THE BEHAVIOURAL ECOLOGY OF CHIMPANZEES IN 
GOMBE NATIONAL PARK, TANZANIA 
by 
Richard Walter Wrangham 
St.John ' s College 
Dissertation submitted to the 
University of Cambridge for the 
degree of Doctor of Philosophy 
Sub-Department of Animal Behaviour 
High Street, Madingley 
Cambridge 
UNIVERSITY 
LIBRARY 
\MdR D~E 
January 1975 
R 
PREFACE 
The study reported here was carried out from January 1972 
to January 1975 based ,at the Sub-Department of Animal Behaviour, 
Madingley, Cambridge. I am grateful to Professor T. Weis-Fogh and 
Dr. H. Lissmann for the use of facilities. The data were collected 
during field work in Gombe National Park, Tanzania from May 1972 to 
September 1973. I am grateful to the authorities of the Tanzania 
National Parks for their permission and cooperation, and to 
Dr. J. van Lawick-Goodall for providing facilities in the Gombe 
Stream Research Centre. 
I acknowledge with gratitude the Grant Foundation of New 
York for their financial support throughout the study. 
At Madingley I was much indebted, for his encouragement, 
advice and criticism, to my supervisor Professor R.A. Hinde, F.R.S. 
Many other ' members of the Sub-Department kindly helped and advised 
me. I am grateful to J.J. Cherfas and H.C.M. Meekings for writing 
computer programs, to L.D. Barden for assistance in the preparation 
of diagrams and to Mrs. P. Lister for her meticulous typing. 
Many of the data analysed here were collected by members of 
the Gombe Stream Research Centre as part of a long-term programme 
co-ordinated by Professor D.A. Hamburg, of Stanford University, 
Professor R.A. Hinde and Dr. J. van Lawick-Goodall. For their 
contributions and valuable discussions I am grateful to all these, and 
especially to Rugema Bambaganya, Adriano Bandora, Harold Bauer, David 
Bygott, Hassani Bitura, Stewart Halperin, Hamisi Matama, Hilali Matama, 
Bill McGrew, Juma Mkukwe, Esilom Mpongo, Yahaya Ntabilio, Anne Pusey, 
David Riss, Kassin Selemani, Yasini Selemani, Margaretha Thorndahl 
and Caroline Tutin. Dr. H. Kraemer kindly assisted by correcting 
l 
ii 
mistakes incurred during computer transcription of data collected by 
these and others. The use of data collected by others is specifically 
acknowledged in the t~xt. 
My gratitude is also due to Dr. J.B. Gillett and Dr. C. Kabuye 
of the East African Herbarium, Dr. T. Nishida and the late D.F. Vesey-
Fitzgerald for identification of plant specimens. 
I declare that the work reported in this dissertation is 
entirely original, except where expressly stated to the contrary; that 
no part of this thesis has been or is being submitted to any other 
University; and that the text does not exceed 80,000 words. 
a 
SUMMARY 
Chapter I 
The aims of this study are to understand certain relationships between 
the behaviour and ecology of chimpanzees. By studying one population 
over time many complicating factors are eliminated. The behaviour of 
females changes according to their reproductive state: in order to 
focus attention on behavioural responses to environmental change, only 
adult males, in two communities, were selected as target individuals. 
Chapter 2 
The study population and its habitat ~ described and recording methods 
are outlined and discussed. 
Chapter 3 
The aim 1n describing feeding behaviour is to identify and explain 
seasonal cpanges. Chimpanzees are found to have an intimate knowledge 
of their environment and its food sources; the data do not support 
previous suggestions that the function of dispersal and aggregation 1S 
to locate foods. Evidence from a series of dependent measures 
indicated that more food was available in the second of two dry season 
periods of observation. At this time parties were larger. Apparently 
when food is scarce the behaviour of individuals is directed to 
maximising feeding efficiency; when food is abundant males may forego 
maximisation of feeding efficiency for the sake of increased 
reproductive effort. 
Chapter 4 
All observations of chimpanzees eating animals in Gombe National Park 
are listed. The factors affecting predation by chimpanzees are 
considered, and it is concluded that opportunity is the most important 
III 
· IV 
determinant of predatory interest. The frequency of predation on 
mammals is calculated, and prey spec~es are found to suffer mortality 
rates comparable to the prey of true carnivores. Food transfer during 
meat - eating is considered to be the outcome of competition for an item 
which is not easily divisible and can be carried. 
Chapter 5 
Data are presented on the ranging behaviour of individuals over one 
day, four days and one year, and on the directions from which 
individuals entered the artificial feeding area. It is found that 
despite considerable overlap individuals use different core areas. 
Although males may be regarded as forming distinct communities, some 
females do not conform to male community'rules': for instance they 
associate peacefully with males of more than one community. A new 
model of chimpanzee dispersion is proposed. 
Factors af f ecting travel patterns are discussed. Larger ranges 
occurred when more food was available, and are suggested to represent 
increased reproductive effort. 
Chapter 6 
Some chimpanzees were g~ven bananas at an artificial feeding area: 
the possible influence of artificial feeding is discussed. Though its 
consequences may be widespread they were reduced ~n this study by 
appropriate methods of observation and analysis. 
The form of chimpanzee social structure is suggested to depend on the 
feeding strategies of mothers. This principle is applied to other 
apes and species with similar social structures. 
