BEHAVIOURAL ECOLOGY OF THE SILVER LEAF MONKEY, Trachypithecus auratus sondaicus, IN THE PANGANDARAN NATURE RESERVE, WEST JAVA, INDONESIA

by

KAREN MARGARETHA KOOL

A dissertation submitted to the University of New South Wales for the degree of Doctor of Philosophy

School of Biological Science University of New South Wales Sydney, New South Wales Australia March, 1989 SR PT02 Form 2 RETENTION

THE UNIVERSITY OF NEW SOUTH WALES

DECLARATION RELATING TO DISPOSITION OF PROJECT REPORT/THESIS

This is to certify that I . .1�9.-.C:eo...... MQ;.r.tQ.J:£'\hP..-. .. .J�l.... being a candidate for the degree of 1)!=>.d:o.r.--.of.-.Y.hSos.Of.hy .. am fully aware of the policy of the University relating to the retention and use of higherdegree project reports and theses, namely that the University retains the copies submitted for examination and is free to allow them to be consulted or borrowed. Subject to the provisions of the Copyright Act, 1968, the University may issue a project report or thesis in whole or in part, in photostat or microfilm or other copying medium.

In the light of these provisions I declare that I wish to retain my full privileges of copyright and request that neither the whole nor any portion of my project report/thesis be published by the University Librarian and that the Librarian may not authorise the publication of the whole or any part of it, and that I further declare that thispreservation of my copyright privileges shall lapse from the ... :fit.�t ...... day of ...3.o.n ...... 9..r: { ...... 19.�.<:\...... unless it shall previously have been extended or revoked in writing over my hand.

I also authorise the publication by University Microfilms of a 350 word abstract in Dissertation Abstracts International ( applicable to doctorates only).

Signature ... ::.:.,.,.,-· ...... Witness......

Date .... :1-.o... M.<\£.ch.... i.. <). i.9 ...... Trachypithecus auratus sondaicus iii

CERTIFICATE OF ORIGINALITY

I hereby declare that this thesis is my own work and that, to the best of my knowledge and belief, it contains no material previously published or written by another _person nor material which to a substantial extent has been accepted for the award of any other degree or diploma of a university or other institute of higher learning, except where due acknowledgement is made in the text of the thesis.

Karen Kool

March, 1989 iv

ACKNOWLEDGEMENTS

This study owes its inception to two people. Dr J. MacKinnon suggested the silver leaf monkey Trachypithecus auratus as a suitable species for study and recommended Pangandaran as a study area. Dr D.B. Croft indicated a willingness (where others hesitated) to supervise a student planning to conduct research overseas. I am extremely grateful to both these people for encouraging me to pursue a long-held ambition. Dr C.P. Groves also provided invaluable advice and encouragement at the outset and provided me with an extensive reading list so that I could become acquainted with primate ecology. Dr Y. Ruhiyat showed me the first T. auratus I was to see.

Lembaga llmu Pengetahuan Indonesia (Indonesian Institute of Sciences) and Perlindungan dan Pelestarian Alam (P PA) branch of the Forestry Department were my sponsors in Indonesia. I thank these institutions for providing the necessary permission to undertake research in Indonesia. I particularly wish to thank Dr Rubini Atmawidjaja, Mr Napitapulu, Mrs Moertini Atmowidjojo, Mr Ubus Wadju Maskar and the P P A staff at Pangandaran.

I am very grateful to Dr J. MacKinnon and Dr H. Rijksen as my overseas supervisors.

Numerous people assisted me in the field. Foremost was Mr R. John Fowler who gave so much of his time and energy during the two years in Indonesia. John Fowler was invaluable for his assistance with mapping the home-range areas of the study groups. He also assisted in the initial surveys of T. auratus groups in the Pangandaran Nature Reserve, in conducting censuses of T. auratus, in estimating tree heights, in quadrat sampling of the vegetation, in focal-animal sampling (as a "scribe") and in setting up a computer database. Others who assisted me in the field were Dr R. J aremovic; Mr R. Hill, Dr D.B. Croft, Kuswan, Ms H. Kool, Ms M. McCoy and Ms M. Crouch. My father, Mr A.F. Kool, deserves a special mention for his assistance, particularly for his forbearance and humour during the night observations. Uce assisted greatly in the vegetation sampling Acknowledgements v

and, as well as being a superb tree climber, Uce's sense of humour made this task enjoyable. My sincere gratitude goes to all these people. I also wish to thank Mr and Mrs P. Hillegers, Dr F. Smiet, Ms S. Berg, and Mr and Mrs E. de Haas for logistical support in Bogor and Jakarta.

Meteorological data were collected daily by the P PA staff at Pangandaran. In particular I wish to thank Mr Deden for his conscientious efforts. I am also grateful to staff at Pemeriksaan Hujan, Pangandaran and Kantor Meteorologi, Cilacap for making their meteorological records available to me.

Plant identification was performed at the Bogor Herbarium with the assistance of the herbarium staff. Dr R. Geesink and Dr E.F. Vogel (from the Rijksherbarium) assisted with the classification of a few botanical specimens. Dr Kostermans assisted with the identification of Cassia species and Prof. Corner with the identification of Ficus samples. Mr Sandot (P P A, Pangandaran) shared with me his knowledge on the uses of plants at Pangandaran.

Analyses of soil samples were conducted under the guidance of Dr M.D. Melville and Mr C.A. Meyers (Geography Department, The University of New South Wales). Chemical analyses of plant samples were performed under the supervision of Dr S. Cork at the Commonwealth Scientific and Industrial Research Organization laboratories, Canberra, and advice on methods was provided by Dr P.J. Waterman. I am extremely grateful to Dr S. Cork for his assistance and co-operation in allowing me to use his facilities for these analyses. Mr R. Engel (Wool and Pastoral Sciences, The University of New South Wales) analysed samples for nitrogen content. Dr A. Bagnara and Tatiana (Biochemistry Department, The University of New South Wales) assisted with labelling bovine serum albumen with Remazol brilliant blue R for the protein precipitation assay.

I have benefitted greatly from verbal and/or written discussions of ideas with Drs D.B. Croft, E.L. Bennett, A.G. Davies, C.P. Groves, P.J. Acknowledgements vi

Waterman, S. Cork, J. MacKinnon, K. MacKinnon, H. Rijksen, B. van Lavieren, T. Struhsaker, A.J. Whitten, R. Jaremovic, B. Wannan, G. Belovsky, B. Fox, D. Woodside, M. Elgar, Mr P. Hillegers, Prof. T.J. Dawson, Prof. I. Hume, Assoc. Prof. R. Crozier, Mr D. Maitland, Ms G. da Silva, Mr T. Clancy and Ms N. Marlow. My thanks go to all these people. Liz Bennett has been remarkable for her encouragement, support and diligence in reading drafts of a few chapters for which I am especially grateful. My thanks go to Dr D.B. Croft, Prof. D.C. Sandeman and Mr A.F. Kool for attending to administrative issues in my absence.

Computing advice and assistance was provided by Dr D.B. Croft, Mr R. Griffiths and Mr G. Nippard. Mr P. Gillespie entered the focal­ animal data on to a computer database and constructed the maps presented in this thesis. Ms L. Hamence typed the reference list and Ms P. Green typed the tables. Ms L. Kool created the drawing presented as the frontispiece - an excellent image of the adult male m one of the study groups. The photographs were produced by the faculty photographic unit, The University of New South Wales.

The project was funded by a Commonwealth Postgraduate Research A ward. Additional funding was provided by the Joyce W. Vickery Scientific Research Fund (Linnean Society of New South Wales) and by Mr A.F. Kool.

Finally, I wish to thank Ibu and Bapak Herlan, Cucu, Lia and Pera for being a family to me and for their hospitality and kindness throughout my stay in Pangandaran. The friends and acquaintances in Pangandaran are not forgotten. My father and (late) mother, Mrs Emma W. Kool, are especially appreciated and thanked for their understanding, encouragement and support throughout my studies. vii

CONTENTS

PAGE

Frontispiece 11 Certificate of Originality iii Acknowledgements iv Contents vii List of Figures xii List of Tables xvi ABSTRACT XX

CHAPTER 1: INTRODUCTION

1.1 Colobine monkeys 1

1.2 The leaf monkeys (or langurs) of Asia 2 1.2.1 General 2 1.2.2 The colobines of Java 2

1.3 Studies of colobine ecology 5

1.4 Optimal foraging theory 7

1.5 Aims of the study 12

1.6 Summary 13

CHAPTER 2: METHODS

2.1 Introduction 15

2.2 The study area 15 2.2.1 Location, history and general topography 15 2.2.2 Selection of the study area 18 2.2.3 Climatic conditions 19 2.2.4 Edaphic conditions 23 2.2.4.1 General 23 2.2 .4 .2 Analysis of soil samples 24 2.2.5 Mammals found in the Pangandaran Nature Reserve 27

2.3 The T. auratus study groups 27 2.3.1 Selection of T. auratus study groups 27 2.3.2 Mapping of home-range areas 31 2.3.3 Habituation of T. auratus study groups 31

2.4 Monthly schedule 32 2.4.1 General 32 2.4.2 Scan sampling 35 2.4.3 Focal-animal sampling 38 2.4.4 Remainder of each sampling month 39

2.5 Statistical analysis 40 2.5 .1 Weighting of data 40 2.5.2 Statistical analysis 41

2.6 Summary 41 Contents viii

CHAPTER 3: VEGETATION OF THE STUDY AREA

3.1 Introduction 44

3.2 Methods 44 3.2.1 Botanical plots 44 3.2.2 Plant part production 50 3.2.3 Phytochemistry 53

3.3 Botanical structure and composltwn 56 3.3.1 Structure of the GRP21 and GRP3 home-range areas 56 3.3.2 Botanical composition of the GRP21 and GRP3 home- range areas 65 3.3.2.1 Comparison between the GRP21 and GRP3 home­ range areas 65 3.3.2.2 Cluster analysis 70 3.3.3 Comparison with other areas in South-east Asia 72

3.4 Plant part production 75

3.5 Phytochemistry 78 3.5 .1 Phytochemistry of mature leaves in the two study sites 78 3.5.2 Comparison with other areas in Asia and Africa where colobines have been studied 81

3.6 Summary 83

CHAPTER 4: GROUP SIZE, POPULATION DENSITY, BIOMASS AND SOCIAL ORGANIZATION

4.1 Introduction 85

4.2 Methods 92 4.2.1 CA censuses 92 4.2.2 TW censuses 94 4.2.3 Social organization of T. auratus 95

4.3 Population density and biomass of T. auratus in the Pangandaran Nature Reserve 95 4.3.1 Population density and biomass in the CA 95 4.3.2 Population density and biomass m the TW 97 4.3.3 Evaluation of methods 98 4.3.4 Comparison between CA and TW 98

4.4 Social organization 100 4.4.1 Group composition and behaviour of bisexual groups 100 4.4.1.1 Composition of T. auratus groups 100 4.4.1.2 Study groups 100 4.4.1.3 Reasons for one or more adult males m bisexual groups 103 4.4.1.4 Group composition of other colobines 103 4.4 .1.5 Adult males 104 4.4.1.6 Adult females 105 4.4.1.7 Infants 107 Contents ix

4.4.2 Solitary males, all male and predominantly male groups 108 4.4.2.1 General 108 4.4.2.2 Trachypithecus auratus 109 4.4.2.3 Other colobines 110 4.4.3 Infanticide 110 4.4.4 Sexual dimorphism 111

4.5 Discussion 112

4.6 Summary 116

CHAPTER 5: ACTIVITY BUDGETS

5.1 Introduction 118

5.2 Methods 118

5.3 Activity budgets of GRP3 and GRP21 119 5.3.1 Comparisons between GRP3 and GRP21 119 5.3.2 Monthly variation in activity budgets 122

5.4 Activity budgets of other groups in the Taman Wisata 123

5.5 Comparison between focal-animal sampling and scan sampling methods 123

5.6 Activity budgets according to age/sex class 125

5.7 Diurnal variation in activity 128 5.7.1 Are T. auratus strictly diurnal? 128 5. 7.2 Diurnal variation in activity 129

5.8 Comparison with activity budgets of other colobines 131

5.9 Discussion 133

5.10 Summary 134

CHAPTER 6: FEEDING BEHAVIOUR

6.1 Introduction 136

6.2 The digestive system of colobines 136

6.3 Methods 138

6.4 Food selection 139 6.4.1 Species 139 6.4.1.1 Comparison of species eaten by GRP21 and GRP3 139 6.4.1.2 Selection ratios 149 6.4.2 Food items 155 6.4.3 Food parts 160 6.4.4 Minor food sources 164

6.5 Monthly variation in feeding 165

6.6 Dietetic diversity 178 Contents X

6.7 Diurnal variation in feeding with respect to items in diet 184

6.8 Discussion 185

6.9 Summary 191

CHAPTER 7: FOOD SELECTION IN RELATION TO PHYTOCHEMISTRY

7.1 Introduction 194

7.2 Methods 196

7.3 Plant chemicals influencing food selection by T. auratus 196

7.4 Discussion 198 7 .4.1 Feeding strategies - comparison between T. auratus and other colobines 198 7.4.2 Role of tannins and phenolics in food selection 202 7.4.3 Optimization 204

7.5 Summary 207

CHAPTER 8: HOME-RANGE AREA AND RANGING BEHAVIOUR

8.1 Introduction 208

8.2 Methods 209

8.3 Home-range area 210 8.3.1 Estimates of home-range area 210 8.3.2 Home-range overlap 212 8.3.3 Number of 0.25 ha quadrats entered daily 217 8.3.4 Number of 0.25 ha quadrats entered monthly 217

8.4 Day-range length 219 8.4.1 Daily range use 219 8.4.2 Monthly range use 221

8.5 Correlates of ranging behaviour and quadrat use 222 8.5.1 Diet 222 8.5.1.1 Major food species 222 8.5.1.2 Proportion of different items in diet 224 8.5.2 Night trees 227 8.5.3 Rainfall 227 8.5.4 Defence of home-range area 228 8.5 .5 Discussion of the determinants of ranging behaviour by T. auratus 230

8.6 Discussion 230 8.6.1 Comparison between GRP3 and GRP21 230 8.6.2 Comparison with other colobines 232 8.6.2 .1 Estimates of home-range area 232 8.6.2.2 Correlates of ranging behaviour, home-range size and territoriality 233

8.7 Summary 236 Contents xi

CHAPTER 9: CONCLUDING DISCUSSION 238

REFERENCES 243

APPENDIX I Stem basal area as an indicator of foliage biomass 269

APPENDIX II Vegetation species with authorities 272

APPENDIX III Species enumerated in the vegetation sampling of the GRP3 and GRP21 home-range areas with corresponding frequency (number of stems per hectare) and biomass (percentage basal area (b.a.)) 277

APPENDIX IV Chemical measures (as % dry weight) of vegetation in the GRP3 and GRP21 home­ range areas 282

APPENDIXV Breakdown of Trachypithecus auratus diet (as unweighted % feeding time) to the level of food part 285 xii

LIST OF FIGURES

FIGURE PAGE

1.1 Sonograph of the loud call of an adult male Trachypithecus auratus sondaicus 4

2.1 Location of study area in West Java, Indonesia 16

2.2 The Pangandaran Nature Reserve 17

2.3 Rainfall as monthly means and ranges (1980-1985) - taken at two nearby locations outside the Nature Reserve 19

2.4 Monthly rainfall (August 1984 - November 1985) at the Pangandaran Nature Reserve, Pangandaran and Cilacap 21

2.5 Per cent humidity (0700 h and 1500 h) and mmxmum and maximum temperature as daily average each month. Readings taken at the Pangandaran Nature Reserve 22

2.6 Mean daily mmtmum and maximum temperature each month (1980-1985) - readings taken at Cilacap 22

2.7 Mean daily per cent humidity each month (1980- 1985) - readings taken at Cilacap 23

2.8 Approximate locations of the home-range areas of the two study groups 30

3 .1 Location of vegetation quadrats m the GRP21 home-range area 46

3.2 Location of vegetation quadrats m the GRP3 home-range area 47

3.3 Species/area curve for vegetation in the GRP21 home-range area 49

3.4 Species/area curve for vegetation in the GRP3 home-range area 50 List of figures xiii

3 .5 Standard curves for gallic and tannic acid 55

3.6 Frequency distribution of trees (2 5 m height) in 10 em girth size classes in the home-range areas of a) GRP21 b) GRP3 57

3. 7 Frequency distribution of trees (2 5 m height) in 10 em girth size classes in the GRP3 home-range area a) non-plantation b) plantation 59

3. 8 Frequency distribution of trees m 5 m height classes in the home-range areas of a) GRP21 b) GRP3 61

3. 9 Frequency distribution of trees iri 5 m height classes in the GRP3 home-range area a) non-plantation b) plantation 62

3 .1 0 The proportion of trees (2 5 m height) in 10 em girth size classes bearing Hanas a) GRP21 b) GRP3 63

3.11 The proportion of trees (2 5 m height) in 10 em girth size classes in the GRP3 home-range area bearing lianas a) non-plantation b) plantation 64

3.12 Hierarchical cluster analysis (Norusis 1985) using average linkage between groups according to presence/absence of species in 10x10 m quadrats in the GRP3 home-range area 7 1

3 .13 Proportion of trees which bore young leaves, mature leaves, fruits and flowers in the phenology sample for the GRP21 home-range area 76

3 .14 Proportion of trees which bore young leaves, mature leaves, fruits and flowers in the phenology sample for the GRp3 home-range area 7 7

4.1 Six kilometre transect used for censusing Trachypithecus auratus in the Cagar Alam 93

4.2 Locations of large Trachypithecus auratus groups in the Taman Wisata 9 9

5.1 Activity budgets: GRP3 and GRP21 120 List of figures xiv

5.2 Proportion of time (as unweighted percentage) spent in major activities (median daily percentage and I Q range) 121

5.3 Activity budgets of each age/sex class m GRP3 (median and IQ range) 126

5.4 Proportions of time (as %) spent in major activities each hour 130

6.1 Proportions of vanous items m the total diet 157

6.2 Proportions of various items in the monthly diet (as unweighted % feeding time) 166

6.3 Item diversity (using Shannon-Weaver index) and the proportion (as % feeding time) of fruit in the GRP3 diet 181

6.4 Item diversity (using Shannon-Weaver index) and the proportion (as % feeding time) of fruit and flowers in the GRP3 diet 181

7.1 Mean content (as percentage dry weight ( + SD)) of protein (PROT), fibre (ADF), condensed tannins (CT), total phenolics (TP); % digestibility (CDIG); and protein precipitating capacity (PP) in leaves and fruits available to Trachypithecus au rat us (G RP21 and GRP3 combined) at Pangandaran 197

8.1 Cumulative number of 0.25 ha quadrats entered by GRP21 and GRP3 throughout the study 211

8.2 MAP(95) utilization distribution for GRP3 (smoothed by the Fourier transform method) 213

8.3 MAP(95) utilization distribution for GRP21 (smoothed by the Fourier transform method) 214

8.4 Two-dimensional representation of home-range use by GRP3 (Fourier transform method) 215

8.5 Two-dimensional representation of home-range use by GRP21 (Fourier transform method) 216 List of figures • XV

8.6 Frequency distribution of the number of 0.25 ha quadrats entered each day by GRP21 and GRP3 throughout the study 21 8

8. 7 Monthly number of 0.25 ha quadrats entered each day by GRP21 and GRP3 (mean + SD) 21 9

8.8 Frequency distribution of day-range lengths of GRP21 and GRP3 220

8.9 Mean (+ SD) monthly day-range lengths (m) of GRP21 and GRP3 221

8.10 Home-range use by GRP3 in November 1984 (Fourier transform method) 225

9.1 Weighted mean protein/fibre levels in mature foliage against colobine biomass for six study areas 240 xvi

LIST OF TABLES

TABLE PAGE

2.1 Soil analysis 26

2.2 Trachypithecus auratus age/sex classes and descriptions 28

2.3 Descriptions of T. auratus activities 33

2.4 Time and weather criteria used during monthly scan samples 38

3.1 Species sampled for observation of plant part production 52

3.2 Families sampled in the GRP21 and GRP3 home­ range areas with corresponding biomass (as basal area (%)) and abundance (as stem density (%)) 66

3.3 Fifteen species with highest biomass at each study site 68

3 .4 Frequency of species which accounted for l 5% each of all trees enumerated (n) in the vegetation quadrats 69

3.5 Proportion (as percentage) of total basal area contributed by the ten families with the highest biomass at Sepilok, Kuala Lompat and Pangandaran study areas 7 4

3.6 Phytochemical analyses of mature leaves m the home-range areas of GRP3 and GRP21 79

3. 7 Weighted values for chemical measures of vegetation (mature leaves) in the two home-range areas 80

3.8 Weighted phytochemical data for mature leaves at six study areas 82

4.1 Age/sex composition of groups m the Taman Wisata (April 1984) 101 List of tables xvii

4.2 Mean group size of colobines 113

4.3 Colobine body weights and biomass at six study sites 114

5.1 Proportion of time (as sum of proportion of each scan) spent in the major activities by GRP3 and GRP21 (overlapping months only i.e. five months for each study group) 121

5.2 Monthly variation in activity budgets (for major activities) usmg Kruskal-Wallis one-way anova 122

5.3 Proportion of time (as sum of proportion of each scan) spent in major activities by GRP3 and two other T. auratus groups located in the Taman Wisata 123

5.4 Time (min) spent in the major activities by each of the age/sex classes (excluding infants) in GRP3, using the focal-animal sampling method 125

5. 5 Results of Normal Scores test on proportion of time spent in each of the major activities by the vanous age/sex classes 127

5. 6 Proportion of time (as sum of proportion of each scan) spent in major activities by GRP3 during the October monthly scan sample (Day 1 - Day 5) and on the day (D-N) following the night during which observations of activity were made 12 9

5.7 Colobine activity budgets (as mean %) 132

6.1 Food items and parts determined with abbreviations, and diagnostic features where appropriate 140

6.2 Food tree species of GRP3 and GRP21 141

6.3 The ten most abundant species (as % basal area) and corresponding stem density (as frequency per ha) within the study sites, and the proportion of T. auratus diet (as % feeding time) contributed by these species 1 4 5 List of tables xviii

6.4 Comparison between basal areas (b.a.) of food tree species and non-food tree species 145

6.5 Abundance (as % basal area) of 12 of the top 15 food tree species for GRP21 which did not rank in the top 15 food tree species for GRP3 14 7

6.6 Abundance (as % basal area) of 12 of the top 15 food tree species for GRP3 which did not rank in the top 15 food tree species for GRP21 148

6. 7 Families of food trees eaten by GRP3 and GRP21 15 0

6. 8 U nweighted % feeding time and weighted % feeding time compared for top twenty food species for a) GRP3 b) GRP21 151

6. 9 Selection ratios (S.R.) for top twenty food tree species 153

6.10 Items of species contributing 2 one per cent to the total diet (as unweighted % feeding time) a) GRP3 b) GRP21 155

6.11 Comparison between proportion of items m the daily diet of GRP3 and GRP21 158

6.12 Proportion (Percentage of feeding time (unweighted cf. weighted)) of items in the diet 159

6.13 Proportion of parts of items in the diet (as unweighted % feeding time) 161

6.14 Comparison between months of high and low dietary intake of fruits, and of young leaves and leaf buds 167

6.15 Species items which each contributed more than five per cent to a monthly diet and proportion of fruit, flowers, and young leaves in each monthly diet 169

6.16 Correlation between dietetic diversity (as species diversity or item diversity), using the Shannon-Weaver information index, and the List of tables · xix

proportion of various items in the diet (as % feeding time) of GRP21 and GRP3 180

6.17 Dietetic diversity (as species diversity) and home-range diversity (as species diversity determined from quadrat sampling of vegetation) for GRP3 and GRP21 183

6.18 The diet of twelve colobine species (percentage feeding time) 189

7.1 Chemical measures for Tectona grandis leaves and contribution to diet (of different parts) 201

8.1 Possible correlates of day-range length and number of quadrats entered each day by GRP3 (n = 40) and GRP21 (n = 25) (Spearman rank) 226

I .1 Correlations between basal area and canopy measures for species with a sample of trees l 5 2 7 0 XX

ABSTRACT

The ecology of the silver leaf monkey, Trachypithecus auratus sondaicus, was studied in the Pangandaran Nature Reserve, West Java, Indonesia. The Reserve comprises two sections: the Cagar Alam (CA) (Nature Reserve proper) consisting largely of secondary forest; the Taman Wisata (TW) (public-use zone) consisting of mixed secondary forest and plantation areas (teak, mahogany and acacia). Two groups of similar age/sex class composition were selected for detailed study; one in the CA (GRP21) and one in the TW (GRP3). Their behavioural ecology was compared with respect to the structure, floristics and phytochemistry of the habitats.

Trachypithecus auratus formed groups of 6-21 individuals with one or two adult males. Home-range areas were actively defended. Major food items were young leaves and fruits. Several species were asynchronous in production of reproductive parts and young leaves such that food was generally abundant. Food items tended to be lower in fibre content and were more digestible than vegetation which was not eaten. Phenolics and condensed tannins were not a major determinant in food selection. The protein/ ADF ratio did not explain selection of specific parts of food items by T. auratus nor was the difference in biomass of T. auratus in the CA and TW attributable to the ratio of these measures in mature foliage.

GRP21 had a larger home-range area, travelled further in a day, had a higher dietary intake of fruit and spent more time travelling and resting than GRP3. GRP3 spent more time feeding than GRP21. Day-range length was positively correlated with the proportion of fruit in the diet for both groups and negatively correlated with the proportion of leaves in the GRP3 diet.

Young teak (Tectona grandis) leaves were the single most common food item for GRP3. Teak trees were distributed in large, relatively homogeneous patches. By comparison, major dietary items of GRP21 were distributed in smaller, widely spaced, patches. xxi

The TW supported a higher biomass of T. auratus distributed in larger groups than in the CA. In plantation areas, teak leaves (not a highly preferred dietary item) provided a constant source of food which was exploited when preferred items were scarce. Chapter 1: Introduction 1

CHAPTER 1 INTRODUCTION

1.1 COLOBINE MONKEYS

The Old World monkeys, the Cercopithecidae, are divided into two sub-families, the Colobinae and the Cercopithecinae (Napier and Napier 1967). Originally from Africa, members of both sub-families had radiated and evolved into a variety of forms throughout South­ east Asia by the Middle Pleistocene (Medway 1970a).

The Colobinae are characterized by:

1. Large, sacculated stomachs (Hill 1968) with a diverse array of microflora which ferment food (Bauchop and Martucci 1968). A reticular groove along the dorsal region of the fermentation chambers of the stomach channels milk past the site of microbial activity to the pyloric section of the stomach enabling efficient digestion of milk in young animals (Black and Sharkey 1970);

2. Cheek teeth with increased relief (cf. Cercopithecinae) (Kay and Rylander 1978);

3. Reduced first digit on the forelimb (Napier and Napier 1967);

4. Long-distance male vocalisations (Napier and Napier 1967);

5. Flamboyant natal coat (Napier and Napier 1967);

6. Postnatal infant transfer (Napier and Napier 1967).

Included in the Colobinae are the Colobus and Procolobus monkeys of Africa (Groves 1970; Kingdon 1971 ), the leaf monkeys (Presbytis and Trachypithecus) and the odd-nosed monkeys (Nasalis, Pygathrix, Rhinopithecus) of Asia (Groves 1970; Medway 1970a). The Procolobus traditionally included only the olive colobus, and the Colobus included the black-and-white species along with the red colobus species (Groves 1970; Thorington and Groves Chapter 1: Introduction 2

1970). Kuhn (1972) placed the red colobus species with the olive colobus in Procolobus, retaining the Colobus genus for the black­ and-white species only. The former classification is used in this thesis although it is acknowledged that the latter classification is regammg recognition (e.g. Whitesides et al. 1988; C.P. Groves pers. comm.).

1.2 THE LEAF MONKEYS (OR LANGURS) OF ASIA

1.2.1 General

The leaf monkeys of Asia have formerly been regarded as two taxa, Presbytis sensu stricto and Trachypithecus Reichenbach 1862 (Medway 1970a) at the specific level, sub-generic or generic level (Pocock 1935, 1939; Washburn 1944; Hooijer 1962). They are now generally accepted as two distict genera (Weitzel and Groves 1985) and are regarded as such in this study.

The Presbytis genus includes species which were formerly frequently referred to as belonging to the "aygula-melalophos" group such asP. rubicunda, P. melalophos, P. entellus, P. aygula, P. thomasi, P. frontata and P. hose i. Trachypithecus includes the leaf monkeys which were formerly regarded as the "c ristatus" group such as T. obscura, T. johnii, T. senex and T. cristata (Pocock 1935; Hill 1939; Hooijer 1962; Thorington and Groves 1970).

Trachypithecus spp. have relatively larger stomachs than do Presbytis spp. (Chivers and Hladik 1980). It has been postulated that the relatively larger size of stomachs in Trachypithecus spp. may be a dietary adaptation enabling a greater intake of foliage.

1.2.2 The Colobines of Java

Two colobines are found on Java. Presbytis comata (Desmarest 1822) (Weitzel and Groves 1985), formerly P. aygula, is locally known as the "surili". Presbytis comata is represented in West Java Chapter 1: Introduction 3

asP. c. comata and in Central Java as P. c. fredericae (Weitzel and Groves 1985). Adults have been described as having an iron-grey coloured back, black head and cheeks and white ventral surface (Medway 1970a). Neonates are pale fawn or white (Medway 1970a). Presbytis comata occurs only in high altitude forest (van Bemmel 1978).

The other leaf monkey of Java, Trachypithecus auratus (E. Geoffroy St. Hilaire 1812) (Weitzel and Groves 1985), is locally referred to as the "lutung". This species was formerly regarded as being conspecific with T. cristata of Burma, Thailand, Indo-China, Cambodia, South Vietnam, South Laos, West Malaysia, Bangka, Belitung, the islands of the Riau and Lingga Archipelagos, Serasan Island, Sumatra and Borneo (Napier 1985). Three sub-species are recognized: T. a. auratus from East Java, T. a. sondaicus from West Java and T. a. kohlbruggei from Bali (Weitzel and Groves 1985).

The· pelage of adult T. auratus is black to silvery-black with a red colour morph having been observed coexisting with black T. auratus in East Java (Medway 1970a; van Bemmel 1978). Females are distinguished from adult males by a pale, cream patch in the pelvic region (Roonwal and Mohnot 1977). Infants have an orange/red natal coat. Trachypithecus auratus sondaicus was the study animal of the research reported here; future reference to T. auratus refers to this sub-species unless otherwise indicated.

Trachypithecus auratus sondaicus has a mean head and body length of 571 mm (465-650) and mean tail length of 742 mm (680- 810) for males (Napier 1985). Trachypithecus cristata measure 465-560 mm head and body length, have a tail length of 686 mm and males weigh 7.1 kg (female weight is approximately 89% of male weight) (Roonwal and Mohnot 1977). Thus, T. auratus are somewhat larger than T. cristata (Napier 1985; Weitzel and Groves 1985). Mean weights for T. auratus are not available but presumably individuals of this species weigh slightly more than T. cristata individuals. Chapter 1: Introduction 4

Adult male loud calls have, in the past, been chosen as a characteristic on which to differentiate species (Struhsaker 1970; Wilson and Wilson 1975; Oates and Trocco 1983). Gibson-Hill (1949) gave a descriptive account of. T. cristata calls. The adult male loud call was described as a call which "one can only describe as a "bray" slightly reminiscent of a donkey but shorter, harsher and more gutteral and, of course, less in volume" (Gibson-Hill 1949 p. 49). Furuya (1961-2) also described the loud call of the adult male.

A sonograph of the loud call of the adult male in one of the T. auratus study groups is shown in Figure 1.1. Davies (1984) presented a sonograph of a P. rubicunda adult male loud call at Sepilok and stated that initial nasal squeaks were followed by a longer rattled coda. Presbytis melalophos adult males at Kuala Lompat had a similar call to P. rubicunda although they had a longer coda and staccato "drumming" prior to the nasal squeaks (Davies 1984). The loud call of the adult male T. auratus is noticeably different to that of P. rubicunda; the T. auratus call is considerably shorter in duration and lacks the initial nasal squeaks (although three squeaks occur mid-way during the call).

- 8

freq (KHz)

duration of call = 0.6 s Figure 1.1 Sonograph of the loud call of an adult male Trachypithecus auratus sondaicus Chapter 1: Introduction 5

Trachypithecus auratus are widespread on Java, ranging from lowlands to high elevations (Medway 1970a), and are found m lowland swamp forest and mangroves and in secondary forest and Tectona grandis (teak) plantations where the latter border with "natural" forest (van Bemmel 1978; pers. obs.). Trachypithecus auratus is similar in this respect to T. cristata which typically occurs in coastal and riverine habitats (Wilson and Wilson 1976) and in secondary hill forests (Wilson and Wilson 1973) in Sumatra, and in the coastal, mangrove forests of West Malaysia (Furuya 1961-2; Southwick and Cadigan 1972).

1.3 STUDIES OF COLOBINE ECOLOGY

Early studies of colobines examined social organization (reviewed by Hrdy 1980) and attempted to relate similarities and differences in social structure of various species to taxonomic relatedness. This gave rise to studies of an ecological nature (for example, Hladik and Hladik 1972; Clutton-Brock 1974a; Struhsaker 1975; Curtin 1976). The importance of dietary requirements on social structure was examined and, as well as comparing species of colobines, involved comparisons of colobines (compound-stomached primates) with simple-stomached primates (for example, gibbons). Models were formulated in an attempt to relate patterns of ranging behaviour and social organization to the distribution of food (Crook and Gartlan 1966; Eisenberg et al. 1972; Clutton-Brock and Harvey 1977a). Differences in the degree of folivory and frugivory were regarded as being of vital importance in that this factor was likely to reflect variation in digestive physiology. Furthermore, a frugivorous diet may reflect a dependence on a seasonal food source whereas a folivorous diet may show less dependence on seasonal availability of food. This in turn may affect primate density and abundance (Eisenberg et al. 1972).

Degree of folivory alone did not adequately explain differences m social structure and biomass between simple-stomached and compound-stomached primates (Bennett 1983 ). Consequently, the possible effects on diet selection of other factors important in Chapter 1: Introduction 6

gastrointestinal physiology, were examined. In particular, the detoxification capacity of the colobine foregut fermentation chamber may permit colobines to eat more toxic foods than monogastric primates. This may affect selection of fruits and flowers as well as leaves (Hladik 1977; Clutton-Brock and Harvey 1977b; McKey 1979).

Recent long-term studies of colobines have examined the phytochemistry of vegetation in the habitat in addition to observing the feeding and ranging behaviour of primates (Hladik 1977; Oates et al. 1977; Gartlan et al. 1978; McKey 1978; Oates et al. 1980; Wrangham and Waterman 1980; Bennett 1983; Davies 1984; Davies et al. 1988). This has provided a greater understanding of interspecific similarities and differences in feeding ecology and helped to elucidate which factors are due to taxonomic (physiological) relatedness and which reflect environmental variation. Two intensive studies of colobines in South-east Asia have been conducted. Bennett (1983) studied P. melalophos, in the Krau Game Reserve, West Malaysia and Davies (1984) studied P. rubicunda in Sepilok, Sabah.

The ecology of T. obscura has been studied in West Malaysia (Curtin and Chivers 1978; Curtin 1980; MacKinnon and MacKinnon 1980), however, these studies did not include an examination of diet with respect to phytochemistry. The study reported here is, therefore, the first detailed ecological study, incorporating an analysis of phytochemistry with respect to diet selection, of a Trachypithecus sp. in South-east Asia. Two other studies of Trachypithecus spp. have been conducted: Oates et al. (1980) performed an ecological study of T. johnii in Kakachi, Southern India; Hladik (1977) studied T. senex (and P. entellus) in Sri Lanka. One other long-term ecological study of a colobine in Indonesia has been conducted. Ruhiyat (1983) studied P. comata in the submontane forests of Kamojang, West Java but this study did not include a phytochemical component. Chapter 1: Introduction 7

1.4 OPTIMAL FORAGING THEORY

Optimal foraging theory (OFf) began in 1966 when MacArthur and Pianka, and Emlen in two separate papers published in the American Naturalist, proposed that "prey selectivity could be understood as driven by a tendency, selected through evolution, to maximize net energy gained per unit time feeding" (Schoener 1987 p. 5). Prey selectivity has since been taken to include patch choice.

Since its inception, the rate of publications involving optimal foraging escalated in an exponential growth form, reaching a plateau in 1982 (Pyke 1984; Gray 1987; Schoener 1987). The original theory has expanded to include other parameters, such as the depletion of patches rather than the absolute consumption thereof, known as the "compression hypothesis" (Schoener 1987). Variation in the "currency" maximized has also been considered.' For example, an optimal diet may involve a maximization in nitrogen (protein) intake (Owen-Smith and N ovellie 1982) rather than energy.

Furthermore, constraints have been taken into consideration which may be important under situations where resources are limited. Schoener (1969, 1971) coined the terms "time mininizer" and "energy maximizer" where the former involves mininizing the time required to gain a fixed energy requirement and the latter involves gathering as much energy as possible within a fixed time constraint; both maximize energy per time spent feeding. This dichotomy has been useful in the construction of alternative diet models (for example, Belovsky 1978).

Nutrients may also pose a constraint m that a diet based on energy maximization will only be optimal if adequate amounts of limiting nutrients are also obtained (Westoby 1974; Altmann 1984; Schoener 1987). Belovsky (1978) developed a linear programming model in which he included nutrients as a constraint. A limitation of this model is that a patch is taken to refer to an area of one food type only. A further constraint included in the Belovsky (1978) Chapter 1: Introduction 8

model is the maximal amount of food that can pass through the gut per unit time.

Freeland and Janzen (1974) considered the effects of secondary compounds on diet selection, proposing that diet diversification may be necessary to ensure that damaging concentrations of certain toxins are avoided. Dietary diversity may also, in itself, aid detoxification of these secondary compounds.

As is discussed further in Chapters 6 and 7, studies of colobine feeding behaviour have preferentially referenced Freeland and Janzen (1974) when discussing diet in terms of selection. Unfortunately, Freeland and Janzen ( 197 4) did not acknowledge convential OFT and consequently several studies of colobine feeding behaviour similarly have not placed their work in the context of OFT nor provided explanations for divergence from conventional OFT. The "currency" considered has usually been protein (although it has generally been considered that energy may be the "currency" in selection of fruits in the diet), and the constraints considered have largely been fibre content and condensed tannins.

An apparent limitation in the colobine studies is the nature and number of constraints taken into consideration. The main reason for this is the complexity of both the primate and the environment in which it lives. Colobines feed extensively on both foliage (leaves and flowers) and fruits, and dietary requirements (and corresponding constraints of different food items) of colobines are, therefore, likely to be considerably more complex than the dietary requirements and intake of the "generalist herbivores" studied by Belovsky (1978).

Linear optimization models assume that nutrients and toxins exert their effects independently and do not, therefore, take into account interactive effects as are known to occur. For example, gut absorption of sugars and amino acids depends on. sodium concentration in the intestine (Altmann and Wagner 1978). Whilst Belovsky's model (1978) accounted for the use of broad classes of vegetation (i.e. aquatic versus terrestrial) in his study on the diet of Chapter 1: Introduction 9

moose, the model failed to account for the use of individual species within the terrestrial habitat (Belovsky 1981). This may have been due to the effects of toxins (Stephens and Krebs 1986).

Stephens and Krebs (1986), in a survey of the literature, reported three views regarding decision, constraint and currency for herbivore diets. They were: rate-maximizing (according to nutrient constraints); avoiding toxins; selecting complementary nutrients to enable a balanced diet as when feeding on abundant low-quality food (Crawley 1983). The former two have already been discussed. Little support for selection of complementary nutrients has been found as herbivores apparently are unable to detect specific nutrients with the exception of sodium and water (and possibly sugar) (Robins 1983; Stephens and Krebs 1986). Deficiencies of other essential nutrients are compensated for by learning to avoid diets with adverse consequences or by learning to increase the dietary intake of particular foods high in certain nutrients (Westoby 1974). Stephens and Krebs (1986) argue that food selection through avoidance learning is unlikely in herbivores as: meals are not discrete entities; meals usually consist of more than one food type; adverse consequences may take a long time to develop whereby it would be difficult to determine which of the previous food items caused the adverse effects; food items may have some adverse consequences but still be an essential dietary component.

Stephens and Krebs (1986) suggested that avoidance of toxins can also be incorporated into a rate-maximizing model where digestion-inhibitors such as tannins and fibre (Harborne 1982) are treated as constraints dependent on stomach capacity and/or passage rate. Similarly, alkaloids (as specific poisons), which may limit the amount of dietary intake for an item, may be regarded as a constraint in their "anti-nutrient" capacity (Stephens and Krebs 1986).

The above forms of toxins can only be treated within conventional rate-maximizing models when "considering assimilated rather than gross intake (which is what should be done in any Chapter 1: Introduction 10

case)" (Stephens and Krebs 1986 p. 123). This is one of the major problems regarding the applicability of traditional OFT to colobine feeding behaviour. As is discussed further in Chapter 7, determination of, for example, tannin content and protein content in vegetation samples does not permit a determination of the amount of protein available to the animal; nor does this provide information on the protein-precipitating capacity (whether dietary protein or "monkey" protein) of the tannins involved which would affect the quantity of protein assimilated.

Other limitations in the application of linear program modelling (Belovsky 1978) to colobine diets are: patches in colobine environments are not readily viewed as consisting of one food type only; extensive feeding studies of captive colobines, including determination of food passage rates, have not been conducted so it has not been possible to include this as a constraint in a model of optimal foraging.

There are, however, valid reasons for exammmg the possible effects of secondary compounds on diet selection of colobines. Vegetation in colobine habitats, particularly in tropical rain forests with often poor soil conditions, has been found to contain considerable proportions of secondary compounds (Janzen 1974). This may function as a defence against predation by folivores (Janzen 1974, 1975; Gartlan et al. 1980).

The precise mode of action of secondary compounds on dietary intake, or on primate physiology, remains unknown (Mole and Waterman 1987). Recent colobine studies (Oates et al. 1980; Choo et al. 1981; McKey et al. 1981; Davies et al. 1988) have, therefore, included a digestibility assay in the phytochemical analyses of vegetation samples. This, in conjunction with determination of protein content, condensed tannin content and fibre content has provided a greater understanding of diet selection in colobines. Similarly, analysis of protein precipitation capacity (see Sections 3.2.3 and 7 .4.2) provides further information on the effect of secondary compounds on diet selection. Chapter 1: Introduction 11

A limitation of static short-term OFfs is that, by their nature, they do not take into account long-term considerations such as the maintenance of a body weight at or above a minimal weight (Schoener 1987). Katz (1974) developed a model based on long­ term optimal foraging and predicted that the proportion of time spent feeding would vary seasonally. The Freeland and Janzen ( 197 4) theory may also be extended to incorporate a seasonal component. When "preferred" food items are less available, dietary diversity may be expected to increase to avoid an intake of large concentrations of particular toxins. In long-term ecological studies of colobines these factors have been considered (see Chapter 6) and relationships found in support of these theories where environments are strongly seasonal (with a couple of exceptions, discussed in Section 6.6).

A major criticism of OFT has been the explanation of deviations from predictions of OFT. For example, sampling, in which a forager appears to gauge the quality of food items (Stephens and Krebs 1986; Schoener 1987) (for example, the stage of maturity of fruits) or patches, may be accountable for any deviation where contrary information is lacking (Schoener 1987). Whilst, in theory, sampling behaviour may be essential for "an animal that is going to come close to foraging optimally but is not prescient" (Schoener 1987 p. 27) the objection is levelled where the concept is used a posteriori (Gould and Lewontin 1979; Lewontin 1979). Certainly, under field conditions where environmental variables, such as the type of food available, are not manipulated and controlled, it is often difficult to determine a priori how an animal would be expected to behave under the hypothesis of optimal foraging. However, this criticism ignores the large number of a priori studies (see Schoener 1987) which have found support for OFf.

Another criticism of OFT has been in the assumption that maximizing "fitness" (reproductive success) results in the maximization of foraging efficiency. Gray (1987) claimed that where food is not a limiting resource there is no reason for animals to maximize net energy uptake or any other "currency". This premise is further considered in Chapters 6 and 7 when comparing Chapter 1: Introduction 12

diet selection by colobines where food resources are scarce, with other colobines found in habitats where food resources are comparatively abundant. It is important to note here that food scarcity may reflect high levels of digestion inhibitors such that much of the vegetation in an area may be unsuitable as food for colobines. Under these conditions, food may be selected to maximize a nutrient such as protein (for example, P. rubicunda in Sepilok (Davies 1984)). Where food is comparatively abundant this may mean that, at certain times of year, there is an abundance of potential food for colobines. At other times of the year food may be comparatively scarce. In other words, seasonal availability of food means that food is limited in certain times of the year and not in others. Under such conditions, food is still regarded as a limiting resource although nutrient intake may not necessarily be maximized throughout a year.

Optimization models generally examme only one design feature at a time and, in so doing, assume that different design aspects are effectively independent. For example, it would be assumed that rate-maximization of food intake would not interfere with finding a mate (Stephens and Krebs 1986). In summary, whilst support for classical OFT has been reported in the literature, there are several complications and unjustified assumptions in the models which render application to colobine feeding behaviour as dubious. With the present state of knowledge it is more appropriate to test the effect of specific components on diet selection; for example, to examine whether items eaten have higher protein to fibre ratios than items not consumed.

1.5 AIMS OF THE STUDY

- To describe the vegetation of the study area at Pangandaran;

- To describe the feeding and rangmg behaviour of T. auratus in order to assess how this relates to the vegetation composition, phenology and phytochemistry of the habitat; to compare these results: a) for two study groups to obtain a measure of Chapter 1: Introduction 13

intraspecific variation m foraging ecology and to assess the importance of plantation species on the diet of a group occurring m a mixed plantation/secondary forest habitat, and b) with studies of other colobines to examine food availability and selection by colobines;

- To discuss the adaptive significance of foraging behaviour of T. auratus at Pangandaran;

- To determine group size and population density of T. auratus in the Pangandaran Nature Reserve; to compare biomass of T. auratus with other colobines in other areas and discuss which features of the habitat are likely to influence colobine biomass.

These aims are pursued as follows: general description of study area and methods used in this study are described in Chapter 2; detailed analysis of vegetation composition, phenology and phytochemistry of the two study sites is presented in Chapter 3; structure ·and composition of T. auratus groups is described in Chapter 4 along with analyses of population density and biomass; activity budgets are described in Chapter 5 and act as a precursor to the study of feeding behaviour (Chapters 6 and 7) and ranging behaviour (Chapter 8). A concluding discussion of the determinants of biomass, feeding and ranging behaviour is given in Chapter 9. This thesis deals largely with arboreal colobines, particularly Colobus, Presbytis and Trachypithecus spp.

1.6 SUMMARY

1. The Cercopithecidae compnses two sub-families, the Colobinae and the Cercopithecinae.

2. The leaf monkeys of Asia consist of two distinct genera, Presbytis and Trachypithecus. Chapter 1: Introduction 14

3. The study animal was T. auratus sondaicus; subsequent reference to T. auratus refers to this sub-species unless otherwise specified.

4. Recent long-term ecological studies of colobines have incorporated an analysis of phytochemistry in an attempt to explain diet selection. The study reported here is the first to include a phytochemical component for a Trachypithecus sp. in South-east Asia; two other studies of Trachypithecus spp. (T. johnii in Kakachi, Southern India, Oates et al. 1980; T. senex in Sri Lanka, Hladik 1977) have been conducted.

5. Optimal foraging theory (OFT) has been the source of much discussion and debate; an outline of OFT is presented and its applicability to colobine feeding studies assessed.

6. The aim of this study was to describe the foraging ecology of two groups of T. auratus in the Pangandaran Nature Reserve, West Java, and to relate this to vegetation parameters in the habitat. This would provide information: on intraspecific variation in foraging ecology and the importance of plantation species on the diet of a group occurring in mixed plantation/secondary forest; for a comparison of diet selection by T. auratus with other colobines; for an assessment of the adaptive significance of foraging behaviour. A further aim was to determine the population density and, thereby, biomass of T. auratus in the Pangandaran Nature Reserve. Habitat variables which may affect colobine biomass could then be assessed. Chapter 2: Methods 15

CHAPTER 2 METHODS

2.1 INTRODUCTION

Fieldwork was conducted from March 1984 to December 1985 inclusive. General methods and the monthly schedule employed are described here. Specific methods used for each component of the study are presented in the appropriate, subsequent chapters.

2.2 THE STUDY AREA

2.2.1 Location, History and General Topography

Pananjung Pangandaran Cagar Alam (subsequently referred 'to as the Nature Reserve) is located on the southern coast of West Java, Indonesia (Fig. 2.1) at latitude 7°43'S longtitude 108°40'E. The elevation above sea level is 2-150 m with an average of about 100 m (Blower et al. 1977). The Nature Reserve is 530 ha in area and is located on a peninsula bounded by the sea on all sides except the north, where the boundary cuts across the narrow isthmus of land (approx. 200 m) which connects the pen-insula with the mainland. Formerly a Game Reserve, it was declared a Nature Reserve in 1961 to protect the abundant Rafflesia patma (Blower et al. 1977), a plant species of botanical importance found only on Java (R. Gee sink, pers. comm.).

The Nature Reserve is divided into two sections: a public use zone, the Taman Wisata (subsequently referred to as TW), which consists of 37.7 ha; the nature reserve proper, the Cagar A lam (subsequently referred to as CA), which constitutes the remainder of the Nature Reserve and which is closed to the general public (Fig. 2.2). An area of about 20 ha within the TW was planted with Tectona grandis (teak) in 1932 and in 1936 and. subsequently cut. In 1957 the Forestry Department again planted the area with T. grandis, Swietenia macrophylla (mahogany) and Acacia Chapter 2: Methods 16

Java Sea

West Java

Indian Ocean Pong ond oro n

N Scale (km) + 0 50 100

Figure 2.1 Location of study area in West Java, Indonesia

auriculiformis, the latter two being exotic species (Blower et al. 1977). These plantations are now well established reaching at least 15 rn in height. Approximate locations of the plantation areas within the TW are shown in Figure 2.2. The CA and other areas in the TW largely contain secondary forest. Chapter 2: Methods 17

Batu Mandi

B.atu. ME'ja

CAGAR

ALAM

Tg. Batu Sodas

Legend

0- Streams Tg. Cimanggu j

Seal• (m) 0 100 200

mm -SH'iete./1/t! mttcrophy/ltt -k~;~J - T.grttnd~·s ond Smttcrophylltt s 0 - Seconde~ry forest - Border of Temen Wlsete

Figure 2.2 The Pangandaran Nature Reserve Chapter 2: Methods 18

2.2.2 Selection of the Study Area

The Nature Reserve was chosen as the location for the study for the following reasons:

1. Several groups of Trachypithecus auratus were known to occur m the Nature Reserve (J.R. MacKinnon, pers. comm.);

2. Due to the vegetation structure and comparatively low canopy levels, it was probable that T. auratus would be more readily visible here than in many other forests in Indonesia. In a study which is heavily dependent on making observations of a primate species, good visibility is an important consideration;

3. The TW is frequented by Indonesian tourists so it was expected that T. auratus groups found there would be at least partially habituated to the presence of humans. Although possibly partially habituated, T. auratus were known not to interact directly with humans, for example, they did not accept or take food from people (J.R. MacKinnon and K.S. MacKinnon, pers. comm.). Their diet, therefore, consisted entirely of vegetation found in the Reserve;

4. The Nature Reserve was dissected by walking trails (of varying width and clarity) enabling access to most parts of it. Thus, the initial phase of the study, in which a search was conducted for suitable T. auratus study groups, could proceed without the time­ consuming process of cutting trails. Similarly, transect censuses for estimating T. auratus population density within the Nature Reserve (see Chapter 4) could be conducted along pre-existing trails;

5. The proximity of Pangandaran to Bogor and Jakarta would be convenient for the numerous trips that would have to be made to the Bogor Herbarium (for identification of plant specimens), and to Lembaga Ilmu Pengetahuan Indonesia (LIP!) (The Indonesian Institute of Sciences), the Indonesian sponsor of the study. Chapter 2: Methods 19

2.2.3 Climatic Conditions

Blower et al. (1977) described the climate of the Pangandaran Nature Reserve as a typical monsoon climate with the wettest season between October and March, coinciding with the north-west monsoon, and the driest months being from July to September during the period of the South-East Trades. Brotoisworo (1979) described the climate as being "everwet". Due to the variability in rainfall patterns (discussed below) the latter description is more apt.

The monthly rainfall (as means and ranges) for two nearby locations, outside the Nature Reserve, for 1980-1985 is shown in Figure 2.3. At Pangandaran, the readings were taken at 0700 h daily at the Rainfall Office, approximately 500 m north-west of the entrance to the Nature Reserve. The second set of readings were taken at the Department of Meteorology, Cilacap, approximately 40 km east of Pangandaran (also on the coast), at 0700 h daily. The latter are included here as a comparison with Pangandaran as subsequent meteorological data (temperature, humidity and wind) are not available for Pangandaran outside the study period. There is considerable variation in monthly rainfall from year to year for

1000 T I 900 Ill I Pangandaran I Cilacap 800 --...--· T T 700 I I -E I I I E T 600 I I T - 500 cu 400 -£: T ...cu 300 200 I I 100 I ,L J. 0 JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC month Figure 2.3 Rainfall as monthly means and ranges ( 1980-1985) - taken at two nearby locations outside the Nature Reserve Chapter 2: Methods 20

both locations (Fig. 2.3), without the distinct (and predictable) wet/dry seasons known to many tropical regions. The driest months are between February and August and the wettest months between September and January.

During the study period, climatological data were collected in the Pangandaran Nature Reserve, approximately 50 m from the entrance to the reserve. Rainfall was measured daily at 0700 h (Nylex Corp. Ltd 1000 rain gauge) (Fig. 2.4). Trends reported for the three locations (Pangandaran Nature Reserve, Pangandaran and Cilacap) are similar on a month to month basis (Fig. 2.4) although variation exists for a given month even between readings taken from the Pangandaran Nature Reserve and Pangandaran, two locations approximately 500 m apart. For example, in October 1984 the difference in rainfall between these two locations was 130 mm. This indicates that the rainfall is very localised. During the study period the wettest month was September 1984 with the driest months being August and September 1985. The rainfall in the Pangandaran Nature Reserve for the perio~ from August 1984 to July 1985 (inclusive) was 4557 mm. This is a high rainfall value when compared to other areas, in Malaysia, where primates have been studied (e.g. Raemaekers et al. 1980; Davies 1984) but not unusually high for Pangandaran.

Being virtually surrounded by sea, Pangandaran is greatly influenced by cool sea breezes, by occasional storms and by strong westerly winds with rough seas (Blower et al. 1977). Wind direction and speed recorded at the Department of Meteorology, Cilacap (Belfort Instrument Co. anemometer), showed a tendency for south-easterly winds between March and November and westerly winds for December, January and February. The westerlies may sometimes bring gale-force winds. For example, in February 1985 a reading of 29 knots was registered at Cilacap. The wind was similarly strong at the Pangandaran Nature Reserve where 22-29 knots were recorded (Dwyer wind meter), for the period from 20 February to 23 February inclusive. Chapter 2: Methods 21

1000

Ill Pangandaran 800 • Cilacap II Pangandaran Nature Reserve -E E 600

-~ -c 400 ,_~

200

AUG SEP OCT NOV DEC JAN FEB MARAPAMAY JUN JUL AUG SEP OCT NOV 1984 1985

month

Figure 2.4 Monthly rainfall (August 1984 - November 1985) at the Pangandaran Nature Reserve, Pangandaran and Cilacap

Daily maximum and minimum temperatures were taken at the Pangandaran Nature Reserve with a maxima-minima thermometer. Little variation was found on a daily or monthly basis (Fig. 2.5). The maximum daily temperature was between 31-35° C and the minimum between 22.5-25.5° C. Temperatures recorded at the Department of Meteorology, Cilacap (R. Fuess thermometers), showed similar trends although values were lower, with the maximum daily temperature varying between 29-32° C and the m1mmum temperature varying between 22-23.5° C (Fig. 2.6).

Per cent humidity was determined at the Pangandaran Nature Reserve at 0700 h and 1500 h daily (Seawell Mason's wet and dry bulb hygrometer). Humidity varied little on a daily or monthly basis. However, humidity was greater at 0700 h than at 1500 h, ranging from 92.5% to 96.5% at 0700 h and varying between 88.5% and 94% at 1500 h (Fig. 2.5). Per cent humidity recorded at the Department of Meteorology, Cilacap (R. Fuess wet/dry thermometer) Chapter 2: Methods 22

100

80

a 0700 h O::e humidity --0 0 -- 60 1500 h CD~ ..::s"'O ·- -as ·-E .. ::s 40 a.CD.c Ea: .!o • • • • • • • • 20 • • • • • m maximum temperature minimum

JUL AUG SEP CCT 'teN rEC JAN FEB MAR APR MAY JUN JUL AUG SEP CCT t'D/ DEC

1984 1985 month Figure 2.5 Per cent humidity (0700 h and 1500 h) and mtmmum and maximum· temperature as daily average each month. Readings taken at the Pangandaran Nature Reserve

40

35 .- • • • 30 • • .____. • .. (.) • • -0 - 25 ...... £! (I) m- m EY ---m-- '"Ill IY m m .. 4!! 1!1" m ::s 20 -..ftl (I) a. E 15 (I) m minimum - 10 maximum 5

0 JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

month Figure 2.6 Mean daily mmtmum and maximum temperature each month ( 1980-1985) - readings taken at Cilacap Chapter 2: Methods 23

also showed little variation on a monthly basis but was lower (82- 84%) than at the Pangandaran Nature Reserve (Fig. 2.7). Values for humidity at Cilacap were based. on daily averages of readings taken· at 0700 h, 1300 h and 1800 h. The lower overall mean daily humidity at Cilacap may reflect the inclusion of readings taken at 1800 h which were not taken at the Pangandaran Nature Reserve.

100

80

-0~ - 60 >- -"0 E ::J 40 .c

20

JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

month

Figure 2. 7 Mean daily per cent humidity each month ( 1980-1985) - readings taken at Cilacap

2.2.4 Edaphic Conditions

2.2.4.1 General

The Pangandaran peninsula, which constitutes the Nature Reserve, was formed in the Miocene period and consists of breccia and limestone formations. Most of the CA consists of dark grey breccia material which is covered by a topsoil layer of red-yellow podsolic soils, yellow podsols and brown latosols with lithosol in the riverbeds and along the coast (Blower et al. 1977). Chapter 2: Methods 24

Alluvial deposits of marine ongm, mainly sand and sandy soils, are found on the isthmus and in the coastal flats to the north. Limestone formations are found in the TW, the northern end of the CA and on the south-east coast of the CA (Blower et al. 1977). They are present as isolated outcrops or as massive uplifted cliffs which in some cases possess large caves with numerous pillar-like formations of deposited limestone. While there are no exposed outcrops in the central plateau of the CA this area may also be of limestone formation as the river water draining it is weakly alkaline (Blower et al. 1977).

2.2.4.2 Analysis of Soil Samples

A sample of topsoil (5-10 em) was collected from the forest floor of two sites within the study area (one sample from each of the T. auratus study groups' home-ranges (see Section 2.3.1)). The soil samples were air-dried and retained in sealed plastic containers.

Prior to analysis, the samples were ground with a mortar and pestel to ensure homogeneity of samples. Sub-samples were oven­ dried at 105° C for 24 h. Analyses were performed at the School of Geography, the University of New South Wales. Where possible, the methods used approximated those employed by Davies and Baillie (1988) so results were comparable. Samples were analysed in duplicate.

Soil pH and conductivity were measured electrometrically in a 1 :5 soil:water supension that had been shaken (in a spinner) for 30 min and then centrifuged at high speed (Labofuge 6000 Heraeus centrifuge), at 5,000 rpm for 3 min. Loss on ignition was measured after heating to 550° C in a muffle furnace for 30 min. Organic carbon was determined by the modified Walkly Black method using acidified potassium dichromate, and titrating against ferrous ammonium sulphate. Soil phosphorus content was determined using the Lactate method: modified Holford and Colwell methods. Phosphorus was extracted with 0.02 M calcium lactate in 0.01 M Chapter 2: Methods 25

hydrochloric acid and the contents measured (Pye Unican SP900 series atomic absorption spectrophotometer).

Exchangeable Na+, K+, Ca2+ and Mg2+ were extracted using the method described by Tucker (1985). Glycerol was added to the soil samples which were spun for 30 min of end-on-end spinning. Samples were centrifuged at high speed and the supernatants removed. The exchangeable cations were extracted by leaching with 0.5 M ammonium sulphate and the contents measured (Pye Unican SP900 series atomic absorption spectrophotometer).

Soil texture was determined by moistening and kneading a small amount of soil in the palm of the hand and assessing the behaviour of the bolus and of the ribbon produced by shearing between thumb and forefinger.

Results of the soil analyses are shown in Table 2.1. The soil samples were alkaline with a pH 7.8 and conductivity of 0.30 mS cm-1. They were high in organic matter, organic carbon and phosphorus. They also had high levels of exchangeable cations with calcium as the dominant exchangeable cation, which was to be expected considering the limestone formation of the study area (Section 2.2.4.1). Site 2 showed a trend towards higher levels for all measures, except calcium, than site 1 (Table 2.1). However, as only one soil sample was collected from each site, further comparisons between the soils of the two study sites are not warranted.

The soils were appreciably higher in all measures than the poor quality dipterocarp forest soils in Sabah, Northern Borneo (Davies and Baillie 1988). The Pangandaran soils also had higher levels of exchangeable cations than alluvial, dipterocarp and heath forest soils in Sarawak and similar levels to soil samples from a forest over limestone (Proctor et al. 1983 ). High values for exchangeable cations in the topsoil suggest a high level of soil fertility (Edwards and Grubb 1982); although no simple relationship between soil nutrient concentrations and vegetation biomass or species richness has been found (Proctor et al. 1983). For example, species-rich dipterocarp forests, like many primary rain forests, occur on very Chapter 2: Methods 26

Table 2.1 Soil Analysis

SAMPLE LOCATION

ANALYSIS SITE 1 SITE 2 (GRP3) (GRP21) pH (in water) 7.8 7.8 Electrical conductivity (ms cm-1) 0.30 0.30 Loss on ignition (%) 26.5 28.4 Organic carbon (%) 10.77 11.67 Phosphorus (soluble) (ppm) 18.06 23.66

Exchangeable 0.14 0.16 cations (meq%) 0.38 0.69 45.70 44.72 3.30 8.58

Soil texture Organic, loam, Organic, loam, fine, sandy fine, sandy

Approximate clay content (%) 25 25

All results are expressed as units of oven-dried soil. poor soils (Golley 1983; Proctor et al. 1983). However, soil nutrient concentration may affect forest biomass in secondary forests where long-term nutrient accumulation in living matter, and efficient nutrient cycling, have not become established (Proctor et al. 1983). Therefore, the rich soil in the Pangandaran Nature Reserve may provide high nutrient concentrations for secondary forest plant growth.

