frugivore resources that did not occur on the transect or 

only at very low densities. The complete phenological transect consisted of 423 

individuals of 95 species (74 genera; 43 families) sampled over two botanical transects 

covering 2320x10m² (Fig. 1). Eighteen trees died naturally or were cut during the 

sampling period, resulting in a complete data set of 405 individuals. The different 

vegetation types, such as primary, secondary and swamp forest as well as abandoned 

tavy (slash and burn areas) were represented in the phenological transect. 

Phenological data were recorded once a month as of January 2000 up to present. 

Here we present data from January 2000 through December 2002. The different 

categories considered are the following: 

Leafing: no leaves, presence of young leaves, full of leaves, leaf fall; 

Flowering: no flowers, flower buds, open flowers, fallen flowers; 

Fruiting: no fruits, fruit buds, unripe, ripe or fallen fruits, dry fruits of last season. 

100m 

campsite 

route nationale 

Sainte Luce – 

Mahatalaky 

100m 

17

Chapter 1 

For analyses we narrowed these categories down to complete leaf fall, presence or 

absence of young leaves, flowers, ripe and other fruits. Observations on leafing, flowering 

and fruiting were always made by the same team of two field assistants, who scanned the 

canopy with binoculars and checked the litter below for fallen flowers and/or fruits. A 

species was scored flushing, flowering or fruiting if at least one individual of this species 

was in this phenophases. To differentiate between unripe and ripe fruits, we focused on 

differences in colour, size and consistency. 

No attempt was made to estimate overall fruit production. Neither did we use 

quantitative scores nor relative abundance of leaves, flowers and fruits as we lacked 

previous knowledge on crop sizes. Furthermore the high variability of crop sizes intra- 

and inter-specifically in time, related to tree size, makes it difficult to objectively quantify 

these reproductive events. Moreover, various measures are used in different studies 

which makes comparison problematic and according to Chapman et al. (1992a) interobserver 

variability is high. 

We further subdivided our species sample by life form into large trees (>6m), small 

trees (

Phenology 

Statistical analyses 

Because phenological data are not independent in time and not normally distributed, we 

used non-parametric statistics for repeated measures. To test for inter-annual variability 

in phenophases we used Kendall’s Coefficient of Concordance, while the Friedman test 

(ANOVA) showed whether significant differences exist in the number of species flushing, 

flowering or fruiting among years. Spearman Rank Correlation Analyses were used to 

examine overall relationships in phenophases, climatic and feeding data. Chi-square 

analyses were conducted to compare between flowering and fruiting on the basis of 

different classes of synchronicity and regularity, as well as to compare between life forms 

used in fruit trail and phenology. Mann Whitney U tests compared the number of species 

with ripe and other fruits between two phenological methods. All statistical tests were 

carried out according to Siegel (1956) with the statistical software SAS and STATISTICA 

for Windows. 

RESULTS 


There was considerable year to year variation in the seasonal distribution of rainfall and 

temperature. No significant correlations could be found for both variables between years 

(Table 1, Fig. 2). Mean annual rainfall during 2000-2002 was 2690mm (±228 SD). Figure 

2 gives monthly distribution in rainfall and monthly average temperature, indicating an 

obvious wet season from November through February. No clear dry season could be 

detected. Driest months were September and October (average 79mm ±37 SD). During 

July 2000 rainfall was unusually high for this time of the year. Temperatures were highest 

from December through March after which they declined to a minimum in July and 

August. Mean monthly temperature was 23°C (±2.3 SD, N=30) ranging from 18.5 to 

25.6°C. As expected for a site below the Tropic of Capricorn, we found a considerable 

difference in day length in Sainte Luce (Fig. 2) with a minimum in summer solstice (June: 

10.6h) and a maximum in winter solstice (December: 12.6h). 

Phenology 

Overall, we monitored the fruiting phenology of 423 individuals of 95 plant species 

belonging to 74 genera and 43 families (Appendix I). The median number of individuals 

per species was five, ranging from one to nine. Most represented plant families were 

Clusiaceae, Flacourtiaceae, and Myrtaceae, with six plant species each and Rubiaceae 

and Euphorbiaceae with five species. The majority of the remaining families were 

represented by only one or two species. Most studied plants were large trees (71%) 

followed by small trees and shrubs (23%). Vines and epiphytes only made up 5% and 1% 

respectively of the complete sample. 

The flowering peak occurred always at the beginning of the rainy season, mainly in 

November, when on average 38 species (±7 SD, range 33-47) were blooming. Of all 

flowering events (N=2054) during three years, 15% (N=309) occurred in November. 

Flowering patterns over years were significantly correlated (W=0.46, df=2, N=12, P

Chapter 1 

Fig. 2. Seasonal variation in the climate of the littoral forest of Sainte Luce is given together with 

monthly rainfall, mean temperature and day length recorded over a three year period (2000-2002). 

Table 1. Spearman rank correlations for interannual patterns within temperature, 

rainfall, flushing, flowering and fruiting. rs-values are given. * P

Phenology 

The highest number of fruiting species were recorded from November through 

February of each year, the fruiting peak coinciding with high rainfall. In general, fruiting 

lagged behind flowering by one or two months. Among 2832 effective fruiting events 

during 36 months, 36% was concentrated between November and February of each year. 

Kendall’s Concordance showed that fruiting patterns among years were significantly 

associated (W=0.53, df=2, N=12, P

Chapter 1 

80 

70 

60 

50 

40 

30 

20 

10 

0 

60 

55 

50 

45 

40 

35 

30 

25 

20 

15 

10 

Fig. 3. Phenological flushing, flowering, and fruiting (all fruits and ripe fruits) are given for the three 

years with indication of the monthly average of number of species present. 

Comparison between phenological transect and fruit trail data 

Over a 12-month period a total of 113 species were monitored fruiting on the fruit trail. 

These species belonged to 75 genera and 48 families. Fourteen plant species could not 

be identified. The fruit trail and phenological transect have 60 species in common. Large 

trees made up 43% of the fruit trail species, while for smaller trees, shrubs, vines, herbs 

and epiphytes this was 57% (App. I). The distribution of growth types within phenological 

transect and fruit trail were significantly different (X²=16.33, df=1, P

Phenology 

Table 2. Spearman rank correlations of three year data between number of species with young 

leaves (YL), flowers (FL), fruits (FR), ripe fruits (RIPE), temperature (TEMP), rainfall (RAIN) as 

well as day length (DL). rs-values are given. * P

Chapter 1 

Fig 4. Comparison of data from the phenological transect and fruit trail is shown with indication of 

the number of species with overall ripe fruits present as well as the distinction between life forms. 

DISCUSSION 


A pronounced island-wide environmental seasonality is said to characterise Madagascar 

(Morland 1993). This is true in particular for the ambient temperature that appears to vary 

in consistent, seasonal cycles throughout Madagascar, which corresponds with data from 

our study site. On the contrary, rainfall does vary widely among regions in Madagascar 

24 

25 

Fruit trail (N=113 sp) 

N sp large trees others all trees 

20 

15 

10 

5 

0 

35 

N sp 

30 

25 

20 

15 

10 

5 

0 

Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan 

Phenological transect (N=95 sp) 

large trees others all trees 

Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan

Phenology 

(Table 4). While the North and West are characterised by prolonged annual dry seasons 

and the South by sparse and irregular rainfall, the East has no clear dry season (Morland 

1993; Kappeler and Ganzhorn 1994; Hemingway 1998; Britt 2000; Vasey 2000). When 

comparing our rainfall data with those from eastern low alititude forests Nosy Mangabe, 

Masoala and Betampona (Morland 1993; Britt 2000; Vasey 2000) and the inland mid 

altitude rainforest Ranomafana (Overdorff 1993a; Hemingway 1998), all sites have their 

driest months in September and October, with a prolongation into November for the 

Masoala peninsula. The mid altitude forest of Ranomafana approaches a more seasonal 

climate as it has relatively fewer rain from April through September. Low altitude coastal 

forests are clearly more aseasonal when considering rainfall which is conform to our 

study site. Thus, no seasons were considered in the littoral forest of Sainte Luce. It 

seems to follow the definition of an aseasonal tropical wet climate (Morellato et al. 2000), 

as it is characterised by mean monthly temperatures of at least 18°C, annual rainfall 

above 2000mm and either no dry season or a short drier period of less than four 

consecutive months with rainfall below 100mm per month. The situation in Sainte Luce 

corresponds with that in other tropical wet forests worldwide. In summary, rainfall seems 

more important in determining local seasonality rather than temperature which has less 

fluctuation (Sakai et al. 1999) and, overall seasonality is said to be lower in humid than in 

dry forests (Koptur et al. 1988). 

Phenological transect 

The phenology of the community as a whole followed a flowering, fruiting and ripe fruiting 

pattern which was repeated from year to year in a regular seasonal cycle, while for 

leafing patterns inter-annual variability is high. However, the number of species flowering 

and fruiting differed substantially from year to year. Over three years flushing, flowering, 

fruiting and ripe fruiting were all inter-correlated and the peak of all phenophases 

occurred in the same period (November and February). This corresponds with the fact 

that in many evergreen species flushing and flowering occur close in time as they are 

both triggered by solar irradiance and young leaves and flowers develop from the same 

new shoots (Van Schaik et al. 1993). 

We found an important quantitative difference between availability of ripe versus all 

fruits, which is probably related to the fact that unripe fruits were often aborted. Unripe 

fruit abortion is considered common in the tropics (Smyte 1970; Medway 1972). This can 

occur due to several reasons such as climate (Janson and Chapman 1999), insect 

predation (Koptur et al. 1988) or other factors. Abnormal high rain in July 2000 resulted in 

fruit abortion in many species as the number of ripe fruits was considerably lower in July 

and August 2000 compared to other years. 

In our data, the large differences between number of species flowering compared to 

the ones fruiting is not biologically meaningful and is an artefact of the methodology. This 

is probably because the category of all fruits includes fruit buds, unripe, ripe and even 

rotten fruits as well as old dry fruits from the previous season, which leads to higher 

numbers of species fruiting during an ‘artificially’ prolonged time. As for flowers, monthly 

scoring is probably not sufficient (De Block pers. comm.) as flowers in general have a 

short existence only and thus may be missed. As a matter of fact, in five species no 

flowers were observed even though these species fruited. 

25

Chapter 1 

Table 4. Data on phenology in different sites throughout Madagascar (ND: no data available, 

RNF: Ranomafana). 

Study site Forest type Climate Phenology sample 

Location Altitude (m) 

Rainfall (mm) 

Duration 

Ambatonakolahy mid altitude rainforest max rain: Feb-Mar 511 ind 72 sp 

NE Madagascar 450-650 min rain: Sep Dec '90-Nov '91 

4200 5 mo 

Vatoharanana RNF mid altitude rainforest max: Dec-Mar 104 ind 26 sp 

21°02-25'S 47°18-37'E 1125 min: Sep-Oct Jul '88-Aug '89 

2300 1 yr 

Vatoharanana RNF mid altitude rainforest max: Dec-Mar 1354 ind 127 sp 

21°02-25'S 47°18-37'E 1100 min: Sep-Oct Jan '91-Jul '92 

2300 1.5 yr 

Daraina dry forest max rain: Dec-Mar 499 ind 150 sp 

13°14'S 49°39'E 400-1100 min rain: Sep Jul '90-Jun '91 

1445 1 yr 

Kirindy dry deciduous forest rainy season: Dec-Feb 80 ind 56 sp 

20°04'S 44°40'E 18-40 dry season: Apr-Oct 78-'87 

800 9 yr 

Anjamena dry semideciduous forest rainy season: Dec-Apr 19 sp 

16°03'S 45°55'E 1 

ND dry season: May-Nov Oct '94-Sep '95 

1189 1 yr 

Ampijoroa dry semideciduous forest rainy season: Dec-Mar 317 ind 

16°19'S 46°49'E 1 

75-390 dry season: May-Oct Jun '96-Jun '97 

1771 1 yr 

BezaMahafaly dry deciduous forest rainy season:Nov-Mar 10 plots (2x50m) 

SW-Madagascar ND dry season: Apr-Oct Feb '99-Feb '00 

866 9 mo 

Lokobe primary lowland rainforest aseasonal 278 ind 

13°23'S 48°18'E 0-430 max rain: Nov-May Nov '92-Dec '93 

2356 1 yr 

Montagne d'Ambre mid altitude rainforest aseasonal ND 

NW Madagascar 850-1474 

3585 

max rain: Dec-Mar 

Nosy Mangabe primary lowland rainforest aseasonal ND 

NE Madagascar 0-331 

3806 

max rain: Jan-May 

Sainte Luce littoral forest aseasonal 423 ind 95 sp 

24°45'S 47°11'E 0-20 max rain: Dec-Mar Jan '00- Dec '03 

2690 3 yr 

1 general geografic data are from the authors or from Preston (1991). 

26

Table 4 Continued 

Flushing Flowering Fruiting Literature cited 

constant peak: Aug high: Oct-Nov Rigamonti (1993) 

low: July 

Phenology 

high: Jan, May peak: Aug, Nov-Dec peak: Aug-Oct, Feb Overdorff (1992,1993a, 1993b) 

low: Jun-Oct low: Apr-Aug Meyers and Wright (1993) 

corr with rainfall corr with rainfall peak: Oct-Dec Hemingway (1996,1998) 

low: Apr-Jul 

peak: wet season peak: Oct-Nov peak: Dec-Mar Meyers and Wright (1993) 

peak: Nov-Dec peak: Oct-Dec all year available Sorg and Rohner (1996) 

peak: Dec-Mar peak: Oct peak: Oct Curtis (1997) 

low: Jun-Oct low: Jun 

peak: Dec-Feb peak: Jan, Sep, Feb peak: Jun-Jul, Sep, May Rasmussen (1999) 

peak: Nov-Jan peak: Nov peak: Apr-Jun Yamashita (2002) 

ND ND low: Mar-Jun Andrews and Birkinshaw (1998) 

ND similar to RNF similar to RNF Freed (1996) 

in Vasey (2000) 

ND peak: Dec-Jan peak: Nov-Feb Andrianisa (1989) 

in Vasey (2000) 

peak: Nov-Jan peak: Nov peak: Nov-Jan Bollen and Donati (this study) 

low: Aug-Oct 

27

Chapter 1 

Relation with abiotic factors 

Climatic parameters such as temperature, rainfall, day length, cloud cover and irradiance 

are the proximate factors that affect the timing of phenophases (Smythe 1970; Van 

Schaik et al. 1993) and among them rainfall has often been identified as the principal 

external factor, directly or indirectly controlling the period rhythms of tropical forests 

(Medway 1972; Lieberman 1982; Sorg and Rohner 1996; Morellato et al. 2000). 

However, transect-wide flowering and fruiting patterns were not related to rainfall in 

Sainte Luce as opposed to many other sites (Hemingway 1996; Sorg and Rohner 1996). 

On the contrary, leaf emergence did correspond positively with rainfall as in most other 

tropical forests (Lieberman 1982; Hemingway 1996, 1998; Rasmussen 1999). Flowering 

was not correlated with temperature nor with rainfall, while fruiting was positively 

associated with temperature and for “ripe fruits” this correlation was even stronger. 

Smythe (1970) mentioned that organic composition is slow throughout the dry season but 

as rains begin, the combination of high temperature and high relative humidity allows 

decomposition of the accumulated materials to proceed very quickly. The dropping of 

seeds at a time when there is a sudden rather brief abundance of nutrients may increase 

the probability of survival of seedlings, which may explain this correlation. Day length, our 

rough estimation of solar irradiance, was highly correlated with all three phenophases. 

This result is in agreement with the findings of Van Schaik et al. (1993). They highlighted 

the importance of irradiance in enhancing photosynthetic processes during flushing and 

flowering. Given the latitude of Sainte Luce, at the southernmost range for a rain forest, 

this abiotic factor could be particularly important in triggering phenological cycles here. 

Caution has to be taken while interpreting these results as both temperature and rainfall 

data are incomplete and impact of edaphic factors was not studied. Besides external 

factors, plant species obviously also have endogenous rhythms and their expressions are 

affected by changes in internal tree functions that are not necessarily linked with 

environmental patterns (Marco and Paez 2002). 

Synchronicity 

Our results showed that there is a complete gradient from species with strong 

seasonality, either annually or continuously flowering and fruiting to those with weak 

seasonality without obvious periodicity or an extended flower and fruit production 

throughout most of the year. This high variability in synchronicity and flower-fruit 

periodicity has also been found in other tropical sites (Gautier-Hion et al. 1981; Van 

Schaik 1986; Van Schaik et al. 1993). However, we can extract some general trends at 

our site. Almost half of all species monitored had flowers and fruits each year, which 

indicates that there is some stable level of food availability for nectarivores and 

frugivores. While flowering is annual in about 39% of the species, for fruiting this 

percentage is reduced to 31%. This difference together with the higher number of species 

that has irregular fruiting, indicates that many species failed to set fruits even though 

flowers were present. Janson and Chapman (1999) mentioned that in species that 

attempt to reproduce every year, fruiting failure is common in many years due to between 

year climatic variation. Absence of pollination, climatic and other environmental as well as 

physiological factors may be responsible for this (Medway 1972, Koptur et al. 1988). The 

category of irregular species may include species with bi-annual and tri-annual rhythms, 

but long-term data are needed to reveal these longer flower-fruit intervals. Extended 

flowering and fruiting occur in most tropical forest at higher or lower percentages 

depending on the site (Frankie 1975; Koptur et al. 1988; Van Schaik et al. 1993; Sakai et 

28

Phenology 

al. 1999; Morellato 2000). It makes biologically sense that this occurs more for fruiting 

than flowering as unripe fruits often take a longer time to mature than flowers. 

Comparison between phenological transect and fruit trail 

By selecting trees above a specific size as in the phenological transect, one makes the 

assumption that trees with a DBH smaller than 5cm are unable of producing fruit. In this 

way all smaller trees, shrubs, herbs, epiphytes and vines are excluded, although they 

may influence overall fruiting patterns. Our data revealed that fruiting patterns in general 

were indeed correlated between both methods. This trend was found as well in a study by 

Wallace and Painter (2002) when comparing between both methods. However, when 

considering ripe fruits only it was evident that in Sainte Luce during austral winter smaller 

vegetation forms carried more ripe fruits than the phenological transect showed. Thus, 

fruit trail data compensate for the underestimation of smaller trees and shrubs in 

phenological transects. As both methods are complementary it is useful to include both to 

get the global picture and reveal different phenological trends for large trees and other life 

forms. Furthermore fruit trails have the advantage of adding a density effect to the data 

as this is a more quantitative phenological approach where all trees in fruit encountered 

on the transect are scored each month. On the contrary, during phenological transects, 

only phenophases of a fixed number of tree species are scored, which represent only a 

sub-sample of all tree present in an ecosystem selected on their DBH. Moreover, 

phenology transects are especially appropriate for scoring resource availability for 

arboreal frugivores, while fruit trails are more appropriate for terrestrial frugivores 

(Wallace and Painter 2002). In this respect we can explain the lag in time in periods of 

fruit scarcity between both methods as in phenology only fruits in the canopy are scored, 

which are earlier available than the subsequently fallen fruits, that are included in fruit-ontrail-counts 

(cf. Wallace and Painter 2002). To conclude, combining both methods is ideal 

to get insight into the complete food availability for the frugivorous guild as a whole. 

Furthermore, our results show the importance of differentiating between unripe and ripe 

fruits in both methods, which often lacks in other studies, to indicate the ‘true’ severity of 

bottlenecks in fruit availability that frugivores face in tropical forests. 

Comparison with other Malagasy sites 

Since methodologies and representation of data differ between sites, we restricted our 

comparison to general phenological patterns. In eastern humid forests (Ranomafana and 

Sainte Luce, Table 4) flushing coincides with the wettest period of the year. In the dry 

semi-deciduous and deciduous forests the emergence of young leaves starts at the end 

of the dry season and attains a peak at the onset of the rainy season (Table 4). Given the 

much stronger seasonal climate of dry forests, where flushing is restricted in time, young 

leaves anticipate the first rains and create the possibility for plants to maximally benefit of 

the favourable rainy season. In all sites flowering seems to occur in October and 

November. As mentioned before, flushing and flowering occur close in time and are both 

triggered by high irradiance (Van Schaik et al. 1993), which is elevated in these months 

when rainfall is ubiquitously low and the sky rarely covered. Fruiting patterns are much 

more diverse between sites and peaks differ from site to site. In Ranomafana, Nosy 

Mangabe, Montagne d’Ambre and Sainte Luce fruiting occurred at the same time as 

flowering and flushing (Table 4). The dry forests may have fruiting peaks in austral 

summer (Meyers and Wright 1993), winter (Curtis and Zaramody 1998; Rasmussen 

1999; Yamashita 2002) or throughout the year (Sorg and Rohner 1996). In summary, the 

29

Chapter 1 

Malagasy rainforests display an associated pattern for all phenophases, while in the dry 

forests phenophases are more spaced in time (Kappeler and Ganzhorn 1994; Curtis 

1997; Rasmussen 1999). Therefore, the lean fruit period in the littoral forest and mid 

altitude rainforest does not appear to be compensated by flower and/or young leaf bursts. 

This pattern together with significant inter-annual variation in food availability may explain 

the lower frugivore biomass in humid forests as opposed to dry forests (Ganzhorn et al. 

1999b; Donati 2002). 

CONCLUSION 

While the littoral forest can be considered aseasonal regarding climate, there are 

however consistent year to year patterns in flowering and fruiting. Intra-annual differences 

occur in phenophases leading to periods of abundance and scarcity. Typically for the 

littoral forest and other Malagasy humid forests in general but very different from dry 

Malagasy forests, is the inter-correlation of all phenophases that peak almost 

simultaneously in the wettest period (November-February). Irradiance seems to be the 

most important abiotic factor in triggering phenophases, due to its importance in 

photosynthetic processes especially at this extreme southern latitude for a tropical wet 

forest. Fruit trails and phenological transects have shown to be complementary methods 

by unravelling fruiting patterns of different life forms and different sub-samples of all 

present plant species, including both fruit data on canopy and ground level. 


This study was carried out under the Accord de Collaboration between Department of Animal 

Biology and Department of Anthropology of the University of Antananarivo, the Institute of Zoology 

of Hamburg University and QIT Madagascar Minerals (QMM). Our thanks go out to QMM for 

providing logistics and infrastructure at the campsite. In particular we acknowledge Manon 

Vincelette, Jean-Baptiste Ramanamanjato, Laurent Randrihasipara of the QMM Environmental and 

Conservation Team for following up phenology and climatic data (QMM unpubl. data) at the site 

after we left and sharing these data with us. Many thanks as well to Give Sambo and Dauphin 

Mbola, the two local field assistants who helped with the collection of phenology and fruit trail data. 

The first author is supported by a grant of the Belgian Fund for Scientific Research, Flanders 

(FWO). We further thank the Flemish Government for structural support to the Centre for Research 

and Conservation (CRC) of the Royal Zoological Society of Antwerp (RSZA). The second author 

was supported by a doctoral grant of the Italian Ministry for Scientific Research (MURST) and the 

University of Pisa. Many thanks as well to Jörg Ganzhorn, Linda Van Elsacker and Silvana 

Borgognini Tarli for their support during our research. 

30

Phenology 

Appendix I. Overview of plant species included in both phenological transect (PH) and fruit trail 

(FT) with indication of life form (LT: large tree, ST: small tree or shrub, V: vine, E: epiphyte). 

Family, species and vernacular name are given. For species without vernacular names, codes 

(x1 or FT1) were given. 

Family Name Species name Vernacular name PH FT Life form 

Anacardiaceae Poupartia chapelieri sisikandrongo x x LT 

Campnosperma micranteia roandria x LT 

Rhus thouarsii kangy x LT 

Annonaceae Monanthotaxis cf. malacophylla vahihazo x x V 

Polyalthia capuronii menapeka x V 

Polyalthia madagascariensis fotsivavo x LT 

Polyalthia sp.1 fotsivavo géante x x LT 

Apocynaceae Cabucala madagascariensis tandrokosy x ST 

Araliaceae Cuphocarpus aculeatus voatsilana marécage x x LT 

Schefflera rainaliana voatsilana x x LT 

Arecaceae Dypsis fibrosa boakandambo x x ST 

Dypsis prestoniana boakabe x x LT 

Dypsis scottiana raosy x ST 

Asteraceae Senecio sp. witte pluisjes x ST 

Bignoniaceae Phyllarthron ilicifolium zahambe x LT 

Ophiocolea delphinensis akondronala x ST 

Burseraceae Canarium boivinii ramy x x LT 

Canellaceae Cinnamosma madagascariensis vahabatra 3eM x LT 

Capparaceae Crataeva obovata belataka x x ST 

Physena madagascariensis FT 85 x ST 

Celastraceae Mystroxylon aethiopicum voavoantatsimo x x ST 

Elaeodendron sp. aramboazo x LT 

Polycardia phyllanthoides fandrianakanga x LT 

Clusiaceae Garcinia chapelieri haziny tomate x LT 

Garcinia cf/aff. madagascariensis disaky kely x x ST 

Mammea bongo disaky be x LT 

Symphonia fasciculata haziny voany be x LT 

Symphonia sp. haziny fleur rouge x x LT 

Calophyllum sp. vitano x LT 

Combretaceae Terminalia fatraea katrafa x x LT 

Connaraceae Agelaea pentagyna rehiba vahimainty x x V 

Dichapetalaceae Dichapetalum sp. vahikatepoka x V 

Dilleniaceae Dillenia triquetra varikanda x LT 

Ebenaceae Diospyros myriophylla forofoka x LT 

Diospyros sp.1 hazomainty blanc x ST 

Diospyros sp.2 hazomainty x x ST 

Diospyros sp.3 hazomainty kely x ST 

Diospyros sp.4 FT 82 x ST 

Elaeocarpaceae Elaeocarpus alnifolius sanga x LT 

Ericaceae Vaccinium eminense tsilantria x ST 

Erythroxylaceae Erythroxylum buxifolium fangora sp.1 x ST 

Erythroxylum nitidilum fangora sp.2 x LT 

Euphorbiaceae Anthostema madagascariensis bamby x LT 

Blotia leandriana x225 x ST 

Blotia mimosoides fantsikaitra x LT 

31

Chapter 1 

Appendix I Continued 


Euphorbiaceae Euphorbia tetraptera famanta x ST 

Macaranga perrieri mocarana x LT 

Uapaca ferruginea voapaky lahy x x LT 

Uapaca littoralis voapaky vavy x x LT 

Uapaca thouarsii voapaky vavy ZJ x LT 

genus indet. randramboay x ST 

Fabaceae Cynometra cf. cloiselii mampay x x LT 

Intsia bijuga harandrato x x LT 

Phylloxylon xylophylloides sotro x x LT 

Flacourtiaceae Bembicia uniflora bemalemy x x LT 

Homalium albiflorum tapinandro 1 

x LT 

Homalium albiflorum lapivatra 1 

x LT 

Homalium louvelianum ramirisa x LT 

Homalium planiflorum hazofotsy x x ST 

Homalium sp. marakoditra x x LT 

Scolopia orientalis zoramena x x LT 

Grossulariaceae Brexia sp. kambatrikambatri x x ST 

Hamamelidaceae Dicoryphe stipulaceae zorala x LT 

Hippocrataceae Salacia madagascariensis voatsimatra x x V 

Icacinaceae Apodytes sp. nov. hazomamy an ala x LT 

Lauraceae genus indet. varongy sp2 x LT 

Beilschmiedia madagascariensis resonzo x LT 

Cryptocarya sp. tavolohazo x x LT 

Ocotea sp. varongy be x LT 

Liliaceae Dracaena reflexa var. nervosa falinandrobe 1 

x x ST 

Dracaena reflexa var. nervosa falinandrokely 1 

x ST 

Dracaena reflexa var. nervosa tavolobotroka 1 

x ST 

Loranthaceae Bakerella ambongoensis velomihanto sp1 x x E 

Bakerella sp. velomihanto sp2 x E 

Loganiaceae Anthocleista longifolia lendemilahy x x ST 

Menispermaceae Burasaia madagascariensis faritsaty x x ST 

Monimiaceae Tambourissa purpurea ambora 1 

x x ST 

Tambourissa purpurea amboralahy 1 

x ST 

Moraceae Trilepisium madagascariense beronono x x LT 

Myristicaceae Brochoneura madagascariensis mafotra x x LT 

Myrsinaceae Monoporus spathulatus FT 88 x V 

Myrtaceae Eugenia cloiselii ropasy sp.1 x x LT 

Eugenia sp.1 ropasy sp.2 x LT 

Eugenia sp.2 ropoaky x x LT 

Syzygium emirnense rotry sosimaro x ST 

Syzygium sp.1 rotry ala x LT 

Syzygium sp.2 rotry mena x x LT 

Ochnaceae Campylospermum obtusifolium hazombato x ST 

Diporidium ciliatum sakambolava x x ST 

Oleaceae Jasminum kitchingii vahifotsy kely x V 

Noronhia cf. lanceolata hazondraotry x ST 

Noronhia ovalifolia zorafotsy x x LT 

32

Phenology 



Oleaceae Noronhia sp.1 belavenoka x x LT 

Olea sp. vahabatra x x LT 

Pandanaceae Pandanus aff. longistylus fandranabo x LT 

Pandanus dauphinensis vakoanala x ST 

Pandanus rollotii fandranabotonboky x x LT 

Pittosporaceae Pittosporum polyspermum x202 x ST 

Rhopalocarpaceae Rhopalocarpus coriaceus tsilavimbinanto x x LT 

Rubiaceae Canthium variistipula fantsikaitramainty x x ST 

Cremocarpon lantzii x220 x ST 

Ixora sp. x203 x ST 

Morinda cf. umbellata vahilengo x V 

Peponidium sp. fantsikaidroka x LT 

Mapouria aegialodes x210a x ST 

Psychotria sp. tanatananala x ST 

Mapouria sp. x210 x ST 

Pyrostria sp. fantsikaitrafotsy x LT 

Rothmannia mandenensis taholagna x x LT 

Saldinia littoralis mangavoa x ST 

Tarenna thouarsiana FT 62 x ST 

Tricalysia cf. cryptocalyx hazongalala x x ST 

Tricalysia sp. kotofotsy x ST 

Rutaceae Vepris eliottii ampoly 1 

x x LT 

Vepris eliottii ampolylahy 1 

x x LT 

Sapindaceae Macphersonia radlkoferi sanirambaza x LT 

Filicium decipiens lahinvoatsilana x LT 

Plagioscyphus jumellei ambirimbarika pionair x ST 

Tina thouarsiana sagnirambavy x x LT 

Tinopsis conjugata sagnira sp.3 x LT 

Sapotaceae Donella delphinensis hazomteraka x LT 

Fauchera hexandra natohetiki x x LT 

Sideroxylon beguei var. sabouraui ambirimbarika x x LT 

Sarcolaenaceae Leptolaena sp. fotonbavy x x LT 

Sarcolaena multiflora meramaintso x x LT 

Schizolaena elongata fotondahy x x LT 

Smilaceae Smilax anceps fandrikatani x V 

Sphaerosepalaceae Podocarpus madagascariensis harambilo x LT 

Taccaceae Tacca leontopetaloides tavolo x H 

Theaceae Asteropeia micraster fanolamena x x LT 

Asteropeia multiflora fanolafotsy x x LT 

Ulmaceae Trema orientalis andrarezona x LT 

Verbenaceae Clerodendrum sp. nofotrako marecage x x LT 

Violaceae Rinorea pauciflora memboloa x ST 

? ? FT 100 x ST 

? ? liane fleur jaune x V 

? ? FT 112 x ST 

? ? FT 20 x ST 

? ? FT 49 x ST 

33

Chapter 1 



? ? FT 51 x ST 

? ? FT 8 x ST 

? ? FT 80 x ST 

? ? FT 91 x ST 

? ? FT 93 x LT 

? ? FT 95 x ST 

? ? FT 96 x ST 

? ? menahi x ST 

? ? vahifotsy be x x V 

1 as indicated by their vernacular name certain plant species correspond to the same scientific name. 

They could represent different ecotypes of the same species or different species that have no taxonomic 

names yet. As this is difficult to affirm at the moment we preferred including all plant species as separate 

units throughout this paper. 

34

varika 

fanihy 

AU BOUT DU MONDE 

stock papier 

hygienique 

mpondiky 

graines sautées 

vazana 

aide 

morale 

bohaky 

compagnie 

pour la nuit? 

Kadoffe, 

raviny feno?


Drawing © Giuseppe Donati 2000 

‘Ataovy toy ny voankazo an ala, 

ka ny mamy no atelemo, ny mangidy aloavy’ 

Do as with the fruits of the forest, 

the sweet ones you swallow, the bitter ones you spit out

Tree dispersal strategies 

Tree dispersal strategies in the 

littoral forest of Sainte Luce 

(south-east Madagascar) 

AN BOLLEN, LINDA VAN ELSACKER , JÖRG GANZHORN 

OECOLOGIA (SUBMITTED) 

ABSTRACT 

Zoochory is the most common mode of seed dispersal for the majority of plant species in 

the tropics. Based on the assumption of tight plant-animal interactions several 

hypotheses have been developed to investigate the origin of life history traits of plant 

diaspores and their dispersers, such as species-specific co-evolution, the low-high 

investment model (low investment in single fruits but massive fruiting to attract many 

different frugivores versus high investment in single fruits and fruit production for 

extended periods to provide food for a few specialised frugivores), and the evolution of 

dispersal syndromes which represent plant adaptations to larger taxonomic groups of 

dispersers. To test these hypotheses the frugi-granivorous vertebrate consumers and 

dispersal strategies of 34 tree species were determined in the littoral forest of Sainte Luce 

(SE-Madagascar) with the help of fruit traps and tree watches. The impact of fruit 

consumers on the seeds was determined based on detailed behavioural observations. 

Phenological, morphological and biochemical fruit traits from tree species were measured 

to look for co-variation with different types of dispersal. There was no evidence for 

species-specific co-evolution nor any support for the low/high investment model. However 

diaspores dispersed by birds, mammals or both groups (mixed dispersed tree species) 

differ in the size of their fruits and seeds, fruit shape, and seed number, but not in 

biochemical traits. Five large-seeded tree species seem to depend critically on the largest 

lemur, Eulemur fulvus collaris, for seed dispersal and recruitment. However, this does not 

represent a case of tight species-specific co-evolution. It rather seems to be a 

consequence of the extinction of larger frugivorous birds and lemurs which also might 

have fed on these large fruits. It seems that the species-poor guild of frugivores in 

Madagascar and in the littoral forest in particular did not allow the evolution of specialised 

dispersal strategies. In particular, the low species diversity of avian frugivores resulted in 

significantly few bird fruits compared to other sites. 

Introduction 

In tropical forests, active transport of seeds by animals (zoochory) is the most common 

means of seed dispersal, involving more than 75% of all plant species (McKey 1975; 

Charles-Dominique et al. 1981; Howe and Smallwood 1982; Janson 1983; Gautier-Hion 

et al. 1985; Jordano 1992). Seed dispersal away from the parent tree seems essential for 

37

Chapter 2 

the successful establishment of seedlings (e.g. Janzen 1970; Connell 1971; Howe and 

Smallwood 1982; Terborgh et al. 2001). Therefore attracting frugivores is crucial for a 

plant in order to ensure reproduction. 

Several hypotheses have been proposed to explain the evolution of zoochory. First, 

under the assumption that the evolution of life history traits of plants, their diaspores and 

their consumers are mutually dependent, the most restrictive hypothesis assumes very 

tight co-evolutionary relationships between one single fruit and frugivore species. So far 

no evidence has been found in support of this hypothesis (Howe and Smallwood 1982; 

Herrera 1984; Howe 1984; Gautier-Hion et al. 1985; Fisher and Chapman 1993; 

Chapman 1995; Erikkson and Ehrlen 1998; Lambert and Garber 1998). Secondly, from 

the plants’ point of view McKey (1975) postulated different patterns of resource 

investment in plants that rely on multiple versus specialised seed dispersers. According 

to this model low investment plants (generalists) invest little in single fruits but display 

large fruit crops during a short fruiting period to attract a large variety of opportunistic 

frugivores. In contrast, the high investment plants (specialists) have fruit pulp with higher 

nutrient content, a more limited fruit production and extended fruiting seasons that attract 

few specialists (Howe 1979). Again not much support could be found for this idea so far 

(Dowsett-Lemaire 1988; Wenny 2000; Wütherlich et al. 2001; but see Wheelwright 1986). 

Thirdly the hypothesis of dispersal syndromes postulates broad morphological 

adaptations of fruit traits associated with different consumer taxa, mostly distinguishing 

diaspores dispersed by birds, mammals or both groups (mixed fruits). This model 

emphasizes the taxonomy and phylogenetic heritage of dispersers with their associated 

sensory capacities (Van der Pijl 1969; Janson 1983; Knight and Siegfried 1983; Gautier- 

Hion et al. 1985; Martinez del Rio 1994; Corlett 1996; Kalko et al. 1996; Korine et al. 

2000; Pizo 2002; Voigt 2001). 

The forests of Madagascar provide good opportunities to test the above hypotheses. 

There is a high level of floral and faunal endemism in Madagascar with 96% endemics 

among the tree species (Schatz 2001), more than 50% of its birds (Langrand 1990) and 

90% of its mammals (Goodman et al. in press). This results in communities with an 

evolutionary history, which is largely independent from the communities for which the 

hypotheses have been developed and tested and therefore allows independent tests for 

these hypotheses. For this, we selected 34 different tree species of the evergreen littoral 

forest of south-eastern Madagascar in order to answer the following questions; 

1. Which frugivores feed on the different fruiting tree species and what is their impact on 

the seeds? 

2. Is there evidence for tight species-specific co-evolution? 

3. Can tree species be categorized as low or high investment species according the 

McKey’s hypothesis? 

4. Are there trees which rely on certain taxonomic groups for dispersal and if so: do fruits 

with different dispersal strategies vary significantly in their morphological and biochemical 

attributes? 

METHODS 

Study site and the frugivorous guild 

The littoral forest of Sainte Luce (24º45'S 47º11'E) is located in south-eastern 

Madagascar, 50km north of Fort-Dauphin. The first author collected data between 

November 1999 and February 2001 in a 377-ha forest fragment, called ‘S9’. Average 

38


annual rainfall is 2,690mm. Mean monthly temperature is about 23°C and ranges from 

12°C to 33°C. For a more detailed description of the study site we refer to Bollen et al. (in 

press, Chapter 4). At Sainte Luce, the trophical guild of frugivores consists of 13 strictly or 

partially frugivorous vertebrate species (Table 1). 

Focal tree species 

For this study 34 tree species, which are thought to be important food sources for certain 

animal species were chosen (App. I). Fruit traps were installed under 29 tree species and 

tree watches were carried out for 27 tree species. Herbarium specimens of all taxa were 

collected. Vernacular ‘antanosy’ names were provided by local research assistants. 

Scientific names were obtained after determination of voucher specimens at the national 

herbaria of Antananarivo with the help of botanists from Missouri Botanical Garden (App. 

I). Voucher specimens were deposited at the Missouri Botanical Garden of Antananarivo 

(Madagascar). 

Fruit traps 

In order to estimate relative fruit production and consumption, fruit traps were placed on 

the ground under the tree in the zone of fruit fall (following Goodman et al. 1997a). In total 

29 tree species were studied, fruit traps being placed under one individual per tree 

species. Each fruit trap was 1m² in size and was made out of black plastic sheeting. Fruit 

traps were blocked at the sides with upturned edges to avoid loss of the content due to 

rain washing. The number of fruit traps per individual tree depended on the crown size, 

on average being one trap per 3.2m² (±1.6 SD) of the crown area. Total fruit crop size 

was obtained by extrapolating the data from the area sampled by the traps to the total 

crown area. Fruit traps were inspected and emptied every other morning between 06h00- 

10h00 throughout the fruiting period. Ideally, fruit traps were installed before fruit ripening 

had started and were removed at the end of the fruit production. For five tree species, 

fruit traps were installed shortly after the onset of fruiting, due to a delay in noticing that a 

species was fruiting (Table 2). 

Analyses of the contents of each fruit trap involved counting and checking the 

condition of fruits and seeds, which were coded using the following categories; 

Neutral effect: intact fruits (unripe, ripe or rotten) 

pulp partially eaten, intact seeds 

pulp completely consumed, intact seeds 

Dispersal: empty fruit husks, pulp eaten and seeds swallowed 

Predation: partially eaten seeds or empty seed husks with gnaw marks 

Percentages of all categories were calculated per tree species. Additionally, defecated 

seeds from the focal tree or other tree species present in the fruit traps were scored as 

well. Several animals leave distinctive feeding marks on the discarded fruit and seed 

remains. Therefore it was often possible to determine the consumer species or at least 

the larger taxonomic group. 

Tree watches 

Observations on feeding assemblages of frugivores were carried out during the peak 

fruiting period of 27 tree species. These so-called ‘tree watches’ (Chapman and 

Chapman 1996; Scharfe and Schlund 1996; Goodman et al. 1997a; Böhning-Gaese et al. 

1999) consisted of intermittent 36h observations (two cycles 06h00-00h00) of one 

39

Chapter 2 

40 

Table 1. The frugivore guild at Sainte Luce with indication of their diet (F: frugivorous, G: granivorous, O: omnivorous), their activity 

budget (D: diurnal, N: nocturnal, C: cathemeral), feeding height (A: arboreal, T: terrestrial), body mass (in g) and length (in cm) 

Role 3 

Body 1,2 

Length 

Body 1 

mass 

Feeding 

height 

Active 

and their potential role as seed dispersers (D) or seed predators (P) in this ecosystem. 

Family 

Scientific name 

English name 

Diet 

D 

D 

P 

D 

P 

P 

32 

28 

28 

24 

35 

50 

215 

ND 

190 

45 

218 

ND 

A 

A 

T 

A 

A 

A 

D 

D 

D 

D 

D 

D 

F 

F 

G 

F 

G 

G 

Malagasy Green Pigeon 

Malagasy Blue Pigeon 

Madagascar Turtle Dove 

Madagascar Bulbul 

Lesser Vasa Parrot 

Greater Vasa Parrot 

Treron australis 

Alectroenas madagascariensis 

Streptopelia picturata 

Hypsipetes madagascariensis 

Coracopsis nigra 

Coracopsis vasa 

AVES 

Columbidae 

Pycnonotidae 

Psittacidae 

P 

P 

15-23 

10-16 

100 

88 

AT 

AT 

N 

N 

O 

O 

Black Rat 

Webb's Tuft-Tailed Rat 

Rattus rattus 

Eliurus webbi 

MAMMALIA 

RODENTIA 

Muridae 

Nesomyinae 

D 

23-27 

500-750 

A 

N 

F 

Madagascar Flying Fox 

Pteropus rufus 

CHIROPTERA 

Megachiroptera 

Pteropodidae 

D 

D 

D 

D 

40-47 

12.5 

25 

20 

PRIMATES 

Lemuridae Eulemur fulvus collaris 

Collared Brown Lemur F C A 2000-2300 

Cheirogaleidae Microcebus rufus 

Brown Mouse Lemur O N A 42 

Cheirogaleus major 

Greater Dwarf Lemur O N A 443 

Cheirogaleus medius 

Fat-tailed Dwarf Lemur O N A 119-282 

1 

Data from Langrand (1990), Fietz and Ganzhorn (1999), Garbutt (1999), Ganzhorn et al. (1999a), Donati (2002). 

2 

Body length (BL) is total length for birds and bats but head/body length for lemurs and rodents. 

3 

Data from Goodman et al. (1997b); Ganzhorn et al. (1999a); Bollen and Van Elsacker (2002a); Donati (2002); Bollen (unpubl. data).


individual tree per species. Six species were observed for less than 36 hours due to a 

very short fruiting period or heavy rains during the fruiting peak (Table 3). Observations 

were made within 10m of the focal tree, which allowed an unobstructed view of 

approximately 70-100% of the canopy. During the day binoculars (Leica 10x42) were 

used while at night a headlight (Petzl), which reflects the tapetum lucidum of the eyes of 

lemurs, flying foxes and rats, aided observations. Additionally, if moonlight conditions 

were favourable, night goggles (Litton Electron Devices) were used as well. Tree watches 

provided data on the animal species feeding in the tree and their handling and feeding 

behaviour. This allowed us to classify them as seed-dispersers, seed droppers or fruit 

pulp consumers and seed predators. This behaviour could be easily observed during the 

day but not always at night when one needed to rely on evidence such as falling fruit 

husks, fruits or seeds. Species visiting the tree without eating were not considered in the 

analyses. 

According to the model of dispersal syndromes, tree and fruit characteristics evolved 

in response to the community of frugivores and their taxonomic affinities. Both literature 

and detailed observations during tree watches allowed us to classify the different 

consumer species as seed dispersers, fruit-pulp consumers and/or seed predators 

(Goodman et al. 1997a; Ganzhorn et al. 1999a; Bollen and Van Elsacker 2002a, Chapter 

3, 3a; Donati 2002). Based on this knowledge, the following dispersal groups were 

distinguished; 

-mixed fruits: eaten by both birds and mammals 

-bird-fruits: eaten by fruit pigeons and/or bulbul 

-mammal-fruits: eaten by lemurs and flying foxes 

-fruits eaten only by Eulemur fulvus collaris 

Morphological and biochemical characterisation 

The number of seeds per fruit were counted, fruits and seeds were weighed fresh using 

spring or electronic balances and measured using callipers with 0.01g and 0.01mm 

precision, respectively. In addition the ratio between length and width (L/W) was used as 

an index of fruit shape (Pizo 2002). Most fruits are typically zoochorous, including soft 

and juicy drupes or berries. 

Ripe fruits were dried in the sun or in a drying oven, ground to pass through a 2mm 

sieve, and dried again overnight at 50-60°C prior to analyses. Lipids were determined by 

the Soxleth method. Total nitrogen (N) was determined using the Kjeldahl procedure. 

Multiplying N by 6.25 converted total nitrogen to crude protein. However it should be 

noted that a conversion factor of 6.25 overestimates the protein actually available for 

frugivores in some fruits (Levey et al. 2000). Soluble carbohydrates and procyanidin 

(condensed) tannins were extracted with 50% methanol. Concentrations of soluble sugar 

were determined as the equivalent of galactose after acid hydrolization of the 50% 

methanol extract. This measurement correlates well with concentrations obtained with 

enzymatic analyses of glucose, fructose and galactose (Ganzhorn and Tomaschewski 

unpubl. data). Concentrations of procyanidin tannin were measured as equivalents of 

quebracho tannin (Oates et al. 1977; Porter and Hemingway 1990). Samples were 

analysed for acid detergent fibre (ADF) (Goering and Van Soest 1970; Van Soest 1994) 

modified according to the instructions for use in an ANKOM FIBER ANALYZER. 

Biochemical analyses were carried out at the Institute of Zoology, Department of Ecology 

and Conservation (University Hamburg). 

41

Chapter 2 

Statistical analyses 

To visualise dispersal types and to test for fruit trait co-variation consistent with these 

dispersal types, two principal component analyses (PCA) were conducted using 

morphological and biochemical fruit trait data of 29 tree species. As data were not 

normally distributed, morphological and biochemical traits were log-transformed prior to 

analyses. Factor loadings were used to determine the strength of association of each fruit 

trait with each principal component. Mann Whitney U tests, Spearman rank correlations 

and Kruskal Wallis were carried out according to Siegel (1956) with the statistical 

software SAS for Windows. 

RESULTS 

Fruit traps 

Considering fruit trap analyses, there was substantial variation in the three main 

categories (neutral, dispersal, predation) among different tree species (Table 2, Fig. 1). 

Based on all fruit trap data, a substantial proportion (median: 70%, quartiles 39-98%) of 

seeds remained under the parent plant; including unripe, ripe and rotten fruits (47%) that 

had not been eaten as well as partially eaten fruits that still contained their seed (13%) or 

intact seeds (9%) that were dropped after the fruit pulp had been swallowed (Table 2). In 

all these cases the animals had a neutral effect on the seeds, since they were neither 

dispersed nor predated. For 18 species the number of fruits or seeds dropped under the 

parent plant was higher than 50% (Fig. 1). Evidence of seed dispersal involved empty 

fruit husks (median 9%, quartiles 0-61%) that were discarded after consumers swallowed 

the fruit pulp and seeds. This proportion was relatively large (>50%) for only eight tree 

species. The proportion of predated seeds was low (median 2%, quartiles 0-5%) and 

exceeded 5% for only six tree species (Table 2, Fig. 1). 

Faecal droppings, collected in fruit traps, provided complementary but nonquantitative 

information on seed dispersal. Ninety percent of all faecal samples collected 

under focal trees were found to contain seeds from other tree species, thus indicating 

seed dispersal. Of the 29 tree species sampled, seeds of 13 species were found in 

droppings of Eulemur fulvus collaris and one species in droppings of Alectroenas 

madagascariensis. For the majority of these species (9 out of 14) hardly any empty fruit 

husks were found in the fruit traps as proof of seed dispersal (0-3%) (Table 2) because 

dispersed seeds involve completely swallowed fruits, which are not accounted for in this 

method. For the other remaining 5 species, more empty fruit husks (42-89%) were 

retrieved in the fruit traps, indicating that, here, consumers most often scooped out and 

swallowed fruit pulp and seeds and discarded the remaining fraction. 

Fourteen plant species had the majority of their fruits eaten. Non-eaten fruits were 

most abundant in eight species and equal percentages of eaten and non-eaten fruits 

occur for 7 species (Table 2). Identification of the consumers was based on faecal 

droppings or feeding marks. On the species level, faecal droppings of E. f. collaris and A. 

madagascariensis are obviously distinguishable and recognizable by size and 

consistency. The bill mark is typical for both Coracopsis spp. Stripped off pulp parts are 

typical marks of Pteropus rufus’ sharp teeth. Lemur tooth marks can be species-specific 

based on their size, but were most often assigned to larger taxonomic group of nocturnal 

lemurs (involving Cheirogaleus spp. and Microcebus rufus) or lemurs (involving all four 

lemur species) (App. II). Rodents leave typical gnawing marks, but these do not always 

42


allow identification to species-level. The other frugivorous bird species most often 

swallow the fruit entirely and thus leave no identifiable marks. 