CONTENTS 
Preface 
Sununary 
Table of contents 
List of figures 
List of plates 
CHAPTER 1 INTRODUCTION 
1. Aims 
2. Theoretical background 
i. Primate social structures 
ii. Problems of behavioural ecology of chimpanzees 
~~~. The applicability of functional explanations 
3. Practical difficulties 
~. Monitoring the environment 
~~. Monitoring social behaviour 
CHAPTER 2 STUDY AREA AND METHODS 
1. Study area 
~ . Gombe National Park 
~~. Vegetation 
~~~. Climate 
iv. Fauna 
v. The research centre 
2. Study population 
i. Individuals 
~~. Age classes 
3. Observation methods 
~ . Recording methods 
ii. Selection of targets 
~~~. Data recorded 
~v. Terminology 
4. Data from other sources 
~. Camp records 
ii. Group Travel charts 
5. Samp ling prob lems 
6. Statistics 
CHAPTER 3 FEEDING BEHAVIOUR 
Section I INTRODUCTION 
Section 11 TIME SPENT IN DIFFERENT ACTIVITIES 
1. Introduction 
2. Activity budgets 
i. Seasonal variation 
~~. Diurnal variation 
~~~. Activities in different habitats 
~v. Variation between days 
v. Variation between age-sex classes 
3. Time spent eating different foods 
~. Variation between months 
~~. Variation between age-sex classes 
v 
Page 
i 
iii 
v 
x~ 
xii 
1.1 
1.1 
1.4 
1.7 
1.8 
1.10 
2.1 
2.1 
2.3 
2.3 
2.4 
2.5 
2.5 
2.5 
2.7 
2.8 
2.8 
2.9 
2.9 
2.11 
2.12 
3.1 
3.3 
3.4 
3.6 
3.6 
3.7 
3.8 
3.9 
3.10 
Section III FOOD SELECTION 
1. Introduction 
2. Variation in diet 
i. Between populations 
ii. Between seasons 
~~~. Between ' days 
3. Factors affecting food selection 
i. Time of day 
ii. Individual differences 
~~~. Social factors 
Section IV HOW FOODS WERE FOUND 
1. Introduction 
2. Foods found without communication 
~. By chance 
~~. By knowing where to look 
~~~. By travelling to a productive area 
iv. By returning to known food sources 
3. Communication about foods within parties 
~. Incidental communication 
~~. Vocal communication 
4. Communication about foods between parties 
~. Food-calling and its effect 
~~. Individual variation in the frequency of 
food-calling 
~~~. Frequency of food-calling in relation to 
food quality 
~v. Seasonal variation in the frequency of 
food-calling 
Section V FEEDING METHODS 
1. Introduction 
2. Inspection 
~. Methods of inspection 
~~. Timing of inspection 
3. Food manipulation 
~. Variation with food items 
~~. Individual variation 
~~~. Rate of eating 
Section VI FEEDING IN PARTIES 
1. Introduction 
2. Feeding sites of different individuals 
~. Spacing 
~~. Intra-specific competition for feeding sites 
3. Social behaviour affecting time spent feeding 
4. 
5. 
~. Greeting interactions 
~~. Sex 
iii. 
~v. 
Party 
~. 
ii. 
~~~. 