A high soil pH may be detrimental in inducing mmor element deficiencies; for example, magnesium and potassium uptake may be impaired by an excess of calcium (Hesse 1971). It is not known to what extent this may apply to the Pangandaran Nature Reserve soils. Chapter 2: Methods 27

2.2.5 Mammals Found in the Pangandaran Nature Reserve

In addition to T. auratus, mammals found in the Nature Reserve include: the rusa deer (Cervus timorensis); barking deer (Muntiacus muntjak); lesser mouse deer (Tragulus javanicus); banteng (B os javanicus) (possibly hybridized with domestic stock); palm civet (Paradoxurus hermaphroditus); giant squirrel (Ratufa bicolor); flying foxes (Pteropus vampytus); flying lemur (Cynocephalus variegatus); slow loris (Nycticebus coucang); and the crab-eating macaque (Macaca fascicularis). Potential predators such as leopards are no longer found in the Nature Reserve.

2.3 THE T. AURATUS STUDY GROUPS

2.3.1 Selection of T. auratus Study Groups

A survey was made of all T. auratus groups found within the TW to obtain estimates of group size and composition and the area used by each group (i.e. the home-range area) (see Chapter 4). Table 2.2 lists the age/sex classes and descriptions for each class. A group was then selected for detailed study on the basis of the following criteria:

1. The selected group was representative of the groups found within the TW with respect to group size and composition i.e. an "average" group was sought;

2. The home-range area of the group encompassed both plantation areas, as well as areas with vegetation of secondary ongm similar to the vegetation found within the CA;

3. The group was in an area away from the mam tourist trail in the TW to enable unhindered observations of the group.

The group selected in the TW is referred to as GRP3. At the time of selection, GRP3 consisted of: one adult male (AM); one adult Chapter 2: Methods 28

Table 2.2 T rachypithecus auratus Age/Sex Classes and Descriptions

Age/Sex Class Abbreviation Appearance Comments

Infant IO Orange natal Usually clinging to coat and un­ adult female, occasionally pigmented skin to other group members. on face, hands Generally within 1m of and feet. female. Dependent on female for food and during travel, IC Changing coat although when older colour; skin occasionally seen to pigmented. nibble plant items. IC-= largely The tail, face, hands and orange feet are usually the first IC-+= 50% orange areas to change from the 150% black orange natal coat to the IC+= largely black black colour of the adults, along with a mid-dorsal strip IB Black coat running down the back to the base of the tail. The infant stage lasts approximately 8-9 months although the time spent in different infant classes is variable (approx 4-5 months old when IB).

Juvenile J Colouration like Plays, feeds and travels adult, < 50% of independently. adult body length. Approximate age: 10-15 months.

Sub-adult SA Smaller than adult but > 50% of adult body length.

Adult female AF Fully grown AFl with infant AF2 = without infant

Adult male AM Fully grown

Where visibility conditions made it possible to determine the sex of an observed animal the sex was recorded e.g. JM = juvenile male. There were no sub-adult males in the two study groups. Chapter 2: Methods 29

female with infant (AFl + IC+ ); s1x adult females without infants (AF2); two sub-adult females (SAF); three juveniles (J).

Similarly, surveys were made of several groups found within the CA (see Chapter 4) to obtain estimates of group size and composition and the home-range area for each group. As groups located in the CA were not habituated, this involved extensive following of various groups to obtain realistic estimates of group composition. Generally, group composition was most successfully determined when a group progressed in a linear fashion from one area to another, all individuals in the group using the same path of travel as, for example, when jumping from one tree to another.

A group in the CA was selected on the basis of the following criteria:

1. Similar group s1ze and composition to GRP3 to avoid bias caused by differences in group structure when comparing the activity budgets, feeding and ranging behaviour of the two groups;

2. Similar general topography of the group's home-range area to that of GRP3 with the exception that there were no plantation areas in the CA. This enabled subsequent analysis of the importance of plantation species on the diet of T. auratus.

GRP20 was selected. However, after mapping of the home-range area and during the first month of observation, this group became inappropriate for further study as the size and composition of the group had changed dramatically since selection. When first selected, GRP20 consisted of one adult male, eight adult females, three sub-adults, three juveniles and one infant. During the first month of observation this group consisted of one adult male and six adult females. Reasons for the change in group size and composition were not determined but it is suspected that some members of the group may have transferred to neighbouring groups. Data collected on GRP20 have not been included in subsequent analyses. Chapter 2: Methods 30

Therefore, another group in the CA was selected on the basis of the previous criteria. This group is known as GRP21. The age/sex structure of GRP21 at the time of selection was: one adult male (AM); one adult female with infant (AFl + IC+); seven adult females without infants (AF2); two sub-adult females (SAF); three juveniles (J). Approximate locations of the home-range areas of the two study groups are shown in Figure 2.8.

Figure 2.8 Approximate locations of the home-range areas of the two study groups ..A..= GRP3 e= GRP21 Chapter 2: Methods 31

2.3.2 Mapping of Home-Range Areas

Maps were constructed from the initial estimates of the home­ range areas for each of the two study groups. Home-range areas were divided into 50x50 m quadrats whose perimeters were marked at regular intervals by red paint on tree trunks, and at 25 m intervals by a double ring of red paint. The comers of the quadrats were indicated by numbered tags nailed to trees. The quadrats were correspondingly marked on maps of the home-range areas and each quadrat allocated a reference number. Therefore, the position of GRP3 and GRP21 (see Section 2.4.1), at any given time, was ascertained to within 5 m accuracy. Where necessary, undergrowth was cleared to enable easy access to the quadrat lines.

2.3.3 Habituation of T. auratus Study Groups

GRP3 was largely habituated at the start of this study due to frequent exposure to Indonesian tourists visiting the TW. This group was still somewhat shy, however, when followed in areas away from marked trails. Members of the group would move away from the observer (typically juveniles and adults with infants), watch the observer and sometimes issue soft calls indicating a disturbance.

It was not long, however, before this group was fully habituated. During the mapping of the home-range area (Section 2.3.2) and during the pilot study (described in Section 2.4.1) the group was exposed to the presence of humans in all areas of the study site. Consequently, by July 1984 GRP3 no longer appeared disturbed by my presence nor did individuals change their behaviour in any discernible way when numbered tags were nailed to tree trunks m close proximity to them. Systematic observations of GRP3, therefore, began in July.

GRP21 was not habituated at the start of the study. In addition to the mapping period (in August 1984) more time was needed to habituate this group before observational data could be regarded as Chapter 2: Methods 32

being representative of their "normal" behaviour. During August, September and October 1984 when I was not observing GRP3 or collecting other data (see Section 2.4), I followed GRP21 in order to habituate them to my presence and to make initial observations of their behaviour and use of their home-range area. In November 1984 the group was habituated and collection of observational data began.

2.4 MONTHLY SCHEDULE

2.4.1 General

All observations in the field were made using Nikon 9x25 binoculars or Fielder 10x50 binoculars. The main method used for collecting observational data of T. auratus was scan sampling (Altmann 1974). This method has been used extensively in other studies of colobine ecology (Clutton-Brock 1975a; Struhsaker 1975; Oates 1977a; McKey 1978; Marsh 1979b; Bennett 1983; Davies 1984). In each scan sample, the behaviour of all animals visible at a given instant was recorded. The interval between successive scans in previous studies has varied between 10 min and 20 min.

A pilot study was conducted for GRP3 in June/July 1984 to determine the following: the frequency at which positions of individuals in the group needed to be plotted on a map in order to give an accurate estimate of distance travelled within a day (see Chapter 8); the use of canopy levels to determine the minimum tree height used so that subsequent enumeration of trees in the vegetation analysis (see Chapter 3) would include all trees of heights occupied by T. auratus;. activities of T. auratus to define behaviour categories.

Locations of individuals were plotted at 10-min, 20-min, 30-min and hourly intervals. Of these intervals, the estimate of distance travelled in a day was the same when locations were plotted at 10- min and 20-min intervals, both showing greater accuracy in estimating the real distance covered than estimates obtained when Chapter 2: Methods 33

plotting locations at the longer intervals. Locations of individuals, therefore, were subsequently plotted at 20-min intervals.

Activities observed were: feeding; moving; clinging; inactive; grooming; playing; vocalizing; other. Descriptions for each activity are given in Table 2.3.

Attempts were made at recogmzmg individuals but this generally proved to be unsuccessful. The single adult male m each study group could usually be recognized, as could infants due to the colour patterning of their coats as they changed from the orange colour of their natal coats to the black colour of the adults. However, it was not possible to recognize consistently adult females, sub-adults and juveniles on an individual basis. The unpigmented area in the female's pelvic region forms a distinct pattern which

Table 2.3 Descriptions of T. auratus Activities

ACTIVITY (+ abbreviation) DESCRIPTION

Feeding:Browsing (FB) Moving foliage in search of food or holding food. Feeding:Ingestion (FI) Masticating food.

Moving:Away (MA) Travelling when in the process of moving between trees as in group progression. Moving:Within (MW) Travel within a tree e.g. when moving from one food source to another in the same tree.

Clinging (C) Ventral clinging of infant to another, usually an adult female, when the other was engaged in activity other than I. Inactive (I) When sitting or lying down and not engaged in any other activity. Inactive:Huddling (IH) Two individuals, inactive but facing one another and arms around each other in embracing position. Inactive:Clinging (IC) Ventral clinging of infant to another, usually an adult female, when the other was I. Chapter 2: Methods 34

••••• Table 2.3 (cont.)

ACTIVITY ( + abbreviation) DESCRIPI'ION

Social Grooming:Groomer (GR) Grooming - when an individual manipulated the body or head hair of another. Social Grooming:Groomee (GE) Grooming - the individual whose body or head hair was being manipulated by another. Self Grooming (GS) Grooming oneself e.g. scratching.

Playing (P) When chasing or being chased along a circular route, wrestling or hanging suspended from a branch - could be solitary or with one or more other individuals. Vocalizing (V) When an animal was seen to vocalize - notes were made to describe the context of the call e.g. adult male loud call: neighbouring group in close proximity. Other (0) Any activity other than those listed above e.g. copulation, intergroup encounters.

Where an animal was observed to feed whilst engaging in another activity as well e.g. moving, the activity scored was feeding. Accounts of this were rare.

may aid in recogmtwn of individuals (Brotoisworo 1979) but this is rarely seen, being largely hidden from view by foliage. Attempts to mark individuals by "spraying" them with a brightly coloured dye proved unsuccessful as no evidence of the dye was present on the animals' coats after 24 h. Recognition was therefore limited to the level of age/sex class determination.

For comparative purposes, the focal-animal sampling method (Altmann 1974) was also used to obtain observational data of T. auratus. This was only possible for GRP3 where the topography and vegetation were such that a given individual could be observed for a prolonged period of time. Chapter 2: Methods 35

2.4.2 Scan Sampling

In previous studies of colobines, behaviour has been recorded at monthly intervals with a consecutive five-day sample period for each month (Clutton-Brock 1975a; Struhsaker 1975; Oates 1977a; McKey 1978; Marsh 1979b). A five-day sample period is considered large enough to allow differences in ranging and feeding behaviour to be sampled and related to variations in the habitat between sample periods. In this study, five-day sample periods for each month (referred to as monthly scan samples) were conducted when the scan sampling method was used.

Eight monthly scan samples were conducted for GRP3 from July 1984 to February 1985 and five monthly scan samples were taken for GRP21 from November 1984 to March 1985. It was not possible to conduct further monthly scan samples for the following reasons:

1. Gale-force winds in February 1985 (Section 2.2.3) uprooted numerous large trees in the GRP3 home-range area resulting in a change in ranging (and presumably feeding) patterns for this group. A monthly scan sample for GRP3 was commenced in March but abandoned after two days. Breaks in the canopy caused the monkeys to descend to ground level on several occasions (normally an infrequent occurrence) and it was thought that this may influence home-range use. Trachypithecus auratus were apprehensive and nervous when on the ground, constantly looking around. When tourists were nearby, T. auratus retreated and climbed trees. A comparison of the feeding and ranging behaviour for this month with previous months was not possible as the structure of the vegetation had changed markedly, rendering an analysis of food availability on feeding and ranging behaviour as inadequate. Data collected for this month have been excluded from analyses;

2. GRP21 was not fully habituated until October 1984 and hence monthly scan samples for this group commenced in November; Chapter 2: Methods 36

3. To fulfil requirements of my candidature at The University of New South Wales it was necessary to return to Australia for a four- · week period in April/May 1985;

4. As no systematic data had previously been collected on the structure and composition of the vegetation in the Pangandaran Nature Reserve it was necessary to devote the remaining time in Indonesia, from mid-May 1985 to mid-December 1985 to this aspect of the study.

Ecological studies of colobines have usually been conducted over a twelve-month period. The choice of twelve months is an arbitrary figure conforming to the length of a calendar year. Where study areas have typical wet/dry seasons it is important to sample for the two seasons. However, the climate at Pangandaran is unpredictable on a monthly basis and does not conform to distinct wet/dry seasons (Section 2.2.3). Figure 2.4 shows that sampling between August 1984 and March 1985 included months of high and low rainfall thereby providing a representative sampling period with respect to climate. Whilst discussion of results obtained in this study is restricted to the sampling period (as in any study) it is considered that the sampling period adopted here is adequate for analysis of ecological variables in an area where there are no marked seasons. Where comparisons between GRP3 and GRP21 are subsequently made, analyses are restricted to the sampling period between October 1984 and February 1985 (GRP3), and November 1984 and March 1985 (GRP21) unless otherwise stated.

During monthly scan samples, scans were made from the time the animals became active in the morning (usually shortly after dawn) to the time they became inactive (usually at dusk). The study group was located on the afternoon preceding the monthly scan sample and followed until dusk so that the group could easily be located the following morning.

Scan samples, taken at 10-min intervals, encompassed all individuals which were observed during a 5-min period. This included all animals immediately visible and those seen in the Chapter 2: Methods 37

ensuing minutes whilst actively searching for other group members. This aimed to reduce bias in overrepresentation of animals engaged in conspicuous activities. Each animal recorded during each scan was termed an observation, and a total of 23,248 observations (with an average of 8.5 animals' activity every 10 min) were recorded during the study for GRP3 and a total of 9,290 observations (with an average of 5.7 animals' activity every 10 min) were recorded for GRP21.

Occasionally, the group was lost during the day as the result of intermingling with a neighbouring group or because of heavy rain obscuring the sounds of the monkeys moving through the trees. Such incidences were rare and it was usually possible to re-locate the group within an hour.

Data collected during scan samples included information on the following for each individual observed: time; weather; the 0.25 ha quadrat in which the individual was located; the age/sex class of the individual; the activity in which engaged (see Section 2.4.1); the height of the individual above ground level; the height of the individual within the tree canopy. A list of time and weather criteria is given in Table 2.4. Where an animal was observed feeding, information was recorded on the tree number and species of the food tree, and on the item(s) and part(s) eaten. This is explained further in Chapter 6. Each food tree was numbered with a metal tag, mapped, and where the species was not known at the time of the scan sample, botanical samples were later collected and identified (Section 2.4.4).

Scan samples were also taken of other habituated groups located m the TW in order to determine whether the activity budget of GRP3 was typical of other groups found in the TW. Techniques used were the same as those described above for monthly scan samples for GRP3 except that the scan sample period for the other groups encompassed two half days per month (from August 1984 to February 1985 inclusive) rather than five full days. Scan sampling of the other groups was done in the week following the monthly Chapter 2: Methods 38

Table 2.4 Time and Weather Criteria Used During Monthly Scan Samples

1. Time: In 10-minute intervals, following west Indonesian Standard Time. (GMT + 7 hours)

2. Weather: a) Sunny b) High overcast c) Low overcast d) Rain - drizzle: drops of rain falling through canopy intermittently - light rain: consistent rain but not heavy enough to warrant wearing of raincoat by observer - moderate rain: drops of rain falling quickly through canopy and raincoat worn by observer - heavy rain: downpour of rain and difficult to see or hear anything

scan sample for GRP3. Scan sampling of other groups in the CA was not possible as no other groups there were habituated.

2.4.3 Focal-Animal Sampling

The focal-animal sampling technique provided a companson with the scan sampling technique. Age/sex class variation in activity budgets was assessed from data obtained by this method. Age/sex classes were sampled in proportion to their representation in GRP3 (with the exception of infants which were excluded) for a two-day period each month, starting on the third day following the monthly scan sample for GRP3. Three age/sex classes were sampled in a given day. An individual from one of the three selected age/sex classes was chosen at random, from all relevant individuals visible at the time, and observed for a 15-min period. An individual from each of the two remaining age/sex classes was similarly selected and observed for a 15-min period each. This process was repeated every hour with the same age/sex classes being selected, although not necessarily the same individuals. Chapter 2: Methods 39

The focal-animal sampling technique involves continuous observation of an individual over a pre-determined length of time. Therefore, two people were required for this method - an observer and a scribe. The monkeys in this group were fully habituated and showed no signs of disturbance when the observer related data to the scribe by talking. Activities recorded were the same as for the scan sampling method (Table 2.3). Locations of individuals sampled were plotted at the beginning and end of each 15-min sampling period.

2.4.4 Remainder of Each Sampling Month

The remaining time during each month was spent m the following activities:

1. Censuses for estimating population densities m the TW and in the CA (Chapter 4);

2. Collection of vegetation samples from numbered food trees for later identification. A botanical reference collection was formed (now housed in the National Herbarium, Canberra, Australia);

3. Visits were made to the Bogor Herbarium where plant samples were identified with the aid of staff at the Herbarium (but under my constant supervision);

4. Observations of plant part production (Section 3.2.2);

5. Samples were collected for phytochemical analysis (Section 3.2.3). Chapter 2: Methods 40

2.5 STATISTICAL ANALYSIS

2.5.1 Weighting of Data

The number of animals observed during a scan varied as a result of their visibility. Where animals were in close proximity to one another or in an area with a relatively· open canopy (for example, large fig trees), a greater number of animals was observed than when the group was more widely spread or in dense foliage. To avoid bias, the number of individuals recorded in a scan was weighted such that each scan contributed one point to the dataset (after Kavanagh 1978). This was done by dividing each observation in a scan by the total number of observations made at the same time. Data were weighted in this fashion whenever the proportion of time spent in different activities was analysed, where proportion of time is the sum of the proportion of scans (sometimes expressed as a percentage of the number of scans). When the proportion of time spent feeding on various species, items or parts was analysed, each feeding observation for a given scan was divided by the number of animals feeding at the same time. The datasets produced are referred to as unweighted data; the combined weighting for each scan being one.

A further weighting of data was carried out to provide a comparison with results obtained using the unweighted dataset. This technique was restricted to analysis of feeding behaviour and involved weighting each feeding observation for a given scan according to the number of individuals sighted at the same time, i.e. each feeding observation was divided by the number of animals recorded at the same time. The dataset produced is referred to as weighted data. In this way, the combined weighting of feeding observations for each scan does not necessarily total one. The contribution to the total dataset for feeding is greater for a scan where several animals were feeding than for a scan where only a few animals were feeding (assuming the same number of animals was observed in a scan). In the previous method (unweighted data) the contribution to the total feeding dataset would be the same for this example (i.e. each scan equals one datapoint). Chapter 2: Methods 41

2.5.2 Statistical Analysis

Data were largely analysed by the application of non-parametric statistical methods because of the lack of normality and the scale of measurement used. Independence of data is assumed in these tests. However, it is recognised that scan data may be interdependent within scans (e.g. activities of group members may be synchronized) and between scans (i.e. the activity of an animal at any time may influence what it does ten minutes later). Statistical procedures and nomenclature follow those of Siegel (1956) and Conover (1980). Statistical tests were two-tailed unless stated to the contrary and included the Mann-Whitney U test, Spearman rank correlation coefficient (r8 ), X2 test, Kruskal-Wallis one-way anova (H), the van der Waerden Normal Scores test and multiple comparisons test.

In addition, the Student's t-test (t) has been used where appropriate in analyses of vegetation parameters along with linear regression analysis and hierarchical cluster analysis following the Statistical Package for the Social Sciences (SPSSx) (Norusis 1985). Multiple regression analysis (Norusis 1985) was used to assess possible correlations between ranging behaviour and vegetation factors.

The abbreviation "n.s." (for not significant) denotes a probability greater than five per cent i.e. where p > 0.05.

2.6 SUMMARY

1. Fieldwork was conducted from March 1984 to December 1985 in the Pangandaran Nature Reserve, West Java.

2. The Nature Reserve is located on a peninsula and consists of 530 ha. The Nature Reserve is divided into two sections: a public use zone, the Taman Wisata (TW), consisting of 37.7. ha; and the nature reserve proper, the Cagar Alam (CA), constituting the remainder of the Nature Reserve. The TW contains an area of approximately 20 Chapter 2: Methods 42

ha of plantation composed of Tectona grandis, Swietenia macrophylla and Acacia auriculiformis.

3. The Nature Reserve has a high annual rainfall (4557mm during the study period) with no distinct wet/dry seasons, although rainfall varies from month to month with little predictability from year to year. Being virtually surrounded by sea, the Nature Reserve is greatly influenced by sea breezes and winds. South-easterly winds are most common between March and November and westerly winds between December and February. Maximum and minimum daily temperatures varied little on a daily or monthly basis. The maximum daily temperature range was 31-35° C and the minimum range was 22.5-25.5° C. Per cent humidity remained fairly constant between days and months but was greater at 0700 h (92.5-96.5%) than at 1500 h (88.5-94%).

4. Breccia and limestone formations dominate the Nature Reserve. Topsoil samples collected from two sites indicated an alkaline soil with a pH 7.8 and conductivity of 0.30 mS cm- 1. The soil samples were high in organic matter, organic carbon, phosphorus and exchangeable cations (N a+, K+, Ca2+ and Mg2+) with calcium as the dominant exchangeable cation.

5. During the initial phase of the study, surveys were made of T. auratus groups in the TW and CA. GRP3 was selected as the study group in the TW. This group consisted of fourteen individuals: one adult male; seven adult females; two sub-adult females; three juveniles; one infant. GRP20 was initially selected as the study group for the CA but then discarded due to changes in the size and composition of this group. Data on GRP20 have been excluded from analysis. GRP21 was then selected as the study group in the CA. This group consisted of fifteen individuals: one adult male; eight adult females; two sub-adult females; three juveniles; one infant. Maps were constructed of the study groups' home-ranges and the groups were habituated.

6. The main method used for collecting observational data of T. auratus was scan sampling. Scans were conducted at 10-min Chapter 2: Methods 43

intervals, from the time the animals became active m the mormng to the time they became inactive (usually at dusk), for five consecutive days per month (referred to as a monthly scan sample). Data were recorded on all monkeys observed within a 5-min period. Eight monthly scan samples were conducted for GRP3 and five monthly scan samples for GRP21. Whilst it would have been preferable to collect further monthly scan samples this was not possible. Reasons for this are presented. However, the sampling period adopted is considered adequate for analysis of ecological parameters in an area where there are no marked seasons and where the sampling period included months of high and low rainfall. During monthly scan samples, locations of individuals in the study groups were plotted at 20-min intervals.

7. Recognition of individuals (except for adult males and infants) proved impossible; animals observed were identified to age/sex class.

8. Scan samples, encompassing two half days per month, were taken of other habituated groups in the TW to determine whether the activity budget of GRP3 was typical of other groups in the TW. As no other groups in the CA were habituated it was not possible to do the same to provide a comparison for GRP21.

9. The focal-animal sampling method provided a companson with the scan sampling method and was used to assess age/sex class variation in activity budgets. Two days per month were spent in focal-animal sampling of GRP3, starting on the third day following each monthly scan sample for this group. Three age/sex classes were observed for 15 min each per hour for a day, with age/sex classes being sampled in proportion to their representation in GRP3 (except for infants which were excluded). Focal-animal sampling was not possible for GRP21 as the topography and vegetation did not permit continuous observation of a selected individual over a prolonged period of time.

10. Data were weighted to reduce bias m data collection. Chapter 3: Vegetation of the study area ' 44

CHAPTER 3 VEGETATION OF THE STUDY AREA

3.1 INTRODUCTION

The vegetation of the Pangandaran Nature Reserve is largely of coastal, lowland, secondary origin. Prior to 1922 there was sporadic shifting cultivation on the peninsula so much of the vegetation post­ dates this period. There are, however, small pockets of vegetation believed to be of primary origin; they are those areas which would have been inaccessible and/or of sacred significance to the local people. Blower et al. (1977) described the secondary forest as having a closed canopy at an average of 20-30 m above ground level and an understorey which, although luxuriant in the valley bottom, was often open and dry along the ridges. Trees of 50 em diameter and above are probably relics of the original primary forest and include species such as Buchanania arborescens, Canarium sp., Artocarpus elasticus and Ficus variegata (Blower et al. 1977).

A previous study of vegetation in the Pangandaran Nature Reserve in relation to the ecology of Bos javanicus (Sumardja and Kartawinata 1977) had been made. However, the study concentrated on the herb/shrub layers and the tree layers had not been studied in detail previously. Thus, vegetation, particularly the tree layers, in the GRP3 and GRP21 home-range areas, was analysed. Vegetation was assessed with regard to structure, composition (floristics), phenology of plant part production and phytochemistry.

3.2 METHODS

3.2.1 Botanical Plots

Structure and compositiOn of the vegetation m the two study sites was determined using the quadrat sampling method (Gysel and Lyon 1980). In vegetation analysis of previous studies of colobine ecology, a variety of techniques has been used. For example, Marsh ( 1981 a) enumerated all trees > 10 m in height m Chapter 3: Vegetation of the study area 45

the entire study area and then assigned the trees to quadrats according to the location of their trunks. Bennett (1983) enumerated trees 1 30 em at breast height (g.b.h.) within 5 m of either side of transects which traversed the length and breadth of the Presbytis melalophos study group's home-range area. The strip enumeration technique had also been used by Struhsaker (1975) and McKey (1979). Davies (1984) selected two large plots in the home-range area of the P. rubicunda study group and enumerated all trees 1 30 em g.b.h. within the plots.

From the pilot study (Section 2.4.1) it was found that the minimum tree height used by Trachypithecus auratus was 5 m. All trees 2 5 m in height in the selected quadrats, therefore, were enumerated. During subsequent observations of T. auratus the monkeys were observed to move in the layer < 5 m in height, and on a few occasions to the ground, but instances of this were rare. An enumeration of trees 2 5 m in height, therefore, reflected the forest layers frequently used by T. auratus.

Quadrats measuring 10x10 m were rando·mly selected with the aid of a table of random permutations. A 10x10 m grid was superimposed on the home-range area maps (Section 2.3.2) to enable accurate location of the selected quadrats in the field. Quadrats were delineated in the field with the aid of a compass (Suunto Co., Helsinki), tape measure and a 40 m rope. The quadrat method was used for the following reasons:

1. This was a more accurate method for determining vegetation composition than using the transect method because

a) quadrats were placed randomly rather than enumerating trees along a predetermined transect line, and b) clearing for trails may influence composition of the lower canopy;

2. Home-range areas were comparatively small and accessible (in large study areas the transect method may be preferable); Chapter 3: Vegetation of the study area 46

3. Species/area curves could be drawn and updated as random selection of vegetation quadrats continued so that it would be possible to determine when the vegetation had been adequately sampled.

Figures 3.1 and 3.2 depict the locations of the vegetation quadrats in the GRP21 and GRP3 home-range areas, respectively .

.r./' I I Cikimel L ' Cl eeri ng ' \ \ ~~----- [3] ~ I3SJ --- -.' '------.,\' ~ ill \ ~ml \ lndien ~rm '' Oceen ~6 ~

Legend Scel e (m)

~ -beech 0 50 ~ -trecks [!;] -epprox .border of home-reng e eree

Figure 3.1 Location of vegetation quadrats in the GRP21 home-range area Chapter 3: Vegetation of the study area 47

IIillllll (ill] ~ [ill] m11mmn [ill] [ill] [[ill

mmm ffilill1ffim .ffil_ ----

Legend

(:::}::::}:::::::::] - Be e c h N II -Rocky Heed! end Scele (m) 0 10 20 3040 50 ~ - Peths, trecks

- Hi 11

, ...... _ ...... ] - Approximete border of home-renge eree

Figure 3.2 Location of vegetation quadrats in the GRP3 home-range area Chapter 3: Vegetation of the study area 48

Each tree within a vegetation quadrat was numbered with a tag and the position plotted on a grid map. Measurements recorded for each tree were:

1. Tree height (rn). This was estimated in 5 rn categories with the aid of a bamboo pole measuring 3 rn in length; estimates of height were generally made independently by two observers and where estimates differed by more than 2 rn they were determined again;

2. Forest layer of the canopy: upper layer (l 15 rn); middle layer (10-14 rn); lower layer (i 9 rn);

3. The presence/absence of lianas;

4. Girth at breast height (g.b.h.) (in ern).

Diameter at breast height (d.b.h.) (in ern) was calculated from g.b.h. Basal area at breast height (in crn2) was calculated, assuming a circular cross-sectional area. In terms of primate feeding behaviour, tree biomass can be considered as an index of food available. It is not possible to measure the weight of all foliage (or of fruits or flowers) on a tree but stern basal area is a good practical indicator of foliage biomass (Anon 1981; Appendix I).

Vegetation samples were collected for each tree enumerated. This was generally done by Uce, in his capacity as a tree climber, and frequently with the aid of the 3 rn bamboo pole to which a knife had been fastened at a 45° angle to the tip, or with a powerful slingshot.

Vegetation quadrats analysed in the early stage of the study included collection of data and samples of all Iiana species in these quadrats. From observations of T. auratus, however, it was apparent that the monkeys rarely included lianas in their diet. Consequently, towards the end of the study, when most of the vegetation quadrat sampling was done, lianas were not sampled. Data on lianas have not been included in the analyses of vegetation. Chapter 3: Vegetation of the study area 49

However, where T. auratus were observed to feed on liana species (see Section 6.4.1), samples were collected where possible.

The species/area curves (number of species (cumulative) against the number of vegetation quadrats) are shown in Figures 3.3 and 3.4 for GRP21 and GRP3, respectively. The curve fitted to the entire dataset for GRP21 showed a continuing increase, i.e., the number of new species encountered increased as more vegetation quadrats were sampled (Fig. 3.3). The curve for the entire dataset for GRP3 also showed a continuing increase in the number of new species encountered as more vegetation quadrats were sampled although the increase in new species encountered, after 43 quadrats had been sampled, was slight as judged by the "levelling-off' of the curve.

When all spec1es contributing i 0.1% each to the total basal area were eliminated from the dataset the species/area curves for both groups levelled off. This occurred after 62 and 37 quadrats had been sampled for GRP21 and GRP3, respectively (Fig. 3.3 and Fig. 3.4). Thus, the increase in new species encountered with further

120

G) > 100 -Cll ::I E ::I 80 u -, CD u 60 ,a.CD

0 40 -... CD _g -a-complete dataset E 20 z::I ---+- spp. contributing ~ 0.1% to total basal area 0 0 10 20 30 40 50 60 70 80

Number of quadrats Figure 3.3 Species/area curve for vegetation in the GRP21 home-range area Chapter 3: Vegetation of the study area 50

-CD ~ 80 C'CI -:::s E :::s -(,) 60 u"'CD CD 40 c. "' 0 - 20 Complete dataset ~ CD spp. contributing~ 0.1% D E to total basal area z:::s 0 0 10 20 30 40 50 60 Number of quadrats

Figure 3.4 Species/area curve for vegetation in the GRP3 home-range area sampling of vegetation was largely attributed to species which were rare in the study areas. Due to time constraints, further sampling of the vegetation was not pursued. In all, 80 quadrats (0.80 ha) in the GRP21 home-range area and 53 quadrats (0.53 ha) in the GRP3 home-range area were sampled. This represented c. 12% of the home-range areas in each site.

Six hundred and ninety six trees belonging to 116 species and at least 36 families were sampled in the GRP21 home-range area and 354 trees belonging to 78 species and at least 32 families were sampled in the GRP3 home-range area. A list of all species identified (including food tree species not sampled in the vegetation quadrats (see Chapter 6)) with authorities is given in Appendix II.

3.2.2 Plant Part Production

The phrasing "plant part production" is used to be consistent with the terminology of previous colobine studies (e.g. McKey 1979; Marsh 1981a; Davies 1984). It should be noted, however, that plant "part" in the phenological sense is synonymous with "item" in subsequent chapters on feeding behaviour and food selection (Chapters 6 and 7). Chapter 3: Vegetation of the study area 51

The study of the cycles of plant part production has been conducted by researchers of arboreal mammals in other South-east Asian forests (e.g. Raemaekers et al. 1980; Bennett 1983; Davies 1984). Observations of plant part production were made each month on the day following the monthly scan sample for GRP3 and GRP21, to assess availability of food.

Availability scores were given for the presence of leaf buds, young leaves, mature leaves, floral buds, flowers, unripe fruits and ripe fruits. Due to the difficulty in determining whether fruits or flowers are present for Ficus spp., the inflorescence or compound fruit is sometimes referred to as a syconium or fig (Weelek Chew pers. comm.). They are here classed as fruits. Scores ranged from 0 for none, to 1 for a few, to 2 for many.

Observations were made on seven species m the GRP3 home­ range area with a sample of five trees for each species (Table 3.1). Three species were selected because they were extensively eaten in the first monthly scan sample; one species was selected because it was extensively eaten in the second monthly scan sample; three species (Tectona grandis, Swietenia macrophylla and Acacia auriculiformis) were selected because they were abundant in the plantation areas and also featured in the diet of GRP3 in the first monthly scan sample (Table 3.1). In addition, observations were made on five large fig trees, representing four species. Five trees for each of these species could not be located in the GRP3 home- range area. These fig species were selected on account of their size; three of the species were also eaten by GRP3 in the first monthly scan sample (Table 3.1).

Observations were made on four species m the GRP21 home­ range area with a sample of five trees for each species. These species were extensively eaten in the first monthly scan sample for GRP21 (Table 3.1). In addition, observations of plant part production were made for five large fig trees in the GRP21 home­ range area, representing two species. Again, it was not possible to observe five trees for these species due to their rarity. Both species Chapter 3: Vegetation of the study area 52

Table 3.1 Species Sampled for Observation of Plant Part Production

Study Site Species Number of Reason(s) for trees sampling

GRP3 Clausena excavata 5 Extensively eaten by Kleinhovia hospita 5 T. auratus in first Pterospermum javanicum 5 monthly scan sample

Eugenia zippelianum 5 Extensively eaten by T. auratus in second monthly scan sample

Acacia auriculiformis 5 Abundant in plantation Swietenia macrophylla 5 areas and eaten in Tectona grandis 5 first monthly scan sample

Ficus benjamina Large size

Ficus glomerata 1 Large size and eaten Ficus sinuata 1 in first monthly scan Ficus sumatrana 2 sample

GRP21 Dysoxylum caulostachyum 5 Ganophyllum falcatum 5 Extensively eaten by Hernandia peltata 5 T. auratus in first Sterculia coccinea 5 monthly scan sample

Ficus sumatrana 3 Large size

Ficus benjamina 2 Large size and eaten in first monthly scan sample

were selected on account of their large s1ze; one species was also eaten in the first monthly scan sample (Table 3.1).

More extensive sampling of plant part production on a monthly basis would have been preferable but was limited by time constraints. Even so, those observations made on a few select vegetation species, provided a test of whether T. auratus were selecting certain vegetation species when preferred food items were available (Chapter 6). This technique provides an alternative to the method in which a large number of trees are sampled from many species with often very few replicates per species (for example, Bennett 1983). The latter provides information on general plant Chapter 3: Vegetation of the study area 53

part productivity in a habitat and is useful in exammmg seasonal availability. The method used here focussed on plant part productivity of a few species selected on account of their importance in the vegetation composition (i.e. large stze or abundant species) and/or their importance as dietary items.

3.2.3 Phytochemistry

An analysis of phytochemical components considers both the distribution of nutrients and the distribution of plant defence compounds. Nutrients include both protein, carbohydrates and minerals. Carbohydrates occur in different forms with varying degrees of digestibility and it is consequently difficult to determine the amount of energy available to a herbivore (Bennett 1983). Gross energy determined by bomb calorimetry does not provide an estimate of available energy and this method was therefore not used. Also, the distribution of minerals in vegetation was also not assessed,· but has been discussed elsewhere in relation to colobine feeding behaviour (for example, Hladik 197 8; Waterman et al. 1980). Distribution of nutrients was restricted, therefore, to an analysis of protein content. Leaves are generally regarded as the richest sources of protein (Hladik 1977; Oates et al. 1980; Waterman et al. 1980; Bennett 1983).

Plant defence compounds include toxins and digestion inhibitors (Rhoades and Cates 1976). Toxins incorporate alkaloids, cardiac glycosides, saponins, sesquiterpenes, dipterpenes and non-protein amino acids (Freeland and Janzen 197 4; Rosenthal and Jan zen 1979; Bennettt 1983 ). Toxins are generally found in r-selected species, that is, plants found in marginal habitats and the first stages of succession (Lebreton 1982). Toxins vary greatly in structure and action and it is not feasible, therefore, to conduct a generalised assay (Bennett 1983). Analysis of alkaloids is usually restricted to a determination of presence/absence (Oates et al. 1980). As it is widely acknowledged that foregut microbes of colobines are capable of detoxification (Hladik 1977; McKey 1978; Waterman 1984), Chapter 3: Vegetation of the study area 54

toxins are not considered to be a maJor factor in food selection for colobines and an analysis of toxins has therefore not been included.

Digestion inhibitors include fibre and tannins. Digestibility of plants is inversely correlated with fibre content (van Soest 1977; Waterman et al. 1980; Choo et al. 1981). Tannins are usually found among k-strategists, that is forest species, particularly primary forest species (Lebreton 1982). Tannins precipitate alkaloids and there is, therefore, a negative correlation between abundance of alkaloids and tannins in plants (Lebreton 1982). Influence of tannins on food selection by colobines is discussed in Chapter 7.

Phytochemical parameters of the vegetation in the two study sites were assessed by analysing mature leaves of common species for their nitrogen, fibre, condensed tannin and total phenolic content. All results are expressed as percentage dry weight. Where possible, methods used closely approximated those of P. Waterman as it is in his laboratory that phytochemistry has been analysed for other colobine studies (Oates et al. 1980; Waterman et al. 1980; Choo et al. 1981; McKey et al. 1981; Davies et al. 1988; Waterman et al. 1988). After collection, leaves were sun-dried and then stored m sealed brown paper bags. Prior to analysis, samples were ground and dried to a constant weight in an oven at 50° C. Samples were analysed in duplicate.

Nitrogen (N) was measured using the Kjelfoss method (modified Kjeldahl method) at Wool and Pastoral Sciences, The University of New South Wales. Protein levels were calculated as N x 6.25. Other analyses were conducted at Division of Wildlife and Rangelands Research - Commonwealth Scientific and Industrial Research Organization (CSIRO), under the guidance of S. Cork. Fibre (lignin and cellulose) was measured as acid detergent fibre (ADF) using the van Soest (1963) method with "filtrex" units (Moir 1982). The ADF method has been shown to be highly correlated with the neutral detergent fibre (NDF) and lignin methods (Waterman and Choo 1981). Chapter 3: Vegetation of the study area 55

Phenolics (and tannins) were extracted with 50% methanol. Whilst 70% acetone (as used in P. Waterman's laboratory) may extract a few more phenolics (Bate-Smith 1973) than methanol, methanol probably extracts more than the animal does (Mole and Waterman 1987). Total phenolics (TP) were measured by the Polin­ Denis method (Oates et al. 1980) using gallic acid as a standard. Condensed tannins (CT) were measured using the vanillin method (Burns 1971) with catechin as a standard. This method has shown a consistent high correlation with the Bate-Smith method frequently used in other studies (Waterman and Choo 1981).

As standards were not the same for the latter two analyses, the results of the TP and CT assays cannot be compared directly with each other (Oates et al. 1980; Waterman et al. 1980). Similarly, comparisons of the results presented here with those of other colobine studies cannot be made based on absolute values as different standards were used (catechin cf. quebracho tannin and gallic acid cf. tannic acid) and the analyses were conducted in different laboratories. Standard curves for gallic acid and tannic acid, based on values obtained on the same day under similar conditions at CSIRO laboratories, are shown in Figure 3.5. Higher

2

y = 0.07 + 0.63x -E c ~ 1 -"'" Ill .c < y =0.05 + 0.54x

m Tannie acid • Gallicacid

0~------r------~------~~------~------o.o 1.0 2.0 3.0 mg/ml Figure 3.5 Standard curves for gallic and tannic acid Chapter 3: Vegetation of the study area 56

values for total phenolic content were obtained when using gallic acid as a standard. Gallic acid is more pure than tannic acid, the latter consisting of a mixture largely of gallic acid and ellagic acids. It is not known how the tannic acid used in P. Waterman's laboratory compared with the tannic acid used in the above comparison.

Digestibility of leaves was assessed by the in vitro pepsm­ cellulase digestibility assay (CDIG) (Choo et al. 1981). This method results in the breakdown of all protein, cellulose and smaller carbohydrate molecules.

The biological effects of tannins lie in their capacity to bind and precipitate proteins (McManus et al. 1981). An absence of a correlation between protein-precipitating capacity and total phenolic or proanthocyanidin content has been reported and the incorporation of a protein-precipitation assay recommended, as such an assay measures the property of tannins in their capacity to act as plant defensive compounds (Martin and Martin 1982). Protein precipitation (PP) was measured by the Asquith and Butler (1985) method, using bovine serum albumen (BSA) covalently labelled with Remazol brilliant blue R, a blue dye. Results were expressed as percentage of tannic acid equivalents which would precipitate BSA.

3.3 BOTANICAL STRUCTURE AND COMPOSITION

3.3.1 Structure of the GRP21 and GRP3 Home-range Areas

Tree density (of trees 2 5 m in height) was 870 trees ha- 1 for GRP21 and 668 trees ha-1 for GRP3. The stem basal area was 41.0407 m 2 ha- 1 for GRP21 and 80.8768 m2 ha- 1 for GRP3. The two study sites differed with respect to the frequency distribution of trees according to girth size classes (Fig. 3.6; X2 = 314, 12 d.f., p < 0.001). Two hundred and ninety eight trees per hectare (c. 34% of all trees enumerated) in the GRP21 home-range area had girths < 20 em (Fig. 3.6). There was also a high frequency of trees in the GRP21 Chapter 3: Vegetation of the study area 57

300

C!l .c ,._ Q) a) GRP21 c. 200 Ill Q) -,_Q) 0 -... 100 Q) .Q E :::1 c:

0 OClCl ClClClClClClClClClClClClClClClClO C\JC\JC') ~~m~oocno-NM~~m~oocno 1 I 1 I I ~~~~~~~~~~~N v 0 0 000000 I I I I I I I I I I C\JC') ~~m~oocnooooooooooN o-C\JC')~~W~OOCl girth at breast height (em)

200

.cC!l ... b) GRP3 Q) c. Ill Q) -,._Q) 100 0 -,._ Q) .Q E :::1 c: 0 OClCl ClClClClClClClClClClClClClClClClO C\JC\JC') ~~m~oocno-NM~~m~oocno I 1 1 t 1 ~~~~~~~~~~~N voo 000000 I I I I I I I I I I C\JC') ~~m~oocnooooooooooN 0...-C\JC')~~W~OOCl girth at breast height (em)

Figure 3.6 Frequency distribution of trees (~ 5 m height) in 10 em girth size classes in the home-range areas of a) GRP21 b) GRP3 Chapter 3: Vegetation of the study area 58

home-range area with girths 20-30 em. By comparison, there were fewer trees per hectare in the GRP3 home-range area with girths < 30 em (very few < 20 em) and a greatt',!r number of trees per hectare with girths in the range 40-140 em (Fig. 3.6). When excluding trees with girths < 20 em there was a density of 572.5 trees ha- 1 with a mean girth of 58 em and a median girth of 32 em in the GRP21 home-range area (n = 458), and a density of 654.7 trees ha- 1 with a mean girth of 77 em and a median girth of 50 em in the GRP3 home-range area (n = 347). Therefore, while the GRP21 home­ range area had a higher density of trees per hectare when considering all girth classes, trees were, on average, significantly larger in the GRP3 home-range area (t = 2.35, 803 d.f., p < 0.02).

The vegetation in the GRP21 home-range area is typical of a more "natural" forest in which there are a few tall emergent trees and many small trees which grow in areas where there is a break in the canopy (as is caused by tree-fall) and light has penetrated to the forest floor (Janzen 1975). The vegetation structure in the GRP3 home-range area largely reflects the artificial environment in which plantation trees have been fairly evenly spaced in certain areas and maintained by regular clearing of undergrowth in the early stages of development. Consequently, the greater frequency of trees in larger girth size classes reflects the presence of plantation trees and the virtual absence of trees with girths < 20 em reflects the effect of undergrowth clearance.

Similarly, the plantation and non-plantation areas within the GRP3 home-range area differed with respect to the frequency distribution of trees according to girth size classes (Fig. 3.7; X2 = 63, 8 d.f., p < 0.001). A quadrat was classed as a plantation quadrat where S. macrophylla and/or T. grand is had a higher biomass (as basal area) than all other species combined. On this basis c. 45% of the GRP3 home-range area was regarded as plantation. Plantation quadrats collectively formed large, relatively uniform patches. Trees with girths < 50 em were more frequent in the non-plantation areas and trees with girths in the range 60-100 em were more frequent in the plantation areas (Fig. 3.7). Excluding trees with girths < 20 em, there were 350.9 trees ha- 1 in the non-plantation Chapter 3: Vegetation of the study area 59

120

cu ..c: 100 a) non-plantation .... G) c.. 80 Ill Cl) ,_Cl) - 60 -0 ... 40 Cl) .0 E :::l 20 1::

0 ommmmmmmmmmmmmmmmmmo N~~v~~~~mo~N~v~w~~mo V f I 1 I 1 I I ~~~~~~~~~~~N 000000001 I I I I I I I I I N~v~w~~mooooooooooN o..-N~v~w~~m

girth at breast height (ern)

80

ttl b) plantation ..c: .... 60 Cl) c.. Ill G) ....Cl).... 40 -0... Cl) D 20 E :::l 1:: 0

girth at breast height (em)

Figure 3.7 Frequency distribution of trees (~ 5 rn height) in 10 ern girth size classes in the GRP3 home-range area a) non-plantation b) plantation (see text for definitions of plantation and non-plantation) . Chapter 3: Vegetation of the study area 60

areas with a mean girth of 81 em and a median girth of 39 em (n = 186) and 303.8 trees ha- 1 with a mean girth of 73 em and a median, girth of 68 em (n = 161) in the plantation areas. The higher mean girth in the non-plantation area was due to an exceptionally large Ficus sinuata tree.

Frequency distributions for tree heights showed similar trends to that reported above for girth size classes. The two study sites differed with respect to the frequency distribution of trees according to height (Fig. 3.8; X2 = 118, 7 d.f., p < 0.001) as did the non-plantation and plantation areas of the GRP3 home-range area (Fig. 3.9; x 2 = 58, 7 d.f., p < o.001).

Trees < 15 m in height were more frequent in the GRP21 home­ range area than in the GRP3 home-range area and trees 15-30 m in height were more frequent in the GRP3 home-range area (Fig. 3.8). Within the GRP3 home-range area there was a higher frequency of trees < 15 m in height in the non-plantation areas and a higher frequency of trees 15-30 m in the plantation areas (Fig. 3.9).

The main canopy layer was between 5-20 m and 5-25 m in the GRP21 and GRP3 home-range areas, respectively (Fig. 3.8). Both areas had a small proportion of tall, emergent trees, the canopies of which were generally not continuous. Where the canopies were widely separated, primates could not travel between emergents and would descend to lower canopy levels when moving between trees. Where small canopies could not support a T. auratus group, the group would split and travel, feed or rest, in neighbouring trees.

The proportion of trees bearing lianas for each 10 em girth size class in the GRP21 and GRP3 home-range areas is shown in Figure 3.10. A greater proportion of trees had lianas in the GRP21 home­ range area (78.8% of all trees) than in the GRP3 home-range area (43.2% of all trees). A larger proportion of trees in the non­ plantation areas of the GRP3 home-range area had lianas (55.9% of all trees) than trees in the plantation areas (28.6% of all trees) (Fig. 3.11 ). Chapter 3: Vegetation of the study area 61

600

co ..c: 500 a) GRP21 ,_ c.CD 400

II) CD -,_CD 300 0 -.... 200 CD .c E 100 ::J c 0 5-9 1 0-14 15-1 9 20-24 25-29 30-34 35-39 tree height (m)

300

ca ..c: b) GRP3 .... c.CD 200 II) CD -,_CD 0 -... 100 CD .c E ::J c 0 5-9 10-14 15-19 20-24 25-29 30-34 35-39 tree height (m)

Figure 3.8 Frequency distribution of trees in 5 m height classes in the home-range areas of a) GRP21 b) GRP3 Chapter 3: Vegetation of the study area 62

200

Cl:l .c: 150 .... a) non-plantation Cl) c..

1/) Cl) -~ 100 0 -.... Cl) .Q 50 E ::J c

0 5-9 10-14 15-19 20-24 25-29 30-34 35-39 tree height (m)

100

Cl:l b) plantation .c: 80 .... Cl) c.. 1/) 60 Cl) -,_Cl) 0 40 -,_ Cl) .Q E 20 ::J c

0 5-9 10-14 15-19 20-24 25-29 30-34 35-39 tree height (m)

Figure 3.9 Frequency distribution of trees in 5 m height classes in the GRP3 home-range area a) non-plantation b) plantation (see text for definitions of plantation and non-plantation) Chapter 3: Vegetation of the study area 63

120

a) GRP21

Ill 100 ctl t:: ctl 80 Cl t:: .... ctl Cl) .0 60 Ill Cl) ....Cl) 40 -..... 0

20

0 0 0 0> 0> 0> 0> 0> 0> 0> 0> 0> Ol 0> 0> 0> 0> 0> 0> 0> 0> 0 C\J v C\J C") " 0 C\J C") " 0 ' ' ' ' ' ' ' ' ~ ~ ~ C\J 0 0 0 0 0 0 0 0 ' ' ' ' ' C\J C") " 0 0 0 0 ' 0 0 0 ' 0 ' 0 ' 0 ' /\I 0 C\J C") v l() (0 f'.. CXl 0>

girth at breast height (em) .120

100 b) GRP3 Ill ctl t:: .!!! 80

Cl t:: ;: ctl 60 Cl) .0

Ill Cl) 40 ....Cl)

0 - 20

0 0 0 0> 0> 0> 0> 0> 0> 0> 0> 0> 0> 0> 0> 0> 0> 0> 0> 0> 0> 0 C\J C\J C") v l() (0 f'.. CXl 0> 0 C\J C") v l() (0 f'.. CXl 0> 0 v ' ' ' ' ' ' ' ' ~ ~ ~ ~ ~ ~ C\J 0 0 0 0 0 0 0 0 ' ' ' C\J C") v l() (0 f'.. CXl 0> 0 0 0 0 ' 0 ' 0 ' 0 ' 0 ' 0 ' 0 ' /\I 0 C\J C") v l() (0 f'.. CXl 0>

girth at breast height (em}

Figure 3.10 The proportion of trees (2 5 m height) in 10 em girth size classes bearing Hanas a) GRP21 b) GRP3 0 = no trees sampled for this girth size class Chapter 3: Vegetation of the study area 64

120

a) non-plantation 100

(I! "'s::: ~ 80

60

"'Q) -...Q) 40 -0 20 0~

0 0 0 0~~~~~~~~~~~~~~~~~~0 NNM~~~~OO~O NM~~~~OO~O vOOOOOOOO,,,,,,.,, .N NM~~~~ro~oooooooooo~ 0 NM~~

girth at breast height (em) 120 b) plantation 100

(I! "'s::: (I! 80

C'l s::: -.:: ~ 60 .0

"'Q) -~ 40 -0 20

0 0 0~~~~~~~~~~~~~~~~~~0 NNM~~~~OO~O NM~~~~OO~O vOOOOOOOO~,, -r;-. "7~,,, C\J NM~~~~ro~oooooooooo~ 0 NM~~~~OO~

girth at breast height (em)

Figure 3.11 The proportion of trees (l 5 m height) in 10 em girth size classes in the GRP3 home-range area bearing lianas a) non-plantation b) plantation (refer to text for descriptions of plantation and non-plantation) 0 = no trees sampled for this girth size class Chapter 3: Vegetation of the study area 65

The proportion of trees with Hanas in the GRP21 home-range area was not least for the smallest and largest trees (Fig. 3.10a) as has been reported elsewhere in South-east Asia (Davies 1984). As lianas were recorded on a presence/absence basis in this study it is not known whether the relative biomass of Hanas varied with tree girth size. The relatively lower proportion of trees with Hanas in the GRP3 home-range area (Fig. 3.10b) and the lower proportion of trees with lianas in the plantation areas when compared with the non-plantation areas (Fig. 3.11), particularly for the lower girth size classes, is again probably explained by the practice of clearing undergrowth in the plantation areas. Furthermore, lianas grow well in areas where trees fall and gaps are produced in the canopy, so that light penetrates (Jan zen 197 5). Such conditions were more evident in the GRP21 home-range area and in the non-plantation areas of the GRP3 home-range area.

3.3.2 Botanical Composition of the GRP21 and GRP3 Home­ Range Areas

3.3.2.1 Comparison Between the GRP21 and GRP3 Home-Range Areas

As well as being structurally diverse, the floral compositiOn of the Pangandaran Nature Reserve was also diverse. This may be due to environmental instability (Connell 1978; Moore 1983). High species diversity is maintained in a non-equilibrium state, for example, in tropical rain forests (Connell 1978). However, this may also apply to secondary forests in which advanced successional stages of forest maturation are present (Ewel 1983). Families present in the GRP21 and GRP3 home-range areas, with corresponding biomass (as basal area (%)) and abundance (as stem density (% )) are listed in Table 3.2.

Trees of the Moraceae contributed the most to the total biomass m each study site (42.5% for GRP21 and 49.0% for GRP3 home-range areas) but abundance was low in each site (2.9% and 2.5%, respectively) (Table 3.2). Trees of the Meliaceae were both abundant and had high biomass in each site, although appreciably Chapter 3: Vegetation of the study area 66

Table 3.2 Families Sampled in the GRP21 & GRP3 Home-Range Areas with Corresponding Biomass (as Basal Area (%)) and Abundance (as Stem Density (%))

GRP21 GRP3

Family % basal % stem Family % basal % stem area density area density

Moraceae 42.47 2.87 Moraceae 48.99 2.54 Hernandiaceae 21.16 1.87 Meliaceae 22.15 23.73 Meliaceae 5.20 10.06 S terculiaceae 10.72 3.67 Verbenaceae 4.28 3.02 Verbenaceae 7.99 17.51 Sterculiaceae 3.42 4.89 Ebenaceae 1.39 8.19 Tiliaceae 2.70 1.01 Euphorbiaceae 1.31 8.19 Myrtaceae 2.31 8.76 Hernandiaceae 1.18 0.85 Ebenaceae 2.22 7.04 Rutaceae 1.16 7.06 Lauraceae 2.07 5.17 Fabaceae 1.15 3.67 Malvaceae 1.93 0.43 Melastomataceae 0.57 3.67 Flacourtiaceae 1.36 6.90 Flacourtiaceae 0.45 2.83 Euphorbiaceae 1.33 8.19 Lauraceae 0.39 1.41 Rubiaceae. 1.31 3.02 Annonaceae 0.35 3.39 Anacardiaceae 1.00 0.72 Arecaceae 0.31 2.26 Bombacaceae 1.00 0.72 Myrtaceae 0.24 1.70 Oleaceae 0.98 4.89 Ulmaceae 0.24 0.85 Burseraceae 0.97 1.44 Sapindaceae 0.23 1.13 Sapindaceae 0.87 5.89 Anacardiaceae 0.22 1.41 Melastomataceae 0.63 2.01 Bignoniaceae 0.21 0.57 Fabaceae 0.55 2.59 Burseraceae 0.14 0.57 Rhizophoraceae 0.49 4.74 Tiliaceae 0.13 0.57 Ulmaceae 0.33 0.86 Urticaceae 0.09 1.13 Annonaceae 0.23 1.01 Sapotaceae 0.08 0.28 Rutaceae 0.20 2.01 Oleaceae 0.06 0.28 Urticaceae 0.20 0.86 Malvaceae 0.05 0.28 Clusiaceae 0.17 0.14 Casuarinaceae 0.04 0.28 Connaraceae 0.17 0.14 Lythraceae 0.03 0.28 Apocynaceae 0.14 0.57 Monimiaceae 0.02 0.28 Leeaceae 0.07 1.29 Pandanaceae 0.02 0.28 Pandanaceae 0.07 0.43 Theaceae 0.02 0.28 Lythraceae 0.06 0.14 Lecythidaceae 0.01 0.28 Myristicaceae 0.06 0.43 Pi ttosporaceae 0.01 0.28 Vitaceae 0.05 0.14 Myrsinaceae 0.04 0.43 Lecythidaceae 0.02 0.29 Sapotaceae 0.01 0.14

U nideni tified 0.06 0.43 U nideni tified 0.05 0.28 Chapter 3: Vegetation of the study area 67

more abundant in the GRP3 home-range area where this family included S. macrophylla, one of the plantation species. Trees of the Euphorbiaceae and Ebenaceae were also abundant in both areas but contributed comparatively little to the total biomass. Trees of the Verbenaceae were present in both sites but considerably more abundant in the GRP3 home-range area where this family included T. grandis, one of the plantation species (Table 3.2).

Trees of families which had a high biomass but were low in abundance were Hernandiaceae (GRP21) and Sterculiaceae (GRP3) (Table 3.2). In the GRP21 home-range area, trees of Myrtaceae, Flacourtiaceae, Sapindaceae and Lauraceae each accounted for > 5% of stem density but contributed little to the total biomass. In the GRP3 home-range area trees of Rutaceae accounted for > 5% of all trees enumerated but contributed little to the total biomass because they were small trees (Table 3.2).

There were 145 species per hectare m the GRP21 home-range area and 147 species per hectare in the GRP3 home-range area. The species which contributed the most to the total biomass in the GRP21 home-range area was Ficus sumatrana (24.7%) and F. sinuata had the highest biomass of all species in the GRP3 home­ range area (28.9%) (Table 3.3). Both species were represented by large individuals and frequency (as the number of trees per hectare) was, consequently, low (Table 3.3). More than 50% of the total biomass in each site was accounted for by three species; each site including two Ficus spp. in the top three species. H ernandia peltata also ranked in the top three for GRP21 and, like the Ficus spp., contained large individual trees with a comparatively low frequency of trees per hectare (Table 3.3). Of the plantation species, S. macrophylla ranked in the top three for the GRP3 home­ range area and T. grandis ranked fourth (Table 3.3); both species also occurred at a high frequency (Table 3.4).

Of the 15 species with the highest biomass at each site, trees of F. sumatrana, H. peltata and Diospyros javanica were common to both sites, although more frequent and higher in biomass in the GRP21 home-range area (Table 3.3). Table 3.3 Fifteen Species with Highest Biomass at Each Study Site GRP21 GRP3 Family Biomass Frequency Family Biomass Frequency Species (as basal (as no. of trees Species (as basal (as no. of trees area(%)) ha-l) area(%)) ha-l) Moraceae Moraceae Ficus sumatrana 24.73 11.25 Ficus sinuata 28.86 1.88 Hernandiaceae Meliaceae Hernandia peltata 21.16 16.25 Swietenia macrophy/la 19.98 115.09 Moraceae Moraceae Ficus fistulosa 10.39 6.25 Ficus sumatrana 13.26 1.89 Moraceae Verbenaceae Artocarpus elasticus 6.14 2.50 Tectona grandis 7.56 103.77 Meliaceae Moraceae Aphanamixis grandifolia 3.89 12.50 Ficus glomerata 6.65 1.89 Verbenaceae Sterculiaceae Vitex pubescens 3.04 17.50 Kleinhovia hospita 5.93 11.32 Sterculiaceae Sterculiaceae Sterculia coccinea 2.85 6.25 Heritiera littoralis 3.09 1.89 Tiliaceae Sterculiaceae Microcos paniculata 2.70 8.75 Pterospermum javanicum 1.70 11.32 Malvaceae Hernandiaceae Thespesia populnea 1.93 3.75 l/ernandia peltala 1.18 5.66 Lauraceae Fabaceae Cinnamomum iners 1.84 35.00 Acacia auriculiformis 1.11 18.87 Ebenaceae Meliaceae Diospyros javanica 1.21 52.50 Dysoxylum caulostachyum 1.09 26.42 Moraceae Rutaceae Ficus variegata 1.17 1.25 Clausena excavata 0.91 41.57 Verbenaceae Meliaceae Vitex pinnata 1.12 7.50 Melia azedarach 0.73 11.32 Myrtaceae Euphorbiaceae Eugenia zippelianum 1.04 25.00 Antidesma bunius 0.73 3.77 Bombacaceae Ebenaceae Bombax valetonii 1.00 6.25 Diospyros javanica 0.69 16.98 Table 3.4 Frequency of Species Which Accounted for l 5% Each of All Trees Enumerated (n) in the Vegetation Quadrats n = 696 (GRP21), 354 (GRP3)

GRP21 GRP3

Family Species Frequency Family Species Frequency (no. of trees ha-l) (no. of trees ha-l)

Meliaceae Dysoxylum caulostachyum 68.75 Meliaceae Swietenia macrophylla 115.09

Ebenaceae Diospyros javanica 52.50 Verbenaceae Tectona grandis 103.77

Rhizophoraceae Gynotroches axillaris 41.25 Rutaceae C lausena excavata 41.51

Lauraceae Cinnamomum iners 35.00

Anacardiaceae Buchanania arborescens 33.75

Sapindaceae · Mischocarpus sundaicus 33.75 Chapter 3: Vegetation of the study area 70

The most common species (as number of trees per hectare) are given in Table 3.4. Only species which accounted for 1 5% each of all trees sampled in the vegetation quadrats have been included. Three species, only, in the GRP3 home-range area, accounted for 2 5% each of all trees sampled: S. macrophylla ; T. grandis; Clausena excavata (Table 3.4). These species also had high biomass in the GRP3 home-range area and were absent in the GRP21 home-range area (Table 3.3).

Six species were frequent in the GRP21 home-range area: Dysoxylum caulostachyum; Diospyros javanica; Gynotroches axillaris; Cinnamomum iners; Buchanania arborescens; Mischocarpus sundaicus (Table 3.4). Of these species, only D. javanica and C. iners also had a high biomass (Table 3.3). Dysoxylum caulostachyum, D. javanica and C. iners were also present in the GRP3 home-range area. Diospyros javanica (Table 3.3) and C. iners were lower in biomass and less frequent in the GRP3 home-range area while D. caulostachyum was less frequent (Table 3.4) but had a higher biomass (Table 3.3) in the GRP3 home-range area. Buchanania arborescens and M. sundaicus were not found in the GRP3 home-range area and neither was G. axillaris. A small region within the GRP21 home-range area was mangrove. Mangroves typically contain species of the family Rhizophoraceae (Watson 1928), to which G. axillaris belongs.