Table 2. Percentages of the different categories used in fruit trap analyses, involving neutral or 

dropped, dispersed and predated seeds, eaten and non-eaten fruits are given. For these different 

categories highest percentages are given in bold. An asterisk indicates non-quantitative dispersal 

evidence by faecal dropping, which leads to underestimation of this category. 

(P and S part. Eaten: pulp and seeds partially eaten, FS: faecal seeds, 

Am: Alectroenas madagascariensis, Efc: Eulemur fulvus collaris). 

Impact on seed: Predation 

Category: Intact P part Intact Sum Empty FS S part Eaten Not 

fruits eaten 1 seeds 2 

husks eaten eaten 

Apodytes dimidiata 3 

21 66 13 100 0 0 79 21 

Brexia sp. 19 2 4 25 69 7 81 19 

Brochoneura acumineata 14 1 15 76 10 86 14 

Burasaia madagascariensis 8 24 24 56 9 36 93 8 

Canarium boivinii 54 16 20 90 8 2 46 54 

Canthium variistipula 3 

23 5 72 100 0* Efc 0 77 24 

Cinnamosma madagascariensis 5 3 31 39 61 0 95 5 

Diospyros sp. 3 

11 10 32 53 42* Efc 5 89 11 

Dypsis prestoniana 53 15 30 98 0* Am 2 47 54 

Elaeocarpus alnifolius 54 36 9 99 0* Efc 1 46 55 

Eugenia cloiselii 67 24 9 100 0* Efc 0 33 67 

Eugenia sp. 82 6 12 100 0* Efc 0 18 82 

Garcinia cf. madagascariensis 63 7 70 28 3 37 63 

Leptolaena multiflora 95 2 97 3* Efc 0 5 95 

Olea sp. 79 18 1 98 0* Efc 3 21 79 

Poupartia chapelieri 4 1 24 29 69* Efc 2 96 4 

Rothmannia mandenensis 31 31 69 0 100 0 

Sarcolaena multiflora 11 11 89* Efc 1 89 11 

Schizolaena elongata 78 14 1 93 3* Efc 4 22 79 

Scolopia orientalis 35 12 2 49 45 7 65 35 

Syzygium sp.1 54 21 2 77 23 1 47 54 

Syzygium sp.2 3 

Neutral effect Dispersal 

58 20 21 99 0* Efc 1 42 58 

Terminalia fatraea 49 34 8 91 9 1 51 49 

Tina thouarsiana 40 3 2 45 1 54 60 40 

Uapaca ferruginea 49 49 50* Efc 2 52 49 

Uapaca littoralis 45 1 6 52 48* Efc 1 55 46 

Uapaca thouarsii 64 18 82 2 16 36 64 

Vepris eliotii 21 7 28 67 5 79 21 

Vepris fitoravina 10 5 15 85 0 90 10 

median 47 13 9 70 9 2 55 46 

quartiles 18-59 3-21 9-24 39-98 0-61 0-5 42-8614-58 

1 

This category refers to fruits of which the pulp is partially eaten, but intact seeds remain. 

2 

This category refers to fruits of which all pulp and husk are eaten, but intact seeds were dropped. 

3 

For these five tree species fruit traps were installed shortly after the onset of fruiting, due to a delay in 

noticing that this species was fruiting. 

43

Chapter 2 

Overall fruit trap data documented 77 plant-animal interactions of which the majority 

(n=63) could be identified on feeding marks alone. Additional interactions (n=14) were 

revealed through analyses of faecal samples. For seven interactions, faecal samples 

confirmed data from feeding marks. Of all interactions, 65% (n=50) could be assigned to 

a single consumer species, while 35% could only be assigned to a larger taxonomic 

group, such as lemurs in general (n=9), nocturnal lemurs (n=10) or rodents (n=8). 

100% 

Fig 1. Indication of percentages of different categories of seed dispersal, predation, and neutral 

seed dropping per plant species. X includes six plant species being Cv, Ad, Ec, E, Ea, S2. For 

abbreviations of tree species, see Appendix I. 

Tree watches 

Tree watches were carried out for 27 species, for a total of 928 observation hours 

(median of 36h/species) during 107 observation days (median 6 days/sp., range 3-10 

days/sp.) at the peak of the fruiting periods (Fig. 2). Observational data on the feeding 

behaviour revealed whether species had a neutral, positive or negative impact on the 

seeds. For Coracopsis nigra, Eliurus webbi and Rattus rattus the destruction of seeds is 

very clear. Streptopelia picturata feeds on seeds on the ground and is likely a seed 

predator but detailed feeding observations of this very shy dove were not possible. The 

lemur and flying fox species can act as seed dispersers but often also drop seeds under 

the parent plant during fruit handling or after swallowing fruit pulp and hence have a 

neutral effect on seeds. Hypsipetes madagascariensis, Treron australis and Alectroenas 

madagascariensis act as seed dispersers swallowing all fruits entirely (Bollen et al., 

Chapter 3). 

If we consider all combinations of consumer-plant species interactions (n=100) of the 

plant species that were included in both fruit trap analyses and tree watches, most of 

these (n=62) were confirmed by both tree watches and fruit trap analyses, even if some 

of the fruit trap data only referred to the larger taxonomic group (App II). In general, tree 

watches further refined the fruit trap data to species level but also added 24 new 

44 

90% 

80% 

70% 

60% 

50% 

40% 

30% 

20% 

10% 

0% 

X Dp- 

O 

Lm Se Tf Cb Ut S1 G Bm D Ul Uf So Tt Cm Rm Pc Ve B Vf B a Sm 

neutral dispersal predation


consumer-plant interactions to the list. Fruit trap data detected 14 interactions that were 

not confirmed by tree watches and these involved mainly shy consumer species that 

were difficult to observe. In Appendix II, twelve tree species were not considered, as data 

on these species had not been sampled by both fruit traps and tree watches. 

Data on general phenology of ripe fruits (Bollen and Donati, Chapter 1) are presented 

in Figure 2 to indicate when the fruiting peak of the different focal tree species occurred 

and how this related to the overall fruit availability of canopy tree species in the littoral 

forest. The number of fruit tree species selected each month more or less corresponds 

with the monthly fruiting diversity throughout the year 2000. 

Dispersal strategies 

It is difficult to actually ‘test’ co-evolution but this paradigm is based on tight interactions 

between one single fruit and disperser species. In this respect there are five plant species 

that are exclusively dispersed by Eulemur fulvus collaris, namely Canarium boivinii 

(ramy), Diospyros sp. (hazomainty), Eugenium sp. (ropasy sp. 2), Rothmannia 

mandenensis (taholagna), Cinnamosma madagascariensis var. namoronensis (vahabatra 

3eM). These fruits were significantly heavier (fruit weight: Z=3.26, P=0.0011) and longer 

(fruit length: Z=3.38; P=0.0007) than the other fruits. No significant difference could be 

found for any of the other morphological and biochemical traits. 

According to McKey's (1975) model, high investment trees have small crop size, long 

fruiting period and few seed dispersers, whereas the opposite is valid for low investment 

trees. However no significant correlations could be found among these traits (fruiting 

period-crop size: rs=0.18, P=0.41; fruiting period-number seed dispersers: rs=-0.08, 

P=0.70; crop size-number of seed dispersers rs=0.13, P=0.53). Investment was 

considered to be represented by the concentrations of nutrients such as sugars, lipids 

and protein. The only significant correlation found was among protein and lipid content 

(rs=0.51, P=0.005) and among sugar content and fruiting period (rs=-0.43, P=0.04). 

Sugar, lipid and protein content were not correlated with the number of seed dispersers 

as the model predicts. Any division in low and high investment trees thus seems to be 

arbitrary here and does not represent a valid classification to test the McKey model. 

Based on data from fruit traps and tree watches, 29 tree species were divided into 

species in which fruits were eaten and dispersed only by Eulemur fulvus collaris, by birds, 

by mammals or by both groups. This classification into disperser ‘syndromes’ is based on 

the taxonomic composition of the consumers (Table 3). Mammal fruits account for 55% 

(n=16) of all species, while both mixed (n=5) and specialist (n=5) tree species accounted 

for 17% each and bird dispersed species (n=3) for 11%. First we tested for correlations 

among all fruit traits, both morphological (n=6) and biochemical (n=5). Using sequential 

Bonferroni adjustment only four Spearman rank correlations remained significant: fruit 

weight and fruit length (rs=0.76, P

Chapter 2 

Table 3. Morphological and biochemical traits used for PCA analyses. Weights are given in 

g, length in mm. Fruit shape is fruit length divided by fruit width. Biochemical components are 

given in percentages dry weight. The disperser type we assigned the plant species is given 

in the last column. ND: no data available, Efc refers to fruits that are dispersed by Eulemur fulvus 

collaris only. 

Morphological 

Plant species Number Fruit Fruit Seed Seed Fruit 

seeds weight length weight length shape 

Apodytes dimidiata 1 0.45 12.34 0.23 10.61 1.55 

Brexia sp. 1 1.57 20.19 0.21 15.12 1.64 

Canarium boivinii 1 9.69 31.14 4.17 27.01 1.29 

Canthium variistipula 2 0.31 7.80 0.06 6.0 0.85 

Cinnamosma madagascariensis 10 6.22 21.99 0.14 8.42 0.98 

Dyospiros sp. 5 16.48 30.71 1.89 19.74 0.93 

Dypsis prestoniana 1 0.59 14.85 0.34 12.92 1.85 

Eugenia cloiselii 1 1.59 13.20 1.26 11.64 0.90 

Eugenia sp. 1 4.20 23.81 1.64 15.33 1.33 

Leptolaena multiflora 2 0.06 5.60 0.01 2.90 1.19 

Ludia antanosarum 6 1.04 12.47 2.98 3.22 1.07 

Macaranga perrieri 1 0.04 4.55 0.03 3.06 1.00 

Olea sp. 1 0.90 16.98 0.80 15.85 1.56 

Polyalthia madagascariensis 1 0.30 12.22 0.12 7.80 1.73 

Polyscias sp. 1 0.04 5.03 0.01 3.64 1.31 

Poupartia chapelieri 1 0.54 15.43 0.36 15.24 1.52 

Rothmannia mandenensis 100 35.33 40.36 0.03 4.35 1.04 

Sarcolaena multiflora 5 0.67 14.11 0.01 2.73 1.32 

Schizolaena elongata 2 0.77 8.89 0.01 3.05 0.63 

Scolopia orientalis 3 0.52 10.52 0.02 3.75 1.11 

Syzygium sp.1 1 0.64 10.23 0.64 9.32 0.90 

Syzygium sp.2 1 0.54 9.55 0.31 6.56 0.98 

Terminalia fatraea 1 0.37 13.19 0.13 8.12 1.91 

Trema orientalis 1 0.02 3.24 0.01 2.17 1.00 

Uapaca ferruginea 3 1.42 13.59 0.19 10.66 1.03 

Uapaca littoralis 3 4.86 23.63 0.52 15.03 1.19 

Uapaca thouarsii 3 1.67 12.53 0.22 9.64 0.88 

Vepris eliotii 3 0.57 9.86 0.04 6.84 1.05 

Vepris fitoravina 2 8.15 8.47 0.16 6.76 1.15 

46



Plant species Fat Crude 

Biochemical 

Sugar Tannin ADF Disperser 

protein type 

Apodytes dimidiata 2.85 6.86 64.06 0.00 ND mixed 

Brexia sp. 2.82 2.69 18.23 0.00 18.64 mammal 

Canarium boivinii 12.98 9.19 2.17 0.00 38.59 Efc 

Canthium variistipula 4.91 5.63 18.18 0.18 20.1 mammal 

Cinnamosma madagascariensis 4.96 5.31 26.01 1.74 12.44 Efc 

Dyospiros sp. 0.55 3.13 6.53 0.55 27.69 Efc 

Dypsis prestoniana 3.04 7.19 15.37 0.16 16.62 mixed 

Eugenia cloiselii 2.11 7.31 31.24 0.20 19.59 mammal 

Eugenia sp. 1.17 3.94 18.48 0.00 24.8 Efc 

Leptolaena multiflora 2.24 5.63 2.95 0.16 35.61 mammal 

Ludia antanosarum 1.24 2.88 23.38 0.39 21.71 mammal 

Macaranga perrieri 4.51 5.38 2.87 0.00 42.32 bird 

Olea sp. 1.69 3.75 38.53 0.14 22.39 mammal 

Polyalthia madagascariensis 1.23 5.44 49.44 0.48 22.46 mixed 

Polyscias sp. 2.12 4.19 17.8 0.00 46.4 bird 

Poupartia chapelieri 0.65 5.69 12.63 0.00 14.97 mammal 

Rothmannia mandenensis 0.32 4.63 8.12 0.13 35.68 Efc 

Sarcolaena multiflora 3.93 4.25 14.98 0.15 34.47 mammal 

Schizolaena elongata 2.22 6.24 26.57 0.00 13.72 mammal 

Scolopia orientalis 0.75 3.06 33.38 0.35 12.78 mammal 

Syzygium sp.1 7.13 4.38 31.94 0.19 17.88 mammal 

Syzygium sp.2 3.36 4.94 43.36 1.07 19.07 mixed 

Terminalia fatraea 3.11 7.81 16.40 0.38 35.68 mixed 

Trema orientalis 44.67 13.97 3.38 0.15 16.69 bird 

Uapaca ferruginea 5.73 5.88 2.26 0.00 51.01 mammal 

Uapaca littoralis 2.05 4.44 7.49 0.4 29.93 mammal 

Uapaca thouarsii ND ND ND ND ND mammal 

Vepris eliotii 14.74 7.89 6.42 0.00 17.21 mammal 

Vepris fitoravina 7.37 5.38 19.82 1.23 15.39 mammal 

47

Chapter 2 

depressed seeds (Fig. 3a). The distribution of the species along the first axis shows 

slightly more variation than on the second axis where the majority of fruits have few 

seeds and a spherical or elongated fruit shape. The first axis clearly separates the birdspecies 

from all other species. They have smaller and lighter fruits and seeds. 

Furthermore the mixed fruits have average or intermediate sizes while the specialist fruits 

are clearly larger and more variable in shape and seed number. The mammal fruits are 

intermediate in size, all rather spherical in shape and few-seeded. So the different 

disperser types can be more or less separated into groups by both axes. 

The first two factors of the PCA conducted with the biochemical traits accounted for 

73% of the total variance (Table 4). The first axis of the PCA included parameters 

associated with sugar and crude protein content and explained 43% of the variance 

(eigenvalue=2.17). The second axis was determined by concentrations of fat and acid 

detergent fibre and explained 30% of the variance (eigenvalue=1.52; Table 4). When 

considering the different disperser types no clear patterns arise according to these axes 

(Fig. 3b). Thus, chemical traits failed to group species according to disperser type. 

To support the conclusions of this descriptive analyses, a Kruskal-Wallis test was 

carried out for the disperser syndromes with the first two principal components of each 

PCA. For the morphological parameters, the four disperser syndromes showed a 

significant difference for PCA1 (P=0.001), meaning traits related to fruit size but not for 

PCA2 (seed number, fruit shape; P=0.31). For the biochemical parameters no significant 

difference could be found among the disperser syndromes based on PCA1 (P=0.14), 

involving sugar and crude proteins, nor for PCA2 (P=0.31) involving fat and ADF (Table 

4). 

Table 4. Principal component analyses for morphological and biochemical traits of 29 tree 

species. Each factor represents an ordination axis. Kruskal Wallis test results of the 

disperser type with the principal components given as well (**P

n species with ripe fruits 

35 

30 

25 

20 

15 

10 

5 

0 

O 

Tt 

Ec 

Bm 

To 

X 

Ps 

Lm 

Vf 

D 

Pc 

Sm 

Ad 

Tf 

Dp 


Fig. 2. Number of tree species bearing ripe fruits in 2000. The height of the fruiting peak of the tree 

species involved in tree watches is indicated as well. X includes the following seven species: So, 

La, S1, Cb, Mp, Ve, Se according the abbreviations in Appendix I. 

DISCUSSION 

In the present study we investigated three hypotheses for evidence of co-evolution 

between life history traits of plants, their diaspores and animal consumers. These were 

tight, species-specific interactions, different investment patterns of plants in their fruits in 

relation to the specialization of dispersers and dispersal syndromes as adaptations to 

taxonomically diverse groups of dispersers with different sensory capabilities (colour 

vision in birds, olfaction in mammals). 

There was no evidence for tight co-evolution between specific tree and consumer 

species. This is consistent with findings of most studies (Howe and Smallwood 1982; 

Howe 1984; Gautier-Hion et al. 1985; Herrera 1986; Fisher and Chapman 1993; 

Chapman 1995; Erikkson and Ehrlen 1998; Lambert and Garber 1998). Most plant 

species do not depend on one single disperser species. The only possible indication of 

co-evolution in our study are the five tree species for which Eulemur fulvus collaris is the 

only seed disperser. However, this lemur species is a very opportunistic feeder. A 

comparative study between the dry deciduous forest of Kirindy and the littoral forest of 

Sainte Luce (Bollen et al. in press, Chapter 4) confirms the absence of co-evolutionary 

plant-animal interactions here and shows that this lemur species has a rather high dietary 

flexibility. However, even though the dietary breath of E. f. collaris is quite large, this 

species is a sequential specialist and as such selects two or three dominant fruit species 

each month (Donati 2002). Canarium boivinii and Eugenia sp. make up important 

S2 

Dec 99 Feb 00 Apr 00 Jun 00 Aug 00 Oct 00 Dec 00 

Jan 00 Mar 00 May 00 Jul 00 Sep 00 Nov 00 

Cm 

P 

E 

Uf 

Ul 

49

Chapter 2 

Fig. 3. Ordination plot of the 29 tree species on the first two axes of a prinicipal component 

analyses of (a) morphological traits and (b) biochemical traits. The asterisks stand for dispersal by 

E. f. collaris only, the diamonds for mammal species, the open circles for bird species and the open 

squares for mixed species. 

50 

- fruit shape 

number of seeds 

fat 

ADF 

5 

4 

3 

2 

1 

0 

-1 

-2 

-3 

3 

2 

1 

0 

-1 

-2 

-3 

-5 -4 -3 -2 -1 0 1 2 3 4 

fruit length 

fruit weight 

seed length 

-3 -2 -1 0 1 2 3 4 

- sugar 

crude protein


portions of the monthly diet in Sainte Luce, being 23% of the total diet in January 2001 

and 12% in July and August 2000 respectively (Donati 2002). The remaining three 

species are rather marginal dietary items. Nevertheless, co-evolution in the strict sense 

does not seem to occur here as it is mainly the large fruit size and weight that physically 

excludes the other frugivores with smaller gape size. E. f. collaris is simply the last of the 

remaining large-bodied lemur species in this littoral forest that can ingest these largesized 

seeds, thus matching the situation of brown lemurs in some of Madagascar's dry 

deciduous forests (Ganzhorn et al. 1999a). Many large-bodied frugivores have 

disappeared in Madagascar recently and the extinction of at least 16 large lemur species 

in the Holocene could have included some specialised seed dispersers (Godfrey et al. 

1997). All these ‘specialist’ tree species depend critically on Eulemur fulvus collaris for 

seed dispersal and recruitment. Even though these lemurs often drop the large seeds (up 

to 30mm seed length) under the parent plant, occasionally seeds are swallowed and 

defecated or dropped some distance away from the parent plant. Thus in terms of 

conservation these relationships are of crucial importance to conserve the integrity of the 

littoral forest. 

Our attempt to test McKey's model was problematic for a variety of reasons. First, the 

predictions are qualitative rather than quantitative in nature. ‘Investments’ are difficult to 

specify and may not be the same at nutrient-poor as at nutrient-rich sites or at sites of 

differing seasonality. Also, it is problematic to decide whether short but massive fruiting 

might actually be less expensive for a tree than extended fruiting over longer periods of 

time. Furthermore, the model was developed for bird-dispersed trees in the Neotropics 

and its validity largely depends on the composition of the frugivore guild. With as few as 

eight vertebrate seed dispersers in the littoral forest any ranking or subdivision into 

specialists and generalists is likely to show too much variation to be detected statistically 

in descriptive field studies. For the littoral forest of Madagascar, it might be risky for any 

tree species to depend on only one of these few frugivores and therefore most tree 

species seem to be characterised by a mixture of general traits from both the low and 

high investment model. Even though in general there is little evidence for the McKey 

model (but see Wheelwright 1986), the depauperate frugivore guild of Madagascar might 

not be suitable to test this concept. 

Of the three hypotheses to be tested, the distinction into dispersal syndromes was 

the only one that could be supported by the present data. Fruits consumed and dispersed 

by birds and mammals differ distinctly in fruit and seed size and weight, fruit shape and 

seed number. Moreover, in a similar PCA analysis, Pizo (2002) found the same 

importance of fruit size, fruit width, seed length (PCA1), fruit shape and seed number 

(PCA2) which distinguished primate fruits from the bird and mixed fruits. Bird fruits tend to 

be smaller and more elongated than primate fruits. Our results thus agree with studies on 

fruit syndromes in other assemblages of plants and animals in different regions (Janson 

1983; Knight and Siegfried 1983; Gautier-Hion et al 1985; Pizo 2002; Voigt et al. 2001). 

The rather uniform results suggest that these syndromes are biologically meaningful. 

However for biochemical traits no significant differences could found, which corresponds 

to the findings of Pizo (2002) and Corlett (1996). Even though it has been shown that 

mammals favour fruits rich in sugars while birds prefer fruits with high lipid and protein 

content (Snow 1981; Fleming et al. 1987; Debussche and Isenmann 1989; Galetti 2000), 

the present study only showed a slightly lower sugar and higher protein content in fruits 

eaten by birds. 

51

Chapter 2 

On a community level the low number of frugivores seems to have a profound impact 

on the composition and relative contribution of functional groups to regional ecosystems. 

Neotropical sites have very speciose frugivorous guilds with more than 50 bird species 

(Wheelwright 1986; Galetti and Pizo 1996). In contrast Madagascar has very few 

frugivores and specifically very few frugivorous bird species (Fleming et al. 1987), which 

seems to be reflected in the striking low number of ‘actual’ bird fruits. These 

circumstances obviously narrow down the options for tree species to specialise on certain 

bird species in Sainte Luce. In a comparison of fruits in deciduous forests of Madagascar 

and South Africa, Bleher and Böhning-Gaese (2001) and Voigt et al. (2001) showed that 

the latter has much more bird-dispersed fruits, whereas in Madagascar more mammal 

fruits exist. This is consistent with our findings from the humid evergreen littoral forest. 

The lack of bird fruits is opposed to the findings in India (Ganesh and Davidar 2000), 

Hong Kong (Corlett 1996), La Selva in Costa Rica (Levey et al. 1993) and Malawi 

(Dowsett-Lemaire 1988). At all these sites the majority of fruits are dispersed by birds or 

by both birds and mammals. In our dataset, there are only slightly more mixed fruits and 

the majority are mammal fruits, again indicating low dependence for most tree species on 

birds but high dependence on lemurs and flying foxes for seed dispersal. 

Even though dispersal strategies may include some specifically selected 

morphological traits known as syndromes, general traits make up the bulk of the floral 

diversity in the littoral forest. In particular the nutritional reward for the animals does not 

seem to be taxa related in Sainte Luce, as is also the case in Corlett (1996) and Pizo 

(2002). Many tree species attract their seed dispersers by more or less generalist fruit 

traits. However we have to be cautious as Zamora (2000) stressed that in communitywide 

studies ‘the noise often overwhelms the pattern’ and thus the diffuse co-adaptations 

we found may be the result of the complexity of interactions, few strong ones but many 

weak ones. Thus we cannot exclude the possibility that certain strong interactions were 

overlooked in this study, even though the impact of the species poor frugivore guild 

seems to be determining in this ecosystem. 

In summary, it seems that in the littoral forest of Madagascar the combination of life 

history traits of tree species have not been shaped under the constraints imposed by 

vertebrate seed dispersers upon the trees as co-evolution and the low-high investment 

model of McKey (1975) state. A classification based on taxonomic affiliation of seed 

dispersers does provide a more clear pattern supporting the idea that animals eat what is 

available and what they can swallow and digest, mainly based on size characteristics. 

The observed links between traits of fruits and seeds and their consumers may be more a 

consequence of the morphological and physiological heritage and constraints of the 

consumers but not the result of co-evolution. The lack of tight co-evolutionary interactions 

makes sense in Madagascar, as the community of vertebrate frugivores is so species 

poor that there might have been few options for co-evolution. It could have been too 

dangerous for a tree species to rely on a single animal species for seed dispersal. Or 

simply, the species poor community of frugivores in Madagascar might not have had a 

large enough impact to produce specific tree traits. Nevertheless, some large seeds can 

only be dispersed by E. f. collaris. These tree species are likely to suffer from the 

extinction of the larger frugivorous lemur species. 

52



We are QMM very grateful for providing logistics and infrastructure at the campsite. In particular we 

thank Manon Vincelette, Jean-Baptiste Ramanamanjato, Laurent Randrihasipara of the QMM 

Environmental and Conservation Team. Plant determinations were carried out with the help of the 

botanists Faly Randriatafika and Johny Rabenantoandro of the Missouri Botanical Garden (MBG, 

Antananarivo). Voucher specimens are currently at MBG in Antananarivo. This manuscript 

benefited greatly from numerous suggestions of two anonymous reviewers. This study was carried 

out under the “Accord de Collaboration” between the Departments of Animal Biology and 

Anthropology of the University of Antananarivo and the Institute of Zoology of Hamburg University 

and QMM. I also thank Dries Bonte and Willem Talloen for providing help with statistics. The first 

author is supported by a grant from the Belgian Fund for Scientific Research, Flanders (FWO). We 

thank the Flemish Government for structural support to the CRC of the RZSA. 

Appendix I. List of abbreviations used in Figures 1 and 2 with family, scientific and vernacular 

name. 

Abbreviation Family name Scientific name Vernacular name 

Pc Anacardiaceae Poupartia chapelieri sisikandrongo 

Pm Annonaceae Polyalthia madagascariensis fotsivavo 

P Araliaceae Polyscias sp. voatsilana 

Dp Arecaceae Dypsis prestoniana boakabe 

Cb Burseraceae Canarium boivinii ramy 

Cm Canellaceae Cinnamosma madagascariensis vahabatra 3eM 

G Clusiaceae Garcinia cf/aff. Madagascariensis disaky kely 

Tf Combretaceae Terminalia fatraea katrafa 

D Ebenaceae Diospyros sp.2 hazomainty 

Ea Elaeocarpaceae Elaeocarpus alnifolius sanga 

Mp Euphorbiaceae Macaranga perrieri mocarana 

Uf Uapaca ferruginea voapaky lahy 

Ul Uapaca littoralis voapaky vavy 

Ut Uapaca thouarsii voapaky lahy ZJ 

La Flacourtiaceae Ludia antanosarum zorafotsy 

So Scolopia orientalis zoramena 

B Grossulariaceae Brexia sp. kambatrikambatri 

Ad Icacinaceae Apodytes dimidiata hazomamy 

Bm Menispermaceae Burasaia madagascariensis faritsaty 

Ba Myristicaceae Brochoneura acumineata mafotra 

Ec Myrtaceae Eugenia cloiselii ropasy sp.1 

E Eugenia sp. ropasy sp.2 

S1 Syzygium sp.1 rotry ala 

S2 Syzygium sp.2 rotry mena 

O Oleaceae Olea sp. vahabatra 

Cv Rubiaceae Canthium variistipula fantsikaitramainty 

Rm Rothmannia mandenensis taholagna 

Ve Rutaceae Vepris eliotii lahinampoly 

Vf Vepris fitoravina fitoravina 

Tt Sapindaceae Tina thouarsiana sanirambavy 

Lm Sarcolaenaceae Leptolaena multiflora fotonbavy 

Sm Sarcolaena multiflora meramaintso 

Se Schizolaena elongata fotondahy 

To Ulmaceae Trema orientalis andrarezona 

53

Chapter 2 

Appendix II. Plant-animal interactions based on fruit traps (T), droppings (F), tree watches 

(W). The number of fruit traps, observation hours, crop size and fruiting period are given as 

well as the number of consumer, disperser and predator species. The abbreviations are Am 

for Alectroenas madagascariensis , Hm for Hypsipetes madagascariensis , Ta for Treron 

australis, Efc for Eulemur fulvus collaris , L for lemurs, NL for nocturnal lemurs, Cm for 

Cheirogaleus medius , CM for Cheirogaleus major, Mr for Microcebus rufus, Pr for Pteropus 

rufus, C for Coracopsis spp., Sp for Streptopelia picturata, Em for Eliurus myoxinus and Rr 

for Rattus rattus. 

Impact on seed: Fruit Tree Crop Fruiting Number Number 

Category: traps watches size period consumer disperser 

(N) (h) (N fruits) (N days) species species 

Apodytes dimidiata 2 36 309 19 7 6 

Brexia sp. 1 - 461 143 4 3 

Brochoneura acumineata 3 - 1324 41 1 0 

Burasaia madagascariensis 2 36 114 49 4 2 

Canarium boivinii 4 36 772 412 2 1 

Canthium variistipula 1 - 1280 59 4 3 

Cinnamosma madagascariensis 2 26.5 1 

1445 59 1 1 

Diospyros sp. 3 25 1 

676 73 1 1 

Dypsis prestoniana 1 36 1006 51 8 6 

Elaeocarpus alnifolius 2 - 749 53 3 1 

Eugenia cloiselii 2 27.5 1 

395 17 2 2 

Eugenia sp. 2 36 1712 107 1 1 

Garcinia sp. 1 - 46 26 2 1 

Leptolaena multiflora 4 36 959 71 5 4 

Ludia antanosarum - 31 1 

- - 5 4 

Macaranga perrieri - 36 - - 2 1 

Olea sp. 5 36 4816 67 5 3 

Polyalthia madagascariensis - 36 - - 5 5 

Polyscias sp. - 36 - - 5 3 

Poupartia chapelieri 3 36 800 45 6 4 

Rothmannia mandenensis 2 - 21 37 2 1 

Sarcolaena multiflora 3 36 3010 53 6 4 

Schizolaena elongata 5 36 1654 36 3 2 

Scolopia orientalis 2 30.5 1 

5476 54 5 3 

Syzygium sp.1 3 36 1769 62 5 3 

Syzygium sp.2 4 36 2416 59 7 6 

Terminalia fatraea 2 36 2018 145 7 5 

Tina thouarsiana 4 36 4395 62 5 2 

Trema orientalis - 36 - - 3 3 

Uapaca ferruginea 3 36 381 143 6 4 

Uapaca littoralis 4 36 6800 328 5 3 

Uapaca thouarsii 3 - 1586 295 4 3 

Vepris eliotii 2 31 1 

1612 55 3 3 

Vepris fitoravina 3 36 7086 44 4 3 

1 

No complete set of 36 hours could be obtained due to a very short fruiting period or difficult climatic conditions. 

54


Appendix II Continued 

Seed dispersers Number Seed predators 

Birds Lemurs 

Bats predator Birds Rodents 

Am Hm Ta Efc L NL Cm CM Mr Pr species C Sp Em Rr 

W W W T TW TW TW 1 TW 

T T T T 1 

T 

1 

T 

W T W 2 T T 

TW 2 T T 

F T TW W 1 T 

TW 0 

FW 0 

FW W W T W W W 2 TW W 

FT 2 T T 

FW T W 0 

TFW 0 

T 1 

T 

FW T W W W 1 W 

W W W T 1 W 

W 1 W 

TFW T W W 3 W T T 

W W W W W 0 

W W W 2 W W 

FW T W W W 2 W T 

T 1 

T 

TFW T W W W 2 W W 

TFW W 

1 W 

TW T TW W 2 W T 

TW T W W 2 T T 

W W TFW T W W W 1 W 

W W W W TW 2 W T 

W T TW 3 TW TW TW 

W W W 0 

FW T W W W 2 T T 

TFW T TW W 2 T T 

T T T 1 T 

TW T TW W 0 

W T T W W 1 TW 

55


Lemur drawing © Stephen Nash 1989 

‘Ataovy dian-tana: jerena ny aloha, 

todihana ny afara’ 

Behave like the chameleon: 

look forward and observe behind

Frugivore guild 

Relations between fruits and disperser 

assemblages in the littoral forest 

of south-east Madagascar: 

a community level approach 

AN BOLLEN, LINDA VAN ELSACKER, JÖRG GANZHORN 

JOURNAL OF TROPICAL ECOLOGY (SUBMITTED) 

ABSTRACT 

Interactions among fleshy fruits and frugivore assemblages are presented from a oneyear 

study in the littoral forest of Sainte Luce, south-eastern Madagascar. This 

community level approach allows us to evaluate the relative contribution of different 

frugivores to seed dispersal and predation. For this, interactions between 136 consumed 

fruit species and 13 frugivorous species were studied. Fruit and seed size are the most 

important physical factors determining food selection of all consumer species. While birds 

favour lipid-rich fruits, mammals seem to avoid them. The lemur species that go into 

hibernation clearly prefer sugar rich fruit pulp. In general, there is substantial dietary 

overlap among consumer species and animals seem to be quite flexible to eat whatever 

is available. This might be related to unpredictable fruit availability, which in turn, might be 

one of the reasons for the evolution of the depauperate frugivore guild here. Nevertheless 

all frugivores have different impacts on seed dispersal. Eulemur fulvus collaris is 

particularly important for the dispersal of large-seeded species. Birds and flying foxes 

ensure genetic exchange and plant regeneration between and outside forest fragments. 

In terms of conservation, heterogeneous seed transport is particularly important for this 

severely degraded littoral forest. 

INTRODUCTION 

Interactions between fleshy-fruited plant species and the community of vertebrate 

frugivores have been studied in the tropics worldwide (Leighton 1982; Gautier-Hion et al. 

1985; Dowsett-Lemaire 1988; Corlett 1996; Kitamura et al. 2002; Ingle 2003), where 

zoochorous plant species make up the majority of the flora (Howe and Smallwood 1982; 

Fleming et al. 1987). The fleshy pulp of endozoochorous fruits attracts its consumers by a 

wide array of morphological traits and offers a nutritional reward for potential seed 

dispersers. In general, fruits are eaten and dispersed by a variety of animals even though 

some fruit traits are more likely to attract one taxonomic group of potential dispersers 

than another (Howe 1984; Gautier-Hion et al. 1985; Herrera 1987; Dowsett-Lemaire 

1988; Terborgh 1990; Jordano 1992, 1995; Fisher and Chapman 1993; Erikkson and 

Ehrlen 1998). In this respect diffuse and broad co-adaptations are revealed when 

analysing feeding selection resulting in ‘fruit character syndromes’ (Van der Pijl 1969; 

Gautier-Hion et al. 1985). However, the definition of syndromes is problematic as traits 

are defined differently in each study. Therefore it is necessary to evaluate diets of 

59

Chapter 3 

different consumer species, unbiased, by looking for morphological and biochemical traits 

that may or may not indicate certain feeding preferences instead of trying to fit 

preconceived syndromes. 

Frugivorous animals have been shown to be important for seed dispersal and forest 

regeneration in Madagascar. Most field studies in Malagasy forests have focused on the 

feeding ecology and dispersal role of frugivore species such as lemurs (Ralisoamalala 

1996; Scharfe and Schlund 1996; Dew and Wright 1998; Overdorff and Strait 1998; 

Birkinshaw 1999, 2001; Ganzhorn et al. 1999a) or flying foxes (Bollen and Van Elsacker 

2002a, Chapter 3a), while others have studied the association between focal tree species 

and their frugivore consumers (Scharfe and Schlund 1996; Goodman and Ganzhorn 

1997; Goodman et al. 1997a; Böhning-Gaese et al. 1999). However, apart from some 

combinations of literature and field studies (Phillipson 1996; Bleher and Böhning-Gaese 

2001; Voigt 2001) no community wide dispersal study has been carried out in 

Madagascar up to now. This current study focuses on interactions between the 

community of vertebrate frugivores present in the littoral forest and the plant species they 

consume, with emphasis on fruit morphology and nutrient content of each plant species 

and with respect to the individual roles of consumers as seed dispersers or predators. 

The study on fruit-frugivore interactions is particularly relevant for Madagascar as this 

island has a high percentage of botanical and faunal endemism (Lowry et al. 1997; 

Schatz 2001) and at the same time a rather depauperate frugivore community (Langrand 

1990; Mittermeier 1994; Goodman et al. 1997a). An attempt was made to unravel 

aspects of animal-plant interactions that determine the dynamics of the littoral forest, 

which presently suffers from severe fragmentation and degradation. Given these aspects, 

it is important to understand these interactions to urgently integrate them in conservation 

management plans for this area. 

The following research questions are addressed: 

1. Which plant species are included in the diet of the frugivores present in the littoral 

forest? 

2. On the basis of which morphological and biochemical fruit and seed 

characteristics do frugivores select their food resources and are certain feeding 

preferences prevalent? 

3. To what extent does dietary overlap occur between these frugivores? 

4. What is the impact of these frugivores on the fruits they eat? Can they be 

considered as efficient seed dispersers, rather neutral seed droppers or more 

destructive seed predators? 

STUDY SITE 

This research was conducted by the first author from November 1999 through January 

2001 in a 377-ha forest fragment (S9) of the littoral forest of Sainte Luce (24º45'S 

47º11'E, south-east Madagascar). At the moment, the south-eastern littoral forest is 

represented by 2500 ha only, which is located in the surrounding area of Fort-Dauphin 

(Petriky, Mandena, Sainte Luce) (Ganzhorn et al. 2001; Vincelette pers. comm.). The 

most intact littoral forest can be found at Sainte Luce, which includes forest fragments 

ranging in size from 3 up to 377ha (Fig. 1). Littoral forest grows on sandy soils and occurs 

within 3km of the coast (Dumetz 1999). A relatively open or non-continuous canopy 

characterises this forest, which is 6 to 8m in height with emergents up to 18m (Dumetz 

1999). The diameter at breast height (DBH) of trees rarely exceeds 30 to 40cm 

60


(Rabevohitra et al. 1996; Dumetz 1999). Average annual rainfall at this site is about 

2690mm, with a marked rainy season from November through February while no clear 

dry season could be detected (Bollen and Donati, Chapter 1). Mean monthly temperature 

is 23°C (QMM unpubl. data). Fruit production is seasonal, with a peak in abundance of 

ripe fruits in December and January and with periods of fruit scarcity that differ strongly 

inter-annually (Bollen and Donati, Chapter 1). 

24°S 

MADAGASCAR 

PETRIKY 

Fort-Dauphin 

48°E 

MANDENA 

10 km 

SAINTE 

LUCE 

N 

FORT- 

DAUPHIN 

Fig. 1. On the left Madagascar is shown with indication of the south-eastern zone. In the middle a 

detail of the littoral forests (Sainte Luce, Mandena and Petriky) is given together with a detail of the 

main forest fragments of Sainte Luce on the right. 

METHODS 

Plant species studied 

Ripe fruits were collected in the study area throughout the research period. Consequently 

morphological characteristics were measured in the field station and biochemical 

components were analysed in the lab. In this chapter only plant species are included, 

which have their fruits or seeds consumed by at least one vertebrate species. The nonzoochorous 

fruits were left out. Therefore the dataset used here is a subset of the 

complete one (N =175) used elsewhere (Bollen and Van Elsacker 2002, Chapter 3a; 

S8 

S7 

S6 

S17 

S9 

SAINTE 

LUCE 

61

Chapter 3 

Bollen et al. in press, Chapter 4). The full dataset is available from the first author by 

request. 

Herbarium specimens of all taxa were collected and deposited at the Missouri 

Botanical Garden of Antananarivo (Madagascar). Local research assistants provided 

Antanosy names. Scientific names were obtained after determination of voucher 

specimens at herbaria of Antananarivo (FOFIFA, Tsimbazaza) with the help of botanists 

from Missouri Botanical Garden (App. I). 

Morphological characteristics 

Discrete variables used to characterise fruits were: 

Growth form: large tree (>6m), small tree (


analyses and analyses of fruit trap contents (78x1m² traps under 29 tree species) (Bollen 

and Van Elsacker 2002a, Chapter3a; Bollen et al., Chapter 2). For Pteropus rufus faecal 

droppings were collected weekly under the roost site year-round. Dietary data on the 

rodents resulted mainly from identifying gnawing marks on seed remains collected at 

feeding sites. Since Eulemur fulvus collaris was studied intensively in parallel to the 

present study (Baldi 2002; Donati 2002; Morelli 2002), fruits eaten by this species may be 

more thoroughly sampled than other species. The con-generic species Cheirogaleus 

medius and C. major could not always be distinguished during observations and are 

treated as Cheirogaleus spp. in the analyses. The same applies to rodent species, 

Eliurus webbi and Rattus rattus (introduced), as not all gnawing marks could be attributed 

to a single species. Details of the diets of the various species are given in Appendix I. For 

the frugivorous bird species, diets are treated separately, but for feeding selection 

Alectroenas madagascariensis, Treron australis, and Hypsipetes madagascariensis were 

combined, as they are the only seed dispersing birds at our site. 

Dietary overlap was calculated among pairs of frugivores using Sørensen’s similarity 

index (Krebs 1989). This index generates a value ranging from 0 to 1, with 0 representing 

no overlap and 1 representing complete overlap. The dataset comprises accounts of fruit 

species consumed by different frugivore consumers along with the impact they have on 

the seeds. Based on this, frugivores can be classified into the following categories: seed 

dispersers or fruit consumers (D), neutral or pulp consumers (N) and seed predators (P) 

according to Gauthier-Hion et al. (1985) and Debussche and Isenmann (1992). The first 

group disperses intact seeds by endozoochory through droppings or synzoochory 

through regurgitation, while the second group eats fruit pulp but drops the seeds under 

the parent plant. The last group eats and destroys the seeds. It is difficult to assign a 

certain frugivore to one category only, as one species may have different impacts on the 

same and on different plant species. The stage of ripeness of the consumed fruits was 

scored as well. To differentiate between unripe and ripe fruits changes in size, colour and 

consistency were looked at. 

Data analyses 

Most of the variables measured have highly skewed distributions so the median value is 

given instead of the mean. For the same reason non-parametric statistics were used. Chisquare 

analyses were conducted to compare discrete fruit traits in the diet with those in 

the overall dataset, whereas Mann Whitney U tests were carried out to control for feeding 

preferences when comparing continuous traits between food and non-food items. 

Afterwards sequential Bonferroni corrections were performed on the significance levels 

(Rice 1989). To understand which factors influence the diet of the different frugivores 

separately a logistic generalised model was applied in which morphological and 

biochemical variables were included as fixed factors. Dietary data were used as binomial 

response variables (0=non-food, 1=food item) in a generalised mixed linear model with 

logit link (glimmix procedure in SAS 8.1.) with forward procedure retaining significant 

variables. As Cheirogaleus spp. and Microcebus rufus go into torpor in austral winter, 

comparison of their diet and total dataset available were restricted to the fruits that were 

present during their active period. Statistical significance was accepted for α≤0.05 for all 

tests. All statistical tests were carried out according to Siegel (1956) with the statistical 

software SAS for Windows. 

63

Chapter 3 

64 

Body 1,2 

Body 

lenght (cm) 

1 

Table 1. List of frugivorous vertebrate species at Sainte Luce, with indication of their diet (F: frugivorous, G: granivorous, O: omnivorous), their 

activity budget (D: diurnal, N: nocturnal, C: cathemeral), feeding height (A: arboreal, T: terrestrial), body mass and length. ND: no data available. 

Family Scientific name 

English name 

Diet Activity Feeding Group size 

height (N individuals) mass (g) 

32 

28 

28 

24 

35 

50 

215 

ND 

190 

45 

218 

ND 

3-8 

3-12 

1-2 

5-15 

3-15 

3-15 

A 

A 

T 

A 

A 

A 

D 

D 

D 

D 

D 

D 

F 

F 

G 

F 

G 

G 



Madagascar Turtle Dove 

Madagascar Bulbull 



Treron australis L. 

Alectroenas madagascariensis L. 

Streptopelia picturata Temminck 

Hypsipetes madagascariensis Müller 

Coracopsis nigra L. 

Coracopsis vasa Shaw 

AVES 

Columbidae 

Pycnonotidae 

Psittacidae 

15-23 

10-16 

100 

88 

1-2 

1-2 

AT 

AT 

N 

N 

O 

G 

Black Rat 

Webb's Tuft-Tailed Rat 

Rattus rattus L. 

Eliurus webbi Ellerman 

MAMMALIA 

RODENTIA 

Muridae 

Nesomyinae 

23-27 

500-750 

250-300 

A 

N 

F 


Pteropus rufus Tiedemann 

CHIROPTERA 

Pteropodidae 

PRIMATES 

Lemuridae Eulemur fulvus collaris E. Geoffroy Collared Brown Lemur F C A 3-10 2000-2300 40-47 

Cheirogaleidae Microcebus rufus E. Geoffroy Brown Mouse Lemur O N A 1 

42 12,5 

Cheirogaleus major E. Geoffroy Greater Dwarf Lemur O N A 1-3 443 25 

Cheirogaleus medius E. Geoffroy Fat-tailed Dwarf Lemur O N A 1-3 119-282 20 

1 

Data from Langrand (1990), Fietz and Ganzhorn (1999), Goodman et al. (1997b), Goodman et al. (in press), Ganzhorn et al. (1999a), Donati (pers. comm. 2002). 

2 Body length is total length for birds and bats but head/body length for lemurs and rodents.


RESULTS 

Lemurs fed on most fruiting species, their diet comprised 119 plant species. Birds, 

rodents, and Pteropus rufus consumed 55, 50, and 39 plant species, respectively. 

Different methods of collecting dietary data influenced the outcome of diet lists (Table 2). 

In general direct observations (systematic or opportunistic) resulted in the largest 

numbers of feeding records. However, five species were difficult to observe. Due to 

hunting pressure, observations of P. rufus at night with a headlight were not routinely 

possible. Both rodent species could be observed only rarely as they detect the observer 

by smell. As explained in the methods, this bias could be limited by systematically 

collecting faecal droppings and identifying gnawing marks year-round. Treron australis 

and S. picturata were very shy and flew away upon detecting the observer. Moreover 

droppings of these bird species as well as of Hypsipetes madagascariensis were found 

only rarely. For Alectroenas madagascariensis and E. f. collaris faecal droppings could be 

collected more easily, but much less often for the smaller nocturnal lemurs. Characteristic 

feeding marks were helpful in particular to identify food species of Coracopsis nigra, both 

rodent and all lemur species. 

Table 2. Number of consumed plant taxa scored per consumer species is given with 

indication of the different methods (O:observations, F: faecal analyses; T: traces). 'Ripeness' 

indicates the stage of ripeness (R: ripe, UR: unripe) at which fruit species were consumed. 

The effect on seeds by the consumer species is indicated (D: dispersal, N: neutral seed dropping, 

P: predation, ?: unknown). 

Dietary diversity 

N N N 

Sampling 

effort 

Ripeness Effect on seeds 

species genera families O F T R UR D N P ? 

Treron australis 9 9 9 7 4 0 9 0 9 0 0 0 

Alectroenas madagascariensis 18 17 14 16 13 0 18 0 18 1 0 0 

Streptopelia picturata 13 13 11 13 0 0 13 0 0 0 0 13 

Hypsipetes madagascariensis 21 20 17 21 1 0 21 0 21 2 0 0 

Coracopsis nigra 37 32 22 36 2 9 26 24 4 8 36 0 

Eulemur fulvus collaris 111 76 43 93 67 21 111 25 100 36 27 0 

Cheirogaleus spp. 39 31 20 37 7 24 39 0 28 24 0 0 

Microcebus rufus 41 33 24 33 6 25 41 0 27 20 0 4 

Pteropus rufus 39 27 21 7 34 5 39 0 37 12 0 1 

rodents 50 37 31 3 1 47 50 0 4 2 49 0 

Morphological Characteristics 

The complete data set of all food species involved mainly large canopy tree species 

(59%). Most common plant families are Rubiaceae (10%), Euphorbiaceae (5%), and 

Flacourtiaceae, Myrtaceae, Annonaceae, and Areceae (each with 4%). Berries and 

drupes were the most common fruit types (83%) with a soft and juicy pulp (62%). Dull 

coloured fruits (green, brown, yellow-orange 68%) with odour (65%) made up the majority 

of the fruits. Other dominant features were indehiscent fruits with a thin husk (77%) and 

seeds could be either protected (54%) or not (46%). The median number of seeds per 

fruits was 2 (quartiles 1-4) and median fruit weight was 1.23g (0.49-5.23g), fruit length 

was 15.43mm (0.49-5.23mm), and seed length was 8.36 mm (4.85–14.42mm). 

Taxonomically, Rubiaceae was the dominant plant family in most diets and 

Euphorbiaceae, Areceae, and Annonaceae were important as well but to a lesser extent. 

There appeared to be no clear taxonomic preferences within the diet of all frugivores. The 

65

Chapter 3 

dominant plant families seemed to be represented in the diets as represented in the 

overall sample. The same is true for growth form (Table 3). Most consumers favour 

berries and drupes, but for rodents there was a trend for selecting drupes. Even though 

soft and juicy fruit made up most of the sample, nocturnal lemurs and flying foxes 

selected this pulp type significantly more often than other pulp types. The latter also 

favoured arillate fruits (Table 3). For colour a selection was noticed towards red and 

purple fruits by all frugivore bird species, whereas mammals ate whatever colour was 

available. Most fruits in the littoral forest had an odour and were probably selected in this 

way, except for Coracopsis nigra, which fed mainly on odourless fruits (Table 3). For the 

nocturnal lemur species fruits with a thin husk were favoured. The other consumers did 

not seem to avoid the few dehiscent and thick-husked fruits present (Table 3). In all 

animal species no difference in seed protection could be found between diet and the 

overall dataset (Table 3). The few preferences indicated above appeared to be nonsignificant 

after correction by sequential Bonferroni. 

Flying foxes were the only consumer species, which preferred multi-seeded fruits, 

while the diet of the other animals did not differ from what was available (Table 4). Initially 

many significant preferences could be found related to fruit and seed size and weight. 

However after sequential Bonferroni adjustment, only the frugivorous birds seem to select 

significantly smaller and lighter fruits. Coracopsis nigra also prefers lighter fruits and 

Pteropus rufus smaller seeds. Contrarily the rodents clearly favour heavier fruits and 

larger seeds. (Table 4). 