Party 
Territorial behaviour 
Affiliative behaviour 
size and food availability 
Party size in relation to banana feeding 
Variation of party size at different foods 
Seasonal variation in party s~ze 
size and time spent feeding 
vi 
Page 
3.11 
3.12 
3.16 
3.17 
3.19 
3.21 
3.22 
3.23 
3.24 
3.25 
3.26 
3.27 
3.30 
3.31 
3.34 
3.35 
3.36 
3.37 
3.38 
3.38 
3.40 
3.42 
3.44 
3.45 
3.46 
3.47 
3.48 
3.52 
3.53 
3.53 
3.54 
3.55 
3.57 
3.58 
3.61 
Section VII SEASONAL CHANGES 
1. Introduction 
2. Evidence of changes in overall food availability 
~. Changes in the diet 
ii. Weight changes 
iii. Comparison between changes ~n weight and 
numbers licking salt 
3. Changes in social behaviour 
i. Feeding behaviour in relation to parties 
ii. Individual relationships 
Section VIII DISCUSSION 
1. Maximization of feeding efficiency 
~. Item selection at a food source 
ii. Diet selection 
~~~. Methods of finding food 
2. Seasonal changes in environmental variables 
~. Habitat structure 
~~. Climate 
iii. Food availability 
iv. Food distribution 
3. The relationship between food availability 
social structure 
4. The significance of apparently cooperative 
behaviour in finding food 
~. Food-calling 
~~. Party lea~ership 
CHAPTER 4 PREDATION 
Section I INTRODUCTION 
and 
Section II PREDATION ON ANIMALS OTHER THAN MAMMALS 
1. Introduction 
2. Invertebrates 
~. Types eaten 
~~. Sex differences 
3. Vertebrates 
1. Fish 
u. Amphibians 
~~~ . Repti les 
~v. Birds 
Section III FACTORS AFFECTING PREDATION ON MAMMALS 
1. Introduction 
2. Factors affecting predation on all spec~es 
~. Age-sex class of chimpanzees 
~~. Individual differences 
~~~. Time of day 
~v. Time of year 
v. Location 
v~. Size of prey 
v~~. Vulnerability 
3. Vulnerability of different prey species 
~. Red colobus monkey 
ii. Bushpig 
~~~. Bushbuck 
vii 
Page 
3.64 
3.64 
3.68 
3.71 
3.72 
3.72 
3.74 
3.75 
3.78 
3.81 
3.81 
3.81 
3.82 
3.82 
3.86 
3.88 
4 . 1 
4.3 
4.3 
4.4 
4.5 
4.5 
4.5 
4.5 
4.6 
4.7 
4.7 
4.8 
4.8 
4.10 
4.10 
4.11 
4.11 
4.14 
4.15 
~v. Redtail and blue monkeys 
v. Baboons 
v~. Other mammals 
Section IV FREQUENCY OF PREDATION ON MAMMALS 
1. Introduction . 
2. Principles of estimation of predation rate 
~. Omission of habituated baboons as 
prey species 
~~. Calculation method 
3. Estimation of predation rates 
~. Annual frequency of predations, 
regardless of prey species 
ii. Relative frequency of each species as prey 
iii. Annual predation rate on each spec~es 
iv. Amount of meat eaten annually 
Section V PREDATORY BEHAVIOUR 
1. Introduction 
2. Behaviour before eating 
~ . Ki lling prey 
~~. Behaviour immediately after a kill 
3. Meat-eating behaviour 
~. Evidence that meat is a preferred food 
ii. Overt aggression over meat 
iii. Individual success in getting meat 
~v. Methods of eating meat 
v. Relation to baboons 
Section VI DISCUSSION 
1. The significance of predation to chimpanzees 
2. The significance of predation to the 
prey species 
3. Why do chimpanzees share meat? 
~. Why do individuals try to get meat 
from others? 
~~. Why do possessors allow other individuals 
to take meat? 
4. Sharing of other foods 
CHAPTER 5 RANGING BEHAVIOUR 
Section I INTRODUCTION 
Section 11 ONE- DAY RANGE 
1. Methods 
2. Factors affecting length of range 
~. Sex 
ii. Community 
U~. Individual 
~v. Age 
v. Camp 
v~. Season 
v~~. Miscellaneous 
3. Comparison with other studies 
4. Circularity 
viii 
Page 
4.17 
4.18 
4.21 
4.24 
4.24 
4.25 
4.26 
4.29 
4.30 
4.34 
4.35 
4.36 
4.37 
4.37 
4.38 
4.42 
4.46 
4.47 
4.47 
4.51 
4.53 
4.56 
4.60 
5.1 
5.2 
5.4 
5.4 
5.4 
5.5 
5.6 
5.6 
5.7 
5.8 
5.9 
Section III FOUR-DAY RANGE 
1. Methods 
2. Factors affecting the size of the four-
day range 
1.. Sex 
1.1.. Conununity 
iii. Individual 
LV. Season 
v. Physical condition 
3. Centres of activity 
i. Males (not consorting) 
1.1.. Females (not consorting) 
1.1.1.. Consort pairs 
4. Four- day range in relation to year range 
Section IV YEAR RANGE 
1. Introduction 
2. Methods 
1.. Analysis 
1.1.. Sampling problems 
1.1.1.. Statistical problems 
3. Independent males 
1.. Size of year range 
1.1.. Location of year range 
4. Independent females 
i. Size of range 
1.1.. Location of range 
5. Juveniles 
i. Size of range 
1.1.. Location of range 
6. Comparison with other studies 
Section V DIRECTIONS OF ENTRY INTO CAMP 
1. Methods 
1.. Data source 
1.1.. Analysis 
1.1.1.. Interpretation of results 
2. Overall directions of entry 
3. Mean angles of individuals 
1.. Males 
1.1.. Females 
1.1.1.. Sex differences 
4. Seasonal variation 1.n directedness 
1.. Males 
1.1.. Females 
1.1.1.. Sex differences 
Section VI INTERACTIONS BETWEEN INDIVIDUALS 
OF DIFFERENT COMMUNITIES 
1. Introduction 
2. Between males 
1.. Types of interaction 
ii. Location of interaction 
1.1.1.. Frequency of interaction 
3. Between males and females 
4. Between females 
Page 
5.10 
5.11 
5.12 
5.12 
5.13 
5.13 
5.14 
5.14 
5.15 
5.15 
5.16 
5.16 
5.17 
5.18 
5.18 
5.19 
5.21 
5.22 
5.23 
5.24 
5.24 
5.26 
5.27 
5.28 
5.30 
5.31 
5.32 
5.34 
5.35 
5.35 
5.36 
5.36 
5.37 
5.40 
5.41 
5.42 
5.45 
Section VII TIMING AND DIRECTION OF TRAVEL 
1. Circadian rhythms 
2. Factors influencing direction of travel 
~. Location of food sources 
~L Availability of "sleeping trees" 
iii. Topography 
~v. Long-distance calls 
v. Behaviour of other individuals 
Section VIII DISCUSSION 
1. The concept of the community 
~. Review 
~~. Evidence in favour of the concept of 
the bisexual community 
~~~. Evidence in favour of the concept of 
the male- only community 
iv. Synthesis 
2. A functional explanation of ranging 
behaviour 
CHAPTER 6 OVERVIEW 
Section I INTRODUCTION 
Section 11 THE INFLUENCE OF ARTIFICIAL FEEDING 
1. Introduction 
2. Community structure 
i. Community ranges 
~~. Individual ranges 
3. Relation to feeding behaviour 
~. Importance of bananas as food 
~~. Predation 
iii. The nature of camp as a food source 
iv. Seasonal variation 
Section III COMPARISONS WITH OTHER SPECIES 
1. Introduction 
2. Closely related species 
3. Distantly related species 
~. Introduction 
ii. Spider monkeys 
iii. Social carnivores 
~v. Wallabies 
Appendix 3.1 Monthly variation in time spent 
eating different foods 
Appendix 3.2 Foods recorded in Gombe National Park 
1. Plant foods 
2. Animal foods 
3. Miscellaneous foods 
Appendix 4.1 Assumptions underlying calculation of 
frequency of attendance at predations 
References 
x 
Page 
5.45 
5.47 
5.47 
5.50 
5.50 
5.51 
5.51 
5.54 
5.55 
5.57 
5.60 
6.1 
6.1 
6.2 
6.2 
6.3 
6.4 
6.5 
6.6 
6.7 
6.8 
6.15 
6.15 
6.17 
6.18 
A.l 
A.13 
A.19 
A.20 
A.22 
R.l 
Fig. No. 