A complete list of all species enumerated m the vegetation sampling, with corresponding frequency and biomass, for each study site is given in Appendix III.

3.3.2.2 Cluster Analysis

Hierarchical cluster analysis (Norusis 1985) was used in an attempt to sub-divide each study site into habitat types according to similarities between 10x10 m quadrats as assessed by the vegetation sampling. The determination of habitat types would have been useful in subsequent analyses of ranging and feeding behaviour. UPGMA Cluster Analysis (Norusis 1985) was performed Chapter 3: Vegetation of the study area 71

distance between combined clusters

0 5 10 15 20 25

48 49 43 j 44 42 J 24 52 1--- 1..5. 53 I .§1 40 12 31 - 15 22 _j 21 6 I-t- 28 ... 30 r- Q) 46 I-- ,g l E 47 :;, 5 J c 20 19 tt f-- 7 I -...as 10 "C 36 - as 37 1- :;, - tr 11 17 - 18 E 4 ~ 0 25 ,.. 27 I )( 2 8 0 I ,.. 29 1--- ~ 50 I ll 9 1 34 13 f- 26 I 32 +-- t-- 23 I i.1 38 39 r 35 16 33 I

Figure 3.12 Hierarchical cluster analysis (Norusis 1985) using average linkage between groups according to presence/absence of species in 10x10 m quadrats in the GRP3 home-range area (quadrats containing Tectona grandis are underlined) Chapter 3: Vegetation of the study area 72

on species' biomass, and presence/absence of species, using both the squared euclidean distance and Dice similarity measures.

Discrete clusters within the GRP21 home-range area were not found. Instead there were either many small clusters of quadrats or one large cluster with a few very small clusters. Similarly, discrete clustering on either biomass or presence/absence of species, leading to separation of groups of quadrats from one another, was not apparent in the GRP3 home-range area, with one possible exception. Quadrats containing T. grandis tended to form a cluster although not all quadrats with this species were included m the cluster (Fig. 3.12). Habitat types were not discerned for the remaining region of the GRP3 home-range area.

The absence of cluster formation distinguishing habitat types in each study site reflects the diversity of the vegetation indicative of a heterogeneous environment.

3.3.3 Comparison With Other Areas in South-East Asia

The coastal, lowland, secondary forest of the Pangandaran Nature Reserve is not directly comparable to other forests (largely primary) in South-east Asia where primates have been studied. However, as the ecology of T. auratus is compared with other colobines in subsequent chapters, broad comparisons are made between the vegetation of the GRP21 home-range area and the forests at Sepilok, Sabah and Kuala Lompat, Peninsular Malaysia where P. rubicunda (Davies 1984) and P. melalophos (Bennett 1983), respectively, were studied, as these are the species most frequently referred to in subsequent chapters in comparative analyses with T. auratus. Data on the vegetation of the GRP3 home­ range area have not been included in these comparisons as plantation areas were not present in other study areas.

At Sepilok, the density of trees (for trees with girths 1 30 em) for two plots measuring 0.93 ha and 0.32 ha was 442 trees ha- 1 and 669 trees ha- 1, respectively, and the basal area was 44.0170 m2 Chapter 3: Vegetation of the study area 73

ha-1 and 58.5203 m2 ha- 1, respectively (Davies 1984). The mam tree canopy was at a height of 20-40 m (Davies 1984). The density of trees (for trees with girths .?: 30 em) in the Kuala Lompat forest, for transects covering an area of 0.92 ha, was 546 trees ha- 1 and the basal area was 30.4128 m2 ha- 1 (Bennett 1983). The main tree canopy was between 5-20 m (Bennett 1983). By comparison, the density of trees (for trees with girths .?: 30 em) in the GRP21 home­ range area was 317.5 trees ha- 1 and the basal area was 39.3078 m2 ha-1. The main tree canopy was between 5-20 m.

The proportion of trees bearing lianas in the Kuala Lompat study area was 81% (Bennett 1983), and in plot 1 and plot 2 at Sepilok, 37% and 21% of trees, respectively, had lianas (Davies 1984). By comparison, 79% of trees in the GRP21 home-range area had lianas. The forest at Pangandaran, therefore, was structurally more similar to the forest at Kuala Lompat. However, there were fewer trees with girths .?: 30 em at Pangandaran, although greater in basal area.

Species composition varied greatly between the three study areas and comparison of floristics of the study areas, therefore, is restricted to an analysis of family composition. Basal area is more appropriate than stem density in such comparative analyses as basal area more closely reflects abundance as potential food sources for colobines. The proportions of total basal area contributed by trees of species collectively comprising the highest biomass for ten families at Sepilok (Davies 1984), Kuala Lompat (Bennett 1983) and Pangandaran are presented in Table 3.5. The non-riverine plot at Kuala Lompat has been excluded from the analyses as data on P. melalophos were not collected from this site (Bennett 1983).

The most noticeable differences between the study areas are: the vast abundance of Dipterocarpaceae (and, to a much lesser degree, Lauraceae) in the Sepilok study area; the abundance of Leguminosae in the Kuala Lompat study area; the abundance of Moraceae and Hernandiaceae at Pangandaran (Table 3.5). Fabaceae (= Leguminosae (s.l.)) was also present in the Pangandaran Nature Reserve but was of less importance, contributing < 1% to the total basal area. These families are of particular relevance to studies of Chapter 3: Vegetation of the study area 74

Table 3.5 Proportion (as Percentage) of Total Basal Area Contributed by the Ten Families With the Highest Biomass at the Sepiloka, Kuala Lompata and Pangandaran Study Areas (number of stems given in parentheses)

Family Sepilok Kuala Lompat Pangandaran

flat plot ridge plot riverine plot GRP21 study site

Anacardiaceae 3.2(15) 1.9(10) 8.8(24) Annonaceae 5.5(43) Combretaceae 3.1(4) Dipterocarpaceae 59.1 ( 104) 59.5 (66) 3.7(7) Ebenaceae 1.4(17) 2.2(49) Euphorbiaceae 3.4(39) 1.9( 19) 5.5(67) Fagaceae 1.7(1) Flacourtiaceae 6.0(37) 1.4(9) Hernandiaceae 21.2(13) Lauraceae 9.9(42) 15.9(10) 2.1(36) Leguminosae 1.8(7) 24.5(67) Malvaceae 1.9(3) Meliaceae 1.8(21) 5.2(70) Moraceae 42.5(20) Myrtaceae 1.7(9) 4.2(27) 2.3(61) Olacaceae 1.3(1) Rubiaceae 3.1(3) Sapindaceae 5.2(37) S terculiaceae 1.7(6) 5.2(12) 3 .4(34) Tiliaceae 2.3(16) 2.7(7) Urticaceae 4.0(7) Verbenaceae 1.2(10) 4.3(21)

TOTAL 90.2(299) 90.0( 143) 69.7(295) 87 .8(314)

a From Waterman et al. 1988

colobine ecology. Leguminosae (Bennett 1983), Moraceae and Hernandiaceae (Section 6.4) are important food sources for colobines whereas Dipterocarpaceae tend to be avoided by colobines as a food source (Davies 1984). A negative correlation between the proportion of Dipterocarpaceae in a forest and primate density has been reported (Marsh and Wilson 1981a; Davies 1984) while a positive correlation has been found between the proportion of Leguminosae in the forest and the density of P. melalophos and T. obscura (Marsh and Wilson 1981 a). Chapter 3: Vegetation of the study area 75

3.4 PLANT PART PRODUCTION

Flowers, fruits, young leaves and mature leaves were available m the GRP3 and GRP21 home-range areas throughout the study, although the number of trees bearing fruits and flowers varied monthly (Fig. 3.13 and Fig. 3.14). Fruit production in the GRP21 home-range area was highest (72%) during November and lowest (16%) in February (Fig. 3.13). In the GRP3 home-range area, fruit production was lowest in October and December (30-35%) and varied between 50-70% in other months (Fig. 3.13).

Generally, fewer trees bore flowers than fruits in any one month. Flower production in the GRP21 study site was lowest in November and February (12-16%) and varied between 20-30% in other months (Fig. 3.13). In the GRP3 study site, flower production was lowest in September and December (10-25%) and varied between 25-40% during the remaining months (Fig. 3.14). The number of trees with young leaves and mature leaves was consistently high (Fig. 3.13 and Fig. 3.14).

These results do not provide a comprehensive description of plant part production for the study sites. Firstly, sample sizes were small for both groups and secondly, scoring on a presence/absence basis ignores the relative abundance of an item in the canopy of any one tree. For example, from Figure 3.13 it may appear that young leaves and mature leaves were equally, highly available. This was not the case. Mature leaves were generally far more abundant (frequently scoring 2 for abundance (Section 3.2.2)) than young leaves. Furthermore, young leaves and leaf buds have not been considered separately in Figures 3.13 and 3.14 and there is, therefore, no distinction made between very young leaves (which may be preferred by T. auratus (Section 6.5)) and tougher, older young leaves approaching mature leaves in size and texture. Further discussion of interspecific and intraspecific variation in availability of food items is deferred to Section 6.5 where monthly variation in food items eaten is assessed with regard to food availability. Chapter 3: Vegetation of the study area 76

30

20

% [[] Flowers/flower buds 10

NOV DEC JAN FEB MAR

50

% • Fruit

0 NOV DEC JAN FEB MAR

100

% 50 I2J Young leaves/leaf buds

NOV DEC JAN FEB MAR

100

% 50- D Mature leaves

.

o~----~------~.------T-.------~------4----- NOV DEC JAN FEB MAR Month

Figure 3.13 Proportion of trees which bore young leaves, mature leaves, fruits and flowers in the phenology sample for the GRP21 home-range area (n = 25) Chapter 3: Vegetation of the study area 77

40

30

o/o 20 [I] Flowers/flower buds

10

JUL AUG SEP OCT OOV DEC JAN FEB

50 o/o • Fruit

0 JUL AUG SEP OCT OOV DEC JAN FEB

100

o/o 50 [21 Young leaves/leaf buds

JUL AUG SEP OCT OOV DEC JAN FEB

100.

o/o 50 • 0 Mature leaves

JUL AUG SEP OCT OOV DEC JAN FEB Month Figure 3.14 Proponion of trees which bore young leaves, mature leaves, fruits and flowers in the phenology sample for the GRP3 home-range area (n = 40) Chapter 3: Vegetation of the study area 78

Figure 3.13 and Figure 3.14 therefore illustrate the following:

1. Presence of all items throughout the study;

2. Variation in monthly availability which was not the same for the two study sites and not obviously related to rainfall patterns (Section 2.2.3). For example, there was no correlation between the proportion of trees in the phenology sample of the GRP3 home­ range area bearing fruits (rs = - 0.39, n = 8, p > 0.05) or flowers (rs = - 0.14, n = 8, p > 0.05) and monthly rainfall.

3.5 PHYTOCHEMISTRY

3.5.1 Phytochemistry of Mature Leaves in the Two Study Sites

Results of phytochemical analyses for species comprising 71.5% of the total basal area in the GRP3 home-range area and 62.1% of the total basal area in the GRP21 home-range area are presented in Table 3.6. Both means and medians are presented: means provide the basis for comparison with other studies (Section 3.5 .2) and medians were used in comparisons between the home-range areas of the two groups. Results for all species analysed are given in Appendix IV.

There was a tendency for the vegetation in the GRP3 home­ range area to have a higher nitrogen, and therefore protein, content and a lower ADF content. However, CDIG was lower and TP and CT contents higher in mature leaves from the GRP3 home-range area. PP capacity also tended to be higher in the vegetation of the GRP3 home-range area (Table 3.6). None of these differences were significant (Mann-Whitney U).

Weighted values for each chemical measure were calculated from the equation I (Pi x X j)IIP i where Pi is the proportion of the basal area contributed by species i and Xi is the chemical measure for species i (Newbery et al. 1980; Waterman et al. 1988) Table 3.6 Phytochemical Analyses of Mature Leaves in the Home-Range Areas of GRP3a and GRP21 b

Chemical Analysisc

Nitrogen ADF TP CT CDIG PP

Mean 2.34 35.63 5.14 2.94 40.27 4.09 (SD) (0.71) (9.98) (4.25) (3.45) (17.51) (3.82)

GRP3 Median 2.20 36.66 3.43 0.83 46.47 1.89 (IQ) ( 1.8-2.85) (26.11-45.06) (1.53-9.06) (0.47-6.17) (25.53-52.11) (0.90-7 .32)

Mean 1.91 37.93 2.66 1.59 45.80 2.14 (SD) (0.68) (10.93) (2.24) (2. 71) (13.88) (1.79)

GRP21 Median 2.00 39.19 1.49 0.32 47.49 1.27 (IQ) (1.45-2.35) (27 .87-49.68) (1.12-5.30) (0.12-2.24) (32.01-54.80) (0.67-3.91) an=9, basal area = 71.52% bn=l2, basal area = 62.12% CRefer to Section 3.2.3 for descriptions of abbreviations for chemical analyses

SD = standard deviation IQ = interquartile range Chapter 3: Vegetation of the study area 80

(Table 3. 7). Weighted values account for the proportion of basal area sampled and are, therefore, a better indicator of the chemical profile of the vegetation in a given area. Trends in differences between the vegetation of the two home-range areas were similar to differences based on mean (and median) values (Table 3.6) except that mature leaves in the GRP3 home-range area were more digestible than vegetation in the GRP21 home-range area when based on weighted values (Table 3. 7).

TP and CT content were positively correlated for mature leaves in the GRP3 home-range area (rs = 0.734, n =12, p < 0.01) and in the GRP21 home-range area (rs = 0.701, n = 9, p < 0.05). This indicates that a consistent proportion of phenolics were laid down as condensed tannins. A similar correlation has been reported in other phytochemical studies; for example, for vegetation in Sepilok (Davies 1984). However, the relationship between TP and CT is not always clear, as evidenced by the lack of correlation between these two variables in the vegetation at Kuala Lompat (Bennett 1983).

Linear regression analysis was used to examine the relationship between CDIG and other chemical measures. Data on vegetation from the two study sites were combined to increase sample size. A significant negative relationship was found between CDIG and ADF

(r = -0.66, F 1•20 = 15.02, p < 0.01). CDIG was not correlated with any other measure including TP and CT. Previous studies have similarly reported that CDIG was more strongly correlated with ADF than any

Table 3.7 Weighted Values for Chemical Measures of Vegetation (Mature Leaves) in the Two Home-Range Areas Chemical Analysisa

Home-range N Prot ADF CDIG cr TP PP area

GRP3 2.14 13.38 34.42 51.06 3.45 4.90 4.36

GRP21 1.80 11.03 36.71 46.54 0.96 2.38 2.04

0 Refer to Section 3.2.3 for descriptions of abbreviations for chemical analyses Chapter 3: Vegetation of the study area 81

other measure, although TP has been found to also have an influence in some studies (Choo et al. 1981; Waterman et al. 1983; Waterman et al. 1988).

The relationship between PP capacity and other measures was similarly assessed by linear regression analysis. PP was highly correlated with CT content (r = 0.83, F 1,20 = 42.24, p < 0.001) and with TP content (r = 0.90, F 1,20 = 85.11, p < 0.001) a reflection of the correlation between these two variables as described earlier. No correlation with ADF or protein content was found. The effects of protein-precipitation capacity of phenolics and tannins is discussed in relation to feeding behaviour in Chapter 7.

3.5.2 Comparison With Other Areas in Asia and Africa Where Colobines Have Been Studied

A comparison between the phytochemistry of mature leaves in the secondary forest in the GRP21 home-range area, with mature leaves at five other study areas where colobines have been studied, was made (Table 3.8). Values given were adapted from Waterman et al. (1988) and were weighted to account for the proportion of basal area sampled (see Section 3.5.1).

Two ratios are given: protein/ ADF; protein/(ADF + CT) (Table 3.8). These ratios are considered to be important in the role of plant chemistry in influencing food selection by colobine species (McKey 1978; McKey et al. 1981; McKey and Waterman 1987; Davies et al. 1988) and ultimately for their influence on colobine biomass (Waterman et al. 1988), aspects which are discussed further in Chapters 7 and 9. The higher the ratio obtained, the better the quality of the leaf.

The secondary forest at Pangandaran had lower protein/ ADF and protein/(ADF + CT) ratios than the forest at Kibale but had higher ratios than for the remaining study areas. Differences in these ratios may reflect differences in chemical defence of foliage against herbivory. Soil quality may influence chemical profiles of Chapter 3: Vegetation of the study area 82

Table 3.8 Weighted Phytochemical Data for Mature Leaves at Six Study Areas (adapted from Waterman et al. 1988a)

Studv Area

Douala-Edea Kibale Kakachi Kuala Sepilok Pangandaran Lompat GRP21

% b.a.c sampled 56.3 87.4b 88.1 49.2 78.0 62.1 Protein/ ADF 0.202 0.510 0.242 0.242 0.167 0.414 Protein/(ADF+CT) 0.187 0.492 0.207 0.214 0.146 0.405

0 Waterman et al. (1988) included data from the following sources in calculations of weighted values: Douala-Edea (Newbery et al. 1986), Kibale (Struhsaker 1975; Oates 1977a), Kakachi (Oates et al.1980) hweighted data based on number of stems (Struhsaker 1975) as basal area data not available

CBasal area foliage in that trees growing on poor soils may invest more heavily in chemical defence (fibre and tannins) because of the greater cost in replacing leaves eaten by herbivores (Janzen 1974; Newbery et al. 1980; Waterman et al. 1988). Differences in soil quality between Sepilok and Kuala Lompat are considered to be the factor most likely to explain differences in the chemical profiles for these two study areas (Waterman et al. 19 8 8).

Colonising species, by investing less m chemical defence, may comprise relatively high food quality (McKey et al. 1981; Coley 1983) when compared with climax species. Possibly this factor may explain why the chemical profile of vegetation in the secondary forest at Pangandaran was of higher quality than the chemical profile of vegetation in four other study areas (Table 3.8). For example; Colobus guereza in Kibale tend to select leaves of colonising tree species containing lower levels of fibre and being more easily digested than those of evergreen species (Oates 1977 a). Interestingly, the weighted values for protein/ ADF and, to a lesser degree, protein/(ADF + CT) for the vegetation in the GRP3 habitat were very similar (0.415 and 0.373, respectively) to the GRP21 home-range area (Table 3.8). Other factors which may influence plant chemistry include leaf life-span (Baranga 1983; Coley 1983; Chapter 3: Vegetation of the study area 83

Janzen and Waterman 1984), topography, rainfall, altitude, light and temperature (Waterman et al. 1988).

3.6 SUMMARY

1. Randomly selected 10x10 m quadrats were used to describe the structure and composition of the GRP21 (0.80 ha sampled) and GRP3 (0.53 ha sampled) home-range areas.

2. All trees in the quadrats 2 5 m in height were enumerated in the vegetation analysis.

3. Six hundred and ninety stx trees of 116 species were sampled in the GRP21 study site and 354 trees of 78 species were sampled in the GRP3 study site.

4. There was a higher frequency of trees per hectare in the GRP21 home-range area than in the GRP3 home-range area. Generally, however, trees were smaller in the GRP21 home-range area. Trees in the non-plantation areas of the GRP3 study site were generally smaller than trees in the plantation areas.

5. Trees of the Moraceae collectively comprised the highest biomass (as basal area) of any one family in each study site. Trees of the Meliaceae were the most abundant (as stem density). By comparison, Dipterocarpaceae had the highest biomass in Sepilok, Sabah where P. rubicunda was studied (Davies 1984) and trees of the Leguminosae had a high biomass in the study area at Kuala Lompat, Peninsular Malaysia where P. melalophos was studied (Bennett 1983).

6. Flowers, fruits, young leaves and mature leaves were available throughout the study although the number of trees bearing fruits and flowers varied monthly.

7. Phytochemical analyses of mature leaves in both study sites included: nitrogen, acid detergent fibre (ADF), total phenolics (TP), Chapter 3: Vegetation of the study area 84

condensed tannins (CT), digestibility (CDIG) and protein- precipitation capacity (PP). Trees of species comprising 71.5% of the total basal area in the GRP3 study site and 62.1% of the total basal area in the GRP21 study site were analysed. Differences in these measures for the vegetation of the two sites were not statistically significant but vegetation in the GRP3 home-range area tended to have higher nitrogen, TP and CT levels and lower levels for ADF and CDIG than the vegetation in the GRP21 home-range area. Weighted values, taking into account the proportion of basal area sampled, showed similar trends except that CDIG was now greater in vegetation in the GRP3 home-range area than in vegetation in the GRP21 home-range area.

8. The secondary forest at Pangandaran had higher protein/ ADF and protein/(ADF + CT) values than four (out of five) other study areas where colobines have been studied.

9. The weighted protein/ADF and protein/(ADF + CT) ratios m vegetation of the GRP3 and GRP21 study sites were very similar. Chapter 4: Group size, population density, biomass and social organization 85

CHAPTER 4 GROUP SIZE, POPULATION DENSITY, BIOMASS AND SOCIAL ORGANIZATION

4.1 INTRODUCTION

There is broad variation in group size and social organization of primate species. For example, Pongo pygmaeus (orangutan) form groups which consist of solitary males and mother-infant pairs for most of the year (MacKinnon 1977) whereas Papio cynocephalus (baboon) form groups with a mean size of eighty (De V ore and Hall 1965). Social organization varies from one based on a monogamous mating system as for the gibbon Hylobates lar (MacKinnon 1977) to one based on the more common polygynous system in which groups consist of a harem structure with a single adult male per group (for example, Presbytis melalophos (Bennett 1983)) or several adult males per group (for example, Colobus badius (Clutton-Brock 1974)). The population density and biomass of primates is also highly variable between species.

Crook and Gartlan (1966), in an attempt to explain the variety m social organization of primates, placed primate species into five categories and concluded that there was similarity in social organization (including group size) of species sharing the same diet and habitat type. This paper was regarded as providing evidence that social organization and ecology are closely related. Gartlan (1968) and Denham (1971) also concluded that social structure is habitat dependant rather than species-specific.

Clutton-Brock (1974a) reported that the differences within most groupings are more striking than the differences between them. For example, within "arboreal leafeaters" (Crook and Gartlan 1966) average group size varied from < 10 to > 50 and home-range size varied from < 0.02 to > 1 km2 . Clutton-Brock (1974a) reasoned that different species may react to environmental pressures in different ways and gave as an example the difference in social organization of C. badius and C. guereza; both species are folivorous, arboreal and forest-dwelling. Chapter 4: Group size, population density, biomass and social organization 86

Colobus badius lives in large, multimale troops of 2 40 animals with large (1 km2) home-range areas. Colobus guereza lives in smaller groups of 5-10 animals, frequently with a single male per group, and defends small home-range areas of < 0.2 km2. Colobus badius ate flowers, fruit and young leaves of a variety of tree species (although selective in choice of food), a food source which was not evenly distributed and varied seasonally. When these items were not available, C. badius ate mature leaves. Colobus guereza fed almost exclusively on two tree species which, although also available to C. bad ius, were not used extensively as a food source by the latter. When young leaves of these two species were not available, C. guereza ate mature leaves of one of these species although shoots, flowers and fruits of other species (eaten by C. badius) were available. Struhsaker and Oates (1975) similarly found that C. guereza ate more mature leaves and fruit and had a less diverse diet than C. badius.

Clutton-Brock (1974a) concluded that small group size in C. guereza permits the animals to minimise range size, thereby increasing their ability to defend their food supply efficiently. Colobus badius, by contrast, requires a larger home-range area to provide sufficient supplies of acceptable food being derived from a wide variety of species with clumped distributions and subject to seasonal availability. The range size required to support even a small group of C. badius would be too large to be efficiently defended and, therefore, selection has favoured larger group size. Larger group size may have other benefits including group knowledge of the distribution of food, and regular use of different parts of the home-range by formation of sub-groups (Crook 1966; Wrangham 1975) to maintain leaf growth at an acceptable stage and to maximise the reaction of the tree to cropping. Large group size may also enhance the detection of, and defence against, predators (Clutton-Brock 1974a).

Large group size and large home-range area has been associated with clumped, unstable food supplies in several other primate species. For example, Hladik (1977) in a study of P. entellus and Trachypithecus senex in Sri Lanka found that P. entellus, which Chapter 4: Group size, population density, biomass and social organization · 87

lives in larger groups with larger home-range areas than T. senex, fed to a greater extent on fruit and to a lesser degree on mature foliage. Presbytis entellus also utilised a wider variety of foods. However, in some colobine groups, a correlation between dietary diversity and group size does not exist. For example, P. melalophos has a more diverse diet than T. obscura (Curtin 1976, 1980; MacKinnon and MacKinnon 1978) but group size is similar for the two species (Bennett 1983).

Rather than argue against a relationship between social organization and ecology, Clutton-Brock (1974a) emphasised the influence of comparatively small differences in habitat and feeding behaviour on the social organization and group size of closely related species.

In a review of 100 primate species, Clutton-Brock and Harvey (1978) found that terrestrial species were of greater body weight than arboreal species and that folivores were heavier than frugivores. This was regarded as similar to the association between diet and body size of antelopes, where browsers (on a relatively nutritious diet) are usually smaller than grazing species (Jarman 1974). There was a significant negative association between body weight and population density, but a positive correlation with species biomass, both in frugivores and folivores. Folivorous species showed the highest biomasses (Clutton-Brock and Harvey 1977b, 1978). Within the colobines this relationship is poor. For example, P. melalophos and T. obscura, while having similar body weights, have different biomasses at Kuala Lompat (Marsh and Wilson 1981b), and P. entellus and T. senex have similar biomasses at Polonnaruwa (Hladik 1977) although they have different body weights. Furthermore, while T. obscura is more folivorous than P. melalophos its biomass at Kuala Lompat was less (Marsh and Wilson 1981b).

Little difference m average group size was found between folivores and frugivores but on comparison of closely related species a tendency for larger group size for frugivorous species was noted (Clutton-Brock and Harvey 1978). Whilst this generalization Chapter 4: Group size, population density, biomass and social organization 88

holds in a comparison of the frugivorous P. entellus with the more folivorous T. senex (Hladik 1977) it is not true of all closely related colobines. For example, P. melalophos (Bennett 1983) and P. rubicunda (Davies 1984) live in smaller groups than the more folivorous T. auratus (see Chapter 6).

The relationship between body size and degree of frugivory is similarly confused. Presbytis entellus is of greater body weight than T. senex (Roonwal and Mohnot 1977) and T. cristata, P. melalophos and P. rubicunda are of similar body weight (Roonwal and Mohnot 1977; Bennett 1983; Davies 1984). Clearly, a simple relationship between body size and group size with degree of frugivory/folivory in the leaf monkeys of Asia does not exist. Again, small differences in habitat and feeding ecology may correlate, at least partially, with differences observed in body weight and group size of closely related species.

Harvey et al. ( 1986) have suggested using brain stze rather than body size as a variable when comparing primate life history variation. Body size may vary considerably during adult life whereas brain size is more constant. Deviations from the allometric relationship of brain on body size relate to diet suggesting that diet variables may provide clues about factors influencing differences in brain size (Harvey et al. 1986). For example, smaller brain size of folivores when compared with frugivores may be associated with smaller home-range areas requiring less processing and storage of information (Clutton-Brock and Harvey 1980; Milton 1981; Harvey et al. 1986). Alternatively, lower metabolic rates of folivores may be associated with smaller brain size (Harvey and Bennett 19 8 3; Harvey et al. 1986).

Colobine biomass m South-east Asia has been related to the abundance of dipterocarps. Where the proportion of dipterocarps IS large, the proportion of vegetation comprising potential food trees Is low (Marsh and Wilson 1981b; Caldecott 1983; Davies 1984). Recently, an apparent positive correlation between colobine biomass and the protein to digestion inhibitor ratio of mature leaves in the habitat has been reported (Waterman et al. 1988). It is argued that Chapter 4: Group size, population density, biomass and social organization 89

colobines living in forests where mature leaves have a relatively high protein to digestion inhibitor ratio are able to eat mature leaves when preferred foods are scarce. Colobines living in forests where mature leaves have a low protein to digestion inhibitor ratio are unable to do this (Waterman et al. 1988)).

A problem with interpretation of social organization and ecological factors is that it is not possible to determine cause and effect (Davies and Krebs 1978). For example, is large home-range size a consequence of large group size or vice versa? Is large group size an adaptation for finding clumped food sources or a consequence of feeding on clumps of food? This problem is compounded when comparing the social. organization and ecology of different species as phylogenetic, morphological or physiological factors may place constraints on these variables (Struhsaker 1969; Clutton-Brock 1974a; Crook et al. 1976).

An examination of intraspecific differences m group size and home-range area for various habitats has provided more direct information on the effect of ecology on social organization. Napier and Napier (1967) in a review of social behaviour in Presbytis and Trachypithecus species found that group size of T. cristata was larger in drier more open areas than in forest areas. A similar association was reported for P. entellus; but a reverse trend in group size was noted for T. senex.

Marsh (1979b) found that C. badius occurred in considerably larger groups in rain forest than in more seasonal environments (semi-deciduous) or in montane forest. Struhsaker and Leland (1979) reported that group size in red colobus is related primarily to differences in overall food density acting on the rate of juvenile mortality. Marsh ( 1979b) questioned this as he considered it doubtful that food density is lower in riverine forests than in rain forests. Whilst tree densities are lower in riverine forests, the production of new leaves and fruit is often greater because of the higher proportion of colonizer species which are favoured in an unstable habitat (Opler 1978). The density of C. badi us was similar for both habitats (200-300 animals km-2). Marsh (1979b) Chapter 4: Group size, population density, biomass and social organization 90

concluded that home-range areas and group size were smaller m nvenne forests because of the low diversity of tree species m drier habitats decreasing the benefits of ranging widely.

Dunbar and Dunbar (1974) in a study of C. guereza m Ethiopia found a correlation between population density and preferred food tree availability (Ficus and Celtis genera). There was a poor correlation between group size and range size within different habitats. Dunbar and Dunbar (1974) inferred that the capacity for increasing home-range area affects group size. Where population density was low (Budongo) there was room for expansion in range size as an increase in group size necessitated. At this site, therefore, there was a correlation between group size and range size. At Bole, where population density was high, range size was more or less fixed and group size was presumed to be limited by range size and the availability of food. At this site, therefore, an increase m group size could only be compensated by the group splitting.

Similarly, P. entellus in India occurring in low densities at Orcha and Kaukori had comparatively larger home-range areas (384 and 768 ha, respectively) than groups occurring at Dharwar where the population density was 30 times higher and ranges averaged 18.4 ha (Yoshiba 1968).

Conversely, group size may remain constant but home-range area may decrease in size as population density increases (for example, C. guereza in the Kibale Forest, Uganda (Oates 1977a)). Reduced home-range size and high population density may be associated with low numbers of interspecific competitors (Clutton­ Brock and Harvey 1978) or intense intraspecific competition (Chivers 1969).

Gartlan (1973) summarised factors which are likely to affect social organization as follows:

1. fixed, genetically controlled factors which are relatively inflexible to environmental influence; Chapter 4: Group size, population density, biomass and social organization · 91

2. density-dependent factors; for example, territoriality;

3. transient factors which depend on a particular or a changing environment; for example, distribution of food which may be affected by seasonal factors (or toxicity (Coelho et al. 1976)).

In addition, group size and social organization may be influenced by predation pressure (MacKinnon 197 4; Tilson 1977; van Schaik and van Hoof 1983) and the avoidance of disease (Freeland 1976, 1979).

Group size (and home-range area) may be seen as a compromise of selection pressures operating on individuals (Wrangham 1979). An individual's behaviour is influenced by social as well as ecological pressures (Wrangham 1988). Population density and biomass for a given species are, in turn, the result of group size and home-range area. In this chapter, group size, population density and biomass of T. auratus in the Taman Wisata and Cagar Alam are presented and possible ecological influences on these parameters outlined. These influences are further examined in subsequent chapters in comparing the activity budgets (Chapter 5), feeding (Chapters 6 and 7) and ranging (Chapter 8) behaviour of the two study groups. A final appraisal of the relationship between social organization and ecology is given in the concluding chapter.

Selection operating on individuals will also influence other aspects of social organization such as group composition, allomothering behaviour and occurrence of infanticide (Clutton­ Brock and Harvey 1976). For the sake of completeness, these and other aspects of social organization in T. auratus are also examined and comparisons made between T. auratus and other colobines, particularly Trachypithecus and Presbytis species. Chapter 4: Group size, population density, biomass and social organization 92

4.2 METHODS

4.2.1 CA Censuses

The line transect or strip (Caughley 1977) method was used to estimate the number of T. auratus groups in the CA. This method has been recommended for surveying primate populations (Anon 1981) and has been widely used in Africa (for example, Struhsaker 1975; McKey 1978; Whitesides et al. 1988) and in South-east Asia (for example, Wilson and Wilson 1975; Marsh and Wilson 1981a; Davies 1984).

A 6 km transect was selected (Fig. 4.1) and marked at 50 m intervals with a ring of red paint around the nearest tree trunk. This transect was selected as being representative of the vegetation and topography in the CA, including both coastal and inland areas, with the latter including gullies, ridge tops and large flat areas typical of the central region of the CA. The transect line was along pre-existing paths so as to avoid additional cutting of trails. As poaching of primates was non-existant in the reserve, bias introduced as a result of T. auratus avoiding cleared trails was considered unlikely.

Six censuses were conducted in May 1984 and one or two censuses per month were conducted in subsequent months between June 1984 and December 1985 (excluding April and May 1985). A total of 35 censuses were obtained (total length of transect samples = 210 km). Transects were traversed between 0700-1100 h (first 3 km of the transect) and 1400-1800 h (the latter 3 km of the transect) and were not conducted on rainy days when detectability and observations of T. auratus may have been impaired. The rate of travel was 40 min km-1. The starting point of the census alternated between the two end points of the transect to avoid possible bias which may have resulted if sections of the transect were consistently traversed at the same time of day.

When a T. auratus group was encountered, the following data were collected: time of day; weather conditions (see Section 2.4.2 Chapter 4: Group size, population density, biomass and social organization 93

Batu Mandi

Tg. Batu Pepet

Legend

~- Streams T g. Cimanggu ~- Grazing areas with overgrowth D Boundary of Taman JJiisettJ scale (km) ~- Transect track end distance from 0 start (km) Figure 4.1 Six kilometre transect used for censusing Trachypithecus auratus in the Cagar Alam for details); observer-to-monkey distance for the first animal sighted; perpendicular distance from track to first monkey sighted; height (above ground level) of monkey; initial activity (i.e. the activity in which the monkey was engaged when first observed); and cue by which the group was detected. The location, time and direction of travel were marked on a map. An estimate was made, where possible, of group size and the age/sex class (refer to Section 2.3.1 for descriptions of categories) composition. Sightings of single animals were not included in analyses of group density.

Group density was determined by the Fourier series estimator on ungrouped data (Burnham et al. 1980) which takes into account transect length and the perpendicular distance from track to first monkey sighted for each group. Group density determined in this way may be an overestimate where the first animal sighted is the individual on the group's periphery nearest the observer and where Chapter 4: Group size, population density, biomass and social organization 94

group s1ze and/or group spread is large (Whitesides et al. 1988). However, as neither group size nor group spread of T. auratus was large, and the first individual sighted was not necessarily the individual nearest the observer, this was a relatively minor problem.

The Fourier estimate was compared with the density of groups calculated using the "effective distance" (D) estimated from the histogram-inspection method (after Whitesides et al. 1988) where

~ D = Nr FD

N1 = total number of sightings of groups Nr = number of sightings of groups at distances less than the fall-off distance FD = fall-off distance i.e. the first interval at which the number of groups detected dropped to half or less that of the immediately previous interval

Gt and group density estimate = 2(S72 + D)L1 where G = number of groups seen S = group spread

L 1 = total length of transect samples

4.2.2 TW Censuses

Due to the comparatively smaller area of the TW it was possible to census all T. auratus groups occurring here. This was achieved by two observers simultaneously traversing parallel lines with a maximum distance of 50 m between lines. Twenty-five metres was regarded as an acceptable maximum sighting distance. Mapped positions of groups, direction of travel by each group and notes on the size and composition of groups encountered by each observer ensured that a group was not counted twice by different observers. Data collected were the same as for the CA censuses described Chapter 4: Group size, population density, b.iomass and social organization · 95

previously. After group density in the TW was known, three censuses using the line transect method (transect length = 3.15 km) · were conducted to determine the accuracy of this method in determining group density.

4.2.3 Social Organization of T. auratus

Trachypithecus auratus groups in the TW were generally more easily observed than groups occurrring in the CA. This was because several groups in the TW were, at least partially, habituated to the presence of humans. Also, the vegetation in the TW was generally more open than in the CA and observation conditions were correspondingly better in the TW. Analysis of social organization was, therefore, largely restricted to the two study groups, other groups in the TW and occasional reference to groups in the CA.

4.3 POPULATION DENSITY AND BIOMASS OFT. AURATUS IN THE PANGANDARAN NATURE RESERVE

4.3.1 Population Density and Biomass in the CA

An average of four groups were recorded per census (n = 35; SD = 3.43; variance = 1.85) at a mean perpendicular distance of 15.9 m (SD = 14.52; SE = 1.23) from the track. The density of groups in the CA was estimated at 0.17 groups ha-l (SE = 0.02; 95% CI = 0.12- 0.21) when using ungrouped data in the Fourier series estimator. This corresponds to a density of 17 groups km-2 with an average home-range area of 5.9 ha per group (assuming home-range areas did not overlap). The Fourier series cannot calculate a goodness of fit on ungrouped data. However, an identical density estimate was obtained from grouped data for which observed values did not differ from expected values (X 2 = 4.44, 3 d.f., p > 0.20). Excluding areas cleared for grazing for the banteng, Bos javanicus, the CA consists of 440.5 ha. On this basis, there were estimated to be 75 T. auratus groups in the CA. Chapter 4: Group size, population density, biomass and social organization 96

By comparison, the group density estimate obtained using the effective distance estimated from the histogram-inspection method (Whitesides et al. 1988) was 11.3 groups km-2 where the fall-off distance was 10 m and the group spread was assumed to be 20 m.

No difference was found between the mean number of groups sighted in the morning and the number of groups sighted in the afternoon (t = 1.82, 28 d.f., n.s.) nor between the mean number of groups sighted for the two halves of the transect census (3 km each) (t = 0.29, 28 d.f., n.s.). Bias introduced as a result of the time of sampling was, therefore, considered negligible. Furthermore, groups were considered to be fairly evenly distributed throughout the CA since the mean number of groups sighted in the two halves of the transect census was not significantly different.

Accurate counts of the number of individuals in groups m the CA proved difficult as these groups were not habituated. However, reliable estimates were obtained for six groups during the first six censuses which were conducted on three consecutive days. These groups were regarded as distinct groups because they were located at a distance of 2 500 m from one another and it was, therefore, unlikely that one group was recounted on a different day. The groups consisted of 10, 7, 15, 10, 11 and 9 individuals yielding a mean of 10.3 individuals per group (SD = 2.7). This gives a population density of 175 animals km-2 (using the group density estimate obtained from the Fourier method).

Davies (1984) provides the weight of an average individual in a group to overcome the problem of differences in group size and body size between different colobine species, i.e. the weight of an average individual was considered more appropriate than adult female weight in the calculation of biomass. However, it is not clear how Davies (1984) calculated the weight of an average individual as weights of sub-adults, juveniles and infants were not given for any of the species listed and adult male and/or adult female weights were not presented for five (of the twelve) species. The weight of an average individual for T. auratus is not known, as weights for infants, juveniles and sub-adults are not available for either Chapter 4: Group size, population density, biomass and social organization 97

T. auratus or the closely related (and similar sized) T. cristata. However, for the purpose of subsequent comparisons of colobine biomass at different localities (Section 4.5), a value of 4.5 kg is used as an average individual weight of T. auratus as seven of the eight Presbytis and Trachypithecus spp. listed by Davies (1984) had an average individual weight of 4.5 kg. On this basis, biomass of T. auratus in the CA was estimated at 788 kg km-2.

4.3.2 Population Density and Biomass in the TW

Seven T. auratus groups were always located in the TW and four groups had home-range areas which extended partially into the TW and partially into the CA. Groups were recognised on the basis of: location, number of individuals; age/sex class composition (refer to Section 2.3.1); characteristic individuals (for example, an adult female in GRP2 had a distinctive kink in her tail); and colouration of infants.

In April 1984 reliable counts of nine groups were obtained. The number of individuals in these groups was 21, 21, 14, 13, 10, 17, 15, 13 and 6 (Table 4.1) (x = 14.4; SD = 4.9).

During three transect censuses (conducted along a trail measuring 3.15 km in length) 4, 5 and 3 groups (X = 4) were recorded at a mean perpendicular distance of 25.0 m. The density of the groups in the TW was thereby estimated at 0.25 groups ha-l. This corresponds to a density of 25 groups km-2 with an average home-range area of 3.9 ha per group. As the TW consists of 37.7 ha, from the censuses it was estimated that there were 9. 7 groups in the TW. This corresponds closely to the number of groups known to exist in the TW as discussed above. The line transect or strip method and the Fourier series estimator were therefore considered reliable in the estimate of group density obtained for the CA. Biomass of T. auratus (again based on average individual weight) in the TW was estimated at 1620 kg km-2. Chapter 4: Group size, population density, biomass and social organization 98

4.3.3 Evaluation of Methods

Whitesides et al. (1988), in a comparison of methods used for estimating primate densities, found no significant differences between estimates produced by the hazard-rate model, estimates derived from the transect-width estimation technique (histogram­ inspection method and graphical method) and estimates derived from the calibration techniques. However, the Fourier series model was not recommended (Whitesides et al. 1988).

Application of the Fourier model in estimating group density of T. auratus showed good agreement with known group density in the TW (Section 4.3.2). Furthermore, the observed probability density function for the CA was in close agreement with the expected density function generated by the Fourier series (Section 4.3.1). This is unlikely to have been true for five of the seven primate species examined by Whitesides et al. (1988) as the maximum number of groups sighted for each of these spectes was not in the distance interval nearest the transect line. Therefore, comparisons of group density in the TW and CA, and subsequent comparisons of biomass, are based on group density estimates derived from the Fourier model and not from the transect-width estimation technique (histogram-inspection method) which underestimated group density (Section 4.3 .1 ).

4.3.4 Comparison Between CA and TW

Trachypithecus auratus groups in the TW had, on average, a greater number of individuals than groups in the CA (14.4 and 10.3 individuals per group, respectively), although this difference was not significant (Mann-Whitney U test). Conversely, home-range areas were, on average, smaller in the TW than in the CA (3.9 ha and 5.9 ha, respectively). Consequently, biomass of T. auratus in the TW was twice that in the CA (1620 kg km-2 and 788 kg km-2, respectively). Chapter 4: Group size, population density, biomass and social organization 99

As groups in the CA were not habituated, it is possible that group sizes were underestimated. Even if groups in the CA consisted of the same average number of individuals as groups m the TW (i.e. 14.4 individuals per. group) biomass of T. auratus in the TW would still be 1.5 times T. auratus biomass in the CA.

Indian Indian Oceen Ocean

s s s

~gend ~ - Tectontl grt~ndis N OJ] - Act~citl tlt/riclllif'ormis ~ - SJoJ·'ietenitl mt~croph!JIIt! ~ - T.grt~ndis and S.mt~crophyllt~ + 0 - Secondary forest ~ -Border of Taman Wiseta D - GRPl Seal• (m) D - GRP2 0 100 200 D - GRP6 - Approx. home-range boundary ~ of GRP3

Figure 4.2 Locations of large Trachypithecus auratus groups in the Taman Wisata (coloured lines enclose all sightings for each group) Chapter 4: Group size, population density, biomass and social organization · 100

An interesting association is apparent when considering group s1ze and location of groups in the TW. Three groups (GRP1, GRP2, GRP6) were noticeably larger than the other groups in the TW (Section 4.3.2). The approximate home-range areas of these groups incorporated large sections of plantation, particularly Tectona grandis (Fig. 4.2). The GRP3 home-range area also contained plantation, but to a lesser extent (Fig. 4.2). Home-range areas of remaining groups in the TW contained little or no plantation. When the median size of the former three groups (21 monkeys) was compared with the median size of groups found in the CA (1 0 monkeys) group size was significantly greater for TW groups associated with large areas of T. grandis (U = 18, n 1 = 3, n2 = 6, p < 0.05).

4.4 SOCIAL ORGANIZATION

4.4.1 Group Composition and Behaviour of Bisexual Groups

4.4.1.1 Composition of T. auratus Groups

Age/sex class composition of nine groups in the TW is presented m Table 4.1. Group size varied between 6-21 with seven groups having one adult male and two groups having two adult males. Where two adult males were present, one adult male was clearly dominant; the other adult male remaining on the periphery of the group and/or being displaced by the former. Several groups had one or two sub-adult males. Adult females formed the largest age/sex class in all groups. Juveniles and infants formed the remainder in each group.

4.4.1.2 Study Groups

In June 1985, fourteen months after the counts given in Table 4.1, GRP3 consisted of 17 individuals: the number of adults was the same; the number of sub-adult females had risen by two to four, presumably recruited from the juvenile class; there were three (') ::r' Table 4.1 Age/Sex Composition of Groups in the Taman Wisata (April 1984) ~ 't:l ...CD "'i ~ GROUP ADULTS AM AFI AF2 u SUB SAM SAF u JUVENILES JM JF u INFANTS 0 c B TOTAL ADULTS 0 "'i 0 't:l= 11 4 4 2 4 2 2 2 2 4 2 2 21 "'j:;j" ~ 't:l 13 2 10 4 3 2 2 0 2 2 21 't:l =iii" ...c;· 3 8 6 2 2 3 2 14 1:1 Q. CD 1:1 4 10 2 4 3 2 2 0 13 ..."'..... ':= c;·cr SA 4 2 0 6 e "' 6 3 2 0 3 10 ~ SB 2 1:1 Q.

"'0 6 1 1 2 8 3 2 2 2 17 e:0 0 "'i 7 8 2 3 2 4 2 2 2 15 OQ ~ 1:1 j:;j" a 14 6 2 3 2 3 3 2 13 c;· 1:1

AM = adult male u = unidentified JF = juvenile female AFt = adult female with infant SAM = sub-adult male 0 = orange infant ..... AF2 = adult female without infant SAF = sub-adult female c = changing infant (orange/black) .....0 JM = juvenile male B = black infant Chapter 4: Group size, population density, biomass and social organization 102

juveniles, presumably one juvenile from the April count and two recruits from the infant class one of which was present in April and the other of which (a male) was born in October 1984; two infants which were born between March and June 1985 outside the period during which this group was observed on a monthly basis. In November 1987, during a brief visit to Pangandaran, GRP3 consisted of one adult male, two adult females with infants, five adult females without infants, three sub-adult females and two juveniles. Therefore, over a period of three and a half years, the group count only fluctuated between 14 and 17. It is not known what group changes occurred between June 1985 and November 1987, however, the adult male was recognized and was the same male present in June 1985.

In July 1984, GRP21 consisted of 15 individuals: one adult male; one adult female with an orange/black infant; seven adult females without infants; two sub-adult females; three juveniles. One infant was born in October 1984 (as for GRP3). In June 1985 the group count was also 15 and the age/sex composition the same as in June 1984 except that there was one more sub-adult female (possibly recruited from the juvenile class) and there were no infants. Presumably, a juvenile or infant had died (or emigrated) in the interim period.

The above is the simplest explanation for changes in group composition which occurred after the period of intensive study. It is possible that immigration and emigration may also have resulted in the changes described but, in the absence of continual observation of the groups and the impossibility of individual recognition, it is not known to what extent immigration and emigration affected group composition. Throughout the intensive study period, however, the only changes noted were the births in October and the recruitment of the infants present at the start of the study to the juvenile class. Chapter 4: Group size, population density, biomass and social organization 103

4.4.1.3 Reasons for One or More Adult Males in Bisexual Groups

Curtin and Dolhinow (1978) proposed that one-male groups are favoured under stressful conditions and that the more usual form of multiple males in bisexual groups occurs under less stressful conditions. Generally, however, the presence of multiple males has been attributed to increased predation pressure. Clutton-Brock and Harvey (1978) challenged this view in speculating that the number of males in a group may depend on the costs and benefits to a dominant male rather than being a response to predation pressure. In an (initially) single male group, the cost of allowing an extra male to join the group or of consistently driving the "intruder" away, would depend on the benefit of having another male in the group (for example; protection of females and young against other intruders) which may also depend on the genetic relationship of the two males. Where the "intruder" is related to the resident male this would presumably increase his chance of being accepted into the group. From a genetic viewpoint, a related male is more likely to assist the resident male in protecting offspring and assisting group cohesion where this may enhance inclusive reproductive success. Furthermore, if the intruder were to sire offspring in the group, a genetic benefit to the resident male would still accrue where the two males are closely related.

4.4.1.4 Group Composition of Other Colobines

Groups with a single adult male (Bernstein 1968; Wolf 1980) and groups with 2-3 adult adult males (Furuya 1961-2) have been reported for the closely related T. cristata.

Colobines usually having unimale bisexual groups include: T. obscura (although up to three adult males in a group has been observed) (Curtin 1980; MacKinnon and MacKinnon 1980; Raemaekers and Chivers 1980); T. johnii (Poirier 1969; Horwich 1972; Oates et al. 1980); T. senex (Rudran 1973); P. aygula (Rodman 1978); P. comata (Ruhiyat 1983); P. hosei (Payne and Davies 1982); P. melalophos (Curtin 1976; Bennett 1983); Chapter 4: Group size, population density, b~omass and social organization ·1 04

P. rubicunda (Payne and Davies 1982; Davies 1984; Supriatna et al. 1986); P. thomasi (Rijksen 1978; Gurmaya 1986); P. pileata (Green · 1981; Islam and Husain 1982); C. angolensis (Groves 1973); C. guereza (Dunbar and Dunbar 1976; Oates 1977a, b; Struhsaker and Leland 1979).

Colobus badius rufomitratus live in groups with 1-2 adult males (Marsh 1979a, b), C. satanas groups contain 1-3 adult males (McKey 1978; McKey and Waterman 1982) and C. badius tephrosceles groups have between 1-11 adult males (Clutton-Brock 1972; Struhsaker 1975). Presbytis phayrei groups have 1-4 adult males (Mukherjee 1982). The number of males in P. entellus groups ts highly variable, ranging from one to many (Sugiyama 1964; Dolhinow 1972; Starin 1978; Hrdy 1980). Presbytis potenziani (Tilson and Tenaza 1976; Anon 1980) and Nasalis concolor (Tilson and Tenaza 1976; Watanabe 1981) form monogamous pairs and, as such, there is only one adult male per group.

4.4.1.5 Adult Males

The loud call of an adult male T. auratus (described in Section 1.2.2) was issued in response to a disturbance, usually when a neighbouring T. auratus group was in close proximity. Sometimes this preceded a chase through the canopy involving two adult males, one the resident male and the other an intruder male from a neighbouring group. Physical contact between adult males was not observed and, after the chase, one of the groups would withdraw from the immediate area where the conflict had taken place.

On several occasions a chase between two adult males ensued without any prior calling. On these occasions it appeared that an intruder male had come within 20 m of the group without being noticed. The first sign of unrest was the squealing of juveniles and infants, upon which either the resident male or the intruder male commenced chasing the other. Chapter 4: Group size, population density, biomass and social organization 105

At other times, adult males of neighbouring groups issued loud calls alternately which were not followed by a chase. If a component of the loud call is indicative of the size (or strength?) of an adult male, the function of these calls may be to indicate whether a potential intruder should enter a pursuit or not (Rijksen pers. comm.). The loud call may also function in the maintenance of intergroup spacing where groups are territorial.

Intergroup displays by adult males are common m colobines (Poirier 1974; Dunbar and Dunbar 1976; Marsh 1979b; Struhsaker and Leland 1979; Bennett 1983; Davies 1984). In addition, loud calls have also been associated with time of day. For example, adult male P. melalophos calls were most frequent around dawn, dusk (and at intervals during the night) (Curtin 1980; Bennett 1983). Loud calls of adult male T. auratus were not issued at specific times of day.

The behaviour of adult males calling and chasing is similar to that described for P. me lalophos (Bennett 1983). However, P. melalophos (Bennett 1983) were not regarded as highly territorial as there was considerable overlap in home-range areas. The function of the adult male call for P. melalophos was, therefore, considered to be the deterrence of rival males and the attraction of females (Bennett 1983). Conversely, there was little overlap in the home ranges of T. auratus groups (see Chapter 8) and it is considered that a main function of the adult male call for this species was in defence of home-range boundaries. Similarly, adult males of T. cristata and T. obscura vocalize less than P. me lalophos (Wilson and Wilson 1976) but home ranges of the former two species have little or no overlap (Chivers 1973; Curtin 1976; Wilson and Wilson 1976).

4.4.1.6 Adult Females

Two aspects of the social behaviour of adult females, of interest here, are receptiveness of adult females and allomothering. On two separate occasions, whilst observing the study groups, an adult Chapter 4: Group size, population density, biomass and social organization 106

female T. auratus was seen to make rhythmic side-to-side head movements. On one occasion the female then presented her hindquarters to the male, the male mounted and copulated. The female continued the head movements during the mounting. This sequence of events was repeated on the following day. On the other occasion, mounting and copulation were not observed. This behaviour is identical to a description given by Bernstein (1968) for T. cristata. Bernstein (1968) considered that the head shaking of the female indicated solicitation for copulation by a receptive female. Only one other incidence of copulation was observed for T. auratus and this was not preceded by the side-to-side head movements of the female.

Allomothering, in which an infant is temporarily cared for by an individual other than the mother is common among colobines (Horwich and Manski 1975; review by Hrdy 1978). Bennett (1982) recorded the incidence of allomothering to twins born to a P. melalophos female. It is generally regarded that allomothering provides experience in handling of infants for nulliparous females. Some allomothers are reckless and negligent towards infants, such behaviour usually resulting in the infant returning to its mother. The combined social and genetic benefits derived from allomothering behaviour may account for the origins of this behaviour in species where intragroup feeding competition is reduced (McKenna 1979).

Allomothering in T. auratus groups was also common although, m the absence of individual identification, it was generally not possible to determine which was the mother and which the allomother where the latter was also an adult female. Sub-adults and juveniles were also observed allomothering. Juveniles, in particular, often appeared incompetent and this usually resulted in the infant squealing and being retrieved by an adult female. Adult males were not observed to physically care for infants. Chapter 4: Group size, population density, biomass and social organization 107

4.4.1.7 Infants

From observations of infants in the T. auratus study groups and other groups in the TW it was found that infants have either completely orange natal coats at birth or have predominantly orange natal coats with a darker mid-dorsal strip extending down into the tail. Infants lack pigmentation in the face, hands and feet at birth. After one or two weeks these areas are pigmented and similar to that of other age/sex classes. Hair colour change usually begins at the head, tail, limbs and mid-dorsal strip, the lateral and ventral surfaces of the body being the last areas to take on the silver/black colour of the adults. The time for complete colour change is about four months, after which the infant remains dependent on an adult female for a further four to five months. A similar period of time for infant colour change has been noted for T. cristata (Bernstein 1968; Medway 1970b; Roonwal and Mohnot 1977).

The infant attracts most attention from other group members when it is largely orange in colour. One infant, less than three days old, was repeatedly "snatched" from an adult female by other group members. This was usually shortlived as the infant would scream upon which it was retrieved by the adult female.

A record was kept of all infants sighted during transect censuses in the CA, during counts of groups in the TW and during non­ systematic observations. Age of infants was estimated from coat colour. From observations of 31 infants it appeared that infants were born throughout the year with possible birth peaks in October (16% of births) and March (32% of births). Further systematic investigation of a large number of groups is required to determine whether birth peaks occur at certain times of year or whether the apparent peaks were a result of non-systematic sampling in this study.

The presence of a strict breeding season may relate to seasonal influences but evidence for this is scant. Birth peaks have been indicated for a few colobine species (Struhsaker and Leland 1986) Chapter 4: Group size, population density, biomass and social organization · 108

but not for others. For example, Medway (1970b) found no evidence of a breeding season for T. cristata in Malaya as births occurred with equal frequency throughout the year. However, Furuya ( 1961-2) reported that, in western Thailand, births in T. cristata groups occurred between December and May. Presbytis melalophos in West Malaysia produced infants throughout the year (Bennett 1983) as did P. thomasi in North Sumatra (Gurmaya 1986), whereas P. rubicunda in Sabah did not give birth between February and August (Davies 1984). Two birth peaks per year have been reported for T. johnii (Poirier 1970) and C. badius tephrosceles (Struhsaker 1975).

4. 4. 2 Solitary Males, All Male and Predominantly Male Groups

4.4.2.1 General

Where bisexual groups consist of one (or a few) adult males and predominantly adult females, as for most colobines, one or more of the following factors would account for the difference in the sex ratio:

1. differential sex ratio at birth;

2. increased male mortality rate;

3. emigration of males from bisexual groups.

Evidence for the first two factors is scant, however, it appears that the latter factor may account, at least partly, for differences in the sex ratio of bisexual groups. Consequently, solitary males and all male (or predominantly male) groups would be expected.

In considering the reasons for the formation of all male groups, as opposed to males remaining solitary, the following advantages have been considered likely to exert an influence (Bennett 1983): Chapter 4: Group size, population density, biomass and social organization 109

1. defence against predators (Struhsaker 1969);

2. mutual asistance m take-overs of bisexual groups (Wolf and Fleagle 1977);

3. protection against harem males which may harass solitary males thereby removing the solitary males from valuable food sources (Crook and Gartlan 1966);

4. efficient foraging (the same advantages would apply as were discussed for increased group size in bisexual groups in Section 4.1).

A possible reason for females with infants leaving a bisexual group and joining an all male group (giving rise to predominantly male groups) is to reduce the risks of infanticide occurring. This is discussed further in Section 4.4.3.

4.4.2.2 Trachypithecus auratus

One solitary sub-adult male was located in the TW in January 1985; in July 1984 a sub-adult male and an adult male were found together and apparently not associated with another group (i.e. no group present within 100 m radius). No other sightings of solitary or all male groups were recorded for the TW.

In the CA, solitary males were sighted on twelve occasions whilst conducting transect censuses. On the basis of locality, these sightings were regarded as being of two individuals. On one occasion two adult males were observed together at a different location to the sightings of the solitary adult males.

An all male group was sighted in July 1984, neighbouring GRP21. This group consisted of six adult males, two sub-adult males and two juvenile males. In December 1984 the count of this group was the same. In June 1985 this group consisted of four adult males, one sub-adult male, one juvenile male and two females - an adult Chapter 4: Group size, population density, biomass and social organization 110

and a juvenile. It is not known what became of the remainder of this group nor from where the females originated.

4.4.2.3 Other Colobines

Whilst Bernstein (1968), in a study of T. cristata, did not observe any solitary males or all-male groups, the occurrence of solitary males and all male groups is widespread amongst colobines. For example; solitary males have been seen for T. cristata (Furuya 1961-2) and P. melalophos (Bennett 1983), and all-male groups have been observed for P. phayrei (Mukherjee 1982), P. rubicunda (Davies 1984), P. melalophos (Bennett 1983) and P. thomasi (Gurmaya 1986).

The existence of predominantly male groups is also widespread. Rudran (1973) observed an immature female in an otherwise all male group of T. senex Female transference has been noted for C. badius rufomitratus (Marsh 1979a), C. badius tephrosceles (Struhsaker and Leland 1979), T. johnii (Poirier 1969), P. rubicunda (Davies 1984) and P. melalophos (Bennett 1983).

4.4.3 Infanticide

Observations on several colobine species show that resident males in bisexual groups have been ousted from their groups by an intruder male(s). Shortly after this event the "disappearance" or death of infants in the group has been noticed. Where the death of infants has resulted, the "new" adult male has been considered responsible. This is referred to as infanticide. Numerous accounts of infanticide have been reported (for example; in P. entellus (Mohnot 1971; Hrdy 1974; Makwana 1979; Sommer and Mohnot 1985), T. senex (Rudran 1973), T. cristata (Wolf and Fleagle 1977; Wolf 1980), C. bad ius (Struhsaker and Leland 1985) and see review by Hausfater and Vogel (1982)) although many are based on circumstantial evidence. Infanticide was not observed in T. auratus groups. Chapter 4: Group size, population density, biomass and social organization 111

Hrdy (1977, 1980) claimed that infanticide was evidence for sexual selection, in being an extension of competition among males. This claim has since been a source of dispute and criticism (for example, Schubert 1982). Boggess (1979) stated that there was no support for Hrdy's (1977, 1980) claim and that infanticide was a maladaptive behaviour occurring in isolated and rare situations or in populations characterized by extreme crowding. Chapman and Hausfater (1979) developed a mathematical model and concluded that infanticide would only be advantageous for resident males for particular lengths of tenure within bisexual groups and would, therefore, only become fixed in populations where tenure length provides this advantage. Possibly, adult males in populations characterised by crowding have short tenure lengths due to pressure exerted by males outside bisexual groups. Under such conditions, infanticide may be advantageous. Recently, the occurrence of infanticide in populations of P. entellus has been associated with unimale group structure and not with a high density (Newton 1988).

Proximate causes may explain abuse of infants after male­ takeover of a group without inferring sexual selection. Rijksen ( 1981) in a study of hamadryas baboons postulated that infant killing resulted occasionally from an action which had as its proximate goal the achievement of social control and dominance. Whilst this study dealt with captive animals and the infants were abused (and subsequently died) by their biological father, certain aspects may extend to field conditions. The lack of familiarity between the new male and the females, their nonacceptance of him, their attention drawn largely by their infants and the state of the new adult, may lead to attacks on infants (Angst and Thommen 1977) presumably aimed at recognition of the dominance of the adult male and thereby achieving social control.

4.4.4 Sexual Dimorphism

Polygynous species usually show a greater degree of sexual dimorphism in body size than monogamous species (Clutton-Brock Chapter 4: Group size, population density, ~iomass and social organization • 112

and Harvey 1977a). Leutenegger (1978) and Leutenegger and Cheverud (1982) noted a high degree of predictability from large body size to strong sexual dimorphism but a comparatively low degree of predictability for a few small- and medium-sized species, for example, Presbytis spp. Within colobines, there is a slight tendency for predominantly folivorous species to show a greater degree of sexual dimorphism than species which predominantly feed on fruits (and flowers) presumably because the latter items are located on smaller branches which would not support large body weight (Bennett 1983). Adult female weight for P. melalophos, a predominant frugivore, is 98% of the adult male body weight (Bennett 1983). Adult female weight for T. cristata, which is presumed to be predominantly folivorous, is 89% of the adult male body weight (Roonwal and Mohnot 1977). A further increase in sexual dimorphism for folivorous species is probably limited by accessibility to food supplies. Young leaves, located on terminal branches, may also be difficult to obtain where branches are not sufficiently flexible to permit bending by a primate using its hands, but too light to support a heavy load.