Biochemical Characteristics 

Water was the dominant constituent of fresh pulp (median 76.0%). On a dry mass basis, 

both acid (22.6%) and neutral (32.0%) detergent fibre contents were high. The median 

sugar content was 19.2%. Median lipid content of fruits was 3.1%, total nitrogen 0.9%, 

and extractable protein 2.8%. Tannin values were very low in our dataset with a median 

value of 0.2%. Enzymatic analyses of all fruits yielded median values of 3.6% 

saccharose, 1.8% glucose, and 1.8% fructose. 

As for fats, frugivorous birds seemed to select fruits with a high lipid content, while 

the opposite was true for E. f. collaris and P. rufus (Table 4). Neither total nitrogen nor 

extractable protein seemed to influence fruit choice for any of the consumer species, nor 

did water content or acid detergent fibre. Cheirogaleus spp. and Microcebus rufus 

selected fruits with high sugar content but this trend was not significant (Cheirogaleus 

spp. P=0.06, M. rufus P=0.19). However when looking at saccharose, glucose, and 

fructose concentrations separately, preferences were significant for Cheirogaleus spp. 

(Table 4). The same trend existed for M. rufus (saccharose P=0.17, glucose P=0.11, 

fructose P=0.056). Tannins were consumed as present in the overall database but 

Coracopsis nigra included fruits with significantly higher tannin content. In the diet of M. 

rufus neutral detergent fibre was significantly lower than in fruits, which were not 

consumed (Table 4). None of these preferences remained significant after sequential 

Bonferroni adjustment. 

66

Table 3. Chi-square results comparing fruit and seed traits in the diets of the different consumer species to the overall data set. 

Significant differences are in bold (* P

Chapter 3 

Table 4. Morphological and biochemical characteristics of food and non-food items of the 

different frugivore species. For comparison of fruit selection by Cheirogaleus spp. and 

Microcebus rufus only those fruits were considered that were present during the months when 

the lemurs were not in torpor. Values are medians, quartiles and sample size. Z-values are 

based on Mann Whitney U tests ( * P



Parameters 

Eulemur fulvus collaris Cheirogaleus spp. 

Food Non food Z Food Non food Z 

Seed number 2.00 1.50 -0.70 1.50 1.00 -0.09 

1.00-5.00 1.00-2.25 1.00-3.00 1.00-4.00 

105 20 38 74 

Fruit weight 1.46 0.72 -2.65** 0.66 1.79 -3.16** 

0.56-5.86 0.21-157 0.32-1.63 0.79 

103 24 39 73 

Fruit length 15.5 13.95 -1.02 12.34 19.10 -3.29** 

10.52-25.18 9.13-21.76 9.22-16.42 12.42-30.59 

109 24 39 78 

Seed length 8.36 7.90 -1.15 8.19 10.27 -2.07* 

5.31-15.05 3.37-10.91 4.39-11.96 5.35-18.27 

104 20 36 75 

Lipid 2.77 6.73 2.65** 2.82 3.27 -0.80 

1.73-5.17 3.74-21.27 1.87-5.07 1.85-6.01 

84 12 35 49 

Total nitrogen 0.86 0.83 -0.28 0.86 0.85 -0.68 

0.65-1.17 0.57-1.07 0.60-1.12 0.63-1.26 

88 12 35 52 

Extractable 2.82 3.02 0.21 3.15 2.78 0.57 

protein 1.59-4.26 1.62-4.01 1.64-4.65 1.68-4.18 

88 12 35 52 

Sugar 18.39 31.42 0.99 27.84 12.84 1.88 

7.52-35.99 10.38-40.64 13.81-40.68 6.51-33.76 

88 12 35 52 

Tannin 0.19 0.22 0.70 0.16 0.21 -1.03 

0-0.51 0.12-0.66 0.00-0.39 0.00-0.58 

88 12 35 52 

NDF 32.73 30.00 -1.37 31.22 34.36 -1.28 

24.11-47.91 21.75-33.55 24.08-41.77 25.62-52.33 

78 12 33 48 

ADF 22.77 22.30 -0.77 21.71 24.95 -1.45 

17.11-35.68 16.99-27.80 16.99-29.93 17.39-37.02 

78 12 33 48 

Water content 76.50 71.00 -0.86 75.86 74.04 0.69 

70.00-83.25 65.00-79.00 71.00-82.50 63.04-80.75 

76 15 31 52 

Saccharose 3.60 6.54 -0.06 5.05 0.99 -2.33* 

0.92-14.10 1.02-15.58 2.48-16.27 0.38-7.54 

48 8 28 20 

Glucose 1.77 2.15 -0.37 1.92 0.39 -2.60* 

0.39-6.25 0.35-5.67 1.09-5.75 0.10-2.00 

48 8 28 20 

Fructose 1.75 1.90 -0.44 2.41 0.53 -2.58* 

0.58-7.71 0.09-7.52 1.22-7.05 0.07-1.45 

48 8 28 20 

69

Chapter 3 


Parameters 

Microcebus rufus Pteropus rufus 

Food Non food Z Food Non food Z 

Seed number 2.00 1.00 0.26 3.00 1.00 3.39*** 

1.00-3.00 1.00-3.00 1.00-30.50 1.00-2.00 

37 75 39 86 

Fruit weight 0.63 1.94 -3.16** 0.98 1.37 -0.91 

0.30-1.64 0.82-6.43 0.43-3.82 0.54-5.85 

38 74 38 89 

Fruit length 12.28 19.69 -3.22** 13.59 16.56 -1.10 

9.59-15.15 12.70-28.75 9.94-21.09 10.64-25.41 

38 79 39 94 

Seed length 7.46 11.83 -2.87** 6.31 10.30 -3.67*** 

3.79-10.58 5.98-17.75 3.74-8.38 6.77-15.78 

37 74 38 86 

Lipid 3.12 2.69 -0.48 2.52 3.79 -2.07* 

2.19-5.60 1.79-5.17 1.94-3.90 1.84-8.50 

32 52 36 60 

Total nitrogen 0.83 0.86 -0.74 0.79 0.87 -1.62 

0.58-1.12 0.64-1.18 0.52-1.14 0.67-1.18 

32 55 37 63 

Extractable 2.90 2.86 -0.18 2.86 2.53 0.79 

protein 1.72-4.18 1.65-4.58 2.18-3.96 1.55-4.39 

32 55 37 63 

Sugar 21.60 15.37 1.29 18.18 20.98 -0.36 

10.66-41.88 6.59-33.5 8.12-37.23 7.76-37.53 

32 55 37 63 

Tannin 0.20 0.18 0.27 0.28 0.16 1.05 

0.00-0.42 0.00-0.52 0.10-0.78 0.00-0.52 

32 55 37 63 

NDF 27.77 34.36 -1.99* 35.06 31.35 1.25 

23.27-41.92 27.56-51.56 27.04-47.91 22.97-44.34 

31 50 34 56 

ADF 20.10 25.98 1.70 25.84 22.45 1.46 

17.08-31.04 18.85-36.56 19.08-35.99 16.71-32.81 

31 50 34 56 

Water content 76.00 73.00 1.54 78.24 74.00 1.14 

71.75-83.25 63.00-81.00 71.75-83.00 64.77-83.5 

32 51 32 59 

Saccharose 3.95 1.15 1.38 3.56 3.95 0.38 

2.34-20.22 0.61-11.52 0.96-13.66 0.91-17.74 

23 25 29 27 

Glucose 1.91 0.50 1.60 1.70 1.91 0.07 

1.07-5.68 0.17-3.50 0.45-5.66 0.36-5.86 

23 25 29 27 

Fructose 1.97 0.79 1.91 1.57 1.93 0.04 

1.19-9.17 0.10-5.42 0.64-6.16 0.13-9.77 

23 25 29 27 

70


Parameters 

Rodents 

Food Non food Z 

Seed number 1.00 2.00 -0.92 

1.00-3.00 1.00-5.00 

48 77 

Fruit weight 3.16 0.92 3.53*** 

0.82-8.03 0.29-2.90 

50 77 

Fruit length 19.04 13.20 3.24** 

13.29-30.38 8.08-21.82 

50 83 

Seed length 12.06 6.90 3.99*** 

7.77-18.89 3.89-11.73 

48 76 

Lipid 2.85 3.22 -0.01 

1.75-6.92 2.12-5.12 

35 61 

Total nitrogen 0.86 0.86 0.77 

0.68-1.21 0.59-1.15 

35 65 

Extractable 2.53 2.83 -0.09 

protein 1.84-4.06 1.57-4.34 

35 65 

Sugar 18.23 21.27 -0.95 

6.98-32.66 10.15-40.02 

35 65 

Tannin 0.18 0.20 -0.95 

0.00-0.44 0.00-0.65 

35 65 

NDF 33.70 31.22 0.58 

24.03-47.45 23.90-46.24 

33 57 

ADF 23.03 22.51 0.49 

17.21-35.68 17.08-33.87 

33 57 

Water content 73.69 77.00 -1.20 

65.00-81.25 70.50-84.00 

40 51 

Saccharose 3.51 5.22 -0.85 

0.75-7.37 1.05-15.75 

20 36 

Glucose 1.27 2.33 -1.01 

0.41-2.27 0.39-6.30 

20 36 

Fructose 1.34 2.37 -0.95 

0.56-3.57 0.37-11.93 

20 36 


71

Chapter 3 

Besides testing feeding selection for each variable separately, we additionally aim to 

understand which factors influence the diet of the different frugivores. According to the 

final significant model of Eulemur fulvus collaris, growth form (F2, 120= 5.75, P=0.004) and 

seed protection (F1 120=6.66, P=0.011) seem to determine most whether a fruit is eaten. 

Their diet is characterised by fruits from large trees with a hard kernel (Table 5). For 

Cheirogaleus spp. fruit length (F1, 115= 10.15, P=0.002) is the only significant determinant, 

whereas for M. rufus this is seed length (F1, 109= 8.84, P=0.004). Obviously smaller fruits 

and seeds determine the diet of both lemur species (Table 5). Conspicuous colours (F5, 

89=3.25, P=0.009) and large fat contents (F1, 89= 6.33, P=0.013) determine whether 

frugivore birds eat a certain fruit. Coracopsis nigra prefers odourless (F1, 128=6.54, 

P=0.012) and small fruits (F1, 128= 6.45, P=0.012). Fat content (F1, 87= 6.97, P=0.009) and 

seed length (F1, 87=15.11, P=0.0002) seem to determine the presence of fruits in the diet 

of P. rufus. In particular fruits with low lipid content and small seed length are most 

abundant. Finally, the diet of rodents is characterised by a significant interaction between 

seed length and seed protection (F1, 114=5.73, P=0.018), indicating that large seeds, 

which often have a hard kernel predominate their diet (Table 5). The goodness-of-fit of 

the different model is significant (X²Efc=92.41; X²Ch=134.49; X²Mr=130.51; X²fb=82.85; 

X²Cn=136.04; X²rod=128.19; P


Diet overlap 

Frugivorous animal species shared between 2 and 42 plant species. This corresponds to 

a Sørensen’s similarity index of 0.08 to 0.67 (Table 6). Bird and lemur species had more 

fruit species in common than flying foxes and rodents. The highest similarity index was 

found between M. rufus and Cheirogaleus spp. (Table 6), followed by P. rufus and M. 

rufus. The lowest index was found between Treron australis and the rodents. In general, 

dietary overlap among mammals was larger than among birds or between mammals and 

birds (Table 6). These results have to be interpreted with caution in particular for those 

species which dietary diversity is underestimated, such as T. australis and S. picturata. 

Both inter- and intra-specific interactions as well as polyspecific feeding associations 

were observed. 

Table 6. Diet overlap between each pair of consumer species. The number of species consumed 

by each frugivore is shown in italics. Above diagonal is the number of species shared between 

pairs, below the diagonal is the dietary overlap values calculated according to Sørensen's 

similarity index. 

N Hm Am Ta Sp Cn Efc Ch 1 Mr Pr Rod 1 

N 21 18 9 13 37 111 39 41 39 50 

Hypsipetes madagascariensis (Hm) 21 10 5 5 11 14 8 10 9 3 

Alectroenas madagascariensis (Am) 18 0.51 7 4 10 14 9 8 9 8 

Treron australis (Ta) 9 0.33 0.52 2 5 6 5 4 4 3 

Streptopelia picturata (Sp) 13 0.29 0.26 0.18 8 10 7 7 5 2 

Coracopsis nigra (Cn) 37 0.38 0.18 0.22 0.32 27 18 16 14 10 

Eulemur fulvus collaris (Efc) 111 0.21 0.21 0.10 0.16 0.36 35 35 35 42 

Cheirogaleus spp. (Ch) 39 0.27 0.32 0.21 0.27 0.47 0.47 27 20 18 

Microcebus rufus (Mr) 41 0.32 0.27 0.16 0.26 0.41 0.46 0.67 23 19 

Pteropus rufus (Pr) 39 0.30 0.32 0.17 0.19 0.37 0.47 0.51 0.57 14 

Rodents (Rod) 50 0.08 0.24 0.02 0.06 0.23 0.52 0.40 0.42 0.31 

1 

Congeneric species are treated together, as well as both rodent species which could not 

always be attributed to a single species 

Dispersal and predation role 

The ‘true’ frugivorous bird species Alectroenas madagascariensis, Treron australis, and 

Hypsipetes madagascariensis dispersed seeds of most of the species they fed on. Ripe 

fruits were nearly always swallowed and thus dispersed (Table 2). Streptopelia picturata 

was considered a seed predator feeding on seeds on the ground but due to its shy 

nature, feeding behaviour could not be observed in detail and thus the fate of consumed 

seeds remains unclear. Coracopsis nigra occasionally dropped fruits under the parent 

plant or flew away with intact fruits but for the majority of their food resources, they acted 

as seed predators, either destroying seeds directly with their beak or feeding on unripe 

fruits (Bollen and Van Elsacker, Chapter 3b). In contrast, E. f. collaris is an important 

seed disperser for a large number of plant species. During feeding this species was 

messy, swallowing numerous seeds while dropping others under the parent plant and this 

happened for a third of their consumed plant species (Table 2). Furthermore, for some 

plant species it ate unripe fruits, thus destroying the seeds this way. Cheirogaleus spp. 

and M. rufus as well as P. rufus act as seed dispersers for smaller seeds and as seed 

droppers for larger ones. P. rufus may participate in dispersal at short distances as they 

carry larger fruits to nearby feeding roost and drop the seeds there. Rodents often tear off 

73

Chapter 3 

and do not consume the fibrous flesh, which surrounds nuts. Both rodent species clearly 

preyed on seeds of numerous species, but intact and even germinated seeds of four 

species were found at feeding sites. No secondary seed dispersal through caching could 

be detected (Table 2). 

DISCUSSION 

Food selection based on taxonomy, morphology, biochemistry and phenology 

Even though in the literature numerous examples exist of plant families that are typically 

consumed by birds such as Lauraceae Myristicaceae, Areceae, Burseraceae, and others 

(McKey 1975; Snow 1981; Corlett 1998; Oliveira et al. 2002) or by flying foxes such as 

Moraceae, Anacardiaceae, Guttiferae, Myrtaceae, Areceae, and Sapotaceae (Marshall 

1983; Banack 1998; Corlett 1998) no clear dominant plant families could be found within 

the diets of any of the consumer species. The families best represented correspond with 

the dominant plant families of the littoral forest. This may be linked either to the fact that 

Madagascar has a high percentage of endemic plant species which may alter the overall 

floral composition compared to other tropical forests (Schatz 2001) or that the 

depauperate guild of frugivores does not discriminate its diet taxonomically. 

The main food preferences for all consumers are related to morphological fruit traits. 

Worldwide in studies on fruit-frugivore interactions size, colour, and fruit protection 

seemed to be the most important traits in fruit selection. Among these, fruit and/or seed 

sizes were found to be most significant and related to the body size of the consumers. 

The gape size of frugivorous birds limits the maximum seed diameter of fruits they can 

swallow, while mammals have teeth and have other means to eat larger fruits (Fleming et 

al. 1987; Dowsett-Lemaire 1988; Corlett 2002). For this reason, in our study, as in most 

studies, birds seem to select significantly smaller fruits than mammals (Snow 1981; 

Janson 1983; Knight and Siegfried 1983; Gautier-Hion et al. 1985; Wheelwright 1986; 

Herrera 1987; Debussche 1988; Debussche and Isenmann 1989; Jordano 1995; Corlett 

1996; Lambert 2002; Pizo 2002; Carlo et al. 2003). Eulemur fulvus collaris in this respect 

is very important in the local community as it is the only seed disperser of the larger fruit 

species. Nevertheless fruit size distinction is weakened by the fact that some bird species 

can pick up broken parts of fruits and eat larger soft fruits in a piecemeal fashion 

(Kitamura et al. 2002; Pizo 2002). As a result some studies did not find a significant 

difference in size of bird and mammal consumed fruits (Gautier-Hion et al. 1985; Voigt 

2001; Poulsen et al. 2002). 

As for fruit protection, mammals are able to feed on fruits with a thick husk 

(Birkinshaw 2001; Kitamura et al. 2002; Lambert 2002), whereas birds select more 

frequently those with a thin husk or dehiscent fruits (Janson 1983; Gautier-Hion et al. 

1985; Kitamura et al. 2002). However, it seems that not all mammals eat thick-husked 

fruits. In our study E. f. collaris and P. rufus did, whereas the Cheirogaleus spp. and M. 

rufus nevertheless preferred thin-husked fruits. Another important trait for feeding 

selection is supposed to be represented by the external colour of ripe fruits, though little 

agreement exists on this topic. One consistency throughout all studies, including this one, 

is that birds eat many red and black fruits (McKey 1975; Charles-Dominique et al. 1981; 

Janson 1983; Knight and Siegfried 1983; Gautier-Hion et al. 1985; Debussche 1988; 

Dowsett-Lemaire 1988; Horvitz et al. 2002; Kitamura et al. 2002; Poulsen et al. 2002; 

Voigt et al. 2001). For primates in particular, different results have been found which is 

likely related to their different visual acuity in different parts of the world. Lemurs have no 

74


or limited colour vision, whereas New World primates have dichromate and Old World 

primates trichromate vision (Jacobs 1995; Dominy and Lucas 2002; Surridge et al. 2003). 

Nevertheless in many studies dull coloured fruits such as green, brown, yellow, and 

orange are eaten by mammals and/or primates (Janson 1983; Knight and Siegfried 1983; 

Terborgh 1983; Gautier-Hion et al. 1985; Dowsett-Lemaire 1988; Birkinshaw 2001; 

Kitamura 2002; Poulsen et al. 2002) but at some sites, including ours, these species do 

not avoid bright colours either (Janson 1983; Knight and Siegfried 1983; Gautier-Hion et 

al. 1985; this study). Few studies include odour in their dataset, as this is a subjective 

measure. Nevertheless it is common knowledge that birds forage upon visual detection 

whereas mammals make use of olfactation. For the remaining morphological fruit traits 

scored, there does not exist a general understanding in literature, nor a clear food 

preference for certain consumer species in our dataset. 

While most studies on fruit preferences concentrate on morphological traits, 

biochemical characteristics are rarely considered (Janson et al. 1986; Herrera 1987; 

Ganzhorn 1988; Jordano 1995; Izhaki 2002; Pizo 2002). The most common finding is that 

mammals in general avoid lipid-rich fruits whereas birds may favour them (McKey 1975; 

Debussche and Isenmann 1989; Jordano 1995). This differential preference for lipids 

corresponds indeed with our findings of frugivorous birds, E. f. collaris, and P. rufus. The 

only other trend found was that nocturnal lemurs that go into torpor favour sugar-rich fruit 

pulp. For Cheirogaleus medius this preference during prehibernation fattening was 

already described by Bonnaire and Simmen (1994) and Fietz and Ganzhorn (1999). In 

Sainte Luce, fruits are relatively low in tannin values (as elsewhere in Madagascar: 

Iaconelli and Simmen 2002) and the consumer species did not seem to avoid fruit 

species with high tannin content. Bairlein (1996) mentioned that tannins in general seem 

to be less detrimental for avian frugivores than others, which seems to be true for 

Coracopsis nigra at our site. Besides the differences mentioned above, the remaining 

chemical components varied little among consumer species and did not seem to 

determine their feeding selection as was shown by other studies (Corlett 1996; Pizo 

2002). 

As mentioned in the results, several of these food preferences do no longer remain 

significant after sequential Bonferroni adjustment. There has been a great deal of 

discussion on whether to apply these adjustments or not. Recently Moran (2003) has 

come up with some arguments for the rejection of this adjustment. All the same, we 

believe that several of our initial significant findings, which are no longer significant after 

rigorous adjustment for increasing type I errors, do represent biological meaningful 

results which are confirmed by similar results from several other studies. 

Selection criteria of granivores ought to differ from those of frugivores based on other 

traits, such as seed length and protection as granivores target the seed instead of the 

pulp. To look for their nutritional preferences, biochemical analyses of seeds are 

necessary. The finding that drupes dominate the diet of rodents may indeed indicate a 

protection of the large seeds against seed predation. 

Besides the obvious fruit traits phenology further has a strong impact on food 

composition. During bottlenecks, feeding preferences are probably not as prevalent as 

when fruit is abundant. Therefore frugivores have to be flexible in their dietary strategy by 

enlarging their home range and/or modifying their activity pattern (such as E. f. collaris, 

Donati 2002), leaving the study area (like Coracopsis vasa), foraging outside the littoral 

forest (like P. rufus) (Bollen and Van Elsacker 2002a, Chapter 3a) or going into torpor 

when food is scarce (such as Cheirogaleus spp. and M. rufus). Goodman and Ganzhorn 

75

Chapter 3 

(1997) already pointed out that frugivores in Madagascar might have adapted to lean 

periods by having broad diets (parrots and lemurs) and moving considerable distances in 

search of food (P. rufus and fruit pigeons). For post-dispersal granivores food availability 

is less problematic as seeds are available much longer than fruits. 

Thus in general few traits consistently determine food selection of the thirteen 

consumer species in the littoral forest. There may be several reasons for this. First, most 

frugivores in Sainte Luce seem to be flexible to eat what is available. Secondly, the large 

dietary overlap among frugivores at our site indicates diffuse mutual relationships 

between plant and consumer species, which is similar to most other study sites (Terborgh 

1983; Gautier-Hion et al. 1985; Dowsett-Lemaire 1988; Debussche and Isenman 1989; 

Kitamura et al. 2002; Carlo et al. 2003). In this respect, Fleming (1979) pointed out that in 

the Paleotropics dietary overlap is generally higher than in the Neotropics. This is 

probably linked to the higher spatio-temporal patchiness of fruit resources here that 

favours dietary generalisation, higher inter-specific dietary overlap, and fewer coexisting 

species if food levels become critically low. This may indeed be true for the littoral forest 

where high inter and intra-annual differences can be found in ripe fruit availability (Bollen 

and Donati, Chapter 1). This might represent one of the reasons why Madagascar's guild 

of frugivores is depauperate (Fleming et al. 1987; Langrand 1990; Ganzhorn et al. 1999b; 

Wright 1999) and has higher levels of dietary overlap than found at other sites. 

Diet of frugivores and their role in seed dispersal and/or predation 

E. f. collaris has a highly frugivorous diet at Sainte Luce (74.0% ripe fruits, 5.4% unripe 

fruits) (Donati 2002), which corresponds with findings from other studies on Eulemur 

species (Overdorff 1993a,b; Rigamonti 1993; Curtis 1997; Birkinshaw 1999, 2001). E. f. 

collaris can be considered a sequential specialist, feeding on a wide array of endemic 

fruit species (111 species) but with only two to three dominant fruit species each month 

(Donati 2002). Their relatively large home range (up to 100ha) and extensive day range 

lengths (1500-3500m) (Donati 2002) indicate that long distance seed dispersal within a 

fragment is likely. Furthermore, being the largest frugivores they represent a large 

proportion of the frugivore biomass in this ecosystem. They eat high amounts of fruit 

throughout the year and are the only ones that are able to swallow larger seeds. All these 

aspects suggest that indeed E. f. collaris is one of the most important seed dispersers in 

this ecosystem. The only limiting factor is that this species is reluctant to cross the 

grassland between fragments and thus rarely disperses seeds across the boundaries of 

the fragment. This important role in seed dispersal has been found for other Eulemur and 

lemur species (Ralisoamalala 1996; Scharfe and Schlund 1996; Dew and Wright 1998; 

Ganzhorn et al. 1999a; Birkinshaw 1999, 2001; Britt 2000). 

The smaller nocturnal lemurs seem to have a less diverse fruit diet and can be 

considered more omnivorous (Martin 1973; Petter et al. 1977; Hladik et al. 1980; 

Tattersall 1982), even though several studies in different forest types found that a high 

proportion of fruit is included in the diet of Microcebus rufus (Wright and Martin 1995; 

Atsalis 1999). Fietz and Ganzhorn (1999) recorded 25 fruit species in the diet of 

Cheirogaleus medius in the western dry deciduous forest of Kirindy (CFPF). Atsalis 

(1999) scored 24 fruit species of M. rufus in the mid-altitude humid forests of 

Ranomafana. Compared with these numbers our list of food items likely represents the 

bulk of their diet in the littoral forest. These lemurs are smaller in body size, eat less, 

occupy limited ranges (1-4ha, Fietz 1999; Atsalis 2000), and have a rather limited gape 

size. Furthermore in Sainte Luce as in other Malagasy forests (Fietz and Ganzhorn 1999; 

76


Schmid 2000) Cheirogaleus spp. and M. rufus go into torpor from May through October, 

which restricts their feeding and thus also dispersal activity to austral summer. These 

species are often found at forest edges and in secondary forest throughout Madagascar 

(Petter et al. 1977) as well as in Sainte Luce. This makes them important seed dispersers 

for small-seeded plant species that fruit during austral summer. So far, few data are 

available on the dispersal role of these nocturnal lemurs but Wright and Martin (1995), 

Ganzhorn and Kappeler (1996), Dew and Wright (1998), and Atsalis (1999) have 

suggested that they are indeed possible seed dispersers for small and medium sized 

fruits, which are found intact in their droppings. 

Pteropus rufus feeds on a wide array of endemic plant species in Sainte Luce and is 

the most important long distance seed disperser in the littoral forest. A colony of 300 

Pteropus rufus eats high quantities of fruit each night, defecates during flight, and covers 

great distances between sleeping and feeding roosts (up to 50km) (Bollen and Van 

Elsacker 2002a, Chapter 3a). By bridging isolated forest fragments P. rufus helps to 

ensure genetic exchange between plant populations and forest fragments and even 

regeneration in clearings. Passage through the gut does not seem to destroy the seeds 

(Bollen and Van Elsacker 2002a, Chapter 3a). 

Few data exist on the diet of both Malagasy fruit pigeon species. However, other 

studies in the Paleotropics recorded a frugivorous diet for seed dispersing fruit pigeons 

(Van der Pijl 1969; Corlett 1998; Dowsett-Lemaire 1988; Oliviera et al. 2002), even 

though Snow (1981) described Treron australis as a seed predator of Ficus seeds. 

Goodman et al. (1997b) mentions as well that a large component of the diet of T. 

australis ate Ficus fruits. Fruit pigeons have large home ranges and can wander daily 10- 

12km from their roost (Dowsett-Lemaire 1988; Fleming 1992). In Sainte Luce, they feed 

on ripe fruits only and always swallow the entire fruit, digesting only the fleshy parts and 

defecating the seeds, as observed by Van der Pijl (1969), Corlett (1998) and Oliveira et 

al. (2002). In the littoral forest fruit pigeons seem to be efficient seed dispersers of 

numerous plant species because they have wide gapes (Corlett 1998), cover great 

distances and use secondary and disturbed habitats dispersing also seeds from pioneer 

and heliophil species from perches in a range of habitats. Both short and long-distance 

seed dispersal occurs. Preliminary results of germination trials demonstrate that passage 

though the digestive tract of Alectroenas madagascariensis does not have a negative 

impact on germination (Bollen unpubl. data). 

Hypsipetes madagascariensis is an important seed disperser as well, feeding at 

lower heights than fruit pigeons and parrots. This species swallows ripe fruits entirely and 

defecates the seeds unharmed (Birkinshaw 2001; Bollen unpubl. data). Bulbuls are very 

tolerant to disturbance and often the most common frugivores and dispersal agents in 

degraded tropical and subtropical forests including urban areas (Corlett 1998). In Sainte 

Luce this species could be encountered in intact forest as well as on forest edges. 

Only limited data are available on the diet of Coracopsis nigra (Goodman et al. 

1997a; Hampe 1998; Dowsett 2000), thus the 37 food species that were recorded in 

Sainte Luce seem to be quite representative. Most data are from C. nigra, which is 

present at the site year round, whereas C. vasa is encountered less frequently and during 

certain periods of the year only. Coracopsis nigra destroys seeds of many endemic plant 

species. They can be considered pre-dispersal seed predators in this ecosystem. 

Occasional seed dispersal may occur but for few species and on rare occasions only. 

Even though substantial differences occur between the numerous parrot species present 

worldwide, they are often referred to as granivores (Janzen 1981; Jordano 1983; Clout 

77

Chapter 3 

1989; Galetti 1993; Saini et al. 1994; Pizo et al. 1995; Corlett 1998; Renton 2001). For 

Madagascar Goodman et al. (1997a; 1997b) observed seed predation of Ficus seeds by 

Coracopsis spp., while Böhning Gaese et al. (1999) described C. nigra as an occasional 

seed disperser for Commiphora guillaumini. 

According to the literature most Streptopelia spp. feed largely on dry seeds and fruits 

and destroy most of the seeds they swallow in their muscular gizzards (Corlett 1998). 

However, S. decaocto is said to regurgitate some large seeds (Corlett 1998). S. picturata 

was most often seen feeding on the ground but no details on its feeding behaviour could 

be obtained at our study site. Nevertheless this species is suspected to be a postdispersal 

seed predator destroying most seeds, which is confirmed by Goodman et al. 

(1997b). Obviously our list of food items sampled for this bird is largely underestimated 

compared to their actual diet. 

Before humans settled, the only seed predators in Sainte Luce were Coracopsis spp., 

S. picturata, and Eliurus webbi. The latter now has to contend also with a non-native 

rodent species (Rattus rattus) even though there are no indications for food or habitat 

competition between these two rodent species (Ramanamanjato and Ganzhorn 2001; 

Ganzhorn 2003). Evidence of post-dispersal predation by rodents destroying the seeds of 

fifty plant species was found, but Goodman and Sterling (1996) and Ganzhorn et al. 

(1999a) suggested that native Malagasy rodents may store seeds in their burrows but no 

evidence has been found so far for Eliurus webbi in the littoral forest. Goodman (1994) 

describes a burrow of E. webbi in Andringitra containing 20 seeds of Cryptocarya sp. of 

which only half were eaten. In Sainte Luce a few seeds were observed to escape total 

destruction when germinating from the rodents’ food piles. This diet list can serve as a 

first indication on the rodents’ food species in the littoral forest, taking into account that 

completely digested seeds are missing here. Nevertheless by targeting seeds of selected 

species, rodents can significantly alter forest composition (DuPuy 1996; Spehn and 

Ganzhorn 2000). 

Besides the described frugivores (Table 1), there are also other animal species in 

Sainte Luce that occasionally feed on fruits. Bird species such as Coua caerulea 

(Cuculidae) and Zosterops maderaspatana (Zosteropidae) (also observed by Goodman 

et al. 1997a) could be seen feeding on fruits in Sainte Luce, as well as the terrestrial 

mammal Tenrec ecaudatus (Tenrecidae) and the fruit bats Rousettus madagascariensis 

and Eidolon dupreanum (Pteropodidae), but the latter species were very difficult to track. 

Local people may also act as seed dispersers when collecting several fruit species (n=12, 

App. I). As secondary seed dispersers Pheidole spp. (Myrmicinae) seem to be important 

and they were observed to transport intact seeds of at least 20 species to their ground 

nests, thus matching the important role of ants in other forests of Madagascar (Voigt et al. 

2002; App. I). 

As mentioned above, the dispersal quality of the frugivores differs substantially. 

Frugivores are active at different times of the day and year, forage at different heights, 

have distinct feeding behaviour, gut passage rates, and seed shadows. While birds may 

defecate from nearby or far away perches either within primary forest or in the clearings, 

flying foxes defecate during flight or under feeding or sleeping roosts, rodents 

concentrate seeds at burrows or feeding sites and lemurs move seeds within the forest 

fragment only. In the end, the combined action of a variety of fruit-eating vertebrates with 

distinct seed shadows produces a very heterogeneous transport of seeds, which is very 

important in the severely fragmented and degraded littoral forest to ensure regeneration 

of most plant species within and outside forest fragments. So even though dietary overlap 

78


among consumers based on species partitioning is high in Sainte Luce, the different 

frugivore species are not ecologically redundant due to their differential impact. Moreover, 

the actual diet overlap may be lower than described here because proportional use of diet 

items should be included (Poulsen et al. 2002). In terms of conservation E. f. collaris is of 

crucial importance for the dispersal and regeneration of large seeded plant species, P. 

rufus for long distance dispersal across fragment boundaries and frugivore bird species 

for enhancing succession and regeneration of plants in degraded areas. 

Although our dietary records for each consumer species are not exhaustive, it is 

assumed that the interactions reported here are an unbiased sample and thus that the 

patterns found provide an accurate picture of how interactions are organized in the littoral 

forest on a species-specific as well as on a community level. On the one hand, speciesspecific 

fruit choice is to a limited extent determined by a particular set of morphological, 

biochemical, and phenological traits. However, on the other hand, there is substantial 

dietary overlap among species. Unpredictable food availability in the littoral forest may 

have led to this diet generalisation and to a species-poor frugivore guild as well. 

However, each frugivore has its own particular impact in the seed dispersal of plant 

species. As the littoral forests become more fragmented, the remaining patches become 

increasingly isolated, and heterogeneous seed transport within and between forest 

patches becomes more critical for long-term species survival. 


We are very grateful to QMM (QIT Madagascar Minerals) for providing logistics and infrastructure at 

the study site. In particular we thank Manon Vincelette, Jean-Baptiste Ramanamanjato, and 

Laurent Randrihasipara of the QMM Environmental and Conservation Team. Special thanks go to 

Giuseppe Donati, Nicoletta Baldi, and Valentina Morelli from the University of Pisa (Italy) for making 

their data on fruit eating by Eulemur fulvus collaris available. This draft benefited greatly from 

suggestions of Steve Goodman, for which many thanks. This study was carried out under a 

Collaboration Agreement between the Department of Animal Biology and Department of 

Anthropology of the University of Antananarivo and the Institute of Zoology of Hamburg University 

and QMM. The first author is supported by a grant from the Belgian Fund for Scientific Research, 

Flanders (FWO). We thank the Flemish Government for structural support to the CRC of the RSZA. 

79

Chapter 3 

Appendix I. All plant species that are included in the diet lists of the consumer species with indication of the method how these data were 

gathered (O:observation, T: feeding marks; F: faecal droppings). Abbrevations of the species are Efc for Eulemur fulvus collaris, Ch for 

Cheirogaleus spp., Cm for C. medius, CM for C. major, Mr for Microcebus rufus , Am for Alectroenas madagascariensis, Ta for Treron 

australis, Sp for Streptopelia picturata, C for Coracopsis nigra , Hm for Hypsipetes madagascariensis, Pr for Pteropus rufus, Rod for both rodent 

species, Rr for Rattus rattus and Ew for Eliurus webbi. SD refers to secondary seed dispersal by Pheidole sp. and Hss stands for humans. For 

certain plant species no antanosy names existed, so codes (such as x205, FT62) or other descriptions to distinguish between species were used. 

Vernacular 

Lemurs Birds Bats Rats Other 

Family name Species name 

Antanosy names Efc Ch Cm CM Mr Am Ta Sp C Hm Pr all Rr Ew SD Hss 

? 

? 

fanotabe 

F 

? 

? 

sarikafe 

O 

? 

? 

vahifotsy be 

O 

? 

? 

x205 

T 

? 

? 

x209 

OF 

Anacardiaceae Campnosperma micranteium roandria 

OF 

Poupartia chapelieri 

sisikandrongo OF OT OT OT OT 

O 

T O 

Protorhus cf. lecomtei 

kangy 

F 

Annonaceae Monanthotaxis cf. malacophylla vahimbotany OF OT OT 

O 

T 

Polyalthia capuronii 

menapeka 

OT 

T 

Polyalthia madagascariensis fotsivavo 

OF OT OT OT OF O OF F T 

Polyalthia sp.1 

fotsivavo géante O 

Polyalthia sp.2 

fotsivavo marécage OF 

Apocynaceae Cabucala madagascariensis tandrokosy 

OT 

O 

O O 

Araliaceae Polyscias sp. 

voatsilana sp2 OF 

OF O O O O F 

Schefflera rainaliana 

voatsilana sp1 O 

O O O 

Arecaceae Dypsis fibrosa 

boakandambo OF 

T T T 

Dypsis nodifera 

raotry 

OF OT OT OT 

TF 

Dypsis prestoniana 

boakabe 

OF OT OT OT OF O O O O 

O 

Dypsis saintelucei 

telopolombilany OF 

Dypsis scottiana 

raosy 

OF OT OT 

F F 

T O 

Bignoniaceae Ophiocolea delphinensis akondronala 

F 

T 

O 

Phyllarthron ilicifolium 

zahambe 

OF 

Phyllarthron sp. 

zahambe manongaroa F 

Burseraceae Canarium boivinii 

ramy 

O 

T T 

Canellaceae Cinnamosma madagascariensis vahabatra 3eM OTF 

F 

Capparaceae Crataeva obovata 

belataka 

OTF 

T 

Celastraceae Mystroxylon aethiopicum voavoantatsimo OT 

O 

Polycardia phyllanthoides fandrianakanga F 

Clusiaceae Garcina chapelieri 

haziny tomate OT 

Clusiaceae Garcinia cf/aff madagascariensis disaky kely 

OT 

T 

Psorospermum revolutum harongampanihy 

T 

OT 

O 

O O 

Combretaceae Terminalia fatraea 

katrafa 

OF OT OT OF O OTF T O 

80

Appendix 1 Continued 

Other 

SD Hss 

Rats 

Rr Ew 

all 

Bats 

Pr 

Hm 

C 

O 

Birds 

Sp 

Ta 

Am 

Mr 

Lemurs 

Ch Cm CM 

Efc 

T 

T 

T 

T 

OT 

O 

O 

F 

O 

O 

T 

O 

OF 

OF 

OF 

OTF 

OF 

OF 

OT 

OTF 

O 

OF 

OT OTF OTF 

F 

OF 

OF 

T 

O 

O 

O 

F 

O 

O 

T 

T 

T 

T 

T 

T 

T 

T 

OF 

OTF 

F 

OT 

OT 

OT 

OTF 

OT 

OT 

OT 

OTFOT 

OTFOT 

OTF 

OF 

OF 

TF 

O 

O 

O 

O 

O 

O 

O 

O 

OF 

O 

T 

O 

T 

T 

T 

T 

T 

T 

T 

T 

T 

T 

F 

F 

OF 

F 

OT 

O 

OT 

OT 

OT 

F 

OT 

OF 

OT 

OT 

OF 

OT 

OT 

OT 

OF 

O 

T 

O 

OF 

O 

O 

O 

O 

OT 

OF 

T 

F 

T 


F 

OF 

OF 

T 

T 

T 

O 

O 

O 

T 

O 

F 

F 

F 

F 

F 

T 

F 

OT 

O 

O 

OTF 

Vernacular 

Antanosy names 

rehiba 

vahihazo 

hazomainty blanc 

hazomainty 

x201 

sanga 

tsilantria 

fangora sp1 

fangora sp2 

bamby 

famanta 

mocarana 

x225 

voapaky lahy 

voapaky vavy 

voapaly lahy ZJ 

mampay 

sotro 

fandramana 

bemalemy 

ramirisa 

hazofotsy 

zorafotsy 

zoramena 

kambatrikambatri 

zorala 

voatsimatra 

hazomamy marécage 

hazomamy an ala 

tavolohazo février 

resonzo 

tavolohazo novembre 

varongy 

lokaza 

falinandro 

tavolobotroka 

lendemilahy 

lendemibe 

velomihanto sp.1 

Species name 

Agelea pentagyna 

Dichapetalum sp. 

Diospyros sp.1 

Diospyros sp.2 

Elaeocarpus alnifolius 1 

Elaeocarpus alnifolius 1 

Vaccinium emirnense 

Erythroxylum braxifolium 

Erythroxylum nitidilum 

Anthostema madagascariensis 

Euphorbia tetraptera 

Macaranga perrieri 

Suregada baronii 

Uapaca ferruginea 

Uapaca littoralis 

Uapaca thouarsii 

Cynometra cf. cloiselii 

Phylloxylon xylophylloides 

Aphloia theiformis 

Bembicia uniflora 

Homalium louvelianum 

Ludia antanosarum* 

Ludia antanosarum* 

Scolopia orientalis 

Brexia sp. 

Dicoryphe stipulaceae 

Salacia madagascariensis 

Apodytes dimidiata 

Apodytes sp. nov. 

Beilschmiedia madagascariensis 1 

Beilschmiedia madagascariensis 1 

Cryptocarya sp. 

Ocotea sp. 

Barringtonia butonica 

Dracaena reflexa var. nervosa 1 

Dracaena reflexa var. nervosa 1 

Anthocleista longifolia 

Anthocleista madagascariensis 

Bakerella ambongoensis 

Family name 

Connaraceae 

Dichapetalaceae 

Ebenaceae 

Elaeocarpaceae 

Ericaceae 

Erythroxylaceae 

Euphorbiaceae 

Fabaceae 

Flacourtiaceae 

Grossulariaceae 

Hamamelidaceae 

Hippocrateaceae 

Icacinaceae 

Lauraceae 

Lecythidaceae 

Liliaceae 

Loganiaceae 

O 

OF 

Loranthaceae 

81

Chapter 3 

Other 

SD Hss 

Rats 

Rr Ew 

82 

all 

Bats 

Pr 

F 

Hm 

O 

C 

O 

Birds 

Sp 

Ta 

Am 

F 

Mr 

OF 

Lemurs 

Ch Cm CM 

OF 


Efc 

O 

O 

T 

T 

T 

T 

OT 

OT 

O 

O 

O 

T 

T 

T 

T 

F 

F 

F 

F 

F 

O 

O 

O 

OF 

O 

O 

O 

O 

O 

O 

O 

O 

F 

OF 

OF 

O 

O 

F 

O 

OF 

O 

OF 

OF 

T 

T 

OT 

OT 

O 

O 

O 

OF 

OF 

O 

O 

O 

O 

O 

O 

O 

T 

T 

T 

O 

O 

O 

T 

O 

O 

O 

O 

O 

O 

O 

O 

O 

OF 

O 

O 

O 

O 

T 

O 

O 

OF 

OF 

OF 

O 

T 

T 

T 

T 

T 

O 

OT 

OT 

OT 

OT 

F 

OF 

O 

OF 

O 

F 

O 

F 

F 

F 

F 

O 

OT 

OT 

OT 

OT 

O 

O 

O 

O 

T 

O 

OF 

OF 

O 

OF 

O 

O 

F 

O 

Vernacular 

Antanosy names 

velomihanto sp.2 

voatrotoky 

sarigoavy 

faritsaty 

amborabe 

ambora 

amboralahy 

aviavy 

fihamy 

nonoka 

beronono 

tsilaka 

mafotra sp.1 

mafotra sp.2 

taratasy 

ropasy sp.1 

ropasy sp.2 

guavy 

rotry ala 

rotry mena 

hazombato 

vahifotsy kely 

hazondraotry 

belavenoka 

x236 

vahabatra 

fandranabo 

vakoanala 

fandranabotonboky 

tainbarika 

Rubiaceae ZJ 

fantsikaitramainty 

x203 

x210a 

x210 

vahilengo 

lehengobe 

fantsikaidroka 

fantsikaitra 

Species name 

Bakerella sp. 

Tristemma mauritianum 

Malleastrum mandenense 

Burasaia madagascariensis 

Tambourissa castri-delphinii 

Tambourissa purpurea 1 

Tambourissa purpurea 1 

Ficus baroni 

Ficus guatteriifolia 

Ficus pyrifolia 

Trilepisium madagascariense 

Myrica spathulata 

Brochoneura acuminata 

Brochoneura madagascariensis 

Embelia incumbens 

Eugenia cloiselii 

Eugenia sp. 

Psidium guayave 

Syzygium sp.1 

Syzygium sp.2 

Campylospermum obtusifolium 

Jasminum kitchingii 

Noronhia cf. lanceolata 

Noronhia sp.1 

Noronhia sp.2 

Olea sp. 

Pandanus aff. longistylus 

Pandanus dauphinensis 

Pandanus rollotii 

genus indet. 

Canthium sp. 

Canthium variistipula 

Ixora sp. 

Mapouria aegialodes 

Mapouria sp. 

Morinda cf. umbellata 

Morinda rigida 

Peponidium sp. 

Plectronia densiflora 

Family name 

Loranthaceae 

Melastomataceae 

Meliaceae 

Menispermaceae 

Monimiaceae 

Moraceae 

Myricaceae 

Myristicaceae 

Myrsinaceae 

Myrtaceae 

Ochnaceae 

Oleaceae 

Pandanaceae 

Rubiaceae

Vernacular 

Species name 

Antanosy names Efc Ch Cm CM Mr Am Ta Sp C Hm Pr all Rr Ew SD Hss 

Psychotria sp.1 

tanatananala 

O 

T 

Pyrostria sp. 

fantsikaitrafotsy O 

Rothmannia mandenensis taholagna 

OTF 

F T 

O 

Saldinia littoralis 

mangavoa 

O O 

O 

O O 

Tarenna thouarsiana 

FT 62 

F 

Tricalysia cf cryptocalyx hazongalala 

F 

O 

O F 

Vepris eliotii ampoly 

OF OT OT O F O F 

lahinampoly 

OF O 

O 

T 

fitoravina 

OTFOT 

OT OT 

OT T 

sanirambaza 

O O 

O 

ambirimbarika pionair 

T 

sanirambavy 

O 

OT O OT O 

sagnira sp3 

OTF 

OT 

natohetiki 

O 

ambirimbarika O 

OT F 

fotombavy 

OF OF OF OF O 

O 

meramaintso OTFOT 

OT OT OT O OT F 

O O 

fotonday 

OF O 

O 

fandrikatany 

F 

tsilavimbinanto F 

ravenala 

O 

tavolo 

OTF 

T 

fanolafotsy 

O 

T 

andrarezona 

O O O 

O 

nofotrako 

OF 

1 

Vepris eliotii 1 


Lemurs 

Birds Bats Rats Other 

Family name 

Rubiaceae 

Rutaceae 

Vepris fitoravina 

Sapindaceae Macphersonia radlkoferi 

Plagioscyphus jumellei 

Tina thouarsiana 

Tinopsis conjugata 

Sapotaceae Mimusops commersonii 

Syderoxylon beguei 

Sarcolaenaceae Leptolaena sp. 

Sarcolaena multiflora 

Schizolaena elongata 

Smilaceae Smilax anceps 

SphaerosepalaceaeRhopalocarpus 

coriaceus 

Strelitziaceae Ravenala madagascariensis 

Taccaceae Tacca leontopetaloides 

Theaceae Asteropeia multiflora 

Ulmaceae Trema orientalis 

Verbenaceae Vitex chrysomallum 

1 

as indicated by their vernacular name several species correspond to the same scientific name. They could represent different ecotypes of one same species or 

be different species that have no taxonomic names yet. As this is difficult to affirm at the moment we preferred treating all vernacular names as separate species 

throughout this paper. 


83


Feeding ecology of Pteropus rufus 

(Pteropodidae) in the littoral forest of 

Sainte Luce (south-east Madagascar) 

AN BOLLEN, LINDA VAN ELSACKER 

ACTA CHIROPTEROLOGICA 4(1): 33-47 (2002) 

ABSTRACT 

This paper examines bat-plant interactions by focusing on the fruit diet and food selection 

of flying foxes (Pteropus rufus) in the littoral forest fragments of Sainte Luce, south-east 

Madagascar. Analyses of faecal samples and opportunistic observations revealed 40 

endemic plant species in the diet. The flying foxes mainly eat odoriferous ripe and juicy 

berries. No particular fruit colour was predominant in their diet. Both multi-seeded and 

single-seeded fruits are eaten. Small seeds (1–3.5mm seed length) are usually 

swallowed whole. Passage through the digestive tract of the flying foxes does not reduce 

the germination rate of seeds nor the percentage of seeds germinated. This study 

indicates that the role of flying foxes in both short and long distance seed dispersal for a 

large number of endemic tree species of the littoral forest should not be underestimated 

when designing reforestation programs in particular or conservation action plans in 

general. 

INTRODUCTION 

Pteropus rufus is an endemic flying fox in Madagascar and belongs to the Old World 

family Pteropodidae (Megachiroptera). The genus Pteropus lives mainly on islands 

(Cheke and Dahl 1981; Banack 1998). Its representatives are almost entirely frugivorous, 

feeding mostly on fruit pulp, juices, nectar and occasionally also on leaves (Marshall 

1983). Pteropus rufus occurs predominantly in the humid forests in the east and the 

north. Most roost sites are found in the coastal lowlands (Racey pers. comm). 

Although some fragments of littoral forest can be found along the north-eastern coastline 

of Madagascar, most of it is situated in the south-east. In this report we concentrate on 

this south-eastern region and more in particular on the littoral forest of Sainte Luce (Fig. 

1). This type of forest has been considerably reduced in size over time (Ramanamanjato 

2000). Between 1950 and 1995, 3,400 ha, almost half of what was present in 1950, has 

disappeared. This represents a deforestation rate of 760 ha every 10 years (Mir 

Télédétection Inc. 1998). At present only highly degraded forest remnants and very few 

intact forest fragments remain. 