3.1 
3.2 
3.3 
3.4 
3.5 
3.6 
3.7 
4.1 
4.2 
4.3 
4.4 
4.5 
4.6 
4.7 
4.8 
5.1 
5.2 
5.3 
5.4 
5.5 
5.6 
5.7 
5.8 
5.9 
5.10 
5.11 
5.12 
5.13 
5.14 
D 
LIST OF FIGURES 
Title 
Seasonal change in diet 
Die t variety: in re lation to observation time 
Aggregation probability in relation to food-
calling 
Diurnal variation in aggregation size ~n 
relation to availability of bananas 
Effect of camp on party size 
Feeding probability and party size 
Seasonal change in association patterns 
Diurnal variation in frequency of predation 
Monthly variation in predation frequency 
Clumping of predations in months 
Frequency of predation in different parts of 
the study area 
Age-sex classes of colobus killed by 
chimpanzees 
Annual number of recorded baboon predations 
Frequency of predations in relation to time 
spent with prey spec~es 
Frequency of predations in relation to 
numbers of prey spec~es 
Four- day ranges of Northern males 
Four- day ranges of Southern males and 
Northern females 
Four-day ranges of consort pa~rs, pregnant 
female and wounded male 
Reference map and key 
Year ranges 
Observed year range size in relation to 
number of observation points 
Male community ranges 
Directions of entry into camp: seasonal 
differences 
Directions of entry into camp: seasons 
combined 
Comparison of entry directions and year 
ranges: Northern males 
Four-day range (detail) 
Circadian rhythms in travelling, party 
stability and calling 
Models of chimpanzee dispersion 
Models of seasonal change in chimpanzee 
dispersion 
x~ 
After page 
3.9 
3.16 
3.34 
3.55 
3.59 
3.62 
3.72 
4.8 
4.8 
4.9 
4.10 
4.14 
4.18 
4.31 
4.32 
5.11 
5.11 
5.11 
5.11 
5.18 
5.18 
5.21 
5.30 
5.30 
5.31 
5.39 
5.45 
5.52 
5.61 
Plate No. 
2.1 
3.1 
3.2 
3.3 
3.4 
4.1 
5.1 
5.2 
5.3 
5.4 
LIST OF PLATES 
Title 
Kakombe Valley 
Adult male feeding ~n outer branches 
Peering 
Reassurance as a result of competition 
for feeding sites 
Interruption of feeding by greeting 
interaction 
A1bizia ? gummifera emerg~ng from 
forest canopy 
Monitoring behaviour 
Reassurance behaviour on the edge of 
the community range 
Dead female 'stranger' 
Travel line 
xii 
After page 
2.3 
3.10 
3.41 
3.48 
3.53 
4.12 
5.37 
5.39 
5.43 
5.51 
"The chimp is the only primate who has achieved that arcadian 
existence of primal innocence which we once believed \l7as the 
paradise that man had somehow lost" 
" 
Ardrey, 1967. 
measured by human standards, the chimpanzee of course 
appears as something horrible, a diabolical caricature of 
ourselves ... as common, as vulgar, as no other animal but 
a debased human being can be" 
Lorenz, 1963. 
"oo- aah, ooh-waah, ooh-waarh, oo-ahoo-wahooo, aaargh!" 
Jomeo, 1973. 
1.1 
Chapter 1 
INTRODUCTION 
1. Aims 
The aims of this study were: 
Ca) To examine the influence of changing food availability 
on the behaviour of individual chimpanzees, ·Pan troglodytes 
schweirtfurthii, and hence on their social structure. 
(b) To understand the significance to individual chimpanzees 
of feeding behaviour which is apparently altruistic (sharing meat and 
glV1ng long-distance calls at food sources) . 
(c) To describe the ranging patterns of individual chimpanzees. 
(d) To understand the adaptive significance of chimpanzee 
social structure. 
These problems are examined below (2/ii). 
2. Theoretical background 
1. Primate social structures 
(a) Comparisons between spec1es 
The aims of this study are concerned with understanding how 
behaviour is affected by the environment. A successful approach to 
understanding differences 1n social behaviour is to compare species 
with a close evolutionary relationship (insects: Wilson 1971; birds: 
Crook 1965, Lack 1968; rodents: Smith 1968, Barash 1974; carnivores: 
Kleiman & Eisenberg 1973; ungulates: Jarman 1974; primates: references 
given below). This method has two obvious advantages. First, differences 
between related species may be sufficiently small to allow a close 
correlation of social and ecological variables. Second; the extent 
to which the form of social structure is influenced by phylogenetic 
UN.S,.", 
LIBRARY 
CAMIIUOGE 
z 
1.2 
variables can be examined (Eisenberg 1963, Struhsaker 1969). Problems 
ln understanding the relationship between ecology and social structure 
of primates are therefore briefly reviewed. 
Distantly related species, such as certain primates and 
carnlvores (Schaller & Lowther 1969, Kruuk 1972, Eisenberg 1973), may 
respond in similar ways to similar ecological pressures. However, 
selection of suitable species for comparison with chimpanzees must 
follow the identification of relevant behavioural and ecological 
variables. 
A number of studies of primates have correlated environmental 
features with differences in social behaviour (DeVore 1963, Hall 1965, 
Crook & Gartlan 1966, Struhsaker 1969, Crook 1970a, Denham 1971, Jolly 
1972a, Eisenberg et al. 1972). The correlations tend to be weak 
(Crook 1970a) and have poor predictive power (Clutton-Brock 1972). 