Body weights for T. auratus are not available but are likely to be similar to T. cristata body weights (see Section 1.2.2). Degree of sexual dimorphism for these two species is, therefore, also likely to be similar.

4.5 DISCUSSION

Group size ofT. auratus varied between 6-21. Trachypithecus cristata group size is also variable. In Malaya, group size of T. cristata has been found to vary between 12-21 (Medway 1970a), 22-48 (Furuya 1961-2) and 20-51 (Bernstein 1968). In western Thailand, group size has varied between 9-30 individuals (Fooden 1971) and in Sumatra a mean group size of 11 has been reported (Wilson and Wilson 1976).

Mean group sizes for several colobines are presented m Table 4.2. Generally, mean group size for Presbytis spp. and Chapter 4: Group size, population density, biomass and social organization 113

Trachypithecus spp. is between 5-15 (the obvious exception being P. entellus which may have considerably larger groups). There is a slight tendency for Trachypithecus spp. to have larger group sizes than Presbytis spp.

Biomass of colobines (calculated using average individual weight (Davies 1984; see Section 4.3.1) for six study sites where long-term

Table 4.2 Mean Group Size of Colobines

Species Mean group size Reference

Rainforest Trachypithecus johnii 15 Poirier 1969, 1970; Horwich 1972; Oates et al. 1980. T. obscura 14 Curtin 1980; MacKinnon and MacKinnon 1980; Raemaekers and Chivers 1980. Presbytis aygula (P.comata) 8 Rodman 1978; Ruhiyat 1983. P. hosei 7 Payne and Davies 1982. P. melalophos 15 Curtin 1976; Bennett 1983; Johns 1983. P. potenziani 5 Tilson and Tenaza 1976; Anon 1980. P. rubicunda 7 Payne and Davies 1982; Davies 1984. P. thomasi 9 Rijksen 1978; Gurmaya 1986. Colobus badius tephrosceles 50 Clutton-Brock 1972; S truhsaker 197 5. C. guereza 8 Dunbar and Dunbar 1974, 1976; Oates 1977a, b; Struhsaker and Leland 1979. C. satanas 15 McKey 1978; McKey and Waterman 1982. Gallery Forest C. badius rufomitratus 18 Marsh 1979 a, b. Deciduous Forest P. pileata 7 Green 1981; Islam and Husain 1982. Seasonal T. senex 9 Rudran 1973. P. entellus 5-120 Sugiyama 1964; Dolhinow 1972; (highly variable) Starin 1978; Hrdy 1980.

15

10 Chapter 4: Group size, population density, biomass and social organization 114

studies have been conducted (Table 4.3), varied markedly. In an attempt to explain these observed differences in biomass the influence of plant chemistry on colobine population densities has been assessed (Waterman et al. 19 8 8). This aspect is discussed further in Chapters 7 and 9.

Features which may influence biomass include availability of food, predation, disease, parasitism and irregular catastrophic events (Bradbury and Vehrencamp 1976 and see Section 4.1). Several of these factors may interact to set the population densities observed (Cant 1980). Differences between T. auratus biomass in the CA and TW are unlikely to be due to differences in disease and

Table 4.3 Colobine Body Weights and Biomass at Six Study Sites

Body weight (kg) Biomassa Total biomass of colobines Study site Species Adult male Adult female (kg km-2) (kg km-2)

RainfQrcst Douala-Edea Colobus satanas 10.5 153 153

Kibale C. guereza 7.0 10.8 64 C. badius 7.0 10.0 1785 1849

Kakachi Tracypithecus johnii 11.0-14.0 11.0 532 532

Kuala Presbytis Lompat melalaphos 5.9 5.8 486 T. obscura 7.4 5.5 390 876

Sepi1ok P. rubicunda 6.3 6.0 49 P. hosei 6.3 6.0 15 64

Sc~Qndar:x:: Fmcst Pangandaran T. auratus 7.1 6.3b 788b 788

0 Biomass estimates are based on the calculation of population density x group weight (where weight = number of animals per group x body weight of average individual - see Section 4.3.1) (Adapted from Waterman et al. 1988). hweights for T. auratus are not available and are therefore based on T. cristata weights (see Section 1.2.2.). Chapter 4: Group size, population density, biomass and social organization 115

parasitism, considering the proximity of the two areas and the overlap in group sizes. Predation pressure is also unlikely to account for differences in group size and biomass. Predation of T. auratus was not observed during this study and there was no indication that T. auratus have been hunted in the past by humans. However, it is not known whether T. auratus emigrated or died when the plantation areas were last logged over thirty years ago. Generally, though, langur group densities decline by emigration during logging but recover within 5-10 years to levels close to those in the original forest (Marsh and Wilson 1981 b; Johns 1986; Marsh et al. 1987). Therefore, it seems unlikely that the greater density of T. auratus in the TW is a latent (or gradual) response to the catastrophic event of logging.

The largest groups were located m areas with large stands of plantation (Section 4.3.4). Dramatic changes in weather conditions (a catastrophic event) are likely to have a similar effect on groups in the CA and TW with the following exception. In marked seasonal areas, T. grandis is deciduous and consequently has no leaves for a period of time (van Bemmel 1978). Throughout this study, spanning a period of two years, T. grandis always had mature and young leaves available. Pangandaran is not characterised by distinct seasons (Section 2.2.3), however, it is conceivable that during an extreme dry period T. grandis may lose all its leaves and this may be catastrophic to groups living in plantation areas if T. grand is forms a considerable proportion of their dietary intake. This is little more than speculation as such conditions are not known to have occurred in Pangandaran.

Variation in T. auratus biomass between the two areas may, therefore, be largely due to differences in food availability and food patch size. Larger groups occupying relatively small home ranges may be associated with the presence of large patches of food sources (Strier 1987). This aspect is examined in subsequent chapters based on observations of the two study. groups which were the same in group size and age/sex class composition but differed in habitat and home-range area. Chapter 4: Group size, population density, biomass and social organization ·116

4.6 SUMMARY

1. From 35 transect (6 km length) censuses in the CA, an estimate of 0.17 T. auratus groups ha-l (17 groups km-2) was obtained using the Fourier series estimator. This gives an average home-range area of 5.9 ha per group. Average group size was 10.3 animals; biomass was estimated at 788 kg-2.

2. There were seven T. auratus groups living wholly m the TW and four groups had home-ranges extending partially into the TW and partially into the CA. The density estimate was 0.25 groups ha-l (25 groups km-2), showing good agreement with known population density in the TW. Average home-range area of groups in the TW was 3.9 ha and there was a mean of 14.4 animals per group. Biomass was estimated at 1620 kg km-2.

3. The largest groups were found in the TW in home-range areas containing plantation, particularly Tectona grandis. Differences in the biomass of T. auratus in the CA and the TW may largely be due to differences in food availability.

4. Trachypithecus auratus group size varied between 6-21; groups usually had one adult male.

5. During the period of intensive study, the only social change which occurred in the study groups was the birth of an infant in October 1984 in each of the groups.

6. Group size for GRP3 varied between 14-17 over a period of three and a half years.

7. Adult males issued loud calls, as is common amongst Asian colobines. These were not synchronised with time of day.

8. When born, infants have an orange coat; time. for colour change to the silver/black of adults is approximately four months. Infants were born throughout the year. Chapter 4: Group size, population density, biomass and social organization 117

9. Solitary males, all-male and predominantly-male groups were observed. Female transference to a formerly all male group was noted. Chapter 5: Activity budgets 118

CHAPTER 5 ACTIVITY BUDGETS

5 .1 INTRODUCTION

The proportion of time an animal, or a group of animals, spends in various activities determines an activity budget. Activity budgets of Trachypithecus auratus were calculated to compare this species with other colobines and to compare the two study groups, GRP3 and GRP21. Comparisons have largely dealt with the major activities of moving (as in group progression), feeding and inactivity; they are a precursor to further comparisons made in analyses of feeding (Chapter 6 and Chapter 7) and ranging (Chapter 8) behaviour.

5.2 METHODS

Scan sampling (Altmann 1974) was the mam method used to determine activity budgets. Focal-animal sampling (Altmann 1974) was used to provide a companson with results obtained using the former method and to examine age/sex class variation in activity budgets. These methods have been described fully in Sections 2.4.1, 2.4.2 and 2.4.3. Descriptions used in the classification of activities and age/sex classes are found in Section 2.4.1 (Table 2.3) and Section 2.3.1 (Table 2.2), respectively.

Infants have not been included in analyses of activity because they may introduce bias into an analysis of group activity due to their comparatively rapid development and changing degree of dependence on an adult female. Infants were not observed during focal-animal sampling. Chapter 5: Activity budgets 119

5. 3 ACTIVITY BUDGETS OF GRP3 AND GRP21 (SCAN SAMPLING)

5.3.1 Comparisons Between GRP3 and GRP21

Activity budgets for both study groups are shown in Figure 5.1. Feeding (FI + FB), moving (MW + MA) and inactivity (IH + I) were the major daily activities for both GRP3 and GRP21. The remaining activities: playing; vocalising; self grooming; social grooming; other (for example, intergroup encounters) contributed only small proportions to the total activity budgets (Fig. 5.1).

Inactive (huddling), in which two individuals embraced one another while resting, has been reported for other colobines (for example, Presbytis melalophos and T. obscura (Curtin 1976); P. rubicunda (Davies 1984)). Inactive (huddling) was observed more often in scans of GRP3 than GRP21 (Fig. 5.2) (U = 450, n1 =

25, n2 = 25, p < 0.01). Observations of other T. auratus groups, collected on an opportunistic basis throughout the study, suggest that this was a more common form of resting for GRP3 than for other T. auratus groups. Reasons for this are unknown and inactive (huddling) is subsequently classed under "inactive", being the broader category to which this behaviour belongs.

When comparing the proportional distribution of time spent m the major activities (Table 5.1), the two study groups differed from one another (X 2 = 22.97, 3 d.f., p < 0.001). The amount of time spent moving (MW), as when travelling from one food source to another within the same tree, was similar for both study groups (Fig. 5.2). However, GRP21 spent more time (as unweighted % of total activity budget) travelling (MA), as in group progression, than GRP3 (Fig. 5.2) (U = 424, n1 = 25, n2 = 25, p < 0.05). GRP21 also spent more time inactive than GRP3 (Fig. 5.2) (U = 448, n1 =

25, n2 = 25, p < 0.01). GRP3, however, spent a greater proportion of time feeding than GRP21 (Fig. 5.2) (U = 488, n1 = 25, n2 = 25, p < 0.01). Chapter 5: Activity budgets 120

a) GRP21

• F (feeding) II MW (moving within tree canopy) Iii MA (moving-group progression) fZJ IH (inactive (huddling)) 0 I (inactive (excl. IH)) • 0 (other) m p (playing) b) GRP3 fii V (vocalising) I2J GS (self-grooming) 0 G (social grooming)

Figure 5.1 Activity budgets: GRP3 and GRP21 Chapter 5: Activity budgets 121

60

50

40

CD E - 30 0~ 20

10

0 F MW MA IH

Activities Figure 5.2 Proportion of time (as unweighted percentage) spent in major activities (median daily percentage and IQ range) Activity abbreviations same as in Fig. 5.1 except that I includes IH

Table 5.1 Proportion of Time (as Sum of Proportion of Each Scan) Spent in the Major Activities by GRP3 and GRP21 (Overlapping Months Only i.e. Five Months for Each Study Group)

Activitya GRP3 GRP21

Feeding (FI + FB) 499.9 370.1

Inactivity (I + IH) 789.3 839.5

Moving (MA + MW) 307.0 335.2

All other activities 80.8 62.7 (P, V, GS, G + 0) a refer to text for details of activity abbreviations

When examining the daily activity budgets (using maJor activities) of each study group, significant differences m proportional representation of activities were found (H = 58.07, 2 d.f., p < 0.001 (GRP3); H = 95.72, 2 d.f., p< 0.001 (GRP21)). GRP3 and GRP21 both spent most of their time inactive (48% (median) Chapter 5: Activity budgets 122

for GRP3 and 53% (median) for GRP21 (Fig. 5.2)). Feeding was the second most common activity (27% (median) for GRP3 and 24% (median) for GRP21) and travelling occupied the least time (13% (median) for GRP3 and 17% (median) for GRP21) of the major activities considered (Fig. 5.2).

5.3.2 Monthly Variation in Activity Budgets

The proportion of time spent inactive varied monthly for both GRP3 and GRP21 (Table 5.2). However, neither the amount of time spent feeding, nor the amount of time spent moving (MA), varied monthly for either of the T. auratus study groups (Table 5.2). The latter is of relevance to an ecological study of feeding behaviour when exammmg monthly variation in food availability and is discussed further in Section 6.8.

Table 5.2 Monthly Variation in Activity Budgets (for Major Activities) Using Kruskal-Wallis One-Way Anova

Study Group Activity No. of Sample size lP d.f. Probability months per month (days)

GRP3 Feeding 8 5 8.17 7 n.s.b

Moving (MA) 8 5 14.02 7 n.s.b

Inactive 8 5 16.69 7 p < 0.01

GRP21 Feeding 5 5 7.34 4 n.s.b

Moving (MA) 5 5 6.23 4 n.s.b

Inactive 5 5 11.61 4 p < 0.05 aH approximates x2 distribution bn.s. = not significant Chapter 5: Activity budgets 123

5.4 ACTIVITY BUDGETS OF OTHER GROUPS IN THE TAMAN WISATA

Observational data were collected (scan sampling) for two other groups (GRPl and GRP2) in the TW to provide a comparison with the GRP3 activity budget (Section 2.4.2). The proportional distribution of time spent in the major activities (Table 5.3) did not differ for the three T. auratus groups (X 2 = 11.43, 6 d.f., n.s.). The GRP3 activity budget may, therefore, be considered representative of other T. auratus groups found in the TW. Unfortunately, due to the lack of other habituated groups in the CA, it was not possible to make a similar comparison for GRP21.

Table 5.3 Proportion of Time (as Sum of Proportion of Each Scan) Spent in Major Activities by GRP3 and Two Other T. auratus Groups Located in the Taman Wisata

Activitya GRP3 GRP2 GRP1

Feeding (FI + FB) 807.2 54.0 54.2

Inactivity (I + IH) 1196.1 54.9 61.4

Moving (MA + MW) 571.9 45.7 42.5

All other activities 138.9 9.4 8.5 (P, V, GS, G + 0)

a Refer to text for details of activity abbreviations

5.5 COMPARISON BETWEEN FOCAL-ANIMAL SAMPLING AND SCAN SAMPLING METHODS

The proportions of time (as %) spent feeding (FI + FB), moving (MA + MW), inactive (I + IH) and in all other activities (P, V, GS, G + 0) by GRP3, as estimated by the scan sampling and focal-animal sampling methods, were calculated for each month. The monthly median for the proportion of time spent feeding was estimated at 30% and 32% for the scan and focal-animal sampling methods, respectively. The proportion of time spent in all other activities Chapter 5: Activity budgets ·124

was estimated at 5% and 6% for the scan and focal-animal sampling methods, respectively. The differences in estimates for these activities, using the two methods, were not significant (U =

30, n1 = 7, n2 = 7, n.s. and U = 30, n1 = 7, n2 = 7, n.s., respectively).

Estimates of the proportion of time spent moving and inactive, however, varied between the two methods. The (median) proportion of time spent moving was estimated to be 20% and 9% for the scan and focal-animal sampling methods, respectively (U =

48, n 1 = 7, n2 = 7, p < 0.01); and 41% and 53% as the proportion of time spent inactive for the scan and focal-animal sampling

methods, respectively (U = 44, n 1 = 7, n2 = 7, p < 0.05). As age/sex classes were sampled in proportion to their true representation in GRP3 for both the focal-animal sampling method (Section 2.4.3) and the scan sampling method (X 2 = 6.48, 4 d.f., n.s.), differences are not due to sampling variation of the age/sex classes for the two methods.

Focal-animal sampling provides a more accurate estimate of activity budgets than scan sampling as the former method involves continuous observations of animals for predetermined periods of time. By comparison, therefore, scan sampling overestimated time spent moving and underestimated time spent inactive. Animals moving are more likely to be noticed and recorded, when conducting instantaneous scan samples, than animals which are inactive. Attempts were made to overcome this potential bias by routinely scoring all animals observed instantaneously as well as those encountered within five minutes (Section 2.4.2). Possibly this reduced further bias but the bias was not eliminated completely.

As estimates of the proportion of time spent feeding did not differ significantly between the two methods, subsequent analyses based on the proportion of time spent feeding using the scan sampling method (for example, the proportion of time spent feeding on various species and items (see Chapter 6)) are regarded as providing a true representation of feeding budgets. As there is no reason to suppose that bias introduced from the Chapter 5: Activity budgets 125

scan sampling method in estimating the proportion of time spent moving and inactive, was different for the two study groups, activity budgets of the two groups may be justifiably compared (as was presented in Section 5.3.1).

5. 6 ACTIVITY BUDGETS ACCORDING TO AGE/SEX CLASS (FOCAL-ANIMAL SAMPLING)

The distribution of time (as the sum of minutes) spent in the major activities for each of the age/sex classes in GRP3 (Table 5.4) differed significantly (X2 = 359.31, 12 d.f., p < 0.05). Activity budgets of each of the age/sex classes are shown in Figure 5.3. The van der Waerden (Normal Scores) test (Conover 1980, p. 318) was used to test the hypothesis that all age/sex classes had the same distribution functions for each of the major activities. The null hypothesis was rejected for the following activities: feeding; inactive; other (Table 5.5). A multiple comparisons test (Conover 1980, p. 319) showed that adult females (without infants) differed significantly from the adult male and juveniles in the amount of time spent feeding (Table 5 .5). The adult male and juveniles spent less time feeding than adult females (Fig. 5.3).

Table 5.4 Time (Min) Spent in the Major Activities by Each of the Age/Sex Classes (Excluding Infants) in GRP3, Using the Focal-Animal Sampling Method

Age/Sex Class a

Activity AM AFl AF2 SAF J

Feeding 242 316 819 298 320

Moving 80 95 193 119 184

Inactive 652 486 1101 550 554

Other 53 30 36 63 169

a Age/sex class abbreviations defined in Table 2.2 (Section 2.3.1) Chapter 5: Activity budgets 126

20

10 Other

0

20

10 Moving

0 C1) -C) 'C :J 40 .Q -

I" T 30 ..L - 20 ~ ::>- Feeding > 0 10 - -m 0

70 ~0 60

50

40 Inactive 30

20

10

0 AM AF1 AF2 SAF J Age/Sex Class

Figure 5.3 Activity budgets of each age/sex class in GRP3 (median and IQ range) Chapter 5: Activity budgets 127

Table 5.5 Results of Normal Scores Test on Proportion of Time Spent in Each of the Major Activities by the Various Age/Sex Classesa

Activity d.f.C Probability Multiple test comparisonsd

Moving 6.90 4 0.141

Feeding 9.64 4 0.047 AF2 >AM AF2> J

Inactive 12.97 4 0.012 AM > AF1 AM > AF2 AM> SAP AM> J

Other 16.45 4 0.003 J >AM J > AF1 J > AF2 J >SAP AF2 >SAP a Age/sex class abbreviations defined in Table 2.2 (Section 2.3.1) bapproximates x2 c d.f. = no of age/sex classes minus one d approximates t-distribution (1-tailed) a = 0.025, 34 d.f. e n.s. = not significant

Adult males of other colobine species have similarly been reported to spend less time feeding than adult females (Clutton­ Brock 1977b; Bennett 1983) and this has been attributed to the cost in energy of pregnancy and lactation in females (Clutton­ Brock 1977b). The age/sex class AFl in this study included females with very young infants as well as females whose infants were approaching independence and feeding on vegetation. This may explain why, although AFls tended to spend more time feeding than the AM (Fig. 5.3), this difference was not significant. The greater proportion of time spent feeding by AF2s (Fig. 5.3), however, may be related to the cost of pregnancy.

There were no significant differences in the amount of time spent moving by the different age/sex classes (Table 5.5). However, juveniles showed a tendency to spend a greater proportion of time moving than the remaining age/sex classes (Fig. 5.3). This is largely due to the eratic movements of juveniles Chapter 5: Activity budgets "128

whilst jumping from one tree branch to another, when other members of the group were inactive or feeding.

Juveniles differed from all other age/sex classes in the amount of time spent in "other" activities (Table 5.5). The greater proportion of time spent in these activities by juveniles (Fig. 5.3) is due to the amount of time that juveniles spent playing. The differences in the proportions of time spent feeding and in other activities, by the different age/sex classes, is reflected in the proportion of time spent inactive (Fig. 5.3). Adult males spent a greater proportion of time inactive than all other age/sex classes (Table 5.5).

5.7 DIURNAL VARIATION IN ACTIVITY

5.7.1 Are T. auratus Strictly Diurnal?

To determine whether T. auratus are strictly diurnal, scan samples were conducted throughout the night during an evening following a monthly scan sample (October) for GRP3. Activities observed outside the scan samples were also recorded. Feeding did not occur between 1810-0610 h. Animals were inactive (as I or IH) between 1810-0610 h with the following exceptions: on two occasions, a juvenile and adult female moved a distance of 10 m and then resumed an I or IH posture; on three occasions the same juvenile vocalised - the first occasion may have been due to disturbance caused by the onset of rain, the second vocalisation occurred at 0530 h (at dawn) and the third at 0600 h; at 0540 h and 0550 h these same two individuals moved (MW and MA, respectively). Other group members did not engage in any activity other than I or IH until 0610 h when the group became active (predominantly MW).

Scan samples were conducted throughout the following day (referred to as D-N). If the night observations made were atypical of nightly activity, activity during the following day may have differed from "usual" daily activity. The proportion of time (as Chapter 5: Activity budgets 129

the sum of proportion of each scan) spent in the major activities during D-N and during the five days constituting the October monthly scan sample, are shown in Table 5.6. The proportional distribution of time spent in the major activities for these days did not differ significantly (X 2 = 13.74, 15 d.f., n.s.).

Table 5.6 Proportion of Time (as Sum of Proportion of Each Scan) Spent in Major Activities by GRP3 During the October Monthly Scan Sample (Day 1 - Day 5) and on the Day (D-N) Following the Night During Which Observations of Activity were Made

Activitya Day 1 Day 2 Day 3 Day 4 Day 5 D-N

Feeding (FI + FB) 18.6 26.7 17.4 19.4 12.4 12.7

Inactivity (I + IH) 35.5 23.1 33.9 29.8 33.1 26.6

Moving (MA + MW) 11.4 10.8 13.5 11.7 12.0 16.2

All other activities 3.2 5.9 4.0 2.7 5.1 3.6 (P, V, GS, G + 0) a Refer to text for details of activity abbreviations

Whilst observations of nightly activity were collected for one night only, it is considered unlikely that T. auratus engaged in major bouts of activity during the night. When conducting monthly scan samples, most (if not all) of the animals in the study group were inactive when I completed observations at dusk. Similarly, when I relocated the group on subsequent mornings, the group had frequently not moved away from the "night tree".

5.7.2 Diurnal Variation in Activity

The proportions (as %) of time spent feeding, moving (as in group progression (MA)) and inactive during each hour of the day, as determined from the focal-animal sampling method, are shown in Figure 5.4. Immediately apparent is the variability for any given hour for each of the activities. In general terms, feeding tended to peak early in the morning and late in the afternoon as Chapter 5: Activity budgets 130

100

80

60

Inactive 40

20

cu 0 -0) "'0 :s Jl 40

~ 20 Moving .:!::.::: u -ca 0

0~

Feeding

06 08 10 12 14 16 18

time of day {h)

Figure 5.4 Proportions of time (as %) spent m major activities each hour n = 14 (mean± SD) Where hour = e.g. 08 this includes time period between 0800-0859 h Chapter 5: Activity budgets 131

did movmg (MA), while inactivity was greatest around the middle of the day with a secondary peak around 0800-0900 h.

The variability in activity each hour reflects the lack of conformity from day to day. Periods of feeding and moving were typically followed by periods of inactivity, sometimes exceeding one hour in duration. However, the duration of feeding and moving bouts, as well as the time interval between such bouts, was highly variable.

5.8 COMPARISON WITH ACTIVITY BUDGETS OF OTHER COLO BINES

The activity budgets of seven colobine species are given in Table 5. 7. Unfortunately, direct comparisons of activity budgets between species were not possible as:

1. Definitions of activity have varied between studies (Clutton­ Brock 1974b);

2. The method of scan sampling has varied between studies (Clutton-Brock 1974b). For example, Struhsaker (1975) determined activity budgets from "sustained activities" i.e. the first activity in which an animal engaged five seconds after the fixed interval; activity budgets in this study were determined from "instantaneous" scan sampling;

3. Differences in age/sex class composition of study groups may bias activity budgets based on group activity (Marsh 1978a);

4. Differences in visibility conditions between study areas may result in varying degrees of bias when using the scan sampling method. For example; visibility conditions in this study resulted in an overestimate of the proportion of time spent moving and an underestimate of the proportion of time spent inactive (Section 5.5). Chapter 5: Activity budgets ·132

Table 5.7 Colobine Activity Budgets (as mean %)

Species Feeding Inactive Moving Source

Colobus badius 30.0 47.8 7.2 Marsh 1981b rufomitratusa C. badius 42.3 28.7 9.7 Marsh 1978a rufomitratusb C. badius 45.3 34.2 9.0 Struhsaker 1975 rufomitratusb C. guerezab 19.9 57.4 5.4 Oates 1977a C. satanas 22.5 54.2 3.6 McKey and Waterman 1982 Presbytis comata 29.3 63.0 4.7 Ruhiyat 1983 P. melalophos 33.0 46.0 21.0 Raemaekers and Chivers 1980 Trachypithecus 22.9 52.0 16.2 GRP21: this study auratus T. auratus 29.6 44.1 15.2 GRP3: this study T. obscura 34.0 46.0 20.0 Raemaekers and Chivers 1980

aT ana River study area bKibale study area

Intraspecific variation m activity budgets may be considerable and further render interspecific comparisons as invalid. For example; Colobus badius rufomitratus from the Tana River (Marsh 1981 b) spent considerably less time feeding and more time inactive than the same sub-species in Kibale (Marsh 1978a) (Table 5.7).

The only generalization regarding activity budgets of colobine species that is warranted is that most time is spent inactive, feeding is second in the proportion of time it occupies and moving occupies the least time of the major activities considered (Table 5.7). Colobus badius rufomitratus in Kibale was an exception to this generalization in that feeding was the most common activity with inactivity being second in the proportion of time it occupied (Struhsaker 1975; Marsh 1978a (Table 5.7)). Chapter 5: Activity budgets 133

5.9 DISCUSSION

Both GRP3 and GRP21 engaged primarily in inactivity. Feeding, followed by moving (as in group progression) were the second and third most common activities, respectively. The proportioning of these three major activities in the total activity budget is similar to that of several other colobines.

When comparing the activity budgets of the two study groups, it was found that GRP21 spent more time moving (MA) and inactive than GRP3 and less time feeding than GRP3. The GRP21 home-range area was larger than that of GRP3 (Chapter 8) and GRP21 may, therefore, have spent more time patrolling and defending this area against neighbouring T. auratus groups. Another reason why GRP21 spent a greater amount of time moving (MA) may be that, in a larger, more diverse (see Section 6.6) home-range area, distances between food sources may be greater and more time may therefore be required to move from one food source to another. But why did GRP21 spend less time feeding (and more time inactive) than GRP3?

Marsh (1978a) found that C. badius living in marginal habitats spent less time feeding and postulated that this may have been due to climatic differences between the study sites or because a diet of low diversity (as is typical of marginal habitats) requires longer to digest and the intake of food is, therefore, also lower. Similarly, P. melalophos in logged forest at Sungai Tekam spent less time feeding than P. melalophos prior to logging (Johns 1986). After logging, fewer fruit were available and, hence, these primates ate more foliage which takes comparatively longer to digest (Johns 1986). Due to the geographical proximity of GRP3 and GRP21 in the Pangandaran Nature Reserve, differences in climatic factors are not considered to be an important variable. It may seem, therefore, that the GRP21 home-range area is "marginal" when compared with the GRP3 home-range area. This is not the case. Dietetic diversity was similar for both study groups, however, diversity of vegetation was far greater in the GRP21 home-range area (Section 6.6). It is argued that GRP3, Chapter 5: Activity budgets 134

living in a "marginal" habitat, spent more time feeding in order to maintain dietetic diversity (or perhaps dietetic diversity was maintained by frequent sampling of foods to test their "acceptability") and/or to obtain sufficient nutrients from poorer (but "acceptable") food sources e.g. midribs of young Tectona grandis leaves (Section 6.4.3 and Section 6.8).

Possibly, the quality of vegetation, as potential food for a colobine, was greater at Pangandaran than in the study sites referred to by Marsh (1978a). Davies (1984) found that as the proportion of seeds (a preferred food source) in the diet of P. rubicunda decreased, the proportion of time spent feeding increased (and dietetic diversity also increased). Therefore, where there is a decrease in the availability of preferred foods (whether on a seasonal or locality basis) (some) colobines increase their food intake where "acceptable" (although not preferred) foods are readily available.

Monthly variation in the proportion of time spent feeding and moving by T. auratus was not found. In an environment where high quality food is only seasonally available, monthly variation m the proportion of time spent feeding by a colobine species has been reported (Davies 1984). As is discussed further in Section 6.8, this does not apply to T. auratus in the Pangandaran Nature Reserve. The absence of monthly variation in the proportion of time spent moving (MA) may be due, at least in part, to the relative abundance of "acceptable" food in the Pangandaran Nature Reserve (Section 6.8) and/or to the comparatively small home-range areas of GRP3 and GRP21 (when compared with home-range size of other colobines - see Chapter 8) such that food sources, even if patchily distributed, may be readily encountered.

5.10 SUMMARY

1. GRP3 and GRP21 both spent most of their time inactive. Feeding was the next most common activity and moving (as m Chapter 5: Activity budgets 135

group progression) occupied the least time of these maJor activities.

2. GRP21 spent more time moving (as in group progression) and inactive than GRP3 but spent less time feeding than GRP3.

3. The amount of time spent inactive varied monthly for both GRP3 and GRP21, however, monthly variation was not found in the proportions of time spent moving and feeding.

4. The GRP3 activity budget was similar to that of two other T. auratus groups found in the TW.

5. Scan sampling provided an accurate estimate of the proportion of time spent feeding but underestimated the proportion of time spent inactive and overestimated the proportion of time spent movtng.

6. The mam age/sex class differences in acttvtty budgets were that adult females spent more time feeding than adult males and juveniles; juveniles spent more time in "other" activities due to time spent playing; adult males spent more time inactive than other age/sex classes.

7. Trachypithecus auratus are diurnally active.

8. There was considerable variation within daily activity budgets when considering the major activities on an hourly basis. Chapter 6: Feeding behaviour . 136

CHAPTER 6 FEEDING BEHAVIOUR

6.1 INTRODUCTION

An important aspect of any ecological study is the availability of food and selection thereof by the animal under study. Several ecological studies of colobines have been undertaken (see Section 1.3) and their feeding behaviour analysed with respect to dietary composition (both species and items). Detailed ecological data have been collected on Colobus badius (Clutton-Brock 1972; Struhsaker 1975; Marsh 1978b), C. guereza (Oates 1977a) and C. satanas (McKey et al. 1981) in Africa. In Asia, comprehensive ecological studies of Presbytis melalophos and P. rubicunda (Bennett 1983; Davies 1984; Davies et al. 1988), Trachypithecus senex and P. entellus (Hladik 1977), T. obscura (Curtin 1980) and T. johnli (Oates et al. 1980) have been conducted. This study is the first detailed study of a Trachypithecus species in Indonesia.

This chapter describes the diet of T. auratus in the Pangandaran Nature Reserve and compares the diets of GRP21 and GRP3 in order to determine intraspecific variation in feeding behaviour, related to differences in habitat composition and availability of food. A further aim in this chapter is to compare the feeding behaviour of T. auratus with the feeding ecology of other colobines.

6.2 THE DIGESTIVE SYSTEM OF COLOBINES

All colobines are characterized by having complex stomachs (Hill 1952; Amerasinghe et al. 1971) which consist of four parts: the sacculated and enlarged presaccus and saccus (the forestomach), the short pars pylorica (the pyloric section) and the long tubus gastricus (the tubular section) which connects them (Kuhn 1964; Bauchop 1978). The forestomach is normally maintained at a pH value between 5 and 7 (Bauchop and Martucci 1968; Bauchop 1971, 1978). The forestomach contains bacteria which ferment the food (Kuhn 1964; Bauchop and Martucci 1968; Moir 1968; Ohwaki et al. Chapter 6: Feeding behaviour 137

1974) in much the same way as ciliate protozoa in large ruminants (Kuhn 1964; Ohwaki et al. 1974) and bacteria in small ruminants (Giesecke 1970).

The microflora (bacteria or protozoa) convert carbohydrates to short-chain volatile fatty acids (VFAs) (Bauchop 1978; Parra 1978) through anaerobic fermentation (Kuhn 1964; Bauchop and Martucci 1968; Ohwaki et al. 1974; Kay et al. 1976). The VFAs then pass through the stomach wall and become available to the animal as an energy source (Bauchop and Martucci 1968; Bauchop 1978). Cellulose-digesting bacteria are present in colobine forestomachs (Bauchop and Martucci 1968), but the extent of cellulose digestion 1s still unknown (Oates 1977a; Bauchop 1978; Davies et al. 1988). In addition to the VF As, the bacteria in the forestomach may be used as nutrition by the host when they pass into the acidic, pyloric section of the stomach, and the small intestine, where they are digested (Cuthbertson and Hobson 1960).

Bennett (1983) lists the advantages of forestomach fermentation known to occur in ungulates and also considered to operate in colobines. These advantages include the following: the bacteria in the forestomach synthesize all essential vitamins with the exception of A and D (Bauchop 1978); the bacteria can use non-protein nitrogen for growth (Bauchop 1978) and therefore the forestomach fermenter can obtain high quality protein (through digestion of bacteria) from a limited protein intake (Hungate 1968; Moir 1968; Janis 1976); urea is used in protein synthesis thereby decreasing the formation of urine which may aid in water conservation (Bauchop 1978); bacteria may deactivate plant toxins in the forestomach before absorption occurs which may enable the animal to ingest food which may otherwise be detrimental (Freeland and Janzen 1974; Oates et al. 1977; Waterman 1984).

A major problem, or disadvantage, of a forestomach fermentative system occurs when concentrate foods are eaten m large quantities. VFAs may be produced too rapidly to be absorbed resulting in a decrease in the forestomach pH, referred to as "acidosis" (Goltenboth 1976; Davies et al. 1988). Chapter 6: Feeding behaviour 138

For colobines and other animals with ruminant-like digestive systems, the intake of food largely depends on the volume of the forestomach (Baile and Forbes 1974). Only when food has been digested as thoroughly as possible in the forestomach can it pass to the rest of the digestive system (Janis 1976). In this way the nutritional value of slowly-digestible foods is maximised and the chance of toxins passing into the rest of the digestive system, before being deactivated, is reduced (McKey et al. 1981). The rate of VFA production in captive T. cristata was found to be constant for ten hours after feeding (Bauchop and Martucci 1968). This is in sharp contrast to primates with simple, spherical stomachs which have much faster food passage rates (Parra 1978). This difference m digestive physiology is likely to influence dietary intake.

6.3 METHODS

Various methods have been used in previOus studies of colobine feeding ecology to quantify food intake. These have included the frequency with which foods are eaten (Struhsaker 1975; McKey et al. 1981; Marsh 1981a), the proportion of items in the stomachs of dead animals (Gautier-Hion 1980), the weight of items ingested (Hladik 1977) and the proportion of time spent eating various species and items (Clutton-Brock 1975a; Oates et al. 1980; Bennett 1983; Davies 1984). Hladik (1977) showed that estimates of food intake differ when estimating the weight of items ingested and when calculating the proportion of time spent feeding on various items.

Whilst visibility conditions of the monkeys were good in this study (Section 2.2.2) it still proved to be extremely difficult to observe a single animal during a feeding bout in order to count the number of items eaten. Frequently, single items were not selected during a feeding bout but rather a handful or mouthful of selected items were eaten in quick succession. This was often the case when either young leaves and flowers or small fruits in clumped distributions were eaten. On such occasions it was impossible to Chapter 6: Feeding behaviour 139

count the number of food items ingested. No attempt was therefore made to quantify diet by measuring feeding rates and weighing foods (Chivers 1974; Hladik 1977; Raemaekers 1978).

The scan sampling method (Section 2.4.1 and Section 2.4.2), with unweighted data (Section 2.5.1), was used to quantify food intake by determining the proportion of time spent feeding on various species, items and parts. Oates (1977a) showed that there was no significant difference between three scan sample techniques in quantifying food intake by C. guereza. One of the techniques used by Oates ( 1977 a) was similar to the scan sampling method used m this study (Section 2.4.2). Therefore, results obtained from this study of T. auratus are comparable to those of other colobine studies based on the scan sample technique.

A list of food items and parts determined with abbreviations, and diagnostic features where appropriate, is given in Table 6.1. The major items determined were: fruits (unripe or ripe); flowers; flower buds; calyx of flowers; leaf buds; young leaves; mature leaves. The major parts determined were: seeds and pericarps of fruits; basal part of lamina (approximately 30% of lamina), apical tip of lamina, whole lamina, petiole, stems/stalks and mid-ribs of leaves.

6.4 FOOD SELECTION

6.4.1 Species

6.4.1.1 Comparison of Species Eaten by GRP21 and GRP3

A list of the species eaten by GRP21 and GRP3 throughout the study is given in Table 6.2. Species are listed in ranked order for the percentage of feeding time (unweighted). Whilst the number of species eaten was large for both groups (88 species for GRP3 and 49 species for GRP21), only 23 species for GRP3 and 22 species for GRP21 contributed > 1% each to the total feeding time. The top 15 species contributed 70.61% and 73.03% to the total diet of GRP3 and Chapter 6: Feeding behaviour '140

Table 6.1 Food Items and Parts Determined with Abbreviations, and Diagnostic Features where Appropriate

Item Abbre­ Diagnostic Part Abbre­ viation features viation

Ripe fruit BR BR:Colour change · Unripe fruit BU from unripe; Seeds s Ripe + Unripe BA dehisced Pericarp (may F Fruit (unknown B BU:Colour; be dry, woody, whether ripe or immature seeds fleshy) unripe) Whole fruit w

FLOWERS

Flower F Flower bud FB Flower & flower FA bud Calyx FP

LEAVES LEAVES

Mature leaf LM Colour, size, turgidity different to LY Basal portion c (c. 30%) of lamina Young leaves LY L Y:leaves of Apical tip of A LYl unknown or mixed lamina LY2 size classes Lamina (whole) B LY3 LY1-LY4= Petiole p LY4 smallest-largest Leaf stem/stalk s (often connecting several petioles) Leaf bud LB Petiole & leaf PS stalk Young leaves (LYl LC Lamina & petiole BP only), & leaf buds Lamina & petiole BS Young leaves LE & leaf stalk (mixed size Mid-rib M classes) & leaf buds Young leaves & LD mature leaves Leaf of unknown L maturity

No item NI No part NP (i.e. unknown) (i.e. unknown) Chapter 6: Feeding behaviour 141

Table 6.2 Food Tree Species of GRP3 and GRP21 (Species listed in ranked order for the unweighted % feeding time)

a) GRP3

Rank Family Species % feeding time al Verbenaceae Tectona grandis 14.51 a2 Moraceae Ficus sinuata 8.50 a3 S terculiaceae Pterospermum javanicum 8.46 4 Sterculiaceae Kleinhovia hasp ita 7.88 as Ulmaceae Celtis philippensis 5.53 6 Rutaceae Clausena excavata 4.12 7 Meliaceae Swietenia macrophylla 3.49 ag Moraceae Ficus sumatrana 3.34 9 Myrtaceae Eugenia zippelianum 2.91 10 Fabaceae Dalbergia latifolia 2.35 11 Fabaceae Acacia auriculiformis 2.23 12 Moraceae Ficus sp. (515) 2.10 a13 Moraceae Ficus glomerata 1.91 14 Moraceae Ficus melinocarpa 1.83 a15 Verbenaceae Vitex pinnata 1.45 16 Rubiaceae Anthocephalus cadamba 1.35 a16 Sterculiaceae Heritiera littoralis 1.35 a18 Lauraceae Cinnamomum iners 1.32 19 Fabaceae Cynometra ramiflora 1.22 a2o Rubiaceae Tricalysia singularis 1.21 21 Euphorbiaceae Bridelia monoica 1.12 a22 Fabaceae Erythrina variegata 1.08 23 Sapotaceae Planchonella obovata 1.04 24 Euphorbiaceae Mall at us multiglandulosa 0.96 a25 Rubiaceae Guettarda speciosa 0.93 a26 Meliaceae Dysoxylum caulostachyum 0.91 a27 Verbenaceae Vitex pubescens 0.69 a28 Moraceae Ficus fistulosa 0.66 29 Euphorbiaceae Baccaurea javanica 0.56 30 Moraceae Ficus septica 0.55 31 Euphorbiaceae Claoxylon polot 0.54 a32 Moraceae Ficus benjamina 0.53 33 Bignoniaceae Radermachera gigantea 0.51 34 Meliaceae Dysoxylum densiflorum 0.49 35 Theaceae Ternstroemia patens 0.48 36 Moraceae Ficus drupacea 0.44 37 Euphorbiaceae Antidesma sp. 0.42 38 Oleaceae Linociera ramiflora 0.40 39 Ebenaceae Diospyros truncata 0.36 a4o Fabaceae Cassia siamea 0.31 41 Clusiaceae Garcinia sp. 0.29 41 (v)Fabaceae Phanera fulva 0.29 41 Myrtaceae Decaspermum fructicosum 0.29 41 Sapotaceae Planchonella ling gensis 0.29 45 Ebenaceae Diospyros javanica 0.28 46 509 Unidentified sample 0.27 46 Malvaceae Hibiscus tiliaceus 0.27 48 513 Unidentified sample 0.27 Chapter 6: Feeding behaviour 142

••••• Table 6.2a (cont.)

Rank Family Species % feeding time

49 Elaeocarpaceae Elaeocarpus glaber 0.23 50 Annonaceae Cananga odorata 0.22 a 5o Rutaceae Acronychia laurifolia 0.22 50 337 Unidentified sample 0.22 53 Moraceae Ficus hispida 0.21 54 Moraceae Ficus subulata 0.20 55 Euphorbiaceae Mallotus philippensis 0.19 56 Sapindaceae Harpullia cupanioides 0.17 58 Moraceae Ficus obscura var. scaberrima 0.16 58 Moraceae Ficus racemosa var. elongata 0.16 59 Bignoniaceae Dolichandrone spathacea 0.15 59 ( v )Connaraceae Rourea minor 0.15 59 Euphorbiaceae Mallotus ricinoides 0.15 59 Tiliaceae Microcos paniculata 0.15 59 543 Unidentified sample 0.15 64 Ebenaceae Diospyros sp. 0.11 a64 Hernandiaceae Hernandia peltata 0.11 64 c Unidentified sample 0.11 67 Burseraceae Canarium hirsutum 0.07 67 Clusiaceae Garcinia balica 0.07 a67 Euphorbiaceae Antidesma bunius 0.07 67 Euphorbiaceae Antidesma montanum 0.07 67 Euphorbiaceae Suregada glomerata 0.07 67 Flacourtiaceae Casearia tuberculata 0.07 67 Myristicaceae Horsfieldia sp. 0.07 67 (v)Rubiaceae Randia longiflora 0.07 67 ( v) V er benaceae Clerondendrum serratum 0.07 67 A Unidentified sample 0.07 67 (v)339 Unidentified sample 0.07 67 433 Unidentified sample 0.07 79 357 Unidentified sample 0.05 80 Euphorbiaceae Mallotus moritzianus 0.04 80 Myrtaceae Syzygium polyanthum 0.04 ago Sterculiaceae Sterculia coccinea var. coccinea 0.04 80 Violaceae Rinorea sp. 0.04 84 Moraceae Artocarpus elasticus 0.03 ag5 Myrtaceae Eugenia sp. 0.02 85 485 Unidentified sample 0.02 87 Meliaceae Melia azedarach 0.01 87 Verbenaceae Vitex cofassus 0.01

Unknown lianas (no samples collected) 0.62 Unknown trees (no samples collected) 2.85

Insects 0.01

Lianas are preceded with (v) in family column aspecies also eaten by GRP21 Chapter 6: Feeding behaviour 143

.•... Table 6.2 (cont.) b) GRP21

Rank Family Species % feeding time bl Meliaceae Dysoxylum caulostachyum 10.05 b2 Moraceae Ficus benjamina 7.45 b3 Fabaceae Erythrina variegata 7.37 b4 Moraceae Ficus sinuata 6.84 b5 Rubiaceae Guettarda speciosa 6.68 b6 S terculiaceae Sterculia coccinea var. coccinea 5.53 b6 Lauraceae Cinnamomum iners 5.53 b8 Hernandiaceae Hernandia peltata 4.36 b9 Verbenaceae Vitex pubescens 3.76 b1o Rubiaceae Tricalysia singularis 3.30 11 Rubiaceae Nauclea sp. 3.25 b12 Verbenaceae Vitex pinnata 2.38 bl3 Fabaceae Cassia siamea 2.36 14 Moraceae Ficus bracteata 2.20 b15 Moraceae Ficus sumatrana 1.97 16 Verbenaceae Vitex glabrata 1.81 17 Euphorbiaceae Glochidion macrocarpum 1.49 18 Tiliaceae Schoutenia ovata 1.42 b19 Moraceae Ficus fistulosa 1.35 20 Sapindaceae Ganophyllum falcatum 1.26 21 Oleaceae Linociera sp. 1.21 22 Anacardiaceae Mangifera indica 1.15 23 Anacardiaceae Spondias cytherea 0.94 b24 Euphorbiaceae Antidesma bunius 0.79 25 Euphorbiaceae Mallotus oblongifolius 0.75 b26 Myrtaceae Eugenia sp. 0.67 b27 Sterculiaceae Pterospermum javanicum 0.64 28 Melastomataceae Memecylon sp. 0.63 28 Meliaceae Dysoxylum sp. 0.63 30 ( v )Connaraceae Agelaea macrophylla 0.55 30 Sapindaceae Arytera litoralis 0.55 32 Fabaceae sp. 0.47 b32 Moraceae Ficus glomerata 0.47 32 Rubiaceae Anthocephalus obtusa var. major 0.47 32 Sterculiaceae Pterospermum diversifolium 0.47 36 Malvaceae Thespesia populnea 0.46 b37 Ulmaceae Celtis philippensis 0.41 b38 S terculiaceae H eritiera littoralis 0.37 39 Clusiaceae Garcinia dulcis 0.31 39 Myrtaceae Rhodamnia cinerea 0.31 b39 Rutaceae Acronychia laurifolia 0.31 b39 Verbenaceae Tectona grandis 0.31 Chapter 6: Feeding behaviour •144

..... Table 6.2b (cont.)

Rank Family Species % feeding time

43 Urticaceae Dendrocnide microstigma 0.28 44 Moraceae Ficus sp. 0.25 45 Moraceae Poikilospermum suaveolens 0.21 46 Moraceae Ficus variegata 0.16 46 (v)4066 unidentified sample 0.16 46 Sapindaceae H ebecoccus cupanioides 0.16 49 (v)4014 unidentified sample 0.04

Unknown lianas (no samples collected) 1.14 Unknown trees (no samples collected) 4.38

Lianas are preceded with (v) in family column hspecies also eaten by GRP3

GRP21, respectively. The number of Iiana species m the diet (Table 6.2) was low for both study groups unlike for some colobines, notably P. rubicunda (Davies 1984), for which lianas were a major dietary component contributing approximately one third of the total diet.

Trachypithecus auratus engaged in major feeding bouts when most or all of the group were feeding on a particular species for an extended period of time. Sometimes, when the majority of the group were resting or travelling, an individual within the group would pluck single food items. This explained, at least partly, why several species accounted for small proportions of the total feeding time.

The T. auratus study groups, particularly GRP3, tended to eat the more abundant species (as % basal area) (Table 6.3). However, some of these species, especially Ficus spp. for GRP3, were quite rare when considered as the number of stems per hectare (Table 6.3). Both T. auratus groups spent more time feeding on species whose members had greater basal areas than other species in the study sites (Table 6.4). Therefore, while food tree species may not have been common (as the number of stems per hectare) within the study sites, T. auratus fed largely in trees with large canopies. This enabled individuals in a group to feed simultaneously on the same food source. Chapter 6: Feeding behaviour 145

Table 6.3 The Ten Most Abundant Species (as % basal area) and Corresponding Stem Density (as Frequency per ha) Within the Study Sites, and the Proportion of T. auratus Diet (as % Feeding Time) Contributed by These Species a) GRP3

Species % b.a. freq. ha-l % feeding time

Ficus sinuata 28.86 2 8.50 Swietenia macrophylla 19.98 115 3.49 Ficus sumatrana 13.26 2 3.34 Tectona grandis 7.56 104 14.51 Ficus glomerata 6.65 2 1.91 Kleinhovia hospita 5.93 11 7.88 H eritiera littoralis 3.09 2 1.35 Pterospermum javanicum 1.70 11 8.46 Hernandia peltata 1.18 6 0.11 Acacia auriculiformis 1.11 19 2.23

b) GRP21

Species % b.a. freq. ha-l % feeding time

Ficus sumatrana 24.73 11 1.97 Hernandia peltata 21.16 16 4.36 Ficus fistulosa 10.39 6 1.35 Artocarpus elasticus 6.14 3 0 Aphanamixis grandifolia 3.89 13 0 Vitex pubescens 3.04 18 3.76 Sterculia coccinea 2.85 6 5.53 Microcos paniculata 2.70 9 0 Thespesia populnea 1.93 4 0.46 Cinnamomum iners 1.84 35 5.53

Table 6.4 Comparison Between Basal Areas (b.a.) of Food Tree Species and Non-Food Tree Species

Variable 1 Variable 2 T. auratus u p group

b.a. food b.a. non-food GRP21 1752 32 80 <0.01 tree species tree species GRP3 1088 36 41 < 0.01 Chapter 6: Feeding behaviour 146

GRP3 and GRP21 ate 22 species in common. Sixty-six species (75% of species eaten by GRP3) were eaten exclusively by GRP3 and 27 species (55% of species eaten by GRP21) were eaten exclusively by GRP21 (Table 6.2). The difference in species eaten may be due to differences in the vegetation composition of the two study sites (Section 3.3.2), differences in availability of preferred items of a species for the two study sites, differences in preferences for certain species and, possibly, preferences for individual trees by the monkeys. The relative importance of these factors in diet selection IS discussed below.

Differences in the duration of the study for the two sites (Section 2.4.2) may also have had an influence. This study did not aim to provide an exhaustive list of food tree species for T. auratus but rather to consider food selection during the study period. Due to the nature of plucking food items occasionally and because of the nature of the forest where all species do not flower and fruit annually it may take several years to compile such a list.

Thirteen of the top 15 food tree species for GRP21 were also eaten by GRP3 (Table 6.2) but only three of these species, F. sinuata, F. sumatrana and Vitex pinnata, also ranked in the top 15 food tree species for GRP3 and contributed 8.50%, 3.34% and 1.45%, respectively, to the GRP3 diet. Abundance (as % basal area) of the remaining 12 species (of the top 15 food tree species for GRP21), in the home-range areas of the two groups, is given in Table 6.5. Nauclea sp. and F. bracteata, which were not eaten by GRP3, were absent, or very rare, in the GRP3 home-range area. The remaining species listed in Table 6.5, with the exception of Dysoxylum caulostachyum, were also rare in the GRP3 home-range area as they either contributed only a small proportion to the total basal area or were not sampled in the vegetation quadrats.

Dysoxylum caulostachyum, whilst more abundant in the GRP3 than the GRP21 home-range area (Table 6.5), was generally not available as a potential food source for GRP3. Many trees of this species in the GRP3 home-range area were immature and few mature trees with fruits were present. The item of this species Chapter 6: Feeding behaviour 147

Table 6.5 Abundance (as % Basal Area) of 12 of the Top 15 Food Tree Species for GRP21 which Did Not Rank in the Top 15 Food Tree Species for GRP3

Family Species % basal area GRP21 GRP3

Meliaceae Dysoxylum caulostachyum 0.62 1.09 Moraceae Ficus benjamina 0.03 rare Fabaceae Erythrina variegata rare rare Rubiaceae Guettarda speciosa 0.75 rare Sterculiaceae Sterculia coccinea 2.85 rare Lauraceae Cinnamomum iners 1.84 0.03 Hernandiaceae Hernandia peltata 21.16 1.18 Verbenaceae Vitex pubescens 3.04 0.04 Rubiaceae Tricalysia singularis 0.43 rare Rubiaceae Nauclea sp. rare rare/absent Fabaceae Cassia siamea 0.42 rare Moraceae Ficus bracteata · rare rare/absent which was preferred by GRP21, was fruit (see Section 6.4.2). Whilst phenological condition was not assessed for this species in the GRP3 home-range area, D. caulostachyum fruits are conspicuous (orange fruits hanging in "bunches" from tree trunks) and I would have noticed them if present as I was aware of the location of D. caulostachyum trees. Therefore, 12 of the top 15 food tree species for GRP21 were either not eaten by GRP3, or accounted only for small proportions of the feeding time, as they were either absent from the GRP3 home-range area or in low abundance (with the exception of D. caulostachyum).

Seven of the 15 top food tree species for GRP3 were also eaten by GRP21. Of these, F. sinuata, F. sumatrana and V. pinnata also ranked in the top 15 for GRP21, whereas Pterospermum javanicum, F. glomerata, Celtis philippensis and Tectona grandis were of lesser rank. Pterospermum javanicum, F. glome rata and T. grandis were present in lower abundance in the GRP21 home-range area while C. philippensis was present in greater abundance than in the GRP3 home-range area (Table 6.6). Of the 15 top food tree species for GRP3 which were not eaten by GRP21, Kleinhoviahospita, Clausena excavata, Dalbergia latifolia, Ficus sp. (515) and F. melinocarpa were either present in low abundance in the GRP21 home-range area (i.e. lower % basal area than for GRP3) or rare/absent (Table Chapter 6: Feeding behaviour ·148

Table 6.6 Abundance (as % Basal Area) of 12 of the Top 15 Food Tree Species for GRP3 which Did Not Rank in the Top 15 Food. Tree Species for GRP21

Family Species % basal area GRP3 GRP21

S terculiaceae Pterospermum javanicum 1.70 0.36 Moraceae Ficus glomerata 6.65 rare Ulmaceae Celtis philippensis 0.24 0.32 Verbenaceae Tectona grandis 7.56 0.12 S terculiaceae Kleinhovia hospita 5.93 rare/absent Rutaceae Clausena excavata 0.99 0.15 Meliaceae Swietenia macrophylla 19.98 absent Myrtaceae Eugenia zippelianum 0.03 1.04 Fabaceae Dalbergia latifolia rare rare/absent Fabaceae Acacia auriculiformis 1.11 absent Moraceae Ficus sp. rare rare/absent Moraceae Ficus melinocarpa rare rare/absent

6.6). Acacia auriculiformis and Swietenia macrophylla, two of the plantation species, were known to be absent from the GRP21 home­ range area. Eugenia zippelianum was present in greater abundance m the GRP21 home-range area (Table 6.6).

Therefore, when comparing the top 15 food tree species for both GRP3 and GRP21 to determine preferential selection for species across the two study sites, only C. philippensis and E. zippelianum can be considered as they ranked high in the GRP3 diet and were low (C. philippensis) or absent (E. zippelianum) in the GRP21 diet. Both species were more abundant in the GRP21 home-range area. Selection ratios (see Section 6.4.1.2) for C. philippensis were 23.04 and 1.28 for GRP3 and GRP21, respectively; and 97.00 and 0 for E. zippelianum for GRP3 and GRP21, respectively, suggesting that while these species were highly selected by GRP3 this was not so for GRP21. However, in the absence of phenological data for the species for both study sites it is not known whether preferred items of these species were available as a potential food source to GRP21 during the study so it can not be concluded that GRP21 preferred other species over these two species. Differences in the species composition of the diet (top 70% feeding time i.e top 15 species) for GRP3 and GRP21 were, therefore, generally related to differences m the vegetation composition of the two study sites and possibly to Chapter 6: Feeding behaviour 149

differences in availability of preferred items for a given species m the two study sites.

A list of the families whose members were eaten by the two groups, in ranked order of (unweighted) ·% feeding time, is given m Table 6. 7. The family accounting for the greatest proportion of the % feeding time for both GRP3 and GRP21 was Moraceae (c. 20% for both groups). Whilst the species contributing the top 70% to the diet varied between GRP3 and GRP21 (Table 6.2), five of the top six families were the same for both groups and contributed 67.49% and 57.06% to the GRP3 and GRP21 diets, respectively.

A comparison between weighted and unweighted values (Section 2.5.1) was made for the top 20 food species for both GRP21 and GRP3 (Table 6.8). As these values are derived from the same data they could not be compared statistically. The two sets of results are similar (Table 6.8). The weighted values are more precise in that they take into account (and weight accordingly) scans in which few of the animals observed were feeding (Section 2.5.1). Unweighted data have generally been used in analyses of recent colobine studies (e.g. Bennett 1983 (although the equivalent of unweighted data as determined for T. auratus (see Section 2.5.1) is that referred to by Bennett as weighted data); Davies 1984). Therefore, for the benefit of comparison, further reference to food species relates to unweighted % feeding time.

6.4.1.2 Selection Ratios

A method for determining the degree of selection is through the calculation of selection ratios (S.R.) (McKey et al. 1981) where

per cent of feeding time on species A S.R. ==per cent of total basal area accounted for by species A

Where the ratio is greater than one, the animals ·ate a given species to a greater extent than would be expected if they were feeding at random, i.e. they were being selective in what they ate. Selection ratios were calculated for the top 20 food species for GRP3 and Chapter 6: Feeding behaviour 150

Table 6.7 Families of Food Trees Eaten by GRP3 and GRP21 (Families listed in ranked order for unweighted % feeding time)

GRP3 GRP21 Rank Family % feeding Rank Family % feeding time time

1 Moraceae 20.63 1 Moraceae 20.90 2 S terculiaceae 17.73 2 Rubiaceae 13.70 3 Verbenaceae 16.75 3 Meliaceae 10.68 4 Fabaceae 7.49 4 Fabaceae 10.20 5 Ulmaceae 5.53 5 Verbenaceae 8.27 6 Meliaceae 4.89 6 Sterculiaceae 7.01 7 Rutaceae 4.34 7 Lauraceae 5.53 8 Euphorbiaceae 4.19 8 Hernandiaceae 4.36 9 Rubiaceae 3.56 9 Euphorbiaceae 3.03 10 Myrtaceae 3.26 10 Anacardiaceae 2.09 11 Sapotaceae 1.34 11 Sapindaceae 1.97 12 Lauraceae 1.32 12 Tiliaceae 1.42 13 Ebenaceae 0.75 13 01eaceae 1.21 14 Bignoniaceae 0.66 14 Myrtaceae 0.98 15 Theaceae 0.48 15 Melastomataceae 0.63 16 Oleaceae 0.40 16 Connaraceae 0.55 17 Clusiaceae 0.36 17 Malvaceae 0.46 18 Malvaceae 0.27 18 Ulmaceae 0.41 19 Elaeocarpaceae 0.23 19 Clusiaceae 0.31 20 Annonaceae 0.22 19 Rutaceae 0.31 21 Sapindaceae 0.17 21 Urticaceae 0.28 22 Connaraceae 0.15 22 Tiliaceae 0.15 24 Hernandiaceae 0.11 25 Burseraceae 0.07 25 Flacourtiaceae 0.07 25 Myristicaceae 0.07 28 Violaceae 0.04

Not identified 1.91 Not identified 1.34 Unknown (no Unknown (no samples collected) 2.85 samples collected) 4.38

Insects 0.01 Chapter 6: Feeding behaviour 151

Table 6.8 Unweighted % Feeding Time and Weighted % Feeding Time Compared for Top Twenty Food Species for a) GRP3 and b) GRP21 a} GRP3 Unweighted % feeding time Weighted % feeding time

Species % Rank % Rank

Tectona grandis 14.51 1 15.26 1 Ficus sinuata 8.50 2 8.41 4 Pterospermum javanicum 8.46 3 10.28 2 Kleinhovia hospita 7.88 4 8.59 3 Celtis philippensis 5.53 5 6.03 5 C lausena excavata 4.12 6 3.92 7 Swietenia macrophylla 3.49 7 3.95 6 Ficus sumatrana 3.34 8 3.52 8 Eugenia zippelianum 2.91 9 2.31 10 Dalbergia latifolia 2.35 10 2.13 11 Acacia auriculiformis 2.23 11 1.99 13 Ficus sp. (515) 2.10 12 2.42 9 Ficus glome rata 1.91 13 2.00 12 Ficus melinocarpa 1.83 14 1.40 15 Vitex pinnata 1.45 15 1.16 19 Anthocephalus cadamba 1.35 16 Heritiera littoralis 1.35 16 1.53 14 Cinnamomum iners 1.32 18 1.33 16 Cynometra ramiflora 1.22 19 1.19 18 Tricalysia singularis 1.21 20 Erythrina variegata 1.20 17 Planchonella obovata 1.11 20

Unidentified spp. 2.77 1.47 Chapter 6: Feeding behaviour •t52

••••• Table 6.8 (cont.) b} GRP21 Unweighted % feeding time Weighted % feeding time

Species % Rank % Rank

Dysoxylum caulostachyum 10.05 1 11.01 1 Ficus benjamina 7.45 2 7.45 3 Erythrina variegata 7.37 3 9.23 2 Ficus sinuata 6.84 4 6.85 4 Guettarda speciosa 6.68 5 6.39 5 Cinnamomum iners 5.53 6 6.13 6 Sterculia coccinea 5.53 6 5.03 7 Hernandia peltata 4.36 8 4.76 8 Vitex pubescens 3.76 9 3.56 10 Tricalysia singularis 3.30 10 3.03 11 Nauclea sp. 3.25 11 3.57 9 Vitex pinnata 2.38 12 1.98 15 Cassia siamea 2.36 13 2.93 12 Ficus bracteata 2.20 14 2.20 13 Ficus sumatrana 1.97 15 2.03 14 Vitex glabrata 1.81 16 1.58 16 Glochidion macrocarpum 1.49 17 1.24 20 Schoutenia ovata 1.42 18 Ficus fistulosa 1.35 19 1.42 18 Ganophyllum falcatum 1.26 20 1.47 17 Linociera sp. 1.27 19

Unidentified spp. 4.38 3.16

GRP21 (Table 6.9), except where the species were sufficiently rare that they did not occur in the vegetation quadrats.