In Sainte Luce a colony of 300–350 individuals of P. rufus inhabits the littoral forest 

fragment ‘S6’ (225ha) (Lewis Environmental Consultants 1992b). This colony has been 

located there for at least 10 years and according to the local people even longer. These 

flying foxes are very easily disturbed when approached as a consequence of severe 

hunting pressure and frequent bush fires in the area (Bollen pers. obs.). Currently there 

85

Chapter 3a 

are at least two other roost sites of flying foxes in the region. The largest one contains 

800–1000 individuals and is found in a private reserve, Berenty (25°00’S, 46°18’E; 

Ramanamanjato pers. comm., Fig. 1). In this gallery forest the flying foxes are more or 

less protected from hunting. Another roost site is found in a sacred forest ‘Enato 

Anandrano’ (24°55’S, 47°00’E) where the flying foxes are protected by the local fady, a 

Malagasy taboo related to the presence of tombs of the ancestors (Ramanamanjato pers. 

comm.). There is no information on the colony size here and its status as it is forbidden to 

enter these forests. In the lowland Anosyennes, up north of Sainte Luce, in Marovony and 

Analalava, two small roost sites, with less than 50 individuals each, were observed in 

isolated forest remnants (Lewis Environmental Consultants 1992b). The current status of 

both bat populations and forest fragments however is unknown (Fig. 1). 

Due to the high degree of fragmentation and degradation of the littoral forest, long 

distance seed dispersal is important to ensure genetic exchange between plant 

communities of different forest fragments. At present not much information is available on 

the diet of Pteropus rufus in these littoral forests. Therefore, the main goal of this 

research is to investigate whether they act as important seed dispersers in this 

ecosystem by determining which plant species are eaten by these flying foxes. This study 

forms part of an extensive ecological research project on the mutual dependence of the 

frugivorous-granivorous guild and the littoral forest flora, more in particular on seed 

dispersal and predation. Because of this, the focus of this study is on frugivory only and 

not on nectarivory, and pollen analyses were not carried out. 

METHODS 

Research site and study period 

The littoral forest of Sainte Luce is located in south-east Madagascar (24º45’S, 47º11’E) 

and is considered to be dense humid evergreen forest (Koechlin et al. 1974). It 

corresponds to the same forest type as mountain rain forest but grows on sandy soils and 

is always found within 2–3km of the coast at an altitude of 0–20m (Lewis Environmental 

Consultants 1992a). Field research was conducted by the first author between November 

1999 and February 2001. During this research period annual precipitation was 2,487mm 

with a mean temperature of 23°C, ranging between 12°C and 33°C. 

The forest fragments of Sainte Luce are considered the least degraded of all. In 1991 

they represented a total area of about 1,947ha. A group of 20 fragments can be 

distinguished, separated by plains of grassland and swamps. The five larger fragments 

(S6, S7, S8, S9, and S17) range in size from 190 to 377 ha (Lewis Environmental 

Consultants 1992a). Distances between these five fragments vary from 1.5 to 5km. Most 

of them have been separated from each other at least since 1950 and have since then 

systematically declined in size due to human impact on the edges (Lewis Environmental 

Consultants 1992a). Today S6, S7, and S8 are further degrading by recent tavy (slash 

and burn followed by cultivation), bushfires, and selective logging (Bollen pers. obs.). 

Fruit diet: faecal analyses and observations 

Most of our data on the fruit diet of P. rufus was obtained by collecting and analysing 

faecal samples. From January 2000 till January 2001 the day roost of the P. rufus colony 

in S6 was visited once a week to collect faecal samples under the roost trees. On each 

visit as many droppings as possible were collected randomly with a minimum number of 

five droppings containing seeds. These samples were analysed for seed content, seed 

86


number per species and seed viability upon returning at the field station. A reference 

collection of the seeds gathered from several fragments of primary and secondary forest 

was made. It should contain the majority of fruits available within the littoral forest during 

our study period. This reference collection was used to identify the seed species during 

faecal analyses. For several Ficus spp., seeds were too similar to allow identification at 

species level. For these species identifications were based on the characteristics of the 

seedlings. 

Observations on feeding behaviour were obtained during tree watches and during 

opportunistic encounters with flying foxes. Because of the difficulty of approaching flying 

foxes at night these data are limited. The following parameters were scored: visitation 

length, defined as the time elapsing between arrival of the first flying fox until departure of 

the last one in the feeding tree, number of individuals feeding and if possible feeding 

behaviour. Ejecta pellets collected under the tree were described by their characteristic 

feeding marks. 

BERENTY 

Ambosary 

ANALALAVA 

ENATO 

ANANDRANO 


10 km 

SAINTE LUCE 

MAROVONY 

Fig 1. On the left a detail of the littoral forests (Marovony, Analalava, Sainte Luce, Mandena and 

Petriky) in the south-east is shown with the five Pteropus rufus colonies’ roost sites indicated and 

on the right a detail shows the five biggest forest fragments of the Sainte Luce littoral forest with 

indication again on the bat’s day roost (S6) and the campsite (S9). 

Food selection: fruit and seed characterisation 

The following characteristics of fruits and seeds were noted for a total number of 175 

individual plant species available in the forest throughout our research period: growth 

form, fruit type, external colour at ripeness, odour, pulp type, fruit and seed length and 

mass, number of seeds per fruit, water content and fruit skin thickness. Large trees (>6m 

in height), small trees (

Chapter 3a 

Table 1. Overview of the fruit species consumed by Pteropus rufus with indication of plant 

family, species or gender name and vernacular name, type of evidence (O: observation,F: faecal 

analyses, M: gnawing marks), growth form, fruit type, colour at ripeness, odour (S: strong; 

P: present, A: absent), pulp type, fruit length (mm) and weight (g), seed length (mm) and weight 

(g), number of seeds (NS) and percentage of watercontent (H2O). Fruit skin thickness was 

identical for all taxa (easily opened by nail) except for Rothmannia mandenensis (by knife only). 

Family Taxon Vernacular Evid Growth 

name form 

Annonaceae Polyalthia madagascariensis fotsivavo F large tree 

Araliaceae Polyscias sp. voatsilana F large tree 

Arecaceae Dypsis prestoniana boakabe OFM large tree 

Dypsis nodifera raotry MF small tree/shrub 

Bignoniaceae Ophiocolea delphinensis akondronala M small tree/shrub 

Canellaceae Cinnamosma madagascariensis vahabatra F large tree 

Combretaceae Terminalia fatraea katrafa OMF large tree 

Ericaceae Vaccinium emirnense tsilantria F small tree/shrub 

Euphorbiaceae Uapaca ferruginea voapaky lahy OF large tree 

Uapaca thouarsii voapaky lahy ZJ F large tree 

Uapaca littoralis voapaky vavy OMF large tree 

Ludia antanosarum 1 hazofotsy F small tree/shrub 

Ludia antanosarum 1 zorafotsy F large tree 

Flacourtiaceae Scolopia orientalis zoramena OF large tree 

Lauraceae Beilschmiedia madagascariensis resonzo M large tree 

Ocotea sp. varongy F large tree 

Liliaceae Dracaena reflexa var. nervosa 1 falinandro F small tree/shrub 

Dracaena reflexa var. nervosa 1 tavolobotroka F small tree/shrub 

Loranthaceae Bakerella ambongoensis velomihanto sp.1 F epiphyte 

Bakerella sp. velomihanto sp.2 F epiphyte 

Loganiaceae Anthocleista madagascariensis lendemibe F large tree 

Anthocleista longifolia lendemilahy F small tree/shrub 

Monimiaceae Tambourissa purpurea 1 ambora F small tree/shrub 

Tambourissa purpurea 1 amboralahy F small tree/shrub 

Moraceae Ficus baronii aviavy F large tree 

Ficus guatteriifolia fihamy F large tree 

Ficus pyrifolia nonoka F large tree 

Myrtaceae Syzygium sp.2 rotry mena OMF large tree 

Rubiaceae Canthium variistipula fantsikaitramainty F large tree 

Tricalysia cf. cryptocalyx hazongalala lahy F small tree/shrub 

Tricalysia sp. hazongalala vavy F small tree/shrub 

Rothmannia mandenensis taholagna F large tree 

genus indet. tainbarika F large tree 

Ixora sp. x203 F small tree/shrub 

Mapouria aegialodes x210a F small tree/shrub 

Mapouria sp.2 x210 F small tree/shrub 

Rutaceae Vepris elliotii ampoly F large tree 

Sapotaceae Sideroxylon beguei var. saboureaui ambirimbarika MF large tree 

Sarcolaenaceae Sarcolaena multiflora meramaintso F large tree 

Saxifragaceae Brexia sp. kambatrikambatri F small tree/shrub 

1 as indicated by their vernacular name in three cases we found two plant species corresponding to the same 

scientific name. They could represent different ecotypes of the same species or they might be different species 

that have no taxonomic names yet. As this is difficult to affirm at the moment we preferred including all vernacular 

as separate units in our diet list and the 40 plant species are also treated as separate species throughout this paper 

88



Fruit Colour Odour Pulp Fruit Seed 

NS H20 type type length weight length weight (%) 

berry orange A soft & juicy 12.2 0.3 7.8 0.12 1 72 

drupe green A jiucy & hard 5.0 0.1 3.6

Chapter 3a 

Pulp type could be juicy or arillate pulp, fibrous pulp or pulpless. Fruit and seed length 

and mass were measured with callipers and an electronic scale ‘Kernbalans NM60’ with 

respectively 0.01mm and 0.01g precision. These measures along with the number seeds 

were each subdivided in three classes. The water content of the fruits was calculated by 

comparing fresh mass and dry mass, after three days of drying in an oven. The fruit skin 

thickness was divided into the following categories; easily cut by fingernail, by a knife or 

by a secateur. 

We used X 2 -analyses to compare the flying foxes’ food selection with the overall fruit 

availability. A herbarium of all fruit species collected during our study was made in the 

field and identified by Dr. Johny Rabenantoandro at Missouri Botanical Garden in 

Antananarivo. 

Germination trials 

The seed viability after passage through the digestive tract was tested by simple 

germination trials, in which three different treatments were used; defecated seeds, control 

seeds and control fruits. Ripe control seeds and fruits were collected on or under the 

parent plant within a restricted time frame. All seeds were sown at 1cm depth in plastic 

pots filled with 8cm sterile sand and a 1.5cm humus layer on top. Pots were placed under 

a shed for protection from direct sunlight. The faecal seeds were still surrounded by their 

faecal matrix, when sown. For each treatment sowing was done at the same time and 

under the same overall ecological conditions in order to standardize procedures. The 

germination rate, defined as time to first germination, and percentage germinated seeds 

were scored weekly over a period of at least six months. 

RESULTS 

Fruit diet: faecal analyses and observations 

Over a 13-month period at least 40 fruit species (27 genera, 21 families) were eaten by P. 

rufus (Table 1). The flying foxes were observed to eat six fruit species while evidence of 

38 species was found in the faecal samples. Eight species were identified as eaten by 

flying foxes based on the marks on the ejecta pellets found under fruit trees. At least five 

seed species remained unidentified in the faecal samples and therefore are not shown in 

Table 1. The family Rubiaceae, including genera Canthium, Rothmannia, Tricalysia, 

Mapouria, Ixora, and one unidentified genus, was predominant, representing eight fruit 

species and thus 20% of the plant species in the diet of the flying foxes. Of the remaining 

20 families, eleven were represented by only one species, six by two species and three 

by three species. The cumulative distribution curve shows that at the end of the 12-month 

period the curve starts to level out ending in a record of 40 species eaten (Fig. 2). 

January was sampled the second time after completing a year cycle, which did not result 

in a higher number of species sampled. 

Five particular plant species, Ficus guatteriifolia, Syzigium sp.2, Terminalia fatraea, 

Uapaca thouarsii, and Uapaca littoralis were found to be important food sources for the 

flying foxes during the study period. This importance was based on the larger number of 

droppings found containing their seeds, the evidence on ejecta pellets, multiple feeding 

observations or because these species were eaten for at least four successive months, 

often much longer (Table 2). 

Typically one seed species per dropping was found. Out of the 410 faecal samples 

collected only four samples had two different species of seeds. No sample contained a 

90


larger number of species. Twenty-three plant species (61%), from a total of 38 food 

species identified by faecal samples were represented only by one or two seeds per of 

dropping. More than 50 seeds per dropping were found for 13% of the food plant species. 

The remaining 26% of the plant species had two to eight seeds in droppings. 

Most of the direct encounters occurred from April through June while the flying foxes 

were feeding mainly on fruits of Terminalia fatraea, Dypsis prestoniana and Syzigium 

sp.2, and on flowers of Ravenala madagascariensis. Out of a total of 47 independent 

observations (Table 3), the median number of individuals feeding together was 1 (range 

1–15) and the median visitation length was 2 minutes (range 1–48min). Sometimes 

feeding animals would take off carrying one or more fruits to a neighbouring tree, eat 

them and then return to the original fruit tree. On other occasions individuals would 

remain eating in the fruit tree itself. The longest visitation length (48min) and largest 

group size (N=15) were recorded in T. fatraea. During observations it was also clear from 

the falling seeds and ejecta pellets that flying foxes most often only suck out juices from 

the pulp and then systematically drop or spit out the remaining pulp fraction and seeds. 

This feeding behaviour was noticed for T. fatraea, D. prestoniana, and Syzigium sp.2. 

Only few data of feeding on flowers were recorded. On one occasion flower petals of 

Ludia antanosarum (Flacourtiaceae) were noticed to be abundantly present in the faecal 

samples during May while this tree species is blooming. Furthermore the flying foxes 

were observed feeding on the nectar of Ravenala madagascariensis (Strelitziaceae) 

flying from one tree to another on several occasions. 

Number of species 

45 

40 

35 

30 

25 

20 

15 

10 

5 

0 

J F M A M J J A S O N D J 

cumulative distribution 

monthly distribution 

Months 

Fig. 2. The monthly and cumulative distribution of the number of plant species eaten by Pteropus 

rufus over a 13-month period. 

91

Chapter 3a 

92 

Table 2. Temporal distribution of feeding patterns of the bats per plant taxon per month from January 2000 to January 2001 with 

indication of type of evidence (F: faecal samples, O: observation, M: recognizable feeding marks on ejecta pellets) per month. 

Number 

of seeds/dropping 

Max 

- 

- 

- 

8 

- 

- 

ND 

3 

3 

- 

3 

2 

2 

4 

- 

- 

ND 

- 

- 

11 

50+ 

10 

- 

- 

- 

300+ 

Min 

- 

- 

- 

4 

- 

- 

ND 

1 

1 

- 

1 

1 

1 

1 

- 

- 

ND 

- 

- 

2 

27 

1 

- 

- 

- 

5 

Median 

1 

1 

1 

6 

1 

1 

ND 

1 

1 

100+ 

1 

2 

1 

2.5 

8 

4 

ND 

1 

1 

9 

50+ 

4 

1 

1 

100+ 

100+ 

Faecal 

samples 

(N) 

2 

4 

1 

2 

1 

1 

- 

11 

3 

1 

34 

43 

9 

2 

1 

1 

- 

1 

1 

3 

36 

4 

1 

1 

12 

148 

J 

D 

N 

F 

F 

OF OF 

OF 

M 

F 

Number of droppings per taxon is also given. The five most important species are highlighted. 

'Feeding' months 

Taxon 

J F M A M J J A S O 

Dracaena reflexa var. nervosa 'falinandro' 

F 

Dracaena reflexa var. nervosa 'tavolobotroka' 

F F 

Polyalthia madagascariensis 

F 

Polyscias sp. 

F 


OM 

F 


MF 

Ophiocolea delphinensis 

M 

Cinnamosma madagascariensis 

F F F 

Terminalia fatraea 

M OM OM M 

Vaccinium emirnense 

F 


O F OF OF 

Uapaca thouarsii 

F 


M F F F F 

Ludia antanosarum 'hazofotsy' 

F 

F 

Ludia antanosarum 'zorafotsy' 

F 


O 

O 

Beilschmiedia madagascariensis 

Ocotea sp. 

Bakerella sp. 

F 

Bakerella ambongoensis 

F F F 

Anthocleista madagascariensis 

F 


F F 

Tambourissa purpurea 'ambora' 

F 

Tambourissa purpurea 'amboralahy' 

F 

Ficus baronii 

F 

Ficus guatteriifolia 

F F F F F F F F F F F 

F 

F 

F 

F


Number 

of seeds/dropping 

Max 

- 

2 

4 

- 

- 

7 

- 

2 

- 

- 

5 

1 

- 

- 

Min 

- 

1 

1 

- 

- 

1 

- 

1 

- 

- 

5 

3 

- 

- 

Median 

100+ 

1 

2 

1 

1 

1 

2 

1 

1 

1 

5 

1 

9 

1 

Faecal 

samples 

(N) 

25 

44 

10 

1 

1 

5 

1 

4 

1 

1 

1 

3 

1 

1 

J 

D 

N 

'Feeding' months 

M J J A S 

Taxon 

F 

F 

F 

F 

O 

F 

F 

F 

A 

M 

F 

J 

F 

F 

F 

F 

OMOF 

F 

F 

F 

F 

F 

F 

F 

F 

F 

F 

F 

F 

F 

MF 

F 

F 

Ficus pyrifolia 

Syzygium sp.2 

Canthium variistipula 

Tricalysia cf. cryptocalyx 

Tricalysia sp. 

Rothmannia mandenensis 

Tainbarika 

Ixora sp. 

Mapouria aegialodes 

Mapouria sp. 

Vepris elliotii 

Sideroxylon beguei var. saboureaui 

Sarcolaena multiflora 

Brexia sp. 

Number of individuals 

Table 3. The number of observations (N) per tree species with indication of the median, 

range, and average deviation (AD) of the number of individuals feeding in the tree. 

Visitation lengths (in min) are also shown. 

Visitation length 

N 


AD 

0.19 

3.38 

- 

1.44 

- 

- 

- 

1.31 

Range 

1-2 

1-15 

1 

1-7 

1 

1 

1 

1-15 

Median 

1.00 

1.00 

1.00 

1.00 

1.00 

1.00 

1.00 

1.00 

AD 

7.01 

14.16 

0.32 

2.12 

1.00 

0.50 

1.00 

5.44 

Range 

1-15 

1-48 

1-2 

1-15 

1-3 

1-2 

1-3 

1-48 

Median 

5.00 

3.00 

1.00 

2.50 

2.00 

2.00 

2.00 

2.00 

9 

7 

5 

20 

2 

2 

2 

47 


Terminalia fatraea 

Ravenala madagascariensis 

Syzygium sp. 




All species together 

93

Chapter 3a 

Table 4. Comparison of morphological characteristics of fruits (with the corresponding sample 

size and frequency distribution) in the diet of P. rufus and the overall fruits available (n=152-175) 

Fruit parameters 

Total fruit 

availability 

Pteropus 

rufus' diet 

Statistics 

N % 

Growth form 

N % X² df P 

Larger trees 94 54 23 58 1.42 2 NS 

Small trees and shrubs 62 35 15 37 

Others 19 11 

Fruit type 

2 5 

Berry 83 50 25 63 4.64 3 NS 

Drupe 51 30 11 27 

Capsule 21 13 1 2 

Others 12 7 3 8 

Colour of ripe fruits 

Yellow 22 13 5 13 6.00 6 NS 

Orange 10 6 4 10 

Red, pink 29 17 9 23 

Puple, blue 16 9 5 13 

Brown 44 25 9 23 

Green 39 22 8 20 

Others 14 8 0 0 

Odour 

Absent 59 35 16 39 3.30 2 NS 

Present 47 28 6 15 

Strong 63 37 

Pulp type 

18 44 

Juicy 124 72 35 88 12.56 3


Food selection: exploited versus available food items 

In order to gain some insight into the flying foxes’ food selection, the different variables of 

several fruit and seed parameters were compared for the 175 available and 40 exploited 

food species (Table 4). Focusing on exploited food species only, it was apparent that 

mainly large trees and to a lesser extent small trees and shrubs are exploited for their 

fruits. No herbs and vines occur in the diet list. Berries are the fruit type most represented 

in the diet followed by drupes. Fruits with a strong odour are predominant in the diet, 

while all colours are present in approximately similar percentages. Fruits with many tiny 

seeds as well as one- to two-seeded fruits are well represented. In general fruit skin 

thickness is minimal and most fruits have a water content over 60%. Furthermore juicy 

fruits with a length between 10–30mm are most common in the list, but no particular fruit 

mass was most abundant. Seed length is often smaller than 10mm and seed mass less 

than 0.1g. The threshold for seed swallowing at our study site is as much as 10mm, with 

4.4mm being the median diameter (N=38). 

Much of the differential use of fruits can be explained by a differential availability. 

There is only a significant difference between observed and expected values for the 

parameters pulp type, seed length and seed mass (Table 4). Fruits with juicy pulp are 

clearly preferred. Fibrous fruits and fruits without pulp, even though available, are not 

consumed by the flying foxes at all. Fruits with seed length smaller than 10mm are 

preferred to longer seeded-fruits. The most preferred seed mass is under 0.1g, but the 

0.1–1.0g category still makes up one third of their food choice, while seeds heavier than 

1g seem to be avoided. After sequential Bonferroni adjustment (Rice 1989 but see Moran 

2003) none of these preferences remained significant. 

Germination trials 

Our faecal analyses show that seeds of at least 38 plant species pass through the 

digestive tract. Due to the scarcity of simultaneous presence of defecated seeds, control 

seeds and fruits, it was not always possible to obtain the same number of duplicates or 

the same number of seeds for all treatments. 

None of the defecated seeds looked damaged. Only five species provided enough 

seeds and fruits at the same time to start a germination experiment (Table 5). Passage 

through the digestive tract had neither a negative nor a positive impact on the 

germination rate and percentage of seeds germinated. It appears that seeds from intact 

control fruits take more time to germinate than seeds of faecal samples and control 

seeds. Numbers were too small, however, to allow statistical comparison. 

DISCUSSION 

Fruit diet 

Quantitatively 

The diet of P. rufus studied at Sainte Luce consists of 40 endemic plant species of the 

littoral forest. Even though our data set represents the most complete information 

available on the fruit diet of P. rufus in littoral forests today, it is probably an 

underestimation of their overall fruit diet for several reasons. First, by focusing mainly on 

faecal sample content, larger seeds, that are often spat out and less commonly eaten 

food species can be missed. Secondly, exotic species that are neither important nor 

typical for the littoral forest were omitted from our study. It is likely that the five seed 

species that could not be identified may represent seeds from such exotic species. They 

may also be fruit species eaten in other forest types and were as such not present in our 

95

Chapter 3a 

reference collection. It cannot be excluded that the mountain rain forest, with a different 

floral composition (Koechlin et al. 1974) lies within the foraging range of the colony 

studied. Finally, there might also be an important temporal bias since a number of tree 

species in these littoral forests do not fruit annually but bi-annually or even less often 

(Randrihasipara pers. comm.; Bollen and Donati, Chapter 1). One year of sampling is not 

enough to establish the complete fruit diet of P. rufus. Nevertheless, our species 

accumulation curve indicates that a large proportion of the diet is indeed already known. 

Long-term studies are needed to further complete the diet list. 

Table 5. The number of weeks needed for the first germination (a) and the percentage of 

germinated seeds (b) for the three treatments: faecal seeds, control seeds, and control fruits. 

The values given here are median, average deviation (AD), and sample size (N). 

a) 

Germination rate 

Species 

Faecal seeds Control seeds Control fruits 

Median AD N Median AD N Median AD N 

Ludia antanosarum 2.5 - 1 3.5 0.5 2 7 - 1 

Terminalia fatraea 25 - 1 22 0.9 3 26 1.2 5 

Syzygium sp. 7 - 1 6 0.9 3 10 6.8 3 

Ficus guatteriifolia 3.5 2.9 8 7 - 1 15.5 2.5 2 

Rothmannia mandenensis 8 - 1 23 - 1 

2 

- 0 

b) 

Percentage germinated 

Species 

Faecal seeds 

Control seeds Control fruits 

Median AD N Median AD N Median AD N 

Ludia antanosarum 25 - 1 62.5 12.5 2 33 - 1 

Terminalia fatraea 60 - 1 30 15.6 3 30 8.0 5 

Syzygium sp. 50 - 1 80 11.1 3 60 13.3 3 

Ficus guatteriifolia 16 9 8 

1 

- 1 

1 

- 2 

Rothmannia mandenensis 50 - 1 40 - 1 

2 

- 0 

1 

as these seeds are very tiny and numerous in fruits it is impossible to know exactly which percentage 

had germinated in both control seeds and control fruits. 

2 

ripe fruits fall apart in small pieces and therefore we could not sow complete fruits for comparison. 

Quantitative data on the diet of P. rufus in other parts of Madagascar are limited. In the 

gallery forest of Berenty the diet of P. rufus contains only 13 plant species, both flowers 

and fruits included (Long and Racey submitted). This much lower number can probably 

be best explained by the lower plant diversity of the much drier gallery forest. Racey et al. 

(in prep) have studied P. rufus in Madagascar for several years at different sites in 

Madagascar and their data on feeding ecology resulted in a diet list of 38 plant species 

for P. rufus with two thirds of these being fruit resources. Racey and Nicoll (1984) 

mention a fruit diet of 18 species for Pteropus seychellensis in the Seychelles and Parry- 

Jones and Augee (1991) found 22 fruit species in the diet of Pteropus poliocephalus. An 

extensive literature survey by Marshall (1983) resulted in a list of 62 plant genera 

consumed for their fruits by all Pteropus spp. (N=67) together. All these numbers 

demonstrate that our diet list including 40 species belonging to 28 genera is quite 

extensive. 

As for the quantitative data of our observations, we believe there may be a bias on 

the visitation length scored, as flying foxes would fly away when they detected our 

96


presence. On the other hand, visitation length could indeed be relatively short because of 

a particular feeding behaviour of these flying foxes to consume fruits in ‘dining roosts’ and 

not in the fruit tree itself on some occasions. This behaviour was also observed for other 

frugivorous flying foxes (Marshall 1983) and for Neotropical frugivorous bat species 

(Morisson 1980; Kalko et al. 1996), even though Kalko et al. (1996) mention a rather 

sedentary feeding mode for Afrotropical flying foxes, which could also be observed in 

Sainte Luce on other occasions. 

Taxonomically 

Of the five plant species mentioned in the results, Ficus guatteriifolia (Moraceae) is likely 

to be the most important one in the diet of P. rufus. This species is available year-round 

due to intra-specific asynchrony in flowering and fruiting and faecal analyses show it was 

consumed the whole year. It forms a staple food and is likely to be a keystone resource 

(definition according to Mills et al. 1993) for these flying foxes. Ficus spp. are also the 

food taxon that is most frequently regarded as important in literature on the diet of the 

flying foxes (Pteropodidae) in the Paleotropics (Cheke and Dahl 1981; Marshall 1983; 

Fujita and Tuttle 1991; Banack 1998) and of fruit bats (Phyllostomatidae) in the 

Neotropics (Heithaus et al. 1975; Fleming et al. 1977; Morrison 1980; Kalko et al. 1996). 

By the same token, Syzigium sp.2 (Myrtaceae) could be considered a keystone fruit 

species as well. It is a very common and widespread tree species, which provides food 

from May up to January. The fruiting cycle of an individual tree is only about two to three 

months long but since the fruiting pattern in the different forest fragments shows a delay 

of a few months, the fruits are available in the area for as long as nine months. In this 

way certain plant species are accessible only to flying frugivores for extended periods. 

Terminalia fatraea (Combretaceae) is a highly preferred food item and one of the 

plant species of which most observations were recorded and large amounts of macerated 

fruit pulp were found under the trees in the morning. From April to July T. fatraea is eaten 

in very large quantities. There may be a parallel between Terminalia catappa consumed 

in the Masoala Peninsula (Hutcheon 1994) and T. fatraea eaten in Sainte Luce. Both act 

as important food species during times of fruit scarcity and therefore play a crucial role in 

the diet of P. rufus. Terminalia catappa has also been regarded as an important food 

species for Pteropus spp. by Cheke and Dahl (1981), Fujita and Tuttle (1991) and 

Banack (1998). 

Several authors mention that fruit bats feed on a taxonomically non-random subset of 

fruits. These so-called ‘bat fruits’ are mainly plant species of following plant families 

Moraceae, Myrtaceae, Sapotaceae, Arecaceae, Piperceae, Solanaceae, Anacardiaceae, 

Guttiferae, Leguminosae, and Combretaceae (Marshall 1983; Fleming 1987; Banack 

1998; Corlett 1998). Our data set contains several of these families. Surprisingly the 

family Rubiaceae, not mentioned by any author, is taxonomically represented by the 

highest species number in our diet list. On the contrary, the plant families Guttiferae and 

Anacardiaceae, both considered ‘bat families’ by Marshall (1983) and Corlett (1998), are 

not eaten by P. rufus in Sainte Luce even though available. Certain plant families might 

indeed be considered ‘bat families’ but the actual diet may still vary greatly according to 

forest type and fruit availability within. 

97

Chapter 3a 

Food selection 

Growth form 

A closer look at fruit characteristics for both eaten and available species seems to point at 

certain food preferences, although at the same time other, non-significant, results were 

rather unexpected. For example according to the analyses the flying foxes did not prefer 

to eat in large trees, which was unexpected as bats were most often observed and 

supposed to feed in large trees (Fleming et al. 1987). The often larger fruit crop available 

at the same time in large trees, the more accessible position and the more easily 

detectable resources both for bats and researchers are probably responsible for this 

result. All important food resources in our diet list involve large trees, but shrubs and 

smaller trees still account for almost 40%. The latter occur in larger numbers in 

secondary forest and are more easily detected and eaten there. Old World pteropodids 

are known to be primarily canopy feeders (Fleming et al. 1987) and prefer primary forest 

to secondary forest (Banack 1998). In our data set most of the fruit species eaten grow in 

intact and primary forest. Both pioneer plant species as well as species from a later 

successive phase were exploited. 

Fruit type versus pulp type and water content 

The flying foxes eat mainly juicy berries with a high water content. Observations, ejecta 

pellets, and literature (Marshall 1983) confirm that fruit juices dominate the diet of flying 

foxes. As indicated in Table 4, only for the parameter ‘pulp type’ a difference was found 

between eaten and available fruits, not for fruit type and water content but these 

parameters are often inter-correlated. In most tropical forests 50–90% of the plant 

species depend on animals for their dispersal (Howe and Smallwood 1982; Fleming et al. 

1987) and among typical endozoochorous fruits juicy berries with a high water content 

form a large proportion, which stands also for the majority of available fruits in the humid 

littoral forest of Sainte Luce. 

Odour and colour 

Most fruits eaten have an odour and even a strong one, which can be related to the bats’ 

well-developed olfactory senses especially used for locating food (Marshall 1983; Kalko 

et al. 1996). Odour was a feature of most (65%) of the fruits available in the forest, which 

may explain why analyses revealed that there was no significant preference for this trait. 

This abundance of odoriferous fruits is probably because a large amount of the available 

fruit species are dependant on mammals for seed dispersal and scent in general is also 

of major importance for mammals when locating and selecting ripe fruits. Besides good 

smell, the flying foxes have also developed large eyes and thus good vision which might 

further help them to locate fruits at night (Marshall 1983). Obviously colour is of little 

relevance since they feed and forage at night and all nocturnal mammals are colour-blind 

(Corlett 1998). This is confirmed by the fact that selection of fruits in favour of a particular 

colour was not observed. 

Size versus mass 

Based on the number of seeds per dropping we presumed that tiny seeds of multi-seeded 

fruits with a length up to 1–3.5mm are likely to be automatically swallowed together with 

the fruit pulp. As for larger one- to two-seeded fruits, with seed length between 3.5– 

20mm, fruit skin and seeds are most often spat out. This feeding behaviour of dropping 

larger seeds and swallowing tiny seeds together with fruit juices was also mentioned by 

98


Marshall (1983). Occasionally some seeds (up to 15mm length) are swallowed as well. 

For all three Uapaca sp., with seed length over 9.6mm, seeds were often swallowed but 

this is probably due to the pulp texture and slippery seeds. The threshold for swallowing 

seeds is reported as being less than 3.2mm diameter for a 600g Pteropus sp. and 

between 3–5mm for flying foxes in general (Corlett 1998; Shilton et al. 1999), which is 

smaller than the 10mm recorded at our study site. 

Our analyses showed no particular fruit mass preference. The fact that Pteropus spp. 

may transport fruits of over 200g (Marshall 1983) means that they are probably not 

limited by masses up to 50g, being the maximum fruit mass that was scored. 

‘Bat fruits’ 

According to Fleming (1979), Marshall (1983), Stashko and Dinerstein (1988), Thomas 

(1988) and Korine et al. (1998) ‘bat fruits’ can be morphologically characterised as 

variable in size with a green or dull colour, a strong and musty odour, high water content, 

pendant position or held away from the foliage. This description corresponds with our 

results meaning that the flying foxes’ food species in Sainte Luce include fruits of all size, 

having no particular conspicuous colour, a strong odour and high water content. 

However, compared to the overall database of available fruits in this forest several of 

these variables are simply characteristic for the majority of fruits. Thus real food selection 

or clear preference cannot be established. Therefore it is important that future studies 

also focus more on all available food resources in an ecosystem rather than studying the 

diet of the bat species only. This way it will be possible to draw broader conclusions on 

real food preferences and typical ‘bat fruits’ compared to the wide array of fruits available. 

The exclusive role of flying foxes in seed dispersal 

Quantitatively important short and long distance seed dispersal 

Pteropus rufus feeds on a huge variety of fruits, which makes this species potentially an 

important seed disperser for a large number and diverse set of endemic plant species in 

the littoral forest. Compared to other frugivores in the littoral forest it is the only one 

capable of long distance seed dispersal since foraging may occur up to 50km away from 

the roost site (Thomas, 1988) thereby bridging isolated forest fragments. This ensures 

genetic exchange between plant populations of different forest fragments, and for very 

small fragments no longer inhabited by other mammal seed dispersers, only flying foxes 

can disperse these fruits. Long distance seed dispersal happens mainly between 

successive feeding trees (0.3–8.3km apart) or even further away between foraging areas 

and roost sites (up to 50km apart) for all ingested seeds (Morrison 1978; Korine et al. 

1999). 

Gut passage rate in flying foxes is often only about 30 minutes (Morrison 1980), 

although there is also evidence for gut retention of food for large periods (>12h or >18h) 

in Pteropodidae (Shilton et al. 1999), which increases the possibility for long distance 

seed dispersal. As digestion can be rapid, large quantities of food can be processed 

every night. This is necessary because being flying mammals, these flying foxes have 

rather high-energy requirements and may eat at least the equivalent of their own body 

mass each night (Marshall 1983; Shilton et al. 1999). In addition, they are very numerous 

in the area. All this probably leads to a massive consumption of fruits and possible 

dispersal of seeds every night by a large number of animals. 

99

Chapter 3a 

Germination experiments 

Izhaki et al. (1995) noted a positive effect of passage through bats’ guts on germination 

rate of the ingested seeds. Nevertheless, Palmeirim et al. (1989), Kalko et al. (1996), 

Iudica and Bonaccorso (1997), Korine et al. (1998), Shilton et al. (1999) as well as this 

study did not find such a relation. There seemed to be no positive or negative impact on 

germination rate and germination percentage of defecated seeds. The results on 

percentage of plants that finally germinated after 6 months are difficult to interpret and 

appear very variable. More extensive experiments under more controlled ecological 

conditions and with more replicas should be carried out to confirm these first preliminary 

data. But, even if germination itself does not profit from the digestive process, more 

important to consider is the distance covered by the flying foxes during gut passage. 

Threats and conservation options 

Unfortunately in Sainte Luce not only the fragmented habitat is at risk, but its inhabitants 

among which the flying foxes, are seriously threatened as well. During our field research 

the whole colony moved once from their original roost site to another one, five kilometers 

west, where they remained from February through May 2000. Afterwards the colony 

returned to the first roost site. At both roost sites, rocks, long sticks and injured patagia 

were indications of severe hunting pressure. Several bush fires in the nearby area 

perturbed the colony even more. Too much harassment might cause the colony to divide 

into smaller groups, and settle elsewhere, leaving the littoral forest deprived of its only 

capable long distance seed disperser. 


We would like to acknowledge QMM (QIT Madagascar Minerals) for providing logistics and 

infrastructure at the campsite. In particular we thank Manon Vincelette, Jean-Baptiste 

Ramanamanjato, Laurent Randrihasipara of the QMM Environmental and Conservation Team and 

Prof. Dr. Jörg Ganzhorn (University of Hamburg) for their support in this research. We are very 

grateful to Dr. Johny Rabenantoandro of Missouri Botanical Garden (Antananarivo) for assistance 

with the identification of the plant species. We also thank Prof. Paul A. Racey (Univ. Aberdeen) for 

providing important literature and for his extensive revisions and useful suggestions. Finally we owe 

great thanks to Giuseppe Donati and the Malagasy assistants in the Sainte Luce campsite, without 

them this research would have been a lot harder. This study was carried out under a Collaboration 

Agreement between Department of Animal Biology and Department of Anthropology of the 

University of Antananarivo and the Institute of Zoology of Hamburg University and QMM. First 

author is supported by a Research Assistantship of the Belgian Fund for Scientific Research, 

Flanders (FWO). We thank the Flemish Government for structural support to the CRC of the RZSA. 

100


The feeding ecology of Coracopsis nigra 

(Psittacidae) in the littoral forest of Sainte 

Luce (south-east Madagascar) 

AN BOLLEN, LINDA VAN ELSACKER 

OSTRICH (SUBMITTED) 

ABSTRACT 

This paper describes the diet and feeding ecology of the parrot species Coracopsis nigra 

in the littoral forest of Sainte Luce, south-eastern Madagascar. Forty plant species were 

recorded being eaten by these parrots over a 14-month study period. C. nigra is an 

opportunistic feeder and eats a large variety of flowers (10%), ripe and unripe seeds 

(68%) and fruits (22%). Of all fruit species consumed, the majority (70%) of fruits are also 

eaten in unripe condition, which may lead to an advantage over potential food 

competitors. Detailed observation of their feeding behaviour shows that generally seeds 

are destroyed, and as such they are considered primarily as pre-dispersal seed 

predators. As granivores their role in the ecosystem is rather negative but the impact of 

this damage seems to be limited due to high dietary overlap with seed dispersing 

frugivores. Among granivores, the different seed predators seem to have occupied 

separate trophical niches based on fruit and seed size and weight, feeding height and 

activity pattern. 

INTRODUCTION 

Worldwide, frugivorous animals play an important role in forest dynamics in terms of seed 

dispersal and forest regeneration. In Madagascar, most studies in this field have focused 

on the dispersal role of lemurs (Ralisoamalala 1996; Scharfe and Schlund 1996; Dew and 

Wright 1998; Overdorff and Strait 1998; Birkinshaw 1999; Ganzhorn et al. 1999a). At 

present, hardly any information is available on the potential role of the frugivorous bird 

species in seed dispersal or predation. The Malagasy frugivore avifauna is strikingly 

depauperate compared with continental areas and other large tropical islands (Fleming et 

al. 1987). In the littoral forest of Sainte Luce (south-east Madagascar), there are only six 

frugivorous bird species being Streptopelia picturata or the Malagasy Turtledove, Treron 

australis or Malagasy Green Pigeon, Alectroenas madagascariensis or Madagascar Blue 

Pigeon (Columbidae), Hypsipetes madagascariensis or Madagascar Bulbul 

(Pycnonotidae) and two species of Coracopsis (Psittacidae), C. nigra, the Lesser Vasa 

Parrot, and C. vasa, the Greater Vasa Parrot. In this study, we focus on one of the largest 

frugivorous bird species in the region; Coracopsis nigra for which little information is 

available on its feeding ecology (Hampe 1998; Dowsett 2000). Many Malagasy animal 

species today are under threat of extinction, and this may have a major impact on 

ecosystems. This is particularly relevant for the littoral forest, as only small forest 

remnants have survived due to human degradation of this habitat and its subsequent 

fragmentation. These forest fragments may not be large enough for certain animal 

101

Chapter 3b 

species to persist (Ganzhorn et al. 2000) and therefore, it is critical to identify effective 

seed dispersers in this ecosystem that might aid forest regeneration. Conversely we also 

want to gain insight into seed predation and the traits that determine food selection of the 

different granivorous species. 

METHODS 

Between November 1999 and February 2001, field data were collected by the first author 

on the feeding ecology of Coracopsis nigra in the 377-ha littoral forest fragment of Sainte 

Luce, called S9 (24º45'S 47º11'E), located in extreme southeast Madagascar. For a 

detailed description of the study site, I refer to Bollen and Van Elsacker (2002a, Chapter 

3a). Coracopsis species are naturally forest birds but also inhabit certain degraded areas 

(Dowsett 2000). Both C. nigra and C. vasa were observed at our site, mostly in single and 

occasionally in mixed species formations. However all feeding observations presented in 

this paper are from C. nigra, which is present year-round and in high densities. C. vasa 

was only observed on few occasions during austral summer (December 2000 - January 

2001) and seems to migrate into the area sporadically. 

Diets were assessed by direct feeding observations through tree watches (36hwatches) 

and more casual observations while walking along transects. Feeding and 

handling behaviour were described in detail in order to determine the role of C. nigra in 

seed dispersal and/or predation. More indirect methods such as faecal analyses and 

identifying fruit trap contents further contributed to the completion of the diet list (for 

details on methodology, see Bollen et al., Chapter 2). Parrots’ bill marks on the rejected 

fruit parts are easily recognizable and analyses of Coracopsis’ faecal samples enabled us 

to evaluate the condition of seeds after gut passage. 

All fruiting species encountered during the study site were characterised according to 

the following variables; fruit type (berry, drupe, capsule, other), pulp type (juicy, fibrous, 

none), seed protection (none, hard seed coat), seed number, fruit and seed length and 

weight. The latter were measured with callipers and a ‘Kernbalans NM60’ scale with a 

precision of 0.01mm and 0.01g respectively. Fruits were considered ripe when seeds 

were fully developed, often coinciding with a change in colour, odour or texture. 

Most of the variables measured have highly skewed distributions so the median value 

is given instead of the mean. For the same reason non-parametric statistics were used, 

such as Spearman rank correlation, Kruskal Wallis test and Contingency tables. 

Statistical significance was accepted for α≤0.05 for all tests. All statistical tests were 

carried out according to Siegel (1956) with the statistical software SAS for Windows. 

RESULTS 

C. nigra was recorded feeding on 40 plant species (36 genera, 25 families; Table 1), of 

which 39 are endemic and one, Psidium guajava, is exotic. Three plant species, 

Symphonia sp., Rhopalocarpus coriaceus and Eugenia sp. are exploited for their flowers 

only, while both the flowers and fruits of Polyscias sp. are eaten. All other plant species 

listed in Table 1 are visited for their fruits or seeds. Summarizing, 68% of the food 

species served as a seed source, 22% are eaten for seed and pulp and 10% for flowers 

(Table 1). 

The range of plant species consumed by C. nigra includes berries (31%) and drupes 

(47%) as well as capsules and other fruit types (22%). Most consumed fruits have a juicy 

102


fruit pulp (70%) but the parrots also consume fibrous (8%) as well as fruits without pulp 

(22%). The majority of fruits is one-seeded (57%), but several multi-seeded fruits are 

consumed as well. Fifty-four percent of all species have a hard kernel to protect the seed, 

while 46% has none. The median fruit length is 12mm (quartiles: 8-17mm) and median 

fruit weight is 0.6g (quartiles 0.2-1.0g). For seeds, median length is 7mm (4-11mm) and 

weight 0.1g (0.0-0.3g). Of all the fruit species consumed (N=37), 13 (35%) are taken both 

ripe and unripe, 13 (35%) in an unripe state only, and 11 (30%) only when ripe. 

Tina thouarsiana (N obs.=13; Sapindaceae) and Macaranga perrieri (N obs.=8; 

Euphorbiaceae) are important food items that are eaten from November through January, 

when overall fruit resources are abundant in the littoral forest (Table 1, Fig. 1). On the 

contrary, Bembicia uniflora (N obs.=28; Flacourtiaceae) fruits during the period of fruit 

scarcity and is probably the main food source within the diet of C. nigra. The parrots 

consume both ripe and unripe seeds of this plant species for at least six consecutive 

months. Another essential food source for numerous frugivorous species among which C. 

nigra is Syzigium sp.2 (N obs.=18; Myrtaceae), which also fruits when fruit availability is 

low. This plant species is very common and abundantly present in the littoral forest and 

may be considered a keystone species for numerous frugivores in this ecosystem. When 

comparing the phenology of C. nigra’s food items with the overall phenology of the forest, 

they are not at all correlated (rs=-0.22, P=0.48, N=13). Thus the patterns of fruit 

abundance and scarcity within the ecosystem are not reflected within the diet of the 

parrots. The monthly dietary diversity of C. nigra (median 18 species/month) seems to be 

quite stable throughout the year with a little less species eaten from June through 

October 2000. 

Our observations revealed that in general C. nigra is a very wasteful eater: it eats 

quickly and drops a considerable quantity of consumable food. Fruits are generally held 

by one foot and cut in half with the bill. Seeds are often only partially eaten, while pulp or 

fruit husks are discarded. Bill marks on discarded pulp show that only the seeds are 

eaten. In the few faecal samples collected (N=6), only parts of the seed coat could be 

recovered. The seed itself was digested entirely. 

Of all selected fruit items, the majority (N=33) is consumed by seed dispersers as 

well (Table 1). The four remaining species are actually not even zoochorous but rather 

autochorous fruits. Nearly half (N=17) of the food species are consumed by other seed 

predators as well such as Streptopelia picturata, Rattus rattus and Eliurus webbi (Table 

1). When comparing fruit and seed sizes within the diet of turtledoves, parrots and 

rodents, there is a significant size effect that separates these three groups of seed 

predators (Table 2). S. picturata selects significantly smaller and lighter fruits and seeds 

than C. nigra, which in turn eats fruits that are smaller than those consumed by the 

rodents. The number of seeds does not differ among these species. Apparently all 

granivorous species select mainly one-seeded fruits. Other traits determining food 

selection for granivores may be fruit type, pulp type and seed protection. There is no 

significant difference for any one of these traits however certain trends are clear (Table 

2). For example S. picturata eats more berries while C. nigra and the rodents clearly 

select more drupes, which then further explains the high proportion of fruits with seed 

protection in their diet. As far as pulp type, C. nigra eats several fruits with no pulp while 

this is rarely the case for the other seed predators. 

103

Chapter 3b 

Table 1. Diet list of C. nigra with indication of the plant part eaten (S: seed, P: pulp, fl: flower), 

the condition when eaten (UR: unripe, R: ripe), month when observed eaten (ME; 1: January, 

2: February , etc.) and months when available (MA) as food source. ‘Evid.’ refers to the method 

by which data were obtained (S: systematic tree watches, O: opportunistic observations, 

T: feeding marks, F: faecal droppings). Other consumer species are also indicated for each 

plant species, subdivided in seed dispersers and seed predators (L: lemurs, Pr: Pteropus rufus, 

Am: Alectroenas madagascariensis , Ta: Treron australis , Hm: Hypsipetes madagascariensis , 

Sp: Streptopelia picturata , R: rodents, Ew : Eliurus webbi , Rr: Rattus rattus ). The following 

fruit traits are given as well; fruit and pulp type, seed protection, seed number (NS), fruit and seed 

length (in mm) and weight (in g). 

Family name Scientific name 

Part Ripeness 

eaten 

ME MA Evid. 