Difficulties in obtaining useful generalisations appear to be 
methodological (inadequacy of descriptive terms: Crook 1970a, Clutton-
Brock 1972, Richard 1973) and conceptual (emphasis of survival rather 
than reproductive success: Goss-Custard et al. 1972). Furthermore, 
since differences in a multitude of adaptive characteristics cause 
species to respond differently to similar environmental pressures, 
simple correlations should not be expected (Crook 1970a, Hinde 1974). 
The most detailed recent review of primate social structures 
(Crook 1972) considered how their relationship to environmental 
pressures was mediated through social selective pressures. Simple 
models were constructed relating particular environmental and social 
variables (e.g. food availability and number of individuals in group). 
However, the adaptive significance of different social structures 
could only be understood when comparing closely related species: no 
overall correlations were made. It appears likely that further 
progress will be made as increasingly relevant social and ecological 
parameters .are identified (e.g. Clutton-Brock 1972). However, the 
relationship of environmental and social variables can be understood 
more directly by causal studies: these must be made within species, 
and may provide insights into functional problems. 
(b) Comparisons within species 
1.3 
Comparison of different populations of a single species clearly 
indicates that features of social structure are correlated with 
ecological factors (e.g. Gartlan & Brain 1968, Richard 1973) . By 
controlling for phylogenetic variables such studies yield firmer 
correlations and have more explanatory power than comparisons between 
species. Even here, however, the mechanism by which environmental 
variables influence social behaviour is often uncertain: for instance, 
it was not clear whether differences in territorial behaviour of sifakas 
Propithecus verreauxi were the result of immediate behavioural responses 
or were genetically determined (Richard 1973). 
The mechanisms by which ecological factors affect social 
behaviour are therefore most accessible to study by observation of a 
single population. In many studies environmental change is slight or 
difficult to measure, in which case the adaptive significance of 
behaviour must be interpreted on theoretical grounds. While this may 
be successful, as when sexes are compared (e.g. Rodman 1973), it is 
sometimes difficult to distinguish between alternative explanations. 
For instance, territorial behaviour by gibbons Hy16bates lar has many 
consequences, among which are the exclusive use of food sources and the 
maintenance of a social group: the relative importance of these 1n the 
evolution of territorial behaviour could not be judged (Ellefson 1968). 
Environmental influences on behaviour are most obvious when the 
process of change is actually seen (e.g. Crook 1966~ Kurnrner 1968). 
1.4 
Such instances provide the opportunity for a detailed examination of 
how individual behaviour mediates the relationship between the 
environment and social structure. A major principle of this study was 
therefore to identify ·changing environmental influences and observe the 
responses of a selected number of individual chimpanzees. 
ii. Problems of behavioural ecology of chimpanzees 
(a) It has been r ecognised for some time that chimpanzees form 
temporary associations or parties with variable combinations of different 
age-sex classes (Nissen 1931 and subsequent field studies). At different 
times the average size of parties changes (Azuma & Toyoshima 1962, 
Reynolds & Reynolds 1965, Izawa 1970, Sugiyama 1973, Nishida 1974): 
parties tend to be smaller when food appears to be scarce (loc. cit.). 
However, it is acknowledged that party size is influenced by numerous 
other factors, including vegetation type, population density and social 
relationships (Sugiyama 1973). No studies have presented evidence that 
food availability was the most important determinant of party size. 
In many species it lS difficult to find out what affects the 
tendency of individuals to form parties, since it varles little: many 
factors could therefore be responsible. In chimpanzees, however, it is 
possible to correlate changes in dispersion with independent variables. 
The relationship between food availability and social structure is 
examined here (1) by attempting to isolate and identify a change in 
food availability; and (2) by showing how individual behaviour changed 
as a consequence of the change in food availability. 
(b) Observations of the sharing of meat have been reviewed by 
Teleki (1973); and in a number of study areas observers have suggested 
that individuals call to attract others to share food sources (Reynolds 
& Reynolds 1965, van Lawick-Goodall 1968, Sugiyama 1973). Such 
apparently altruistic behaviour needs to be explained. If there lS a 
1.5 
cost to sharing food, evolutionary theory demands that there is a greater 
benefit associated with it. Benefits could come, for instance, from the 
expectation of a reciprocal act (Trivers 1971), from increased fitness 
of related individuals - (Hamilton 1964) or from attracting potential 
mates. possible benefits are explored by examining the context and 
consequences of sharing food. 
(c) One of the major problems ~n interpreting the significance 
of chimpanzee social structure is that there has been considerable 
confusion about its nature. "Social structure" is used to mean the 
content, quality, and patterning of relationships (Hinde, in press) . 
Itani & Suzuki (1967) were the first to suggest that chimpanzees 
have any permanent social structure beyond the mother-infant relationship. 
They suggested that a "large- sized group" o{ 30-50 individuals is a 
social unit sharing a common home range, and that unstable "small-sized 
groups" were made up from within that unit. The principle of this 
suggestion has since been accepted by most field-workers, though 
nomenclature has varied ("regional population" Sugiyama 1968; "unit -
group" Nishida 1970, Izawa 1970, Nishida & Kawanaka 1972; "community" 
Bygott 1974, van Lawick-Goodall 1974). In this study "community" is 
used to refer to a number of individuals who share the same home range: 
temporary associations are referred to as parties. It now appears that 
the number of individuals in a community varies more than suggested by 
Hani & Suzuki (1967), between 15 (Gombe) and 80 (Nishida & Kawanaka 
1972, Sugiyama 1973). 
Individual relationships within a community were reviewed and 
examined in detail by Bygott (1974). Males assDciate more closely with 
each other than with females, and spend more time together than females 
do. Interactions between communities were found to be mostly agonistic, 
with males from each community acting in coalitions against their 
neighbours. Further studies on individual relationships are in 
preparation (aggression: Bauer, Halperin; adolescents: Pusey; 
families: Thorndahl; sex: Tutin). Studies in Gombe were reviewed 
by van Lawick-Goodall ' (i.n pYe5s). 