Several species were highly selected. A few species appeared not to have been selected, and some had selection ratios less than one (Table 6. 9) indicating that these species may have been avoided as food species. The calculation of selection ratios is misleading m this respect as consideration of the species concerned shows. Chapter 6: Feeding behaviour 153

Table 6.9 Selection Ratios (S.R.) for Top Twenty Food Tree Species a) GRP3

Species % feeding time % basal area S.R.

Tectona grandis 14.51 7.56 1.92 Ficus sinuata 8.50 28.86 0.29 Pterospermum javanicum 8.46 1.70 4.98 Kleinhovia hospita 7.88 5.93 1.33 Celtis philippensis 5.53 0.24 23.04 Clausena excavata 4.12 0.99 4.16 Swietenia macrophylla 3.49 19.98 0.17 Ficus sumatrana 3.34 13.26 0.25 Eugenia zippelianum 2.91 0.03 97.00 Dalbergia latifolia 2.35 rare Acacia auriculformis 2.23 1.11 2.01 Ficus sp. (515) 2.10 rare Ficus glomerata 1.91 6.65 0.29 Ficus melinocarpa 1.83 rare Vitex pinnata 1.45 0.32 4.53 H eritiera littoralis 1.35 3.09 0.44 Anthocephalus cadamba 1.35 rare Cinnamomum iners 1.32 0.03 44.00 Cynometra ramiflora 1.22 rare Tricalysia singularis 1.21 rare b) GRP21

Species % feeding time % basal area S.R.

Dysoxylum caulostachyum 10.05 0.62 16.21 Ficus benjamina 7.45 0.03 248.33 Erythrina variegata 7.37 rare Ficus sinuata 6.84 rare Guettarda speciosa 6.68 0.75 8.91 Cinnamomum iners 5.53 1.84 3.01 Sterculia coccinea var. coccinea 5.53 2.85 1.94 Hernandia peltata 4.36 21.16 0.21 Vitex pubescens 3.76 3.04 1.24 Tricalysia singularis 3.30 0.43 7.67 Nauclea sp. 3.25 rare Vitex pinnata 2.38 1.12 2.13 Cassia siamea 2.36 0.42 5.62 Ficus bracteata 2.20 rare Ficus sumatrana 1.97 24.73 0.08 Vitex glabrata 1.81 rare Glochidion macrocarpum 1.49 rate Schoutenia ovata 1.42 rare Ficus fistulosa 1.35 10.39 0.13 Ganophyllum falcatum 1.26 0.01 126.00 Chapter 6: Feeding behaviour 154

For GRP3, five of the top 20 food tree species appear not to have been selected: F. sinuata, S. macrophylla, F. sumatrana, F. glomerata, and Heritiera littoralis (Table 6.9a). The items largely eaten of these species were: fruits and young leaves (F. sinuata, S. macrophylla); fruits (F. sumatrana); young leaves (F. glomerata); flowers and flower buds (H. littoralis) (see Table 6.10, Section 6.4.2). As mature leaves of these species, which constitute the bulk of the canopy volume for which basal area is an index, were not eaten, the items which were eaten may have been selected as:

1. The abundance of fruits or flowers or young leaves within the tree canopy was low relative to mature· leaves;

2. There were differences m the monthly availability of the items;

3. There was selectivity for particular trees which would mean that the availability of items for these particular trees would have been more important in terms of feeding selection than the availability of items for the species (to which these trees belong) throughout the study site.

Phenological data (Section 3.2.2) are available for three of the five species - F. sinuata, F. sumatrana and S. macrophylla. Selection ratios were calculated for the months in which the items eaten were available. The only selection ratio with a value greater than one was for F. sumatrana (S.R. = 1.14) in January. Trees of these species were not synchronous in production of items considered, however, so selection ratios are likely to be low because abundance (as basal area) will be an overestimate of food actually available to the monkeys at a given time. This becomes further exaggerated where a species contains large individuals as was the case for F. sinuata and F. sumatrana (Table 6.3).

For GRP21, three of the top 20 food tree species appear not to have been selected: Hernandia peltata, F. sumatrana and F. fistulosa. Phenological data are available for H. peltata and F. sumatrana (Section 3.2.2) and, as described for GRP3, selection Chapter 6: Feeding behaviour 155

ratios were calculated for the months in which the items were available. None of the selection ratios was greater than one. Again, these two species are asynchronous in production of items considered and trees of these species are relatively large (Table 6.3) so low selection ratios are not surprising for the reasons stated above for GRP3. Selection of items with respect to species, and availability thereof during the study period, is discussed further m Section 6.5.

6.4.2 Food Items

A breakdown of the total diet (as unweighted % feeding time) to the level of item showed that a wide variety of items from different species were eaten. However, only a small proportion of items of particular species contributed > 1% each to the total diet (Table 6.10). For GRP3, six items from 15 species accounted for 57.80% of

Table 6.10 Items of Species Contributing 1 One Per Cent to the Total Diet (as unweighted % feeding time) a) GRP3 b) GRP21

a GRP3

Family Species % feeding time

V erbenaceae Tectona grandis LY 14.13 Moraceae Ficus sinuata LY 5.41 Sterculiaceae Kleinhovia hospita FA 4.82 Sterculiaceae Pterospermum javanicum F 4.27 Rutaceae Clausena excavata BR 3.21 Sterculiaceae Pterospermum javanicum BR 2.98 Ulmaceae Celtis philippensis FB 2.79 Myrtaceae Eugenia zippelianum BU 2.61 Moraceae Ficus sumatrana BR 2.19 Meliaceae Swietenia macrophylla LY 2.10 Moraceae Ficus sp. LY 2.02 Moraceae Ficus sinuata BU 2.00 unknown NI 1.94 Fabaceae Dalbergia latifolia LY 1.92 Moraceae Ficus glomerata LY 1.45 Verbenaceae Vitex pinnata LY 1.27 Sterculiaceae Heritiera littoralis FA 1.25 Meliaceae Swietenia macrophylla BU 1.23 Sterculiaceae Kleinhovia hospita LY 1.05 Moraceae Ficus melinocarpa LY 1.00 Chapter 6: Feeding behaviour !56

...•. Table 6.10 (cont.) b) GRP21

Family Species % feeding time

Meliaceae Dysoxylum caulostachyum BU 8.20 Fabaceae Erythrina variegata LY 7.24 Moraceae Ficus sinuata LY 6.73 Rubiaceae Guettarda speciosa LY 6.54 Lauraceae Cinnamomum iners LY 4.89 S terculiaceae Sterculia coccinea var. coccinea BU 4.58 Hernandiaceae Hernandia peltata BR 3.97 Moraceae Ficus benjamina BU 3.40 Rubiaceae Nauclea sp. BU 2.73 Fabaceae Cassia siamea LY 2.35 Verbenaceae Vitex pubescens LY 2.24 Moraceae Ficus benjamina BR 2.14 unknown NI 1.97 unknown L 1.88 Moraceae Ficus benjamina LY 1.74 Rubiaceae Tricalysia singularis FA 1.72 Meliaceae Dysoxylum caulostachyum BR 1.61 Moraceae Ficus sumatrana BU 1.57 Verbenaceae Vitex pinnata LY 1.43 Tiliaceae Schoutenia ovata LY 1.41 Euphorbiaceae Glochidion macrocarpum LY 1.25 Moraceae Ficus bracteata BU 1.25 Anacardiaceae Mangifera indica LY 1.14 Verbenaceae Vitex glabrata FA 1.10 Oleaceae Linociera sp. LY 1.04 Moraceae Ficus fistulosa LY 1.03 a Refer to Table 6.1 for definitions of item classes the total diet and for GRP21 five items from 21 species made up 71.43% of the total diet (Table 6.10).

Both GRP3 and GRP21 largely ate young leaves (Fig. 6.1). Fruits were also eaten extensively and, to a lesser degree, flowers (Fig. 6.1). The GRP21 diet had a higher proportion of unripe fruits (29% cf. 13% (median)) and a lower proportion of flowers (7% cf. 14% (median)) than the GRP3 diet (Table 6.11). Mature leaves accounted for < 1% of the total diet for both groups. As mature leaves were the most consistently available of all items during the study period, the monkeys were being selective in their choice of food items. Chapter 6: Feeding behaviour 157

a) GRP3

b) GRP21

II young leaves + leaf buds • other leaf items ml fruits E3 flowers + flower buds 0 unknown

Figure 6.1 Proponions of various items in the total diet (as unweighted % feeding time) Chapter 6: Feeding behaviour 158

Table 6.11 Comparison Between Proportion of Items in the Daily Diet of GRP3 and GRP21

Variable 1 Variable 2 Item u

GRP21 GRP3 Flowers and flower 770 25 40 < 0.01 buds

Young leaves 531 25 40 n.s. and leaf buds

Fruits 643 25 40 n.s. (unripe & ripe)

Unripe fruit 704 25 40 < 0.01

Ripe fruit 533 25 40 n.s.

n.s. = not significant.

Weighted % feeding time for items was calculated and compared with unweighted % feeding time for GRP3 and GRP21 (Table 6.12). The two sets of results are similar. For the same reasons given in Section 6.4.1.1 further reference to food items relates to unweighted % feeding time. Table 6.12 provides a more comprehensive view of items consumed than the summary depicted by Figure 6.1. The majority of fruits eaten were unripe. This was so for both GRP3 and GRP21. Twenty seven per cent of the total diet of GRP3 consisted of fruits of which 15% (of the total diet) were unripe. Thirty seven per cent of the GRP21 diet was fruit, 27% of the total diet being unripe fruits. Chapter 6: Feeding behaviour 159

Table 6.12 Proportion (Percentage of Feeding Time (Unweighted cf. Weighted)) of Items in the Diet

a) GRP3 % feeding time Item a unweighted weighted

LB 2.19 2.07 LY 42.52 41.98 LC 0.15 0.25 LD 0.18 0.25 LM 0.93 0.69 L 3.11 1.93

Total leaves 49.08 47.15

BU 14.91 14.51 BR 11.54 12.76 BA 0.22 0.12 B 0.45 0.70

Total fruits 27.12 28.09

FB 6.50 6.64 F 6.45 6.84 FA 7.02 8.18 FP 0.75 1.08

Total flowers 20.72 22.74

No item 3.07 2.02 (i.e. unknown) Chapter 6: Feeding behaviour 160

•.•.. Table 6.12 (cont.)

b) GRP21 % feeding time Item a unweighted weighted

LB 3.84 4.00 LY 42.04 42.59 LE 1.10 1.29 LM 0.47 0.36 L 3.76 2.88

Total leaves 51.21 51.12

BU 26.80 27.11 BR 9.46 9.98 B 0.79 0.98

Total fruits 37.04 38.07

FB 2.74 3.38 FA 3.94 4.00

Total flowers 6.68 7.38

No item 5.07 3.43 (i.e. unknown) aRefer to Table 6.1 for definitions of item classes

6.4.3 Food Parts

Not all parts of items were eaten m equal proportions (Table 6.13). In order to consume some of the parts, T. auratus adopted particular behaviour patterns. For example, for some fruits where only the seeds were eaten and the rest of the fruit discarded, the monkeys stripped the fruit of its exterior (usually with the teeth) and then consumed the seeds. Where the pericarp was woody this sometimes required considerable handling time before the seeds were exposed and eaten.

A complete breakdown of the diet (as % feeding time) to the part level, for both GRP3 and GRP21, is provided in Appendix V. Trends were difficult to deduce when considering such a diverse diet so Chapter 6: Feeding behaviour •161

Table 6.13 Proportion of Parts of Items in the Diet (as unweighted % feedin'g time) a) GRP3

Item a Parta % feeding time

B s .08 B w .36 BA w .21 BR F 2.65 BR NP .97 BR s .98 BR w 6.92 BU F .43 BU NP .61 BU s 6.07 BU w 7.79 F NP 6.44 FA NP 7.01 FB NP 6.50 FP NP .75 L A .50 L B 1.31 L BP .07 L NP 1.06 L p .14 LB NP 2.19 LC NP .14 LD B .18 lM A .03 lM B .54 lM M .26 lM NP .03 lM p .04 LY B 4.70 LY BS 1.72 LY NP 1.51 LY p .31 LY PS .07 LY1 A .07 LY1 B 2.61 LY1 BP .09 LY1 M .06 LY1 NP .23 LY1 p .03 LY2 A .22 LY2 B 8.76 LY2 BP .09 LY2 c .03 LY2 M 2.39 LY2 NP .07 LY2 p .32 LY3 A .64 Chapter 6: Feeding behaviour 162

••••• Table 6.13 (cont.) a) GRP3

Itema Parta % feeding time

LY3 B 6.07 LY3 BP .01 LY3 c .01 LY3 M 8.77 LY3 NP .06 LY3 p .26 LY4 A .11 LY4 B .75 LY4 BP .02 LY4 M 2.34 LY4 p .05 NI NP 3.07

b) GRP21

B NP .78 BR F 3.73 BR NP .15 BR s 2.18 BR w 3.37 BU F .55 BU NP .94 BU s 16.86 BU w 8.44 FA NP 3.94 FB NP 2.73 L B .65 L NP 2.94 L p .15 LB B .05 LB NP 3.79 LE BS 1.10 LM B .47 LY B 4.25 LY BP .47 LY NP .71 LY1 B 5.62 LY1 NP .22 Chapter 6: Feeding behaviour 163

••••• Table 6.13 (cont.) b) GRP21

Itema Parta % feeding time

LY2 A .03 LY2 B 9.94 LY2 BP .13 LY2 BS .02 LY2 NP .15 LY3 A .45 LY3 B 16.82 LY3 BP .18 LY3 BS .48 LY3 M .31 LY3 p .15 LY4 B 1.88 LY4 BS .07 LY4 NP .08 NI NP 5.06 aRefer to Table 6.1 for definitions of item and part classes specific examples will be given, when appropriate, to illustrate important findings.

Although the diet of both groups of T. auratus may be considered diverse when considering all combinations of species, items and parts eaten, the monkeys were selective in their choice of foods even at the part level. For example, the species most extensively eaten by GRP3 was T. grand is (14.51% of the total % feeding time, Table 6.2). Young T. grandis leaves contributed 14.13% to the total diet (Table 6.10) and the mid-ribs of these young leaves contributed 12.80% to the total diet. In other words, 88% of this top food species for GRP3 consisted of mid-ribs of young leaves. To obtain a mid-rib, a young leaf was generally pulled off a branch with one of the forelimbs, bent in half (from apical tip to base) and held by one or both of the forelimbs while the mid-rib was stripped, chewed and eaten. The mid-rib of the basal portion of the leaf (c. 20% of the total mid-rib) was usually not eaten. The rest of the leaf was then discarded. Feeding on T. grandis leaves was therefore highly selective, young leaves being selected in Chapter 6: Feeding behaviour 164

preference to the more abundant mature leaves and mid-ribs generally being selected over the more readily obtainable leaf blades.

6.4.4 Minor Food Sources

Less than 1% of the total diet for GRP3 (and none for GRP21) was made up of non-plant material. These items in the GRP3 diet were insects (larval stage), which were eaten from a tree trunk (Table 6.2 and Table 6.7).

In addition to observations made usmg the scan sampling method, a single account of an adult female (in GRP3) eating sand was noted. Soil-eating has been recorded for African colobines (Clutton-Brock 1972; Marsh 1978b; Oates 1978) and Asian colobines (Poirier 1970; Ripley 1970; Hladik and Hladik 1972; Bennett 1983; Davies 1984). Soil from termite mounds is also eaten by colobines (Hladik and Hladik 1972; Marsh 1978b; Davies 1984; Davies and Baillie 1988) and it has been suggested that this may aid in the absorption of tannins or toxins (Feeny 1969; Hladik 1977) or may help to alleviate digestive disorders to the forestomach microbes (Oates 1978; Davies and Baillie 1988). Possibly, eating sand functions in a similar way.

Another explanation is that the sand provided an important nutrient which was lacking, or low, in the diet of this individual. The sand eaten was subject to tidal ebbs and flows and was probably coated with salt particles. It may, therefore, have provided a source of sodium. Selection of items high in sodium has been reported for C. guereza in their habit of eating pond plants in small amounts at regular intervals (Oates 1978).

A further possibility is that the sand acted as an antacid agent returning the pH of the forestomach to a neutral pH. This would be beneficial if VF A production was proceeding at a high rate, causing an acidic pH in the forestomach, which could no longer be adjusted Chapter 6: Feeding behaviour '165

by swallowing saliva (Kay et al. 1976) and which could be detrimental to the forestomach microbes (see Section 6.8).

Insect matter was possibly ingested when feeding on vegetation as this would not have been observed. For example, T. auratus ate the ripe fruits of F. sinuata trees; several ripe fruits collected from the forest floor contained fig wasps. Trachypithecus senex which also feeds on ripe Ficus fruits is thought to ingest animal matter in the process (Hladik 1977). Trachypithecus cristata has been observed to eat dry and decayed wood with no separation of insects from the wood (Bernstein 1968; Medway 1970a). Generally, little animal matter is eaten by colo hines (Hladik 1977; Oates 1977 a; McKey 1978; Bennett 1983; Davies 1984) and it is thought that sufficient animal proteins are probably obtained by digestion of the forestomach microbes (Parra 1978).

6.5 MONTHLY VARIATION IN FEEDING

The proportion of various items in the diet varied each month throughout the study (Fig. 6.2). The proportion of fruit in the GRP3 diet was greater during July, August, September, January and February (37% (median)) and lower during October, November and December (12% (median)) (Table 6.14a). The proportion of young leaves and leaf buds in the diet was greater during October, November, December (68% (median)) and lower (31% (median)) in other months (Table 6.14a). When fruit consumption was at its lowest the greatest proportion of the diet consisted of young leaves and leaf buds (Fig. 6.2a). The proportion of "other" leaves (including mature leaves) in the diet was consistently low (Fig. 6.2a). Flower consumption followed a similar trend to fruit consumption for July to November but then showed a gradual increase from December to February, with the highest consumption of flowers in February (Fig. 6.2a).

The proportion of various items in the diet of GRP21 also varied each month. The trends followed, however, were quite different to those for GRP3. Fruit consumption was highest in November and Chapter 6: Feeding behaviour 166

80 a) GRP3

Cl) 60 E

-C) c: 40 "0 Cl) Cl) - 20 0~

0 J A s 0 N 0 J F Month

80 b) GRP21

Cl) 60 ·--E C) c: :0 40 Cl) -Cl) 0~ 20

~ 0 0 : N D J F M Month

---1::1-- Young leaves+ leaf buds Other leaves a Fruit Flowers + flower buds

Figure 6.2 Proportions of various items in the monthly diet (as unweighted % feeding time) Chapter 6: Feeding behaviour 167

Table 6.14 Comparison Between Months of High and Low Dietary Intake of Fruits, and of Young Leaves and Leaf Buds a) GRP3

Variable 1 Variable 2 Item U p

Oct, Nov, Dec Jul, Aug, Sept, Proportion of 34 7 15 25 < 0.01 Jan, Feb fruit in daily diet

Proportion of 35 6 15 25 < 0.01 young leaves and leaf buds in daily diet

b) GRP21

Nov, Dec Jan, Feb, Mar Proportion of 114 10 15 < 0.05 fruit in daily diet

Proportion of 124 10 15 < 0.01 young leaves and leaf buds in daily diet

December for GRP21 (52% (median)) (Table 6.14b) at a time when fruit consumption was low for GRP3 (Fig. 6.2a). The proportion of young leaves and leaf buds in the diet was greater during January, February, March (56% (median)) and lower in November and December (31% (median)) (Table 6.14b). The proportion of "other" leaves (including mature leaves) in the diet was consistently low (Fig. 6.2b). The amount of fruit eaten was least during February, when the amount of young leaves eaten was greatest (Fig. 6.2b). The proportion of flowers in the diet was low for November to February (2% (median)) and greatest (12% (median)) in March.

In a seasonal environment, where fruits are produced for certain months of the year and not in other months, fruits are preferentially selected during the fruiting season (e.g. Davies 1984; Davies et al. 1988). As the Pangandaran Nature Reserve does not have distinct seasons (Section 2.2.3) it is not surprising that peaks Chapter 6: Feeding behaviour 168

m the fruit contribution to the diet of the two T. auratus groups did not coincide. This may be due to local differences in fruit production between the study sites or be related to differences m fruit production for preferred species. Whilst the months in which fruits contributed the most to the diet were not the same for the two groups, there was a similarity between the two groups in that the consumption of young leaves peaked at a time when fruit contribution to the diet was at its lowest (Fig. 6.2).

In an attempt to determine the species which were highly selected, the dietary intake of items of particular species was compared with the availability of fruits, flowers and young leaves for these species each month. All species items contributing > 5% each to a monthly diet, and for which phenological data were available (Section 3.2.2), were examined (Table 6.15).

GRP3 (Table 6.15a) selected the fruits of F. sinuata, S. macrophylla, C. excavata and F. sumatrana as the fruits of these species were eaten whenever available. Ficus sinuata and S. macrophylla were possibly more strongly selected than C. excavata. Clausena excavata, while being as available (as a source of fruit) in a few other months was eaten most in August and September (Table 6.15a) when the former species were not fruiting. Ficus sumatrana was eaten most when available as ripe fruit (January) (Table 6.15a) contributing < 5% to any other monthly diet when available largely as unripe fruit.

Pterospermum javanicum fruits were eaten most by GRP3 in September (Table 6.15a) when most were available as ripe fruit. Pterospermum javanicum fruits were available in other months when they were not eaten (e.g. July). During these months, however, the number of phenology trees bearing fruit was less. This may reflect selection of particular trees for this species. One tree belonging to this species was located on the extreme edge of the GRP3 home-range area. When this tree had many ripe fruits (September) they were extensively eaten by GRP3. A neighbouring T. auratus group also ate the ripe fruits of this tree and always displaced GRP3 from this food source on arrival. Another tree Table 6.15 Species Items which Each Contributed More than Five Per Cent to a Monthly Diet and Proportion of Fruit, Flowers, and Young Leaves in Each Monthly Diet a) GRPJ

Month Fruit items Flower items Young leaves & leaf buds Species % feeding time Species % feeding time Species % feeding time

Ficus sinuata Pterospermum javanicum 14.80 - ripe 6.84 Kleinhovia hospita 6.32 JUL - unripe 6.84 - total 13.68 Total fruit 35.39 Total flowers 22.77 Total young leaves 28.25

Eugenia zippelianum Pterospermum javanicum 23.75 Ficus sinuata 20.82 - unripe 17.64 AUG Clausena excavata - ripe 12.93 Total fruit 35.11 Total flowers 26.65 Total young leaves 36.39

Pterospermum javanicum Celtis philippensis 8.55 Swietenia macrophy/la 16.32 - ripe 23.72 Ficus sinuata 7.69 Clausena excavata Tectona grandis 5.41 SEP - ripe 10.85 - unripe 3.85 - total 14.70 Total fruit 38.41 Total flowers 11.89 Total young leaves 40.08 •••.. Table 6.15 (cont.) n ::r a GRP3 !)) '1:1... (l '"1 & Month Fruit items Flower items Young leaves leaf buds 0\ Species % feeding time Species % feeding time Species % feeding time 't1 (l e:(l Tectona grandis tj Kleinhovia hospita 5.90 19.11 [JQ Ficus sp. (515) 7.74 C1' ocr Ficus melinocarpa 6.35 (l::r !)) Vitex pinnata 5.55 :::. Total fruit 13.69 Total flowers 11.69 Total young leaves 68.34 0c:: '"1

Cynometra ramiflora Tectona grandis 31.42 NOV - unripe 7.83 Ficus sinuata 14.09 Total fruit 18.46 Total flowers 6.38 Total young leaves 68.91

Anthocephalus cadamba 10.76 Tectona grandis 34.00 DEC Bridelia monoica 5.52 Total fruit 10.37 Total flowers 16.76 Total young leaves 65.46

Ficus sumatrana Celtis philippensis 15.51 Tectona grandis 11.45 - ripe 15.16 Kleinhovia hospita 7.06 Ficus sp. (515) 5.28 Swietenia macrophylla - unripe 7.54 JAN Cinnamomum iners - unripe 6.66 Total fruit 39.78 Total flowers 26.94 Total young leaves 28.63 .... -...I Ficus sinuata Kleinhovia hospita 22.25 Tectona grandis 11.55 0 FEB - unripe 8.57 lleritiera littoral is 10.14 Guellarda speciosa 5.23 Total fruit 20.94 Total flowers 37.73 Total young leaves 32.51 n ::r •.... Table 6.15 (cont.) ~ ..."'CD b GRP21 '"I ~ Month Fruit items Flower items Young leaves & leaf buds 'TI Species % feeding time Species % feeding time Species % feeding time CD e:CD 1:1 (JQ Hernanadia peltata Cinnamomum iners 5.91 0" CD - ripe 11.55 ::r Sterculia coccinea ....<"' 0 - unripe 10.73 ='"I Dysoxylum NOV caulostachyum - ripe 9.32 Ficus benjamina - ripe 8.48 - unripe 0.61 - total 9.09 Total fruit 54.62 Total flowers 2.73 Total young leaves 36.56

Dysoxylum Cinnamomum iners 7.27 caulostachyum Guettarda speciosa 8.51 - unripe 19.64 Sterculia coccinea - ripe 0.53 DEC - unripe 10.89 - total 11.42 Ficus benjamina - ripe 2.70 - unripe 4.96 - total 7.66 ..... Total fruit 52.54 Total flowers 5.67 Total young leaves 35.52 -...l..... n t:r ••••• Table 6.15 (cont.) I>) "t:l b GRP21 ...n "1 0\ Month Fruit items Flower items Young leaves & leaf buds Species % feeding time Species % feeding time Species % feeding time 't1 n n ;;·c:lo OQ Dysoxylum Erythrina variegata 12.31 tT n caulostachyum Ficus benjamina 6.10 t:r JAN - unripe 8.85 Linociera sp. 5.13 I>)

Ficus bracteata Ficus sinuata 36.14 - unripe 7.08 Guettarda speciosa 5.31 FEB Erythrina variegata 5.09 Total fruit 14.60 Total flowers 5.53 Total young leaves 63.79

Ficus benjamina Guettarda speciosa 13.10 - ripe 0.35 Erythrina variegata 12.07 - unripe 9.86 MAR - total 10.21 Ficus sumatrana - unripe 5.53 Total fruit 26.71 Total flowers 13.25 Total young leaves 50.01 Chapter 6: Feeding behaviour ·173

belonging to this spec1es, appeanng to be of similar maturity and also bearing ripe fruits, was not eaten by GRP3 despite this tree being well within the GRP3 home-range area and being passed-by on several occasiOns.

Eugenia zippelianum fruits contributed a major component to the GRP3 diet in August (Table 6.15a) and were also eaten in small quantities in July. They were not eaten in December when they were also available in similar quantity to August. No explanation for this is forthcoming, especially considering that the December diet had the lowest fruit intake of all months (Fig. 6.2a) but perhaps the location of the trees on the extreme eastern boundary of the GRP3 home-range area, on a rocky outcrop jutting out over the sea, may have had an influence. Perhaps the availability of food from other species in this extreme location determined selection of E. zippelianum. Phenological data are not available for Cynometra ramiflora or Cinnamomum iners so it is not known whether these species were strongly selected.

Of the ripe fruits which contributed > 5% to any monthly diet (Table 6.15a) F. sinuata and F. sumatrana were eaten as whole fruits. Trees of these species, like many Ficus spp., exhibit asynchronous intraspecific fruiting (Medway 1972; Janzen 1979; Corlett 1987). The fruits of these species were typically available m large quantities on a given tree for a few days only. It was not uncommon to see T. auratus, M acaca fascicularis and sometimes Pteropus vampyrus (flying fox), feeding in the same tree, when the ripe fruits of these species were available, without apparent displacement of one species by another. Fruits were therefore plentiful, and possibly one or more of the animal species feeding on the fruits was acting as a seed-dispersal agent. These fruits were sweet to taste (pers. obs.).

All parts of P. javanicum npe fruits were also eaten by GRP3. The fruits of this species are very woody in appearance and the pericarp not sweet (pers. obs.). The ripe fruits selected had only recently ripened (judging by the splitting of the peri carp). They were eaten one-at-a-time (unlike F. sinuata and F. sumatrana Chapter 6: Feeding behaviour 174

fruits of which several were ingested at a time) and while all parts of the fruit were eaten, the whole fruit was generally not eaten, although seeds were rarely discarded.

Ripe fruits of C. excavata were also eaten by GRP3 (Table 6.15a). Usually only the pericarp was eaten although sometimes the whole fruit was eaten. The parts eaten by GRP3 may have had a high sugar content although they lacked a sweet flavour (pers. obs.). Fruits which may have been sweet, therefore, did not contribute more than 15% to any monthly diet (from Table 6.15a).

The flower items of five species each contributed > 5% to monthly diets of GRP3 (Table 6.15a). Phenological data are not available for three of these species - C. philippensis, Anthocephalus cadamba and H eritiera littoralis. Pterospermum javanicum flower items were selected when available and contributed major proportions to the diet for these months (July and August) (Table 6.15a).

Kleinhovia hospita flower items were available in all months but were not eaten every month and sometimes only contributed small percentages to the total % feeding time for a month. Kleinhovia hospita flowers contributed most to the diet in February (Table 6.15a) which was one of the two months in which they were most abundant (January being the other month). Rather than being a highly selected species, K. hospita flowers provided a standby being eaten to a greater or lesser degree depending on the availability of items of other species. A neighbouring T. auratus group (not the same neighbouring group referred to previously) was observed eating K. hospita flowers from trees which were also exploited as a food source by GRP3, once when GRP3 was 50 m distant. These trees were well within the home-range area of GRP3 and yet GRP3 did not displace this group. This would suggest that either K. hospita flowers were not a highly preferred food source or that the food source was in sufficient abundance not to incite competitive displacement. Chapter 6: Feeding behaviour 175

Young T. grandis leaves were available in all months of the study for GRP3. They were eaten most in the months when fruit and flowers were low in the diet (October to December (Table 6.15a, Fig. 6.2a)). Young leaves of F. sinuata were also available throughout the study but were most abundant in August during a flush of very young leaves when they were eaten extensively (Table 6.15a). During this month, T. grandis contributed the least to the monthly diet (Table 6.15a). Swietenia macrophylla young leaves were available for several months but only contributed > 5% to a monthly diet in September (Table 6.15a) which also coincided with a flush of young leaves. Ficus sinuata and S. macrophylla were, therefore, selected when abundant as very young leaves. The young leaves of T. grand is were not highly selected but rather acted as a standby when preferred young leaves or other preferred items (e.g. fruits of certain species) were not available.

Young leaves of Ficus sp. (515), F. melinocarpa, V. pinnata, Bridelia monoica and Guettarda speciosa may also have been selected preferentially, but no phenological data are available for these species to determine presence/absence of young leaves for these species throughout the study. Their contribution, however, was less than the species considered above and only accounted for 5-8% of any monthly diet (Table 6.15a).

For GRP21, the fruits selected preferentially were the fruits of H. peltata, D. caulostachyum, F. benjamina and Sterculia coccinea as the fruits of these species were always eaten when available, contributing large proportions to the monthly diet (Table 6.15b). In addition, ripe H. peltata fruits were eaten in December and March, contributing < 5% to the total feeding time for these months, months in which ripe H. peltata fruits were not recorded in the phenology sample. H ernandia peltata did not fruit synchronously, therefore, although presumably the majority of trees of this species bore fruit in November as this was when the trees in the phenology sample for this species bore the most fruit. Similarly, S. coccinea contributed a small proportion ( < 5%) to the diet in January when one of the five trees in the phenology sample for this species was carrying a small amount of unripe fruit. Chapter 6: Feeding behaviour 176

The availability of fruits in November and December for the four species mentioned above largely accounts for the high proportion of fruit in the GRP21 diet in these months (Table 6.15b). Dysoxylum caulostachyum fruits were also available and eaten in January, and F. benjamina fruits in March (Table 6.15b). Fruit of these species was not available in February when dietary fruit intake was at its lowest (Fig. 6.2b). In February, only the fruit of F. bracteata contributed > 5% to the monthly diet (Table 6.15b). Phenological data are not available to determine whether this species was selected preferentially in this month or whether the fruits of this species were available in other months when they were not eaten. Similarly, phenological data are not available for Nauclea sp., the fruits of which were eaten only in January (Table 6.15b).

Ficus sumatrana fruits were available in all months except January but were eaten only in December (< 5%) and March (Table 6.15b) when one of the trees in the phenology sample was carrying many unripe fruits, the tree bearing the fruits differing between the two months. This indicates individual tree selectivity for this species as a food source of fruit as the tree recorded as having many unripe fruits in March was the tree which the monkeys ate from during this month. This was the only month in which this tree bore many unripe fruits. The most important species as a source of fruit for GRP21 were, therefore, H. peltata, D. caulostachyum, F. benjamina and S. coccinea, with specific selection for one F. sumatrana tree when carrying unripe fruits in a large quantity.

Of the ripe fruits which contributed > 5% to any monthly diet for GRP21 (Table 6.15b), F. benjamina fruits were eaten whole. Ficus benjamina fruit production and predation thereof was similar to F. sinuata and F. sumatrana described previously for GRP3. Like F. sinuata and F. sumatrana, F. benjamina fruits were also sweet (pers. obs.).

When GRP21 ate the ripe fruits of D. caulostachyum (Table 6.15b), newly ripe fruits (as determined by the narrow split in the pericarp) were selected. On 90% of occasions, only the seeds were Chapter 6: Feeding behaviour •177

eaten and the woody pericarp discarded. Ripe H. peltata fruits were also eaten by GRP21 (Table 6.15b); usually only the fleshy pericarp was eaten and the seeds discarded. The pericarps of H. peltata fruits may have had a high sugar content although not sweet to taste (pers. obs.). Fruits which may have been sweet, therefore, did not contribute more than 20% to the total diet in any month (from Table 6.15b).

Flower items contributed less to monthly diets for GRP21 than corresponding months for GRP3 (Fig. 6.2). Flowers of food tree species did not contribute > 5% each to any monthly diet for GRP21 (Table 6 .15b) and were therefore presumably not strongly selected. Flowers contributed 2-6% to the mean monthly diet for November to February and approximately 13% to the diet in March (Fig. 6.2b). Species were possibly being selected when flowers were available but, if so, the contribution to and effect on diet proportions was considerably less for GRP21 than for GRP3.

Young leaves contributed more to the GRP21 diet when fruit contribution was less (January, February and March). The highest intake of young leaves occurred in February when the fruit contribution was least (Fig. 6.2b). Young leaves contributed approximately 64% to the diet in February and over half of this (36%) consisted of the young leaves of F. sinuata (Table 6.15b). Phenological data for this species are not available for GRP21. However, it is likely that this peak of young F. sinuata leaves in the diet corresponded with a flush of very young leaves, as was described above for GRP3.

Young F. benjamina leaves were available in all months but only contributed > 5% to the monthly diet in January (Table 6.15b). The young leaves of this species were, therefore, not strongly selected. Young leaves of other species contributing > 5% each to any monthly diet- C. iners, G. speciosa, Erythrina variegata and Linociera sp. (Table 6.15b), may have been selected preferentially when available or acted as standbys in the absence of preferred species. However, phenological data were not collected for these species so it is not possible to determine whether they were selected Chapter 6: Feeding behaviour 178

preferentially or not. Their contribution, however, was less than F. sinuata and only accounted for 5-13% each of any monthly diet (Table 6.15b).

6.6 DIETETIC DIVERSITY

A tendency to increase dietary diversity when favoured food is scarce has been reported for C. guereza (Oates 1977a), C. satanas (McKey and Waterman 1982), P. melalophos (Bennett 1983) and P. rubicunda (Davies 1984). This has been regarded as according with optimal foraging theory. When preferred foods are scarce the cost of searching for high-quality items increases to a point where it is no longer cost-effective and it becomes more profitable to eat foods that are less nutritious than preferred foods, but more abundant (Schoener 1971). Foods which are not highly preferred may also contain more toxins than highly favoured items and a diverse diet in times of preferred food scarcity may be advantageous in that small amounts of a range of toxins may be preferable to large quantities of a few (Freeland and Janzen 1974).

Colobus badius tephrosceles showed a reverse trend in that dietetic diversity decreased when preferred foods were scarce. When preferred young leaves were unavailable this species ate mature leaves of a few tree species only (Struhsaker 1975).

The Shannon-Weaver diversity index H' (Pielou 1966) has been used in previous studies as the measure of item and species diversity;

where Pi is the relative abundance of each item or species m the diet and logpi is the natural logarithm of Pi· H' is insensitive, however, because it effectively reduces variance (May 1981). Thus, Chapter 6: Feeding behaviour 179

Simpson's index Bi(Simpson 1949; Fox 1981) has also been used as a measure of diversity, where

Spearman rank correlation coefficients (Siegel 1956) were initially calculated using the Shannon-Weaver diversity index for item diversity and the proportion of fruit in the diet (as % feeding time). The highest correlation coefficient resulted when the four main item classes were considered (i.e. all fruit, all flower items, young leaves and leaf buds, and remaining leaf items) in determining the diversity index, as opposed to all item classes used; and the proportion of all fruit in the diet, as opposed to considering ripe fruit and unripe fruit separately. Consequently, in subsequent calculations of correlation coefficients between item diversity and the proportion of fruit in the diet, the four main item classes and the proportion of all fruit in the diet were used, respectively.

For GRP21, correlation coefficients were calculated for the proportion of fruit in the diet and species/item diversity and the proportion of fruit and young leaves in the diet and species/item diversity (Table 6.16). For GRP3, correlation coefficients were determined for the proportion of fruit in the diet and species/item diversity, and the proportion of fruit and flowers in the diet and species/item diversity (Table 6.16). Flowers were considered for GRP3 as flowers of a few species were preferentially selected by GRP3 but not by GRP21 (Section 6.5).

Where Simpson's index was used as the measure of diversity, species/item diversity was not correlated with any of the variables tested for GRP21 and GRP3. Where the Shannon-Weaver index was used, significant correlations were only found for GRP3 between item diversity and the proportion of fruit in the diet, and item diversity and the proportion of fruits and flowers in the diet (Table 6.16). Chapter 6: Feeding behaviour 180

Table 6.16 Correlation Between Dietetic Diversity (as Species Diversity or Item Diversity), Using the Shannon-Weaver Diversity Index, and the Proportion of Various Items in the Diet (as % Feeding Time) of GRP21 and GRP3

T. auratus Variable 1 Variable 2 group (using Shannon -Weaver index)

Proportion Species diversity 0.26 n.s. of fruit in diet Item diversity 0.31 n.s.

GRP21 Proportion Species diversity 0.35 n.s. (n = 25) of fruit and young leaves Item diversity 0.29 n.s. in diet

Proportion Species diversity 0.22 n.s. of fruit in diet Item diversity 0.51 < 0.001

GRP3 Proportion Species diversity 0.27 n.s. (n = 40) of fruit and flowers in diet Item diversity 0.76 < 0.001

an.s. = not significant

Figure 6.3 is a plot of the datapoints for item diversity and the proportion of fruit in the GRP3 diet. Figure 6.3 shows a wide scattering of datapoints. Thus, whilst a positive correlation was found between item diversity (using the Shannon-Weaver index) and the proportion of fruit in the diet (Table 6.16) the relationship was not tight.

Item diversity increased as the proportion of fruits and flowers m the GRP3 diet increased up until c. 40% when a further increase m the proportion of fruits and flowers in the diet did not appear to result in further appreciable increases in item diversity (Fig. 6.4). Therefore, correlation coefficients were calculated between item diversity and the proportion of fruits and flowers in the diet where the contribution to the diet was ~ 40%, and separately for item diversity and the proportion of fruits and flowers in the diet where Chapter 6: Feeding behaviour '181

1.4

1!1 1!1 Ill 1!1 1!1 1!1 1!1 1.2 1!1 1!1 1!1 1!1 !Eilll lilt:~ IE 1!1 "hP 1.0 1!1 1!1 1!1 1!1 1!1 >- 1!1 1!1 -...(/) 1!1 Cl) 0.8 - 1!1 > 1!1 1!1 '0 1!1 1!1 0.6 1!1 E Cl) 1!1 1!1 0.4 1!1

0.2

0.0 0 10 20 30 40 50

proportion of fruit in diet

Figure 6.3 Item diversity (using Shannon-Weaver index) and the proportion (as % feeding time) of fruit lll the GRP3 diet

1.4

1!1 1!1 1!1 ~ 1!1 1.2 l!l a 1!1 1!1~1!1 Ill . 1!1 1!1 '8 1!1 1!1 1!1 1!1 1!1 1!1 1.0 i:P >- 1!1 1!1 -(/) 1!1 ... 0.8 1!1 Cl) 1!1 1!1 > 1!1 '0 1!1 m 0.6 1!1 E Cl) 1!1 - 1!1 0.4 1!1

0.2

0.0 0 10 20 30 40 50 60 70 80

proportion of fruit + flowers in diet

Figure 6.4 Item diversity (using Shannon-Weaver index) and the proportion (as % feeding time) of fruit and flowers in the GRP3 diet Chapter 6: Feeding behaviour 182

the contribution to the diet was > 40%. A significant correlation (r8 = 0.54, n = 40, p < 0.001) was found for the former but not the latter.

From the results described above and presented in Table 6.16, Figure 6.3 and Figure 6.4, several points are worth noting:

1. Where Simpson's index was used as the measure of diversity, significant correlations were not found for any of the pairs of variables tested;

2. Where the Shannon-Weaver index was used as a measure of diversity a) significant correlations were not found for any of the pairs of variables tested for GRP21; b) significant positive correlations were found for GRP3 when considering item diversity but not for species diversity.

In general, a poor correlation existed between dietetic diversity and the proportion of preferred food items in the diet of both T. auratus groups. Certainly, species and item diversity did not decrease as preferred food items in the diet increased as has been described m other studies of colobine feeding behaviour (e.g. P. melalophos (Bennett 1983); P. rubicunda (Davies 1984)).

Factors which may account for this are the absence of distinct climatic seasons regulating fruit and flower production (Section 2.2.3, Section 6.5) in the Pangandaran Nature Reserve and the availability of acceptable food from a few species when preferred foods were scarce. Consequently, dietetic diversity remained fairly constant within, and between, months. Presbytis rubicunda at Sepilok (Davies 1984) by comparison, for example, lives in a highly seasonal environment and, consequently, the proportion of fruit in the monthly diet varied from c. 30% to 80%. It is not surprising that in such a seasonal environment the preferred food sources, fruits of a few species, were highly selected when available. When these were not available, P. rubicunda ate young leaves of several species (Davies 1984). Chapter 6: Feeding behaviour 183

Dietetic diversity (as species diversity) and home-range diversity (as species diversity determined from quadrat sampling of the vegetation) for both GRP21 and GRP3 are shown in Table 6.17. Dietetic diversity for GRP3 is based on results of the first five monthly scan samples only to be consistent with the number of monthly scan samples analysed for GRP21. Both the Shannon­ Weaver and Simpson's indices, as measures of diversity, are presented.

Table 6.17 Dietetic Diversity (as Species Diversity) and Home-Range Diversity (as Species Diversity Determined from Quadrat Sampling of Vegetation) for GRP3 and GRP21

a) GRP3

n Shannon-Weaver Index Simpson's Index

Dietetic diversity 69 3.22 84.8 X 106

Dietetic diversity 58 3.18 24.2 X 105 incl. only spp. which contributed 2 1% to the total diet

Home-range diversity 8 3 3.47 50.5 X 105

b) GRP21

n Shannon-Weaver Index Simpson's Index

Dietetic diversity 46 3.26 20.4 X 105

Home-range diversity 46 4.10 21.6 X 106

Dietetic diversity for GRP3 was very high when using Simpson's index and greater than home-range diversity. When all species contributing < 1% to the total diet were removed, and Simpson's index recalculated, the dietetic diversity decreased appreciably. This indicates that several rare species (i.e. species not sampled in the vegetation analysis) contributed very small amounts to the total diet (Table 6.17). Using Simpson's index (and the less reliable Chapter 6: Feeding behaviour 184

Shannon-Weaver index), dietetic diversity was similar for the two groups but home-range diversity was far greater for GRP21 than for GRP3. Therefore, GRP21 was more selective in choice of species as food than GRP3.

6.7 DIURNAL VARIATION IN FEEDING WITH RESPECT TO ITEMS IN THE DIET

Primates with simple stomachs generally eat more fruit early m the day and more foliage late in the afternoon. During the night, an energy deficit may accumulate and fruit-eating early in the morning may restore the deficit (Chivers 1975). Eating foliage late in the afternoon may act to keep the digestive system active during the night (Clutton-Brock 1977b). Alternative explanations for changes in food choice throughout the day are provided by Raemaekers (1978) in a discussion of food choice of gibbons.

Primates with compound stomachs do not appear to follow this pattern in feeding behaviour (Bennett 1983). Two main peaks in feeding were observed for T. auratus at Pangandaran: early morning; late afternoon (Section 5.7.2). The proportions of different items in the diet between 0600-0800 h and 1600-1800 h were compared using focal-animal sampling data (Mann-Whitney U). The proportion of leaves in the diet did not vary between the two allotted time periods (U = 112, n1 = 14, n2 = 14, n.s.). Similarly, the proportion of ripe fruit in the diet (U = 107, n1 = 14, n2 = 14, n.s. ), the proportion of unripe fruit (U = 130, n 1 = 14, n2 = 14, n.s.) and the proportion of flowers (U = 102, n 1 = 14, n2 = 14, n.s.) in the diet were similar for both time slots.

The digestive system of colobines IS such that VFA production may continue at a constant rate throughout the night (Section 6.2), releasing nutrients gradually. Consequently, T. auratus would not need to vary the item intake in the diet according to the time of day. Chapter 6: Feeding behaviour ·185

6.8 DISCUSSION

GRP3 spent significantly more time feeding than GRP21 (Section 5.3.1). This suggests that the quality of food was higher for GRP21 than for GRP3. As there was a higher proportion of fruit in the diet for GRP21 (Fig. 6.1, Section 6.4.2) and fruit was strongly selected by both groups when available for certain species (Section 6.5) it is likely that the GRP3 home-range area contained poorer quality food than the GRP21 home-range area.

Differences in high-ranking food species for the two groups were generally found to be related to differences in presence/absence (or relative abundance) of species in the two study sites or, where species were present in both sites, to differences in availability of food items for the two sites (Section 6.4.1.1, Section 6.5). Possibly, at some time in the future, when trees belonging to H. peltata and D. caulostachyum (the fruits of which were selected by GRP21) reach maturity and bear fruit in the GRP3 home-range area, the diet of the two groups may be more similar. During the study period, however, these fruits were generally not available to GRP3 (Section 6.4.1.1).

The proportion of leaves (largely young leaves) in the diets of GRP3 and GRP21 was similar, but the proportion of flowers was not (Fig. 6.1, Section 6.4.2). Flowers contributed a greater proportion to the GRP3 diet than to the GRP21 diet (Section 6.4.2) and flowers of certain species were selected by GRP3 when available while flowers were not strongly selected by GRP21 (Section 6.5). Young leaves contributed most to the diet for both groups during months when the proportion of fruit in the diet was least (Fig. 6.2, Section 6.5). Very young leaves of F. sinuata were selected when available by both groups (Section 6.5). In a home-range area with less fruit available as food, GRP3 compensated by increasing the proportion of flower items in the diet. In addition, use was made of two plantation species, T. grandis and S. macrophylla, which together comprised c. 28% of the vegetation (as % of total basal area). The very young leaves of S. macrophylla were selected when available and the young leaves of T. grandis, available throughout the study, Chapter 6: Feeding behaviour 186

acted as a dietary standby contributing large proportions to monthly diets when preferred items were scarce (Section 6.5). GRP21 did not have a major dietary staple.

The percentage of time spent feeding did not vary between months for the two T. auratus study groups (Section 5.3.2). Bennett (1983) in a study of P. melalophos found that time spent feeding did vary between months but was not correlated with the proportion of foliage in the diet. This is in contrast to Davies (1984) who found that percentage of time spent feeding by P. rubicunda in a dipterocarp forest varied between months and feeding time was positively correlated with the proportion of foliage in the diet. More time was spent feeding during months in which fruit was scarce in order to obtain enough energy from the less digestible foliage diet. This suggests that the Pangandaran Nature Reserve, like the Krau Game Reserve where P. melalophos was studied (Davies et al. 1988; Waterman et al. 1988), had a greater abundance of high quality food available at all times than at Sepilok where P. rubicunda was studied (Davies 1984).

At Sepilok, the ten highest ranking food tree species, accounting for 24% of the P. rubicunda diet, contributed 3.6% to the total basal area (Davies et al. 1988). By comparison, ten tree species accounted for 43% of the P. melalophos diet and contributed 13.7% to the basal area (Davies et al. 1988). At Pangandaran, ten tree species contributed 61% to the diet of T. auratus (GRP21) and accounted for 31% of tree biomass (as basal area).

An interesting comparison between T. auratus and other colobines is the proportion of ripe fruit in the diet. The proportion of sweet fruit in the diet is generally low for colobines and this has been regarded as implementing ecological separation between sympatric primates with simple stomachs and colobines, the former tending to feed selectively on the pulp of succulent fruits. Such dietary differences between colobines and simple-stomached primates have been reported widely for both Asia (MacKinnon 1974; Hladik and Hladik 1972; Chivers 1974; Hladik 1977; Raemaekers 1977; Gittins and Raemaekers 1980; Bennett 1983; Chapter 6: Feeding behaviour 187

Caldecott 1983; Davies 1984) and Africa (Struhsaker 1975; Oates 1977a; Waser 1977; Homewood 1978; Marsh 1978b; Struhsaker and Leland 1979). A simple-stomached primate, M. fascicularis, is found in the Pangandaran Nature Reserve. While data on the dietary composition of M. fascicularis are not available for this study area, sugary, pulpy fruits form a major part of the diet for this species elsewhere (MacKinnon and MacKinnon 1978; Aldrich­ Blake 1980; Mah 1980).

Differences in food preferences for colobines and simple­ stomached primates have been related to digestive physiology, m particular to the forestomach fermentation digestive system of the colobines. Goltenboth (1976) proposed that a large intake of simple sugars (as are present in sweet fruits) may cause hyperacidity of the forestomach fluid in ruminants which may be detrimental to the microflora (Bauchop and Martucci 1968) and thereby to the animal. Hyperacidity results when the VFAs produced by bacterial fermentation build up too quickly for the animal to absorb or for other microflora to process. Davies (1984) stated that the problem of "acidosis" may be overcome by eating fibrous foods which will slow down the rate of fermentation. A lack of foliage in the diet of captive colobines has lead to diarrhoea, bloat and sometimes death resulting from stomach infections (Hollihn 1971).

Many sweet fruits with watery flesh have low concentrations of nutrients (McKey 1975). Consequently, to obtain sufficient nutrients from such food sources, large quantities need to be consumed and passed through the gut quickly (Davies 1984). A fast passage rate is incompatible with a fermentation digestive system (Janis 1976). This may explain why the proportion of ripe fruit (including sweet fruits) in the diets of GRP3 and GRP21 did not differ from one another although the proportion of unripe fruit was greater in the GRP21 diet (Table 6.11, Section 6.4.2). There may be an upper limit to the amount of ripe fruit acceptable in the T. auratus diet.

It may be expected that T. auratus should have a higher proportion of foliage in the diet, to slow down the rate of digestion to prevent acidosis, than other colobines which do not eat sweet Chapter 6: Feeding behaviour 188

fruits. A comparison of the diets of twelve colobine species (Table 6.18) shows that T. auratus had a higher proportion of foliage m the diet (as leaves and flower items) than P. rubicunda, P. melalophos or C. satanas which tend to feed selectively on the seeds of fruits and avoid sweet, fleshy fruits (Davies et al. 1988).

McKey et al. (1981) reported that C. satanas were never observed to eat fig fruits; T. auratus frequently ate the fruits (often ripe) of Ficus spp. Species which contributed major fruit items to the diets of P. melalophos (Bennett 1983) and P. rubicunda (Davies 1984) were synchronous in their production of fruit indicating that these species may involve a "predator-swamping" strategy. At least several species, which contributed largely to the ripe fruit component of the T. auratus diet, were asynchronous in their fruit production and T. auratus may therefore have been involved in seed dispersal for these species.

The proportion of foliage in the T. auratus diet was similar to that of several other Presbytis and Trachypithecus species: T. obscura, T. johnii, T. senex, P. comata and P. entellus; and considerably less than other C olobus spp. (Table 6.18). Trachypithecus obscura, T. johnii, T. senex, P. comata and P. entellus all ate considerable proportions of whole fruits, or whole fruits and seeds. However, it is not known from the data presented in Table 6.18 whether these whole fruits were ripe or unripe, nor whether they were sweet or not. Frequently fruits have not been classed as ripe or unripe due to the difficulty in determining the stage of maturity for some fruits.

Trachypithecus auratus had the lowest proportion of mature leaves in the diet of any colobine studied (Table 6.18). This may seem surprising if the fibre content of the diet should be increased to prevent acidosis when sweet fruits are eaten (Davies 1984 ), since mature leaves generally have a higher fibre content than young leaves (Waterman 1984; Waterman et al. 1988). However, with an adequate supply of young leaves in the study area at Pangandaran (Section 6.5) the necessity to feed on lower nutrient mature leaves probably does not arise. The fibre content of the young leaves n::r Ill Table 6.18 The Diet of Twelve Colobine Species (Percentage Feeding Time) 'l:l.... 0.... 0\ Species Mature Young Flower Total Whole Whole Seeds Total Source '11 leaves leaves items foliage fruits fruits & only fruits 0 0 0. seeds t;• (JQ c::r ::r0 Trachypithecus Ill aural us ;:;·< ~ GRP3a < 1 45 21 67 18 7 27 This study .... GRP21b < 1 47 7 55 16 19 37 This study T. johnii 27 31 9 67 25 25 Oates et al. 1980 T. obscura 22 36 7 65 32 3 35 Curtin 1980 T. senex 40 20 12 72 28 + 28 Hladik 1977 Presbytis comata ? 59 7 66 14 < 1 14 Ruhiyat 1983 P. entellus 21 27 7 55 45 45 Hladik 1977 P. melalophos 7 26 17 50 9 11 26 46 Bennett 1983 P. rubicunda 36 11 48 19 30 49 Davies 1984 P. thomasi ? 32 8 40 58 Gurmaya 1986 Colobus badius temminckii 20 52 8 80 8 8 Gatinot 1977 C. badius rufomitratus 11 52 6 69 25 25 Marsh 1981a C. badius tephrosceles 21 51 12 84 6 6 Struhsaker 1975 C. guereza 12 62 2 76 14 14 Oates 1977a C. satanas 18 21 3 42 53 53 McKey eta/. 1981 a leaves of unknown maturity = 3%; unknown item = 3% b leaves of unknown maturity 4%; unknown item 5% ..... 00 \0 Chapter 6: Feeding behaviour 190

eaten may be adequate to complement the simple sugar intake present in the sweet fruits consumed.

The proportion of sweet fruits in any monthly diet for T. auratus did not exceed 20% of the total monthly diet (Section 6.5). While this proportion of ripe fruit in the diet is high for a colobine it is still considerably low when compared with that of simple-stomached primates. Furthermore, the avoidance of mature, fleshy fruits is not uniform amongst other colobines. For example, P. entellus ate ripe fruits (up to 80-90 % of monthly diet) including Ficus fruits (Hladik 1977). Interestingly, T. senex, with a higher foliage intake than P. e nt ell us, tended to eat fruits (not more than 50% of monthly diet) which were fibrous and desiccated (Hladik 1977).

Presbytis melalophos and P. ente llus are less folivorous than their sympatric counterparts of T. obscura and T. senex, respectively (Hladik 1977; Curtin 1980). Davies (1984), in comparing the diet of P. rubicunda with P. melalophos and P. entellus, stated that species belonging to the Presbytis genus may not be well adapted to eating foliage. The proportion of foliage in the diet for T. auratus would, therefore, be expected to be similar to other species of the Trachypithecus genus such as T. obscura, T. johnii and T. senex and higher than the proportion of foliage in the diet for species belonging to the Presbytis genus such as P. rubicunda, P. melalophos, P. entellus, P. comata and P. thomasi. When comparing the proportion of foliage in the diet of GRP3 and other Trachypithecus species with Presbytis species (Table 6.18) this relationship holds (with the exception of P. comata). When incorporating GRP21 into the comparison the relationship no longer holds as GRP21 is more similar to P. melalophos and P. e nte llus in the proportion of foliage in the diet than to any of the Trachypithecus species or to GRP3.

The shorter duration of the study for GRP21 may have had an influence here. However, it is important to consider the flexibility in the diets even between populations of a single species (e.g. differences in dietary composition between GRP3 and GRP21) and Chapter 6: Feeding behaviour 191

between months. Marked differences in diet in response to changes in food availability can occur (Cant 1980). For example, the largest difference in proportion of foliage in the diet for Presbytis and Trachypithecus species given in Table 6.18 is between P. thomasi (40%) and T. senex (72%), a difference of 32%. Yet GRP3 varied the proportion of foliage in the diet between months by as much as 32%, the proportion of foliage in the diet being 53% in September and 85% in October and November.

Similarly, the proportion of foliage in the diet varies considerably on a monthly basis for other colobines. For example, foliage contributed c. 10% to the P. rubicunda diet in August and > 60% to the diet in June (Davies 1984). Therefore, differences in the degree of folivory for Presbytis and Trachypithecus species may be a function of differences in vegetation composition and availability of food items. However, where species of the two genera are sympatric, Trachypithecus species are more folivorous than Presbytis species (Hladik 1977; Curtin 1980). This is possibly explained by the larger gut size of most Trachypithecus species (Chivers and Hladik 1984).

6.9 SUMMARY

1. Trachypithecus auratus, like all other colobines, has an enlarged forestomach. Bacteria in the forestomach ferment food, converting carbohydrates to VFAs, which are then absorbed.

2. GRP3 ate food from 88 vegetation species. Tectona grandis was the top food species. Young leaves and leaf buds accounted for 45% of the total diet, fruit for 27% and flower items for 21%.

3. GRP21 ate food from 49 vegetation species. Dysoxylum caulostachyum was the top food species. Young leaves and leaf buds made up 47% of the total diet, fruit comprised 37% and flower items comprised 7%. Chapter 6: Feeding behaviour 192

4. The family which was exploited most as a food source by GRP3 and GRP21 was Moraceae. Trees of this family accounted for c. 20% of feeding time of each group.

5. Differences in the diet of the two groups were largely related to differences in vegetation composition and availability (and abundance) of food items for the species common to both sites.

6. There was considerable monthly variation in the intake of different items, although this was not related to seasonal, climatic changes in the study area. Young leaf intake was greatest in months when fruit intake was low.

7. Trachypithecus auratus were selective m their choice of food as: a) Not all vegetation species were eaten. Trachypithecus auratus tended to eat the most abundant (as basal area) species although some food species were scarce in terms of stem density;

b) Mature leaves, which were the most abundant item throughout the study, contributed only a small proportion to the diet;

c) Fruits (GRP3 and GRP21), flowers (GRP3) and very young leaves (GRP3 and GRP21) of a few species were selected preferentially. Sweet, fleshy fruits of a few species were eaten when available, unlike most other colobines which tend to avoid such fruits;

d) Particular parts of items were eaten e.g. seeds only of some fruits, calyx of some flowers, petioles of some young leaves.

8. The GRP3 home-range area had less favoured food items available than the GRP21 home-range area. GRP3 spent more time feeding than GRP21. The proportion of flowers in the diet was greater for GRP3 and the proportion of fruit (unripe) less.

9. GRP3 made use of two abundant plantation species by feeding selectively on very young leaves of Swietenia macrophylla and on Chapter 6: Feeding behaviour •193

the mid-ribs of young T. grandis leaves. Young leaves of T. grand is, available as a source of food throughout the study, were a major staple food of GRP3 and were eaten when preferred foods were scarce. No single species acted as a staple food for GRP21.

10. Dietetic diversity (as item diversity and species diversity) was not negatively correlated with the proportion of preferred food items in the diet as has been reported in several other colobine studies. Chapter 7: Food selection in relation to phytochemistry 194

CHAPTER 7 FOOD SELECTION IN RELATION TO PHYTOCHEMISTRY

7.1 INTRODUCTION

Primates may be classified according to their digestive physiology. Some primates have simple, spherical stomachs (Parra 1978). Others have enlarged regions in the digestive system. In primates with caeco-colic fermentation (for example, Alouatta spp. (Napier and Napier 1967)), both the caecum and ascending colon are enlarged into a fermenting chamber (Bauchop 1978; Chivers and Hladik 1980). In colobines, fermentation of food occurs in the enlarged presaccus and saccus (Kuhn 1964; Bauchop and Martucci 1968; Moir 1968; Ohwaki et al. 1974); hence the common reference to colobines as forestomach or foregut fermenters (Section 6.2).

Food selection is related to the type of digestive system of the animal and the animal's body size (Section 6.2). It has been suggested that foregut fermentation probably evolved in areas where there was an abundance of food of low nutritive value (Hume and Warner 1980). Whilst this may apply for foregut fermenters of comparatively large body size such as the Bovidae and Macropodidae (Hume 1982) which are primarily herbivores, the association is less clear for the smaller sized Colobinae. Colobines are folivores/frugivores, differing between species in the relative proportions of fruits and leaves in the diet. While primates with simple stomachs may make use of low quality succulent fruits (Vellayan 1982), colobines obtain more nutrients from the slow and thorough digestion of higher quality fruit (Davies et al. 1988).

As the amount of fibre in the diet increases, food digestibility decreases as may the level of proteins and other nutrients (Janis 1976). With decreasing body size, the energy and protein requirements of an animal increase relative to body weight (Kleiber 1961) as do the energetic costs of locomotion (Schmidt-Nielsen 1972). Furthermore, relative gut capacity decreases with decreasing body size (Janis 1976; van Soest 1981) as does the Chapter 7: Food selection in relation to phytochemistry 195

volume of the gut involved in fermentation (Chivers and Hladik 1984). The minimum body weight for a foregut fermenter has been estimated at 5 kg (van Soest 1981). Selection for foods of high nutrient quality and low digestion-inhibitor content is therefore expected to be pronounced in foregut fermenters approaching the lower limit in body size for this digestive system. Where vegetation m the habitat is generally of low quality, selection for certain foods 1s expected to be particularly marked.

A good example of food selection in relation to plant chemistry is provided by a study of the closely related species Presbytis rubicunda and P. melalophos (Davies et al. 1988). These species are of similar body weight (6.0-6.5 kg) (Davies et al. 1988), approaching the lower limit for foregut fermentation (van Soest 1981), and their digestive tracts have similar proportions (Chivers and Hladik 1980). Both species selected foliage on the basis of nitrogen content (N) (although only a weak positive influence was reported for P. me lalophos) and digestibility (CDIG) and P. rubi cunda selected leaves low in fibre content. Seeds were probably selected largely for their fermentable carbohydrate content (Davies et al. 1988). Vegetation in the Sepilok study area where P. rubicunda was studied (Davies 1984) was of lower quality than the vegetation in the Kuala Lompat study area (Davies et al. 1988) where P. melalophos was studied (Bennett 1983) and, as expected, P. rubicunda was more selective in food choice than P. melalophos (Davies et al. 1988).