Anacardiaceae Poupartia chapelieri S UR 4 11-4 SOT 

Annonaceae Polyalthia madagascariensis S R 4 11-5 SOT 

Apocynaceae Cabucala madagascariensis S UR 2 whole year OT 

Araliaceae Polyscias sp. fl,PS R 8 7-9 SO 

Schefflera rainaliana P,S R 5 4-5 O 

Arecaceae Dypsis prestoniana S UR & R 3 2-5 ST 

Clusiaceae Symphonia sp. fl 12 12-3 O 

Combretaceae Terminalia fatraea P,S UR & R 3 2-6 ST 

Connaraceae Agelea pentagyna S UR 1 12-1 OT 

Ericaceae Vaccinium emirnense P,S R 1 10-3 OT 

Erythroxylaceae Erythroxylum buxifolium P,S UR & R 5 2-5 OF 

Erythroxylum nitidulum P,S UR & R 8 8 O 

Euphorbiaceae Macaranga perrieri S R 11-1 11-2 S 

Suregada baronii S UR 5,7 4-11 O 

Uapaca littoralis S UR 3,8-11 whole year ST 

Flacourtiaceae Bembicia uniflora S UR & R 4-9 4-11 O 

Homalium louvelianum S UR & R 6,8 6-11 OT 

Ludia antanosarum S UR 1-2 12-3 SOT 

Icacinaceae Apodytes dimidiata S R 4 3-4 ST 

Loranthaceae Bakerella ambongoensis S UR 4 3-4 O 

Bakerella sp. S UR 12 12-3 O 

Myrtaceae Eugenia sp. fl 12 11-12 O 

Psidium guajava P,S R 1 1-2 O 

Syzygium sp.2 P,S R 4,6 4-7 SOT 

Ochnaceae Campylospermum obtusifolium P,S UR & R 2 11-8 O 

Oleaceae Jasminum kitchingii S R 6,9 3-6, 9-11 O 

Noronhia cf. lanceolata S UR & R 8 8-11 O 

Noronhia sp. S UR 7,10 4-1 OT 

Olea sp. S R 12 11-1 ST 

Rubiaceae Cantium variistipula S UR 8,12 4-9 OT 

Morinda cf. umbelluligera S UR & R 7 3-7 O 

Morinda rigida S UR & R 1 12-1 O 

Rutaceae Vepris elliotii S UR 4 whole year SO 

Vepris fitoravina S UR & R 2 2-3 ST 

Sapindaceae Tina thouarsiana S UR & R 11-1 6-3 SOTF 

Tinopsis conjugata S UR & R 12 10-3 OT 

Sapotaceae Sideroxylon beguei var. saboureaui S UR 11 11-3 OT 

Sarcolaenaceae Sarcolaena multiflora S UR 3-4 12-5 OT 

Sphaerosepalaceae Rhopalocarpus coriaceus fl 3-4 2-5 O 

Strelitziaceae Ravenala madagascariensis S R 4,8 4-9 O 

104



Other consumer species 

Dispersers Predators 

Fruit 

type 

Pulp 

type 

Seed 

protection 

NS 

Fruit Fruit Seed Seed 

length weight length weight 

L R drupe juicy hard coat 1 15.43 0.54 15.24 0.36 

L, Pr, Hm, Am R berry juicy hard coat 1 12.22 0.30 7.80 0.12 

L - drupe juicy hard coat 3 9.16 0.10 37.97 1.11 

L, Pr, Am, Ta, Hm Sp, R drupe juicy none 1 5.03 0.04 3.64 0.01 

L, Am, Hm - drupe juicy hard coat 2 4.73 0.08 2.85 0.01 

L, Pr, Am, Hm Sp berry juicy hard coat 1 14.85 0.59 12.92 0.34 

L, Pr, Am R drupe fibrous hard coat 1 13.19 0.37 8.12 0.13 

- - other no pulp none 1 16.89 1.05 12.58 0.32 

L, Pr, Ta, Hm - berry juicy hard coat 100 11.51 0.84 1.45 0.00 

L, Hm - drupe juicy hard coat 1 7.29 0.07 6.86 0.03 

L, Am R drupe juicy none 1 10.97 0.28 8.25 0.09 

L, Hm - drupe no pulp hard coat 1 4.55 - 3.06 0.03 

- Sp capsule no pulp none 4 9.47 0.40 3.95 0.04 

L, Pr R, Ew, Rr drupe juicy hard coat 3 23.63 4.86 15.03 0.52 

L - capsule no pulp none 1 5.34 0.01 - - 

- - capsule no pulp none 1 2.35 0.02 - - 

L, Pr - berry juicy none 6 12.47 1.04 3.22 2.98 

L, Am, Ta, Hm R drupe juicy hard coat 1 12.34 0.45 10.61 0.23 


L, Pr, Am, Hm - berry juicy none 1 15.86 0.66 10.35 0.25 

L - berry juicy none - 24.25 4.60 - - 

L, Pr, Am, Hm Sp berry juicy none 1 9.55 0.55 6.56 0.31 

Am, Ta Sp drupe juicy none 1 28.87 0.25 - - 

L Sp berry juicy none 1 6.49 0.20 4.07 0.07 

Hm - drupe juicy hard coat 3 7.06 0.21 3.39 0.02 

L R, Rr drupe fibrous hard coat 1 20.53 1.79 18.28 - 

L R, Rr, Ew drupe juicy hard coat 1 16.98 0.90 15.85 0.80 

L, Pr - drupe juicy hard coat 2 7.80 0.31 6.30 0.06 

L - berry juicy hard coat 20 11.94 1.75 5.31 0.01 

L - berry juicy hard coat 100 26.97 9.30 7.23 0.04 

L, Pr, Ta - drupe fibrous none 4 10.20 0.82 7.46 0.08 

L R drupe juicy hard coat 2 8.47 8.15 6.76 0.16 

L Sp, Rr, Ew other no pulp none 1 17.88 0.74 3.79 0.03 

L - other no pulp none 1 20.01 1.79 12.02 0.66 


L, Pr Sp capsule juicy hard coat 5 14.11 0.67 2.73 0.01 

- - capsule no pulp hard coat 6.93 0.25 

105

Chapter 3b 

N species fruiting 

Fig 1. An overview of the overall phenology at Sainte Luce is given for the period January 2000- 

January 2001. The number of species with ripe and unripe fruits is given per month. The number of 

C. nigra food species fruiting is also shown. 

DISCUSSION 

To date, little information is available on the diet of the Vasa Parrots. The work of Dowsett 

(2000), mostly on C. nigra, is the most extensive study of Coracopsis carried out at ten 

different Malagasy study sites, involving humid low-altitude and mid-altitude rainforests in 

the east, dry deciduous forest in the west and spiny forest in the south. Their feeding data 

include more than 30 species from 23 plant families, mainly endemic species. As in our 

study, most (60%) of the food species are exploited for seeds, 20% for flowers and 8% 

for pulp. The remainder are eaten for young leaves and plant buds. Their dietary list from 

humid eastern forest sites shows some taxonomic resemblance with ours at genus level 

(Uapaca, Psidium, Macaranga, Dypsis, and Poupartia) and at species level (preying on 

seeds of Ravenala madagascariensis). Similarity was found as well with the observations 

of Goodman et al. (1997b) of C. nigra feeding on Macaranga sp., Sarcolaena multiflora, 

buds of Symphonia sp., blue pericarps of Ravenala madagascariensis and the seeds of 

Ficus sp. and Uapaca sp. in the South-eastern Marosohy forest. Hampe (1998) observed 

C. nigra for one month in the western dry deciduous forest of Kirindy. Eight food species 

were recognized; four involved flowers, three fruits and seeds, and one young leaf 

shoots. Two more studies (Ratsirarson and Silander 1997; Böhning-Gaese et al. 1999) 

106 

60 

50 

40 

30 

20 

10 

0 

Phenology ripe fruits 

Phenology unripe fruits 

Coracopsis food items 

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan


Table 2. Morphological variables within the diet of Coracopsis nigra , Streptopelia 

picturata and the rodents (including E. webbi and Rattus rattus ) are shown. 

A comparison is made for the continuous variables through Kruskal Wallis test and 

for the class variables through contingency tables. 

C. nigra S. picturata rodents Kruskal Wallis 

median (N=37) median (N=13) median (N=50) H df P 

Number of seeds 1 1 1 0.13 2 0.93 

Fruit length 12.28 9.55 19.04 22.35 2

Chapter 3b 

behaviour of parrots worldwide, many authors agree on their role as seed predators 

(Janzen 1981; Jordano 1983; Clout 1989; Galetti 1993; Saini et al. 1994; Pizo et al. 1995; 

Corlett 1998; Renton 2001). C. nigra and possibly also C. vasa can certainly be 

considered as granivores in Sainte Luce rather than frugivores, as they are so often 

referred to in literature. As such they vary from other seed predators in many different 

aspects. C. nigra is a pre-dispersal seed predator, which feeds in the tree crown only 

during daylight on both ripe and unripe fruits, which are medium in size. It can further 

tolerate high tannin contents and eats non-zoochorous fruits as well without any fruit pulp 

at all. S. picturata is a post-dispersal seed predator which eats small seeds of juicy fruits 

on the ground only, either with a protective seed coat or not. The rodents both feed on 

the ground as in the tree at night, but most often carry seeds away to a feeding site. They 

eat the largest seeds of mostly drupes and may occasionally be involved in secondary 

seed dispersal as well. Nevertheless they can be mainly considered as post-dispersal 

seed predators for the majority of consumed fruit species. As shown in Figure 1 C. nigra 

has an almost constant year-round food availability as unripe fruits in particular stay long 

periods on the tree before ripening. On the other hand seeds stay longer periods 

available on the ground than in the tree, which is favourable for both the rodents and S. 

picturata. So, apparently there is a good niche separation among all seed predators in 

this ecosystem, which is mainly based on fruit and seed size, feeding height and activity 

pattern. 

In conclusion, C. nigra destroys seeds of numerous endemic plant species in Sainte Luce 

and should therefore be considered a seed predator in this ecosystem. In general, they 

split and thus destroy the seed with their bill, eat it and reject the surrounding pulp. Seed 

dispersal does occasionally occur when parrots fly away from the parent plant, dropping 

several intact fruits or the remains of soft and juicy berries. Zoochorous tree species have 

evolved phenological, morphological and biochemical flower, fruit and seed 

characteristics to attract efficient pollinators and seed dispersers, which can assure 

successful plant regeneration. However, the behaviour of granivorous animals offers the 

plant no apparent advantage. Nevertheless, most food species consumed by these 

parrots are also eaten by other frugivores, which act as seed dispersers. Dietary overlap 

among granivores is limited and the different seed predators seem to occupy separate 

trophical niches in this ecosystem. 


We would like to thank QMM (QIT Madagascar Minerals) for providing logistics and infrastructure at 

the study site. We are very grateful to Steve Goodman and Jörg Ganzhorn for their extensive 

revisions and useful suggestions. This study was carried out under a Collaboration Agreement 

between the Department of Animal Biology and the Department of Anthropology of the University of 

Antananarivo, the Institute of Zoology of Hamburg University and QMM. The first author was 

supported by a grant from the Belgian Fund for Scientific Research, Flanders (FWO). We thank the 

Flemish Government for its structural support for the Centre of Research and Conservation (CRC) 

of the Royal Zoological Society of Antwerp (RZSA). 

108


Drawing of fruits from Sainte Luce © An Bollen 2003 

’Isavasavana ny tahony, 

tiana hahitana ny fotony’ 

If someone explores the trunk of a tree, 

it means he wants to see the roots

Intersite comparison 

Fruit characteristics in a dry deciduous 

and a humid littoral forest of Madagascar: 

evidence for selection pressure through 

abiotic constraints rather than through 

co-evolution with lemurs 

as seed dispersers. 

AN BOLLEN, GIUSEPPE DONATI, JOANNA FIETZ, 

DOROTHEA SCHWAB, JEAN-BAPTISTE RAMANAMANJATO, 

LAURENT RANDRIHASIPARA, LINDA VAN ELSACKER , 

JÖRG GANZHORN 

FRUITS AND FRUGIVORES (L DEW, J-P BOUBLI) 

(IN PRESS) 

ABSTRACT 

Fruit and seed characteristics are compared between a dry deciduous forest in the west 

and a humid littoral forest in the south-east of Madagascar to discriminate between the 

role of abiotic factors (humidity, climate, soil characteristics) and frugivorous vertebrates 

for the evolution of morphological and biochemical fruit characteristics. The sites differed 

in abiotic conditions but contain very similar communities of frugivorous vertebrates. Fruit 

selection by two lemur species (Eulemur fulvus and Cheirogaleus medius) that are 

important for seed dispersal and that are present at both study sites, was compared 

between sites to examine fixed selection criteria that could give rise to possible 

co-evolution between frugivores and their fruit species on the one hand or to dietary 

flexibility of the frugivores on the other hand. Our results show that most fruit 

characteristics differ significantly between study sites. Food selection by both lemur 

genera in relation to morphological and biochemical fruit characteristics co-varies closely 

with their representation at a given site. These results indicate that morphological and 

biochemical characteristics are more likely the result of abiotic conditions rather than of 

interactions between frugivorous lemurs and their food. 

INTRODUCTION 

The study of interactions between fruits and their vertebrate consumers has generated a 

great deal of interest in recent decades, especially in tropical forests where most plant 

species depend on frugivorous animals for dispersal of their seeds (see Willson et al. 

1989 for a review). Attracting frugivores is crucial for these plants in order to ensure 

reproduction by seed dispersal (Howe and Smallwood 1982). Morphological fruit 

characteristics, such as colour, pulp richness, hardness of the shell, seed size, and 

patterns of spatio-temporal distribution have been interpreted as co-adapted features that 

govern animals' choice of fruit species. 

111

Chapter 4 

Most seed dispersal studies and reviews of correlations between frugivore food 

selection and fruit characteristics have produced little empirical support for tight 

co-evolutionary relationships (Howe and Smallwood 1982; Herrera 1984; Howe 1984; 

Gautier-Hion et al. 1985; Fisher and Chapman 1993; Chapman 1995; Erikkson and 

Ehrlen 1998; Lambert and Garber 1998), as most plant species do not depend on one 

single species of disperser. In most cases a range of taxonomically distinct frugivores 

may consume and disperse the seeds of the same fruiting species (Gautier-Hion et al. 

1985; Herrera 1986; Ganzhorn 1988; Chapman 1995; Chapman and Chapman 1996; 

Bollen et al., Chapter 2, 3). Fruit traits are likely to evolve in response to other selection 

pressures or may perform more than one function (Willson and Whelan 1990). Data from 

the fossil record suggest that morphological fruit traits often have remained relatively 

constant for millions of years (Fisher and Chapman 1993; Chapman 1995). 

Primates represent a major group of mammalian seed dispersers in the tropics. 

Studies have demonstrated that many primate species rely heavily on fruit and that they 

represent a large component of the frugivore biomass (25-40%, Terborgh 1983; Bourlière 

1985; Chapman 1995; Julliot 1996; Chapman and Onderdonk 1998; Lambert and Garber 

1998). In Madagascar, lemurs have been postulated to be important seed dispersers 

(Ralisoamalala 1996; Scharfe and Schlund 1996; Dew and Wright 1998; Overdorff and 

Strait 1998; Birkinshaw 1999, 2001; Ganzhorn et al. 1999a) in particular since the guild of 

frugivorous birds and bats is depauperate in this island as compared to other continents 

(Fleming et al. 1987; Wright and Martin 1995; Goodman and Ganzhorn 1997; Wright 

1997a; Böhning-Gaese et al. 1999; Ganzhorn et al. 1999a). 

In this study, we investigate whether morphological and biochemical fruit 

characteristics can be linked to abiotic conditions or whether there is evidence for 

co-evolution between these fruit characteristics and the main consumers that are involved 

in seed dispersal, i.e. Eulemur fulvus and Cheirogaleus medius. We selected two types of 

forest in Madagascar growing under very different climatic and edaphic conditions: 

evergreen littoral wet forest and dry deciduous forest. Both sites had a similar 

complement of frugivore species, having six genera and five species in common. 

The following predictions were tested: 

1. If fruit characteristics evolved mainly in response to abiotic conditions we expect 

different morphological and biochemical fruit characteristics at the two sites 

2. If fruit characteristics co-evolved in response to selective pressure of consumers 

we expect that characteristics of food items at both sites do not differ, as the guild 

of frugivorous vertebrates is very similar at both sites. 

3. The second prediction listed above requires that selection criteria of frugivores 

are species-specific. We therefore predict that these consumers will have a 

specialised diet irrespective of fruit availability, as is supposed by co-evolution. 

112

Kirindy 

Antananarivo 


Fig. 1. Location of the study sites. 

Madagascar 

Morondava 



METHODS 

Study Sites 

Data were collected at two sites: Sainte Luce (STL) and Kirindy/CFPF (KIR). 

In STL the study site is a 377-ha fragment of humid littoral forest located in southeastern 

Madagascar, 50km north of Fort-Dauphin/Tolagnaro at 24º45'S 47º11'E. Data 

collection was carried out by AB and GD between November 1999 and February 2001 

(Fig. 1, Donati 2002). From January 2000 through December 2000 annual rainfall was 

2,480 mm with four distinct seasons; hot-wet (December-February), hot-dry (March-May), 

cold-wet (June-August) and transitional-dry (September-November) (Donati 2002), but 

substantial inter-annual variation has been recorded (QIT Madagascar Minerals, unpubl. 

data). Mean temperature is about 23°C and ranges from 12°C to 33°C. The littoral forest 

of STL is characterised by a relatively open or non-continuous canopy, which is 6 to 12m 

in height with emergents of up to 20 m (Lewis Environmental Consultants 1992a). The 

diameter at breast height (DBH) of trees rarely exceeds 30-40cm (Rabevohitra et al. 

1996). Littoral forest grows on sandy soils and occurs within 2-3km of the coast at an 

altitude of 0-20m (Dumetz 1999). 

The forest of Kirindy/CFPF is a forestry concession of the Centre Formation 

Professionnelle Forestière de Morondava at 20°04’S 44°40’E, some 60km north of 

Morondava in west-Madagascar). It consists of 12,000ha of dry deciduous forest. Annual 

rainfall averages about 800 mm with a long distinct dry season from April to October 

when most trees lose their leaves. Most rain falls between December and February. 

113

Chapter 4 

Mean temperature is around 24.7°C and relative humidity varies between 58% and 67% 

with an average of 63% (Rakotonirina 1996; Sorg and Rohner 1996). The canopy 

reaches 10-12m in height. Trees with DBH50%) consists of fruits and/or seeds so 

they may be considered as possible seed dispersers. Two lemur species were studied in 

more detail for this study. These were Eulemur fulvus rufus in KIR and E. f. collaris in STL 

and Cheirogaleus medius at both sites. These species, particularly E. fulvus, are 

supposed to represent very important, if not essential, seed dispersers in Malagasy 

forests (Ganzhorn et al. 1999a). 

Phenology and pluviometry 

In STL a phenological transect with a total of 423 individual trees belonging to 95 species 

and 43 families was set up by AB and GD and monitored between January 2000 and 

January 2001. Trees sampled for phenology had a DBH>5cm and an effort was made to 

obtain five individuals per species whenever possible. Twice a month, presence or 

absence of young leaves, flowers, unripe and ripe fruits were recorded. A Tru-Check Rain 

Gauge was installed at the campsite in December 1999. It was checked and emptied 

each morning around 06h00. during the whole study period. For KIR rainfall and 

phenological data were taken from Sorg and Rohner (1996) involving 80 individual trees 

of 56 species (26 families) monitored over several years (1978-1987). For the present 

phenology analysis only large overstory tree species were considered. Small trees, 

shrubs, vines and epiphytes were left out in order to allow comparison between sites. A 

subsample of both phenologies (STL: 54 spp., KIR: 32 spp.) was extracted to include only 

those plant species that had been characterised morphologically (see below). 

Plant and fruit characteristics 

In STL and KIR fruits of 173 and 171 plant species belonging to 58 and 47 families 

Morphological characteristics 

Variables used to characterise fruits were: 

Growth form of parent plant: large tree, small tree and shrub, others 

(including herbs, vines and epiphytes); 

Fruit type: berry, drupe, capsule, pod, samara, synconia, others; 

Pulp type: juicy soft, juicy fibrous, dry fibrous, aril, no pulp; 

(in phenology ‘fleshy’ fruits are characterised as juicy soft, juicy fibrous or arillate; ‘nonfleshy’ 

fruits are dry fibrous or do not have any pulp); 

Colour: yellow-orange, red, purple, brown, green, others (black, grey and white), 

(multicoloured fruits were put in the category of the most conspicuous colour present); 

Odour: absent, present; 

114

Table 1. Frugivorous, granivorous and omnivorous vertebrate species possibly involved in seed dispersal in Sainte Luce 

and Kirindy. Diet: F: frugivorous, FG: frugi-granivorous; Ffo: frugi-folivorous; foF: foli-frugivorous; O:omnivorous; 

Activity: D: diurnal, N: nocturnal, C: cathemeral; Body mass (in g) and Body length (in cm) and their potential role as 

seed dispersers (D) or seed predators (P) in these ecosystems.ND: no data available. 

Impact 

seeds 

Body ab 

length 

Body a 

mass 

Activity 

Diet 

KIR 

STL 

English name 

Scientific name 

Family 

D 

D 

D 

P 

P 

32 

28 

24 

35 

50 

215 

ND 

45 

218 

ND 

D 

D 

D 

D 

D 

F 

F 

F 

G 

G 

x 

x 

x 

x 

x 

x 

x 

x 

x 



Madagascar Bulbull 



Treron australis 

Alectroenas madagascariensis 

Hypsipetes madagascariensis 


AVES 

Columbidae 

Columbidae 

Pycnonotidae 

Psitaccidae 

Coracopsis vasa 

D 

23-27 

500-750 

N 

F 

x 



MAMMALIA 

CHIROPTERA 

Pteropodidae 


PRIMATES 

Indridae Propithecus verreauxi 

Verreaux's Sifaka 

x FoF D 3000 40-47 D,P 

Lemuridae Eulemur fulvus collaris Collared Brown Lemur x 

F C 2000-2300 40-47 D 

Eulemur fulvus rufus 

Red-fronted Brown Lemur 

x FFo C 1600-2100 45 D 

Cheirogaleidae Microcebus murinus 

Grey Mouse Lemur 

x O N 60 12.5 D 

Microcebus berthae 

Berthe's Mouse Lemur 

x O N 30 12.5 D 

Microcebus rufus 

Brown Mouse Lemur 

x 

O N 42 12.5 D 

Cheirogaleus major 

Greater Dwarf Lemur x 

O N 443 25 D 

Cheirogaleus medius 

Fat-tailed Dwarf Lemur x x O N 119-280 20 D 

a 

Data from Langrand (1990), Goodman et al. (1997b), Fietz and Ganzhorn (1999), Garbutt (1999), Ganzhorn et al. (1999a), Donati (2002). 

b Body length is total length for birds and bats but head/body length for lemurs. 

115

Chapter 4 

Number of seeds: 1-2, 3-10, 11-50, 50+; 

Fruit weight: 50g; 

Fruit length: 30mm; 

Seed length: 20mm; 

Fruit protection: dehiscent, indehiscent with thin husk; indehiscent with thick husk; 

Seed protection: no protection; seed coat or lignified kernel; 

Dispersal type: zoochorous (exo- and endo-) or non-zoochorous including anemochorous, 

hydrochorous, autochorous. 

The characterisations were modified based on the original classifications by Gautier- 

Hion et al. (1985) and Lambert and Garber (1998). Epiphytes, vines, shrubs, large (>6m) 

and small trees (


characteristics were analysed with two-way analyses of variance. Data were arcsine 

transformed for these analyses. Statistical analyses were run according to Siegel (1956) 

with the help of SAS and SPSS software. 

RESULTS 

Climate and phenology 

Figure 2 shows phenology and annual precipitation for both study sites. Annual rainfall 

was 2,480mm in STL and 721mm in KIR during the study period. The phenological 

patterns considered in this study differ slightly from the overall pattern at both sites as 

published previously (Sorg and Rohner 1996; Donati 2002) because only a subset of the 

complete phenological dataset was used for comparison. 

In KIR ripe fruits are available year round with a minimum in April. ‘Fleshy’ and ‘nonfleshy’ 

fruit species are equally (50%) represented in KIR (Fig. 2). During the dry season 

(May through October) non-fleshy fruits predominate. As indicated before, in STL there 

are no clearly defined wet or dry seasons. Fruit abundance here is highest from January 

through March, rather limited from April through October with a lean period from June to 

August. The majority (81%) of fruit species in STL are characterised as ‘fleshy’. In 

contrast to KIR, the representation of the ‘non-fleshy’ fruits remains low but fairly constant 

(4-7%) in STL throughout the year. 

Soil conditions 

In the upper layer (A horizon) soils are more acid and contain higher concentrations of 

organic matter, nitrogen and phosphor in STL than in KIR (Table 3). Exchange capacity 

has not been measured for STL. The situation at STL is similar to the data available for 

Ranomafana, an evergreen rainforest site at higher altitude (Ganzhorn et al. 1999b). 

There, growth rate of trees is higher than at Kirindy, probably due to the longer growth 

season. However, the probability of fruiting is reduced, indicating that fruit production is 

associated with higher stress for the trees of the evergreen forest. It is unclear how these 

different constraints affect the type of fruits produced. 

Floristics 

Both datasets have 30 families (40%) and 19 genera (10%) in common but no tree 

species (Table 2). In STL the four most important plant families were Rubiaceae (23 

species), Euphorbiaceae (8), Flacourtiaceae (6), and Myrtaceae (6). They accounted for 

25% of all species. In KIR Fabaceae (16), Euphorbiaceae (14), Tiliaceae (9), Rubiaceae 

(8), and Combretaceae (6) were the five most important plant families. They accounted 

for 31% of the species. The representation of these top eight families is not correlated 

between the two datasets (rs=0.18; P=0.7, N=8). The representation of large and small 

trees, shrubs and other growth forms in the samples did not differ between sites (Table 

4). 

117

Chapter 4 

Fig. 2. Monthly fruit availability and rainfall in Sainte Luce and Kirindy. Fleshy fruits are fruits 

characterised as juicy soft, juicy fibrous or arillate, non-fleshy fruits are dry fibrous or do not have 

any pulp. 

118 

% of sp with ripe 

fruits 

% of sp with ripe fruits 

60 

50 

40 

30 

20 

10 

0 

60 

50 

40 

30 

20 

10 

0 

Phenology (N=54) 

rainfall (2480mm) 


J F M A M J J A S O N D 

Months 

Phenology (N=38) 

rainfall (721mm) 

Kirindy 

J F M A M J J A S O N D 

Months 

non fleshy fruits 

fleshy fruits 

rainfall 

600 

500 

400 

300 

200 

100 

0 

600 

500 

400 

300 

200 

100 

0 

Rainfall (mm) 

Rainfall (mm)


Table 2. Plant families, genera and species sampled in Sainte Luce and Kirindy. In bold the 

most important plant families per site are indicated. There are 30 plant families in common. 

L KIR TOT COM STL KIR TOT COM 

Families (N = 74) N N N N 

Families 

N N N N 

sp sp gen gen 

sp sp gen gen 

Anacardiaceae 3 4 4 1 Meliaceae 1 4 6 0 

Annonaceae 5 0 2 0 Menispermaceae 1 1 2 0 

Apocynaceae 1 3 4 0 Monimiaceae 3 0 1 0 

Araceae 1 0 1 0 Moraceae 5 3 3 1 

Araliaceae 2 0 1 0 Myricaceae 1 0 1 0 

Arecaceae 5 1 2 0 Myristicaceae 2 0 1 0 

Asclepiadaceae 0 1 1 0 Myrsinaceae 1 0 1 0 

Asteraceae 0 2 2 0 Myrtaceae 6 0 3 0 

Asteropeiaceae 1 0 1 0 Ochnaceae 1 1 2 0 

Bignoniaceae 3 4 5 1 Olacaceae 1 1 2 0 

Bombaceae 0 2 1 0 Oleaceae 5 3 4 1 

Boraginaceae 0 1 1 0 Pandanaceae 3 2 1 1 

Burseraceae 1 2 2 0 Passifloraceae 0 2 1 0 

Buxaceae 0 1 1 0 Pedaliaceae 0 1 1 0 

Canellaceae 1 0 1 0 Physenaceae 1 0 1 0 

Capparaceae 1 0 1 0 Pittosporaceae 2 0 1 0 

Celastraceae 2 1 3 0 Podocarpaceae 1 0 1 0 

Guttiferae 5 1 3 1 Ptaeroxylaceae 0 3 1 0 

Combretaceae 1 6 3 1 Rhamnaceae 0 2 2 0 

Connaraceae 1 1 1 1 Rubiaceae 23 8 20 2 

Dichapetallaceae 2 0 1 0 Rutaceae 3 2 2 1 

Ebenaceae 2 4 1 1 Sapindaceae 4 2 5 1 

Elaeocarpaceae 2 0 1 0 Sapotaceae 2 2 3 0 

Ericaceae 1 0 1 0 Sarcolaenaceae 4 2 4 0 

Erythroxylaceae 3 1 1 1 Saxifragaceae 2 0 1 0 

Euphorbiaceae 8 14 13 1 Scrophulariaceae 0 1 1 0 

Fabaceae 3 16 17 0 Simaroubaceae 0 1 1 0 

Flacourtiaceae 7 1 6 1 Solanaceae 0 1 1 0 

Hammamelidaceae 1 0 1 0 Sphaerosepalaceae 1 1 1 1 

Hernandiaceae 0 1 1 0 Sterculiaceae 0 3 3 0 

Hippocrateaceae 1 0 1 0 Strelitziaceae 1 0 1 0 

Icacinaceae 2 0 1 0 Tiliaceae 0 9 1 0 

Lauraceae 4 0 3 0 Ulmaceae 1 0 1 0 

Loranthaceae 3 0 1 0 Verbenaceae 1 3 2 1 

Lecythidaceae 1 1 2 0 Violaceae 1 0 1 0 

Liliaceae 4 1 4 0 UNKNOWN 15 37 0 

Loganiaceae 3 5 2 1 Sum (N species) 173 171 180 19 

Lythraceae 0 1 1 0 Maximum 23 16 20 2 

Melastomataceae 1 0 1 0 Minimum 0 0 1 0 

Average 3 1.81 2.46 - 

119

Chapter 4 

120 

Table 3. Soil characteristics of the littoral forest at Sainte Luce and the dry deciduous forest at Kirindy/CFPF (values are 

medians and range; N=4 for both sites). 

Depth 

C 

N 

Organic 

Horizon pH 

C/N P (ppm) K (ppm) 

(cm) 

(%) (%) 

matter 

Sainte 

4.6 11.3 0.42 30 18.1 2.0 22 

10 A 

Luce 4.1 – 5.9 7.2 – 25.5 0.21 – 0.84 23 – 41 12.4 – 43.9 1.0 – 3.0 17 - 49 

5.7 0.6 0.88 2.07 1.1 

0 

15 

80 B 

5.6 – 7.2 0.6 – 4.9 0.14 – 1.46 0.65 – 4.28 1.0 – 8.35 0 – 3.0 0 - 17 

1 

25 

12 – 35 

13 

6 – 31 

0.55 

0.2 – 0.9 

0.25 

0.2 – 0.3 

3.0 

0.9 – 3.7 

14 

13 - 21 

0.13 

0.03 – 0.15 

1.76 

0.53 – 2.15 

6.4 

6.0 – 6.5 

5.2 

A 

10 

Krindy/ 

CFPF 2 

4.9 – 6.6 

B 

80 

1 Data come from Razafimizanilala (1996). 

2 Data come from Felber (1984).


Comparison of fruit characteristics between sites 

Morphology 

Five out of eleven morphological parameters differed significantly between both study 

sites (Table 4). In STL, berries are the dominating type of fruit followed closely by drupes. 

In KIR drupes are most abundant followed by berries and capsules. Fruit pulp in STL is 

mostly soft and juicy. In KIR the majority of fruits has a rather dry and fibrous pulp. 

Remarkable is the large number of odoriferous fruits in STL while in KIR only one third of 

the fruits was classified as odoriferous. KIR has more dehiscent fruits and thick-husked 

indehiscent fruits than STL where 75% of fruits are indehiscent and thin-husked. 

Concerning dispersal type, zoochorous fruits prevail both in KIR and STL, but KIR has 

more non-zoochorous fruits than STL. However this difference was no longer significant 

after sequential Bonferroni adjustment. There is no significant difference between study 

sites for colour, number of seeds, fruit length, fruit mass, seed length and seed 

protection. 

Chemistry 

The chemical composition of mature fruits differed between sites in most chemical 

variables except for extractable proteins and sugars. Lipid concentrations were 

significantly higher in STL while NDF, ADF, total nitrogen, and procyanidin tannins were 

higher in KIR (Table 5). After rigorous adjustment for Type I errors (Rice 1989), there was 

only a significant difference for NDF, ADF and tannins. 

Comparison of diets of Eulemur and Cheirogaleus between sites 

Morphology 

Fruits eaten by both lemur species at both sites did not differ significantly in growth form, 

number of seeds, fruit length, fruit mass, seed length or seed protection (Table 4). 

However significant differences were found with respect to pulp type and the protection of 

fruits consumed by Eulemur and Cheirogaleus at both sites. The observed difference 

corresponds to the differential availability of fruits with different types of pulp and 

protection at both sites. 

Even though significantly fewer berries but more capsules were available in KIR than 

in STL (Table 4), this difference was not apparent when comparing diets of both lemurs 

between sites. Both species seemed to prefer berries and drupes even when these are 

less common and harder to find. In contrast, proportions of fruit colours did not differ 

significantly between samples eaten by C. medius, although E. fulvus did eat significantly 

more brown and green fruits in KIR and more yellow, orange and red fruits in STL. The 

proportion of odoriferous fruits eaten by both lemur species was higher in STL than in 

KIR, though the difference is not significant in the case of C. medius. KIR also had 

significantly more non-zoochorous fruits than STL, but still zoochorous fruits dominate the 

fruit diet of both lemur species at both sites. 

Chemistry 

Except for higher tannin concentrations in fruits consumed in KIR, none of the 

concentrations of the plant chemicals differed between fruits eaten by C. medius in STL 

and KIR (Table 6). Fruits consumed by E. fulvus contained higher concentrations of fibre 

and tannins in KIR than in STL. These results correspond with the biochemical 

differences in overall fruit availability between sites. Only the difference between fibre 

content remains significant after sequential Bonferroni. 

121

Chapter 4 

Table 4. Morphological characteristics of fruits collected in Sainte Luce and Kirindy and of 

fruits eaten by Cheirogaleus medius and Eulemur fulvus ssp. at the two sites. The X²-values 

were calculated for comparisons between sites; * P



Total Cheirogaleus Eulemur 

database diet diet 

STL KIR STL KIR STL KIR 

Number of seeds (N) 172 151 36 35 107 44 

1-2 97 72 22 16 62 17 

3-10 43 58 9 15 26 19 

11-50 12 9 5 4 

50+ 20 12 14 4 

10+ 5 4 

Unknown 1 20 0 1 0 2 

X² 

7.02; df=3 2.54; df=2 7.39; df=3 

Fruit weight (N) 161 146 35 33 100 39 

10g 15 10 11 3 

>1g 12 18 

Unknown 12 25 1 3 7 7 

X² 

2.38; df=2 2.83; df=1 1.78; df=2 

Fruit length (N) 167 159 36 36 105 45 

30mm 29 29 2 4 20 7 

Unknown 6 12 0 0 2 1 

X² 

3.63; df=2 2.19; df=2 0.27; df=2 

Seed length (N) 160 148 33 35 101 43 

20mm 18 8 11 3 

>10mm 11 11 

Unknown 13 23 3 1 6 3 

X² 

4.52; df=2 0.03; df=1 1.10; df=2 

Fruit protection (N) 173 151 36 35 104 45 

Dehiscent 26 39 0 3 10 5 

Indehiscent thin husk 130 80 83 26 

Indehiscent thick husk 17 32 11 14 

Indehiscent 36 30 

Unknown 0 20 0 1 3 1 

X² 

17.68***; df=2 9.27**; df=1 10.05**; df=2 

Seed protection (N) 158 147 35 34 101 45 

None 82 79 14 18 41 24 

Lignified kernel/seed 76 68 21 16 60 21 

Unknown 15 24 1 2 6 1 

X² 

0.10; df=1 1.16; df=1 2.05; df=1 

Dispersal mode (N) 163 150 35 35 103 45 

Zoochorous 130 104 34 33 93 41 

Non-zoochorous 33 46 1 2 10 4 

Unknown 10 20 1 1 4 1 

X² 

4.50*; df=1 1.16; df=1 2.05; df=1 

123

Chapter 4 

Table 5. Biochemical characteristics of ripe fruits at Sainte Luce and Kirindy. NDF: neutral 

detergent fiber, ADF: acid detergent fiber; Nitrogen: total nitrogen; Tannin: procyanidin tannin. 

Z-values are based on Mann-Whitney-U tests; * P


Table 6. Biochemical characteristics of food and non-food fruits of Eulemur fulvus ssp. and 

Cheirogaleus medius . For comparisons of fruit selection by C. medius only those fruits were 

considered that were present during the months when C. medius were active (i.e. not 

hibernating). Values are medians, quartiles, and sample size. Z-values are based on 

Mann-Whitney-U. * P

Chapter 4 

Chemistry 

At both sites, fruits consumed by C. medius contained higher concentrations of sugar 

than fruits not consumed (Table 6), but this was no longer significant after sequential 

Bonferroni adjustment. In KIR, fruits consumed by C. medius had lower fibre contents 

than fruits that had not been consumed. The only significant difference between non-food 

fruits and fruits consumed by E. fulvus consisted of lower fat concentrations in food 

species consumed at STL. Again this was not significant anymore after adjustment for 

Type I errors (Rice 1989). 

Interactions between lemur food selection and site effects on fruit chemistry 

In order to separate possible effects due to site characteristics from effects of lemur food 

selection on the chemical composition of fruits, two-way analyses were run using ’site’ 

and ‘lemur food’ as fixed independent factors. The results of these analyses are 

consistent with the conclusions above. Site-specific effects are significant for the majority 

of chemicals. According to the two-way ANOVA C. medius consistently searches for fruits 

with high sugar concentrations (Table 8). E. fulvus seems to avoid fruits with high fat 

contents in Sainte Luce only. The site effects persisted once the food items of the two 

lemur species were pooled and contrasted to the fruits that had not been eaten by neither 

species. There were several significant interactions between site and the food effects. 

Cheirogaleus medius avoids high fiber content but this clearly depends on the relative 

availability of fiber content at a certain site, while for Eulemur fulvus tannin concentrations 

in fruits eaten by E. fulvus vary differently at the two sites. Finally when both lemur 

species are pooled together, the lipid content of the consumed fruit species corresponds 

as well with the site specific availability. 

Table 7. The X² results of the comparison between morphological traits of lemur food 

species and the overall representation of these fruit traits within a site; * P


Table 8. Effects of site characteristics and whether or not an item was eaten by lemurs 

according to two-way analyses of variance. Analyses were performed on arcsine transformed 

data. Analyses were run separately for Cheirogaleus medius, Eulemur fulvus , and for fruits that 

had been eaten by either one or both species. 

Values are F-values; * P

Chapter 4 

considering species number (Table 2). Scharfe and Schlund (1996) also concluded from 

their study that in the western forests of Madagascar the majority of fruits are 

autochorous or dispersed by mammals while in the east dispersal by birds (that eat 

mainly berries and drupes) and mammals prevail. Our results concur with these. 

The site-related difference in the representation of fruits with an odour merits further 

consideration. In Malagasy forests, frugivorous diurnal and thus visually oriented bird 

species are poorly represented and most mammalian frugivores of Madagascar are 

cathemeral or nocturnal. Colour is probably less relevant for these lemurs and flying 

foxes while olfactory clues are likely to be important (Schilling 1979; Barton et al. 1995; 

Hladik and Simmen 1996; Bollen and Van Elsacker 2002a, Chapter 3a; Dominy et al. 

2002; Luft et al. pers. comm.). Since comparative data on fruit odour from other forests 

are lacking and taste and smell perception differ largely between individuals and species, 

the present results - which are based on subjective impressions of different human 

individuals - cannot be further interpreted. A more standardised evaluation of olfactory 

clues might be worthwhile in future research. 

With respect to our predictions we can say that given the almost identical set of 

frugivores present at both sites, these large differences in morphological and biochemical 

fruit traits between sites are most likely not a consequence of selection for seed dispersal 

by animals, as far as the particular lemur species compared. They rather represent the 

adaptations of a plant community responding to the need for protection against water loss 

during the long and harsh dry season, typical for dry deciduous forest in Madagascar. 

Comparison of lemur diets between sites and lemur food selection within a site 

Regarding feeding selection within a given site and comparison of diets between sites 

several patterns arise from the datasets. First of all, there are several parameters that 

seem less important for lemur food selection such as growth form, colour, fruit length, 

seed length, number of seeds, seed protection and extractable proteins. They did not 

differ at all between sites and did not influence lemurs’ feeding selection. On the contrary, 

clear feeding preferences were found according to fruit and dispersal type. Both lemurs 

selected almost exclusively zoochorous berries and drupes when fruits with abiotic 

dispersal were also available at both sites. Finally and most remarkably, both lemur 

species display a high dietary flexibility for certain parameters, both morphological (pulp 

type, odour, fruit skin protection) as biochemical (total nitrogen, tannins, ADF and NDF). 

For these parameters they would select food items in correspondence to what is most 

available at a given site. This seems to indicate that these species can switch their diet to 

what is available. This allows them to survive in different forest types on frugivorous diets 

with different nutrient compositions and different morphological traits. 

Overall, from a chemical perspective these lemur species did not show much 

evidence for fruit selection based on consistent chemical properties once site-specific 

characteristics were taken into account. In the present analyses E. fulvus avoid fruits with 

high lipid contents and fruits eaten by C. medius had lower fibre content than the nonfood 

items. These criteria persist even after site-specific effects have been accounted for 

(Table 8). Similarly, the preference of C. medius for fruits with high sugar content also 

persists at both sites. This has been linked to their need to accumulate fat reserves for 

hibernation (Fietz and Ganzhorn 1999). This selectivity however does not result in tight 

co-evolution, as a lot of less sugary fruits are present as well at both sites because other 

seed dispersers also occur and do not necessarily select sugary fruits. 

128


The results of the present study do not support the prediction that morphological and 

biochemical fruit and seed characteristics result from strong specific interactions and 

co-evolution with lemurs. Rather they could be the consequence of abiotic conditions and 

can best be interpreted as the result of an opportunistic and generalist zoochorous 

dispersal strategy of plants. Chapman (1995) has pointed out that weak selection 

pressure on fruit traits could result if primates have highly flexible diets and are not the 

only dispersers available in an ecosystem. Furthermore, large dietary differences 

between neighbouring primate groups or groups living a few hundred kilometres apart are 

not uncommon (Chapman 1995). This matches our findings of selection criteria of 

Eulemur and Cheirogaleus at STL and KIR, which are located 600km apart. Abiotic 

factors influence the phenology and taxonomy at a site and may then indirectly also lead 

to different morphological features and distinct biochemical compositions of food items 

available at each site. 

Considering the predictions outlined above we can summarize our results as follows: 

1. Since the frugivore communities are rather similar at the two sites, abiotic 

conditions rather than specific consumers are more likely to be responsible for 

the variety of morphological and biochemical features in fruits from different forest 

types. 

2. No evidence for co-evolution between these lemurs and fruit traits could be found 

as diets of the same lemur species differed substantially between sites. 

3. Within fleshy fruits, the lemur species considered did not show any persistent 

criteria for fruit selection in general besides few biochemical preferences but 

modified their diet according to fruit availability, even though mutual interactions 

and dependencies of fruits/seeds and their consumers exist. 


We thank L. Dew and J. P. Boubli for their invitation to participate at the symposium ‘Floristics, 

Phenology and Frugivore Communities: A Pantropical comparison’ at the ATB Conference in 

Panama and the ‘Commission Tripartite’ of the Malagasy Government, the Laboratoire de la 

Primatologie et des Vertébrés de E.E.S.S. d'Antananarivo, the Ministère pour la Production 

Animale et des Eaux et Forêts, Missouri Botanical Garden at Antananarivo, the CFPF at 

Morondava, QMM (QIT Madagascar Minerals) and the WWF Madagascar for their collaboration 

and permission to work in Madagascar. The studies were carried out under Accord de Collaboration 

between the Dept. Animal Biology and Dept. Anthropology of the University of Antananarivo, the 

University of Tübingen, the German Primate Center, the Institute of Zoology of Hamburg University, 

the CFPF and QMM. We are grateful to Nicoletta Baldi and Valentina Morelli for providing data on 

Eulemur feeding ecology for STL. B. Rakotosamimanana, C. Ravaoarinoromanga and E. 

Rakotovao provided important help at various stages of the study. I. Tomaschewski and U. 

Walbaum helped with plant analyses. Financial support came from the Belgian Fund for Scientific 

Research, Flanders (FWO) to AB, the University of Pisa to GD, the German Primate Center, DFG 

and DAAD to JF, DS and JUG. We thank the Flemish Government for structural support to the CRC 

of the RZSA. 

129

’Ny valala tsy indroa mandry 

eo am-bavahady’ 

Grasshoppers won’t show up 

twice at your doorstep 


Personal Malagasy stamp made by a young man on the streets of Antananarivo

The Malagasy littoral forest: 

threats and possible solutions 

AN BOLLEN, GIUSEPPE DONATI 

(TO BE PUBLISHED) 

Conservation 

ABSTRACT 

The littoral forest is expected to lose numerous endemic plant and animal species in the 

near future because of deforestation and resultant habitat changes. A great concern is 

the disruption of plant-animal interactions. It can be predicted that alterations in the 

recruitment dynamics of plant species in forest fragments might have unknown 

consequences for their long-term survival. This paper discusses the characteristics of 

animal seed dispersal relevant to the regeneration of the littoral forest. Possible 

management implications are discussed in relation to the existing initiatives. Urgent 

protection of the largest remaining forest fragments (S9, S17), including a flying fox roost 

site (S6) is of great importance. Furthermore it is necessary to install corridors to connect 

the isolated fragments and create plantations to fulfil the need for wood of the local 

people. 

INTRODUCTION 

Worldwide, Madagascar is considered a high priority for global biodiversity protection due 

to high faunal and floral endemism and biodiversity (Mittermeier et al. 1998). The flora of 

Madagascar is one of the richest in the world in comparison to its area (Dumetz 1999). 

About 96% of all plant species present are endemic (Schatz 2001). The avifauna is 

relatively species poor compared to other tropical islands, but the level of endemism is 

also extremely high (52% Langrand 1990). For mammals, reptiles and amphibians 

endemism even raises up to 90% (Garbutt 1999), 95% (Ramanamanjato 2000) and even 

99% (Ramanamanjato 2000) respectively. At the same time Madagascar is one of the 

tropical regions where the effects of deforestation are most worrying (Green and 

Sussman 1990). Ever since humans first reached the island some 2000 years ago, the 

native forests have provided them with animals for food, land for cultivation and wood for 

construction and fuel. Humans have thus dramatically changed the island vegetation. 

According to Kull (2000) dense endemic rainforests cover only 10% of the island, while 

total forest cover is about 23%. Deforestation is proceeding most rapidly in the east, 

where 66% of the original rainforest has been logged or irreversibly converted to land for 

cultivation (Dumetz 1999; Kull 2000). At these rates it is predicted that in the year 2025 

rainforest will only remain on the steepest slopes, in remote areas and nature reserves 

(Kull 2000). Madagascar receives global attention as a hot spot of biological diversity, 

133

Chapter 5 

environmental degradation and conservation action (Mittermeier et al. 1998; Kull 2000), 

but at the moment only 3% of the island has a protected status (Godfrey et al. 1997). 

The littoral forest of south-eastern Madagascar is one of the ecosystems most 

threatened on this island and is reduced to its vestiges with only 2500ha remaining today, 

representing at most 10% of the original forest (Ganzhorn et al. 2001; Vincelette pers. 

comm.). During the last 50 years more than 5000ha have disappeared (MIR 

Télédétection Inc 1998). This forest type has been severely degraded due to charcoal 

production, logging, bushfires and shifting cultivation (slash and burn). At the moment it is 

represented by severely degraded forest fragments, ranging in size from 3 to 377ha. In 

the Fort-Dauphin area the existing littoral forests on sandy soils are Petriky, Mandena 

and Sainte Luce (Fig. 1). Sainte Luce has the highest species diversity and can be 

considered among the most intact littoral forest remaining in eastern Madagascar 

(Dumetz 1999; Rabevohitra et al. 1996). A botanical study by Razafimizanilala (1996) 

and Rabevohitra et al. (1996) in the forest fragment ‘S9’ of Sainte Luce shows that 98% 

of the 189 plant species monitored are endemic for Madagascar and there are 29 

endemic plant species for the south-eastern littoral forest (Lewis Environmental 

Consultants 1992a). 

In order to highlight the importance of the littoral forest before it completely vanishes, 

we present here our current understandings of frugivore-fruit interactions. We further 

discuss the causes of habitat loss and forest fragmentation and imply our knowledge on 

seed dispersal for landscape restoration. Finally suggestions concerning conservation 

applications as well as priorities for future research are made. 

STUDY SITE 

In this paper we focus in particular on the situation of Sainte Luce, where first and second 

author conducted two PhD researches from September 1999 till February 2001. The 

campsite is located in the 377 ha-large forest fragment, called ‘S9’ (24º45'S 47º11'E). 

Littoral forest lies within 5 km of the coast and occurs on sandy soils at an elevation of 0 

up to 20m. It is characterised by a relatively open or non-continuous canopy, which is 6 to 

8m in height and diameter at breast height (DBH) of trees rarely exceeds 30 to 40cm 

(Rabevohitra et al. 1996; Dumetz 1999). This forest type is characterised by an average 

annual rainfall of about 2,690 mm, with a marked rainy season from November through 

February. No clear dry season could be detected (Bollen and Donati, Chapter 1). Mean 

monthly temperature is 23°C (QMM unpubl. data). Fruit production is seasonal, with a 

peak in abundance of ripe fruits in December and January and with periods of fruit 

scarcity that differ strongly inter-annually (Bollen and Donati, Chapter 1) 

THREATS 

Three villages, Ambandrika, Ampanasatomboky and Manafiafy lie within close range of 

the largest forest fragments of Sainte Luce; S6 (225 ha), S7 (206 ha), S8 (190 ha), S9 

(377 ha) and S17 (244 ha)(Fig. 1), which make up half of the remaining south-eastern 

littoral forest. Some 700 villagers depend on the forest for crop growing, timber and nontimber 

forest products for subsistence and commercial activities. The main causes 

responsible for degradation and fragmentation in Sainte Luce are clearance by slash and 

burn, useless bushfires and unsustainable harvesting of logs. Threats in the near future 

further include charcoal production by the Antandroy people, a southern tribe, and 

134

Conservation 

Fig. 1. On the left is a detail of the south-eastern Malagasy region around Fort-Dauphin shown. In 

black the remaining littoral forests are indicated, while the humid montane rainforest (including 

Andohahela National Park) is in grey. On the right we zoom in on the largest forest fragments of the 

littoral forest of Sainte Luce and indicate the Antanosy villages in this area and our study site. 

ilmenite extraction from the sand by a large Malagasy mining company, QIT Madagascar 

Minerals (QMM) owned by QIT-Fer et Titane, a subsidiary of Rio Tinto Zinc. 

Shifting cultivation through ‘tavys’ (slash and burn) contributes directly to the actual 

dramatic levels of deforestation, in particular for forests on sandy soils. Farmers cut a 

patch along the forest border, allow the slash to dry and then burn it in preparation for 

cultivation of maize, manioc or rice. A new patch is created and cultivated every 1-2 

years, since extensive rain quickly depletes the bare sandy soil from minerals. This 

causes a retreat of forest edges and leads to soil sterilisation. After the first clearing, 

grasslands dominated by invasive heath shrubs Erica sp. (Ericaceae) directly replace the 

natural forest (Lewis Environmental Consultants 1992a). As such the soils become even 

more acid and less suitable for regeneration of endemic plant species. Farmers rarely 

monitor their fires allowing them to expand and run their course. As a consequence 

uncontrolled bushfires further result in an alarming progress of deforestation while not 

serving any purpose. Even though the people burn throughout the whole year, these 

activities are more concentrated in periods when winds are stronger (September-October, 

February-March). Whereas in the dry deciduous forest of West Madagascar bushfires 

occur naturally, this is less common in the eastern forests (Kull 2000). 

There is both subsistence and commercial logging, but on a local scale only. People 

cut mainly large and mature trees to get fuel, construction and tool wood. Logging here 

involves mainly unsustainable and wasteful resource use so that many cut logs are 

partially left behind in the forest. The selective removal of timber species changes their 

distribution patterns impeding regeneration. Although the logging pressure was 

considered low up to now in Sainte Luce, some plant species used for fuel were observed 

135

Chapter 5 

at a very low density in apparent pristine area (Donati pers.obs.). Discussions with local 

people and observations of many cut trunks lead us to the conclusion that some plant 

species, like Cinnamosma madagascariensis, are already rare due to selective logging. 

At the end of our study period (December 2000) logging intensified and trucks 

transporting the logs could be observed at night on the national road heading to Fort- 

Dauphin. This short-sighted commercial logging is even more harmful and a serious 

threat to all forests in the surrounding area of Fort-Dauphin as needs for construction 

wood get larger in the city. The only wood people are allowed to take out of the forest by 

law is ‘dead wood’. As such, people have been very inventive in interpreting this law. 