1.6 
There is little information so far on individual ranging 
behaviour within communities. This ~s evidently important, however, 
since an individual's travelling patterns are related to, and may be a 
determinant of, his association patterns. If the distribution of food 
sources influences travel patterns, for instance, it would influence 
party composition and stability in different ways according to 
individual differences in ranging. Factors affecting individual ranging 
behaviour are therefore examined. As a result the present picture of 
community structure will be argued to be inadequate as a description of 
the relation between males and females. For the sake of simplicity, 
however, individuals are treated as belonging to one of two communities 
(Northern or Southern), as described above. 
(d) The unique form of chimpanzee social structure, with its 
apparently open and non-competitive society, has presented a problem 
to the search for correlations between ecology and social structure. 
Crook & Gartlan (1966) attributed the anomalous position of 
chimpanzees to their "pongid status". The significance of "pongid 
status" was interpreted by Kummer (1971) as their capacity for tolerance 
towards rarely seen conspecifics. He noted that chimpanzees were the 
only forest primate with an open society, adjusting party size to 
ecological requirements; ' and he suggested that monkeys could only 
develop such a system under unusually strong selective pressures, such 
as those experienced by gelada Theropithecus ge1ada (Crook 1966a) and 
hamadryas baboons Papiohamadryas (Kummer 1968). However, some forest 
monkeys do have fusion - fission societies within which parties combine 
b 
1.7 
or disperse like chimpanzees (blue monkey Cercopithecus mitis, Aldrich-
Blake 1970; spider monkey Ateles geoffroyi, Eisenberg & Kuehn 1966). 
This suggests that the phylogenetic status of chimpanzees is not so 
important as postulated. 
Similarities between chimpanzees, C. mitis and A. geoffroyi, led 
Eisenberg et al. (1972) to suggest that the best strategy for fruit-
eaters is to forage independently until a food source is found: its 
location should then be "announced" for the benefit of others. This 
suggestion raises important evolutionary questions, as mentioned 
above (b). Furthermore, other spec~es which live in forest and eat 
fruit have different social systems: gibbons live in small closed 
family groups (Ellefson 1968); orang-utans (Pongo pygmaeus) are almost 
entirely solitary (Rodman 1973, MacKinnon 1974). We therefore need to 
explain not only why chimpanzees have their particular type of social 
structure, but also why, among the species facing apparently similar 
ecological problems, the social systems of some are similar and of 
others are not. 
~~~. The applicability of functional explanations 
If behaviour ~s to be interpreted as adapted to an evironment, 
it is clearly important to ask whether the environment ~n which it is 
seen is that in which it evolved. This may not always be so. For 
instance, Dunbar (1973) suggested that the one- male groups of gelada 
baboons, now confined to montane plateaus, are a relict adaptation to 
an earlier lake-shore habitat. 
Chimpanzees currently occupy a variety of habitats from lowland 
evergreen forest to open canopy woodland (Kortlandt 1972), but their 
distribution appears to be limited by the necessity for access to 
riverine forest (Kano 1971), which ~s structurally similar to evergreen 
forest (Richards 1966). During the alternating wet and dry periods of 
b 
1 . 8 
the Pleistocene the proportion of different vegetation types 1n East 
Africa appears to have oscillated, but not their fundamental nature 
(Livingstone 1971). The Miocene Dryopithecinae, considered to be 
ancestors of chimpanzees, are known from open canopy woodlands 
associated with evergreen forest (Kortlandt 1972); fossil fruits from 
this period are similar and often referable to modern genera in East 
Africa (Lind & Morrison 1974). The present habitats of chimpanzees thus 
appear to be similar to those they have inhabited through much of their 
evolution: there is no evidence that they ever lived without access to 
components of the tropical forest . 
One ecological feature of Gombe National Park is clearly 
anomalous. This is the artificial feeding area, where bananas have been 
provided for chimpanzees, and accidentally for baboons Papio anubis, 
Slnce 1962 (van Lawick- Goodall 1968, 1971, Wrangham 1974). Ways in 
which feeding, ranging and social behaviour may have been affected by 
the provision of bananas are discussed in Chapter 6. 
3. Prattical difficulties 
1. Monitoring the environment 
Changes in environmental variables tend to be linked: between 
wet and dry seasons in a tropical forest, for instance, there are changes 
in climate, habitat structure and the distribution and abundance of food 
sources. A major difficulty of ecological studies of primates has been 
the isolation and identification of these factors. 
It is difficult to measure food availability because animals 
requ1re many different nutrients. Hairston et .al. (1960) considered 
that since grass tends to be present all year, large terrestrial 
herbivores do not run short of food. However this ignores the importance 
of particular nutrients. For instance, Sinclair (1970) showed that grass 
L 
1.9 
contained insufficient protein for buffalo (Syncetus caffer) during the 
dry season. Similarly the fact that an ecosystem contains many fruits 
and leaves of a type normally "eaten does not mean that food is adequate 
for the primates there. 
Any attempt to sample food availability must therefore take 
into account variations in food quality: many different nutrients could 
be important, as could the presence of toxins (Free land & Janzen 1974). 
Unfortunately the diversity of diet, food source distribution and item 
quality (between and within food sources) makes this a momentous task. 
Availability is itself a complex concept, depending on the position of 
food items in space, the degree of inter- specific competition and the 
experience of the eater. in knowing which items are valuable foods. These 
problems mean that measurements of food availability are extremely hard 
to obtain, particularly in a large and complex habitat with discretely 
distributed food sources. Any conclusion that food is abundant is 
inevitably ' dubious owing to the difficulties of including all nutrients 
in an analysis, and of estimating the effects on sampled availability of 
competitors for food. 