Food selection in relation to phytochemistry has been reported for several other colobine species (reviewed by Davies et al. 1988). Selection for high protein and/ or high digestibility has been noted for Colobus satanas in Douala-Edea (McKey 1978; McKey et al. 1981) and Trachypithecus johnii in Kakachi (Oates et al. 1980, Waterman and Choo 1981). Selection for plant parts low in fibre has been recorded for C. satanas (McKey 1978; McKey et al. 1981), C. badius tephrosceles in Kibale (Struhsaker 1975; Gartlan et al. 1980; Waterman and Choo 1981) and T. johnii (Oates et al. 1980). Chapter 7: Food selection in relation to phytochemistry 196

The relationship between food selection and tannin content is more confused. Selection for plant parts low in tannin content has been reported for C. satanas (McKey 1978; McKey et al. 1981) and C. guereza in Kibale (Oates 1977a; Oates et al. 1977) but no consistent influence of tannin concentration on food selection was observed for T. johnii (Oates et al. 1980), P. rubicunda and P. melalophos (Davies et al. 1988).

The aim of this chapter is to present data on the influence of phytochemistry on food selection by T. auratus in the secondary forest at Pangandaran and to examine whether food selection by this species followed similar patterns to food selection by other colobines.

7.2 METHODS

Analyses of phytochemical parameters in relation to food selection included fibre (ADF), total phenolics (TP), condensed tannin (CT), protein (PROT), digestibility (CDIG) and protein precipitation (PP). Plant samples collected for analysis were chosen on the basis that they looked identical to food items which were eaten from the same plants. Sample preparation and chemical methods employed have been described in Section 3.2.3. Samples were analysed in duplicate. Due to the comparatively small number of samples analysed, data for the GRP3 and GRP21 home-range areas have been combined to consider general properties of food selection by T. auratus. A complete list of chemical levels in food items analysed is given in Appendix IV along with analytical results of items not eaten.

7. 3 PLANT CHEMICALS INFLUENCING FOOD SELECTION BY T. AURATUS

Levels of all six chemical measures in fruits (eaten and uneaten) and leaves (eaten and uneaten) were compared with Mann-Whitney U tests. None of the differences found were statistically significant, Chapter 7: Food selection in relation to phytochemistry 197

possibly reflecting small sample size. Discussion of results is therefore restricted to qualitative analysis (Fig. 7.1 ). CDIG tended to

16 -.r:: 14 Cl 12 Q) :: 10 8 ~ 6 "0 4

0~ 2 PP 0 16 .r::- 14 Cl 12 ::Q) 10 8 ....>- 6 "0 4

0~ 2 CT 0

14 -.r:: 12 Cl Q) 10 :: 8 ....>- 6 "0 4 ~ 2 0 TP 0

60 .r:: 50 .2' Q) 40 :: 30 ....>- "0 20

~0 10 CDIG 0

60 .r::- 50 .2' Q) 40 :: 30 ....>- "0 20

0~ 10 ADF 0

20 .r::- Cl Q):: 10 ~ • not eaten "0 D eaten

~0 PROT 0 LM LY FRUIT

Figure 7.1 Mean content (as percentage dry weight (+ SD)) of protein (PROT), fibre (ADF), condensed tannins (CT), total phenolics (TP); % digestibility (CDIG); and protein precipitating capacity (PP) in leaves and fruits available to Trachypithecus auratus (GRP3 and GRP21 combined) at Pangandaran LM = mature leaves; L Y ":' young leaves Chapter 7: Food selection in relation to phytochemistry 198

be greater in both fruits and leaves which were eaten when compared with parts not eaten. ADF was less for fruits and leaves eaten, and CT, TP and PP capacity tended to be higher in fruits eaten and lower in leaves which were eaten. Protein content was higher in fruits and ·leaves which were eaten, although only marginally greater (1 %) for leaves eaten when compared with leaves not eaten. Leaves tended to have higher levels of protein, to be more digestible and have less ADF than fruits while CT, TP and PP were present in higher quantities in fruits eaten than in leaves which were eaten (Fig. 7.1).

In summary, it appears that leaves may have been selected for lower fibre content and, to a lesser degree, greater digestibility. Condensed tannins and total phenolics may also have influenced fruit and leaf selection in that they were higher in fruits eaten and lower in leaves eaten when compared with parts not eaten. Analyses of flowers were too few to permit comparisons between flowers eaten and not eaten. However, chemical composition of flowers and their influence on food selection is generally regarded as being similar to that of leaves (Davies et al. 1988).

7.4 DISCUSSION

7 .4.1 Feeding Strategies - Comparison between T. auratus and Other Colobines

The qualitative analyses discussed m Section 7.3 are necessarily tentative due to the small sample size but the results raise a few interesting points. Davies ct al. (1988) found that P. rubicunda selected leaves on the basis of high nitrogen, high CDIG and low ADF content and circumstantial evidence suggested that P. melalophos selected leaves with higher nitrogen and CDIG (although ADF content was similar for leaves eaten and not eaten). Fruit selection was not related to any of the chemical measures tested; however, fruits differed from young leaves in having less protein, lower levels of ADF and being more digestible, from which it was inferred Chapter 7: Food selection in relation to phytochemistry 199

that fruits may have been selected for their storage carbohydrate content (Davies et al. 1988).

This raises the question as to why leaves as food items were not strongly selected by T. auratus on the basis of protein content. Two possibilities are forthcoming:

1. Protein levels in the foliage at Pangandaran were sufficiently high for T. auratus' requirements that selection for protein was superfluous;

2. Leaves were selected for some nutrient not examined m this study.

Support for the first possibility comes from the protein/ ADF ratio of mature leaves, described in Section 3.5.2. This ratio is considered indicative of the role of plant chemistry in influencing food selection by colobines (McKey 1978; McKey et al. 1981; McKey and Waterman 1987; Davies et al. 1988), ultimately influencing colobine biomass (Waterman et al. 1988). The ratio is high at Pangandaran, being the second highest recorded (Table 3.8, Section 3.5.2). The highest value was for Kibale where C. badius and C. guereza were studied (Oates 1977a; Oates et al. 1977; Gartlan et al. 1980; Waterman and Chao 1981). By selecting leaves lower m fibre (and higher in digestibility), protein may be effectively available to T. auratus in adequate amounts over time.

If "acceptable" leaves are comparatively abundant at Pangandaran why were T. auratus not more folivorous? Leaf parts comprised approximately 72% of the diet for C. badius (Struhsaker 1975) and 74% of the diet for C. guereza (Oates 1977a) at Kibale, while approximately 48% of the diet for T. auratus (GRP21 - Sections 6.4.2 and 6.5) and 33% of the diet for P. me lalophos (Bennett 1983) was contributed to by leaf parts. The rank order of the proportion of leaves in the diet for these species corresponds to the rank order of protein/ ADF ratios for the vegetation in the respective study areas (Table 3.8, Section 3.5.2). This ratio may Chapter 7: Food selection in relation to phytochemistry 200

therefore influence the proportion of foliage in a colobine diet. However, it is unlikely to be the sole determinant.

Colobus badius have a greater body weight (Kingdon 1971) than T. auratus (Section 1.2.2) and T. auratus weigh more than P. melalophos (Bennett 1983). With decreasing body size there is an increase in food quality required (Kleiber 1961, Section 7.1) and this may place a limit on the amount of foliage in the diet, particularly with respect to the amount of fibre which can be ingested. Larger body size permits a slower passage rate (Parra 1978; van Soest 1981) and this may result in more energy being obtained from the structural carbohydrates in leaves for the Colobus spp. The smaller bodied P. melalophos and the intermediate T. auratus may, therefore, need to obtain energy requirements from other plant parts as well.

Presbytis melalophos had a higher fruit intake (46%, Bennett 1983) than T. auratus (37%, GRP21, Section 6.4.2). Interestingly, while the proportion of leaves in the diet for GRP21 and GRP3 was similar (48% cf. 46%, Section 6.4.2), GRP21, living in a more diverse environment (Table 6.17, Section 6.6), had a greater dietary intake of fruit (37% cf. 27%). Whilst mean levels of protein tended to be higher in vegetation of the GRP3 home-range area and ADF levels lower, the weighted protein/ ADF ratio was similar for vegetation in the two study sites (Section 3.5.2). Therefore, while the protein/ ADF ratio may provide an indicator of the acceptability of leaves in a habitat as potential food for a colobine, other factors also influence food selection.

Support for the possibility that leaves were selected for a nutrient not examined in this study is obtained from an analysis of chemical measures in Tectona grandis leaves (Table 7.1). Blades of young leaves and mature leaves and midribs of young leaves and mature leaves were analysed separately. The proportion of time spent feeding on these parts is also given in Table 7 .1. Midribs of young T. grandis leaves was the plant part eaten most by GRP3 (Section 6.4.3) but the large difference in feeding time on this part, when compared with the other parts of T. grandis leaves, is not Chapter 7: Food selection in relation to phy.tochemistry •201

accounted for by any of the chemical measures available (Table 7 .1). If the protein/ ADF ratio strictly governed leaf selection, a greater proportion of young leaf blades in the diet would have been expected as midribs had considerably less nitrogen and more ADF than young leaf blades. If leaf selection was primarily based on digestibility, the proportion of mature leaf midribs in the diet would have been greater. As discussed in Section 6.4.3, feeding on midribs of young teak leaves was regarded as highly selective as mature leaves were more abundant than young leaves and manipulation of food was required to extract the midrib from the rest of the leaf.

Mature T. grandis leaves may have presented a physical limitation to T. auratus as food items as the blades had a coarse texture and midribs were woody in appearance. In a study of herbivory and defensive properties of tree species in a lowland tropical forest on Barro Colorado Island, leaf toughness was the variable most highly correlated with levels of herbivory (Coley 1983). However, this does not explain the preference of midribs over blades of young teak leaves as the latter were flacid and noticeably less tough than the midribs. Possibly, therefore, selection of leaves was governed by a mineral or nutrient (for example, carbohydrate content) not considered in this study.

Table 7.1 Chemical Measuresa for Tectona grandis Leaves and Contribution to Diet (of Different Parts)

N ADF Prot/ADF CDIG CT TP pp % feeding time

LY - blades 2.9 35.9 0.505 52.4 0.75 6.31 1.32 1.33 LM - blades 2.8 46.7 0.375 48.5 0.08 1.40 1.31 0.04 midribs - LY 1.6 42.7 0.227 56.4 0.04 0.75 0.39 12.80 midribs - LM 1.3 44.4 0.176 59.4 0.03 0.09 0.26 0.26

aAbbreviations for chemical measures are given rn Figure 7.1 L Y = young leaves LM = mature leaves Chapter 7: Food selection in relation to phytochemistry 202

Fruits eaten by T. auratus were lower in ADF and higher in protein and CDIG than fruits which were not eaten, following the same pattern already described for leaf selection. However, these factors are not considered to be major components in the selection of fruit as food items as CDIG and protein were lower, and ADF higher, in fruits eaten when compared with leaves eaten. These findings are different to those reported for P. melalophos and P. rubicunda (Davies et al. 1988) but the conclusion that fruit selection was not related to the chemical measures tested, remams the same (with the possible exception of TP and CT).

7.4.2 Role of Tannins and Phenolics in Food Selection

The role of tannins and phenolics in food selection by animals IS still a source of debate. Some species, for example, C. guereza (Oates et al. 1977), C. satanas (McKey et al. 1981) and P. melalophos (Bennett 1983 ), select foods low in tannin concentration. TP and CT levels were low in the major dietary items for T. johnii but were not regarded as a major feeding deterrent as both were present in considerable quantities in minor food items (Oates et al. 1980). Presbytis rubicunda preferred seeds with comparatively high tannin content (Davies et al. 1988) and T. auratus tended to select fruits higher in CT and TP content although CT and TP levels were, on average, lower in leaves eaten than in leaves not eaten (Section 7.3).

Cork and Pahl (1984) suggested that this inconsistency may be due to the effect of toxic compounds and/or non-compositional factors overriding the balance between nitrogen, fibre and tannin content in some instances. Cork and Pahl (1984) also indicated that methods of chemical analysis, traditionally used in analysis of diet quality for grazing herbivores, may be inadequate when applied to forest foliage and folivores. The latter seems unlikely to account for differences described in food selection by colobines in relation to tannin and phenolic content as similar discrepancies have been noted for grazing herbivores. Reed and Soller (1987) found that phenolics in Acacia cyanophylla had a negative effect on nitrogen Chapter 7: Food selection in relation to phytochemistry 203

utilization in sheep in that they caused low digestibility, weight loss and negative nitrogen retention. Conversely, phenolics in another Acacia sp. (A. seyal) resulted in an increase in microbial utilization of endogenous nitrogen in the rumen (Reed and Soller 1987). Moreno-Black and Bent (1982) emphasised the importance of spatial, geographical and temporal variables on the production of secondary compounds and concluded that "given the unpredictability of secondary compounds in the environment ... different patterns of adaptation from not only one primate species to another, but also between geographical localities of species" should be expected (Moreno-Black and Bent 1982, p. 33).

CT content, and to a lesser degree TP content, were negatively correlated with digestibility in two study areas in Africa (Waterman et al. 1980). These variables were not correlated in plant samples collected from Pangandaran (Section 3.5.1). Martin and Martin (1982) concluded from their analyses of six oak secies that there was no correlation between PP capacity and TP content. Vegetation at Pangandaran showed a positive correlation between PP capacity and both TP and CT content (Section 3.5.1). These findings further emphasise the structural heterogeneity of phenolics and tannins.

Low levels of tannins may enhance protein digestibility (Davies et al. 1988). Several possibilities have been suggested: tannin­ protein precipitates may protect protein from microbial degradation in the forestomach (Jones and Mangan 1977), the precipitates being broken down in a later stage of digestion (Mole and Waterman 1987) rendering dietary protein available to the primate; tannins may denature proteins thereby enhancing proteolysis by digestive enzymes (Mole and Waterman 1987); tannins may reduce the risk of acidosis by slowing down the rate of fermentation (Goltenboth 1976; Davies et al. 1988). The latter may explain why CT and TP (and PP) tended to be higher in fruits eaten by T. auratus than in fruits which were not eaten (Fig. 7.1). Ripe fruits are generally avoided by colobines (Section 6.9) because a rapid rate of fermentation may cause hyperacidity of the forestomach fluid (Goltenboth 1976; Section 6.9). Ripe fruits formed a substantial proportion of the T. auratus diet (Section 6.4.2) and, possibly, Chapter 7: Food selection in relation to phytochemistry 204

tannins played an important role in slowing down the rate of fermentation when eating fruits.

In a review of the effects of tannins on food selection by mammalian herbivores, Mole and Waterman (1987) noted that there was little evidence to support the claim that tannins effectively reduce digestibility. Tannins (in high quantities) are more likely to have adverse consequences in precipitation of endogenous protein but it is not clear whether the astringent taste of tannins may act as a negative feeding cue to avoid such consequences (Mole and Waterman 1987).

7.4.3 Optimization

Optimal foraging, in which food is selected to maximize some nutrient intake within limitations imposed by constraints has been a focus of studies dealing with foraging ecology in the past twenty years (see Chapter 1 for review). Gray (1987) stated that optimization will only occur where food is limited. In a complex environment such as a tropical forest it is difficult to ascertain whether food is a limited resource (Cant 1980). Coelho et al. (1976) implied that nutritional availability is not a limiting factor in tropical forests. This supposition was based on results obtained from a socio-bioenergetic analysis of the amount of energy and food resources required by a population of two sympatric species (Alouatta villosa (howler monkey) and Ateles geoffroyi (spider monkey)) in Tikal, Guatemala (Coelho et al. 1976). However, two important features were not considered in the Coelho et al. (1976) study. Firstly, seasonal toxicity of plant foods may restrict the availability of potential diets (Glander 1978) and secondly, cropping rates need to be considered. An extensive critique of the Coelho et a l. (197 6) study is provided by Cant (1980).

Hladik (1977), in a study ofT. senex, found that potential food supply was more than ten times what was actually eaten and that variations between different groups were small. Hladik (1977) proposed that where young leaves formed a large part of the diet, Chapter 7: Food selection in relation to phrtochemistry •205

eating one-tenth of the food production might be the maximum permissible without causing over-cropping and endangering the plant food supply. In studies of colobine ecology the nutrient which . is maximised has generally been· considered to be protein and the constraints, digestion inhibitors such as fibre and tannins.

The relationship between condensed tannins and phenolics and food selection in colobines is still unclear (Section 7 .4.2). The strongest association is between food selection and low levels of fibre (for example, C. satanas (McKey 1978; McKey et al. 1981); C. badius tephrosceles (Struhsaker 1975; Gartlan et al. 1980; Waterman and Chao 1981); T. johnii (Oates et al. 1980); P. rubicunda (Davies et al. 1988); and T. auratus (this study)). Fibre is therefore the digestion inhibitor which most consistently has a negative influence on food selection by colobines.

Evidence in support of protein maximisation in food selection by colobines is scant. Colobus satanas ate both foliage and seeds which had a high protein content (McKey 1978; McKey et al. 1981) and P. rubicunda ate young leaves which had higher protein levels (Davies et al. 1988). Protein had a weak positive influence on foliage selection by T. johnii (Waterman and Choo 1981) and P. melalophos (Davies et al. 1988). Other studies of colobine ecology, including the study of T. auratus reported here, have not found a significant correlation between foliage selection and protein levels.

Primate spec1es which are largely frugivorous may feed on young leaves and shoots to increase the dietary intake of protein; however, Hladik ( 1977) stated that protein content is less likely to strongly influence food selection in species whose staple food (leaves) is protein rich. Colobus satanas, P. rubicunda and P. melalophos had the lowest proportion of foliage intake in the diet of all colobines studied to date, except P. thomasi for which phytochemical data are not available (Table 6.18, Section 6.8). This may explain why these three species selected foliage with higher protein content. In the absence of data on dietary protein requirements for colobine species it is not known whether general Chapter 7: Food selection in relation to phytochemistry 206

levels of protein m the vegetation at some study sites (for example, Pangandaran) are sufficiently high (and ADF sufficiently low) such that selection for leaves high in protein is not warranted. However, an excess of protein in the diet may cause problems leading to disorders of the blood and liver (Wolter 1982) so it is likely that when vegetation contains high levels of protein (and low levels of digestion inhibitors) a level is reached beyond which food selection on the basis of increased protein content is unlikely to occur and would actually be disadvantageous to a colobine.

It seems likely, therefore, that a nutrient(s) other than protein may be selected in foliage intake by at least some colobines. Foliage contains many structural carbohydrates (Davies et al. 1988) which may provide an important energy source when broken down by cellulose-digesting bacteria in the stomach (Bauchop and Martucci 1968). When considering all dietary intake this certainly appears to be true. The proportion of fruit in the diet varies across species (see Table 6.18, Section 6.8) contributing as much as 53% to the C. satanas diet (McKey et al. 1981). Only for this species and T. auratus (Section 7 .3) was ingested fruit associated with a higher protein content, although for T. auratus the level of protein in fruits was still less than the mean protein level in leaves.

Fruit selection is generally thought to be associated with storage carbohydrate (such as starch) content (for example, Davies et al. 1988). As some fruits are usually selected by colobines when available, in preferrence to young leaves (Hladik 1977; Bennett 1983; Davies 1984; this study - Section 6.5) this may be an important nutrient in diet selection. Hladik ( 1977) found that P. entellus selected foods rich in soluble glucids. The possible influence of non-structural carbohydrate content on fruit selection by other colobines has not been tested. Where examined, no relationship between gross energy content and fruit selection has been found (for example, Bennett 1983; Davies 1984) but this is not surprising as gross energy content does not provide a measure of available energy to a colobine (see Section 3.2.3). There is a need for feeding trials of captive colobines to determine energy assimilated from different foods, and to relate this to feeding Chapter 7: Food selection in relation to phytochemistry 207

behaviour of colobines in the field. The effects of foregut fermentation on nitrogen metabolism also need to be examined (Davies et al. 1983; Watkins et al. 1985). Only in this way will food selection by colobines be more comprehensively understood.

7.5 SUMMARY

1. Samples of leaves and fruits exploited as food items by T. auratus were analysed for nitrogen content, ADF, CDIG, CT, TP and PP and compared with fruits and leaves not eaten.

2. Differences in chemical measures for items eaten and not eaten were not statistically significant (probably due to small sample size) but trends indicate that leaves were selected for their greater digestibility and lower fibre content. Fruits eaten also had a higher mean level of CDIG and lower mean level of ADF than fruits not eaten but these measures are not considered to be of major importance in fruit selection as CDIG was lower and ADF higher in fruits eaten when compared with leaves eaten. Levels of CT, TP and PP capacity were higher in fruits eaten than in fruits not eaten. The role of tannins and phenolics in food selection is discussed.

3. Leaves (and fruits) were not strongly selected on the basis of protein content. Approximately half the dietary intake of T. auratus was leaves, a protein-rich food source. Possibly, protein levels in foliage at Pangandaran were adequate such that selection for this nutrient was not required. A nutrient other than protein (for example, energy) may have been maximised through food selection.

4. The protein/ ADF ratio may provide an indicator of the acceptability of foliage in a habitat as potential food for a colobine. However, this ratio did not govern food selection by T. auratus at Pangandaran. Chapter 8: Home-range and ranging behaviour 208

CHAPTER 8 HOME-RANGE AREA AND RANGING BEHAVIOUR

8.1 INTRODUCTION

Where areas are clearly circumscribed by a group of animals this adherence to a particular area justifies the use of the term home­ range area (Davies 1984). Well-defined home-range areas have been described for several groups of colobines (e.g. Struhsaker 1975; Hladik 1977; Oates 1977a, Bennett 1983). The size of a primate group's home-range area, the distance travelled in a day and the differential use of areas within the home range may be influenced by factors such as diet and the distribution and abundance of food (e.g. Clutton-Brock 1975b, 1977b; Marsh 1981c; Curtin 1982; Whitten 1982), position of sleeping trees (e.g. Bennett 1983; Gittins 1983); interactions with neighbouring conspecific groups (e.g. Chivers 1969; Struhsaker 1974, 1975; Davies 1984) and weather conditions (e.g. Chivers 197 4; Raemaekers 1980; McKey and Waterman 1982). This chapter aims to determine the factors influencing the ranging behaviour of the Trachypithecus auratus study groups in the Pangandaran Nature Reserve.

A commonly accepted definition of a territory 1s a home-range area which is defended (Noble 1939; Burt 1943). Kaufmann (1983) introduced the concept of territoriality based on space-related dominance. Territoriality has been observed for some groups of colobines (e.g. Bernstein 1968; Struhsaker and Oates 1975 ;) and not for others (e.g. Poirier 1970; Bennett 1983). A further aim in this chapter is to determine whether the home-range areas of GRP21 and GRP3 constitute territories and to discuss the conditions under which defence of home-range areas would be advantageous and the conditions under which the maintenance of a territory is not required or is not cost-effective. Chapter 8: Home-range and ranging behaviou.r 209

8.2 METHODS

Locations of all monkeys visible at twenty minute intervals (Section 2.4.1) were plotted on 1:1,000 scale maps of the study sites during monthly scan samples for GRP3 and GRP21. Datapoints were digitalised and the centre of the group calculated from the mean of all co-ordinates at a given time. The sum of distances between consecutive group-centre points in a day determined the day-range length. Numbered quadrats (0.25 ha) were demarcated in the field and marked on maps of each study site (Section 2.3.2). The quadrat number and number of monkeys present in each quadrat were also recorded at twenty minute intervals.

Home-range area was measured in three ways:

1. The number of 0.25 ha quadrats entered. This method has been widely used in estimating home-range areas of primates (e.g. Clutton-Brock 1972; Struhsaker 1975; Oates 1977a; Marsh 1978b; Bennett 1983; Davies 1984; Caldecott 1986).

2. A map of each study site (see Fig. 3.1 and Fig. 3.2, Section 3.2.1) was overlaid with a 50x25 m2 grid and the area encompassing all locations of the group determined. This method Is essentially the same as the previous method but, being of greater resolution, Is more accurate than the former.

3. The Fourier transform method (Anderson 1982) was used to determine the minimum area probability which encompassed 50 per cent (MAP(50)) and 95 per cent (MAP(95)) of group locations. The MAP(50) estimate represents the core area and the MAP(95) effectively corresponds to the total home-range area.

Quadrat use was assessed by the proportion of time spent m each 50x50 m quadrat, where time is the number of twenty minute intervals at which the locations of T. auratus were plotted. This method of assessing home-range use has been widely employed in previous primate studies (e.g. Rowell 1966; Altmann and Altmann 1970; Clutton-Brock 1972; Struhsaker 1975; Marsh 1978b; Bennett Chapter 8: Home-range and ranging behaviour 210

1983; Davies 1984). Results of observations at a given time were weighted such that the contribution to the total dataset was always one. For example, if eight monkeys were observed in quadrat 60 and two monkeys in quadrat 58 at 0800 h this gave values of 0.8 and 0.2 for quadrats 60 and 58, respectively.

Possible correlations between quadrat use and frequency (number of trees) and biomass (basal area) of ten species, which were eaten most in each month, were examined by multiple regression analysis (Norusis 1985). Other possible correlations between ranging behaviour and diet and environmental variables were assessed by univariate analysis.

8.3 HOME-RANGE AREA

8.3.1 Estimates of Home-Range Area

GRP3 and GRP21 were observed in 23 and 35 0.25 ha quadrats, respectively. Assessment of the cumulative number of quadrats entered with time shows that the total area occupied by both groups had been recorded by the end of the study (Fig. 8.1). The graph levels out after approximately ten days for GRP21 and after six days for GRP3 indicating that home-range areas were small, relatively stable and probably intensively used over short periods of time. Small increases in the curves after the initial levelling out (Fig. 8.1) indicate quadrats in which the monkeys had not previously been recorded. However, this is unlikely to represent a shift in home range as the increment was small and the graph had plateaued rather than showing a sharp gradient with a continual increase in the number of quadrats entered with time.

Home-range area estimated from the number of 0.25 ha quadrats entered and multiplied by 0.25 was 5.75 ha and 8.75 ha for GRP3 and GRP21, respectively. Estimates derived in this way are likely to give an overestimate of actual area used (Struhsaker 1975) by a T. auratus group as they do not take into account lacunae within the home-range area (except where lacunae occupy Chapter 8: Home-range and ranging behaviour 211

a) GRP21

40

"C 30 ...(I) (I) -r:: (I) 20

1/) -...ca "C 10 ca ::s C" ca .c: 0 0 5 10 15 20 25 It) C'll c) Number of days -0 ... (I) .c b) GRP3 E 30 ::s r::

(I) > 20 -ca ::s E ::s 10 (.)

0 5 10 15 20 25 30 35 40 Number of days

Figure 8.1 Cumulative number of 0.25 ha quadrats entered by GRP21 and GRP3 throughout the study Chapter 8: Home-range and ranging behaviour 212

a whole 0.25 ha quadrat), and include quadrats on the periphery of the home-range area even when the area used within the quadrats ts small.

Home-range areas determined by placing a 50x25 m2 grid over each study site and including all grid squares where the study groups were observed were estimated at 4.5 ha and 6.5 ha for GRP3 and GRP21, respectively. This method provides a more realistic estimate of home-range area due to the error in overestimation caused by the former method which is considerable where home­ range areas are small. Consequently, 4.5 ha (GRP3) and 6.5 ha (GRP21) are the home-range estimates used in subsequent comparisons with other colobines. The former method is used when assessing quadrat use as the 0.25 ha quadrats, having been demarcated in the field, provided reference points for all locations of T. auratus and were entered into the computer database.

The MAP(95) yielded home-range area estimates of 2.18 ha and 3.23 ha for GRP3 (Fig. 8.2) and GRP21 (Fig. 8.3), respectively. These estimates are approximately half of the home-range area estimates obtained using the previous method. The Fourier transform method is undoubtedly the most accurate of the three methods tested as it is based on locations of monkeys and is therefore· independent of quadrat size. Furthermore, lacunae within each study site are taken into consideration even where lacunae are small (Fig. 8.4 and Fig. 8.5).

The MAP(50) estimates for core area were 0.50 ha and 0.96 ha for GRP3 and GRP21, respectively, indicating that comparatively small areas were used intensively by both groups. Core areas were dispersed throughout the home-range areas of both groups (Fig. 8.4 and Fig. 8.5).

8.3.2 Home-Range Overlap

Home-range overlap was calculated from the number of 0.25 ha quadrats in the study groups' home-range areas which were Figure 8.2 MA.P(95) utilization distribution for GRP3 (smoolbed by the Fourier transform method) (\ength of each side is '350 rn) 00

figure S.3 MAP(95) utilization distribution for GRP21 (smoothed by the Fourier transform method) (length of each side is 400 ro) Chapter 8: Home-range and ranging behaviour 215

0 100 200 JOO

300 JOO

200 200

100 100

0~------~------~------~------~------_. ______. ______~ 0 0 100 200 JOO

Figure 8.4 Two-dimensional representation of home-range use by GRP3 (Fourier transform method) 0.50 = MAP(50) 0.05 = MAP(95) (Axes indicate distance (m) from arbitrary origin) Chapter 8: Home-range and ranging behaviour 216

0 100 200 300

300 300 0

200 200

100 100

0 0

0 100 200 300

Figure 8.5 Two-dimensional representation of home-range use by GRP21 (Fourier transform method) 0.50 = MAP(50) 0.05 = MAP(95) (Axes indicate distance (m) from arbitrary origin) Chapter 8: Home-range and ranging behaviou~ . 217

entered by other T. auratus groups at some time during the study divided by the total number of 0.25 ha quadrats used by each of the study groups. Of the 23 quadrats used by GRP3, six (26%) were entered by neighbouring T. auratus groups; of the 35 quadrats used by GRP21, seven (20%) were used by other groups. These values are likely to be an overestimate of actual home range overlap as they do not consider partial use of quadrats by other groups. However, estimates of home range overlap determined in this way are useful in comparisons with other colobines where home range overlap has been calculated similarly (see Section 8.6.2.2).

8.3.3 Number of 0.25 ha Quadrats Entered Daily

Frequency distributions of the number of 0.25 ha quadrats entered each day by GRP21 and GRP3 are shown in Figure 8.6. The median number of quadrats entered each day by GRP3 was 9.5 which corresponds to an area of 2.4 ha (53% of the home-range area). The median number of quadrats entered each day by GRP21 was 10 which corresponds to an area of 2.5 ha (38% of the home­ range area). Thus, both groups used substantial proportions of their home-range areas in a day. The number of quadrats entered each day by GRP3 and GRP21, during months in which both groups were observed, did not differ significantly (U = 382.5, n1 = 25, n2 = 25, n.s.).

8.3.4 Number of 0.25 ha Quadrats Entered Monthly

The number of 0.25 ha quadrats entered per day did not differ greatly between months for either group (Fig. 8.7). The number of quadrats entered each month ranged between 15-20 for GRP3 and 21-27 for GRP21, reflecting the difference in total home-range area of the two groups. The mean number of quadrats entered in a month by GRP3 was 16.9 (SD = 1.7) which corresponds to 4.2 ha (94% of the home-range area). The mean number of quadrats entered in a month by GRP21 was 23.6 (SD = 1.96) which Chapter 8: Home-range and ranging behaviour 218

40 a) GRP21 (n =25)

30

Ill >o Ill "C 20 -0 '#.

10

0 5 6 7 8 9 10 11 12 13 14 15

Number of quadrats used each day

5 6 7 8 9 10 11 12 13 14 15 Number of quadrats used each day

Figure 8.6 Frequency distribution of the number of 0.25 ha quadrats entered each day by GRP21 and GRP3 throughout the study Chapter 8: Home-range and ranging behaviour 219

Figure 8.7 Monthly number of 0.25 ha quadrats entered each day by GRP21 and GRP3 (mean + SD)

corresponds to 5.9 ha (91% of the home-range area). Thus, most of the total home-range area of each group was entered each month.

8.4 DAY-RANGE LENGTH

8.4.1 Daily Range Use

Frequency distributions of day-range lengths of the two study groups are shown in Figure 8.8. Generally, both GRP21 and GRP3 Chapter 8: Home-range and ranging behaviour 220

0 300-399 400-499 500-599 600-699 700-799 Day-range length (m)

300-399 400-499 500-599 600-699 700-799 Day-range length (m)

Figure 8.8 Frequency distribution of day-range lengths of GRP21 and GRP3 travelled between 400-700 m in a day. During months in which both groups were observed, the median day-range length of GRP3 was 441.8 m, (IQ range = 408.3-530.6 m) and the median day-range length of GRP21 was 529.7 m (IQ range = 453.6-603.1 m). GRP21

had a significantly greater day-range length than GRP3 (U = 453, n1

= 25, n2 = 25, p < 0.01). Chapter 8: Home-range and ranging behaviou.r 221

8.4.2 Monthly Range Use

Day-range length of both GRP21 and GRP3 varied considerably between months and within months (Fig. 8.9). Day-range lengths did not vary in the same way between the two groups in each month during which observations of both groups were made. Analysis of environmental factors which may affect day-range length of each group are discussed below (Section 8.5).

700 a) GRP21 -E 650 .c::: 600 C'l -c:: 550 CD 500 CD C) c:: 450 ...('(I I >- 400 ('(I c 350 300 oov CEC JAN FEB MAR 1984 1985

700 b) GRP3 650 -E - 600 .c::: C) 550 -c:: CD 500 CD C) c:: 450 ('(I... I >- 400 ('(I c 350 300 JUL AUG SEP OCT oov CEC JAN FEB 1984 I 1985

Figure 8.9 Mean (+ SD) monthly day-range lengths (m) of GRP21 and GRP3 Chapter 8: Home-range and ranging behaviour 222

8.5 CORRELATES OF RANGING BEHAVIOUR AND QUADRAT USE

Due to the variability in day-range lengths within a month (Section 8.4.2) analysis of monthly means may not detect correlations which are present but which can only be detected if examined on a daily" basis. Therefore, subsequent analyses of ranging behaviour and quadrat use were performed on data collected daily unless otherwise specified.

8.5.1 Diet

8.5.1.1 Major Food Species

Results of multiple regression analysis between monthly quadrat use and the frequency and biomass of the ten species which were eaten most in each month, were inconclusive. For example, in the November sample for GRP3, quadrat use was significantly correlated with the biomass of Bridelia monoica and Acacia

auriculiformis (Adjusted r2 = 0.71, F2,8 = 13.37, p < 0.005). The regression was Y = 3.04 + O.lX - 0.05Z, where X = B. monoica and Z = A. auriculiformis. Partial correlations were 0.86 and -0.68 for B. monoica and A. auriculiformis, respectively. Therefore, there was a positive correlation between quadrat use and the biomass of B. monoica and a negative correlation between quadrat use and the biomass of A. auriculiformis, suggesting an avoidance of the latter species. Bridelia monoica accounted for 0.1% of the diet in this month and A. auriculiformis accounted for 3.9% of the diet in this month.

The species which was eaten most in November was Tectona grandis, which accounted for 31.4% of the diet (Section 6.5). However, a correlation between quadrat use in November and the biomass of T. grandis was not found. A clear relationship between quadrat use and biomass of species, as a factor relating to diet, was not apparent using multiple regression analysis. Likewise, species frequency did not adequately explain monthly quadrat use with respect to dietary intake. Chapter 8: Home-range and ranging behaviour 223

Possibly, factors other than diet may explain monthly quadrat use. However, frequency and biomass of food species as an influence on quadrat use can not be discounted on account of results obtained from multiple regression analysis for the following reasons:

1. Vegetation analysis was based on random sampling of 10x10 m plots throughout each study site (Section 3.2.1). Consequently, 0.25 ha quadrats were not sampled in equal proportions and only those quadrats represented by at least three vegetation plots were used in the above analyses. Therefore, while the vegetation in each study site was adequately described by the sampling method used (Section 3.2.1), this may not be true for each 0.25 ha quadrat.

2. All trees ~ 5 m in height were included in the vegetation sampling (Section 3 .2.1) irrespective of maturation stage. Consequently, an absence of a correlation between quadrat use and the biomass or frequency of a species may be caused by an abundance of trees of this species which were not bearing the reproductive parts selected by T. auratus, i.e. frequency or biomass of a species does not necessarily reflect abundance as a food source for T. auratus.

3. Abundance of fruits and flowers for spec1es which are asynchronous in production of reproductive parts will not be reflected in the total biomass and frequency of these species. Thus, food selection and quadrat use may be associated with specific trees m any one month rather than with the biomass or frequency of a species.

4. Selection by T. auratus for specific trees was observed while trees of the same species at a similar maturation stage which were also bearing the selected food item were ignored as a food source by T. auratus (Section 6.5). In this case, quadrat use may be influenced by the biomass of specific trees and, again, not be correlated with species biomass and frequency. Chapter 8: Home-range and ranging behaviour 224

5. Multiple regression analysis is most effective where variance is high i.e. where patches are variable in size; for example, where a species is present in all quadrats but in varying degrees of abundance. Where patches are large and uniform in distribution or where species are rare i.e. thinly distributed throughout the home­ range area, variance is low and multiple regression analysis may not be an effective analytical method. For example. T. grandis was found in large relatively uniform patches in the GRP3 home-range area (Section 3.3.1). A correlation between quadrat use and biomass of T. grandis was not found for November although T. grand is contributed a large proportion to the diet in this month (see above). However, the Fourier transform method (Anderson 1982) included areas in which T. grandis were found in calculation of the MAP(50) (Fig. 8.10). Therefore, (some) areas with T. grandis were extensively used in this month. A lack of correlation between quadrat use and biomass of T. grandis reflects either the uniform distribution of T. grandis and uniform use by GRP3 of quadrats in which T. grand is was found, or the extensive use of some quadrats containing T. grandis and the absence of use of other quadrats containing T. grandis.

Bennett (1983) similarly found that the amount of time spent m a quadrat by Presbytis melalophos was not directly related to the presence of important food species within it and concluded that this may have been because dietary diversity was quite high even though the diet was specialised. This may also apply to T. auratus at Pangandaran. By comparison, dietary diversity of Colobus badius rufomitratus was low (Marsh 1978b) and quadrat use was correlated with the position of major food sources (Marsh 1981 c).

8.5.1.2 Proportion of Different Items in Diet

Day-range lengths and the number of quadrats entered each day by GRP3 and GRP21 were tested on a univariate basis against the proportions (as percentage) of young leaves, fruits and flowers in the diet. Day-range length was significantly positively correlated with the proportion of fruit in the diet for both GRP3 and GRP21 Chapter 8: Home-range and ranging behaviqur 225

0 100 200 .300

300 JOO

200 200

100 100

0~------~------~------.______~------~------~ 0 0 100 200

Figure 8.10 Home-range use by GRP3 in November 1984 (Fourier transform method) Thick lines enclose MAP(50) areas and thin lines enclose MAP(95) areas X = areas with Tectona grandis (Axes indicate distance (m) from arbitrary origin)

(Table 8.1). A significant negative correlation was found between day-range length and the proportion of young leaves in the diet for GRP3, but not for GRP21 although there was a tendency for shorter day-range lengths as the proportion of young leaves in the diet increased as shown by the negative rs value (Table 8.1). Chapter 8: Home-range and ranging behaviour 226

Table 8.1 Possible Correlates of Day-Range Length and Number of Quadrats Entered Each Day by GRP3 (n = 40) and GRP21 (n = 25) (Spearman Rank) a) GRP3

day-range length no. of quadrats entered each day Variable rs p rs pa

% young leaves -0.3603 p < 0.05 -0.0689 n.s. in diet % fruit in diet 0.3160 p < 0.05 -0.1381 n.s. % flowers in diet 0.1540 n.s. 0.1472 n.s. b) GRP21

day-range length no. of quadrats entered each day Variable rs p rs pa

% young leaves -0.3292 n.s. -0.2682 n.s. in diet % fruit in diet 0.3991 p < 0.05 0.3423 n.s. % flowers in diet 0.2360 n.s. 0.0094 n.s. an.s. = not significant

The proportion of flowers in the diet was not correlated with day-range length. The number of quadrats entered was not correlated with any of the variables tested (Table 8.1).

A comparison was made between day-range lengths during months when the proportion of young leaves in the diet was comparatively high with months when the proportion of young leaves in the diet was comparatively low. As there was an inverse relationship between the proportion of fruit and the proportion of young leaves in the diet (Section 6.5), this is effectively the same as comparing months in which the proportion of fruit in the diet was low with months during which the proportion of fruit in the diet was high. The median day-range length of GRP3 ·during months when the proportion of young leaves in the diet was high ( 450.4 m) differed significantly from the median day-range length when young leaf intake was comparatively low (584.9 m) (U = 323, n1 = Chapter 8: Home-range and ranging behaviour 227

15, n2 = 25, p < 0.001). The median day-range length of GRP21 did not differ significantly between months of high and low young leaf intake (U = 104, n 1 = 10, n2 = 15, n.s.) but there was a tendency to travel shorter distances when the proportion of young leaves in the diet was high (463.3 m (median)) than during months when the proportion of young leaf intake was low (557. 7 m (median)).

8.5.2 Night Trees

The proportion of time (active period) spent in quadrats with night trees (i.e. trees in which the monkeys slept during the night) was compared with the proportion of time spent in quadrats which did not have sleeping sights therein. GRP3 spent significantly more time during the day in quadrats with night trees (U = 17 4, n 1 = 11, n 2 = 12, p < 0.001); a similar association between quadrat use and the presence/absence of night trees was hot found for GRP21 (U =

110, n1 = 8, n2 = 27, n.s.).

The proportion of time spent m quadrats with night trees was then tested against the number of nights spent in each of these quadrats. A significant correlation between these two variables was found for GRP3 (r8 = 0.78, n = 12, p < 0.01) but not for GRP21 (r8 = 0.27, n = 8, n.s.).

8.5.3 Rainfall

During fifteen days (37.5%) on which scan sampling of GRP3 was conducted, there was no rainfall. During other days, rainfall varied between 1-129 mm. Rainfall did not occur on sixteen days (64%) during which scan samples were conducted of GRP21 and varied between 1-42 mm during remaining days. Days were categorized according to whether rainfall was ~ 10 mm or > 10 mm. Day-range length and the number of quadrats entered were then compared between categories for each study group. Day-range length did not differ significantly between categories for GRP3 (U = 190.5, n 1 = 14, n2 = 26, n.s.) nor for GRP21 (U = 67, n1 = 6, n2 = 19, n.s.). Chapter 8: Home-range and ranging behaviour 228

Similarly, the number of quadrats entered each day did not differ between categories for GRP3 (U = 182.5, n1 = 14, n2 = 26, n.s.) nor for GRP21 (U = 70, n1 = 6, n2 = 19, n.s.).

8.5.4 Defence of Home-Range Area

If groups travel over large areas to obtain food these areas may be larger than they can defend (Crook 1972), and, on an economic basis, territoriality may not be a viable option (Brown 1964; Brown and Orians 1970; Mitani and Rodman 1979). Mitani and Rodman (1979)'s "index of defendability" (D) was used to determine whether the study groups were capable of defending their home-range areas.

D = ~ 4N1t) where d = day-range length and A = home-range area. Where the index is greater than one, animals are capable of defending areas.

Taking the shortest day-range length of GRP3 (384 m) and of GRP21 (335 m), the defendability index was calculated as 1.6 and 1.2 for GRP3 and GRP21, respectively. Therefore, both groups were capable of defending territories. The degree of overlap in home­ range area with neighbouring groups (Section 8.3.2) and the behaviour observed when neighbouring groups interacted (Section 4.4.1.5) indicate that both groups actively defended their home­ range areas. Interactions between GRP3 and neighbouring groups were observed on 16 days (40%) and between GRP21 and neighbouring groups on 8 days (32%).

Day-range lengths and the number of quadrats entered on days when intergroup encounters occurred were compared with day­ range lengths and the number of quadrats entered on days when intergroup encounters did not occur.

For GRP3, the median day-range length and the median number of quadrats entered did not differ between days in which

intergroup encounters did, and did not, occur (U = 218.5, n 1 = 16, Chapter 8: Home-range and ranging behaviour 229

n 2 = 24, n.s.; U = 200, n1 = 16, n2 = 24, n.s.; respectively). Similarly, day-range length and the number of quadrats entered by GRP21 did not differ significantly between days in which intergroup encounters did, and did not, occur· (U = 72, n1 = 8, n2 = 17, n.s.; U =

89, n1 = 8, n2 = 17, n.s.; respectively). Thus, GRP3 and GRP21 did not range further on days when they encountered neighbouring groups as might be expected if territorial defence is a major determinant in ranging behaviour. However, this possible association between ranging behaviour and territorial defence probably depends on the size of the home-range area. More than 90% of the home-range area was entered in a month by each of the study groups (Section 8.3.4) indicating that both home-range areas were small enough to be regularly occupied throughout, in a comparatively short period of time. Thus, in the course of traversing their home-range areas m response to other factors (food availability and location of night trees (GRP3 only)), territorial boundaries may also have been maintained concurrently.

Intergroup encounters were usually associated with a disputed food source on, or close to, a home-range boundary (Section 6.5). In such instances, one of the groups would withdraw from the feeding site; two or more groups of T. auratus were not observed feeding simultaneously from the same food source. Generally, the presence of another group within a study groups's home-range area was not tolerated and would lead to the displacement of the neighbouring group. Only on rare occasions was another group observed feeding within a study group's home-range area without being displaced. On such occasions the study group was at least 50 m distant (although presumably aware of the presence of the encroaching group) and the food source was in abundant supply. For example, a neighbouring group entered GRP3's home-range area and ate Kleinhovia hospita flowers. GRP3 had fed from the same trees earlier on the same day. This species is synchronous in production of flowers such that several large trees bore flowers at a given time (Section 6.5). Chapter 8: Home-range and ranging behaviour 230

8.5.5 Discussion of the Determinants of Ranging Behaviour by T. auratus

In summary, the main correlation between day-range length and the variables tested, was with the proportions of fruits and young leaves in the diet. On days when fruit intake was high, the monkeys travelled further. Generally, these were the days when they fed from large, widely-dispersed, and comparatively rare food sources (Section 6.5). Conversely, day-range lengths were generally shorter when young leaf intake was high. Young leaves were more abundant in the forest than fruit and, for GRP3, young leaves of T. grandis provided a constant food source (Section 6.5) distributed fairly evenly in large patches (Section 3.3.1).

The number of quadrats entered was not correlated with any of the variables tested. This is probably a reflection of the size of the quadrats, as was discussed in Section 8.3.1.

Location of night trees may have influenced ranging behaviour of GRP3. Generally, large emergents with open canopies were used as night trees by both groups. Whilst the frequency (number per hectare) of emergents was similar for both home-range areas (Section 3.3.1) there were fewer emergents in the GRP3 home-range area on account of the smaller area (Section 8.3.1). Therefore, particular trees were probably selected by GRP3 as sites from which to view the surrounding area. For example, one large tree in a central position within the GRP3 home-range area was used as a night-tree on 25% of occasions.

8.6 DISCUSSION

8.6.1 Comparison Between GRP3 and GRP21

All three methods used in estimating home-range s1ze showed that the GRP21 home-range area was larger than the GRP3 home­ range area (Section 8.3.1). The T. auratus groups were of similar size and age-sex class composition (Section 2.3.1) which raises the Chapter 8: Home-range and ranging behaviour 231

question as to why the GRP3 home-range area was approximately 30% smaller in size than that of the GRP21 home-range area.

The number of species per hectare in the two home-range areas was similar (145 spp ha- 1 and 147 spp ha- 1 in the GRP21 and GRP3 home-range areas, respectively (Section 3.3.2.1)). However, species diversity in the GRP21 home-range area was greater than species diversity in the GRP3 home-range area, while dietary diversity (as species) was similar for the two groups (Section 6.6). But the proportion of the diet made up by the top three and top five species was greater for GRP3 than GRP21. Three species contributed 31.5% and 24.9% to the GRP3 and GRP21 diets, respectively, and five species made up 44.9% and 38.4% of the diets, respectively. In other words, fewer species contributed major proportions to the GRP3 diet than to the GRP21 diet. This indicates that food sources for GRP21 were probably more thinly and widely distributed than food sources for GRP3 considering that major food source trees were frequently rare (as the number of stems per hectare) (Section 6.4.1), and available as a food source (i.e. largely as fruits or young leaves) for short periods of time only (Section 6.5).

It is argued that the smaller home-range area of GRP3 was correlated with the presence (and abundance) of T. grandis stands. Plantation species occupied approximately 45% of the GRP3 home­ range area (Section 3.3.1), of which T. grandis was exploited as a major food source in times of preferred food scarcity (Section 6.5). Young leaves of T. grand is were available throughout the study but only contributed major proportions to the GRP3 diet when fruits, flowers and very young leaves of a few preferred species were not available (Section 6.5). This corresponded with months during which young leaf intake was high (Section 6.5) and explains, at least partly, the negative correlation between day-range length and the proportion of young leaves in the diet.

To follow on, GRP21 ate more fruits than GRP3 (Section 6.4.2) which may be associated with the longer day-range length of GRP21 (Section 8.4.1 ), as fruits are generally rare and widely dispersed in smaller patches (Oates 1986). Furthermore, there was no staple Chapter 8: Home-range and ranging behaviour 232

young leaf source in large patches for GRP21. Therefore, GRP21 probably had to travel further than GRP3 in search of food.

8.6.2 Comparison With Other Colobines

8.6.2.1 Estimates of Home-Range Area

In addition to the number of 0.25 ha quadrats entered by a group of monkeys, another method commonly used in determining home-range areas of primates is the "taut string line" method (Altmann and Altmann 1970) which is the area enclosed by a line placed around the outer limit of all areas entered. Measurement of core areas has been taken to include areas used by a group of primates for more than four observation hours (e.g. Davies 1984) or the area enclosed by a line around the region of most common sightings (Caldecott 1986). As the position of the line or the number of hours assumed to be important are based on arbitrary decisions, these methods were not included in this study.

The method used for determining home-range size has a pronounced effect on the estimate obtained (Section 8.3.1). Therefore, there is a need to standardize the method for determining home-range area if direct comparisons are to be made between various studies. A recommendation is hereby made to use the Fourier transform method (Anderson 1982, Section 8.2, Section 8.3.1). This model provides a more precise estimate of the area used by a group of monkeys than other methods tested (Section 8.3.1) and, therefore, would be of value when making intraspecific and interspecific comparisons between home-range sizes of colobines. However, as forests are generally heterogeneous in composition and structure it is probable that pockets will exist within the general area used by a primate group, which are not entered. Therefore, home-range area determined by the Fourier transform method may not be the same as the estimate of home­ range area determined by conducting transect censuses, for example, as the latter would include pockets not entered by the study animals. Chapter 8: Home-range and ranging behaviol!-r 233

8.6.2.2 Correlates of Ranging Behaviour, Home-Range Size and Territoriality

The environmental variable ·most commonly cited as being correlated with primate ranging behavior is the distribution and abundance of food (e.g. Clutton-Brock 1975b; Marsh 1981c; Curtin 1982; Whitten 1982; Bennett 1983; Davies 1984; Johns 1986; Oates 1986). Food availability may vary on a short-term basis (e.g. seasonally) or alter over a longer period of time. Consequently, ranging behaviour of a population may ·vary with time and populations of the same species occurring in different habitats may show variation in ranging behaviour in response to differences m food availability.

Several colobines (e.g. C. badius tephrosceles in Gombe (Clutton­ Brock 1975b); P. melalophos at Kuala Lompat (Bennett 1983); P. rubicunda in Sepilok (Davies 1984); T. auratus (this study)) travel further when favoured foods are abundant, while others (e.g. P. entellus in the Nepal Himalaya (Curtin 1982) and C. satanas at Douala-Edea (McKey and Waterman 1982)) travel less far when preferred foods are available. This is probably related to the abundance of food during times of preferred food scarcity. Few species were used as food items by C. satanas at Douala-Edea due to the high levels of digestion inhibitors and low levels of protein in the leaves of common species (Gartlan et al. 1980). Consequently, when preferred foods were not available these colobines had to travel further to obtain acceptable foods (McKey and Waterman 1982).

Species which feed extensively on fruits tend to have larger home-range areas than closely related species which are more folivorous. For example, P. rubicunda (Davies 1984; Supriatna et al. 1986) and P. melalophos (Bennett 1983) are frugivorous by colobine standards (Section 6.8) and have larger home-range areas and greater day-range lengths than the more folivorous T. cristata in Sumatra (Wilson and Wilson 1976) and Kuala Selangor, Malaysia (Bernstein 1968), and T. auratus in the Pangandaran Nature Reserve (this study). Chapter 8: Home-range and ranging behaviour 234

Territorial behaviour has also been associated with the seasonal availability and abundance of food as well as to tree species diversity (Struhsaker and Oates 1975; Struhsaker and Leland 1979). Considerable variation within a species has been found with respect to territoriality; for example, C. guereza and P. melalophos have exclusive territories in some areas and extensive overlap in home­ range areas in others (Dunbar and Dunbar 1974; Oates 1977a; Bennett 1983; Johns 1983, 1986). Generally, territoriality has been observed in seasonal forests with comparatively low diversity; that is, in areas where food is probably not abundant (Marsh 1981c; Bennett 1983). For example, P. rubicunda at Sepilok generally displayed territorial behaviour although the defendability index (Mitani and Rodman 1979) was less than one and the study group was not always successful in displacing another group from a food source within the former group's home...:range area (Davies 1984). Food sources at Sepilok were subject to seasonal availability and were scarce due to the abundance of Dipterocarpaceae trees (Davies 1984, Davies et al. 1988, Section 3.3.3). Home-range areas were large (85 .5 ha) and the percentage overlap of home-range area with neighbouring groups was 10% (Davies 1984).

By comparison, P. melalophos at Kuala Lompat were not territorial although home-range areas could have been defended as the defendability index was greater than one (Bennett 1983). Food was comparatively abundant at Kuala Lompat (Davies et al. 1988) and home-range areas comparatively small (29 ha). Percentage overlap of home-range area with neighbouring groups was 79% (Bennett 1983). Although preferred foods were also seasonally available at Kuala Lompat, these favoured foods were from very large trees. Thus, the food supply of a particular tree was limited by the tree no longer producing the favoured food item rather than by the monkeys depleting the food source. Therefore, the cost of sharing the food sources was probably minimal (Bennett 1983).

The number of tree species per hectare at Pangandaran ( 145 spp ha- 1 (GRP21) (Section 3.3.2.1)) was similar to that at Kuala Lompat (158.5 spp ha- 1 (Bennett 1983)) and, likewise, most preferred foods were available for short periods of time only. From this it may be Chapter 8: Home-range and ranging behaviour 235

expected that T. auratus are not territorial. As discussed previously, this was not so. Territoriality may be associated with the adult male's ability to monitor and defend adult females from other adult males (Bennett 1983). While trees exploited as food by T. auratus were often very large such that all group members fed simultaneously in the one tree this was not always the case. When the group was scattered in smaller trees it may have been more difficult for the adult male to monitor and defend the group. Therefore, T. auratus males may have defended a harem by defending a territory.

Territorial behaviour may also occur where food sources are evenly spaced (Crook 1970), where the renewal rate of food items Is high enough to sustain constant harvesting (Struhsaker and Leland 1979) and where food sources are predictably available within a defendable area (Goss-Custard et al. 1972). All three factors characterize the GRP3 home-range area, with regard to the availability of T. grand is, and may account for the territorial behaviour displayed by GRP3.

Furthermore, home-range size of primates decreases as the amount of food in an area increases (Struhsaker 1967; Freeland 1979) up to a point where the home-range area reaches the minimum size necessary to contain sufficient food to sustain a group at all times. Home-range areas of T. auratus at Pangandaran were considerably smaller than home-range areas of most other colobines probably reflecting an abundance of food for this species in this study area. If the home-range areas of both GRP21 and GRP3 were approaching the minimum size required to ensure adequate food supplies at all times, territorial defence would be expected.

The distribution of night trees has been shown to influence the patterns of home-range use by primates (Bennett 1983; Gittins 1983; Davies 1984). For example, quadrat use by P. melalophos at Kuala Lompat was positively correlated with the. number of sleeping sites therein from which it was inferred that the relatively inaccessible and open crowns of night trees may have acted as a defence against predators and/or as a suitable site from which to Chapter 8: Home-range and ranging behaviour 236

produce loud calls (Bennett 1983). Quadrat use by GRP3 (but not for GRP21) was influenced by the location of night trees (Section 8.5.2). However, as night trees were frequently also exploited as food sources, selection for quadrats on the basis of night trees alone, is unlikely. Generally, sleeping sites of rain forest primates are not rare enough to affect daily quadrat use (Raemaekers 1977).

Rainfall did not have an effect on ranging behaviour of T. auratus at Pangandaran. This is similar to findings reported by Struhsaker (1975), Whitten (1982), Bennett (1983) and Davies (1984) in studies of ranging behaviour of primates, but is the opposite to findings reported by Chivers (1974), Raemaekers (1980), McKey and Waterman (1982) and Isbell (1983). However, where rainfall did influence ranging behaviour, the effect was less than the effect of food source distribution (Raemaekers 1980; McKey and Waterman 1982; Isbell 1983). In the long-term, rainfall is likely to influence ranging behaviour in its effect on flower and fruit production (Medway 1972; Whitmore 1975).

8.7 SUMMARY

1. GRP21 had a larger home-range area than GRP3 (6.5 ha and 4.5 ha, respectively). It is likely that the constant availability of T. grandis as a source of food in times of preferred food scarcity enabled GRP3 to have a smaller home-range area.

2. GRP21 travelled further in a day than GRP3. This is probably because food sources were more widely distributed in the GRP21 home-range area.

3. The distribution and abundance of food influenced ranging behaviour. Both groups travelled further (i.e. had greater day­ range lengths) when fruit intake was high. There was a significant negative correlation between day-range length and the proportion of young leaves in the diet for GRP3 (and a similar trend for GRP21, although not significant). Chapter 8: Home-range and ranging behaviour. . 237

4. The location of night trees was correlated with quadrat use by GRP3, but not by GRP21. However, as these trees were also exploited as major food sources by GRP3, selection for quadrats with night trees is unlikely to have been primarily due to their suitability as night-time sleeping sites.

5. Rainfall was not correlated with day-range length or quadrat use.

6. Both groups entered more than 90% of their home-range areas m a month. Defendability indices (Mitani and Rodman 1979) were greater than one and home-range areas were actively defended from neighbouring groups. Each group had exclusive use of most of its home-range area. Home-range overlap with neighbouring groups was estimated at 20% and 26% for GRP21 and GRP3, respectively. Both home-range areas were, therefore, regarded as territories.

7. Home-range areas of GRP3 and GRP21 were small by colobine standards. Possibly, the home-range areas were approaching the minimum size required to ensure adequate food supplies at all times; hence, the maintenance of territories.

8. The use of the Fourier transform method (Anderson 1982) is recommended for determining home-range areas and core areas of primates.

9. Multiple regression analysis was not a useful tool for explaining quadrat use with respect to biomass and frequency of major food items. This may be because there were too many confounding variables or because this method is most appropriate where variance in variables considered is high. Chapter 9: Concluding discussion 238

CHAPTER 9 CONCLUDING DISCUSSION

This was the first ecological study of a colobine spec1es m Indonesia to incorporate an analysis of plant chemistry and its influence on food selection. Of the two genera of leaf monkeys found in Asia, the ecology of Presbytis species is better understood than the ecology of the more folivorous Trachypithecus species (Section 6.1). Previous in depth studies of Trachypithecus species were conducted by Oates et al. (1980) (T. johnii) and Hladik (1977) (T. senex). Therefore, the results of this study provide valuable information for further comparisons between the behavioural ecology of the two genera of Asian leaf monkeys.

The breccia/limestone formation of the peninsula at Pangandaran, on which the Pangandaran Nature Reserve 1s located, provides an alkaline soil rich in exchangeable cations for plant growth (Section 2.2.4.2). The vegetation, largely of secondary origin, is dominated by trees belonging to the family Moraceae (Section 3.3.2) which were exploited by Trachypithecus auratus as major sources of food (Section 6.4.1). Secondary forest tree species, by comparison with primary forest species where most studies of colobine ecology have been conducted, are frequently r-strategists and do not invest heavily in plant defence compounds (Janzen 1974; Lebreton 1982; Section 3.5.1). The climate at Pangandaran was not characterized by distinct wet/dry seasons (Section 2.2.3) and many tree species were asynchronous in production of fruits, flowers and flushes of young leaves (Section 6.5). All of these factors indicate that food for T. auratus was abundant in the Pangandaran Nature Reserve and probably explain the high density of T. auratus in this study area. However, the absence of large predators as a determining factor in population density must be acknowledged when comparisons are made with densities of colobines in other regions where predators may exert a major influence on primate density (i.e. predation pressure is likely to select for an increase in group s1ze for large, diurnal animals such as colobines (Struhsaker 1981)). Chapter 9: Concluding discussion 239

Primarily folivorous, T. auratus also ate large amounts of fruit (Section 6.4.2). Like other colobines, they were not indiscrimant feeders, preferentially selecting the fruits and very young leaves of a few species when these were available (Section 6.5). GRP3 also preferentially selected flowers of a few species (Section 6.5). Fruits and leaves which were eaten were more digestible and lower in fibre content than fruits and leaves which were not eaten. Foliage selection by T. auratus was not correlated with protein content (Section 7 .3).

A few colobines select foliage for their comparatively high protein content (e.g. Colobus satanas (McKey 1978; McKey et al. 1981); Presbytis rubicunda (Davies et al. 1988)) and/or low fibre content (C. satanas (McKey 1978; McKey et al. 1981); C. badius tephrosceles (Struhsaker 1975; Gartlan et al. 1980; Waterman and Choo 1981); T. johnii (Oates et al. 1980); P. rubicunda (Davies et al. 1988)). From these observations, the protein/ ADF ratio of foliage has been considered a major determining factor in the carrying capacity and, hence, colobine biomass for a study area (Waterman et al. 1988). It has been argued that this ratio reflects the acceptability of mature leaves as a food source when preferred foods are seasonally unavailable or when a food shortage occurs as, for example, might eventuate during an extremely dry period. Thus, even though mature leaves generally form a relatively small component of the diet in colobines, where the protein/ ADF ratio in mature leaves is high, mature leaves may be eaten if this becomes necesssary and, consequently, colobine biomass is greater in areas where this ratio is high (Waterman et al. 1988).