They cut several trees in the forest, leave the logs there for a month to afterwards collect 

this ‘dead’ wood (Donati and Bollen pers. obs.). Nowadays they even circumcise the bark 

deep enough so eventually trees will die (Hapke pers. comm.). 

The villagers take several other products out of the forest. Large palms (Dypsis 

prestoniana) are cut at the base only to collect the nervature of their leaves of which traps 

for crabs and langoustines are made. The enormous logs are often left behind unused. 

Canoes are dug out of the largest trees, such as ‘ramy’ Canarium boivinii and ‘vitano’ 

Calophyllum sp. (Table 1) and different vine species serve for fishing gear. People also 

eat the fruits of several endemic tree species, but no real large-scale harvesting occurs. 

During famine, people collect and eat the roots of tavolo (Tacca leontopetaloides) and via 

(Typhonodorum lendleyanum)(Table 1). 

Besides the indirect threat of habitat loss, hunting also has its impact on littoral forest 

dynamics. However in contrast to sites more inland, such as Andohahela NP, Mandena 

and Petriky, the impact of hunting in Sainte Luce is rather limited. Fish rather than bush 

meat make up the largest proportion of animal protein in the daily menu. The majority of 

villagers are fishermen and catching fish (year-round) and shellfish (langoustines only 

April-December) represents their most important income. Bush meat is only eaten on rare 

occasions or during traditional events. Nocturnal lemurs, such as Cheirogaleus major, 

Cheirogaleus medius and the tiny Microcebus rufus, that hibernate during austral winter 

in hollow tree trunks are an easy catch. The larger Eulemur fulvus collaris is hunted by a 

traditional technique called tandroho (Randriamanalina et al. 2000). More specifically a 

strip of forest (50 m²) is cleared, so that canopies are too far apart for the lemurs to cross. 

Long logs are then placed between both ends as the only crossover with two snares in 

the middle. This hunting technique could be at the base of the unbalanced sex-ratio in 

favour of males reported in S9 (Donati unpubl. data), as in this species females often 

have progression priority and as such they are more likely to become victims of these 

traps. E. f. collaris was also captured by the use of row slings, a practice particularly 

frequent in young mans and probably responsible for the cryptic behaviour of these 

lemurs at the beginning of our study. Hunting pressure on E. f. collaris declined drastically 

in the study area (S9) during the presence of researchers based on several discussions 

with the local people. Nevertheless traps were still encountered in September 2000 in 

S17 (Fig. 1). Pteropus rufus is very vulnerable to hunting as it roosts communally and can 

thus be located and killed easily in great numbers. Catapults, long branches and stones 

were often found under their roost site. The animals were harassed during daylight and 

even changed their roost site twice in 2000 (Bollen and Van Elsacker 2002a, Chapter 3a). 

Terrestrial birds are trapped using snares or other ingenuous systems on the ground. 

Fruit pigeons (Treron australis, Alectroenas madagascariensis), parrots (Coracopsis vasa 

and C. nigra) and bulbuls (Hypsipetes madagascariensis) are hunted with arrows or 

catapults. Alternatively, fruit pigeons get stuck onto small branches covered with sticky 

136

Conservation 

latex, which are placed in Ficus or other fruiting trees by adolescent men. Tenrecs are 

hunted with the aid of dogs at dusk. Hunting occurs mainly for food purposes but 

occasionally the animals are traded regionally as well for pets and this in particular for 

Eulemur fulvus collaris and Coracopsis spp. (Bollen pers.obs.). 

Charcoal production is an important activity of the people living in the extreme 

southern Androy. At the moment, charcoal pits are regularly found in Mandena and 

Petriky but not yet in Sainte Luce. It is a very destructive practice. A strip of forest is 

cleared, branches are stapled in a hole in the ground, covered by vegetation and then 

burned for several days. Charcoal is used by numerous people as fuel and represents a 

major income for the Antandroy. It can be expected that in the near future these activities 

will move up north towards Sainte Luce as the forests of Mandena and Petriky get more 

depleted. 

Furthermore QMM represents another threat as mining activities will destroy a large 

part (76%) of the littoral forest (Lewis Environmental Consultants 1992a). In Sainte Luce, 

74% of the remaining forest will vanish (QMM 2001). In November 2001, QMM got 

permission of the Malagasy government to start the mining. For the coming six year all 

infrastructure is being prepared to begin extraction of ilmenite in 2009 in Mandena. 

Petriky and Sainte Luce will be mined respectively 20 and 35 years later. Mining activities 

will last 45 up to 60 years (Vincelette pers. comm.). Populations of numerous plant 

species risk disappearing together with the fauna (Lewis Environmental Consultants 

1992a). Anthropogenic pressure will then further intensify on the few remaining littoral 

forest fragments outside the QMM concession zone. 

As everywhere, population growth is associated with limited resources. Even though 

the local people are responsible for most of the degradation and fragmentation of the 

littoral forest at the moment, there is definitely also willingness from their part to protect 

the forest. The President of Sainte Luce in 2000-2001, Olaf Abel Isaia has done this by 

initiating a traditional dina. These are local agreements for protection of natural resources 

(LOI N°96-025, 1996). At the moment, QMM and delegates from the three villages are 

discussing on a new version of the dina for Sainte Luce, which will probably be finalised 

later this year (Vincelette pers. comm.). 

IMPACT ON ECOSYSTEM AND SEED DISPERSERS 

Plant-frugivore interactions are important components of complex forest communities. 

Vertebrate seed dispersal is a key process in the dynamics of natural vegetation and 

vegetation recovery (Wallace and Painter 2002). Furthermore, frugivores play a vital role 

in the maintenance of biodiversity in tropical forests, where they constitute a large 

proportion of the vertebrate biomass (Fleming et al. 1987). In the same way zoochorous 

tree species make up the bulk of the tropical flora (Howe and Smallwood 1982). Loss of 

fruit-frugivore interactions may thus have profound consequences for conservation 

(Corlett 1998). 

Many animal species in Sainte Luce rely on fruit as an essential food resource and 

conversely provide valuable dispersal services to many of these fruit bearing plants 

(Bollen et al., Chapter 3)(Table 1). As forests become more fragmented the remaining 

patches become increasingly isolated and less accessible for arboreal lemur species. As 

a consequence gene flow and seed dispersal between patches becomes more critical for 

long term survival of many plant species. On the other hand, if fragments get too small or 

hunting increases, the long-term survival of animal species will not be guaranteed either 

137

Chapter 5 

Table 1. A list of plant species from Sainte Luce, Indicating if they are endemic (E), abundant (A), 

common (C) and typical littoral forest (LITT) plant species. The utility of these plant species is 

given as well; medicinal use (M), firewood (F), construction wood (C) or unknown use (X). 

'Special' indicates plant species that are consumed by numerous frugivores (number is given), 

LS means that these are large-seeded plant species for which Eulemur fulvus collaris is the 

only seed disperser and 'key?' refers to a potential keystone species during periods of fruit 

scarcity. 'Food' species for the people are indicated as well, with indication of plant part eaten. 

Family Species Vernacular name Status 1 Utilitair 1 Special Food 

Anacardiaceae Poupartia chapelieri sisikandrongo A C 7 

Protorhus cf. lecomtei kangy LS 

Annonaceae Monanthotaxis cf. malacophylla vahimbotany C 

Polyalthia madagascariensis fotsivavo 8 

Polyalthia capuronii menapeka LS 

Apocynaceae Cabucala madagascariensis tandrokosy FRUIT 

Arecaceae Typhonodorum lindleyanum via ROOT 

Araliaceae Schefflera rainaliana voatsilana sp1 FM 

Polyscias sp. voatsilana sp2 7 

Areceae Dypsis fibrosa boakandambo LITT LS 

Dypsis nodifera raotry LITT 

Dypsis prestoniana boakabe LITT 9-key? 

Dypsis saintelucei telopolombilany LITT RARE 

Dypsis scottiana raosy LITT 6 

Bignoniaceae Ophiocolea delphinensis akondronala EC X FRUIT 

Phyllarthron ilicifolium zahambe E C 

Phyllarthron sp. zahambe manongaroa E C 

Burseraceae Canarium boivinii ramy MC LS 

Canellaceae Cinnamosma madagascariensis vahabatra 3eM LS-RARE 

Capparaceae Crataeva obovata belataka C LS 

Clusiaceae Psorospermum revolutum harongampanihy MF FRUIT 

Calophyllum sp. vitano C 

Garcina chapelieri haziny tomate LS 

Garcinia cf/aff. madagascariensis disaky kely LS 

Combretaceae Terminalia fatraea katrafa C 

Dichapetalaceae Dichapetalum sp. vahihazo LS 

Ebenaceae Diospyros sp.1 hazomainty blanc F LS 

Diospyros sp.2 hazomainy LITT F LS 

Elaeocarpaceae Elaeocarpus alnifolius* sanga LS 

Ericaceae Vaccinium emirnense tsilantria CF 9 FRUIT 

Erythroxylaceae Erythroxylum buxifolium fangora sp.1 F 

Erythroxylum nitidilum fangora sp.2 F 

Euphorbiaceae Uapaca ferruginea voapaky lahy CF 6 

Uapaca littoralis voapaky vavy CF 9 

Uapaca thouarsii voapaky lahy ZJ LITT CF 

Fabaceae Cynometra cf. cloiselii mampay A C 

Phylloxylon xylophylloides sotro E C 

Intsia bijuga harandrato C 

Flacourtiaceae Aphloia theiformis fandramana C 

Bembicia uniflora bemalemy A CF 

Homalium louvelianum ramirisa CF 

Scolopia orientalis zoramena C F 7 

Grossulariaceae Brexia sp. kambatrikambatri C 

Hippocrateaceae Salacia madagascariensis voatsimatra C LS FRUIT 

138

Conservation 


Family Species Vernacular name Status 1 Utilitair 1 Special Food 

Icacinaceae Apodytes dimidiata hazomamy marécage E 9 

Apodytes sp. nov. hazomamy an ala E X LS 

Lauraceae Cryptocarya sp. tavolohazo X 

Ravensara acuminata ? M 

Liliaceae Dracaena reflexa var. nervosa falinandro LITT-C FM 

Dracaena reflexa var. nervosa tavolobotroka LITT-C FM 

Loganiaceae Anthocleista longifolia lendemilahy C C 

Loranthaceae Bakerella sp. velomihanto 6 

Melastomataceae Tristemma mauritianum voatrotoky FRUIT 

Meliaceae Malleastrum mandenense sarigoavy EC 

Monimiaceae Tambourissa castri-delphinii amborabe CFM 

Tambourissa purpurea ambora CFM 7 

Moraceae Ficus guatteriifolia fihamy M 

Ficus pyrifolia nonoka 7 

Myricaceae Myrica spathulata tsilaka M 

Myristicaceae Brochoneura acuminata mafotra sp.1 C CM LS 

Brochoneura madagascariensis mafotra sp.2 C CM LS 

Myrsinaceae Embelia incumbens taratasy M 

Myrtaceae Eugenia cloiselii ropasy sp.1 EC CFM 

Eugenia sp. ropasy sp.2 CFM LS 

Syzygium sp.1 rotry ala C 7 FRUIT 

Syzygium sp.2 rotry mena C 10-key? FRUIT 

Oleaceae Jasminum kitchingii vahifotsy kely C 

Noronhia cf. lanceolata hazondraotry M 

Noronhia sp.1 belavenoka M 

Olea sp. vahabatra A M 7 

Pandanaceae Pandanus dauphinensis vakoanala A C LS 

Pandanus aff. longistylus fandranabo C LS 

Pandanus rollotii fandranabotonboky C LS 

Podocarpaceae Podocarpus madagascariensis harambilo CFM 

Sphaerosepalaceae Rhopalocarpus coriaceus tsilavimbinanto LS 

Rubiaceae Canthium sp. Rubiaceae ZJ CFM 

Canthium variistipula fantsikaitramainty C CFM 6 FRUIT 

Ixora sp. x203 C X 

Plectronia densiflora fantsikaitra M 

Psychotria sp.1 tanatananala F 

Rothmannia mandenensis taholagna X FRUIT 

Tricalysia cf. cryptocalyx hazongalala F 

Rutaceae Vepris eliotii lahinampoly EC CFM 6 

Sapindaceae Macphersonia radlkoferi sanirambaza X 

Plagioscyphus jumelei ambirimarika pionair X 

Sarcolaenaceae Leptolaena multiflora fotombavy C CF 

Sarcolaena multiflora meramaintso LITT-C CF 9 FRUIT 

Schizolaena elongata fotondahy C 

Sphaerosepalaceae Rhopalocarpus coriaceus tsilavimbinanto C X 

Strelitziaceae Ravenala madagascariensis ravenala CM 

Taccaceae Tacca leontopetaloides tavolo ROOT 

Theaceae Asteropeia multiflora fanolafotsy CFM 

Verbenaceae Vitex chrysomallum nofotrako LITT C 

1 data come from Dumetz (1999), Razafimizanilala (1996), Rabevohitra et al.(1996), Koechlin (1974), 

Lewis Environmental Consultants (1992a), QIT Madagascar Minerals (2001). 

139

Chapter 5 

(Ganzhorn et al. 1999a). Madagascar in general, already has a depauperate avian 

frugivore community (Langrand 1990; Goodman et al. 1997a) and lacks larger frugivores 

such as ruminants, ungulates and elephants. Moreover one third of lemur species 

became extinct years ago (Godfrey et al. 1997). Thus emphasizing even more the 

importance of the remaining seed dispersers here. Large frugivores, such as E. f. collaris, 

are often most vulnerable to habitat fragmentation which is conform findings at other sites 

(Johns and Skorupa 1987; Kannan and James 1999). The vulnerability of E. f. collaris is 

reflected by the fact that this species is only present in the two largest and most intact 

fragments, S9 and S17, but absent in all other, smaller fragments in Sainte Luce 

(Ganzhorn et al. 2000). E. f. collaris is in particular important for seed dispersal of 

numerous plant species and it is the only frugivore present here that is able to swallow 

and thus disperse larger seeds (up to 16.5mm diameter, Table 1). Therefore, local 

extinction of this species could ultimately lead to the lack of regeneration of different 

dependant plant species. Other specialist frugivores, such as Pteropus rufus, Treron 

australis and Alectroenas madagascariensis, are very vulnerable as well, in particular 

when important food sources are selectively logged (Table 1). Frugivorous birds and 

flying foxes are the most important mobile seed dispersers bringing seeds into grasslands 

and early succession vegetation. The simple structure of these habitats pose less of a 

barrier to them than it does for arboreal lemurs. Genetic exchange and long distance 

dispersal between and among fragments are less likely to occur, if populations of mobile 

flying dispersers decrease or vanish. 

Our studies showed that no evidence for co-evolution could be found, nor were there 

strong indications for syndromes that attract taxonomic groups by certain morphological 

and/or nutritional traits (Bollen et al., Chapter 2, 3, 4; Bollen et al. in press). Instead we 

found that there is great dietary overlap among frugivore species and that dispersal is 

achieved through redundant systems. Most frugivores seem to eat according to what is 

available, given the limitation of fruit and seed size and certain feeding preferences. 

Animal seed dispersers are vital for the regeneration of littoral forests. Especially since 

losses of ecologically interdependent species will be permanent to an even larger extent 

here as forest fragments are very isolated. Due to co-dependency of frugivores and tree 

species, the hunt on frugivores leads to suboptimal or insufficient dispersal and 

recruitment success of certain plant species (Chapman and Onderdonk 1998; McConkey 

and Drake 2002) while deforestation, forest degradation and fragmentation has a drastic 

impact on the food chain of the frugivorous fauna. 

PROTECTION OF THE LITTORAL FOREST 

Forest 

The remaining intact forest fragments, which act as reservoirs from where indigenous 

floral and faunal species can colonize new habitats are in urgent need for protection. 

Even though the littoral forest only represents 5% of all forest cover in the south-eastern 

region (QIT Madagascar Minerals 2001), the importance of its conservation cannot be 

emphasized enough. At present, Sainte Luce contains the most intact littoral forest 

fragments on sand, which differ strongly in floristic diversity from the one on laterite and 

from the inland montane forest (Rabevohitra et al. 1996; Dumetz 1999). Ideally 

conservation zones should be large enough to encompass the micro-heterogeneity in 

order to avoid their effective elimination. In this respect, the forest fragments S9 and S17 

have the highest conservation priority, representing the two largest and most intact forest 

140

Conservation 

fragments, including both inland and coastal littoral forest. S9 has been subject to 

numerous research projects since 1989 (Vincelette pers. comm.). This fragment contains 

some recent tavy in its northern and north-eastern section (Bollen and Donati pers. obs.). 

Logging has increased during the last years but ‘intact’ primary forest with important 

lemur populations is still present. S17 is a long and narrow forest fragment located on a 

dune system very close to the coast. The southern part is very degraded (Bollen and 

Donati pers. obs.) but the northern part can still be considered very pristine, which is 

probably due to its remote position as it is separated from the villages by a lake and an 

estuarine system (Fig. 1). The extreme northern part (ca. 60ha) is owned by Mr. De 

Heaulme and theoretically supervised by local guards. Both fragments, S9 and S17 are 

very different in appearance and floral composition and include much of the biodiversity 

present in the littoral forest. The conservation zones proposed by QMM involve 190ha of 

S9 and almost complete S17. This will be enforced once the dina gets accepted by all 

involved parties. Additionally we believe S6 is of extreme importance due to the presence 

of the roost site of a colony of flying foxes, which are irreplaceable long distance seed 

dispersers for numerous plant species. The forest fragment itself is extremely degraded 

and victim to numerous unregulated forestry activities. All these fragments together 

comprise about 1000ha of littoral forest. They lie within close proximity to one another 

and comprise inland and coastal littoral forest, mangroves, dunes and marshes, sandy 

beaches, a lake and an estuary providing important refuges for a wide variety of plant and 

animal species (Fig. 1). 

Several studies in Madagascar have demonstrated that it is better to protect a few 

large fragments as opposed to several small ones. Ramanamanjato (2000) found that a 

series of small littoral forest fragments does not provide the biodiversity of reptiles and 

amphibians found in one or a few large fragments. The species number declines 

substantially in fragments smaller than 200-300ha. As for birds, Raherilalao (2001) 

showed that the number of bird species also decreases proportionally with the size of 

forests blocks in Ranomafana. In addition, Ganzhorn et al. (2000) found that the number 

of lemur species present in a fragment is related to its overall size. At the moment, Sainte 

Luce still provides an important refuge for E. f. collaris (S9 density 0.38ind/ha, Banks 

2002), which occurs only in the south-eastern region of Madagascar. Protection of the 

remaining primary forest is of crucial importance in general because Ganzhorn and 

Schmid (1998) found that even 40 year old secondary dry forests in western Madagascar 

are unlikely to provide a suitable habitat even for the smallest seemingly least threatened 

of all lemur species. As for protection measure all logging, hunting and slash and burn 

practices should be banned from these fragments, so that ecosystems can recover. In 

theory, the regional division of the Ministry of Water and Forests should control the 

presence of these activities in the forest but there is a shortage of staff, finances and 

means and the remoteness of Sainte Luce further contributes to the lack of an accurate 

control system. The final management plan should clearly indicate what land could still be 

used for these traditional activities allowing buffer zones close to the main population 

centres. Ideally slash and burn practices should be replaced by more sustainable land 

use practices in the zones adjacent to protected areas. Involvement of local people in 

decision-making for conservation action plans is indispensable. Resource management 

should be urgently improved and control systems should at least partially come from 

within the villages. 

141

Chapter 5 

Animals 

The degree of vulnerability of a given species due to forest fragmentation is likely to be 

related to its tolerance to habitat change and its capability to use or bridge the grasslands 

around the remaining forest fragments. As for frugivores, the large lemur species E. f. 

collaris is most vulnerable as it is reluctant to cross these grasslands. Due to the spatiotemporal 

patchiness of its food resources, it needs large home ranges (up to 100ha) and 

covers long distances daily (1500-3500m, Donati 2002). E. f. collaris only occurs in the 

south-eastern region of Madagascar (Tattersall 1982) and risks to decline severely in 

numbers as a consequence of fragmentation and degradation of the littoral forest and 

through hunting practice. Eulemur fulvus collaris, or Eulemur collaris as debated by some 

(Djletati et al. 1997; Wright 1999; Wyner et al. 1999; but see also Pastorini et al. 2000) is 

listed as a vulnerable taxon by IUCN (Hilton-Taylor 2000) and needs to be urgently 

protected. With respect to this species, there has already been a translocation on several 

groups of E. f. collaris in 1999 in Mandena. Their habitat (fragment M3) was almost 

completely destroyed by producers of charcoal. Therefore the lemurs were captured and 

transferred to an actively protected forest fragment (M15-M16, Mandena Conservation 

Zone). Despite the initial loss of some individuals, the groups seem to have adapted to 

their new habitat (Ramanamanjato pers. comm.). Translocation and re-introduction of 

primates, especially in ‘rescue’ situations such as the one in Mandena will become 

increasingly important as a tool for species conservation (Soorae and Baker 2002). 

Moreover, this practice has a high popular significance and a very high potential for 

educational applications. The nocturnal lemurs are still more abundantly present in 

smaller forest fragments (Ganzhorn et al. 2000) but are restricted to a single fragment as 

well. Another highly vulnerable species is Pteropus rufus, because a population of 250- 

300 flying foxes has its roost in a severely degraded fragment. Therefore, this roost site 

should be included in an actively protected zone. Most frugivorous bird species seem to 

be less threatened and more abundantly present in small, large, intact and degraded 

fragments. 

As frugivores face periods of fruit scarcity it is important to collect long-term data on 

phenology to understand inter-annual patterns and predict periods of fruit abundance and 

scarcity. Interesting as well is to identify keystone species that bear fruit during periods of 

fruit scarcity and supply much of the diet of the frugivores in this forest (Terborgh 1986b). 

Given the short duration of our studies (Bollen and Donati, Chapter 1) we are unable to 

asses true ‘keystone species’ (definition according Terborgh 1986b; Mills et al. 1993) at 

the moment. However, a potential candidate may be Syzigium sp. 2 (Myrtaceae) and to a 

lesser extent Dypsis prestoniana (Arecaceae). Both species fruit when fruit availability is 

low. Syzigium sp. 2 is a large canopy tree species that is very common in the littoral 

forest (Razafimizanilala 1996) with numerous odoriferous purple berries characterised by 

a soft and juicy pulp and thin husk. These fruits are one-seeded with high sugar 

concentrations (43%). This species constituted 80% of the diet of E. f. collaris in June 

2000. Dypsis prestoniana is much less abundant as it used to be but can still be find in 

the more intact parts of S9. This high palm species has a relatively large fruit crop 

considering its small canopy. Within the diet of E. f. collaris fruits of this species 

constituted 20% in April (Table 1). The overall importance of these potential keystone 

species increases if we consider that they are eaten by all frugivorous species present in 

Sainte Luce. Phenology data on important timber species could further be a useful tool to 

more ecologically sustainable initiatives as well (Wallace and Painter 2002). As they can 

142

Conservation 

contribute disproportionately to the diets of certain species, detailed data on food supply 

are indispensable (Chapman and Peres 2001). 

Corridors and plantations 

As degradation and fragmentation are quite advanced in the littoral forest, active 

protection of the remaining intact forest and control of hunting, logging and fires is not 

enough to conserve and restore this ecosystem. Furthermore, natural regeneration via 

secondary forests is too slow to counteract the loss of primary forests. Therefore, it is 

necessary to accelerate the natural recovery process. In this respect, creation of corridors 

that connect isolated primary forest remnants with thin strips of habitat is considered a 

prime target for conservation activities (Ganzhorn et al. 1997; Beier and Noss 1998). In 

this respect, QIT Madagascar Minerals (2001) has installed a corridor in 1999 in between 

M4 and M5 in Mandena with 20% endemic species and 80% exotic species. Corridors 

are valuable conservation tools, promoting increased plant and animal movement among 

patches that will enhance population viability and likelihood of recolonisation, as well as 

facilitation of pollination and seed dispersal (Beier and Noss 1998; Tewksbury et al. 2002) 

As some species readily move between fragments, using habitat corridors, others do not 

(Chapman and Peres 2001). Reasonably, seed of forest plant species will be dispersed 

at greater distances from their source through continuous forest than through open field 

or pasture, so corridors should ideally be contiguous with the native forest seed source 

(Wunderle 1997). 

In Sainte Luce, S9, S17, S6 and S7 are not too far apart and could ideally be linked 

by corridors in the near future so that reforestation can happen from these nuclei. Birds 

and flying foxes are of great importance during first succession stages dispersing pioneer 

and heliophil species. After three years the ground is effectively shaded by their canopy 

that climax species will then predominate in seedling growth, provided the seed source is 

brought into these parcels by flying foxes, fruit pigeons, bulbuls and in a later phase even 

mouse lemurs. As the corridor gets more ample dwarf lemurs and eventually E. f. collaris 

will make use of them as well, dispersing certain seeds. If we depend entirely on natural 

seed dispersal to bring tree species to a site, this may result in a secondary forest 

dominated by a well-dispersed subset of the forest flora. Unassisted succession has 

proven to be better at restoring biomass than biodiversity (Corlett 2002). In this respect 

large-seeded plant species (for example Canarium boivinii, Diospyros sp., Apodytes sp. 

nov.)(Table 1) are less easily dispersed than small-seeded trees and because they have 

fewer dispersers, they require planting in subsequent efforts (Terborgh 1983; Janzen 

1988; Wunderle 1997; Kitamura et al. 2002. Ingle 2003). 

Besides creating corridors, plantations of both native and exotic species can help 

regeneration as well while at the same time providing an alternative wood source for the 

local people. In this respect QIT Madagascar Minerals has carried out experiments in a 

tree nursery for 10 years with different exotic tree species such as Eucalyptus, Acacia 

and Casuarina species (QIT Madagascar Minerals 2001), which were found to be 

suitable in landscape restoration. They even accelerate natural forest succession by 

ameliorating harsh soil and understory microclimatic conditions, suppressing dominant 

grasses, improving soil fertility and nutrient availability and attracting seed dispersers 

(Wunderle 1997; Holl et al. 2000; Corlett 2002). In the near area of Mandena there are 

about 1930ha of plantations, either private, state property or installed by QMM. QMM has 

installed 200ha of plantations in Mandena and 2ha in Sainte Luce. So far there are no 

plantations in Petriky. QMM intends to grow more plantations every year (Vincelette pers. 

143

Chapter 5 

comm.). Mixed species plantations in Ampijoroa have proven to be suitable and to offer 

fairly acceptable habitats for the majority of the extant lemur species. In Mandena 

Cheirogaleus medius and Microcebus murinus were observed to feed on the flowers of 

the exotic Melaleuca quinquenervia (Myrtaceae, QIT Madagascar Minerals 2001). In 

other eastern sites even the larger lemurs, like Eulemur sp., were seen to relate on 

Eucalyptus flowers (Ganzhorn 1985; Overdorff 1988). However the floristic diversity of 

plantations is limited and may thus not provide food year round for all lemur species 

(Ganzhorn 1987; Ganzhorn and Abraham 1991). Therefore plantations, just as corridors, 

should ideally border natural forests. This way economic interest can be combined with 

lemur conservation by planting economically and ecologically valuable trees on presently 

deforested areas (Ganzhorn and Abraham 1991). Nevertheless one should be careful in 

selecting appropriate species for this and the corridors or plantation should not be 

monospecific nor involve dominant invasive species. 

Planting important fruiting trees in corridors, plantations or clearings also improves 

the process of reforestation by attracting frugivores and the seed dispersal they provide. 

Fruit trees can thus be used to accelerate seed dispersal and enrich biodiversity. As 

zoochory is the predominant dispersal mode in the tropics (Howe and Smallwood 1982), 

dietary data on frugivores can influence the choice of fruit species included in planting 

projects. Otherwise, perches or trees in clearings may further attract bird and flying foxes, 

which then again bring certain seeds into these open areas (Holl et al. 2000). In this way 

frugivores can be used to facilitate regeneration, reforestation and vegetation succession 

of tropical forest. 

Development aid 

Together with active protection measures for the littoral forest ecosystem, there is great 

need for an integrated approach combining research, conservation and developmental 

aid in the area. Madagascar is one of the poorest countries in the world. More than 75% 

of its human population lives below the poverty level. Conservation of local fauna and 

flora is a luxury for the Malagasy who struggle to survive from day to day. There is an 

urgent need for alternative resources necessary to provide the people with fuel, 

construction and tool wood. For this, plantations of fast-growing non-invasive species are 

necessary. At the same time conservation management plans and long term integrity of 

protected areas depend critically upon support from rural communities. Cooperation with 

local people, as promoted by the first phase of the National Environment Action Plan 

(1991-1995), is an essential feature to succeed in conservation. 

There are several possibilities to provide economical benefits to the Malagasy. First 

of all, ecotourism may serve as a conservative strategy for extending economic 

opportunities while safeguarding this unique natural environment. The area of Sainte 

Luce provides several options for ecotourism, as there are impressive coastal and inland 

forests, beautiful pristine white sandy beaches, several nearby small islands, gorgeous 

bays and traditional villages. Several groups of E. f. collaris are habituated and can thus 

easily be observed as well as numerous bird, amphibian and reptile species. The main 

problem for ecotourism is the difficult accessibility of Sainte Luce because of the poor 

road system and rough sea conditions during most of the year. Sharing revenues from 

forest entrance may be one way to fully distribute benefits from tourism to residents. 

Furthermore local guides can be trained and organisations could support local economy. 

Mainly small-scale tourism would be advantageous here, so it does not have a negative 

impact on the ecosystem nor lead to increased habitat disturbance or human immigration. 

144

Conservation 

Ferraro (2001) has noticed that benefits from tourism in Ranomafana are only seasonal 

and have been captured by a relatively small subset of the population, which are often 

migrants. This should be avoided by all means. Butterfly farming and beekeeping were 

proposed as other economical opportunities by QMM and the latter was set up by QMM 

in Mandena and Sainte Luce. This initiative was welcomed with great enthusiasm by the 

local people (Vincelette and Ramanamanjato pers. comm.). Different possibilities can be 

thought of as long as they offer benefits for both the local people and the littoral forest. 

Exclusively protected areas may have a negative impact on the livelihood of residents 

that cultivate the land and/or collect and sell fuel and construction wood to the fishermen. 

They need to be compensated by recruitment as tourist guides, forest guards, workers in 

the plantations and so on. Compared to sites more inland, this is only a small percentage 

of the population in Sainte Luce. Another possibility to diminish the locals’ impact on the 

littoral forest in general is to offer alternatives for their fishing gear, such as more 

synthetic materials (nylon or plastics). By the same token, gas or Eucalyptus charcoal 

may be a better alternative for charcoal and indigenous firewood. Overall, it is very 

important that ecologically, economically and socially sustainable solutions to the 

conservation of biodiversity are searched for as well as the wise management of natural 

resources. 

We further believe that environmental education in the villages is of great importance. 

The first and second author had several meetings with the local authorities and gave 

presentations in the local school regarding research activities in the forest. We explained 

the purpose of our studies and the relevance of conserving this ecosystem. These 

meetings always attracted lots of people and should definitely be continued in the future 

to include all parties in the process of decision-making and to improve social acceptance 

of conservation plans. 

Future studies 

Even though dynamics of the littoral forest ecosystem are slowly revealed, much more 

data are necessary. First of all, a community wide study on pollination is needed, as no 

seed setting can occur without it. Secondly there is an urgent need for long-term data on 

phenology in order to predict periods of fruit scarcity and to define important food species, 

which can be consequently planted in corridors and/or plantations to facilitate and 

enhance germination. Thirdly future studies should focus on patterns of seed shadows 

and post-dispersal survival of seeds and seedlings across the full range of habitats 

including both intact as well as degraded landscapes. Managers need to be able to 

predict which species will survive in forest fragments in order to identify which ones are 

potentially most threatened by deforestation and thus should be given priority for planting 

(Chapman and Peres 2001). Furthermore, once we get insight on community wide plantanimal 

interactions, it is necessary to estimate minimum viable population sizes of 

animals and their tolerance to forest fragmentation, translocation or re-introduction, which 

will become a necessary tool in the littoral forest due to the future mining plans. Finally a 

follow up is needed to assess whether they make use of plantations and corridors. Most 

studies in the past have been carried out in primary and protected forest. This approach 

may not serve the interests of conservation as we need to be able to evaluate 

conservation action plans and to reformulate these based on new findings. Future 

research should be carried out by animal ecologists, plant population biologists and forest 

managers collaborating closely together in order to find a better balance between timber 

harvesting, biodiversity conservation and sustainable management. Even though more 

145

Chapter 5 

data are needed we believe this study along with others from the littoral forest provide a 

conclusive and solid data base to aid in formulating conservation plans now and thus 

urgent and active protection should no longer be delayed any further in time. 

CONCLUSION 

The littoral forest is being cleared at an alarming rate while our understanding of its 

ecology is still incomplete. Multi-disciplinary approach is required in the near future to 

unravel further interactions and to refine conservation action plans. Forest restoration is a 

necessary activity but simultaneously efforts should be focused on conserving the 

integrity of existing primary forests, as there is no guarantee that all species will recolonize 

after disturbance. At the moment, less than 3% of Madagascar has a protected 

status (Wright 1997b) and one third of the lemur species has already gone extinct 

(Godfrey et al. 1997). Up to date, the littoral forest is not represented in protected areas. 

This is an important and unique forest type due to its high endemism and biodiversity. 

Understanding forest dynamics and unravelling plant-animal interactions in degraded 

tropical ecosystems is essential if we are to conserve existing forests and accelerate the 

process of reforestation. It is necessary to urgently join forces in order to prevent that our 

and several other studies were a last testimony of what once was the littoral forest. We 

owe it to the future generations to protect this ecosystem, so they can take pleasure in 

exploring it further and so the endemic plant and animal species can persist and thrive 

here. 


We like to thank Manon Vincelette and Jean-Baptiste Ramanamanjato of the QMM Environmental 

team for providing detailed information on the environmental policy in the littoral forest and on the 

current activities of QMM. Many thanks as well to Jörg Ganzhorn, Linda Van Elsacker and Silvana 

M. Borgognini Tarli for their support in this research. The first author is supported by a grant from 

the Belgian Fund for Scientific Research, Flanders (FWO). We thank the Flemish Government for 

structural support to the CRC of the RZSA. The second author was supported by a doctoral grant of 

the Italian Ministry for Scientific Research (MURST) and the University of Pisa. 

146

GENERAL CONCLUSION 

General conclusion 

The main purpose of this study was to gain insight into overall fruit-frugivore interactions 

in the littoral forest of Sainte Luce. It represents the first community-wide approach to 

primary seed dispersal in a Malagasy forest type from the perspective of both the tree 

and the consumer species. 

Three hypotheses concerning evidence of co-evolution between life history traits of 

plants, their diaspores and animal consumers were tested by studying the frugivorous 

vertebrates and the dispersal strategies of 34 tree species. Phenological, morphological 

and biochemical fruit traits from these species were measured to look for co-variation with 

their seed dispersers. No evidence was found for species-specific co-evolution in this 

study. The lack of tight co-evolutionary relationships was suggested as well by many 

other studies (Gautier-Hion et al. 1985; Herrera 1986; Fisher and Chapman 1993; 

Erikson and Ehrlen 1998). Five large-seeded tree species however did seem to depend 

critically on the largest lemur, Eulemur fulvus collaris, for seed dispersal and recruitment. 

This strong dependence however does not represent a case of co-evolution in the strict 

sense but can be interpreted as an indirect consequence of the extinction of the larger 

frugivorous birds and lemurs, which would also have been capable of dispersing these 

large fruits. Nevertheless, in terms of conservation these fruit-frugivore interactions are of 

crucial importance to conserve the integrity of the littoral forest. 

The low-high investment model (McKey 1975) subdivides tree species into 

specialists and generalists but again my results do no support this model. McKey’s model 

was originally developed for bird-dispersed trees in the Neotropics and its validity seems 

to depend largely on the site-specific composition of the frugivore guild (see Wheelwright 

et al. 1984). It seems that the species-poor guild of frugivores in Madagascar did not lead 

to specialised dispersal strategies. Most tree species seem to be characterised by rather 

‘generalist’ fruit traits allowing them to attract as many seed dispersers as possible. This 

way the risk of relying on only one frugivore species is avoided, which may be dangerous 

in an ecosystem with few frugivores. Furthermore, the low species diversity of avian 

frugivores resulted in significantly few bird fruits compared to other sites. 

Of all three hypotheses, the concept of dispersal syndromes (Van der Pijl 1969) 

was supported most clearly, as there were indeed indications that certain morphological 

traits correspond to taxonomic groups of dispersers. I found that diaspores dispersed by 

birds, mammals or both groups (called mixed fruits) differ in their fruit and seed size, fruit 

shape and seed number, but not in biochemical composition. These results agree with 

studies on fruit syndromes in other tropical regions with distinct assemblages of plants 

and animals (Janson 1983; Knight and Siegfriend 1983; Gautier-Hion et al. 1985, Corlett 

1996; Pizo 2002). Nevertheless, dispersal syndromes can only partly clarify the variability 

displayed in tree dispersal strategies at Sainte Luce, while the influence of abiotic factors 

on fruit traits may serve as an additional explanation along with the phylogenetic heritage 

of plant taxa. In this respect, Hampe (2003) mentions the importance of temperature, 

water availability and day length in relation to seed growth, fruit size and sugar or lipid 

content respectively. As such, there may be similar meaningful correlations that I 

overlooked. Overall, efficient plant-disperser interactions do exist in Sainte Luce as well 

147


as in other sites (Gautier-Hion et al. 1985; Dowsett-Lemaire 1988; Herrera 1995, Corlett 

2002; Pizo 2002) without requiring the close co-variation of fruit traits with their dispersers 

as predicted by the tested models. 

To evaluate the relative contribution of the different animal species in relation to seed 

dispersal and predation, the core part of this study concentrated on plant-animal 

interactions on a species-specific as well as on a community level. These mutual 

relationships determine the dynamics of the littoral forest ecosystem. Fruit and seed size 

appear to be the most determining physical traits in food selection of all consumer 

species. While birds show a strong preference for specifically coloured fruits (red, purple 

and black), nocturnal lemurs show clear preference for soft and juicy fruit pulp and thinhusked 

fruits irrespective of colour. Arillate or soft and juicy fruits are also favoured by 

flying foxes. Nutritionally, birds prefer lipid-rich fruits whereas certain mammals (Eulemur 

fulvus collaris, Pteropus rufus) avoid those, which may be linked to their differential 

capacity to digest and assimilate lipids. Mouse and dwarf lemurs select fruits with high 

sugar content. This allows them to prepare and store fat reserves before going into torpor 

(see also Bonnaire and Simmen 1994, Fietz and Ganzhorn 1999). These findings are the 

only indications for diet selection of all frugivores, while for the vast majority of fruit traits, 

both biochemical and morphological, the frugivores consume whatever is available. This 

weak selection pressure represents another reason for the lack of strong mutual 

relationships among fruit traits and dispersers. 

Compared to other study sites worldwide (Fleming 1979; Gautier-Hion et al. 1985; 

Kitamura et al. 2002) dietary overlap among frugivores seems to be rather high in Sainte 

Luce. However, without considering the proportional use of different food items this 

overlap may be overestimated. Dietary overlap among frugivores may be strongly 

influenced by phenology, increasing when fruit is abundant and decreasing when it is 

scarce (Overdorff 1993; Johnson 2002). Phenological data from Sainte Luce show that 

fruiting is highly seasonal and that lean periods differ substantially inter-annually. The 

overall low fruit productivity and high unpredictability of food resources in Sainte Luce 

and other Malagasy forests (Overdorff 1996; Goodman and Ganzhorn 1997; Wright 

1999) may be the at the base of low feeding selection pressure and thus relatively high 

dietary overlap. This also corresponds with the theory of Fleming (1979) that high spatiotemporal 

patchiness in the Paleotropics leads to much higher dietary overlap and at the 

same time to the co-existence of fewer frugivore species as opposed to the 

Neotropics. In this respect, Terborgh (1986) says that periods of fruit scarcity are crucial 

to set the carrying capacity of tropical forests for their frugivore community. During 

bottlenecks frugivores have to switch to other food items (such as young leaves, flowers 

or insects) to compensate for the lack of fruit and to avoid inter-specific competition. In 

Sainte Luce and other Malagasy humid forests phenophases are highly inter-correlated in 

time, which means that alternative diet items are not available either during lean periods. 

As a result many lemur species remain highly frugivorous even when fruits are scarce but 

concentrate on a few important food species (Overdorff 1993; Vasey 2000; Donati 2002; 

Johnson 2002). Some of these plant species may be potential keystone species and 

Ficus species are known to often play this role in the tropics for numerous tropical 

frugivores (Terborgh 1986a; Johnson 2002 but see Gautier-Hion and Michaloud 1989). 

However Goodman and Ganzhorn (1997) mention that Madagascar in general has a 

reduced Ficus diversity. This is also the case for Sainte Luce, where other plant species 

are likely to fulfil this role (Syzigium sp.2, Dypsis prestoniana). 

148


Even though dietary overlap is high and most fruit species are eaten and dispersed 

by several frugivores, the different animal species clearly have a distinct impact on seed 

dispersal. As such, they do not appear to be ecologically redundant in their role within the 

ecosystem. While flying frugivores (fruit pigeons, bulbuls, flying foxes) disperse seeds 

into the clearings and ensure genetic exchange between forest fragments, E. f. collaris is 

the only disperser of large-seeded fruit species. Mouse and dwarf lemurs disperse only 

small-seeded fruit species during austral summer. On the contrary the granivores 

(rodents, turtledoves, parrots) prey on seeds of most fruits they eat. Clearly, 

heterogeneous seed transport is particularly important for a severely fragmented 

ecosystem such as the littoral forest. 

The community-wide and mainly descriptive approach of this study only allowed me 

to unravel general trends in food ecology and determine the particular importance of the 

different species within the ecosystem. This study lacks more detailed quantitative data 

on the animal side of the interactions. Due to this shortcoming niche differentiation of the 

sympatric frugivores has certainly been overlooked. Frugivore species occupy a speciesspecific 

multidimensional niche within the ecosystem, which obviously has its influence as 

well on food selection. Due to competition and particular life history traits not all fruits may 

be ‘truly’ available to all frugivores as was set forward by this study. When understanding 

niche differentiation, it may be easier to detect which other traits may indeed be more 

relevant in determining actual diet choice. 

To check whether previous results are confirmed at other Malagasy forest types, a 

comparison on fruit availability and the feeding ecology of two lemurs species was 

conducted within two completely different forests: the dry deciduous forest in Kirindy 

(west Madagascar) and the humid littoral forest in Sainte Luce (south-eastern 

Madagascar). Both sites differ substantially in abiotic conditions and have a distinct plant 

species composition and phenology. However, the frugivore guild at both sites is 

comparable. As for fruit traits, Sainte Luce has more fleshy zoochorous berries with thin 

husks while dehiscent capsules and indehiscent thick-husked drupes are more abundant 

in Kirindy. Biochemically, lipid concentrations are higher in Sainte Luce whereas fibre, 

tannin and nitrogen contents are higher in Kirindy. Most of the dominant fruit traits in the 

dry deciduous forest represent adaptations against water loss during the long and harsh 

dry season. This stresses once more the importance abiotic factors may have on fruit 

traits. Other studies comparing fruit traits on a larger geographical scale have found a 

potential influence of abiotic factors and phylogeny of plant communities (Hampe 2003, 

Voigt et al. submitted). When comparing feeding selection of two seed dispersers, 

Eulemur fulvus and Cheirogaleus medius within and between sites, there are three 

different trend that can be found. Firstly, there are clear food preferences. Zoochorous 

berries and drupes are strongly preferred, even though other dispersal and fruit types are 

present at both sites. As for nutrients, Eulemur fulvus collaris avoids lipid rich fruits in 

Sainte Luce and Cheirogaleus medius selects fruit with high sugar content both in Sainte 

Luce and in Kirindy as preparation for their torpor. Secondly, many traits such as fruit and 

seed size, growth form, colour, seed number, seed protection and extractable proteins do 

not differ between sites and do not determine lemur food selection. Finally, for several 

traits, both morphological (pulp type, odour, fruit skin protection) and biochemical (total 

nitrogen, tannins, ADF and NDF) the lemur species seem to display a large dietary 

flexibility. For these features, which differ between sites, the animals select according to 

the overall availability at a given site. These results show that overall there is a weak 

149


selection pressure by frugivores on fruit traits and this at both study sites. At the same 

time these frugivores show remarkable regional dietary variation, a phenomenon that has 

been found in other studies on primates as well (Richard and Dewar 1991; Richard 1977; 

Chapman 1995). Thus on a larger geographical scale, the results confirm our previous 

conclusions that fruit traits are more likely to be the result of abiotic conditions rather than 

of interactions with their frugivores. In Kirindy, no evidence could be found either for 

species-specific co-evolution. To summarize there has been no support for tight coevolution 

between fruit traits and dispersers in Madagascar (this study), nor in temperate 

forests (Herrera 1984, 1987; Hampe 2003) or the Neotropics (Janzen 1983; Pizo 2002; 

Silva et al. 2002) and the Paleotropics (Gautier-Hion et al. 1985; Corlett 2002). There 

seem to be three main explanations for this. First of all, research has shown that fruit 

traits are very conservative and strongly phylogenetically determined (Herrera 1986, 

1989, 1995; Jordano 1995, Hampe 2003). Secondly, fruit selection by dispersers is not 

consistent in time and space and has shown to be too weak to shape fruit traits in 

different ecosystems with different frugivore diversity. Worldwide, fruit size seems to be 

the only common selection cue for dispersers. Finally, many studies, along with this one 

show that rather abiotic factors may shape fruit characteristics (Herrera 1986, 1995; 

Hampe 2003; Voigt et al, submitted). It seems thus that we need to look beyond coevolution 

as it may well be that the ‘conservative’ fruit traits have more strongly influenced 

frugivore behaviour than the other way around. 

As mentioned before, one might question the suitability to study seed dispersal in a site 

with a depauperate frugivore community. Indeed, Madagascar is characterized by a 

species poor frugivore guild, with sparsely representation of bird and bat species and the 

complete lack of larger mammals like elephants and ungulates. Besides the composition 

and abundance of the frugivore assemblage varies significantly over time as larger lemur 

species have gone extinct during Holocene and bush pigs and lemur species 

disappeared locally. On an evolutionary time scale these extinctions occurred recent and 

it is unlikely that they may have interfered in shaping fruit traits. The main disadvantage of 

dealing with a depauperate and incomplete frugivore community is that it obviously 

narrows down the chance of finding species-specific co-evolution. Specialized dispersal 

strategies seem rather unlikely in an ecosystem where few frugivores occur. 

Nevertheless it does seem that the frugivore composition is reflected in the 

representation of the different dispersal syndromes (cf. findings Gautier-Hion et al. 1985; 

Ganesh and Davidar 2000; Hampe 2003; Voigt et al. submitted) as very few ‘bird-fruits’ 

are present at the study site. Besides, the models that were investigated, were developed 

in the Neotropics where species-rich frugivore communities are present. Nonetheless no 

evidence could be found there either, nor elsewhere. So in the end, my results from 

Sainte Luce do not seem to contradict with findings from other sites with completely 

different frugivore guilds. The main advantage of this particular frugivore composition is 

that it allowed me to sample data on all frugivores present and to include frugivory by 

bats, birds and rats which has only been poorly studied up to now in Malagasy forests. 

Interesting as well is that following Fleming’s (1979) theory, Madagascar represents an 

extreme situation for the Paleotropics as it has a significantly low and unpredictable fruit 

production and few frugivores. Furthermore, the majority of frugivores and fruiting trees 

studied are endemic, which stresses the importance of providing data on these distinct 

ecosystems as well to complete our knowledge on the tropics worldwide. Besides, insight 

150


in fruit-frugivore interactions in Sainte Luce is crucial to understand forest dynamics here 

and to provide a solid database on which to formulate conservation plans. 

In the closing section of this study, the hazardous situation of the Malagasy littoral forest 

at the moment is described and suggestions are made as to how my findings can be 

used in the design of population habitat viability analyses (PHVA), which will further lead 

to the development of conservation management plans. The littoral forest is expected to 

lose numerous endemic plant and animal species in the near future because of 

continuing deforestation and resultant habitat changes. Of great concern is the disruption 

of plant-animal interactions. Alterations in the recruitment dynamics of plant species in 

forest fragments might have unknown consequences for their long-term survival. As 

forests become more fragmented, the remaining patches become increasingly isolated 

and less accessible for arboreal lemur species, which were found to be important seed 

dispersers for numerous plant species. Consequently, gene flow and seed dispersal 

between patches become more critical for the long-term survival of many plant species. 

Conversely, if fragments get too small or hunting increases, the long-term survival of 

animal species will not be guaranteed either. As Holocene extinctions have shown, large 

frugivores, such as E. f. collaris, are most vulnerable to habitat fragmentation. We now 

know that this species is the only remaining seed disperser of large-seeded plant species 

and its decline and extinction will inevitably lead to a decline and lack of regeneration of 

large-seeded trees. Frugivorous birds and flying foxes are the most important mobile 

seed dispersers bringing seeds into grasslands and early succession vegetation. Genetic 

exchange and long distance dispersal between fragments is less likely to occur if 

populations of mobile flying dispersers decrease or vanish. It is thus not only the threat of 

habitat loss but also the hunting in the littoral forest that will disrupt animal-plant 

dynamics. I believe my findings on seed dispersal can represent a crucial input for 

underlying PHVA of conservation management plans. They will help to indicate priorities 

of action and vulnerable animal and plant species, which need special protection efforts. 

Obviously, it is of crucial importance to urgently protect the largest remaining intact forest 

fragments, which act as reservoirs from where indigenous floral and faunal species can 

colonize new habitats. In this respect, conservation zones are being established by QMM 

at the moment. Furthermore, to accelerate the natural recovery process corridors and 

plantations are being installed. In addition, environmental education in local schools 

should be continued as active cooperation with local people is indispensable to the 

success of sustainable use of land and natural resources and to actively protect the 

remaining littoral forests. 