However, unless we have measurements of food availability, the 
significance of changes ~n other environmental variables can rarely be 
satisfactorily interpreted. It is only safe to assume that the 
ecological cause of a behavioural variant ~s not a change in food 
abundance if a very strong correlation is found: an example is the 
relationship between few cliffs and large sleeping parties of hamadryas 
baboons (Kunnner 1968). 
The solution adopted here was to control for changes in habitat 
structure, climate and predation pressure by considering a single area 
during the same months in successive years. The only expected 
ecological change was therefore in the nature of food sources: failures 
1.10 
of particular food crops are known to occur regularly (van Lawick-
Goodall1968). Details of data collection methods and indirect evidence 
that food availability did change between successive dry seasons are 
given in the relevant ·sections. 
ii. Monitoring social behaviour 
Methods of recording social behaviour must evidently be relevant 
to the problem. The aim was to find out how behaviour was affected by 
the environment. However, an individual's behaviour is also affected 
by age, sex, reproductive condition, number and age of dependent young, 
dominance status and other relationships. In order to control for as 
many of these variables as possible, only adult males were selected as 
subjects for observation. 
Data on other age-sex classes are available from two sources. 
First, the behaviour of all individuals seen with the selected male was 
recorded. Second, data contributed by other observers were used to 
answer specific questions. 
This study is therefore not a general account of chimpanzee 
behaviour in relation to the environment. There are few data except on 
adult males, and some months of the year are not represented at all. 
Good general descriptions of chimpanzee behaviour and ecology are 
available for Gombe (van Lawick-Goodall 1968, 1974) and other areas 
(Mahali Mountains, Tanzania: Nishida 1974; Budongo forest, Uganda: 
Sugiyama 1973) in addition to numerous shorter studies (mentioned 
throughout text). These studies show that ecological conditions may 
very considerably from year to year. Only after many years of 
observation, therefore, can useful generalisations be made about 
chimpanzee behaviour. By using selective recording methods this study 
aims to explain rather than merely describe behaviour: this approach 
may yield principles applicable to other populations and species. 
2.1 
Chapter 2 
STUDY AREA AND METHODS 
1. Study area 
1.. Gombe National Park 
The Gombe National Park lies on the shore of Lake Tanganyika in 
the west of Tanzania, about 10 km north of Kigoma. Its eastern boundary, 
the running crest of hills bordering the lake, varies about 1500 m 
above mean sea level. From there a series of parallel dendritic streams 
flow 3 to 4 km to the lake, 683 m a.m.s.l. The northern and southern 
boundaries are cut fire-breaks. 
The study area (Fig.5.4) was 1.n the northern part of the park, 
and included all areas used by the study population. Its dimensions were 
roughly 3 km x 15 km. 
1.1.. Vegetation 
A detailed description of the vegetation in Gombe is being 
prepared by Clutton-Brock & Simpson, and an extensive account was given 
by Clutton-Brock (1972). The following account is therefore brief. 
Although Gombe National Park lies in the continuous miombo belt 
of central Africa (Lind & Morrison 1974), a mosaic of different vegetation 
types occurs. Two factors are responsible. First, there 1.S more soil 
water than occurs beyond the hills surrounding the lake: most of the 
streams are permanent, and there are springs on some of the valley walls. 
Riverine forest is therefore widespread. 
Second, human influence has been considerable. The Park was 
first gazetted as a Game Reserve in 1942. Until this time, and 
possibly subsequently, trees were cleared to make room for plantations. 
Two consequences are still important. First, the cleared ground 1.S 1.n 
• 
2.2 
the process of recovering its natural vegetation: much of the semi-
deciduous forest is secondary, and where fires have been excluded the 
boundary of grassland has been observed to recede by several metres 
since 1965, as shown by photographs from then. Second, some of the 
plantations were of the oil palm, E1aeis guineensis (Hi1a1i, pers. 
comm.). As in the farmed areas outside the Park, the oil palms mostly 
occur in stands in the valley bottoms. These are an important food 
species of chimpanzees and other primates, but there are indications 
that the palms are not replacing themselves. 
The habitats in Gombe intergrade, and the more types that are 
distinguished the harder their description becomes. For simplicity 
five basic types were classified. These were: 
a) Evergreen forest. Canopy up to about 35 m, many evergreen 
speCles, woody vines common in understory. Largely confined to valley 
bottoms. Anthoc1eista, Newtonia, Monanthotaxis common. 
b) Semi - deciduous forest. Dominant trees very variable, 
including Combretum, Pterocarpus, Ficus, Anisophy11ea. Herbaceous 
shrub layer, often dense. 
c) Brachystegia woodland. Miombo woodland with a single 
continuous canopy at 10- 15 m. Sparse shrub layer of grasses. 
Brachystegia, Uapaca, Monotes, Termina1ia common. 
d) Grassland with scattered trees. Hyparrhenia often dominant 
grass, no continuous tree canopy. Regularly burnt especially on 
upper slopes. 
e) Sand. A strip up to 150 mwide on the lake shore. No 
vegetation. 
Fig. 5.4 shows the approximate distribution of these habitats 
ln the central part of the study area. I am grateful to Juma and 
Yasini for preparing the habitat map. Plate 2.1 is a composite 
2.3 
photograph showing the entire length of Kakombe valley, and 
demonstrates the diversity of vegetation types. 
111. Climate 
Annual and circadian rhythms in climate were described by 
Clutton- Brock (1972). The climate was typical during the study period: 
a) Rain. Monthly rainfall during 1972 and 1973 is shown 
in Table 2.1. 