Data obtained on the vegetation at Pangandaran, the feeding behaviour of T. auratus, and the density estimates of T. auratus at Pangandaran show that this explanation does not account for the difference in biomass of T. auratus in the TW and CA in the Pangandaran Nature Reserve, nor for the difference m colobine biomass in the CA compared with colobine biomass in other study areas (Fig. 9.1). As chemical analyses in this study were not conducted in the same laboratory as analyses of several other colobine studies (Section 3.2.3) the latter discrepancy may, in part, Chapter 9: Concluding discussion 240

2000 Kibale m N. E .lili: TWm 01 .lili:

-Ill Ill C'G 1000 E Kuala Lompat m .2 CA m .Q CD Kakachi m c: ~ .2 Douala-Edea 0 m u Sepilok m 0 0.00 0.10 0.20 0.30 0.40 0.50 Proteln/ADF (based on weighted means) Figure 9.1 Weighted mean protein/fibre levels in mature foliage against colobine biomass for six study areas (Adapted from Waterman et al. 1988)

reflect differences in analytical procedures. However, this does not explain the difference in T. auratus biomass in the CA and TW. Chemical analyses of vegetation in the GRP3 (TW group) and GRP21 (CA group) home-range areas were standardized in the same laboratory (CSIRO - Canberra) and the weighted mean protein/ ADF ratio was very similar (0.415 and 0.414, respectively). Yet, T. auratus biomass in the TW was twice that in the CA (Section 4.3.4).

Furthermore, selection for specific parts of leaves was not directly related to the protein/ADF ratio (Section 7.4.1). Clearly, factors other than the protein/ ADF ratio influence the acceptability of vegetation as food items.

Trachypithecus auratus at Pangandaran were not foraging to maximise protein intake. Possibly, they were maximising another nutrient such as energy in their dietary intake. Bauchop and Martucci ( 1968) found that energy provided by VFA production was 283 kcal day- 1. As the maintenance energy requirement for the 4.5 kg T. cristata used in this experiment was 218 kcal day- 1, the conclusion reached was that structural carbohydrates, which produce the VF As when broken down by the cellulose-digesting bacteria, are a major factor in energy metabolism for this colobine. Chapter 9: Concluding discussion '241

The experimental animal was fed a largely folivorous diet and foliage is known to be high in structural carbohydrates (Bauchop and Martucci 1968). In the field, where the foliage component in the diet may be less it is not known whether the foliage consumed would alone meet the daily energy requirement, but clearly the contribution of foliage to the energy requirement of colobines cannot be overlooked. That storage carbohydrates or fats may be an important component in fruit selection by colobines is a common belief which awaits verification.

The availability and distribution of large food patches may account for some of the variation in ranging behaviour of GRP3 and GRP21 (Section 8.6.1). Large patches of food might permit the formation of larger groups that occupy relatively small home-range areas (Strier 1987). The comparatively large patches of Tectona grandis, an important food source for GRP3 when preferred foods were scarce, may explain the smaller home-range area of GRP3 (Section 8.6.1). The larger size of groups associated with plantation areas in the TW compared with group size in the CA, and the smaller mean home-range area of groups in the TW (cf. mean home-range area of groups in the CA (Section 4.3.4)), may similarly be related to the constant availability of young T. grandis leaves in relatively large patches.

GRP21 travelled further in a day (Section 8.4.1), had a higher proportion of fruit in the diet (Section 6.4.2) and spent more time travelling and resting (Section 5.3.1) than GRP3. GRP3 spent more time feeding than GRP21 (Section 5.3.1). Day-range length was positively correlated with the proportion of fruit in the diet for each group and negatively correlated with the proportion of leaves in the diet of GRP3 (Section 8.5.1.2). There was an inverse relationship between the proportion of fruit and the proportion of young leaves in the diet of each group (Section 6.5) and fruits of a few species were highly selected when available (Section 6.5). These comparisons between the two groups indicate that . T. grandis leaves were not a highly preferred food item but that this tree species influenced the foraging behaviour of T. auratus by acting as a Chapter 9: Concluding discussion 242

"stand-by" food source when preferred food items were scarce or not available.

The results of this study are of potential importance for management of T. auratus. In the management plan for the Pangandaran Nature Reserve (Blower et al. 1977) a recommendation was made · to remove the plantation species from the Taman Wisata and to replace them with natural forest species. As yet, this recommendation has not been implemented. Due to the value of T. grandis as a food source for T. auratus I propose that this plantation species should not be logged. The value of the Taman Wisata as a public-use zone is enhanced by the abundance of T. auratus in this area thereby favouring the retention of T. grandis stands.

Another of the plantation species, Swietenia macrophylla, also abundant within the TW (see Fig. 2.2, Section 2.2.1), provides comparatively little in the way of food for T. auratus (Section 6.4.1 ). Therefore, this tree species could probably be removed from the TW without causing severe disruption to the T. auratus groups in the TW provided that the removal of S. macrophylla trees is gradual such that a relatively continuous tree canopy layer is maintained during the time required for the newly planted species to develop. However, the ecology of other animal species within the TW should be assessed before removing S. macrophylla stands.

Trachypithecus auratus typically lives in mangrove or secondary forest and is rarely found within primary rain forest. This study shows that this colobine species has adapted to a changing environment, surviving at a comparatively high density in mixed plantation/secondary forest, by using T. grandis leaves as a food source in addition to leaves, fruits and flowers of secondary forest species. 243

REFERENCES

Aldrich-Blake, F.P.G. 1980. Long-tailed macaques. IN Chivers, D.J. (Ed.). Malayan Forest Primates, pp.147-165. London: Plenum Press.

Altmann, J. 1974. Observational study of behaviour: Sampling methods. Behaviour, 49: 227-267.

Altmann, S.A. 1984. What is the dual of the energy-maximization problem? Am. Nat., 123: 433-441.

Altmann, S.A. and Altmann, J. 1970. Baboon Ecology: African Field Research. Chicago: University of Chicago Press.

Altmann, S.A. and Wagner, S.S. 1978. A general model of optimal diet. IN Chivers, D.J. and Herbert, J. (Eds). Recent Advances in Primatology, Vol. 1: Behaviour, pp.407-414. London: Academic Press.

Amerasinghe, F.P., van Cuylenberg, B.W.B. and Hladik, C.M. 1971. Comparative histology of the alimentary tract of Ceylon primates in correlation with the diet. Ceylon J. Sci. (Bio. Sci.), 9: 75-87.

Anderson, D.J. 1982. The home range: a new nonparametric estimation technique. Ecology, 63: 103-112.

Angst, W. and Thommen, D. 1977. New data and a discussion of infant killing in old world monkeys and apes. Folia Primatol., 27: 198-229.

Anon. 1980. Saving Siberut: A Conservation Master Plan. Bogar: W.W.F. Indonesia.

Anon. 1981. Gross and net primary production and growth parameters. IN Tropical Forest Ecosystems (UNESCO/UNDP/FAO), pp.233-248. Paris: UNESCO.

Asquith, T.N. and Butler, L.G. 1985. Use of dye-labelled protein as spectrophotometric assay for protein precipitants such as tannin. J. Chern. Ecol., 11: 1535-1544.

Backer, C.A. and Bakhuizen van den Brink, R.C. 1968. Flora of Java. Groningen: Wolters-Noordhoff N.Y. References 244

Baile, C.A. and Forbes, J.M. 1974. Control of food intake and regulation of energy balance in ruminants. Physiol. Rev., 54: 160-214.

Baranga, D. 1983. Changes in chemical compos1t10n of food parts in the diet of colobus monkeys. Ecology, 64: 668-673.

Bate-Smith, E.C. 1973. Tannins of herbaceous legumes. Phytochemistry, 12: 1809-1812.

Bauchop, T. 1971. Stomach microbiology of primates. Ann. Rev. Microbiol., 25: 429-436.

Bauchop, T. 1978. Digestion of leaves in vertebrate arboreal folivores. IN Montgomery, G.G. (Ed.). The Ecology of Arboreal Folivores., pp.193-204. Washington, D.C.: Smithsonian Institution Press.

Bauchop, T. and Martucci, R.W. 1968. Ruminant-like digestion of the langur monkey. Science, New York, 161: 698-700.

Belovsky, G.E. 1978. Diet optimization in a generalist herbivore, the moose. Theor. Pop. Biol., 14: 105-134.

Belovsky, G.E. 1981. Food selection by a generalist herbivore: the moose. Ecology, 62: 1020-1030. van Bemmel, A.C.V. 1978. Loetongs, anders dan andere apen. Artis (Amsterdam), 24: 30-33.

Bennett, E.L. 1982. The traumas of twins: twinning and infant development in a wild arboreal colobine. (Abstract). Int. J. Primatol., 3: 260.

Bennett, E.L. 1983. The banded langur: ecology of a colobine in West Malaysian rain forest. PhD. dissertation, University of Cambridge.

Bernstein, I.S. 1968. The Lutong of Kuala Selangor. Behaviour, 32: 1-16.

Black, J.L. and Sharkey, M.J. 1970. Reticular groove (Sulcus reticuli): an obligatory adaptation in ruminant-like herbivores? Extrait de Mammalia, 34: 294-302. References '2.45

Blower, J.H., Wind, J. and Mulyana, Y. 1977. Proposed Penanjung Pangandaran Reserve management plan 1977-1981. Field Report. of U.N.D.P./F.A.O. Nature Conservation and Wildlife Management Project. Rome: Food and Agriculture Organisation of the United Nations.

Boggess, J. 1979. Troop male membership changes and infant killing in langurs (Presbytis entellus). Folia Primatol., 32: 65-107.

Bradbury, J.W. and Vehrencamp, S.L. 1976. Social organisation and foraging in Emballonurid bats. II. A model for the determination of group size. Behav. Ecol. Sociobiol., 1: 383-404.

Brotoisworo, E. 1979. The lutung (Presbytis cristata) in Pananjung­ Pangandaran Nature Reserve: Social adaptation to space. Doctor of Science dissertation, Kyoto University.

Brown, J.L. 1964. The evolutionary diversity m avian territorial systems. Wilson Bull., 76: 160-169.

Brown, J.L. and Orians, G.H. 1970. Spacing patterns m mobile animals. Am. Rev. Ecol. Syst., 1: 239-262.

Burnham, K.P., Anderson, D.R. and Laake, J.L. 1980. Estimation of density from line transect sampling of biological populations. Wildl. Monogr., 72: 1-202.

Burns, R.E. 1971. Method for estimation of tannin m sorghum gram. Agron. J, 63: 511-512.

Burt, W.H. 1943. Territoriality and home range concepts as applied to mammals. J. Mammal., 24: 346-352.

Caldecott, J.O. 1980. Habitat quality and populations of two sympatric gibbons (Hylobatidae) on a mountain in Malaya. Folia Primatol., 33: 291-309.

Caldecott, J.O. 1983. An ecological study of the pig-tailed macaque in Peninsular Malaysia. PhD. dissertation, University of Cambridge.

Caldecott, J.O. 1986. An ecological and behavioural study of the pig­ tailed macaque. Contrib. Primatol., 21: 1-259.

Cant, J.G.H. 1980. What limits primates? Primates, 21: 538-544. References 246

Caughley, G.G. 1977. Analysis of Vertebrate Populations. Chichester: Wiley and Sons.

Chapman, M. and Hausfater, G. 1979. The reproductive consequences of infanticide in langurs: a mathematical model. Behav. Ecol. Sociobiol., 5: 227-240.

Chivers, D.J. 1969. On the daily behaviour and spacing of howling monkey groups. Folia Primatol., 10: 48-102.

Chivers, D.J. 1973. An introduction to the socio-ecology of Malayan forest primates. IN Michael, R.P. and Crook, J.H. (Eds). Comparative Ecology and Behaviour of Primates, pp.101-146. London: Academic Press.

Chivers, D.J. 1974. The siamang in Malaya: a field study of a primate in a tropical rain forest. Contrib. Primatol., 4: 1-335.

Chivers, D.J. 1975. The behaviour of siamang in the Krau Game Reserve. Malay. Nat. J., 29: 7-22.

Chivers, D.J. 1980. Introduction. IN Chivers, D.J. (Ed.). Malayan Forest Primates, pp.1-27. London: Plenum Press.

Chivers, D.J. and Hladik, C.M. 1980. Morphology of the gastrointestinal tract in primates: Comparisons with other mammals in relation to diet. J. Morpho!., 166: 337-386.

Chivers, D.J. and Hladik, C.M. 1984. Diet and gut morphology in primates. IN Chivers, D.J., Wood, B.A. and Bilsborough, A. (Eds). Food Acquisition and Processing in Primates, pp.213-230. London: Plenum Press.

Choo, G.M., Waterman, P.G., McKey, D.B. and Gartlan, J.S. 1981. A simple enzyme assay for dry matter digestibility and its value m studying food selection by generalist herbivores. Oecologia, 49: 170-178.

Clutton-Brock, T.H. 1972. Feeding and ranging behaviour of the red colobus monkey. PhD. dissertation, University of Cambridge.

Clutton-Brock, T.H. 1974a. Primate social organization and ecology. Nature, Lond. 250: 539-542.

Clutton-Brock, T.H. 1974b. Activity patterns of red colo bus (C olobus badius tephrosceles). Folia Primatol., 21: 161-187. References 247

Clutton-Brock, T.H. 1975a. Feeding behaviour of red colobus and black and white colobus in East Africa. Folia Primatol., 23: 165- 207.

Clutton-Brock, T.H. 1975b. Ranging behaviour of red colobus (Colobus badius tephrosceles) in the Gombe National Park. Anim. Behav., 23: 706-722.

Clutton-Brock, T.H. (Ed.). 1977a. Primate Ecology: Studies of feeding and ranging behaviour in lemurs, monkeys and apes. London: Academic Press Inc. Ltd.

Clutton-Brock, T.H. 1977b. Some aspects of intraspecific vanatwn in feeding and ranging behaviour in primates. IN Clutton-Brock, T.H. (Ed.). Primate Ecology: Studies of feeding and ranging behaviour in lemurs, monkeys and apes, pp.539-556. London: Academic Press Inc. Ltd.

Clutton-Brock, T.H. and Harvey, P.H. 1976. Evolutionary rules and primate societies. IN Bateson, P.P.G. and Hinde, R.A. (Eds). Growing Points in Ethology, pp.195-237. Cambridge: Cambridge University Press.

Clutton-Brock, T.H. and Harvey, P.H. 1977a. Primate ecology and social organization. J. Zool. Lond., 183: 1-39.

Clutton-Brock, T .H. and Harvey, P.H. 1977b. Species differences in feeding and ranging behaviour in primates. IN Clutton-Brock, T.H. (Ed.). Primate Ecology: Studies of feeding and ranging behaviour in lemurs, monkeys and apes, pp.557-584. London: Academic Press Inc. Ltd.

Clutton-Brock, T.H. and Harvey, P.H. 1978. Primate ecology and social organization. IN Clutton-Brock, T.H. and Harvey, P.H. (Eds). Readings in Sociobiology, pp.342-383. Reading: W.H. Freeman & Co. Ltd.

Clutton-Brock, T.H. and Harvey, P.H. 1980. Primates, brains, and ecology. J. Zool. Lond., 190: 309-323.

Coelho, A.M., Bramblett, C.A., Quick, L.B.and Bramblett, S.S. 1976. Resource availability and population density in primates. Primates, 17: 63-80. References 248

Coley, P.D. 1983. Herbivory and defensive characteristics of tree species in a lowland tropical forest. Ecological Monogr., 53: 209-233.

Connell, J.H. 1978. Diversity in tropical rain forests and coral reefs: high diversity of trees and corals is maintained only m a non­ equilibrium state. Science, 199: 1302-1310.

Conover, W.J. 1980. Practical Nonparametric Statistics. 2nd Ed. New York: John Wiley & Sons.

Cork, S.J. and Pahl, L. 1984. The possible influence of nutritional factors on diet and habitat selection by the ringtail possum (Pseudocheirus peregrinus). IN Smith, A.P. and Hume, I.D. (Eds). Possums and Gliders, pp.269-276. Sydney: Australian Mammal Society.

Corlett, R.T. 1987. The phenology of Ficus fistulosa in Singapore. Biotropica, 19: 122-124.

Crawley, M.J. 1983. Herbivory. Oxford: Blackwell Scientific Publications.

Crook, J.H. 1966. Gelada baboon herd structure and movement: a comparative report. Symp. Zoo!. Soc. Lond., 18: 237-258.

Crook, J.H. 1970. The socio-ecology of primates. IN Crook, J.H. (Ed.). Social Behaviour in Birds and Mammals, pp.106-166. London: Academic Press.

Crook, J.H. 1972. Sexual selection, dimorphism and social organisation in the primates. IN Campbell, B.C. (Ed.). Sexual Selection and the Descent of Man, pp. 231-281. Chicago: Aldine Press.

Crook, J.H., Ellis, J.E. and Goss-Custard, J.D. 1976. Mammalian social systems: structure and function. Anim. Behav., 24: 261-274.

Crook, J.H. and Gartlan, J.S. 1966. Evolution of primate societies. Nature, Lond. 210: 1200-1203.

Curtin, R.A. 1982. Range use of gray langurs in highland Nepal. Folia Primatol., 38: 1-18.

Curtin, R. and Dolhinow, P. 1978. Primate social behaviour m a changing world. Am. Sci., 66: 468-475. References . 249

Curtin, S.H. 1976. Niche separation in sympatric Malaysian colobines (Presbytis obscura and Presbytis melalophos.) Ybk. Phys. Anth., . 20: 421-431.

Curtin, S.H. 1980. Dusky and banded leaf monkeys. IN Chivers, D.J. (Ed.). Malayan Forest Primates, pp.107-145. London: Plenum Press.

Curtin, S.H., and Chivers, D.J. 1978. Leaf-eating primates of peninsular Malaysia: The siamang and the dusky leaf-monkey. IN Montgomery, G.G. (Ed.). The Ecology of Arboreal Folivores, pp.441-464. Washington D.C.: Smithsonian Institution Press.

Cuthbertson, D.P. and Hobson, P.N. 1960. Microbiology and digestion. World Rev. Nutr. Diet., 2: 69-99.

Davies, A.G. 1984. An ecological study of the red leaf monkey (Presbytis rubicunda) in the dipterocarp forest of northern Borneo. PhD. dissertation, University of Cambridge.

Davies, A.G. and Baillie, I.C. 1988. Soil-eating by red leaf monkeys (Presbytis rubicunda) in Sabah, Northern Borneo. Biotropica, 20: 252-258.

Davies, A.G., Bennett, E.L. and Waterman, P.G. 1988. Food selection by two South-east Asian colobine monkeys (Presbytis rubicunda and Presby tis melalophos) in relation to plant chemistry. Bioi. J. Linn. Soc., 34: 33-56.

Davies, A.G., Caldecott, J.O. and Chivers, D.J. 1983. Natural foods as a guide to the nutrition of Old World primates. IN Remfry, J. (Ed.). Standards in Laboratory Animal Management, pp.225-244. Potters Bar: U.F.A.W.

Davies, N.B. and Krebs, J.R. 1978. Introduction: Ecology, natural selection and social behaviour. IN Krebs, K.R. and Davies, N.B. (Eds). Behavioural Ecology - An Evolutionary Approach, pp.l-18. Oxford: Blackwell Scientific Publications.

Denham, W.N. 1971. Energy relations and some basic properties of primate social organization. Amer. Anthrop., 73: 77-95.

DeVore, I. and Hall, K.R.L. 1965. Baboon ecology. IN DeVore, I. (Ed.). Primate Behavior: Field studies of monkeys and apes, pp.20-52. New York: Holt, Rinehart and Winston. References 250

Dolhinow, P.J. 1972. The north Indian langur. IN Dolhinow, P.J. (Ed.). Primate Patterns, pp.181-238. New York: Holt, Rinehart and Winston, Inc.

Dunbar, R.I.M. and Dunbar, E.P. 1974. Ecology and population dynamics of Colobus guereza in Ethiopia. Folia Primatol., 21: 188-208.

Dunbar, R.I.M. and Dunbar, E.P. 1976. Contrasts in social structure among black and white colobus monkey groups. Anim. Behav., 24: 84-92.

Edwards, P.J. and Grubb, P.J. 1982. Studies of mineral cycling in a montane rain forest in New Guinea. J. Ecology, 70: 649-666.

Eisenberg, J.F., Muckenhirn, N.A. and Rudran, R. 1972. The relation between ecology and social structure in primates. Science, 176: 863-874.

Emlen, J .M. 1966. The role of time and energy m food preference. Am. Nat., 100: 611-617.

Ewel, J. 1983. Succession. IN Golley, F.B. (Ed.). Tropical Rain Forest Ecosystems - Ecosystems of the World 14A, pp.217-223. Amsterdam: Elsevier.

Feeny, P.P. 1969. Inhibitory effect of oak leaf tannins on the hydrolysis of proteins by trypsin. Phytochemistry, 8: 2119-2126.

Fooden, J. 1971. Report on primates collected in western Thailand, January-April, 1967. Fie/diana (Zoo!.), 59: 1-62.

Fox, B.J. 1981. Niche parameters and species richness. Ecology, 62: 1415-1425.

Freeland, W.J. 1976. Pathogens and the evolution of primate sociality. Biotropica, 8: 12-24.

Freeland, W.J. 1979. Mangabey (Cercocebus albigena) social organisation and population density in relation to food use and availability. Folia Primatol., 32: 108-124.

Freeland, W.J. and Janzen, D.H. 1974. Strategies in herbivory by mammals: The role of plant secondary compounds. Am. Nat., 108: 269-289. References 251

Furuya, Y. 1961-2. The social life of silvered leaf monkeys (Trachypithecus cristatus). Primates, 3: 41-60.

Gartlan, J.S. 1968. Structure and function in primate society. Folia Primatol., 8: 89-120.

Gartlan, J.S. 1973. Influences of phylogeny and ecology on variations in the group organization of primates. IN Menzel, E.W. (Ed.). Symposia of the Fourth International Congress of Primatology, Vol. 1, pp.88-102. Basel: Karger.

Gartlan, J.S., McKey, D.B. and Waterman, P.G. 1978. Soils, forest structure and feeding behaviour of primates in a Cameroon coastal rain forest. IN Chivers, D.J. and Herbert, J. (Eds). Recent Advances in Primatology, Vol. 1: Behaviour, pp.259-267. London: Academic press.

Gartlan, J.S., McKey, D.B., Waterman, P.G., Mbi, C.N. and Struhsaker, T.T. 1980. A comparative study of the phytochemistry of two African rain forests. Biochem. System. Ecol., 8: 401-422.

Gatinot, B .L. 1977. Le regime alimentaire du Colo be bai au Senegal. Mammalia, 41: 373-402.

Gautier-Hion, A. 1980. Seasonal vanatwns of diet related to species and sex in a community of Cercopithecus monkeys. J. Anim. Ecol., 49: 237-270.

Gibson-Hill, C.A. 1949. The calls of the silvered leaf monkey. Malay. Nat. J., 4: 40.

Giesecke, D. 1970. Comparative microbiology of the alimentary tract. IN Phillipson, A.T. (Ed.). Physiology of Digestion and Metabolism in the Ruminant, pp.306-318. Newcastle-upon-Tyne: Oriel Press.

Gittins, S.P. 1983. Use of the forest canopy by agile gibbons. Folia Primatol., 40: 134-144.

Gittins, S.P. and Raemaekers, J.J. 1980. Siamang, lar and agile gibbons. IN Chivers, D.J. (Ed.). Malayan Forest Primates, pp.63- 105. London: Plenum Press.

Glander, K.E. 1978. Howling monkey feeding behaviour and plant secondary compounds: A study of strategies. IN Montgomery, G.G. (Ed.). The Ecology of Arboreal Folivores, pp.561-574. Washington, D.C.: Smithsonian Institution. References 252

Golley, F.B. 1983. Nutrient cycling and nutrient conservation. IN Golley, F.B. (Ed.). Tropical Rain Forest Ecosystems - Ecosystems of the World, 14A, pp.137-156. Amsterdam: Elsevier.

Goltenboth, R. 1976. Non human primates (apes, monkeys and prosimians). IN Klos H.-G. and Lang, E.M. (Eds). The Handbook of Zoo Medicine (translation), pp.46-85. New York: Van Nostrand Reinhold.

Goss-Custard, J.D., Dunbar, R.I.M. and Aldrich-Blake, F.P.G. 1972. Survival, mating, and rearing strategies in the evolution of primate social structure. Folia Primatol., 17: 1-19.

Gould, S.J. and Lewontin, 1979. The spandrels of San Marco and the Panglossian paradigm: a critique of the adaptationist programme. Proc. R. Soc. Lond. (B), 205: 581-598.

Gray, R.D. 1987. Faith and foraging: a critique of the "paradigm argument from design". IN Kamil, A. C., Krebs, J.R. and Pulliam, H.R. (Eds). Foraging Behaviour, pp.69-140. London: Plenum Press.

Green, K. 1981. Preliminary observations on the ecology and behavior of the capped langur, Presbytis pileatus, in the Madhupur Forest of Bangladesh. Int. J. Primatol, 2: 131-151.

Groves, C.P. 1970. The forgotten leaf-eaters, and the phylogeny of the Colobinae. IN Napier, J.R. and Napier, P.H. (Eds). Old World Monkeys - Evolution, Systematics, and Behaviour, pp.555-587. London: Academic Press.

Groves, C.P. 1973. Notes on the ecology and behaviour of the Angola colobus (Colobus angolensis P.L. Sclater 1860) in N.E. Tanzania. Folia Primatol., 20: 12-26.

Gurmaya, K.J. 1986. Ecology and behavior of Presbytis thomasi in northern Sumatra. Primates, 27: 151-172.

Gysel, L.W. and Lyon, L.J. 1980. Habitat analysis and evaluation. IN Schemnitz, S.D. (Ed.). Wildlife Management Techniques Manual. 4th Ed., Revised, pp.305-327. Washington, D.C.: The Wildlife Society.

Harborne, J.B. 1982. Introduction to Ecological Biochemistry. 2nd Ed. London: Academic Press. References ·253

Harvey, P.H. and Bennett, P.M. 1983. Brain size, energetics, ecology, and life-history patterns. Nature, 306: 314-315.

Harvey, P.H., Martin, R.D. and Clutton-Brock, T.H. 1986. Life histories · in comparative perspective. IN Smuts, B.B., Cheney, D.L., Wrangham, R.W. and Struhsaker, T.T. (Eds). Primate Societies, pp.181-196. Chicago: University of Chicago Press.

Hausfater, G. and Vogel, C. 1982. Infanticide in langur monkeys (genus Presbytis): Recent research and a review of hypotheses. IN Chiarelli, A.B. and Corruccini, R.S. (Eds). Advanced Views in Primate Biology (Main lectures of the VIIIth Congress of the International Primatological Society, Florence, 7-12 July, 1980), pp.160-176. Berlin: Springer-Verlag.

Hesse, P.R. 1971. A Textbook of Soil Chemical Analysis. London: John Murray.

Hill, W.C.O. 1939. An annotated systematic list of the leaf-monkeys. Ceylon J. Sci. (B), 21: 277-305.

Hill, W.C.O. 1952. The external and visceral anatomy of the olive colobus monkey (Procolobus verus). Proc. Zoot. Soc. Lond., 122: 127-186.

Hill, W.C.O. 1966. Primates: Comparative Anatomy and Taxonomy, Vol. VI: Catarrhini, Cercopithecoidea. Edinburgh: Edinburgh University Press.

Hladik, C.M. 1977. A comparative study of the feeding strategies of two sympatric species of leaf monkeys: Presbytis senex and Presbytis entellus. IN Clutton-Brock, T.H. (Ed.). Primate Ecology: Studies of feeding and ranging behaviour in lemurs, monkeys and apes, pp.323-353. London: Academic Press.

Hladik, C.M. 1978. Adaptive strategies of primates in relation to leaf-eating. IN Montgomery, G.G. (Ed.). The Ecology of Arboreal Folivores, pp.373-395. Washington, D.C.: Smithsonian Institution Press.

Hladik, C.M. and Hladik, A. 1972. Disponibilites alimentaires et domaines vitaux des primates a Ceylan. Terre et Vie, 26: 149-215. References 254

Hollihn, K-U. 1971. Das Verhalten von Guerezas (Colobus guereza und Colobus polycomos), Nasenaffen (Nasalis larvatus) und Kleideraffen (Pygathrix nemaeus) bei der Nahrungsaufnahme und ihre Haltung. Z. Saugetierkunde, 36: 65-95.

Homewood, K.M. 1978. Feeding strategy of the Tana mangabey (Cercocebus galeritus galeritus) (Mammalia: Primates). J. Zool. Lond., 186: 375-391.

Hooijer, D.A. 1962. Quarternary langurs and macaques from the Malay Archipelago. Zoo!. Verh. Mus. Leiden., 55: 1-64.

Horwich, R.H. 1972. Home range and food habits of the Nilgiri langur Presbytis johnii. J. Bombay Nat. Hist. Soc., 69: 255-267.

Hrdy, S .B. 197 4. Male-male competition and infanticide among the langurs (Presbytis entellus) of Abu, Rajasthan. Folia Primatol., 22: 19-58.

Hrdy, S.B. 1977. Infanticide as a primate reproductive strategy. Am. Sci., 65: 40-49.

Hrdy, S.B. 1978. Allomaternal care and abuse of infants among Hanuman langurs. IN Chivers, D.J. and Herbert, J. (Eds). Recent Advances in Primatology, Vol. 1: Behaviour, pp.169-172. London: Academic Press.

Hrdy, S.B. 1980. The Langurs of Abu. Cambridge, Massachusetts: Harvard University Press.

Hume, I.D. 1982. Digestive Physiology and Nutrition of Marsupials. Monographs on Marsupial Biology. Cambridge: Cambridge University Press.

Hume, I.D. and Warner, A.C.I. 1980. Evolution of microbial digestion in mammals. IN Ruckebusch, Y. and Thivend, P. (Eds). Digestive Physiology and Metabolism in Ruminants, pp.615-634. Lancaster: MTP Press.

Hungate, R.E. 1968. Ruminal fermentation. IN Code, C.F. (Ed.). Handbook of Physiology, Section 6: Alimentary Canal, Vol. V, pp.2725-2745. Washington, D.C.: American Physiological Society.

Isbell, L. 1983. Daily ranging behaviour of red co lobus (C olobus badius tephrosceles) in Kibale Forest, Uganda. Folia Primatol., 41: 34-48. References 255

Islam, M.I. and Husain, K.Z. 1982. A preliminary study on the ecology of the capped langur. Folia Primatol., 39: 145-159.

Janis, C. 1976. The evolutionary strategy of the Equidae and the origins of rumen and caecal digestion. Evolution, 30: 757-774.

Janzen, D.H. 1974. Tropical blackwater rivers, animals and mast fruiting by the Dipterocarpaceae. Biotropica, 6: 69-103.

Janzen, D.H. 1975. Ecology of Plants in the Tropics. London: Edward Arnold.

Janzen, D.H. 1979. How to be a fig. Ann. Rev. Ecol. Syst., 10: 13-52.

Janzen, D.H. and Waterman, P.G. 1984. A seasonal census of phenolics, fibre and alkaloids in foliage of forest trees in Costa Rica: some factors influencing their distribution and relation to host selection by Sphingidae and Saturniidae. Bioi. J. Linn. Soc., 21: 439-454.

Jarman, P.J. 1974. The social organization of antelope m relation to their ecology. Behaviour, 48: 215-267.

Johns, A.D. 1983. Ecological effects of selective logging in a West Malaysian rain-forest. PhD dissertation, University of Cambridge.

Johns, A.D. 1986. Effects of selective logging on the behavioural ecology of West Malaysian primates. Ecology, 67: 684-694.

Jones, W.T. and Mangan, J.L. 1977. Complexes of the condensed tannins of saionfoin (Onobrychis viciifolia Scop.) with fraction-1 leaf protein and submaxillary mucoprotein and their reversal with polyethylene glycol and pH. J. Sci. Food Agric., 28: 126-136.

Katz, P.L. 1974. A long-term approach to foraging optimization. Am. Nat., 108: 758-782.

Kaufmann, J.H. 1983. On the definitions and functions of dominance and territoriality. Biol. Rev., 58: 1-20.

Kavanagh, M. 1978. The diet and feeding behaviour of Cercopithecus aethiops tantalus. Folia Primatol., 30: 30-63.

Kay, R.N., Hoppe, P. and Maloiy, G.M. 1976. Fermentative digestion of food in the colobus monkey, Colobus polykomos. Experientia, 32: 485-487. References 256

Kay, R.F. and Rylander, W.L. 1978. The dental structure of mammalian folivores with special reference to primates and Phalangeroidea (Marsupialia). IN Montgomery, G.G. (Ed.). The Ecology of Arboreal Folivores, pp.173-191. Washington, D.C.: Smithsonian Institution Press.

Kingdon, J. 1971. East African Mammals, Vol. 1. New York: Academic Press.

Kleiber, N. 1961. The Fire of Life. New York: John Wiley & Sons.

Kuhn, H.-J. 1964. Zur Kenntnis von Bau und Funktion des Magens der Schlankaffen (Colobinae). Folia Primatol., 2: 193-221.

Kuhn, H.-J. 1972. On the perineal organ of male Procolobus badius. J. Hum. Evol., 1: 371-378.

Lebreton, P. 1982. Tannins and alkaloids: two alternative phytochemical tactics for deterring herbivores. Terre et Vie, 3 6: 539-572.

Leutenegger, W. 1978. Scaling of sexual dimorphism in body size and breeding system in primates. Nature, Lond. 272: 610-611.

Leutenegger, W. and Cheverud, J. 1982. Correlates of sexual dimorphism in primates: Ecological and size variables. Int. J. Primatol., 3: 387-402.

Lewontin, R.C. 1979. Fitness, survival and optimality. IN Horn, D.H., Mitchell, R. and Stairs, G.R. (Eds). Analysis of Ecological Systems, pp.3-21. Columbus, Ohio: Ohio State University Press.

McKenna, J.J. 1979. The evolution of allomothering behavior among colobine monkeys - function and opportunism in evolution. Amer. Anthrop., 81: 818-840.

McKey, D.B. 1975. The ecology of coevolved seed dispersal systems. IN Gilbert, L.E. and Raven, P.H. (Eds). Coevolution of Animals and Plants, pp.159-191. Texas: University Press.

McKey, D.B. 1978. Soils, vegetation and seed eating by black colobus monkeys. IN Montgomery, G.G. (Ed.). The Ecology of Arboreal F olivores, pp.423-437. Washington, D.C.: Smithsonian Institution Press. References ·257

McKey, D.B. 1979. Plant chemical defenses and the feeding and ranging behavior of Colobus monkeys in African rain forests. PhD. dissertation, University of Michigan.

McKey, D.B., Gartlan, J.S., Waterman, P.G. and Choo, G.M. 1981. Food selection by black colobus monkeys (Colobus satanas) in relation to plant chemistry. Bioi. J. Linn. Soc., 16: 115-146.

McKey, D.B. and Waterman, P.G. 1982. Ranging behavior of a group of black colobus (Colobus satanas) in the Douala-Edea Reserve, Cameroon. Folia Primatol., .39: 264-304.

McKey, D.B. and Waterman, P.G. 1987. Secondary compounds in rain forest plants: patterns of distribution and biological implications. IN Lieth, H. and Werger, J.A. (Eds). Ecosystems of the World, 14A. The Netherlands: Elsevier.

McManus, J.P., Davis, K.G., Lilley, T.H. and Haslam, E. 1981. The association of proteins with polyphenols. J. Chern. Soc. Chern. Commun. Perkin Trans., 00: 309-311.

MacArthur, R.H. and Pianka, E.R. 1966. On optimal use of a patchy environment. Am. Nat., 100: 603-609.

MacKinnon, J. 1974. The behaviour and ecology of wild orang utans (Pongo pygmaeus). Anim. Behav, 22: 3-74.

MacKinnon, J. 1977. A comparative ecology of Asian apes. Primates, 18: 747-772.

MacKinnon, J.R. and MacKinnon, K.S. 1978. Comparative feeding ecology of six primates in west Malaysia. IN Chivers, D.J. and Herbert, J. (Eds). Recent Advances in Primatology, Vol. 1: Behaviour, pp.305-321. London: Academic Press.

MacKinnon, J.R. and Mackinnon, K.S. 1980. Niche differentiation in a primate community. IN Chivers, D.J. (Ed.). Malayan Forest Primates, pp.167-190. London: Plenum Press.

Mah, Y.L. 1980. Ecology and behaviour of Macaca fascicularis. PhD. dissertation, University of Malaya.

Makwana, S.C. 1979. Infanticide and social change in two groups of the Hanuman langur, Presbytis entellus, at Jodhpur. Primates, 20: 293-300. References 258

Marsh, C. 1978a. Comparative activity budgets of red colobus. IN Chivers, D.J. and Herbert, J. (Eds). Recent Advances in Primatology, Vol. 1: Behaviour, pp.249-251. London: Academic Press.

Marsh, C.W. 1978b. Ecology and social organization of the Tana River red colobus, Colobus badius rufomitratus. PhD. dissertation, University of Bristol.

Marsh, C.W. 1979a. Female transference and mate choice among Tana River red colobus. Nature, Lond. 281: 568-569.

Marsh, C.W. 1979b. Comparative aspects of social organization m the Tana River red colobus, Colobus badius rufomitratus. Z. Tierpsychol., 51: 337-362.

Marsh, C.W. 1981a. Diet choice among red colobus (Colobus badius rufomitratus) on the Tana River, Kenya. Folia Primatol., 35: 147- 178.

Marsh, C.W. 1981b. Time budget of Tana River red colobus. Folia Primatol., 35: 30-50.

Marsh, C.W. 1981c. Ranging behaviour and its relation to diet selection in Tana River red colobus (Colobus badius rufomitratus). J. Zool. Lond., 195: 473-492.

Marsh, C.W., Johns, A.D. and Ayres, J.M. 1987. Effects of habitat disturbance on rain forest primates. IN Marsh, C.W. and Mittermeier, R.A. (Eds). Primate Conservation in the Tropical Rain Forest, pp.83-107. New York: Alan R. Liss, Inc.

Marsh, C.W. and Wilson, W.L. 1981a. A Survey of Primates in Peninsular Malaysian Forests. Kuala Lumpur: Universiti Kebangsaan Malayasia and University of Cambridge.

Marsh, C.W. and Wilson, W.L. 1981b. Effects of natural habitat differences and disturbance on the abundance of Malaysian primates. Malays. Appl. Biol., 10: 227-249.

Martin, J.S. and Martin, M.M. 1982. Tannin assays in ecological studies: lack of correlation between phenolics, proanthocyanidins and protein-precipitating constituents in mature foliage of six oak species. Oecolgia, 54: 205-211. References 259

May, R.M. 1981. Patterns in multi-species commumtles. IN May, R.M. (Ed.). Theoretical Ecology - Principles and Applications, pp.197 -227. Oxford: Blackwell Scientific Publications.

Medway, L. 1970a. The monkeys of Sundaland: Ecology and systematics of the cercopithecids of a humid equatorial environment. IN Napier, J.R. and Napier, P.H. (Eds). Old World Monkeys - Evolution, Systematics and Behaviour, pp.513-553. London: Academic Press.

Medway, Lord. 1970b. Breeding of silvered leaf-monkey, Presbytis cristata, in Malaya. J. Mammal., 51: 630-632.

Medway, Lord. 1972. Phenology of a tropical rain forest m Malaya. Biol. J. Linn. Soc., 4: 117-146.

Milton, K. 1981. Distribution patterns of tropical plant foods as an evolutionary stimulus to primate mental development. Amer Anthrop., 83: 534-548.

Mitani, J.C. and Rodman, P.S. 1979. Territoriality: the relation of ranging pattern and home range size to defendability, with an analysis of territoriality among primate species. Behav. Ecol. Sociobiol., 5: 241-251.

Mohnot, S.M. 1971. Some aspects of social changes and infant-killing in the Hanuman langur, Presbytis entellus (Primates: Cercopithecidae), in Western India. Mammalia, 35: 175-198.

Moir, K.W. 1982. Theory and practice of measuring the cell-wall content of food for ruminant and non-ruminant animals. Lab. Practice, 31: 732-733.

Moir, R.J. 1968. Ruminant digestion and evolution. IN Code, C.F. (Ed.). Handbook of Physiology, Section 6: Alimentary Canal, Vol. V, pp.2673-2694. Washington D.C.: American Physiological Society.

Mole, S. and Waterman, P.G. 1987. Tannins as antifeedants to mammalian herbivores - still an open question? IN Waller, G.R. (Ed.). Allelochemicals: Role in Agriculture and Forestry, pp.572- 587. Washington: American Chemical Society Symposium Series.

Moore, P.D. 1983. Ecological diversity and stress. Nature, Lond. 306: 17. References 260

Moreno-Black, G.S. and Bent, E.F. 1982. Secondary compounds in the diet of Colobus angolensis. Afr. J. Ecol., 20: 29-36.

Mukherjee, R.P. 1982. Phayres leaf monkey (Presbytis phayrei) of Tripura, India. J. Bombay Nat. Hist. Soc., 79: 47-56.

Napier, J.R. and Napier, P.R. 1967. A Handbook of Living Primates. London: Academic Press.

Napier, P.R. 1985. Catalogue of Primates in the British Museum (Natural History) and Elsewhere in the British Isles. Part III: Family Cercopithecidae, Subfamily Colobinae. London: British Museum (Natural History).

Newbery, D. McC., Gartlan, J.S., McKey, D.B. and Waterman, P.G. 1986. The influence of drainage and soil phosophorus on the vegetation of Douala-Edea Forest Reserve, Cameroun. Vegetatio, 65: 149-162.

Newton, P.N. 1988. The variable social organization of hanuman langurs (Presbytis entellus), infanticide and the monopolization of females. Int. J. Primatol., 9: 59-77.

Noble, G.K. 1939. Dominance in the life of birds. Auk, 56: 263-273.

Norusis, M.J. 1985. SPssx Advanced Statistics Guide. (SPSS Inc.) New York: McGraw-Hill Book Co.

Oates, J.F. 1977a. The guereza and its food. IN Clutton-Brock, T.H. (Ed.) Primate Ecology: Studies of feeding and ranging behaviour in lemurs, monkeys and apes, pp.275-321. London: Academic Press.

Oates, J .F. 1977b. The social life of a black-and-white colo bus monkey, Colobus guereza. Z. Tierpsychol., 45: 1-60.

Oates, J.F. 1978. Water-plant and soil consumption by guereza monkeys (Colobus guereza): a relationship with minerals and toxins in the diet? Biotropica, 10: 241-253.

Oates, J.F. 1986. Food distribution and foraging behaviour. IN Smuts, B.B., Cheney, D.L., Seyfarth, R.M., Wrangham, R.W. and Struhsaker, T.T. Primate Societies, pp.197-209. Chicago: University of Chicago Press. References ·261

Oates, J.F., Swain, T. and Zantovska, J. 1977. Secondary compounds and food selection by colobus monkeys. Biochem. System. Ecol., 5: 317-321.

Oates, J.F. and Trocco, T.F. 1983. Taxonomy and phylogeny of black­ and-white colobus monkeys. Inferences from an analysis of loud call variation. Folia Primatol., 40: 83-113.

Oates, J.F., Waterman, P.G. and Choo, G.M. 1980. Food selection by the South Indian leaf-monkey, Presbytis johnii, in relation to leaf chemistry. Oecologia, 45: 45-56.

Ohwaki, K., Hungate, R.E., Lotter, L., Hofmann, R.R. and Maloiy, G. 1974. Stomach fermentation in East African Colobus monkeys m their natural state. Appl. Microbial., 27: 713-723.

Opler, P.A. 1978. Interaction of plant life history components as related to arboreal herbivory. IN Montgomery, G.G. (Ed.) The Ecology of Arboreal Folivores, pp.23-31. Washington, D.C.: Smithsonian Institution Press.

Owen-Smith, N. and Novellie, P. 1982. What should a clever ungulate eat? Am. Nat., 119: 151-178.

Parra, R. 1978. Comparison of foregut and hindgut fermentation in herbivores. IN Montgomery, G.G. (Ed.). The Ecology of Arboreal Folivores, pp.205-229. Washington, D.C.: Smithsonian Institution Press.

Payne, J.B. and Davies, A.G. 1982. A Faunal Survey of Sabah. Kuala Lumpur: World Wildlife Fund Malaysia.

Pielou, E.C. 1966. Shannon's formula as a measure of specific diversity: its use and misuses. Am. Nat., 100: 463-465.

Pocock, R.I. 1935. On monkeys of the genera Pithecus (or Presbytis) and Pygathrix found to the east of the Bay of Bengal. Proc. Zoo!. Soc. Lond.: 895-961.

Pocock, R.I. 1939. The Fauna of British India, Including Ceylon and Burma: Mammalia I. London: Taylor and Francis.

Poirier, F.E. 1969. The Nilgiri langur (Presbytis johnii) troop: Its composition, structure, function and change. Folia Primatol., 10: 20-47. References 262

Poirier, F.E. 1970. The Nilgiri langur (Presbytis johnii) of South India. IN Rosenblum, L.A. (Ed.). Primate Behaviour - Developments in Field and Laboratory Research, Vol. 1, pp.251- 283. New York: Academic Press.

Poirier, F.E. 1974. Colobine aggression: a review. IN Holloway, R.L. (Ed.). Primate Aggression, Territoriality and Xenophobia, pp.123- 157. New York: Academic Press.

Proctor, J.J., Anderson, M., Chai, P. and Vallack, H.W. 1983. Ecological studies in four contrasting lowland rain forests in Gunung Mulu National Park, Sarawak. I. Forest environment, structure and floristics. J. Ecology, 71: 237-260.

Pyke, G.H. 1984. Optimal foraging theory: a critical review. Ann. Rev. Ecol. Syst., 15: 523-575.

Raemaekers, J.J. 1977. Gibbons and trees: comparative ecology of the siamang and lar gibbons. PhD dissertation, University of Cambridge.

Raemaekers, J .J. 197 8. Changes through the day m the food choice of wild gibbons. Folia Primatol., 30: 194-205.

Raemaekers, J.J. 1980. Causes of variation between months in the distance travelled daily by gibbons. Folia Primatol., 34: 46-60.

Raemaekers, J.J., Aldrich-Blake, F.P.G and Payne, J.B. 1980. The Forest. IN Chivers, D.J. (Ed.). Malayan Forest Primates, pp.29-61. London: Plenum Press.

Raemaekers, J.J. and Chivers, D.J. 1980. Socio-ecology of Malayan forest primates. IN Chivers, D.J. (Ed.). Malayan Forest Primates, pp.279-316. London: Plenum Press.

Reed, J. and Soller, H. 1987. Phenolics and nitrogen utilization in sheep fed browse. IN Herbivore Nutrition Research - Proceedings from Second International Symposium on the Nutrition of Herbivores, Univ. of Queensland, Brisbane. July 6-10, 1987. Australian Society of Animal Production.

Rhoades, D.F. and Cates, R.G. 1976. Toward a general theory of plant anti-herbivore chemistry. IN Wallace, J.W. and Mansell, R.L. (Eds). Biochemical Interaction Between Plants and Animals, pp.168-213. New York: Plenum Press. References 263

Rijksen, H.D. 1978. A Field Study on Sumatran Orang Utans (Pongo pygmaeus albelii Lesson 1827). Wageningen: H. Veenman and Zonen B.V.

Rijksen, H.D. 1981. Infant killing; a possible consequence of a disputed leader role. Behaviour, 78: 138-168.

Ripley, S. 1970. Leaves and leaf-monkeys: The social organization of foraging in gray langurs Presbytis entellus thersites. IN Napier, J.R. and Napier, P.H. (Eds). Old World Monkeys - Evolution, Systematics, and Behaviour, Part III: Behaviour and Ecology, pp.481-509. London: Academic Press.

Robins, T.C. 1983. Wildlife Feeding and Nutrition. New York: Academic Press.

Rodman, P.S. 1978. Diets, densities, and distributions of Bornean primates. IN Montgomery, G.G. (Ed.). The Ecology of Arboreal Folivores, pp.465-478. Washington, D.C.: Smithsonian Institution Press.

Roonwal, M.L. and Mohnot, S.M. 1977. Primates of South Asia. Ecology, Sociobiology and Behavior. Cambridge: Harvard University Press.

Rosenthal, G.A. and Janzen, D.H. (Eds). 1979. Herbivores, Their Interactions with Secondary Plant Metabolites. London: Academic Press.

Rowell, T.E. 1966. Forest living baboons in Uganda. J. Zool. Lond., 149: 344-364.

Rudran, R. 1973. Adult male replacement in one-male troops of purple-faced langurs and its effect on population structure. Folia Primatol., 19: 166-192.

Ruhiyat, Y. 1983. Socio-ecological study of Presbytis aygula in west Java. Primates, 24: 344-359.

van Schaik, C.P., and van Hoof, J.A.R.A.M. 1983. On the ultimate causes of primate social systems. Behaviour, 85: 91-117.

Schmidt-Nielsen, K. 1972. Locomotion: energy costs of swimming, flying and running. Science, 177: 222-228. References 264

Schoener, T.W. 1969. Optimal size and specialization in constant and fluctuating environments: an energy-time approach. Brookhaven Symp. Biol., 22: 103-114.

Schoener, T.W. 1971. Theory of feeding strategies. Ann. Rev. Ecol. Syst ., 11: 369-403.

Schoener, T.W. 1987. A brief history of optimal foraging ecology. IN Kamil, A.C., Krebs, J.R. and Pulliam, H.R. (Eds). Foraging Behaviour, pp.5-67. London: Plenum Press.

Schubert, G. 1982. Infanticide by usurper hanuman langur males - sociobiological myth. Social Sci. Information, 21: 199-244.

Siegel, S. 1956. Nonparametric Statistics for the Behavioural Sciences. Tokyo: McGraw-Hill/Kogakusha.

Simpson, E.H. 1949. Measurement of diversity. Nature, 163: 688. van Soest, P.J. 1963. Use of detergents in the analysis of fibrous feeds. II. A rapid method for the determination of fibre and lignin. J. Assoc. Official Agric. Chern., 46: 829-835. van Soest, P.J. 1977. Plant fibre and its role in herbivore nutrition. Cornell Vet., 67: 307-326. van Soest, P.J. 1981. Forages, Fiber and the Rumen. Corvallis, Oregon: C. and 0. Books.

Sommer, V. and Mohnot, S.M. 1985. New observations on infanticides among hanuman langurs (P. entellus) near Jodhpur (Rajasthan/India). Behav. Ecol. Sociobiol., 16: 245-248.

Southwick, C.H. and Cadigan, F.C. 1972. Population studies of Malaysian primates. Primates, 13: 1-18.

Starin, E.D. 1978. A preliminary investigation of home range use m the Gir forest langur. Primates, 19: 551-568.

Stephens, D.W. and Krebs, J.R. 1986. Foraging Theory. Princeton, New Jersey: Princeton University Press.

Strier, K.B. 1987. Ranging behavior of woolly spider monkeys, or muriquis, Brachyteles arachnoides.lnt. J. Primatol., 8: 575-591. References Q.65

Struhsaker, T.T. 1967. Ecology of vervet monkeys (Cercopithecus aethiops) in the Masai - Amboseli Game Reserve, Kenya. Ecology, . 48: 891-904.

Struhsaker, T.T. 1969. Correlates of ecology and social organization among African cercopithecines. Folia Primatol., 11: 80-118.

Struhsaker, T.T. 1970. Phylogenetic implications of some vocalizations of Cercopithecus monkeys. IN Napier, J.R. and Napier, P.H. (Eds). Old World Monkeys- Evolution, Systematics and Behaviour, pp.365-444. London: Academic Press.

Struhsaker, T.T. 1974. Correlates of ranging behaviour in a group of red colobus monkeys (Colobus badius tephrosceles). Am. Zool., 14: 177-184.

Struhsaker, T.T. 1975. The Red Colobus Monkey. Chicago: University of Chicago Press.

Struhsaker, T.T. 1981. Polyspecific associatiOns among tropical ram forest primates. Z. Tierpsychol., 57: 268-304.

Struhsaker, T.T. and Leland, L. 1979. Socioecology of five sympatric monkey species in the Kibale Forest, Uganda. IN Rosenblatt, J.S., Hinde, R.A., Beer, C. and Busnel, M.C. (Eds). Advances in the Study of Behaviour, Vol. 9, pp.159-228. New York: Academic Press.

Struhsaker, T.T. and Leland, L. 1985. Infanticide in a patrilineal society of red colobus monkeys. Z. Tierpsychol, 69: 89-132.

Struhsaker, T.T. and Leland, L. 1986. Colobines: infanticide by adult males. IN Smuts, B.B., Cheney, D.L., Wrangham, R.W. and Struhsaker, T.T. (Eds). Primate Societies, pp.83-97. Chicago: University of Chicago Press.

Struhsaker, T.T. and Oates, J.F. 1975. Comparison of the behavior and ecology of red colobus and black and white colobus monkeys in Uganda: a summary. IN Tuttle, R.H. (Ed.). Socio-ecology and Psychology of Primates, pp.103-123. The Hague: Mouton.

Sugiyama, Y. 1964. Group composition, population density and some sociological observations of hanuman langurs. Primates, 5: 7-37.

Sumardja, E.A. and Kartawinata, K. 1977. Vegetation analysis of the habitat of banteng (Bos javanicus) at the Pananjung-Pangandaran Nature Reserve, West Java. Biotrop Bull., 13: 3-43. References 266

Supriatna, J., Manullang, B.O. and Soekara, E. 1986. Group composition, home range, and diet of the maroon leaf monkey (Presbytis rubicunda) at Tanjung Puting Reserve, Central Kalimantan, Indonesia. Primates, 27: 185-190.

Thorington, R.W. (Jr.) and Groves, C.P. 1970. An annotated classification of the Cercopithecoidea. IN Napier, J.R. and Napier, P.H. (Eds). Old World Monkeys - Evolution, Systematics and Behaviour, Part V: Systematic conclusions, pp.629-647. London: Academic Press.

Tilson, R.L. 1977. Social behaviour of Simakobu monkeys and its relationship to human predation. J. Mammal., 58: 202-212.

Tilson, R.L. and Tenaza, R.R. 1976. Monogamy and duetting in an Old World monkey. Nature, Lond. 263: 320-321.

Tucker, B .M. 1985. Laboratory procedures for soluble salts and exchangeable cations in soils. CSIRO Aust. Div. Soils Tech. Pap. No. 47: 1-36.

Vellayan, S. 1982. Chemistry and digestibility of foods of the lar gibbon (Hylobates lar). M.Sc. thesis, Universiti Pertanian Malaysia.

Waser, P.M. 1977. Feeding, ranging and group size in the mangabey Cercocebus albigena. IN Clutton-Brock, T.H. (Ed.). Primate Ecology: Studies of feeding and ranging behaviour in lemurs, monkeys and apes, pp.183-221. London: Academic Press.

Washburn, S.L. 1944. The genera of Malaysian langurs. J. Mammal., 25: 289-294.

Watanabe, K. 1981. Variations in group compositiOn and population density of two sympatric Mentawaian leaf-monkeys. Primates, 22: 145-160.

Waterman, P.G. 1984. Food acqms1t10n and processing by primates as a function of plant chemistry. IN Chivers, D.J., Wood, B.A. and Bilsborough, A. (Eds). Food Acquisition and Processing by Primates, pp.177-211. New York: Plenum Press.

Waterman, P.G. and Choo, G.M. 1981. The effects of digestibility­ reducing compounds in leaves on food selection by some Colobinae. Malays. Appl. Biol., 10: 147-162. References 267

Waterman, P.G., Choo, G.M., Vedder, A.L. and Watts, D. 1983. Digestibility, digestion-inhibitors and nutrients of herbaceous foliage and green stems from an African montane flora and comparison with other tropical flora. Oecologia, 60: 244-249.

Waterman, P.G., Mbi, C.N., McKey, D.B. and Gartlan, J.S. 1980. African rain forest vegetation and rumen microbes: phenolic compounds and nutrients as correlates of digestibility. Oecologia, 47: 22-33.

Waterman, P.G., Ross, J.A.M., Bennett, E.L. and Davies, A.G. 1988. A comparison of the floristics and leaf chemistry of the tree flora in two Malaysian rain forests and the influence of leaf chemistry on populations of colobine monkeys in the Old World. Biol. J. Linn. Soc., 34: 1-32.

Watkins, B.E., Ullrey, D.E. and Whetter, P.A. 1985. Digestibility of a high-fiber biscuit-based diet by Black and White Colobus (C. guereza). Amer. J. Primatol., 9: 137-144.

Watson, J.C. 1928. Mangrove forests of the Malay Peninsula. Malay. Forest Rec., 6: 1-275.

Weitzel, V. and Groves, C.P. 1985. The nomenclature and taxonomy of the colobine monkeys of Java. Int. J. Primatol., 6: 399-409.

Westoby, M. 197 4. An analysis of diet selection by large generalist herbivores. Am. Nat., 961: 290-304.

Whitesides, G.H., Oates, J.F., Green, S.M. and Kluberdanz, R.P. 1988. Estimating primate densities from transects in a West African rain forest: a comparison of techniques. J. Anim. Ecol., 57: 345- 367.

Whitmore, T.C. 1975. Tropical Rain Forests of the Far East. Oxford: Clarendon Press.

Whitten, A.J. 1982. A numerical analysis of tropical ram forest, using floristic and structural data, and its application to an analysis of gibbon ranging behaviour. J. Ecology, 70: 249-271.

Willis, J.C. 1966. A Dictionary of the Flowering Plants and Ferns. 7th Ed. (revised by H.K. Airy Shaw). London: Cambridge University Press. References 268

Wilson, C.C. and Wilson, W.L. 1973. Final report: census of Sumatran primates. University of Washington, Seattle.

Wilson, C.C. and Wilson, W.L. 1975. The influence of selective logging on primates and some other animals in East Kalimantan, Indonesia. Folia Primatol, 23: 245-274.

Wilson, C.C. and Wilson, W.L. 1976. Behavioral and morphological variation among primate populations in Sumatra. Ybk. Phys. Anthrop., 20: 207-233.

Wilson, W.L. and Wilson, C.C. 1975. Species-specific vocalizations and the determination of phylogenetic affinities of the Presbytis aygula-melalophos group in Sumatra. Contemp. Primat. (5th Int. Congr. Primat. Nagoya 1974), pp.459-463.

Wolf, K. 1980. Social change and male reproductive strategy in silvered leaf-monkeys, Presbytis cristatus, in Kuala Selangor, Peninsular Malaysia. Am. J. Phys. Anthropol., 52: 294.

Wolf, K.E. and Fleagle, J.G. 1977. Adult male replacement in a group of silvered leaf-monkeys (Presbytis cristata) at Kuala Selangor, Malaysia. Primates, 18: 949-955.

Wolter, R. 1982. Alimentation et pathologie chez le cheval. Pratique Ve'terinaire Equine, 14: 12-20.

Wrangham, R.W. 1975. The behavioural ecology of chimpanzees in Gombe National Park. PhD dissertation, University of Cambridge.

Wrangham, R.W. 1979. On the evolution of ape social systems. Social Sci. Information, 18: 335-368.

Wrangham, R.W. 1986. Evolution of social structure. IN Smuts, B.B., Cheney, D.L., Wrangham R.W. and Struhsaker, T.T. (Eds). Primate Societies, pp.282-296. Chicago: University of Chicago Press.

Wrangham, R.W. and Waterman, P.G. 1981. Feeding behavior of vervet monkeys on Acacia tortilis and Acacia xanthophloea with special reference to reproductive strategies and tannin production. J. Anim. Ecol., 50: 715-732.

Yoshiba, K. 1968. Local and intertroop variability in ecology and social behaviour of common Indian langurs. IN Jay, P.C. (Ed.). Studies in Adaptation and Variability, pp.217-242. New York: Holt, Rinehart and Winston. '269

APPENDIX I STEM BASAL AREA AS AN INDICATOR OF FOLIAGE BIOMASS

A measure of accuracy in using stem basal area as a relative index of foliage biomass is obtained by taking measurements in the field of canopy size. The following measures of canopy size were recorded for trees whose canopies fell within quadrats, for quadrats constituting 5% of the total home-range area in each study site:

1. Crown depth (m); 2. Crown length (m); 3. Crown breadth (m).

To determine crown breadth and length the distance was measured between two people standing under extreme ends of the canopy diameters. Crown depth was estimated by eye. From these measures the following indices or areas were calculated:

1. Canopy index 1 = crown length x crown depth (after Marsh 1981a). 2. Canopy index 2 = (crown length + crown breadth)/2 x crown depth. This index is more appropriate than index 1 where canopies are elliptical rather than circular in shape. 3. Crown surface area = crown length x crown breadth x 0.7854 (cm2) (after Gysel and Lyon 1980).