To conclude, this study provided a survey on fruit availability and its fluctuations in the 

littoral forest as well as an extensive three-dimensional dataset involving numerous plant 

species with their corresponding phenological, morphological and biochemical traits. Fruit 

diets of all frugivore species were obtained as well. Based on my results, I can conclude 

with great certainty that, in the littoral forest of Sainte Luce, fleshy-fruited plants engage in 

diffuse mutualisms with their dispersal agents. These interactions are quite generalized, 

very ancient and extraordinarily frequent in certain communities (Willson and Traveset 

2000). High unpredictability and asymmetry of interactions, coupled with an important 

influence of abiotic factors, signal that mutual selection pressures between plants and 

seed dispersers are greatly constrained (Levey and Benkman 1999). In Sainte Luce fruiteating 

animals tend to consume many fruit species and likewise the fruits of many plants 

151


are consumed by a wide range of animals, possibly to minimize the effects of the loss of 

one dispersal agent. Abiotic factors seem to be more responsible than biotic ones in 

shaping fruit characteristics. The long-term dynamics of fruits and their dispersers appear 

to be decoupled and the diet choice of frugivores shows a remarkable flexibility towards 

variations in the fruit supply. If frugivore preference had influenced the evolution of fruit 

traits at all it would most probably have acted upon general characteristics, such as fruit 

size. Clearly, this shows that abiotic variables and phylogeny are much more important in 

this ecosystem and thus may outweigh the extent of connections between frugivores and 

fruits. 

Future research should concentrate more on the animal side of these 

interactions, which is essential to understand how niche separation among frugivores is 

organised in space and time. This will lead to a better understanding of what exactly 

determines diet choice. Data on the post-dispersal phase are needed as well to complete 

the dispersal cycle and to comprehend how the assembly and recruitment of plant 

communities are organised in space and time. As clear patterns in one year may 

disappear in the next, long-term data on phenology are needed. At present several 

undergraduate and graduate students are looking at parts of these processes, slowly 

filling the gaps in our knowledge. Even though there is still a lot to explore and to 

investigate, now is the time to act in order to preserve the littoral forest, and to prevent 

the irreversible disappearance of this precious ecosystem. 

152

SAMENVATTING 

Samenvatting 

Dit doctoraat behandelt zaadverspreiding op soorten op gemeenschapsniveau in het 

littoraal regenwoud van Sainte Luce in zuidoost Madagaskar. Het tracht bovendien een 

inzicht te geven in de ecologische relaties tussen enerzijds de gilde van frugivoren en 

anderzijds bepaalde soorten vruchtbomen kenmerkend voor dit ecosysteem. Meer 

specifiek wordt aandacht besteed aan de predispersie- en dispersiefase. 

Zaadverspreiding wordt benaderd zowel vanuit het standpunt van de plantensoorten als 

dat van de diersoorten. In het eerste deel bestuderen we aan de hand van de temporele 

beschikbaarheid van vruchten en hun morfologische en biochemische kenmerken de 

verspreidingsstrategieën van 34 boomsoorten. Een uitgebreid onderzoek wordt hiervoor 

uitgevoerd naar de algemene vruchtbeschikbaarheid in het littoraal regenwoud. 

Vervolgens worden bestaande hypotheses, zoals soortspecifieke co-evolutie, het hoge 

versus het lage investeringsmodel en het concept van verspreidingssyndromen 

onderzocht. Verder hebben we getracht te achterhalen op basis van welke 

vruchtkenmerken de verschillende diersoorten hun voedsel selecteren. Hiervoor werden 

in dit woudtype 173 vruchtensoorten beschreven op morfologie en voedingswaarde. 

Dertien vruchteneters komen in dit ecosysteem voor. Zij behoren tot de volgende 

taxonomische groepen: primaten (4 spp.), vleermuizen (1 sp.), vogels (6 spp.) en 

knaagdieren (2 spp.). Ook bepalen we eveneens de rol die deze verschillende 

diersoorten spelen in het behoud van deze habitat en de regeneratie van typische 

woudsoorten door een actieve deelname aan de verspreiding en/of predatie van zaden. 

Om de geldigheid van deze resultaten binnen een ruimer kader te testen en te 

interpreteren, vergelijken we vruchtkenmerken en voedselkeuze van twee lemuursoorten 

in twee totaal verschillende woudtypes. We kunnen besluiten dat dit werk niet alleen 

aanleiding heeft gegeven tot fundamenteel wetenschappelijke resultaten maar eveneens 

een bijdrage kan leveren in de onderbouwing van beleidsplannen omtrent conservatie. 

Eerst bestuderen we de vruchtbeschikbaarheid in het littoraal regenwoud. Deze wordt 

aan de hand van twee methodes bepaald, met name phenological transects en fruit trails. 

Hoewel dit woudtype een aseizoenaal klimaat kent, blijkt uit de resultaten toch dat er 

duidelijke seizoenale en jaarlijkse fluctuaties optreden in het vruchtaanbod. Jaarlijks is er 

een vruchtpiek van november tot en met februari. Anderzijds zijn de periodes van 

vruchtschaarste minder voorspelbaar, ze verschillen van jaar tot jaar. Frugivoren hebben 

dus af te rekenen met zowel periodes van vruchtovervloed als van -schaarste. In het 

littoraal woud zijn, net als in de andere regenwouden in Madagaskar de phenophases 

(bladgroei, bloei en vruchtrijping) onderling sterk gecorreleerd en bereiken ze hun 

hoogtepunt in de periode van maximale neerslag (november-februari). In droge 

woudtypes daarentegen zijn de phenophases meer gespreid over het hele jaar. Zonlicht 

fungeert als trigger bij fotosynthese en daglengte blijkt sterk gecorreleerd te zijn met alle 

phenophases in Sainte Luce. Variatie in daglengte is een belangrijke factor in dit extreem 

zuidelijk gelegen tropisch woud in tegenstelling tot evenaarwouden waar daglengte 

constant is. Daarenboven is de aanwezigheid van rijpe vruchten ook sterk gecorreleerd 

met temperatuur. Dit verband is hoogstwaarschijnlijk een gevolg van het hoge 

nutriëntengehalte in de bodem in periodes van hoge temperatuur en neerslag. Deze 

153


omstandigheden zijn bovendien ideaal voor de zaadkieming. De resultaten van zowel 

phenological transects als fruit trails tonen aan dat beide methodes aanvullend zijn. In 

tegenstelling tot het noteren van vruchtrijping bij grote bomen in de phenological 

transects, worden in fruit trails alle groeivormen (bomen, struiken, kruiden, epifyten en 

lianen) in rekening gebracht. Uit de resultaten blijkt dat die onderling een verschillende 

impact hebben op het seizoenaal verloop van het integraal vruchtaanbod in een woud. 

Bovendien worden in fruit trails ook de vruchten op de bodem meegeteld. Door 

combinatie van beide methodes wordt een inzicht verworven in de temporele variatie in 

vruchtbeschikbaarheid voor zowel arboreale als terrestrische vruchteneters. 

Vervolgens richten we onze aandacht op verspreidingsstrategieën van 34 boomsoorten. 

Hierbij baseren we ons op enkele basismodellen uit de tropische ecologie, zijnde 

soortspecifieke co-evolutie, het model van lage versus hoge investering (McKey 1975) en 

verspreidingssyndromen. Eerst en vooral gaat men er bij co-evolutie van uit dat één 

bepaalde vruchtensoort verspreid wordt door één enkele vruchteneter. Hierbij zou de 

wisselwerking tussen beide soorten zo sterk zijn dat ze elkaars evolutie beïnvloeden, wat 

mogelijk zou leiden tot extreme vormen van wederzijdse aanpassingen. Volgens het 

model van lage en hoge investering verwachten we dat de lage investeerders of 

generalisten aan de hand van massale vruchtproductie gedurende een korte periode 

zoveel mogelijk zaadverspreiders aantrekken met hun waterige en zoete vruchten. De 

hoge investeerders of specialisten daarentegen produceren minder vruchten maar met 

een hogere voedingswaarde (hoog vet- en eiwitgehalte) en dit gedurende een langere 

vruchtperiode. Hierdoor zouden ze slechts enkele maar wel efficiënte zaadverspreiders 

aantrekken. Uit beide modellen vloeit het meer genuanceerd principe van 

verspreidingssyndromen voort waarbij bepaalde morfologische co-adaptaties in vrucht en 

zaad bepaalde taxonomische diergroepen aantrekken. Om deze modellen en principes te 

testen hebben we aan de hand van fruit traps en tree watches de identiteit van de 

vruchteneters achterhaald evenals hun rol als mogelijke zaadverspreiders en/of – 

predatoren. Fenologische, morfologische en biochemische kenmerken van de 

vruchtensoorten werden bovendien in rekening gebracht om te testen of er al dan niet covariatie 

optreedt tussen deze kenmerken en bepaalde verspreidingsstrategieën. Er werd 

geen bewijs gevonden voor co-evolutie, noch voor het model van lage en hoge 

investering. Desalniettemin kan er toch een onderscheid gemaakt worden tussen wat we 

vogel-, zoogdieren gemengde (zowel vogels als zoogdieren) vruchten kunnen noemen 

op basis van vruchten zaadgrootte, vruchtvorm en zaadaantal per vrucht. 

Voedingswaarde kan niet volgens deze categorieën ingedeeld worden. Vijf boomsoorten 

met grote zaden worden enkel en alleen verspreid door Eulemur fulvus collaris. Hieruit 

concluderen dat dit om co-evolutie gaat is wellicht fout, gezien het eerder een gevolg is 

van het uitsterven van grote frugivore vogel- en lemuursoorten, die ongetwijfeld ook deze 

grote zaden konden verspreiden. Niettegenstaande spreekt het voor zich dat deze 

exclusieve interacties uiterst belangrijk zijn vanuit het standpunt van conservatie. Wat het 

investeringsmodel betreft, blijkt dat dit model gebaseerd werd op vogelvruchten in de 

Neotropics en de geldigheid ervan blijkt dan ook sterk afhankelijk te zijn van de 

samenstelling van de frugivore gilde. De soortenarme groep van vruchteneters zowel in 

heel Madagascar als in het littoraal regenwoud heeft blijkbaar geen aanleiding gegeven 

tot gespecialiseerde verspreidingsstrategieën. Het is voor een bepaalde boomsoort 

mogelijk te riskant om afhankelijk te zijn van slechts één enkele vruchteneter. Als gevolg 

hiervan worden de meeste boomsoorten dan ook gekarakteriseerd door een set van 

154


gemengde en veeleer algemene morfologische en biochemische kenmerken. Anderzijds 

heeft de lage soortendiversiteit van frugivore vogels aanleiding gegeven tot opvallend 

weinig vogelvruchten hier in vergelijking met andere tropische sites. Tot slot kunnen we 

dus stellen dat van de drie modellen het concept van de verspreidingssyndromen het 

meest aannemelijk is in Sainte Luce maar de variatie in verspreidingsstrategieën wordt 

hierdoor maar gedeeltelijk verklaard. 

Het centrale hoofdstuk van dit doctoraat behandelt de vrucht-frugivoor interacties die deel 

uitmaken van het ecosysteem. Zowel het dieet als de selectiecriteria en de overlapping 

van voedselsoorten van de 13 verschillende vruchten- en zaadeters wordt beschreven 

evenals hun respectievelijke rol binnen dit ecosysteem. Vrucht- en zaadgrootte blijken de 

voornaamste kenmerken te zijn in de voedselvoorkeur van alle frugivoren. Vogels eten 

vooral kleine vruchten, terwijl voor de verschillende zoogdieren de gegeten vruchten 

zaadgrootte evenredig toeneemt met de grootte van de mondholte en het lichaam. Terwijl 

vogels voornamelijk rode, paarse en zwarte vruchten selecteren, hebben de nocturne 

lemuren en een duidelijke voorkeur voor sappige en vlezige vruchten met een dunne 

schil. Vliegende honden eten eveneens graag vlezige vruchten. Wat voedingswaarde 

betreft verkiezen vogels vetrijke vruchten terwijl de zoogdieren (Eulemur fulvus collaris en 

Pteropus rufus) deze net vermijden. Dwerg- en muislemuren selecteren vruchten met een 

hoog suikergehalte om vetreserves aan te leggen voor hun torpor. Overlapping van de 

diëten tussen de verschillende vruchteneters is relatief hoog hier ten opzichte van andere 

sites, wat mogelijk kan verklaard worden door de onvoorspelbaarheid van het 

vruchtaanbod in het littoraal regenwoud. Dit zou op zijn beurt een verklaring kunnen zijn 

voor het lage soortenaantal frugivoren dat Madagaskar en onze site kenmerkt. Ook al is 

er een grote overlapping van dieet, de impact van de vruchteneters op zaadverspreiding 

blijkt toch verschillend en hun rol in het ecosysteem evenmin overlappend. Naast de 

vermelde vruchteneters, telt dit woud ook een aantal zaadeters. Zowel endemische als 

exotische rattensoorten, een tortelduifsoort en twee papegaaiensoorten zijn duidelijk 

granivoren daar zij de zaden vernietigen van de meeste vruchten die ze eten. 

Daarentegen zijn de sterk frugivore duiven en bulbuls belangrijke zaadverspreiders voor 

de verspreiding van soorten in open gebieden waar deze planten deel uitmaken van de 

eerste successiefase. E. f. collaris heeft een zeer soortenrijk dieet en is daarenboven als 

grootste vruchteneter vooral belangrijk als exclusieve verspreider van grote zaden 

binnenin een woudfragment. De kleine dwerg- en muislemuren zijn eerder omnivoor en 

verspreiden vooral zaden van kleine vruchten gedurende de periode waarin ze actief zijn. 

Tenslotte zijn de vliegende honden uiterst belangrijk voor zaadverspreiding op lange 

afstand en verzekeren ze bovendien de genetische uitwisseling tussen plantenpopulaties 

en woudfragmenten. Dit heterogeen zaadtransport is enorm belangrijk in een sterk 

gefragmenteerd ecosysteem zoals in Sainte Luce. In twee kleinere hoofdstukken 

besteden we extra aandacht aan de voedselecologie en de rol van de vliegende honden 

als zaadverspreiders en de zwarte papegaaien als zaadpredatoren omdat over beide 

soorten tot hiertoe nauwelijks data beschikbaar zijn voor Madagaskar. 

Tenslotte vergelijken we vruchten zaadkenmerken binnen twee totaal verschillende 

woudtypes in Madagaskar: enerzijds het littoraal regenwoud van Sainte Luce (zuidoost 

Madagaskar), anderzijds het droge bladverliezende woud van Kirindy (west 

Madagaskar). Aan de hand hiervan trachten we de rol van abiotische factoren en 

vruchteneters in relatie te brengen tot de evolutie van morfologische en biochemische 

155


vruchtkenmerken. Beide sites verschillen sterk in abiotische factoren maar hebben een 

vergelijkbare frugivore gilde. Deze studie laat ons toe te testen of onze bevindingen van 

Sainte Luce ook opgaan voor andere sites van Madagaskar en ze vervolgens in een 

ruimer kader te interpreteren. Eerst en vooral is het duidelijk dat de algemene 

vruchtkenmerken sterk verschillen tussen beide sites. Sainte Luce heeft opvallend meer 

sappige steenvruchten en bessen met dunne schil terwijl openspringende doosvruchten 

en steenvruchten met dikke schil meer voorkomen in Kirindy. Biochemisch ligt het 

vetgehalte hoger in vruchten van Sainte Luce, in tegenstelling tot de hogere 

concentraties vezels, tanninen en stikstof in Kirindy. De typische vruchtkenmerken in dit 

droge bladverliezende woud duiden dan ook vooral op aanpassingen tegen watervlies 

gedurende het lange droge seizoen. Wanneer we kijken naar voedselselectie van 

Eulemur fulvus en Cheirogaleus medius binnen en tussen beide sites wordt het duidelijk 

dat kenmerken zoals vruchten zaadgrootte, groeivorm, kleur, zaadaantal, 

zaadbescherming en eiwitgehalte, die niet verschillen tussen de sites, niet echt relevant 

zijn voor de voedselselectie van deze lemuren. Anderzijds is er wel een sterke voorkeur 

voor zoöchore bessen en steenvruchten in beide sites, ook al is er een kwantitatief 

verschil in de aanwezigheid van deze vruchttypes tussen Kirindy en Sainte Luce. Wat de 

voedingswaarde betreft, negeert Eulemur fulvus vetrijke vruchten in Sainte Luce en 

selecteert Cheirogaleus medius vooral vruchten met een hoog suikergehalte en dit zowel 

in Sainte Luce als in Kirindy als voorbereiding voor hun torpor zoals eerder vermeld. 

Tenslotte blijkt dat beide lemuursoorten zich sterk kunnen aanpassen aan de algemene 

beschikbaarheid wat tal van morfologische (pulptype, geur, vruchtwand) en biochemische 

kenmerken (stikstof, tannines, ADF en NDF) betreft. Deze vruchtkenmerken verschillen 

sterk tussen beide sites en de lemuren eten dan ook wat meest aanwezig is in een 

bepaalde site. Voedselkeuze van beide soorten lemuren wordt dus sterk bepaald door 

het algemene vruchtaanbod. Aan de hand van deze resultaten kunnen we dus stellen dat 

vruchtkenmerken wellicht eerder het gevolg zijn van sitegebonden abiotische factoren 

dan het gevolg van een sterk selectieve impact door interacties met hun vruchteneters. 

Net zoals in andere onderdelen van deze studie kan ook deze vergelijking tussen sites 

evenmin bewijsmateriaal leveren voor het bestaan van co-evolutie. Er kan hooguit 

gesproken worden van een zwakke selectieve invloed van de frugivoren op 

vruchtkenmerken. De lemuren zijn duidelijk flexibel genoeg om hun dieet aan te passen 

en te overleven op vruchten met andere morfologische kenmerken en andere 

voedingswaarde, waarvan het aanbod verschilt per regio. 

Uiteindelijk wordt in het afsluitende hoofdstuk de huidige situatie van het littoraal 

regenwoud toegelicht en worden enkele adviezen geformuleerd die relevant zijn voor 

conservatie. Het littoraal woud van Sainte Luce bestaat momenteel enkel nog uit sterk 

gedegradeerde woudfragmenten. Dit woud riskeert dan ook vele endemische planten- en 

diersoorten te verliezen in de nabije toekomst omwille van ontbossing en verdere 

habitatdegradatie. Het ontwrichten van plant-dier interacties is dan ook een van de 

grootste bedreigingen, aangezien veranderingen in de regeneratie van plantensoorten in 

woudfragmenten belangrijke gevolgen kunnen hebben voor het overleven van deze 

soorten op lange termijn. Slash and burn, nutteloze bosbranden en houtkap vormen de 

belangrijkste oorzaken voor wouddegradatie en fragmentatie. Daarenboven is er in de 

nabije toekomst ook nog de bedreiging van houtskool en titaniumontginning in deze 

regio. Door sterke fragmentatie raken de resterende woudfragmenten geïsoleerd 

waardoor ze niet langer toegankelijk zijn voor arboreale diersoorten. Genetische 

156


uitwisseling tussen fragmenten door vogels en vleermuizen is dan ook van extreem 

belang om de dynamiek van deze wouden in stand te houden. Daarenboven hebben 

grote zaden weinig zaadverspreiders en verdienen deze plantensoorten speciale 

aandacht wat conservatie betreft. Het zijn bovendien ook juist die grote vruchteneters met 

hoge habitatvereisten die het meest kwetsbaar zijn en bedreigd worden door 

versnippering van hun habitat. Niet enkel habitatverlies maar ook de jacht op 

verscheidene diersoorten kan leiden tot het uiteenvallen van plant-dier interacties. Aan de 

hand van onze resultaten hopen we toch duidelijk een aantal sleutelsoorten, zowel 

binnen fauna als flora te hebben aangeduid die extra aandacht verdienen in 

natuurbehoud. De bevindingen van dit proefschrift kunnen dan ook als soliede basis 

dienen waarop beleidsvoering rond natuurbeheer ter plaatse zich kan baseren. In eerste 

instantie is het van groot belang de meest intacte en grootste woudfragmenten dringend 

maar vooral actief te beschermen. Zij vertegenwoordigen de laatste reservoirs vanwaar 

planten en dieren nieuwe habitats kunnen koloniseren. Anderzijds is het ook noodzakelijk 

de regeneratie van het woud te stimuleren door het aanleggen van plantages en corridors 

die herbebossing bevorderen. Actieve deelname van de lokale bevolking is hierbij 

onmisbaar en van cruciaal belang voor de slaagkans van het natuurbehoud in deze regio. 

Milieu-educatie in de lokale scholen is in dit verband een ideale manier om de Antanosy 

actief bij deze projecten te betrekken. Alleen zo kunnen wij hen overtuigen dat het 

belangrijk is voor henzelf en vooral voor de toekomstige generaties om hun wouden te 

beschermen. 

Als conclusie hopen we te kunnen stellen dat dit doctoraat erin geslaagd is een eerste 

inzicht te geven in de algemene vruchtbeschikbaarheid en haar fluctuaties in het littoraal 

regenwoud. Eveneens hebben we een driedimensionale dataset van 173 plantensoorten 

samengesteld met hun respectieve fenologische, morfologische en biochemische 

kenmerken. Bovendien hebben we het dieet en de voedselecologie van de verschillende 

vruchteneters beschreven. Uit de resultaten is gebleken dat in het littoraal regenwoud 

van Sainte Luce sappige en vlezige vruchten eerder diffuus geassocieerd zijn met hun 

zaadverspreiders. Een hoge onvoorspelbaarheid en asymmetrie in de interacties, die 

daarenboven sterk beïnvloed worden door abiotische factoren, leggen beperkingen op 

aan mutuele selectieve druk. Vruchteneters consumeren een grote variëteit aan vruchten, 

terwijl vruchten daarentegen vaak door verschillende diersoorten gegeten worden. 

Hierdoor is het risico van te sterke afhankelijkheid beperkt wanneer een welbepaalde 

zaadverspreider lokaal uitsterft. Abiotische factoren (neerslag, bodem, zonlicht, 

temperatuur) zijn duidelijk belangrijker in het scheppen van vruchtkenmerken dan 

biotische factoren (de frugivoren zelf) in het littoraal regenwoud. Bovendien vertonen 

bepaalde frugivore diersoorten vaak een hoge flexibiliteit wat hun voedsel betreft en 

passen zij zich aan de algemene beschikbaarheid aan. Toch zijn er enkele duidelijke 

voedselvoorkeuren van bepaalde taxonomische groepen of soorten, maar die hebben 

blijkbaar geen sterke impact gehad op de ontwikkeling van bepaalde vruchtkenmerken. 

Des te meer omdat de samenstelling en grootte van frugivore groepen sterk kan 

verschillen in de tijd. Dit is zeker het geval in Madagaskar waar talrijke grote vogel- en 

lemuursoorten uitstierven sinds de komst van de mens. Als er al enige duidelijke impact 

heeft bestaan van frugivoren op de evolutie van vruchtkenmerken dan heeft dit vooral 

geleid tot bepaalde algemene morfologische kenmerken. Vooral abiotische factoren 

overheersen de relatie vrucht-frugivoor. Om de cyclus van zaadverspreiding te 

vervolledigen zouden toekomstige studies in Sainte Luce het lot van de zaden na 

157


zaadverspreiding moeten bestuderen. Zo kan de samenstelling en regeneratie van 

plantengemeenschappen in tijd en ruimte nog beter begrepen worden. De rol van 

zaadverspreiders benadrukt het probleem van ‘lege wouden’ waar langlevende bomen 

als ‘levende doden’ getuige zijn van de ontkoppeling van essentiële biotische interacties 

ten gevolge van habitatverstoring en jacht. 

158

RÉSUMÉ 

Résumé 

Cette thèse de doctorat traite des modalités de dispersion des graines au niveau des 

espèces ainsi qu’au niveau de la communauté (floristique et faunistique) de la forêt 

littorale de Sainte Luce au Sud-est de Madagascar. Cette étude tente de mieux 

comprendre les relations écologiques entre la communauté de frugivores et les arbres 

fruitiers qui caractérisent cet écosystème. Plus spécifiquement, nous nous sommes 

concentrés sur les phases de pré-dispersion et dispersion des graines. La dispersion des 

graines est abordée du point de vue de la flore aussi bien que de la faune. Dans un 

premier temps, nous avons étudié les stratégies de dispersion des arbres en nous basant 

sur la disponibilité temporelle des fruits et sur leurs caractéristiques morphologiques et 

biochimiques. Ensuite, nous avons évalué certaines hypothèses telles que le principe de 

co-évolution, le modèle d’investissement haut/bas et le concept des syndromes de 

dispersion. Enfin, nous avons tenté de déterminer, sur base de quelques caractéristiques 

des fruits, comment les différents frugivores sélectionnent leur nourriture. Pour cela, nous 

avons décrit et catégorisé 173 espèces de fruits en nous basant sur leur morphologie et 

leur valeur nutritive. Dans cet écosystème, il existe treize espèces de frugivores 

appartenant à quatre groupes taxonomiques: les primates (4 spp.), les chauve-souris (1 

sp.), les oiseaux (6 spp.) et les rongeurs (2 spp.). Nous avons évalué le rôle que jouent 

ces animaux dans le maintien de cet habitat et dans la régénération des espèces 

forestières typiques au travers de leur contribution à la dispersion et/ou à la prédation des 

graines. Afin de tester la validité de nos résultats de manière plus vaste, nous avons 

étudié les caractéristiques des fruits et la sélection alimentaire des deux espèces de 

lémuriens dans deux types de forêts différentes. Finalement, ce travail n'a pas seulement 

donné lieu à des résultats scientifiques intéressants mais a aussi contribué à l'élaboration 

de plans de gestion pour la conservation de ce milieu. 

Premièrement, nous avons étudié la disponibilité des fruits et ses fluctuations dans une 

forêt littorale du Sud-est de Madagascar. Afin de faire ce peu, nous avons parcouru des 

layons phénologiques et réalisé des relevés mensuels au sol. Malgré que la saisonnalité 

soit peu marquée et qu'une vraie saison sèche soit inexistante dans la forêt littorale de 

Madagascar, les tendances phénologiques interannuelles peuvent être considérées 

comme relativement saisonnières. De plus, une variation intra-annuelle se discerne avec 

des périodes d’abondance et d'autres de pénurie en fruits. Toutes les phénophases 

(feuillaison, floraison, fructification) sont inter-corrélées et culminent à un maximum 

durant la période de novembre à février. Ce phénomène a déjà été observé dans 

d’autres forêts humides malgaches. Cependant, dans les forêts malgaches sèches, les 

différentes phénophases sont relativement espacées dans le temps en raison d’un climat 

plus saisonnier. Apparemment, la photopériode exerce un impact important sur les 

phénophases. La luminosité fonctionne comme un déclencheur pour la photosynthèse et, 

de manière évidente, il existe une plus grande variation de la durée du jour dans un site 

situé aussi extrêmement au sud. Par contre, la durée du jour est constante dans les 

forêts équatoriales. Les précipitations n'ont une importance que pour la feuillaison et la 

température affecte la présence des fruits et surtout des fruits mûrs. Ceci peut être 

expliqué par la haute valeur nutritive du sol durant les périodes de plus haute 

159

Résumé 

température et précipitation. Ces circonstances sont par ailleurs idéales pour la 

germination des graines. Lorsqu'on compare les résultats obtenus par les deux 

méthodologies que nous avons utilisées, la différence de présence de fruits mûrs peut 

être expliquée par une contribution différente des types biologiques rencontrés sur le 

layon phénologique et le relevé au sol. Cette différence de contribution influence 

clairement les tendances de fructification totale. De fait, lors des relevés au sol, les fruits 

rencontrés à même le sol sont aussi comptés. La combinaison des deux méthodes met 

en évidence la variation temporaire dans la disponibilité des fruits pour les frugivores 

arboricoles ainsi que les frugivores terrestres. 

Ensuite, nous nous sommes concentrés sur les stratégies de dispersion de 34 espèces 

d’arbres. Nous nous sommes basés sur certains modèles de l’Ecologie Tropicale : la coévolution, 

le modèle d’investissement haut/bas (McKey 1975) et les syndromes de 

dispersion. Selon le modèle de co-évolution, une espèce de fruit dépend d’une seule 

espèce de frugivore pour sa dispersion. Cette interaction est si forte que les deux 

espèces influencent leurs évolutions réciproquement, ce qui peut mener aux formes 

extrêmes d’adaptations mutuelles. Selon le modèle d’investissement haut/bas, les 

moindre investisseurs ou généralistes attirent les espèces disséminatrices grâce une 

production fruitière importante durant une courte période de fructification. En général, la 

pulpe des fruits de ce genre d'espèces est succulente et sucrée. Par contre, les grands 

investisseurs ou spécialistes produisent moins de fruits mais d’une valeur nutritive plus 

haute avec un taux élevé de lipides et protéines, et ce, pendant une période plus longue. 

De ce fait, elles n'attirent que peu d'espèces disséminatrices mais ces espèces seront 

des disséminatrices efficaces. De ces deux modèles découle des syndromes de 

dispersion qui supposent des co-adaptations morphologiques des fruits et graines attirant 

certains groupes taxonomiques plutôt que d'autres. Afin de tester ces modèles et 

principes, nous avons identifié les consommateurs et déterminé leur rôle en tant que 

disséminateurs ou prédateurs par les méthodes des placettes, des pièges à fruits et par 

un suivi régulier de quelques pieds de différentes espèces. De plus, nous avons 

caractérisé les traits phénologiques, morphologiques et biochimiques des espèces de 

fruit pour tester s’il existait une co-variation de ces traits et de certaines stratégies de 

dispersion. Les résultats ne soutiennent pas l'hypothèse de la co-évolution ni le modèle 

d’investissement haut/bas. Néanmoins, il est quand même possible de distinguer les 

fruits à oiseaux, les fruits à mammifères et les mixtes (mangés par les oiseaux et par les 

mammifères) sur base de la taille des fruits et graines, de la forme des fruits et du 

nombre de graine par fruit. Ces mêmes catégories ne se retrouvent pas quand on 

mesure la valeur nutritive des différentes espèces. Cinq espèces d’arbres à grandes 

graines ne peuvent être disséminées que par Eulemur fulvus collaris. Nous ne pouvons 

dès lors conclure que ces cas représentent un cas de co-évolution. Il s'agirait plutôt d'une 

conséquence de l’extinction des grandes espèces de lémuriens et oiseaux qui étaient 

capables d'aussi avaler ces graines de grande taille. Pourtant ces interactions exclusives 

sont extrêmement importantes du point de vue de la conservation. Quant au modèle 

d’investissement différent, il semble qu'il soit basé sur des fruits à oiseaux des 

Néotropiques et sa validité dépend fortement de la composition de la communauté de 

frugivores. En règle générale, la communauté de frugivores est relativement pauvre à 

Madagascar tout comme à Sainte Luce, ce qui semble avoir mené à des stratégies plutôt 

généralistes de la part des espèces de cette communauté. Il est peut-être top risqué pour 

une espèce d’arbre fruitier de dépendre d’une seule espèce de frugivore pour sa 

160

Résumé 

dispersion dans cette région. Par conséquent, la plupart des espèces d’arbres est 

caractérisée par des traits morphologiques et biochimiques générales et non spécialisés. 

D’autre part, la faible diversité des oiseaux frugivores a pu avoir donné lieu au fait qu'il y 

ait très peu de fruits à oiseaux en comparaison avec d'autres sites. En conclusion, on 

constate que des trois modèles, celui des syndromes de dispersion semble le plus 

plausible à Sainte Luce. Cependant, il n'explique que partiellement la variation dans les 

syndromes de dispersion. 

Le chapitre central de ce doctorat traite des interactions fruit-frugivore qui font partie de 

l’écosystème. On y décrit aussi le régime alimentaire des frugivores que le 

chevauchement des espèces consommées par 13 différentes frugivores et granivores 

ainsi que leurs rôles respectifs dans cet écosystème. La taille des fruits et des graines 

semble être un des facteurs les plus déterminants dans la sélection alimentaire de tous 

les frugivores. Les oiseaux mangent surtout des petits fruits tandis que pour les différents 

mammifères, la taille des fruits et graines augmentent proportionnellement avec la taille 

de leur gueule et de leur corps. Alors que les oiseaux sélectionnent les fruits rouges, 

pourpres et noirs, les lémuriens nocturnes ont une préférence pour les fruits succulents à 

enveloppe fine. Pteropus rufus se délecte aussi de fruits succulents. Quant à la valeur 

nutritive des fruits consommés, les oiseaux favorisent les fruits au contenu en lipides très 

élevé alors que les mammifères (Eulemur fulvus collaris et Pteropus rufus) évite 

simplement ceux-ci. Cheirogaleus spp. et Microcebus rufus sélectionnent des fruits 

sucrés pour créer des réserves de graisses durant les périodes de torpeur. Le 

chevauchement des régimes alimentaires entre les différents frugivores est relativement 

haut comparé aux autres sites, ce qui peut être expliqué par l’imprévisibilité de la 

disponibilité des fruits dans la forêt littorale et, qui peut être le motif du nombre 

relativement bas des espèces frugivores à Madagascar. Même s’il y avait un 

chevauchement important dans les régimes alimentaires, l’impact des différents 

frugivores sur la dispersion des graines serait quand même différent ainsi que leur rôle 

dans l’écosystème. Outre les frugivores déjà mentionnés, il y a aussi des granivores dans 

la forêt littorale (deux espèces de rongeurs endémiques et exotiques, une espèce de 

tourterelle et deux espèces de perroquet noir) qui consomment et détruisent les graines. 

Par contre les pigeons et bulbuls sont des disséminateurs importants qui transportent les 

graines dans les ouvertures de la végétation, là où ces espèces feront partie de la 

première phase de régénération de la forêt. E. f. collaris a un régime alimentaire très 

divers et, étant le plus grand frugivore, cette espèce est surtout importante en tant que 

disséminatrice exclusive des grandes graines à l’intérieur d’un seul fragment de forêt. 

Cheirogaleus spp. et Microcebus rufus sont plutôt omnivores et dispersent surtout les 

graines des petits fruits pendant la période où ils sont actifs. Enfin la chauve-souris 

Pteropus rufus est l'espèce la plus importante pour la dispersion des graines à longue 

distance et assure en plus l’échange génétique entre les populations des plantes des 

différents fragments de forêt. Ce transport hétérogène des graines est essentiel dans un 

écosystème aussi fragmenté que celui de Sainte Luce. Dans deux plus petits chapitres, 

nous attirons l’attention sur l’écologie alimentaire de Pteropus rufus et Coracopsis nigra 

et sur leur rôle en tant que disséminateurs et prédateurs de fruits parce que le régime 

alimentaire de ces deux espèces reste encore très peu connu à Madagascar. 

Finalement, nous comparons les caractéristiques de fruits et graines dans deux types de 

forêt complètement différentes à Madagascar: D'une part, la forêt littorale de Sainte Luce 

161

Résumé 

(Sud-est de Madagascar) et d’autre part, la forêt sèche caducifoliée (Ouest Madagascar). 

Cette étude nous permet de mieux comprendre le rôle des facteurs abiotiques (humidité, 

climat, sol) sur les caractéristiques des fruits et l’influence des frugivores eux-mêmes qui 

exercent une pression sélective sur ces mêmes caractéristiques (interactions fauneflore). 

Les deux sites diffèrent fortement par les facteurs abiotiques qui y prévalent mais 

sont occupés par une communauté de frugivores comparables. Ceci nous permet de 

vérifier si nos constatations et les résultats obtenus lors de cette étude sont valables 

aussi sur d'autres sites malgaches et de les interpréter dans un cadre plus vaste. Il est 

clair que les caractéristiques de fruits sont assez différentes entre les deux sites. A 

Sainte Luce, il y a beaucoup plus de baies et drupes succulentes à enveloppe fine tandis 

qu'à Kirindy, on retrouve plus de capsules déhiscentes et drupes à enveloppe épaisse. 

Biochimiquement le taux de lipides des fruits est plus élevé à Sainte Luce. Les fruits de 

Kirindy ont, quant à eux, un taux plus élevé de fibres, de tannins et de composés azotés. 

Ces caractéristiques sont typiques d'une forêt caducifoliée et indiquent l'existence 

d'adaptations contre la déshydratation pendant la longue saison sèche. Quant à la 

sélection des fruits par les deux genres de lémuriens (Eulemur et Cheirogaleus) dans un 

même site ainsi que entre les deux sites, les résultats montrent que les caractéristiques 

comme la taille de fruits et de graines, le type de végétation dans laquelle le fruit se 

retrouve, la couleur du fruit, le nombre de graines par fruits, la protection des graines et le 

taux des protéines qui ne diffèrent pas entre les deux sites, ne sont pas pertinents pour 

leur sélection alimentaire. Par contre, il existe une claire préférence pour des baies et 

drupes zoochores dans les deux sites même s’il y a une différence du point de vue de 

leur disponibilité entre Kirindy et Sainte Luce. Quant à la valeur nutritive, Eulemur fulvus 

collaris ignore les fruits riches en lipides à Sainte Luce, tandis que Cheirogaleus medius 

sélectionne surtout des fruits sucrés à Sainte Luce et Kirindy pour préparer leur torpeur 

tel que précédemment mentionné. Finalement, nous avons trouvé que pour certaines 

caractéristiques morphologiques (type de pulpe, odeur, enveloppe) et biochimiques 

(azotes, tannins, ADF, NDF), ces espèces de lémuriens peuvent s’adapter étroitement à 

la disponibilité générale des fruits. Ces caractéristiques diffèrent dans les deux sites et 

les lémuriens mangent ce qui est à leur disposition. Sur la base de ces résultats, il 

semblerait que ces caractéristiques soient plutôt influencées par les conditions abiotiques 

spécifiques à chaque site que par les interactions strictes entre lémuriens frugivores et 

les espèces consommées. Cette comparaison affaiblit les possibilités d'existence de la 

co-évolution, ce qui correspond à nos constatations précédentes. Nous dirions plutôt qu’il 

n’y a qu’une faible influence sélective des frugivores sur les caractéristiques de fruits. Les 

lémuriens sont clairement assez flexibles que pour s’adapter à un régime alimentaire 

caractérisé par des fruits à morphologies différentes et avec une autre valeur nutritive, 

selon la disponibilité locale et ponctuelle en fruits de la région. 

Enfin, dans le chapitre final, nous commentons la situation actuelle de la forêt littorale et 

nous nous permettons de formuler certains conseils pertinents pour la conservation. En 

ce moment, la forêt littorale de Sainte Luce est composée de fragments fortement 

dégradés. Cette forêt risque donc de perdre beaucoup d’espèces de plantes et d'animaux 

endémiques dans un futur proche avec pour cause le déboisement et la dégradation 

progressive de l'habitat. L'interruption des interactions entre les plantes et les animaux 

est une des plus grandes menaces pour cette région car elle suppose des changements 

consécutifs dans la régénération des plantes qui peut avoir des conséquences 

importantes pour la survie des espèces à plus long terme. Les cultures sur brûlis, les feux 

162

Résumé 

de brousse et la coupe du bois sont les causes plus importantes de cette dégradation et 

fragmentation de la forêt. Dans le futur, cette région sera en plus sujette à la fabrication 

du charbon et à l’extraction d’ilménite, ce qui représentent quelques menaces de plus à 

long terme. Par une forte fragmentation, les îlots de forêt deviennent de plus en plus 

isolés et moins accessibles pour les animaux arboricoles. L'échange génétique entre les 

populations d'oiseaux et de chauve-souris des différents fragments est crucial pour 

maintenir la dynamique de ces forêts. Les grandes graines ont peu de disséminateurs 

potentiels et méritent une attention spéciale en matière de Conservation. Par ailleurs, ces 

grands frugivores nécessitent de grands habitats et sont donc les plus vulnérables face à 

la fragmentation du leur habitat. Non seulement la perte de l'habitat mais aussi une 

chasse excessive peut mener à l'interruption des interactions entre les plantes et les 

animaux. Nos résultats prétendent avoir indiqué certaines espèces-clés de la faune et de 

la flore qui méritent une attention spéciale et qu'il est urgent de protéger. Certaines 

constatations de cette étude de doctorat peuvent être intégrées concrètement et 

appliquées dans les analyses sur lesquelles se base la gestion de la faune et la 

Conservation. Tout d’abord, il est d’un grand intérêt de préserver activement et 

urgemment les fragments de forêts les plus grands et plus intacts. Ils représentent les 

derniers réservoirs où plantes et animaux peuvent coloniser de nouveaux habitats. 

D'autre part, il est aussi nécessaire de stimuler la régénération de la forêt en installant 

des plantations et corridors qui stimulent le reboisement. Une participation active de la 

population locale est indispensable pour la réussite de chaque plan de conservation dans 

la région. L'éducation environnementale dans les écoles locales est la façon idéale pour 

impliquer les Antanosy dans les projets de protection de l'environment. C’est l’unique 

manière de pouvoir les convaincre de l’importance d'une conservation de leurs forêts 

pour eux-mêmes et les générations futures. 

En conclusion, nous espérons que ce doctorat soit parvenu à faire mieux comprendre la 

disponibilité des fruits et ces fluctuations dans la forêt littorale du Sud-est de 

Madagascar. Nous avons établi un ensemble de données tridimensionnelles de 173 

espèces de plantes avec leurs caractéristiques phénologiques, morphologiques et 

biochimiques respectives. Nous avons aussi décrit le régime et la sélection alimentaire 

de différents frugivores. Les résultats montrent que dans la forêt littorale de Sainte Luce 

les fruits juteux et succulents sont associés à leurs disséminateurs de graines de manière 

diffuse. Une haute imprévisibilité et une asymétrie dans les interactions, qui sont 

fortement influencés par des facteurs abiotiques, limitent les possibilités de pression 

sélective mutuelle. Les frugivores consomment une grande variété de fruits et les fruits 

sont souvent mangés par plusieurs espèces d’animaux. De cette façon, le risque d’une 

trop forte dépendance est limitée lorsqu'un disséminateur particulier disparaît localement. 

Les facteurs abiotiques (précipitations, nature du sol, soleil, température) influencent plus 

les caractéristiques des fruits que les facteurs biotiques (les frugivores eux-mêmes) dans 

la forêt littorale de Madagascar. Certaines frugivores montrent, quant à eux, une grande 

flexibilité, notamment, en ce qui concerne leur nourriture et s’adaptent facilement à la 

disponibilité générale en fruits du site. Néanmoins, il y a des préférences évidentes dans 

la nutrition de certains groupes taxonomiques mais qui n’ont apparemment pas eu 

d'impact important sur le développement de certaines caractéristiques de fruits. De 

même, la composition et diversité d’espèces de frugivores à un site peuvent différer dans 

le temps. Ceci est certainement le cas à Madagascar où de grandes espèces d’oiseaux 

et de lémuriens se sont éteintes depuis l’arrivée des hommes. S'il y a eu un impact clair 

163

Résumé 

des frugivores sur l’évolution des caractéristiques de fruits, cela a plutôt amené à des 

traits morphologiques très généraux. Ce sont surtout les facteurs abiotiques qui 

déterminent la relation fruit-frugivore. Il serait intéressant que les études futures à Sainte 

Luce se concentrent sur le destin des graines après dispersion pour compléter l'étude du 

cycle de dispersion. Ainsi la composition et régénération des communautés des plantes 

dans l’espace et le temps pourraient être encore mieux compris. Le rôle des 

disséminateurs souligne l’importance des forêts vides où des arbres à longévité élevée 

sont comme des reliques, les derniers témoins de l'interruption des interactions biotiques 

mutuelles à cause de la perturbation des habitats et de la chasse. 

164

REFERENCES 

References 

Andrianisa JA (1989) Observations phénologiques dans la Réserve Spéciale de Nosy 

Mangabe, Maroansetra. Mémoires de fin d’études, Ecole Supérieur des Sciences 

Agronomiques, Université d’Antananarivo, Madagascar 

Andrews JR, Birkinshaw CR (1998) A comparison between the daytime and nighttime 

diet, activity and feeding height of the Black Lemur, Eulemur macaco (Primates: 

Lemuridae) in Lokobe forest, Madagascar. Folia Primatol 69:175-182 

Atsalis S (1999) Diet of the brown mouse lemur (Microcebus rufus) in Ranomafana 

national park, Madagascar. Int J Primatol 20:193-229 

Atsalis S (2000). Spatial distribution and population composition of the brown mouse lemur 

(Microcebus rufus) in Ranomafana National Park, Madagascar, and its implications 

for social organization. Am J Primatol 51:61-78 

Bairlein F (1996) Fruit-eating in birds and its nutritional consequences. Comp Biochem 

Physiol 113:215-224 

Baldi N (2002) Il ruola della fitochimica nella scelta alimentare di Eulemur fulvus collaris, 

una proscimmia del Madagascar. Unpublished Master Thesis, Pisa University, 

Italy 

Banack SA (1998) Diet selection and resource use by flying foxes (genus Pteropus). 

Ecology 79:1949–1967 

Banks M (2002) Littoral forest primate fauna in the Tolagnaro (Fort-Dauphin) region of 

south-eastern Madagascar. ONG Azafady Report 

Barton RA, Purvis A, Harvey PH (1995) Evolutionary radiation of visual and olfactory 

brain systems in primates, bats and insectivores. Philos T Roy Soc B 348:381-310 

Beier P, Noss RF (1998) Do habitat corridors provide connectivity? Conserv Biol 

12:1241-1252 

Birkinshaw CR (1999) The importance of the black lemur (Eulemur macaco) for seed 

dispersal in Lokobe Forest, Nosy be. In: Rakotosamimana B, Rakotosamimana H, 

Ganzhorn JU, Goodman SM (eds) New Directions in Lemur Studies. Kluwer 

Academic/Plenum Publishers, New York, pp 189-199 

Birkinshaw CR (2001) Fruit characteristics of species dispersed by the black lemur 

(Eulemur macaco) in the Lokobe Forest, Madagascar. Biotropica 33:478-486 

Bleher B, Böhning-Gaese K (2001) Consequences of frugivore diversity for seed dispersal, 

seedling establishment and the spatial pattern of seedlings and trees. Oecologia 

129:385-394 

Böhning-Gaese K, Gaese BH, Rabemanantasoa SB (1999) Importance of primary and 

secondary seed dispersal in the Malagasy tree Commiphora guillaumini. Ecology 

80:821-832 

Bollen A, Van Elsacker L (2001) Tree dispersal strategies; a variety of possibilities in 

inland littoral forest of SE-Madagascar. Abstract Poster BES meeting, Reading 

(UK), pp 20 

Bollen A, Donati G, Fietz J, Schwab D, Ramanamanjato J-B, Randrihasipara L, Van 

Elsacker L, Ganzhorn JU (2002) Fruit characteristics and fruit choice by frugivorous 

lemurs in a dry deciduous forest and a humid littoral forest of Madagascar. Abstract 

Oral Presentation. ATB-Conference Panama City (Panama), pp 11 

165

References 

Bollen A, Van Elsacker L (2002a) Feeding ecology of Pteropus rufus (Pteropodidae) in 

the littoral forest of Sainte Luce, SE Madagascar. Acta Chiropterologica 4:33-47 

Bollen A, Van Elsacker L (2002b) Seed predation and secondary dispersal in littoral 

forest (Sainte Luce, SE Madagascar): an experimental approach. Abstract Poster. 

ATB-Conference Panama City (Panama), pp 135 

Bollen A, Van Elsacker L (2002c) Focal tree studies: Dypsis prestoniana as an example. 

Abstract. BENELUX Congress of Zoology Antwerp (Belgium), pp 57 

Bollen A, Van Elsacker L (2003) Measuring fruit availability: data from the littoral forest of 

Sainte Luce (SE-Madagascar). Abstract Poster ATB-Conference, Aberdeen 

(Scotland), pp 60 

Bollen A, Donati G, Fietz J, Schwab D, Ramanamanjato JB, Randrihasipara L, Van 

Elsacker L. and Ganzhorn JU (2003) Fruit characteristics in a dry deciduous and a 

humid littoral forest of Madagascar: evidence for selection pressure through abiotic 

constraints rather than seed dispersers. In: Dew L, Boubli JP (eds) Fruits and 

frugivores. Kluwer Publishers, Dordrecht, Chapter 6 (in press) 

Bonnaire L, Simmen B (1994) Taste perception of fructose solutions and diet in 

Lemuridae. Folia Primatol 63:171-176 

Bourlière F (1985) Primate communities: their structure and role in tropical ecosystems. 

Int J Primatol 6:1-26 

Britt A (2000) Diet and feeding behaviour of the black-and-white ruffed lemur (Varecia 

variegata variegata) in the Betampona Reserve, Eastern Madagascar. Folia 

Primatol 71:133-141 

Brugiere D, Gautier J-P, Moungazi A, Gautier-Hion A (2002) Primate diet and biomass in 

relation to vegetation composition and fruiting phenology in a rain forest in Gabon. 

Int. J Primatol 23:999-1024 

Carlo TA, Collazo JA, Groom MJ (2003) Avian fruit preferences across Puerto Rican 

forested landscape: pattern consistency and implications for seed removal. 

Oecologia 134:119-131. 

Chapman CA (1995) Primate seed dispersal: coevolution and conservation implications. 

Evol Anthropol 4:74-82 

Chapman CA, Chapman LJ (1996) Frugivory and the fate of dispersed and nondispersed 

seeds of six African tree species. J Trop Ecol 12:491-504 

Chapman CA, Onderdonk DA (1998) Forests without primates: primate/plant 

codependency. Am J Primatol 45:127-141 

Chapman CA, Peres CA (2001) Primate conservation in the new millennium: the role of 

scientists. Evol Anthropol 10:16-33 

Chapman CA, Wrangham R (1994) Indices of habitat-wide fruit abundance in tropical 

forests. Biotropica 26:160-171 

Chapman CA, Chapman LJ, Wrangham RW, Hunt K, Gebo D, Gardner L (1992a) 

Estimators of Fruit Abundance of Tropical Trees. Biotropica 24:527-531 

Chapman LJ, Chapman CA, Wrangham RW (1992b) Balanites wilsoniana: elephant 

dependent dispersal? J Trop Ecol 8:275-283 

Charles-Dominique P (2001) Relationships between seed dispersal and behavioural 

ecology In: Bongers F, Charles-Dominique P, Forget J-P, Thery M (eds) 

Nouragues:Dynamics and plant-animal interactions. Kluwer Academic Publishers, 

Dordrecht, pp 191-196 

166

References 

Charles-Dominique P, Atramentowicz M, Charles-Dominique M, Gerard H, Hladik A, 

Hladik CM, Prevost MF (1981) Les mammifères frugivores arboricoles nocturnes 

d’une forêt guyanaise: interrelations plantes-animaux. Terre Vie-Rev Ecol A 

35:341-435 

Cheke AS, Dahl JF (1981) The status of bats on western Indian Ocean islands, with 

special reference to Pteropus. Mammalia 45:205–238 

Clout MN (1989) Foraging behaviour of glossy black cockatoos. Aust Wildlife Res 16:467- 

473 

Connell JH (1971) On the role of natural enemies in preventing competitive exclusion in 

some marine animals and rain forest trees. In: den Boer PJ, Gradwell PR (eds) 

Dynamics of populations. Centre for Agricultural Publishing and Documentation, 

Wageningen, pp 289-312 

Corlett RT (1996) Characteristics of vertebrate-dispersed fruits in Hong Kong. J Trop Ecol 

12:819-833 

Corlett RT (1998) Frugivory and seed dispersal by vertebrates in the Oriental 

(Indomalayan) Region. Biol Rev 73:413–448 

Corlett RT (2002) Frugivores and seed dispersal in degraded tropical east asian 

landscapes. In: Levey DJ, Silva RW, Galetti M (eds) Seed dispersal and frugivory: 

Ecology, Evolution and Conservation. CAB International, New York, pp 451-465 

Curtis DJ (1997) The Mongoose lemur (Eulemur mongoz) a study in behavior and 

ecology. PhD Dissertation, University Zurich, Germany. 

Curtis DJ, Zaramody A (1998) Group size, home range use, and seasonal variation in the 

ecology of Eulemur mongoz. Int J Primatol 19:811-835 

Darwin C (1859) On the origin of species by means of natural selection. John Murray, 

London 

Debussche M (1988) La diversité morphologique des fruits charnus en Languedoc 

méditerranéen: relations avec les caractéristiques biologiques et la distribution des 

plantes et avec les disséminateurs. Acta Oecol 9:37-52 

Debussche M, Isenmann P (1989) Fleshy fruit characters and the choices of bird and 

mammal seed dispersers in a Mediterranean region. Oikos 56:327-338 

Debussche M, Isenmann P (1992) A Mediterranean bird disperser assemblage 

composition and phenology in relation to fruit availability. Terre Vie-Rev Ecol A 

47:411-432 

Dew JL, Wright P (1998) Frugivory and seed dispersal by four species of primates in 

Madagascar’s eastern rain forest. Biotropica 30:425-437 

Djletati R, Brun B, Rumpler Y (1997) Meiotic study of hybrids in the genus Eulemur and 

taxonomic considerations. Am J Primatol 42:235-245 

Dominy N, Lucas P (2002) The sensual side of primate food choice. Am J Phys Anthropol 

34:64 

Donati G (2002) L’attivita’ e le sue correlate ecologiche nel lemure bruno dal collare, 

Eulemur fulvus collaris (Lemuridae), nella foresta litorale di Ste Luce (Fort-Dauphin, 

Madagascar). PhD Dissertation, University of Pisa, Italy 

Donati G, Lunardini A, Kappeler PM (1999) Cathemeral activity of red-fronted brown 

lemur (Eulemur fulvus rufus) in the Kirindy Forest/CFPF. In: Rakotosamimana B, 

Rakotosamimana H, Ganzhorn JU, Goodman SM (eds) New Directions in Lemur 

Studies. Kluwer Academic/Plenum Publishers, New York, pp 119-137 

Dowsett-Lemaire F (1988) Fruit choice and seed dissemination by birds and mammals in 

the evergreen forest of upland Malawi. Terre Vie-Rev Ecol A 43:251-285 

167

References 

Dowsett JR (2000) Le statut de perroquets vasa et noir C. vasa et C nigra et de 

l'inséparable a tête grise Agapornis canus à Madagascar. IUCN, Cambridge 

Dumetz N (1999) High plant diversity of lowland rainforest vestiges in eastern 

Madagascar. Biodivers Conserv 8:273-315 

Du Puy B (1996) Faunal interactions with the genus Adansonia in the Kirindy Forest. In: 

Ganzhorn JU, Sorg JP (eds) Ecology and economy of a tropical dry forest in 

Madagascar. Primate Report, 46-1, Göttingen, pp 329-334 

Erikkson O, Ehrlen J (1998) Phenological adaptations in fleshy vertebrate-dispersed fruits 

of temperate plants. Oikos 82:617-621 

Felber HR (1984) Influence des principales propriétés physico-chimiques des sols et de 

la structure des peuplements sur le succès de la régénération naturelle d’essences 

représentatives sur des layons de débardage dans une forêt sèche de la côte 

occidentale de Madagascar. Master thesis, Fachbereich Waldbau, ETH Zurich 

Ferraro PJ (2001) The local costs of establishing protected areas in low income nations: 

Ranomafana National Park, Madagascar. Environmental policy working paper n° 

2001-006. Georgia State University 

Fietz J, Ganzhorn JU (1999) Feeding ecology of the hibernating primate Cheirogaleus 

medius: how does it get so fat? Oecologia 121:157-164 

Fietz J (1999) Demography and floating males in a population of Cheirogaleus medius. In: 

Rakotosamimanana B, Rasamimanana H, Ganzhorn JU, Goodman SM (eds) New 

Directions in Lemur Studies. Kluwer Academic/Plenum Publishers, New York, pp 

159-172 

Fischer KE, Chapman CA (1993) Frugivores and fruit syndromes: differences in patterns 

at the genus and species level. Oikos 66:472-482 

Fleming TH (1979) Do tropical frugivores compete for food? Am Zool 19:1157–1172 

Fleming TH (1987) Fruit bats: prime movers of tropical seeds. Bats 5:3–8 

Fleming TH (1992) How do fruit- and nectar-feeding birds and mammals track their food 

resources. In: Hunter MD, Takayuki O, Price PW (eds) Effects of resource 

distribution on animal-plant interactions. Academic Press, San Diego, pp 355-391 

Fleming TH, Breitwitsch R, Whitesides,GH (1987) Patterns of tropical frugivore diversity. 

Annu Rev Ecol Syst 18:91-109 

Fleming TH, Heithaus ER, Sawyer WB (1977) An experimental analysis of the food 

location behavior of frugivorous bats. Ecology 58:619–627 

Frankie GW (1975) Tropical forest phenology and pollinator plant coevolution. In: Gilbert 

LE, Raven PH (eds) Coevolution of animals and plants. University of Texas, Austin, 

pp 192-207 

Freed BZ (1996) Co-occurrence among crowned lemurs (Lemur coronatus) and 

Sanford’s lemurs (Lemur fulvus sanfordi) of Madgascar. PhD Dissertation. 

Washington University, St. Louis 

Fujita MS, Tuttle MD (1991) Flying foxes (Chiroptera: Pteropodidae): Threatened animals 

of key ecological and economic importance. Conserv Biol 5:455–463 

Galetti M (1993) Diet of the scaly-headed parrot (Pionus maximiliani) in a semideciduous 

forest in southeastern Brazil. Biotropica 25:419-425 

Galetti M, Pizo MA (1996) Fruit eating birds in a forest fragment in southeastern Brazil. 

Ararajuba 4:71-79 

Galetti M (2000) Frugivory by toucans (Rhampastidae) at two altitudes in the Atlantic rain 

forest of Brazil. Biotropica 32:842-850 

168

References 

Ganesh T, Davidar P (2000) Dispersal mode of tree species in wet forests of southern 

Western Ghats. Curr Sci India 80:394-399 

Ganzhorn JU (1985) Utilization of eucalyptus and pine plantations by brown lemurs in the 

eastern rainforest of Madagascar. Primate Conservation 6:34-35 

Ganzhorn JU (1987) A possible role of plantations for primate conservation in 

Madagascar. Am J Primatol 12:205-215 

Ganzhorn JU (1988) Food partitioning among Malagasy primates. Oecologia 75:436-450 

Ganzhorn JU (2003) Effects of introduced Rattus rattus on endemic small mammals in dry 

deciduous forest fragments of western Madagascar. Animal Conserv 6:147-157 

Ganzhorn JU, Abraham JP (1991) Possible role of plantations for lemur conservation in 

Madagascar: food for folivorous species. Folia Primatol 56:171-176 

Ganzhorn JU, Kappeler PM (1996) Lemurs of the Kirindy Forest. In: Ganzhorn JU, Sorg JP 

(eds) Ecology and economy of a tropical dry forest in Madagascar. Primate Report, 

46-1, Göttingen, pp 257-274 

Ganzhorn JU, Schmid J (1998) Different population dynamics of Microcebus murinus in 

primary and secondary deciduous dry forest of Madagascar. Int J Primatol 19:785- 

796 

Ganzhorn JU, Sorg J-P (1996) Ecology and economy of a tropical dry forest. Primate 

Report, 46-1, Göttingen 

Ganzhorn JU, Fietz J, Rakotovao E, Schwab D, Zinner D (1999a) Lemurs and the 

regeneration of dry deciduous forest in Madagascar. Conserv Biol 13:794-804 

Ganzhorn JU, Goodman SM, Ramanamanjato J-B, Ralison J, Rakotondravony D, 

Rakotosamimanana B (2000) Effects of fragmentation and assessing minimum 

viable populations of lemurs in Madagascar. In: Rheinwald G (ed) Isolated 

vertebrate communities in the tropics. Bonner zoologische Monographien Edition. 

Volume 46. Museum Alexander Koenig, Bonn, pp 265-272 

Ganzhorn JU, Lowry PP, Schatz GE, Sommer S (2001). The biodiversity of Madagascar: 

one of the world’s hottest hotspots on its way out. Oryx 35: 346-348 

Ganzhorn JU, Malcomber S, Andrianantoanina O, Goodman SM (1997) Habitat 

characeristics and lemur species richness in Madagascar. Biotropica 29:331-343 

Ganzhorn JU, Wright PC, Ratsimbazafy J (1999b) Primate communities: Madagascar. In: 

Fleagle JG, Janson C, Reed KE (eds) Primate Communities. Cambridge University 

Press, Cambridge, pp 75-89 

Garber PA, Lambert JE (1998) Introduction to primate seed dispersal: Primates as seed 

dispersers: ecological processes and directions for future research. Am J Primatol 

45:3-8 

Garbutt N (1999) Mammals of Madagascar. Pica Press, UK 

Gautier-Hion A, Duplantier J-M, Quris R, Feer F, Sourd C, Decoux J-P, Dubost G, 

Emmons L, Erard C, Hecketsweiler P, Moungazi A, Roussilhon C, Thiollay J-M 

(1985) Fruit characters as a basis of fruit choice and seed dispersal in a tropic 

forest vertebrate community. Oecologia 65:324-337 

Gautier-Hion A, Gautier JP, Quiris R (1981) Forest structure and food availability as 

complementary factors influencing the habitat use by a troop of monkeys 

(Cercopithecus cephus). Terre Vie-Rev Ecol 35:511-536 

Gautier-Hion A, Michaloud G (1989) Are figs always keystone species for tropical 

frugivorous vertebrates? A test in Gabon. Ecology 70:1826-1833 

169

References 

Godfrey LR, Junger WL, Reed KE, Simons EL, Prithijit SC (1997) Inferences about past 

and present primate communities in Madagascar. In: Goodman SM, Patterson BD 

(eds) Natural change and human impact in Madagascar. Smithsonian Institution 

Press, Washington and London, pp 218-256 

Goodman SM (1994) A description of a ground burrow of Eliurus webbi (Nesomyinae) 

and case of cohabitation with an endemic bird (Brachypteraciidae, 

Brachypteracias). Mammalia 58:670-672 

Goodman SM, Sterling EJ (1996) The utilization of Canarium (Burseraceae) seeds by 

vertebrates in the Reserve Naturelle Integrale d’Andringitra, Madagascar. Fieldiana 

Zoology 85:83-99 

Goodman SM, Ganzhorn JU (1997) Rarity of figs (Ficus) on Madagascar and its 

relationship to a depauperate frugivore community. Terre Vie-Rev Ecol 52:321-329 

Goodman SM, Ganzhorn JU, Wilmé L (1997a) Observations at a Ficus tree in Malagasy 

humid forest. Biotropica 29:480-488 

Goodman SM, Pidgeon M, Hawkins AFA and Schulenberg S (1997b) Birds of 

southeastern Madagascar. Fieldiana: Zoological New Series, No 87. Field Museum 

of Natural History, Chicago 

Goodman SM, Ganzhorn JU, Rakotondravony D (in press) Introduction to mammals. In: 

Goodman SM, Benstead JP (eds) The Natural History of Madagascar. The University 

of Chicago Press, Chicago 

Green G, Sussman R (1990) Deforestation history of the eastern rain forests of 

Madagascar from satellite images. Science 248:212-215 

Goering HK, Van Soest PJ (1970) Forage fiber analysis. USDA Handbook. N379, 

Washington DC 

Hampe A (1998) Field studies of the Black Parrot Coracopsis nigra in western 

Madagascar. Bulletin of African Bird Club 5:108-113 

Hampe A (2003) Large-scale geographical trends in fruit traits of vertebrate dispersed 

temperate plants. J Biogeogr 30:487-496 

Heithaus ER, Fleming TH, Opler PA (1975) Foraging patterns and resource utilitzation in 

seven species of bats in a seasonal tropical forest. Ecology 56:841–854 

Hemingway CA (1996) Morphology and phenology of seeds and whole fruit eaten by 

Milne-Edwards' sifaka, Propithecs diadema edwardsi, in Ranomafana National 

Park, Madagascar. Int J Primatol 17:637-657 

Hemingway CA (1998) Selectivity and variability in the diet of Milne-Edwards ‘ Sifakas 

(Propithecus diadema edwardsi): Implications for folivory and seed-eating. Int J 


Herrera CM (1984) Determinants of plant-animal coevolution: the case of mutualistic 

dispersal of seeds by vertebrates. Oikos 44:132-141 

Herrera CM (1986) Vertebrate-dispersed plants: Why don’t they behave the way they 

should. In: Estrada A, Fleming TH (eds) Frugivores and seed dispersal. Dr W Junk 

Publishers, Dordrecht, pp 5-18 

Herrera CM (1987) Vertebrate-dispersed plants of the Iberian Peninsula: a study of fruit 

characteristics. Ecol Monogr 57:305-331 

Herrera CM (1989) Seed dispersal by animals: a role in Angiosperm diversification? Am 

Nat 133:309-322 

Herrera CM (1995) Plant-vertebrate seed dispersal systems in the Mediterranean: 

ecological evolutionary, and historical determinants. Annu Rev Ecol Syst 26:705- 

727 

170

References 

Hilton-Taylor C (2000) 2000 IUCN red list of threatened species. IUCN, Gland, 

Switzerland and Cambridge, United Kingdom 

Hladik CM, Charles-Dominique P, Petter JJ (1980) Feeding strategies of five nocturnal 

prosimians in the dry forest of the West coast of Madagascar. In: Charles- 

Dominique P, Cooper HM, Hladik A, Hladik CM, Pages, E, Pariente GF, Petter- 

Rousseaux A, Schilling A, Peter JJ (eds) Nocturnal Malagasy Primates: ecology, 

physiology and behaviour. Academic Press, New York, pp 41–73 

Hladik CM, Simmen B (1996) Taste perception and feeding behaviour in non-human 

primates and human populations. Evol Anthropol 5:58-71 

Holl KD, Loik M, Lin EHV, Samuels IA (2000) Tropical montane forest restoration in 

Costa Rica: overcoming barriers to dispersal and establishment. Restor Ecol 8:339- 

349 

Horvitz CC, Pizo MA, Bello y Bello H, LeCorff J, Dirzo R (2002) Are plant species that 

need gaps for recruitment more attractive to seed-dispersing birds and ants than 

other species? In: Levey DJ, Silva RW, Galetti M (eds) Seed dispersal and 

frugivory: Ecology, Evolution and Conservation. CAB International, New York, pp 

145-160 

Howe HF (1977) Bird activity and seed dispersal of a tropical wet forest tree. Ecology 

58:539-550 

Howe HF (1979) Fear and frugivory. Am Nat 114:925-931 

Howe HF (1984) Constraints on the evolution of mutualism. Am Nat 125:853-865 

Howe HF (1993) Aspects of variation in a neotropical seed dispersal system. Vegetatio 

107/108:149-162 

Howe HF, Estabrook GF (1977) On intraspecific competition for avian dispersers in 

tropical trees. Am Nat 111:817-832 

Howe HF, Smallwood J (1982) Ecology of seed dispersal. Annu Rev Ecol Syst 13:201- 

228 

Hutcheon JM (1994) The great red island: a future for its bats? Bats 12:10–13 

Humbert H (1936-1984) Flore de Madagascar et des Comores. Imprimerie officielle 

Tananarive, Muséum national d'histoire naturelle Paris, or Typographie Firmin- 

Didot-er C ie 

Iaconelli S, Simmen B (2002) Taste tresholds and supratreshold responses to tannin-rich 

plant extracts and quinine in a primate species (Microcebus murinus). J Chem Ecol 

28:2315-2326 

Ingle NR (2003) Seed dispersal by winds, birds, and bats between Philippine montane 

rainforest and successional vegetation. Oecologia 134: 251-261 

Iudica CA, Bonaccorso FJ (1997) Feeding of the bat, Sturnira lilium, on fruits of Solanum 

riparium influences dispersal of this pioneer tree in forests of northwestern 

Argentina. Stud Neotrop Fauna E 32: 4–6 

Izhaki I. 2002. The role of fruit traits in determining fruit removal in east mediterranean 

ecosystems. In: Levey DJ, Silva RW, Galetti M (eds) Seed dispersal and frugivory: 

Ecology, Evolution and Conservation. CAB International, New York, pp 161-175 

Izhaki I, Korine C, Arad C (1995) The effect of bat (Rousettus aegypticus) dispersal on 

seed germination in eastern Mediterranean habitats. Oecologia 101:335–342 

Jacobs GH (1995) Variations in primate colour vision: mechanisms and utility. Evol 

Anthropol 3:196-205 

Janson CH (1983) Adaptation of fruit morphology to dispersal agents in a Neotropical 

forest. Science 219:187-189 

171

References 

Janson CH, Chapman CA (1999) Resource and primate community structure. In: Fleagle 

JG, Janson C, Reed KE (eds) Primate Communities. Cambridge University Press, 

Cambridge, pp 237-267 

Janson CH, Stiles EW, White DW (1986) Selection of plant fruiting traits by brown capuchin 

monkeys: a multivariate approach. In: Estrada A, Fleming TH (eds) Frugivores and 

Seed Dispersal. Dr W Junk Publishers, Dordrecht, pp 83-92 

Janzen DH (1970) Herbivores and the number of tree species in tropical forests. Am Nat 

104:501-528 

Janzen DH (1980) When is it coevolution? Evolution 34:611-612 

Janzen DH (1981) Ficus ovalis seed predation by an Orange-chinned parakeet 

(Brotogeris jugularis) in Costa Rica. Auk 98:841-844 

Janzen DH (1983) Seed and pollen dispersal by animals: convergence in the ecology of 

contamination and sloppy harvest. Biol J Linn Soc 20:103-113 

Janzen DH (1988) Tropical dry forests. The most endangered major tropical ecosystem. 

In: Wilson OE (ed) Biodiversity. National Academy Press, Washington, pp 130-137 

Johns AD, Skorupa JP (1987) Responses of rain-forest primates to habitat disturbance: a 

review. Int J Primatol 8:157-191 

Johnson SE (2002) Ecology and speciation in brown lemurs: white-collared lemurs 

(Eulemur albocollaris) and hybrids (Eulemur albocollaris X Eulemur fulvus rufus) in 

South-eastern Madagascar. PhD Dissertation, University of Texas, Austin 

Jordano P (1983) Fig-seed predation and dispersal by birds. Biotropica 15:38-41 

Jordano P (1992) Fruits and frugivory. In: Fenner M (ed) Seeds: the ecology of 

regeneration in plant communities. CAB International, Wallingford, pp 105-156 

Jordano P (1995) Angiosperm fleshy fruits and seed dispersers: a comparative analysis 

of adaptation and constraints in plant-animal interactions. Am Nat 145:163-191 

Jordano P (2000) Fruits and frugivory. In Fenner M (ed) Seeds: The ecology and 

regeneration of plant communities. CAB International, Wallingford, pp 125-16 

Julliot C (1996) Fruit choice by red howler monkeys (Alouatta seniculus) in a tropical rain 

forest. Am J Primatol 40:261-282 

Julliot C (1997) Impact of seed dispersal by red howler monkeys Alouatta seniculus on 

the seeding population in the understorey of tropical rain forest. J Ecol 85:431-440 

Kalko EKV, Herre EA, Handley COJ (1996) Relation of fruit characteristics to fruit-eating 

bats in the New and Old World tropics. J Biogeogr 23:565-576 

Kannan R, James DA (1999) Fruiting phenology and the conservation of the great pied 

hornbill (Buceros bicornis) in Western Ghats of Southern India. Biotropica 31:167- 

177 

Kappeler PM, Ganzhorn JU (1994) The evolution of primate societies and communities in 

Madagascar. Evol Anthropol 2:159-171 

Kerner A (1898) The natural history of plants, their form, growth, reproduction, and 

distribution. (English translation by FW Oliver) Holt, New York. 

Kitamura S, Yumoto T, Poonswad P, Chuailua P, Plongmai K, Maruhashi T, Noma N 

(2002) Interactions between fleshy fruits and frugivores in a tropical seasonal forest 

in Thailand. Oecologia 133:559-572 

Koechlin J, Guillaumet J-L, Morat P (1974) La forêt dense humide sempervirente de 

basse altitude et la forêt littorale. In: Koechlin J, Guillaumet J-L, Morat P (eds) Flore 

et végétation de Madagascar J Cramer, Vaduz, pp 105–165 

172

References 

Korine C, Izhaki I, Arad Z (1998) Comparison of fruit syndromes between the Egyptian 

fruit-bat (Rousettus aegyptiacus) and birds in East Mediterranean habitats. Acta 

Oecol 19:147–153 

Korine C, Kalko EKV, Herre EA (1999) Fruit characteristics and factor affecting removal 

in a Panamanian community of strangler figs. Oecologia 123:560-568 

Koptur S, Haber WA, Frankie GW, Baker HG (1988) Phenological studies of shrub and 

treelet species in tropical cloud forests of Costa Rica. J Trop Ecol 4:323-346 

Korine C, Kalko EKV, Herre EA (2000) Fruit characteristics and factors affecting fruit 

removal in a Panamanian community of strangler figs. Oecologia 123:560-568 

Knight RS, Siegfried WR (1983) Interrelationships between type, size, color of fruits and 

dispersal in Southern African trees. Oecologia 56:405-412 

Krebs CJ (1989) Ecological methodology. Harper and Row, New York 

Kull CA (2000) Deforestation, erosion and fire: degradation myths in the environmental 

history of Madagascar. Environ Hist 6:423-450 

Lambert JE (2002) Exploring the link between animal frugivory and plant strategies: the 

case of primate fruit processing and post-dispersal fate. In: Levey DJ, Silva RW, 

Galetti M (eds) Seed dispersal and frugivory: Ecology, Evolution and Conservation. 

CAB International, New York, pp 365-379 

Lambert JE, Garber PA (1998) Evolutionary and Ecological Implications of Primate Seed 

Dispersal. Am J Primatol 45:9-28 

Langrand O (1990) Guide to the birds of Madagascar. Yale University Press, New Haven- 

London 

Leighton M (1982) Fruit resources and patterns of feeding , spacing and grouping among 

sympatric Bornean hornbills (Bucerotidae). PhD, University of California 

Levey DJ, Bissell HA, O’Keefe SF (2000) Conversion of nitrogen to protein and amino 

acids in wild fruits. J Chem Ecol 26:179-1763 

Levey J, Benkman CW (1999) Fruit-seed disperser interactions: timely insights from a 

long-term perspective. Tree 14:41-43 

Levey DJ, Moermond TC, Denslow JS (1993) In: McDade L, Bawa KS, Hepsenheide HA, 

Hartshorn GS (eds) La Selva: ecology and natural history of a neotropical 

rainforest. University of Chicago Press, Chicago, pp 282-294 

Lewis Environmental Consultants (1992a) Madagascar Minerals Project. Preliminary 

Environment Impact Assessment Study. Part I: Natural Environment. Appendix III: 

Flora. Report to QMM. Madagascar Minerals Project, Montreal and Antananarivo 

Lewis Environmental Consultants. (1992b). Madagascar Minerals Project. Preliminary 

Environment Impact Assessment Study. Part I: Natural Environment. Appendix IV: 

Fauna. Report to QMM. Madagascar Minerals Project, Montreal and Antananarivo 

Lieberman D (1982) Seasonality and phenology in a dry tropical forest in Ghana. J Ecol 

70:791-806 

Long E, Racey PA (submitted) Food availability and diet of Pteropus rufus in a gallery 

forest surrounded by sisal plantations in SE Madagascar. J Trop Ecol 

Lowry PPII, Schatz GE, Phillipson PB (1997) The classification of natural and 

anthropogenic vegetation in Madagascar. In: Goodman SM, Patterson BD (eds) 

Natural Change and Human Impact in Madagascar. Smithsonian Institution Press, 

Washington, pp 93-123 

Marco DE, Paez SA (2002) Phenology and phylogeny of animal-dispersed plants in a dry 

Chaco forest (Argentina). J Arid Environ 52:1-16 

173

References 

Marshall AG (1983) Bats, flowers and fruit: evolutionary relationship in the Old World. Biol 

J Linn Soc 20:115–135 

Martin RD (1973) A review of the behaviour and ecology of the lesser mouse lemur 

(Microcebus murinus JF Miller, 1777). In: Michael RP, Crook JH (eds) Comparative 

ecology and behaviour in primates. Academic Press, London, pp 1-68 

Martinez del Rio C (1994) Nutritional ecology of fruit-eating and flower-visiting birds and 

bats. In: Langer P, Chivers DJ (eds) The Digestive System in Mammals: Food, Form 

and Function. Cambridge University Press, Cambridge, pp 103-127 

McConkey KR, Drake DR (2002) Extinct pigeons and declining bat populations: are large 

seeds still being dispersed in the tropical pacific? In: Levey DJ, Silva RW, Galetti M 

(eds) Seed dispersal and frugivory: ecology, evolution and conservation. CABI 

Publishing, Wallingford, pp 381-395 

McKey DS (1975) The ecology of coevolved seed dispersal systems. In: Gilbert E, 

Raven, AP (eds) Coevolution of animals and plants. University of Texas Press, 

Austin, pp 159-191 

Medway FLS (1972) Phenology of a tropical rain forest in Malaysia. Biol J Linn Soc 

4:117-146 

Meyers DM, Wright PC (1993) Resource tracking: food availability and Propithecus 

seasonal production. In: Kappeler PM, Ganzhorn JU (eds) Lemur Social Systems 

and their ecological basis Plenum Press, New York, pp 179-192 

Mills LS, Soulé ME, Doak DF (1993) The keystone-species concept in ecology and 

conservation. Bioscience 43:219–224 

Milton K, Dintzis FR (1981) Nitrogen to protein conversion factors for tropical plant 

samples. Biotropica 13:177-181 

MIR Télédétection Inc. (1998) Etude sur la déforestation dans la région de Fort-Dauphin, 

Madagascar. Rapport Final Soumis à QMM (Qit Madagascar Minerals). 

Madagascar Minerals Project, Antananarivo 

Mittermeier RA (1994) Lemurs of Madagascar. Conservation Intl, Washington DC 

Mittermeier RA, Myers N, Thomsen JB (1998) Biodiversity hotspots and major tropical 

wilderness areas: approaches to setting conservation priorities. Conserv Biol 

12:515-520 

Moran MD (2003) Arguments for rejecting the sequential Bonferroni in ecological studies. 

Oikos 100: 403-405. 

Morellato LPC, Talora DC, Takahasi A, Bencke CC, Romera EC, Zipparro VB (2000) 

Phenology of Atlantic rain forest trees: a comparative study. Biotropica 32:811-823 

Morelli V (2002) Il regime alimentare di Eulemur fulvus collaris nella foresta litorale del 

Madagascar: selezione e disponibilità delle risorse. Master Thesis, Pisa University, 

Italy 

Morland HS (1993) Seasonal behavioral variation and its relationship to thermoregulation. 

In: Kappeler PM, Ganzhorn JU (eds) Lemur Social Systems and their ecological 

basis Plenum Press, New York , pp 193-204 

Morrison DW (1978) Foraging ecology and energetics of the frugivorous bat Artibeus 

jamaicensis. Ecology 59:716–723 

Morrison DW (1980) Efficiency of food utilization by fruit bats. Oecologia 45:270–273 

Moshier SL (1991) Ephemeris v5.1. Public domain software. 

Oates JF, Swain T, Zantovska J (1977) Secondary compounds and food selection by 

Colobus monkeys. Biochem Syst Ecol 5:317-321 

174

References 

Oliviera P, Marrero P, Nogales M (2002) Diet of the endemic Madeira Laurel Pigeon and 

fruit resource availability: a study using microhistological analyses. Condor 

104:811-822. 

Overdorff D (1988) Preliminary report on the activity cycle and diet of the red-bellied 

lemur (Lemur rubriventer) in Madagascar. Am J Primatol 16:143-153 

Overdorff DJ (1992) Differential patterns in flower feeding by Eulemur fulvus rufus and 

Eulemur rubriventer in Madagascar. Am J Primatol 28:191-203 

Overdorff DJ (1993a). Similarities, differences and seasonal patterns in the diet of 

Eulemur rubriventer and Eulemur fulvus rufus in the Ranomafana National Park, 

Madagascar. Int J Primatol 14:721-753 

Overdorff DJ (1993b) Ecological and reproductive correlates to range use in red-bellied 

lemurs (Eulemur rubriventer) and rufous lemurs (Eulemur fulvus rufus). In: 

Kappeler PM, Ganzhorn JU (eds) Lemur Social Systems and their ecological basis 

Plenum Press, New York, pp 167-178 

Overdorff DJ (1996) Ecological correlates to activity and habitat use of two prosimian 

primates: Eulemur rubriventer and Eulemur fulvus rufus in Madagascar. Am J 


Overdorff DJ, Strait SG (1998) Seed handling by three prosimian primates in 

Southeastern Madagascar: implications for seed dispersal. Am J Primatol 45:69-82 

Overdorff DJ and Strait SG (1998) Seed handling by three prosimian primates in 

southeastern Madagascar: Implications for seed dispersal. Am J Primatol 45:69-82 

Palmeirim JM, Gorchov DL, Stoleson S (1989) Trophic structure of a neotropical frugivore 

community: is there competition between birds and bats? Oecologia 79:403–411 

Parry-Jones K, Augee ML (1991) Food selection by grey-headed flying foxes (Pteropus 

poliocephalus) occupying a summer colony site near Gosford, New South Wales. 

Wildlife Res 18:111–124 

Pastorini J, Forstner MR, Martin RD (2000) Relationships among brown lemurs (Eulemur 

fulvus) based on mitochondrial DNA sequences. Mol Phylogenet Evol 16:418-429 

Petter JJ, Albignac R, Rumpler Y (1977) Faune de Madagascar: Mammifères Lémuriens. 

Volume 44. ORSTOM CNRS, Paris 

Phillipson PB (1996) Endemism and non-endemism in the flora of south-west Madagascar. 

In: Lourenco WR (ed) Biogeography of Madagascar. ORSTOM, Paris, pp 125-136 

Pizo MA (2002) The seed dispersers and fruit syndromes of Myrtaceae in the Brazilian 

Atlantic Forest. In: Levey DJ, Silva RW, Galetti M (eds) Seed dispersal and 

frugivory: Ecology, Evolution and Conservation. CAB International, New York, pp 

129-143 

Pizo MA, Simao I, Galetti M (1995) Diet and flock size of sympatric parrots in the Atlantic 

forest of Brazil. Ornitologia neotropical 6:87-95 

Porter LJ, Hemingway RW (1990) Significance of condensed tannins. In: Rowe JW (ed) 

Natural products of woody plants. Springer-Verlag, New York, pp 988-1027 

Poulsen JR, Clark CJ, Connor EF, Smith TB (2002) Differential resource use by primates 

and hornbills: implications for seed dispersal. Ecology 83:228-240 

Preston K (1991) Madagascar: a natural history. Curtis Garratt limited, Oxford 

QIT Madagascar Minerals SA (2001) Projet Ilménite: Etude d’impact social et 

environmental. QMM, Antananarivo 

175

References 

Rabevohitra R, Lowry PP, Randrianjafy H, Razafindrianilana N (1996) Rapport sur le 

projet “Assesment of Plant Diversity and Conservation Importance of East Coast 

Low Elevation Malagasy Rain Forests”. Centre National de la recherché appliquée 

au développement rural CENRADERU-FOFIFA, Missouri Botanical Garden 

Racey PA, Nicoll ME (1984) Mammals. In: Stoddart DR (ed) Biogeography and ecology 

of the Seychelles Islands. Junk Press, Leiden, pp 607-627 

Raherilalao MJ (2001) Effets de la fragmentation de la forêt sur les oiseaux autour du 

parc national de Ranomafana (Madagascar). Terre Vie-Rev Ecol 56:389-406 

Rakotonirina (1996) Composition and structure of the dry deciduous forest on sandy 

soils. Primate Report, 46-1, Göttingen, pp 81-87 

Ralisomalala RC (1996) Role de Eulemur fulvus rufus (Audeberg, 1799) et de 

Propithecus verreauxi (A Grandier 1867) dans la dissemination de graines. Primate 

Report, 46-1, Göttingen, pp 285-293 

Ramanamanjato J-B (2000) Fragmentation effects on reptile and amphibian biodiversity 

in the littoral forest of southeastern Madagascar. In: Rheinwald G (ed) Isolated 

vertebrate communities in the tropics. Bonner zoologische Monographien Edition. 

Volume 46. Museum Alexander Koenig, Bonn, pp 297–308 

Ramanamanjato J-B, Ganzhorn JU (2001) Effects of forest fragmentation, introduced 

Rattus rattus and the role of exotic tree plantations and secondary vegetation for the 

conservation of an endemic rodent and a small lemur in littoral forests of southeastern 

Madagascar. Animal Conservation 4:175-183 

Randriamanalina MH, Rafararano L, Babary L, Laha R (2000) Rapport des enguêtees sur 

le chasses dans les fokotany d’Ivondro, d’Erara et d’Etsilesy. Lemur news 5:11-14 

Rasmussen MA (1999) Ecological influences in two cathemeral primates, the mongoose 

lemur (Eulemur mongoz) and the common brown lemur (Eulemur fulvus fulvus). 

PhD Dissertation, Duke University, North Carolina 

Ratsirarson J, Silander JA (1997) Factors affecting the distribution of a threatened 

Madagascar palm species Dypsis decaryi. Principes 41:100-111 

Razafimizanilala AAM (1996) Introduction à l’étude écologique des forêts sublittorales de 

Sainte-Luce (Tolagnaro). Mémoire de diplôme d’études approfondies en sciences 

biologiques appliquées. Option ‘Ecologie végétale’. Université d’Antananarivo 

Renton K (2001) Lilac-crowned parrot diet and food resource availability: resource 

tracking by a parrot seed predator. Condor 103:62-69 

Rice WR (1989) Analyzing tables of statistical tests. Evolution 43:223-225 

Richard AF (1977) The feeding behaviour of Propithecus verreauxi. In: Clutton-Brock TH 

(ed) Primate ecology. Academic Press, London, pp 72-96 

Richard AF, Dewar RE (1991) Lemur ecology. Annu Rev Ecol Syst 22:145-175 

Ridley HN (1930) The dispersal of plants throughout the world. Reeve, Ashford, Kent 

Rigamonti MM (1993) Home range and diet in red ruffed lemurs (Varecia variegata rubra) 

on the Masoala Peninsula, Madagascar. In: Kappeler PM, Ganzhorn JU (eds) 

Lemur Social Systems and their ecological basis, Plenum Press, New York, pp 25- 

39 

Saini MS, Dhindsa MS, Toor TS (1994) Food of the rose-ringed parakeet Psittacula 

krameri: a quantitative study. Journal of the Bombay Natural History Society 91:96- 

103 

176

References 

Sakai S, Momose K, Yumoto T, Nagamitsu T, Nagamasu H, Hamid AA, Nakashizuka T 

(1999) Plant reproductive phenology over four years including an episode of 

general flowering in a lowland dipterocarp forest, Sarawak, Malaysia. Am J Bot 

86:1414-1436 

Schatz GE (2001) Generic Tree Flora of Madagascar. Royal Botanical Gardens, Kew and 

Missouri Botanical Garden Press 

Scharfe F, Schlund W (1996) Seed removal by lemurs in a dry deciduous forest of 

western Madagascar. Primate Report, 46-1, Göttingen, pp 295-304 

Schatz GE (2001) Generic tree flora of Madagascar. Royal Botanical Gardens, Kew and 

Missouri Botanical Garden Press 

Schilling A (1979) Olfactory communication in prosimians. In: Doyle GA, Martin RD (eds) 

The study of prosimian behavior. Academic Press, New York, pp 461-542 

Schmid J (2000) Daily torpor in the gray mouse lemur (Microcebus murinus) in 

Madagascar: energetic consequences and biological significance. Oecologia 

123:175-183 

Shilton LA, Altringham JD, Compton SG, Whittaker RJ (1999) Old World fruit bats can be 

long-distance seed dispersers through extended retention of viable seed in the gut. 

Proc Roy Soc Lond B Bio 266:219–223 

Siegel S (1956) Non parametric statistics for the behavioural sciences. Mc Graw-Hill 

Book Company. New York 

Silva WR, De Marco P, Hasui E, Gomes VSM (2002) Patterns of fruit-frugivore 

interactions in two Atlantic forest bird communities of south-eastern Brazil: 

implications for conservation. In: Levey DJ, Silva RW, Galetti M (eds) Seed 

dispersal and frugivory: Ecology, Evolution and Conservation. CAB International, 

New York, pp 423-436 

Smith JF (2001) High species diversity in fleshy fruited tropical understory plants. Am Nat 

157:646-653 

Smythe N (1970) Relationships between fruiting sesons and seed dispersal methods in a 

neotropical forest. Am Nat 104:25-36 

Snow DW (1971) Evolutionary aspects of fruit-eating in birds. Ibis 113:194-202 

Snow DW (1981) Tropical frugivorous birds and their food plants: a world survey. 

Biotropica 13:1-14 

Soorae PS, Baker LR (2002) eds. Re-introduction NEWS: Special Primate Issue, 

Newsletter of the IUCN/SSC Re-introduction Specialist Group, AbuDhabi, UAE 

(21): 60 pp 

Sorg J-P, Rohner U (1996) Climate and tree phenology of the dry deciduous forest, the 

Kirindy Forest. Primate Report, 46-1, Göttingen, pp 57-80 

Spehn S, Ganzhorn JU (2000) Influence of seed dispersal by brown lemurs on removal 

rates of three Grewia species (Tiliaceae) in the dry deciduous forest of 

Madagascar. Ecotropica 6:13-21 

Stashko E, Dinerstein E (1988) Methods of estimating fruit availability to frugivorous bats. 

In: Kunz TH. Ecological and behavioral methods in the study of bats. Smithsonian 

Institution Press, Washington DC, pp 221–231 

Surridge AK, Osorio D, Mundy NI (2003) Evolution and selection of trichromatic vision in 

primates. Trends Ecol Evol 18:198-205 

Sussman RW (1975) A preliminary study of the behavior and ecology of Lemur fulvus 

rufus Audebert 1800. In: Tattersal I, Sussman RW (eds) Lemur biology. Plenum 

Press, New York, pp 237-258 

177

References 

Tattersall I (1982) The primates of Madagascar. Columbia University Press, New York 

Terborgh J (1983) Five New World Primates; a study of comparative ecology. Princeton 

University Press, New Jersey 

Terborgh J (1986a) Keystone plant resources in the tropical forest. In Soule ME (ed) 

Conservation Biology: the science of scarcity and diversity. Sinauer, Sunderland, 

pp 330-344 

Terborgh J (1986b) Community aspects of frugivory in tropical forests. In: Estrada A, 

Fleming TH (ed) Frugivores and seed dispersal. W. Jung, Dordrecht, pp 371-384 

Terborgh J (1990) An overview of research at Cocha Cashu Biological Station. In: Gentry 

AH (ed) Four neotropical forests. Yale University Press, New Haven, pp 48-59 

Terborgh J, Lopez L, Nuñez P, Rao M, Shahabuddin G, Orihuela G, Riveros M, Ascanio R, 

Adler GH, Lambert TD, Balbas L (2001) Ecological meltdown in predator-free forest 

fragments. Science 294:1923-1926 

Tewksbury JJ, Levey DJ, Haddad NM, Sargent S, Orrock JL, Weldon A, Danielson BJ, 

Brinkerhoff J, Damschen EI, Townsend P (2002) Corridors affect plants, animals 

and their interactions in fragmented landscapes. PNAS 99: 12923-12926 

Thomas DW (1988) Analysis of diets of plant-visiting bats. In: Kunz, TH (ed) Ecological 

and behavioral methods in the study of bats. Smithsonian Institution Press, 

Washington, D.C., pp 211–220 

Tiffney BH (1984) Seed size, dispersal syndromes and the rise of angiosperms: evidence 

and hypothesis. Annals of Missouri Botanical Garden 71:551-576 

Tutin CEG, Williamson EA, Rogers ME, Fernandez M (1991) A case study of a plantanimal 

interation: Cola lizae and lowland gorillas in the Lope Reserve, Gabon. J 

Trop Ecology 7:181-199 

Van der Pijl L (1969) Principles of dispersal in higher plants. Springer Verlag, Berlin. 

Van Schaik CP (1986) Phenological changes in a Sumatran rain forest. J Trop Ecol 

2:327-347 

Van Schaik CP, Terborgh JW, Wright SJ (1993) The phenology of tropical forests: 

Adaptive Significance and Consequences for Primary Consumers. Annu Rev Ecol 

Syst 24:353-377 

Van Soest PJ (1994) Nutritional ecology of the ruminant. Cornell University Press, Ithaca, 

New York 

Vasey N (2000) Niche separation in Varecia variegata rubra and Eulemur fulvus albifrons: 

I. Interspecific patterns. Am J Phys Anthropol 112: 411-431 

Voigt F (2001) Einfluss von Samenausbreitern auf die Häufigkeit von Fruchttypen in 

tropischen Baumgemeinschaften. Master Thesis, Philips-Universitât-Marburg, 

Germany 

Voigt FA, Burkhardt JF, Verhaagh M, Böhning-Gaese K (2002) Regional differences in ant 

community structure and consequences for secondary dispersal of Commiphora 

seeds. Ecotropica 8: 59-66 

Voigt FA, Bleher B, Fietz J, Ganzhorn, Schwab D, Böhning-Gaëse K (submitted) A 

comparison of morphological and chemical fruit traits between two sites with different 

frugivore assemblages. Oecologia 

Wallace RB, Painter RLE (2002) Phenological patterns in a southern Amazonian tropical 

forest: implications for sustainable management. Forest Ecol Manag 160:19-33 

Wallace AR (1879) Colour in Nature. Nature 19:580-581 

Wang BC, Smith TB (2002) Closing the dispersal loop. Trends Ecology Evol 17:379-385 

178

References 

Wenny DG (2000) Seed dispersal of a high quality fruit by specialized frugivores: high 

quality dispersal? Biotropica 32:327-337 

Wheelwright NT (1986) A seven year study of individual variation in fruit production in 

tropical bird-dispersed tree species in the family Lauraceae. In: Estrada A, Fleming 

TH (eds) Frugivores and seed dispersal. Dr W Junk Publishers, Dordrecht, pp 19- 

35 

Wheelwright NT, Haber NA, Murray KG, Guindon C (1984) Tropical fruit-eating birds and 

their food plants: a survey of a Costa Rican lower montane forest. Biotropica 16: 

173-192 

Willson MF & Traveset A (2000) The ecology of seed dispersal. In: Fenner M. (ed) Seeds. 

The ecology of regeneration in plant communities. CABI Publishing, Wallingford, pp 

85-110 

Willson MF, Whelan CJ (1990) The evolution of fruit colour in fleshy-fruited plants. Am 

Nat 136: 790-809 

Willson MF, Irvine AK, Walsh NG (1989) Vertebrate dispersal syndromes in some 

Australian and New Zealand plant communities, with geographic comparisons. 

Biotropica 21:133-147 

Witmer MC, Van Soest PJ (1998) Contrasting digestive strategies of fruit-eating birds. 

Funct Ecol 12:728-741 

Wright PC (1997a) Behavioral and ecological comparisons of Neotropical and Malagasy 

primates. In Kinzey WG (ed) New World Primates: Ecology, Evolution and 

Behavior, Aldine de Gruyter, New York, pp 127-141 

Wright P (1997b) The future of biodiversity in Madagascar. In: Goodman SM, Patterson 

BD (eds) Natural change and human impact in Madagascar. Smithsonian Institution 

Press, Washington and London, pp 381-405 

Wright PC (1999) Lemur traits and Madagascar ecology: coping with an island 

environment. Yearbook Phys Anthropol 42:31-72 

Wright PC, Martin LB (1995) Predation, pollination, and torpor in two nocturnal primates: 

Cheirogaleus major and Microcebus rufus in the rain forest of Madagascar. In: 

Alterman L, Doyle G, Izard K (eds) Creatures of the Dark: The Nocturnal 

Prosimians. Plenum Press, New York, pp 45-60 

Wunderle JM (1997) The role of animal seed dispersal in acceleration native forest 

regeneration on degraded tropical lands. Forest Ecol Manag 99:223-235 

Wütherlich D, Azocar A, Garcia-Nuñez C, Silva JF (2001) Seed dispersal in Palicourea 

rigida a common treelet species from neotropical savannas. J Trop Ecol 17:449- 

458 

Wyner YR, Absher G, Amato E, Sterling R, Stumpf Y, Rumpler Y, DeSalle R (1999) 

Species concepts and determination of historic gene flow patterns in Eulemur 

fulvus complex. Biol J Linn Soc 66:39-56 

Yamashita N (2002) Diets of two lemur species in different microhabitats in Beza 

Mahafaly special reserve, Madagascar. Int J Primatol 23:1025-1051 

Zamora R (2000) Functional equivalence in plant-animal interactions: ecological and 

evolutionary consequences. Oikos 88:442-447 

The format of the references is according the lay-out of the journal Oecologia 

179

family 

Loganiaceae 

scientific name 


Dracaena reflexa var. nervosa 

Liliaceae 

Sarcolaenaceae 

Rutaceae 

Burseraceae 

Violaceae 

Rubiaceae 


Rubiaceae 

Anacardiaceae 

Erythroxylaceae 

Lauraceae 

Euphorbiaceae 


Euphorbiaceae 

Myricaceae 

Rubiaceae 

Sphaerosepalaceae 

Fabaceae 

Celastraceae 

Rubiaceae 

Ochnaceae 

Rubiaceae 

Taccaceae 

Ebenaceae 

Tricalysia cf. cryptocalyx 

Campylospermum obtusifolium 

nr vernacular name scientific name 

family 

nr vernacular name 

1 hazomamy an ala Apodytes sp. nov. 

Icacinaceae 26 lendemilahy 

2 rotry mena 

Syzygium sp.2 

Myrtaceae 27 falindandro 

3 raotry 


Arecaceae 28 meramaintso Sarcolaena multiflora 

4 fandranabo 

Pandanus aff. longistylus Pandanaceae 29 ampoly 

Vepris eliotii 

5 ramirisa 

Homalium louvelianum 

Flacourtiaceae 30 ramy 

Canarium boivinii 

6 fandrikatany Smilax anceps 

Smilaceae 31 memboloa 

Rinorea pauciflora 

7 fandrianakanga Polycardia phyllanthoides Celastraceae 32 fantsikaitrafotsy Pyrostria sp. 

8 akondronala Ophiocolea delphinensis Bignoniaceae 33 bemalemy 

Bembicia uniflora 

9 amboralahy 

Tambourissa purpurea 

Monimiaceae 34 FT62 

Tarenna thouarsiana 

10 kambatrikambatri Brexia sp. 

Grossulariaceae 35 sisikandrongo Poupartia chapelieri 

11 x219 

Erythroxylum sp. 

Erythroxylaceae 36 fangora sp2 Erythroxylum nitidilum 

12 vahilengo 

Morinda umbellata 

Rubiaceae 37 tavolohazo novembre Cryptocarya sp. 

13 ropasy sp.2 

Eugenia sp. 

Myrtaceae 38 voapaky vavy Uapaca littoralis 

14 x202 

Pittosporum polyspermum Pittosporaceae 39 zoramena 


15 fotsivavo 

Polyalthia madagascariensis Annonaceae 40 mocarana 

Macaranga perrieri 

16 hazomamy marécage Apodytes dimidiata 

Icacinaceae 41 tsilaka 

Myrica spathulata 

17 fanolafotsy 

Asteropeia multiflora 

Theaceae 42 ? 

Canthium sp. 

18 vahabatra 3e M Cinnamosma madagascariensis Canellaceae 43 tsilavimbinanto Rhopalocarpus coriaceus 

19 x201 

Elaeocarpus alnifolius 

Elaeocarpaceae 44 mampay 

Cynometra cf. cloiselii 

20 nofotrako 

Vitex chrysomallium 

Verbenaceae 45 voavoantatsimo Mystrolylon aethiopicum 

21 fotombavy 

Leptolaena sp. 

Sarcolaenaceae 46 hazongalala 

22 tainbarika 

genus indet. 

Rubiaceae 47 hazombato 

23 x225 

Suregada baronii 

Euphorbiaceae 48 fantsikaitramainty Canthium variistipula 

24 fitoravina 

Vepris fitoravina 

Rutaceae 49 tavolo 

Tacca leontopetaloides 

25 voatsilana 

Schefflera rainaliana 

Araliaceae 50 hazomainty 

Dyospyros sp.2 

Each number represent a certain fruit species ('a' refers to the entire fruit and 'b' to the seed). One exception is 4b which referes to a fruitlet.

Fruit-frugivore interactions in a Malagasy littoral forest - Universiteit ...

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?