Correlation coefficients between basal area and each of the three canopy indices were calculated for each species with a sample of trees 1 5 (Table I.l). Generally, basal area was highly correlated with each of the three canopy measures; most correlation coefficients having a value between 0.6 and 1.0. A notable exception was Vitex pubescens for which basal area was not highly correlated with any of the three canopy measures. Trees of this species were large (in height) and canopies variable in size and shape. Appendix I 270

Table 1.1 Correlations Between Basal Area and Canopy Measures for Species with a Sample of Trees l 5

Correlation coefficients of basal area with Species en en b (ii)CJ2C (iii)CAd

Antidesma bunius 0.911 0.878 0.724 Aphanamixis grandifoliaa 0.951 0.931 0.636 Buchanania arborescensa 0.761 0.695 0.923 Canarium hirsutum 0.816 0.876 0.973 Casearia grewaefolia 0.961 0.937 0.630 Cinnamomum iners 0.258 0.882 0.927 Clausena excavataa 0.961 0.953 0.845 Decaspermum fructicosum 0.987 0.984 0.953 Dendrocnide microstigma 0.743 0.709 0.997 Diospyros cauliflora 0.652 0.523 0.576 Diospyros javanicaa 0.875 0.865 0.909 Dysoxylum caulostachyuma 0.914 0.895 0.934 Eugenia lineata 0.850 0.851 0.805 Eugenia zippelianuma 0.990 0.989 0.986 Ficus sumatranaa 0.856 0.873 0.949 Flacourtia inermis 0.675 0.630 0.572 Gynotroches axillarisa 0.712 0.669 0.679 Hernandia peltataa 0.685 0.622 0.473 Linociera sp. 0.784 0.774 0.827 Mallotus oblongifolius 0.936 0.936 0.946 Memecylon myrsinoides 0.594 0.655 0.845 Mischocarpus sundaicusa 0.805 0.815 0.901 Pterospermum javanicuma 0.950 0.959 0.999 Swietenia macrophyllaa 0.705 0.710 0.871 Tectona grandisa 0.773 0.792 0.794 Tricalysia singularis 1.000 1.000 0.999 Vitex pubescensa 0.565 0.483 0.280 aspecies which each comprised l 2% of the total % feeding time of T. auratus and/or were important vegetation species (i.e. l 2% of total basal area and/or l 5% of stem density). ben = canopy index 1. cc12 = canopy index 2. dcA = crown area. (refer to text for details)

Canopy surface area (Clutton-Brock 1972) and canopy volume (Bennett 1983) are good indicators of the amount of food available to colobines. Bennett (1983) showed that both canopy surface area and canopy volume were positively correlated with basal area. The data presented in Table I.l also show strong correlations between Appendix I 271

basal area and crown surface area as well as two other canopy measures. The use of stem basal area as a relative index of foliage biomass was, therefore, considered appropriate in this study as:

1. Correlations between basal area and canopy measures were generally good (Table I.1);

2. Basal area has been widely used as an index of tree size in studies of colobine ecology (e.g. McKey et al. 1979; Caldecott 1980, 1983; Bennett 1983; Davies 1984);

3. The measurement of stem basal area ts comparatively rapid and accurate when compared with the measurement of canopy size which is time consuming and involves two people. 272

APPENDIX II VEGETATION SPECIES WITH AUTHORITIES (nomenclature used follows that of Willis (1966) and Backer et al. (1968))

AIAC!RDIACEAI BDCBAWIA ARBORESCIIS (BL.IBL. AUCABDIACEAE IIAJGIPKRA IIDICA L. AIIAC!RDIACEAE IIAIGIPW LOIGIPES GRIPP, AR!CABDIACEAK SI!IBCARPOS HETEROPBYLLA BL. ARAC!RDIACEAE SPOIDIAS Cn'IIRBA SOli, AIIOilCEAE AIIIIORACBAE CAUIGA ODORATA(LMI)BOOI.F,, TBOKS, AIIOilCIAI POLULTBI! LATBRIPLORA {BL.) IIIG. AIIIOUCIAB PSIODDVARIA RIGOSA (BL.I MDI. APOCYIIACBAE ALUlA REIIIARDTI BL.VAR.LATIPOLIA(BL.)BAIB.F. APOCYUCBAE CDBDA MAIGBAS L. APOCYIIACEAB ORCIOLA BRACBJSIPALA BOOI.F. WCBAB SCIIDAPSUS BBDIRACIOS(Z,'",)MIQ. ARALIA CUE SCIIPPLERA ELLIPTICA(BL. l8ARIIS ARECA CRAB ARDGA 0Br8SIPOLIA BL. EX MART. ARKC!CEAB DliiOIOROPS IIILAIOCBAETB BL. BIGIOIIACDI! DOLICBAIDROII SPATBACEA{L.F.li.SCBOII. BIGIIOIIACEAI! OiOfiLO! IIDICOM (L,) VBIT. BIGIOIIACEAI RADBHIIACBKRA GIGAITIA(BL.IMIQ. IMIIBACACIAB IKUAI VALBTOIII BOCBR. BORAGIIIACBAB CORDIA SP BDUDACUI CADRIIJII ASPBRUM VAR.ASPBRDM BTH. BORSERACBAB C!IAIIUM cf. IIDICUM L. BORSERACUB C!URIOM BIRSOTOM IILLD. BIISEIACDI CADRIOM LITTORALE BL. CASUARIUCEAI CASOARIRA KQ8ISITIFOLIA J.R,, G.PORST. CLUSIACEAE CIA!OfiLOI SP CLUSI!CEAB GAJlCIJIA BALICA MIQ. CLUSIACIAI GARCIRIA DDLCIS{ROIB.IIORI CL8SIACUB GARCIIIA SP COWRACBAE AGILAD MACROPBYLLA(ZOLL,)LI!IB. COIIWCDB CIISTIS PLATAITBA GRIFF. COllAR! CUI lORlA MIIIOR{GABRTI.ILBDB. COIVOLVOLACKAE IP

..... Appendix II (cont.)

KOPBORBIACEAK AITIDBSMA TKTAIDRIII BL. EIIPIORBIACEAK APOROSA MICROSPB!KRA BOOI.P. KOPBORBI!CEAK BACCI URI! JAVIIICA (BL.I M.A. KOPIORBI!CUK BREYIIA VIRGATA (BL.I M.A. KOPIORBI!CEAK BRIDKLIA IIOIOICA ILOUR.IMERR. KOPBORBI!CUK CLAOHLOH POLOT IBOI!II.F .IMERR. KUPRORBIACEAK GLOCBIDIOI KACROCARPOII BL, KOPRORBIACUI GLOCBIDIOI SP KDPBORBI!CEAK GLOCBIDIOI SP KOPIORBIACUK W.LOTDS MORITSIAIOS M.A. KUPIORBIACEAK KALLOTUS MILTIGLAIDDLOSA(BL.IRCBB.P,,IOLL. KOPBORBIACUB IIILLOTOS OBLOIGIPOLIOS M.A. KDPRORBIACEAB IIALLOTOS PILTITUS(GIISILIM.A. KDPBOIBI!CUE IIALLOTDS PBILIPPKISIS ILMII M.A. KUPIORBIACEAB MILLOTUS RICIIOID!S(P!RS.IM.A. KOPIORBIACEAK W.LOTDS SP EJPIORBIACEAK MILLOTDS SP BOPBORBIACEAK IIALLOTOS TILIIEFOLIBS IBL.I M.A. IUPIOIBIACEAK IEOSCOKT!CBIHIA IIIGIIIBOOI.F.IPAI,I.BOFPM. IBPBORBIACUK SIDROPOS AIDROGUOS IL ,) KERR, IIPIORBIACEAK SORIGADA GLOIBRITAIBL.IBAILL. FWCIAK PWBIA FBLVA (BL. II IORTB.IBTB. FABACUK PABA CUI ACACIA AIIICBLIPORMIS A.COII. Kl BTH. PABA CUE CASSIA JAVUICA L. FAllA CUE CASSIA SlAW Lll PABACUB CYIOMETRA RAIIFLORA L. FABACUK DALBERGIA LATIPOLIA ROIB, PABA CUE DIMDIBII UIIBILLA TOll IL .I DC • FAllA CUE KRHDIII VARIBGATA L. PLACODR'!'IACUK CASURIA FLIYOVIRKIS BL. PLACODITIACUK CASBIRIA GRII!EFOLIA VKIT. PLACOUR'!'IACUK CUBA RIA TDDCULATA BL. PLACOUKTIACUI FL&CODITIA IIDiliS ROIB. PLACODHIACEAK SCOLOPIA SPIIOSA (ROIB.I WARB. B!WIDIACEAK BDIIIDIA PILTATA IIKISSII. LAilllCIII BIILSCBIIIBDIA SP LlDIACBAB CIIIAIIOIIOII IIERS RIIII. II BL. LAURACBAE IIOLITSU CASSIA IL.I IOSTBRII. LADBACUI IIOLITSEA SP LA RACIAl IIOLITSEA SP LICITBIDACUK BAUIIGTOIIA GIGIJTOSTICB!A I.KT V. LICYTIIDACIAE BIRRIIGTOIIA RACBMOSIIL.ISPRKIG. LICITBIDACEA£ PLAICBOIIA VU.IDl IBL.l BL. LIBACUE LID IIDICA IBDRM.P ,) KERR. LYTRBACUB LlGDSTROillliA OVlLIPOLIA T. ' B. lllLVACUI BIIISCUS TILIACBUS L. lllLVACEAI TBISP!SIA POPILIIA(L, ISOLAID. U CORREA 274

....• Appendix II (cont.)

IIBLAmiiAtlCDI IIIIBCYLOI BDULI ROIB.VAR.OVAtllll(J.B.SIIItBICLABll! IIIWtCIIltlCDE IIIIBClLOI IIJRSIIOIDI!S BL. IIBLAStcllltlCIAE IIIIBClLOI SP IIILUCDI IIILUCDI AGL&Il BI!PtliDBA l. 1!t V. IIILUCDI lGLAIA SP IIILIACDI AIIOOBA CICCJLLitA BOJB. IIILUCDI lPilllliiiiS GlliDIPOLIA IBL. I lflLP. IIILIACDI DJSODLllll CAILOStlCBYllll IIIQ. IIlLI& CDI DtSODLIJII DIISIFLORDIIIBL.IBIQ. IIILIACIAI otsom• SP IIILIACDI DJSODLUII SP muca1 mu AIIDWCB L. mucw Slll!tDll KACIOPI1LLA IIIG IIOIIIIACDE IIBW COIIACD IBL.I DilL. EX ROOI.F. l tR<*S. IIOIACIAE llfOCAIPOS ILIB!ICUS RI!III.II BL. IIOIACIAI FICUS BIIJliiiA L. IIOIACDB FICIS Bllctll!A IALL. IIIIIQ. IIOIACDI FICUS DIIP.&CD mo. IIOilCIAI FICIS PlstBLOSl 11!111. Bl BL. IIOilCIAI FICIS GLCIII.&t.& IOJB. IIOilCIAI FICUS BISPIDl L.F. IIOIACDB FICliS IIILIIOC.&IPA BL. IIOilCDI FICIS OBSCUil BL.VAB.SCABIBBIIIl(BL.JCOBII!B IIOIACilB FICUS lleaiOSA L.VAR.BLOJGAtA(UICIBABRm IIOIACIAE PI CDS SIP!ICA 8811. F. IIOilCDI PICUS SIIDltl tBUJB, IIOilCill FICUS SP IIOilCDB FICUS SP IIOilCUI PI CIS tiictolll FORSr.G.SSP.GIBBOS!(BL.ICORIER IOIACDI FICUS SP IIOIACIAI PI CIS SP IIOilCDB FICUS SP IIOilCDB FICUS SP IOIACIAB FICUS SP IIOIACDI FICUS SUBILltl BL. IIOilCDI PI COS SliiAtBlll IllQ , IOilCDI FICUS VWIGltl BL. IIODCIAB POIIILOSPBIIII SIIVIOLDS(IL. )IBBB. IIOilCDI StiBBLUS lSPD LOUI. 111lstiCICDB IOISPIILDIA GL&Bil (BL. I lAD. 111IstiCICDI BOISPIILDIA 1111 (GADtl. I llRB. mistlCACDI IOISPIILDIA SP 11UlllCDI liDISU 1111ILIS VAIL mtlCDI! DIWPBIIIDII RlctlCOSII J .l.lG. fORSt. mtlelll BIGIIU LIIIltl (DC. I DutBIB mtlCDI! DGBIIA SP mtlCDI! BUGDil SP 275

..... Appendix II (cont.)

MYRTACEA£ IDGIIIA SP MYRTA CUB I!UGDIA SP MYRTACIA£ DGDIA SP MYR'rACIA£ IOGDIA ZIPPILIAIUII IIIQ. lllRTACBA£ PSIDIOM GUAJAVA L. MYRTACBA£ RBODAIIIIA CIIDU JACI IIYR'rACBAB SUYGIOII POLYAITBDI(IIGBTIKALP. OLIACBAE LIIOCII!RA RA!IlLORA(ROIB.)WALL. OLIACIA£ LIIOCIERA SP OLIACBAB LlG8stlllll SP OLBACIA£ OLEA JAVAIICA IBL.IIROBL. PAIDAIACKAB PAIDAJOS SP PIPBRACIA£ PIPER ADDCIII L. PIPBRACBAB PIPER SP PIT'fOSPORACIAE Pln'OSPOROII FIRI8GIII811 I. AIT. RBUOPIORAC£!1 GYIOTROCBI!S AIILWIS BL. RDBIACIAB liTIOCBPBALUS CADAIIBA III Q. RDBIAC£!1! AJTBOCI!PBALUS OBTUSA BL. VAR.IIlJOB RDBIACIAI! GDmHDA SPICIOSA L. RUBIACIAE 11011 PALDDOSA (BL. I 1011 RDBilCBll! 11011 SP ROBIACIA£ IIORA SP RIBIACDB IAUCLIA SP RBBIAC£!1 PlYmA IIDICA L. RDIACIAI PSYCBOTBIA SP RDBIACBAB RAIDIA LOIGIYLORA LAlli R8BIACBAB TARDIA FUGIAIS IBL.I I. rr V. RIBIACIAB TIICALYSIA SIIGDLARIS(IORTB.)I.SCB. In'A CIA I ACIOIYCBIA LADIPOLIA BL. RBTACBAB CLADSW EICAVATA 81111. F, BirACIAl I!DODIA AI

...•. Appendix II (cont.)

SUICDLIACEAE STERCULIA COCCIIBA JACK VAR.COCCIII!A STERCULIACI!AE STERCULIA OBLOIIGATA R.BR. TIDCIAE EURYA GLABRA (BL. ) IORTB, TBUCI!AE TBRISYROEIIIA PAfEIS(IORTR.ICBOISY TILUCEAB MICROCOS PAIICULAfA L. TILIACEAE SCBOUTDIA OVAtA IORTB. ULIIACEAB CELfiS PHILIPP!ISIS BLAICO OLIIACIAB CELfiS PBILIPPBISIS VAR. WRIGBTII IPLAICB.) SOEPADMO DRTICACEAE DnDIOCIIDE MICROSTIGMAIODD. ICHElf ORTICACEA! VILLIIBRDIBA RUBESCKIS (BL.) BL. VBRBDACEAB CLBIODDDRUM SERIATBM(L,)MOOI VEB.BDACEAE TBCfOIA GRAIDIS L.F. VDBWCEAB VIfll COFASSDS REIIW. EX BL. VERBDACEAE VITEI GLABBATA R.BI. VDBDACEAB VIfll PIIIATA L. VEB.BEIACEAE VIf!l POB!SCIIIS VABL VIOLACBAB RIIIOIBA SP VITACEAE TBfBASTIGIIA MUTABILE IBL.) PLAICI. VI!ACEAE fEfRASTIGIIA PAPILLOSUMIBL.)PLARCB. VI\' ACUI! TBfBASfiGIIA SP VI!ACEAB TBfRASTIGMA TRIFOLIATUM MKRR. 277

APPENDIX III SPECIES ENUMERATED IN THE VEGETATION SAMPLING OF THE GRP3 AND GRP21 HOME­ RANGE AREAS with Corresponding Frequency (number of stems per hectare) and Biomass (percentage basal area (b.a.)) a) GRP3

Family Genus Species freq b.a.

ANACARDIACEAE BOCHANANIA ARBORESCEIIS 5.66 0.19 ANACARDIACEAE MAIIGIFERA LONG IPES 1.88 0.01 ANACARDIACEAE SEMECARPOS HETEROPHYLLA 1.88 0.02 ANNONACEAE POLYALTHIA LATERIFLORA 3. 77 0.02 ANNOIIACEAE PSEUDUVARIA RUGOSA 18.86 0.33 ARECACEAE ARENGA OBTUSIFOLIA 7.54 0. 20 ARECACEAE DAEMONOROPS MELANOCHAETE 7.54 0.11 BIGIIOIIIACEAE DOLICBANDRONE SPATBACEA 1.88 0.19 BIGIIONIACEAE OROXYLUM INDICUM 1.88 0.02 BURSERACEAE CAllARI OM ASPERUM VAR.ASPEROM !. 88 0.13 BURSERACEAE CANARIUM HIRSDTUM 1.88 0.01 CASUARINACEAE CAS DAR INA EQUISETIFOLIA 1. 88 0.04 EBENACEAE DIOSPYROS CAUL IFLORA 20.75 0.52 EBENACEAE DIOSPYROS JAVANICA 16.98 0.69 EBENACEAE DIOSPYROS SP 3. 77 0.02 EBENACEAE DIOSPYROS TRDNCATA 13.20 0.16 EUPHORBIACEAE ANT !DESMA BUN! OS 3.77 0.73 EUPHORBIACEAE BRIDELIA MONOICA 7.54 0.09 EUPHORBIACEAE CLAOXYLON PO LOT 5.66 0.03 EUPBORBIACEAE GLOCHIDIOII SP 1.88 0.01 EUPHORBIACEAE MALLO'!' US MORITZIANUS 9.43 0.14 EUPHORBIACEAE MALLO'!' US MDLTIGLAIIDULOSA 1.88 0.06 EUPHORBIACEAE MALLOTUS OBLONGIFOLIUS 16.98 0.19 EUPHORBIACEAE MALLO'!' US SP 1.88 0.01 EUPHORBIACEAE MALLO'!' US TILIAEFOLIUS 1.88 0.02 EUPHORBIACEAE NEOSCORTECHINIA KING II 3.77 0.03 FABACEAE ACACIA AURICULIFORMIS 18.88 1.11 FABACEAE DESKODIUM OMBELLATOM 5.66 0.04 FLACOURTIACEAE CASEARIA GREWAEFOLIA 11.32 0.40 FLACOURTIACEAE SCOLOPIA SPINOSA 7.54 0.05 HERNANDIACEAE HERNAIIDIA PELTATA 5.66 1.18 LAURACEAE CINNAMOMUK INERS REIHl'/ 5.66 0.03 LAURACEAE NEOLITSEA CASSIA 1.88 0.04 LADRACEAE NEOLITSEA SP 1.88 0.32 LECYTHIDACEAE PLAJICHONIA VALIDA 1.88 0.01 LYTHRACEAE LAGERSTROEMIA OVALIFOLIA 1.88 0.03 MALVACEAE HIBISCUS TILIACEUS 1.88 0.05 KELASTOMATACEAE KEMECYLON EDDLE VAR.OVATUM 7.54 0.33 MELAS'!'OMA'!'ACEAE MEMECYLON MYRSIIIOIDES 16.98 0.24 MELIACEAE 1.88 0.04 MELIACEAE AGLAIA SP 1. 88 0.02 MELIACEAE DYSOXYLDK CADLOSTACHYOM 26.41 1.09 MELIACEAE DYSOXYLUM DENSIFLOROH 1.88 0.29 MELIACEAE MELIA AZEDARACH 11.32 0.73 MELIACEAE SWIETENIA MACROPBYLLA 115.09 19.98 MONIMIACEAE KIBARA CORIACEA 1.88 0.02 MORACEAE ARTOCARPUS ELASTICUS 1.88 0.04 MORACEAE FICUS GLOMERATA 1.88 6.65 278

..... Appendix III (cont.)

Family Genus Species freq b.a.

MORACRAE FICUS HISPIDA 3. 77 0.02 . MORACRAR FICUS SEPTICA 1.88 0.02 MORACRAE FICUS SIRUATA 1.88 28.86 MORACEAE FICUS SP 3. 77 0.14 MORACEAE FICUS SUMATRARA 1.88 13.26 MYRTACEAE DECASPERMUM FRUCTICOSUM 3. 77 0.12 MYRTACEAE EUGERIA LIHEATA 3. 77 0.09 MYRTACEAE EUGERIA ZIPPRLIAHUM 3. 77 0.03 OLEACEAE OLEA JAVARICA 1.88 0.06 PARDARACEAE PAliDAHUS SP 1.88 0.02 PITTOSPORACEAE PITTOSPORUM FERRUGIRRUM 1.88 0.01 RUTACEAE CLAUS ERA EXCAVATA 41.50 0.99 RUTACEAE EUODIA AROMATICA 5.66 0.17 SAPIHDACEAE BARPULLIA CUPAliiOIDES 3. 77 0.20 SAPIRDACEAE HEBECOCCUS PBRRUGIREUS 1.88 0.01 SAPIRDACEAB LBPIDOPBTALUM SP 1.88 0.03 SAPOTACBAE PLARCHOliBLLA LIRGGERSIS 1.88 0.08 STBRCULIACEAE HBRITIERA LITTORALIS 1.88 3.09 STERCULIACBAE KLBiliBOVIA HOSPITA 11.32 5.93 STBRCULIACEAE PTEROSPERMUM JAVARICUM 11.32 1. 70 THEACEAE EURYA GLABRA 1.88 0.02 TILIACEAE MICROCOS PARICULATA 3. 77 0.13 ULMACEAE CELTIS PHILIPPEliSIS 5.66 0.24 DRTICACEAE DEliDROCHIDE MICROSTIGMA 5.66 0.05 URTICACEAE VILLEBRUNEA RUBESCERS 1.88 0.04 VERBENACEAE TECTORA GRAND IS 103.77 7.56 VERBERACEAE VITEX GLABRATA 1.88 0.07 VERBERACEAE VITEX PIRRATA 5.66 0.32 VERBERACEAE VITBX PDBESCENS 5.66 0.04 279

..... Appendix III (cont.)

b) GRP21

Family Genus Species freq b.a.

AHACARDIACEAE BDCHARARIA ARBORESCERS J3.i5 0.82 ANACARDIACEAE MAHGIFERA LONG IPES 1.25 0.01 ANACARDIACEAE SEMECARPDS HETEROPHYLLA 1.25 0.02 ANACARDIACEAE SPONDIAS CYTHEREA 3. 75 0.15 ANHONACEAE 2.50 0.07 ANNONACEAE POLYALTHIA LATERIFLORA 6.25 0.16 APOCYHACEAE CERBERA MAN GRAS 2.50 0.10 APOCYHACEAE URCEOLA BRACHYSEPALA 2.50 0.04 BOMBACACEAE BOMBA X VALETOHII 6.25 1.00 BDRSERACEAE CANARIDM CF. INDICDM 1.25 0.01 BDRSERACEAE CAHARIUM HIRSDTUH 10.00 0.95 BDRSERACEAE CANARIDM LITTORALE 1.25 0.01 CLDSIACEAE GARCIRIA DULCIS 1. 25 0.17 COHHARACEAE AGELAEA MACROPHYLLA 3. 75 0.16 COHNARACEAE CHEST IS PLATANTHA 1.25 0.01 EBENACEAE DIOSPYROS CADLIFLORA 3.75 0.07 EBEHACEAE DIOSPYROS JAVAHICA 52.50 1. 21 EBEHACEAE DIOSPYROS MALABARICA 1.25 0.01 EBENACEAE DIOSPYROS SP 2.50 o. 72 EBEHACEAE DIOSPYROS SP 1.25 0.21 EDPHORBIACEAE ANTIDESMA BUN IUS 7.50 0.13 EDPHORBIACEAE ANT !DESMA MONT ANUM 1.25 0.01 EDPHORBIACEAE ANT !DESMA SP 2.50 0.08 EDPHORBIACEAE APOROSA MICROSPHAERA 2.50 0.81 EUPHORBIACEAE BREYNIA VIRGATA 2.50 0.03 EDPHORBIACEAE BRIDELIA MONOICA 5.00 0.14 EDPHORBIACEAE CLAOXYLOH PO LOT 2.50 0.03 EUPHORBIACEAE MALLOTUS MORITZIANUS 3.75 0.17 EDPHORBIACEAE MALLOTUS OBLONGIFOLIUS 10.00 0.07 EDPHORBIACEAE MALLOTDS PHILIPPENSIS 3.75 0.13 EDPHORBIACEAE MALLOTDS SP 16.25 0.17 EOPHORBIACEAE MAL LOTUS SP 1.25 0.01 EDPHORBIACEAE MALLOTDS TILIAEFOLIUS 6.25 0.07 EOPHORBIACEAE NEOSCORTECHINIA KING II 1.25 0.01 EDPHORBIACEAE SADROPDS ANDROGYHUS 1.25 0.03 EDPHORBIACEAE SUREGADA GLOMERATA 3.75 0.02 FABACEAE CASSIA JAVAHICA 1. 25 0.13 FABACEAE CASSIA SIAMEA 17.50 0.42 FABACEAE DESMODIUH DMBELLATDH 3.75 0.02 FLACOURTIACEAE CAS &ARIA FLAVOVIRENS l. 25 0.01 FLACODRTIACEAE CASEARIA GREWAEFOLIA 6.25 0.04 FLACODRTIACEAE CASEARIA TUBERCULA fA 13.75 0.13 FLACOURTIACEAE FLACODRTIA IHERMIS 15.00 0.48 FLACODRTIACEAE SCOLOPIA SPINOSA 23.75 0.70 HERRAHDIACEAE BERNARDI A PELTATA 16.25 21.16 LADRACEAE BEILSCHMIEDIA SP 1.25 0.01 LADRACEAE CIHHAMOMUH INERS 35.00 1.84 LADRACEAE HEOLITSEA CASSIA 7.50 0.21 280

..... Appendix III (cont.)

Family Genus Species freq b.a.

LADRACEAE NEOLITSEA SP 1.25 0.01 LECYTHIDACEAE BARRINGTONIA GIGANTOSTACHYA 1.25 0.02 LECYTHIDACEAE PLANCHONIA VALIDA 1. 25 0.01 LEEACEAE LEEA INDICA 11.25 0.07 LYTHRACEAE LAGERSTROEMIA OVALIFOLIA 1. 25 0.06 MALVACEAE THESPESIA POPULNEA 3.75 1. 93 MELASTOMATACEAE MEMECYLON EDULE VAR.OVATUM 2.50 0.03 MELASTOMATACEAE MEMECYLON MYRS INO IDES 12.50 0.24 MELASTOMATACEAE MEMECYLON SP 2.50 0.36 MELIACEAE AGLAIA HEPTANDRA 2.50 0.02 MELIACEAE APHANAMIXIS GRANDI FOLIA 12.50 3.89 MELIACEAE DYSOXYLUM CAULOSTACHYUM 68.75 0.62 MELIACEAE MELIA AZEDARACH 3.75 0.67 MORACEAE ARTOCARPUS ELASTIC US 2.50 6.14 MORACEAE FICUS BERJAMINA 2.50 0.03 MORACEAE FICUS FISTULOSA 6.25 10.39 MORACEAE FICUS SUMATRANA 11.25 24.73 MORACEAE FICUS VARIEGATA 1.25 1.17 MORACEAE POIKILOSPERMUM SUAVEOLENS 1.25 0.01 MYRISTICACEAE HORSFIELDIA GLABRA 1. 25 0.03 MYRISTICACEAE HORSFIELDIA IRYA 2. 50 0.03 MYRSIHACEAE ARDISIA HUMILIS 3.75 0.04 MYRTACEAE DECASPERMUM FRUCTICOSUM 12.50 0.15 MYRTACEAE EUGENIA LIHEATA 8.75 0.37 MYRTACEAE EUGENIA SP 15.00 0.47 MYRTACEAE EUGENIA SP 3.75 0. 20 MYRTACEAE EUGENIA SP 2.50 0.02 MYRTACEAE EUGENIA SP 1. 25 0.01 MYRTACEAE EUGENIA ZIPPELIANUM 25.00 1.04 MYRTACEAE PSIDIUM GUAJAVA 1. 25 0.01 MYRTACEAE RHODAMNIA CINEREA 6.25 0.05 OLEACEAE LINOCIERA RAM IF LORA 3.75 0.04 OLEACEAE LINOCIERA SP 12.50 0.42 OLEACEAE LYGUSTRUM SP 26.25 0.52 PANDAHACEAE PANDANUS SP 3.75 0.07 PIPERACEAE PIPER AD UN CUM 1.25 0.02 RHIZOPHORACEAE GYHOTROCHES AXILLAR IS 41.25 0.49 RUBIACEAE GUETTARDA SPECIOSA 7.50 0.75 RUBIACEAE IX ORA PALDDOSA 1.25 0.02 RDBIACEAE IX ORA SP 1.25 0.03 RDBIACEAE PSYCHOTRIA SP 1.25 0.01 RUBIACEAE TAREHHA FRAGRAHS 3.75 0.07 RUBIACEAE TRICALYSIA SINGULARIS 11.25 0.43 RDTACEAE ACRORYCHIA LAURIFOLIA 2.50 0.05 RUTACEAE CLAUSENA EXCAVATA 15.00 0.15 SAPINDACEAE ALLOPHYLUS COBBE 1. 25 0.03 SAPINDACEAE ARYTERA LITORALIS 6.25 0.03 SAPIHDACEAE GAROPHYLLUM FALCATUM 1.25 0.01 281

..... Appendix III (cont.)

Family Genus Species freq b.a.

SAPINDACEAE HARPULLIA CUPANIOIDES 1.25 0.35 SAPINDACEAE HEBECOCCUS FERRUGINEUS 2.50 0.04 SAPINDACEAE LEPISARTHES TETRAPHYLLA 1.25 0.01 SAPINDACEAE MISCHOCARPUS SONDAICUS 33.75 0.38 SAPINDACEAE POMETIA PIHRATA 3.75 0.03 SAPOTACEAE PLANCHONELLA LINGGENSIS 1.25 0.01 STERCULIACEAE PTEROSPERMUM DIVERSIFOLIUM 10.00 0.11 STERCULIACEAE PTEROSPERMUM JAVANICUM 25.00 0.36 STERCULIACEAE STERCULIA COCCIREA VAR.COCCIREA 6.25 2.85 STERCULIACEAE STERCULIA OBLONGATA 1.25 0.10 TILIACEAE HICROCOS PANICULATA 8.75 2.70 ULMACEAE CELTIS PH !LIPPENS IS 5.00 0.32 ULMACEAE CELTIS PHILIPPERSIS VAR. WRIGHTII 2.50 0.01 URTICACEAE DENDROCNIDE MICROSTIGHA 5.00 0.12 URTICACEAE VILLEBRUNEA RUBESCENS 2.50 0.08 VERBENACEAE TECTONA GRAND IS 1. 25 0.12 VERBENACEAE VITEX PINNATA 7.50 1.12 VERBENACEAE VITEX PUBESCENS 17.50 3.04 VITACEAE '!'ETRASTIGMA SP 1.25 0.05 282

APPENDIX IV Chemical Measuresa (as % Dry Weight) of Vegetation in the GRP3 and GRP21 Home-Range Areas (a) Analysis of Mature Leaves

1. GRP3

Species N PROT ADF CDIG CT TP pp % basal area

Ficus sinuata 1.7 10.63 37.79 52.86 0.39 1.36 0.84 28.86 Swietenia macrophylla 2.0 12.19 26.34 60.21 8.98 11.65 11.05 19.98 Tectona grandis 2.8 17.50 46.72 48.46 0.08 1.40 1.31 7.56 Kleinhovia hospita 3.8 23.75 25.87 45.79 1.32 1.69 0.48 5.93 Heritiera littoralis 1.6 10.00 46.46 22.66 9.06 9.79 10.04 3.09 Pterospermum javanicum 2.1 12.82 43.66 11.27 7.55 5.04 5.80 1.70 Hernandia peltata 2.3 14.38 28.01 61.80 0.26 1.65 2.44 1.18 Acacia auriculiformis 2.7 16.57 35.53 28.40 4.79 8.33 6.78 1.11 Dysoxylum caulostachyum 3.1 19.38 24.50 47.15 0.54 4.20 1.17 1.09 Clausena excavata 2.9 18.13 20.48 51.36 0.74 2.65 0.95 0.99 Eugenia zippelianum 1.2 7.50 40.07 21.37 0.91 13.17 7.86 0.03 Ficus benjamina 1.9 11.57 52.18 31.93 0.68 0.80 0.31 <0.01 TOTAL 71.52

2. GRP21

Ficus sumatrana 1.4 8.75 50.57 29.45 0.71 1.99 1.89 24.73 Hernandia peltata 2.4 15.00 19.12 72.69 0.27 1.49 1.27 21.16 Artocarpus elasticus 0.8 5.00 48.79 24.74 0.32 4.33 3.61 6.14 Aphanamixis grandifolia 2.0 12.19 27.32 53.17 0.00 1.13 0.85 3.89 Sterculia coccinea 2.0 12.19 33.13 47.82 8.59 6.26 5.68 2.85 Microcos paniculata 2.3 14.07 39.19 34.57 3.76 6.37 4.20 2.70 Dysoxylum caulostachyum 3.3 20.32 28.42 56.43 0.09 1.29 0.40 0.62 Ficus benjamina 1.5 9.04 51.70 47.49 0.42 1.10 0.54 0.03 Ficus sinuata 1.5 9.07 43.13 45.85 0.15 0.00 0.80 s.{LQl TOTAL 62.12 283

...... APPENDIX IV (cont.)

b) Analysis of Young Leaves, Fruits and Flowers

1. Young leaves

Group Species N Prot ADF CDIG cr TP pp % feeding time

GRP3 Tectona grandis -midribs 1.6 9.69 42.70 56.36 0.04 0.75 0.39 12.80 Tectona grandis -leaf lamina 2.9 18.13 35.91 52.37 0.75 6.31 1.32 1. 33 Ficus glomerata 1.6 9.69 27.27 39.92 0.48 1.78 1.17 1.45 Kleinhovia hospita 3.8 23.75 25.87 45.79 1.32 1.69 0.48 1.19 GRP21 Guettarda speciosa 1.6 10.00 21.87 58.18 1.10 6.32 5.06 6.54

2. Fruits

GRP3 Clausena excavata 2.4 15.01 24.81 47.65 0.95 3.22 Pterospermum javanicum 1.0 6.25 53.42 17.66 1.78 2.47 4.35 2.99 Kleinhovia hospita 2.2 13.75 48.79 27.18 0.85 1.34 0.34 0.13 Acacia auriculiformis 2.0 12.50 43.04 35.04 3.44 6.15 4.85 0.07 Ficus benjamina 1.5 9.07 61.71 23.52 0.09 0.00 0.24 0.00 GRP21 Dysoxylum caulostachyum -seeds 2.8 17.50 14.69 45.86 0.07 0.60 0.27 9.66 Hernandia peltata 1.3 8.13 39.38 60.73 0.63 3.74 Ficus benjamina 0.8 5.00 35.42 26.88 9.24 8.84 12.20 2.14 Sterculia coccinea -seeds 3.3 20.63 15.54 33.97 25.90 20.61 29.12 0.12 Dysoxylum caulostachyum -peri carp of ripe fruit 1.9 11.88 45.38 22.23 0.04 0.27 0.33 0.00 Dysoxylum caulostachyum -unripe fruit 2.1 13.13 40.11 34.90 0.10 0.55 0.37 0.00 Ficus sumatrana 1.1 6.88 61.07 35.84 1.66 3.58 2.79 0.00 Sterculia coccinea -peri carp 1.3 7.82 31.99 3.10 6.23 2.91 0.00 284

••••• APPENDIX IV (cont.)

3. Flowers

Group Species N PROT ADF CDIG cr TP pp % feeding time

GRP3 Kleinhovia hospita 2.4 15.00 39.43 35.72 0.93 2.74 0.71 5.61 Swietenia macrophylla 1.5 9.07 50.31 27.73 ·4.54 4.23 4.92 0.00 GRP21 Hernandia peltata 3.0 18.75 34.25 59.29 0.04 0.58 0.29 0.00

a N = nitrogen CT = condensed tannins PROT = protein TP = total phenolics ADF = acid detergent fibre PP = protein precipitation CDIG = cellulase digestibility . 285

APPENDIX V BREAKDOWN OF TRACHYPITHECUS AURATUS DIET (as unweighted % feeding time) to the Level of Food Part a) GRP3

Family Genus Species Item Part %

AIIOIACBA! CAIAIGA ODOIATA. BU IP .041 AJIOIAC!A! CAWGA ODOR! fA L 8 .145 AIIOilCBA! CWIGl ODOUfA L A .036 BIGIOIIAC!A! DOLICBAIDIOI! SPAHACBA Bl s .145 BIGIOIIACIAI RADBIIIIACIIRA GIGHTBA L!3 8 .168 BIGIOIUC!AB RADBRIIACIW GIGHTBA L8 liP .254 BIGIOIUCBA! lAD DillCiliA GIGm!A LY 8 .021 BIGIOIUC!AB RADDIIACIDA GIGm!A L!l B .061 BIISWCBAI CAIHIOM BIRSftlll Lll B .072 CLDSUC!A! GAICUU BALICA LY2 8 .036 CLISI!CU! GAICIIU BALICA LY3 8 .036 CLISUCBA! GAlCiliA SP Lt2 B .145 CLISI!C!AE GABCIIIA SP L IP .us COIIli!CBAI ROIIBA 111101 LY 8 .us BBDACIAB DIOSPYBOS JAVAIICA FB liP .145 !BIDCBA! DIOSPYROS JAVDICA LYl B .061 DIIACBA! DIOSPIIOS JAVAIICA Ul B .872 DDICW DIOSPIROS SP L!l 8 .114 BBIIACIAB DIOSPYBOS SP LY2 8 .083 DIDCIA! DIOSPYBOS SP Ltl 8 .914 IBIDCBAB DIOSP!IOB ftUICAfA L 8 .967 DDICIAB DIOSPIBOS fBDICAfA L!2 B .836 IBIDCBAB DIOSPIROS TRDJCAfA Lll 8 .us llBIDCIAK DIOSPIBOS fRIICAfA Ll4 8 .109 !LAKOCWACBAB ILAmcHPIS GLABD Ll2 8 .145 IL&IOCliPACBAB !LABOC!RPDS GLABBR LU 8 .811 nPIOIBIACBAK AlfiDBSIIA 8DIIDS Ltl 8 .036 DPIOIBIACBAB Al'riDBSIIA BOlDS Lt2 8 .036 IIPIOUIACBAB AlfiDISIIA NOifAIIIl Ll2 8 .072 DPIOIBIAC!AB AlflDISIIA SP FB IP .012 DPIOIBIACBA! AlfiDBSMA SP Lt2 B .194 BDPBOIBUCBAB AlfiD!SIIA SP LU B .109 DPIORBIACBAK Al'rlDBSIA SP Ltl B .072 DPIODIAC!AK UfiDISIIA SP Ul IP .036 DPIOIBIACIAK BACCADBIA JAVAIICA 81 w .218 BUPIOIBIACBAK BACCADIBA JAYUICA Bl II .087 DPBOUIACBAI BACCAIW JAVUICA LYl B .024 IIPIOIBIACIA! BACCADIBA JAVAIICl Ll2 B .136 IIPIOBBIACIAB BACCADBBA JAYAIICl LY3 8 .868 DPIORBIACIAI BACCAUIBA JAVUICl PB liP .133 KIPIOBBIAC!AI BIIDKLIA IIOIIOICl L A .121 KDPIOIBIACBAB BIIDBLIA IIOIOICA 88 w .296 IUIOIIIACUI BIIDKLIA IIOIOICl LY2 8 .264 KDPIORBIACBAB BRIDILIA IIOIOICA Ll3 B .436 DPIOIBIACIAB CLAODLOIU POLOT Lll A .145 BDPBOIBIACBAI CLAODLOII POLOf L A .036 BBPIORBIACIA! CLAOIILOI POLO'I' Ltl 8 .072 BBPBOBBIACIAB CLAODLOII POLOf L 8 .887 286

•.••• Appendix V (cont.)

Family Genus Species Item Part %

BIPIOIBIACIAI CJ.AODLOI POLOt Lll BP .018 IUI'IOIBIICIAI CJ.AODLOI PO LOt Ll2 BP .091 DPIOIBIACIAI CL&ODJ.OI POLO! LYl BP .ou IDPIOUilCill CLAODLOI PO LOt Ll B .065 IUPIOUIACIAI IIALLOfDS IOII!ZIUBS L12 B .036 IDPIOIBilCill IIWO!IS IIUL!IGLUDBLOSl BU If .072 BIPIOIBIACIAI IIALLO!US llaL!IGLUDOLOSl L12 B .157 IIPIOIBIICIAI IIALLOfDS IIDL!IGLUDDLOSl Lll B .203 llfiOIIIlCIAI IIALLO!DS IIIL!IGLAIDDLOSA Lll A .036 IDPIOIBilCill IIALlOIIS lliLtiGLUDILOSl LU A .087 IIPIOUIICIAB IIALLO!DS IIL!IGLAIDDLOSA L B .121 IUPIOIBIICIAI IIALLO'liS IULtiGLUDOLOSI L12 II .048 IIPIOIIIICIAB IIALLO'I'IS IIILtiGLAIDILOil LYl II .048 DPIOUilCill IIALLOfliS IULtiGLUDDLOSA II IP .109 DPIOIBIACIAI IIALLO!OS IIILtiGLUDOLOSA Lll B .072 IIPIOIBIICIAB IIILLO!U PBILIPPDSIS p IP .194 IIPIORBIACIAB IIALLOfiS IICIIOIDIS Pl IP .145 IUPBOUIACIAB SDIIGADl GLCIIIA!l Lll B .078 PWCill ACACIA lRICBLIFOIIIIS p IP .133 PWCIU lCACU lOIICOLIFOIIIIS PB IP .200 PWCIAI lCACU AOIICULIPOIIIIS LY2 B .045 PWCill ACACU ADIICOLIPOIIIS Lll B .019 Pli&CIAI ACACIA AOIICULIFOIIIIS 81 IP .065 PWCIAI ACACIA AIUCOLIPOIIIS 88 IP .024 PWCIAI ACACIA AOIICULIPOIIIS 88 s .731 FWCIU ACACIA A81ICOLIPOIIIIS Ltl II .425 PABACIAI ACACIA A81ICILIPOIIIS LB IP .178 PWCIAI ACACil AIIICOLIPOIIIS Ll B .024 FWCIAI ACACU AIIICULIPOIIIS Lll B .029 FWCIAI ACACIA AOUCOLIPORIIS Ll2 II .093 PABACIAB ACACIA AOICULIPOIIIS Pl IP .169 PWCIAB ACACIA ARICULIPOBIIS Ll4 II .lt4 PWCIAB wsu SWIIA LYl B .Ott PAIICIII CASSU SUIIIA LU B .114 PWCIAB CASSIA SIJIIIl Ll3 B .011 PWCIAI CASSIA SUIIIA LD B .182 PWCIAI e1DmA DIIIPLOU II IP .Ott FD&el&l CIIDfll UIIFLOU BD I .849 PWCIAI ClaltiA UIIFLOIA LB IP .291 PWCIAI CIICIII!Il UIIPLOIA Lll BP .012 PWCill DILIIIGIA LAtiPOLU LU B .343 PAIICI&I DALBDGU LAtiroLU Lll B .242 PABACIAI DALIIIGil LA!IPOLU Lt4 B .176 PII&Cill DALBDGU LAtiPOLil 88 s .145 PWCIAI DALBDGIA LA!IPOLU II s .119 PWCill DALBIIGU LAflfOLU Lll B .109 PWCIAI DILBDGll LAtiPOLU Lt B .728 PUlCill DILBDGil LltiPOLU Lt PS .812 287

..•.. Appendix V (cont.)

Family Genus Species Item Part %

FWCBAE DALBDGIA LATIFOLIA LY2 A .036 FWCIAI DALBDGIA LATIPOLIA LY3 A .072 FABACIIE DALBIIGIA LA'J'IFOLIA Ll liP .145 PWCIII DALBDGIA LA'J'IFOLIA LM B .024 FWCIAB DALBDGIA LA'J'IFOLIA III liP .072 FABACIIB DALBDGIA LA!I FOLIA L liP .072 FABACBAB DnDIIA VARIBGATA BD If .824 FABACIAI ERt'tBRIIA VARIBGA'J'A B s .084 FABACIIB ERmiiD VAIIBGA'J'A m B .112 PABACW ER!TDID VARIBGATA LU 8 .449 PABA CIA I ER!TBRID VAIIBGATA L12 A .016 FWCBAB ERtmiD VlliiGATA LYl A .016 PWCIIE BmiiiiA VAIIIGATA LY3 B .344 rwcw D!'tBRID VAIIIGATA II IP .012 PABACIIE ERtmiiA VAIIBCATA LY4 B .024 FIBACIAB PBAIERA FILVA F IP .291 FLACOURTIACBAB C.UDIIA TU81RCDLATA L!1 B .072 IIIIUDIACIIB IIIIUDIA PBL'fATA BD .824 IIWIDIACIAB IIWIDIA PIL'fATA L •IP .048 IBUliDIACIIB IIIIUDIA PBL'fA'fA 81 F .036 IISBCH* II IP .809 LAIIACIIE CIJIAII(Illll IIBIS REID L!2 8 .163 LA RACIAl CIIJAII(H( liDS Rllft L!2 A .843 LABIA CIA! CIJIAI(OI IIBIS RIIB PB IP .145 LA IRA CIA I CIIJAII(BI IIBRS RBift BU .965 IIALVACW HIBISCUS TILUCBU LY4 •B .048 IIALVACIAB IIBISCliS TILIACBDS PB IP .200 lilLY& CIA I !IBISCliS ULUCBUS Lit B .811 IIILIACIAI DtSOmlll CAIJ.OS'flCIIII L p .072 IIILUCDI D!SOfiLBII CAILOSTACBt• F IP .ooa IIILIACDI DISOfiLll CADLOSTACBYIII L IP .072 ULUCDI D!SOfiLIIII CAILOSTACB!OII PB IP .097 IIILIACW D!SOmlll CULOS'fAelllJII DR s .072 ULIACDB DISOfiLIII CAILOS'f!CI!lll 81 s .582 IILUCIAI DtSOmllll DDSIFLOIII 81 s .491 IIILIACIAE ULU AIIDWCI 80 .006 IIILIACDB Sllft'DIA llCIOHlLLA L •BP .072 IIILUCDI Sllft'DIA llCIOPI!LLA Lll B .131 IIILUCDI Sllft'DIA lllCIOPDLLA Ll2 8 .182 IlLIA CDI Sllft'DIA lllCIOPIILLA Ll2 A .062 IILIACDI Sllft'DIA IIICIOmLLA L!l A .031 IILUCDI Sllft'DU llCIOPOLLA Ll SL 1. 72 IIILIACIAB Sllft'DIA MCIOHlLLA Ll 8 .072 IlL UClA I Sllft'DU IACIOPIILLA Bl s 1.23 IILIACDI SIII'fiiiA IIACIOPI!LLA L A .872 IOUCBAB ARTOCARPUS ILAS\'ICDS Ll3 B .032 III&Cill FICUS BIIJAIIIA Ll IP .826 IORACIAI PI CIS BDJAIID Bll If .433 288

..... Appendix V (cont.)

Family Genus Species Item Part %

MOUCI&B FICUS BDJliiiiA LY2 B .072 IIOIACIAE FICBS DRBPICI& BR .436 llOIICI&I FICIS FIS!DLOSA LY2 "B .449 IIORICI&B FICUS FISTOLOSA LU B .024 MOIACI&I FICIS FIS!ILOSI LJ1 B .182 MOD CUI FICUS GL

...•. Appendix V (cont.)

Family Genus Species Item Part %

MOI&CIAI FICUS SP LY IP .873 IIOilCill FICIS SP LY B .145 ltORACIAI FICUS SP LB liP .587 IIOIACIAI FICUS SP L liP .072 MOUCIAB FICUS SP LYl B .036 IIOIACIAB FICUS SP LYl IP .165 llOUCilB FICIS SP LY2 B .072 IIOIACIAB FICUS SP LC IP .145 MOUCIAB FICUS SDILATA L 8 .036 ltOUCIAB FICUS SDBILATA LY2 B .139 IIOIACDB FICUS SOBDLATA LY3 B .024 ltOUCilB FICUS SIIIA!RW BD s .072 MOIACIAB FICUS SilltWA 8D If .218 MOIACIAB FICUS SIIIATRW LY2 8 .072 IIOIACill FICUS SIIATIW LYl B .072 IIOUCill FICBS SilltWA 8 If .364 MORA CIA! FICUS SIIIATUJA BR IP .519 MOll WI FICUS SillTWA 8R If 1.68 MOUCill FICUS SlllftW lB liP .218 IOU WI FICUS SilltWA II IP .121 miSTICACBAI IOISPIELDIA SP II IP .072 llll!lCill DBCASPUIUifl FIDCTICOSIII Bl If .291 KYI!lCill IDGIIIA SP F IP .824 IIIRTACill IUGIIIA IIPPBLUIUII BD If .278 llll!lCill IUGDIA IIPPILIAIIII 88 s 2.33 DI!ACill BDGDIA IIPPBLIAIIII p liP .osa mtACIAI ncniA UPPBLIAIIII lB IP .234 IIYITACBAI SlUGIIJII POLUITIIII LYl B .036 OLBACIAB LIIOCIBRA UllrLOU BR If .072 OLIACW LIIOCIW IWIIFLOIA LJl A .145 OLIACBAI LIIOCIERA WIIFLOU LYl B .123 OLBACIAB LIJOCIW WIIFLORl LY4 B .osa RIBIACIAI AI!BOCIPBALDS CADIIIBA lB liP .631 IIIIACW UTIOCBPIIALDS CADAIBA F BP .716 RDIACUB GIITTARDl SPICIOSA LY2 B .311 IOBIACBA! GBmARDA SPICIOSA LY3 B .457 IDIACIAB CliftAIDA SPICIOSA lB IP .066 RIBIACill GO mARDI SPICIOSA LU c .033 BIBIACDB GIITTAIDA SPICIOSA LU c ;019 IBIACU! GIITTAIDl SPICIOSA LJl B .041 IIBIACBAI GlmARDA SPICIOSA LY4 B .012 RDIACilB BAIDIA LOIGIPLOBA L!2 8 .072 RIIIACBAB TIICAL!SIA SIIGJLAIIS L A .242 IDilCill HICAL!SIA SIJGOLAIIS LU B .211 IDIACU! TIICAL!SIA SIIGILUIS LJ p .us RIBIACIAB THCAL!SIA SIIGILDIS p liP .203 IIBIACill TIIClLYSil SIICUWIS LY2 B .072 RDIACW ruCALYSil SIIGULAIIS BD If .145 290

..•.. Appendix V (cont.)

Family Genus Species Item Part %

RRUCIAI TRICALYSIA SIIGILAIIS II IP .048 ROBIACDI TRICALISIA SIIGOLARIS BD liP .018 ROBIA CUB TRICALYSIA SIIGILAIIS LY NP .072 ROBIA CUB TRICALYSIA SIIGDLAIIS LY B .024 ROBIA CUI ftiCALYSIA SIIGVLABIS LY4 B .018 ROTA CUB ACIOIICIIA LAOIIPOLIA BO s .218 RllfACDB CLAUSW BIClYAfl BR liP .109 RUTACDB CLABSDA BICIVATA BR I .227 RltACDI CLADSW BIClVAfl LY IP .043 IUI!ACBAI CLAISDA BICIVATA 80 II .145 RllfACBAI CLABSDA BieAVATA LU B .109 ROTACDI CLAOSDA BICIVATA 88 liP .012 RltACIAI CLAUW IICIVATA LB JP .072 RUTACIAI CLI8SDA BICIVATA BR p 2.57 RUTACIAB CLADSDA BICIVATA BR s .307 RllflCIAB CLADSDA IICIVA'fl F IP .018 RB'flCIAI CLADSDA BICIVATA BB p .436 SlPIID&Cill WPOLLIA CIPAIIOIDBS Lll B .169 SlPOTACIAB PLUCIOIILLA LIIGGDSIS Lll B .029 WO!'ACBAI PLUC!OIBLLA LIIGGDSIS LY2 B .087 SlPOTACIAB PLUCIOIILLA LIIGGDSIS L!l B .029 SAPOTA CIA I PLAICIOIBLLA LIIGGDSIS LU A .145 SlPOTACBAI PLUCIOIILLA OIIOVI'fA LY2 B .286 SlPOTACW PWCIOIILLA OIIOYATA PI IP .400 SAPOTA WI PLUCIOIILLA OBOYATA L8 liP .364 SAPOTICIAB PLUCIOIBLLA OBOVITA Ltl B .072 SfDCBLilCDI BDI'l'IBRA LlftOULIS II IP .072 SfiiCILilCDB BDHIBIA LITTOULIS PA IP 1.25 SftiCULilCDB BDI'fiBD LITTORALIS Ll4 8 .010 SflleDLilCDB BDHIBRA LITTORAL IS LY2 B .014 SHICILIACIAB ILIIIBOYIA IOSPITA Bl IP .091 STDCULIACDB ILIIDOYil BOSPITA Lll B .111 Sfam.JlCDB ILIIDOYil IOSPITA Ll4 B .018 SfDCIIdlCDB ILIIIBOYIA IOSPI'fA Ll B .836 SftiCILilCIAI ILIIDOYIA IOSPI'l'A II IP .269 SfDCDLilCDB lLIIIIOVIl BOSPU'A BR I .036 staCILUCDI ILBIIBOYIA IOSPI'l'A 81 liP .072 SfDCDLilCBAB ILIIDOYIA IOSPI'fl FA liP 4.82 S!IICILUCIAI ILIIDOYIA IOSPITA r IP .131 SfDCDLilCIAB ILIIJIOYIA IOSPHA FB liP .645 SfiiCILilCIAB ILUDOVIA IOSPI'fl 81 I .621 STIICBLIICIAB ILIIIIOYIA IOSPHA L B .148 S!IICILUCIAI ILIIDOYIA IOSPI'l'A Lll p .036 S'liiCILilCI&I ILIIIIOYIA IOSPI'fl LY2 p .326 SfDCILilCIAB ILIIIIOYIA IOSPITA LY4 p .850 staCILilCBAI ILIIDOYIA IOSPI'fl Ll p .148 SHICILIICIAI ILIIJIOVIA BOSPI'fA Ll p .169 SfDCILilCIAI ILIIIIOVIA IOSPI'fl LY3 p .264 291

..... Appendix V (cont.)

Family Genus Species Item Part %

STERCULIA CRAB ILIIIIOVIA BOSPITA LY3 .072 STIICBLIACBAI! PTIIOSPI!llllllll JAVUICIII F "liP 4.27 STDCULIACBAI PHROSPIII811 JAVAIICBII LY2 B .133 STDCBLIACUI PTIIOSPIIIIIIIII JAVAIICIIK L liP .121 STDCULIACI!AI PTIIOSPDIIIIII JAVUICIIII LYl B .024 STIICBLIACBAE PTIIOSPBIIIIIII JAVUICIII FB IP .169 STIIC8LIACI!AI PTIROSPIIIIUII JAVAIICUM FP liP •752 STDCDLilCBAI PTIIOSPIRIIIJII JAVUICOK 81 F .046 STIICDLilCI!AI PTDOSPDillll JAVAIICIII 81 If 2.93 STIRCOLilCIAI STIICDLil COCCIIBA VAR.COCCIIEA L13 IP .036 TIBACUE TmsTIOIIIIl PATDS 80 liP .041 TIUCIAI TEUSTROBIIIA PATDS Bl s .434 TILIACI!AE IIICIOCOS PUICDLATA L p .072 TILilCIAI KICROCOS PUICIL&Tl LI2 B .072 IIAIACIAE CELTIS PIILIPPDSIS .. IP .323 BLUCUE C!L!IS PIILIPPDSIS FB IP 2.79 DLUCI!AI CELTIS PBILIPPDSIS L8 IP .121 UUIACW CELTIS PIILIPPDSIS L!l IP .844 ILUCBAB CIL!IS PIILIPPDSIS LY2 IP .024 BLUCIAI CEL!IS PIILIPPDSIS Lll B .114 DLUCIAB CEL!IS PIILIPPDSIS LY2 B .398 OLUCEAB CEL!IS PIILIPPDSIS II liP .182 BLIIICUB CILTIS PBILIPPDSIS 88 If .732 Dill CIA I C!LTIS PIILIPPDSIS FA IP .617 DLUCIAB CELTIS PIILIPPDSIS Bl liP .072 OlllCEAI! C!LfiS PIILIPPDSIS LU B .014 VIIBIIACUB CLIIOD-811 SIRRATOK Ltl B .072 YDBDACBAK TIC!OIA GIUDIS LU B .383 YIIBD!CBAK TBCTOIA GIUDIS LY2 M 2.25 VBIBDACBAK TIC'fOIA GIUDIS LYl II 8.23 VIRBDACBAB TBCTOIA GDIDIS LY4 A .036 VDBDlCUE TBC'fOIA GIUDIS L!4 II 2.24 YIIBIIACI!AI! TICTOJA GIUDIS Ll2 A .036 VIIBDACBAI TIC'fOIA GIUDIS LY3 B .265 VEIBDlWB TICTOIA GDIDIS LY B .566 VDBDlCEAB TIC!OD GUIDIS Lll f! .264 VDBDlCBAI TIC!OIA GWDIS 811 If .072 VIIBDACW tiC!OIA GBUDIS LU B .036 VDBIIACIAI TBC!OIA GRAIDIS L!4 B .868 VIIBDlCIU TIC!OIA GIUDIS Lll B .036 YIIBDACI!AE Tlt"rrOA GIUDIS LYl II .069 VDBDACBAI TIC!OIA GIHDIS LB liP .029 VIIBIIICUI TICTOIA GIUDIS FA liP .819 VIIBII&CDI Ylfii COPUS OS FB liP .ooa YIRBDACI!AB VI Til PIDATA L 8 .109 VIRBBIACEAB VITII PllllTA LY2 8 .534 'IIDIIACIAB YITBI PIDATA Lt3 8 .583 VIIBDACEAK VITII PIIBTA LU B .853 292

..... Appendix V (cont.)

Family Genus Species Item Part %

VDIDACBAI VI'fii PillA fA LI B .072 VERBBIACIAI VI'fii PIJIATA FB liP .072 VIIBDACEAI VITII PinATA LYl liP .029 VERBBIACIAE VI'fll POBISCDS Ll2 B .169 VDBDACIAI VIm PUBISCDS LY3 B .206 VDBDACIAE Vltll POBISCDS LY B .091 9DBDACIAI VI'fll POBISCDS LJl B .009 VBIBDACIAE VITKI PUBISCDS LM B .072 VDBIIACBAI VITI I PUB!SCDS L B .072 VDBDACBAE 9If!l PUBISCEIS LY4 B .072 9IOLACIA! RIIOW SP 81 If .036 293

....• Appendix V (cont.)

b) GRP21 Family Genus Species Item Part %

DMJIIDIACIU lllmlliA mu(J LY2 B .335 AMIJIIDIACI!AB liiGIPIIA IIU(J ttl B .426 AIICAIDIMDB lllmPIIA II)I(J LY4 B .345 AMCIIIDIMDB IMllPIIA Il)lC! Ltt B .039 AatJIIDIMDB sgnAS C'ft'IIIIIBA Lll B .157 AMIJIIDimK SRII>IAS Cft'IIBI.!EA LY2 B .157 AIICIImiJAB SPill)lAS tmllll!& LY B .235 AIICIII)IMDB SRIIliAS ct'fllf.IIBA LY3 8 .235 AIICAI)JACIAB SP!IIIIAS t'ftiiRIA LY liP .157

..... Appendix v (cont.)

Family Genus Species Item Part %

IIID.'IIII tBIISD lUW8 L!l B .314 IIILVJall ~ lUDfA II B .104 !JISliiiSil IUIIIa L!lSL .039 ...liUSmlft(!AB B.'IUI Sl LY1 B .178 IIWftiiW!AI IBimll SP LllB .078 IIWftiiW!AI SP L!l B .m W.JIC!II .... CIUirlal!ll II 1.45 DIDBII s W.IIf!P D1.IIJIILlll IJIIlll'fDJlll Lll B .078 &JN'IU DIDBit ~ II .157 BDIJII DIII!L1II ~ II •s 8.21 -.JDI· IJIImll ~ II ., .157 BUCJQ DID!Llll SP II s .628 FlaB II If 2.14 ...IIUliiM B8 If 1.41 ... PmiS ... FJ(JS lllllllliA L! 8 .775 IIIKIII IIIICIII FICE IIUliiM Ll2 B .615 IIIICIIII FICII I!DDIIIM LY3 B .209 lllllDI PICIS .... II liP .157 IIIICI'II FilS LYl B .141 ....IIICIIU! Bl .314 11111!11 mE s IIIICIII FICIII ..,.. 1111 If .943 liiDU FICIB lllmlfl II If .943 IIIICIII FICE nmma L!lB .464 FICIII PIBrUJl L!l B .1184 IIIICIII FICE nmma Ll2 8 --IIIIIJII FICE PISIUil L liP ••.314 llldll FICIII W8AB Ll liP .314 11111!11 Plll8 WIIIIATl L!l A .157 liiiCIII Flail SIIITl L! B .314 lllllJII FICII ~ L liP .104 IIIIIJII neas SIDTl 1.1 liP 2.73 llliCIII FICIB Slim Ll1 B 2.55 .... FICE SIIIm Ll2 B .565 FICIII SIDTl L!ll .298 ... FICE SIIITA LilA .Ill ...... FICE ~ Ltt IP .031 llldAB ncm SIIIft LI211P .157 IIIICIII FICIII S.uA Ll liP .052 IIIIICDI FICIII SP II If .251 IIIICIII FR1I .... II If 1.57 FICIII LY3 B .078 FICIII LllB .157 Plt'lll ... II liP .157 --llllt1ll - --IIIICIII FICII VMIDD II liP .157 IIIIMJII IOJIIUIIIIIIIII --IIVBII II Jll .157 roBIJaiPIIIIII .VIUIJ • If .052 llllrDII IQIIIl SP L Jll .288 --IIIIICJM IIIOIIJ! SP II liP .852 295

..... Appendix v (cont.)

Family Genus Species Item Part %

tmrrlCIIAB I!IDIJl SP LY2 B .262 IUB!JCIII IQIIl SP L13 B .026 tmlfACIAE I!IDIJl SP L B .039 MDIBCIU IIIDWII Ciml LY2 B .235 ll1liDCIII liiDlEA Ciaa L13 B .078 (JACIIII LllmBA SP RI II' .157 a.BICIU LIImlll 111 Lfi B .917 a.BICIU JJIX:IIIIA SP LYlB .131 IIIIIAIJII Ml'llHLIS (1ft& YARJUII HI II' .471 IIDIUI!II GlllftD SPDIIA LY B .471 JIIDA(JI8 cal'llllll ftl(IIA Ltt B .251 RII1DI'AI GlllftAD KIIIIA LJ2 B IEDM Glll'l'!lllll SPI!Ria LYlB 3.65•• BIIJJIIU GllftAD RBIIA L!4 B .627 IIIJ1(!U CDiftAD SPII:I(Il LY4 ll' .083 IIDIDIJII l2r.llm SlEC6I L B .141 IIIIBIDil am~~ SPII:ll& L13SL .444 RIBimll (Jift'AD ftii& L!2 SL .022 lliJDCIAI CDiftAD SPDIII LY4 SL .018 IIDIIDB IIIOA SP B8 I 1.79 IIIIIDI IUDA IP 811 s .943 IIIDI!U UDA SP PB ., .471 BWMU IIlJA SP II IP .039 lllllal 'IIICAL1SD SOOQI!JS II IP .471 IIIDtDI 'l'RICALJSD SIRMlS fl If 1.72 IIIJ1(!U 'IIICJLJSD smuRIS II If .471 IIIIJDII ftiCALJSIA SJUUIIS 1111 II .411 IIIBDaiH !IICJLJSD SIIIUIIS L p .157 AID!aliA J.IIIIIWA 1111 s .314 SIPID(JU ami!A LlDILIS Bl s .551 --WlllaiAI (IIIIJ!Jl,ll FALIMIII 1111 s .393 SIPliiCIU caDB!WII r»a!!l LJ2 B .471 SIPliiC!IB IIDIIWII flla!ll II IP .157 SIPliWJII GADII1Wil PAUl!lll WB .235 IIPIIICIII •oo:m

..... Appendix V (cont.)

Family Genus Species Item Part %

srmt'IIJACIAI sna.IA ax:ciiBA VAR.amB L!l B .052 STIIICIIJACJWI S!IDIJl axma m.a:mm LY3 P .157 TILJ.mB saomm ~NATA LYl 1P .188 TIJ.D(DI S(DMJj CNATA LY IP .031 TIIJAaiB saomm omA LYl B .078 TILIAaliB saomm am! LY2 B .509 TILJ.mB saDJI'EA UiATA L!l B .584 TIIJAarU saornm l1lUl LY4 B .022 lW(!AI CUIS AIIJJPI1t!IIIS LY2 B .324 IIIIL!AI auiS miLIPPI!IISIS L!l B .083 lllmJIIIII DlmDJIIll KICDTIQI I'B If .275 VIIII!RIIl!ll ~ GWDIS LY3 II .314 vna QBATl FA liP 1.10 yPIIICIII VIM ~ 88 F .550 ---.-cal VIM

Refer to Table 6.1 (Section 6.3) for descriptions of abbreviations used for item and part classes