Tab le 2.1 
Monthly rainfall 
Data collected by field assistants from a rain-gauge ln lower Kakombe. 
Total rainfall lS shown ln cm. 
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 
1972 ? 19.0 19.9 16.6 8.1 0.0 0.0 0.0 0.2 8.5 12.8 21. 9 
1973 14.2 12.9 18.9 17.1 18.8 0.3 0.0 0.0 0.0 6.1 39.2 27.4 
The period June to September is a regular dry season, and May 
and October generally have less rain than other months. 
b) Temperature. Variation in temperature was slight, between 
19 0 and 300 C (measured one metre above ground, in shade). The air was 
coolest in the early morning, warmed steadily from 0730 to 1300, and 
often remained high until 1700 hours. The hottest months, when daily 
o 
maXlma were regularly 30 C, were August and September. 
c) Humidity. In January 1973 mean relative humidity varied 
between 74% (1300 hrs) and 92% (0700 hrs), compared to 44% (1300 hrs) 
and 78% (0700 hrs) in July 1973. There were no observed differences in 
humidity or temperature between July and September 1972 and 1973. 
lV. Fauna 
The mammals and other relevant animals are listed ln Chapter 4 
(4/11 and 4/111/3) . 
• 
Plate 2.1 Kakombe Valley. The length of the valley (ca. 3 km) is shown from its source in the hills 
to the east (top left) to Lake Tanganyika (right). The photograph covers almost 1800 • 
2.4 
v. The research centre 
Since March 1965 the chimpanzees originally habituated by 
van Lawick-Goodall (1968, 1971) have been the subjects of numerous 
studies. In 1972 the ,staff of the research centre numbered about fifty 
people: many of these had dependents, bringing the total living in the 
Park near to one hundred. The great majority of these lived and worked 
on the sandy strip by the lake, interfering little with the natural 
environment. Up to ten people lived in huts sited individually in 
lower Kakombe valley (Fig.5.4). 
Chimpanzees occasionally took an interest in the huts: a few 
individuals stole clothes and chewed them. Similarly they sometimes 
raided the huts of fishermen living on the lake shore, where they ate 
ash, sacking etc. The response of chimpanzees to houses appeared 
opportunistic: their normal ranging and feeding behaviour was probably 
affected little. 
This was not true of the artificial feeding area (camp) where 
chimpanzees have been provided with bananas since 1962. Wrangham (1974) 
described the methods and some of the consequences of giving bananas. 
Aggregation sizes and the frequency of aggression were closely related 
to the manner of feeding, and fell following a change in the feeding 
method in 1968. Since then individuals have been given bananas once a 
week, or less, when alone or in small parties. 
The feeding area undoubtedly influenced rang~ng patterns. This 
will be discussed in Chapter 6. In order to obtain results which were 
influenced as little as possible by artificial feeding, target 
individuals were not given bananas if they visited camp during 
observation. 
2.S 
2. Study population 
i. Individuals 
The study population included all individuals fully or 
partially habituated to human observers, and are listed in Table 2.2. 
On the basis of association data they were classified into two 
communities (Northern and Southern), following Bygott (1974). Females 
were assigned to the community which included their most frequently 
observed companion. This is a convenience, though it is doubtful whether 
it represented their real association patterns ~n all cases (S/VIII/l). 
~~. Age classes 
Use of the terms infant, juvenile, adolescent and adult was the 
same as van Lawick- Goodall's (1974), roughly representing the ages O-S, 
6-9, 9-13 and over 14 years. Van Lawick-Goodall also recognised a 
category of old age which is here called "past prime". 
Offspring normally travelled with their mothers until 
adolescence. In this study all infants and juveniles were considered 
dependent, and were omitted from many of the analyses. They are 
mentioned specifically where they are included. 
3. Observation methods 
~. Recording methods 
Twelve months were spent ~n the field (May 1972 - January 1973; 
July - September 1973). Recording methods were developed before this 
study began during 14 months as a research assistant to Dr. J. van 
Lawick-Goodall. 
Small checksheets and a tape-recorder were used at all times. 
They were carried in bags strapped to my waist so as to allow both 
hands to be free for climbing or crawling. Individual chimpanzees 
Table 2.2 
Study population, 1972-1973 
Initials in brackets. Numbers show dependent offspring of mothers 
(average for period) . 
Northern community: independent males 
Past prime Prime 
Hugo (HG) Humphrey (HM) 
Mike (MK) Faben (FB) 
Evered (EV) 
Figan (FG) 
Satan (ST) 
Jomeo (JJ) 
Northern community: independent females 
Past prime Adult 
F1o* (FL) +1* Passion (PS) +2 
Nope (NP) +1 
Melissa (ML) +2 
Sprout (SP) +2 
Athena (AT) +1 
Nova (NV) +1 
Pallas (PL) +1 
Miff (MF) +1 
Fifi (FF) +1 
Gigi (GG) 
Winkle (WK) +1 
Southern community: independent males 
Past prime 
Goliath (GO) 
Hugh;~ (HH) 
Prime 
Char lie ' (CH) 
wi lly-Wally (WW) 
De (DE) 
Godi (GI) 
Southern community: independent females 
Adult 
* 
Madam Bee 
Mandy 
Wanda 
Dove 
(ME) +2 
(MD) +2 
(WD) +1 
(DO) +1 
Adolescent 
Sherry (SH) 
Adolescent 
Sparrow (SW) 
P atti (PI) 
Adolescent 
Sniff (SF) 
Adolescent 
Gilka (GK)*** 
Little Bee (LB)** 
Flo died in August 1972, and her juvenile son in September 1972. 
2.6 
Hugh was last seen in November 1972, and was presumed dead. Pa1las' 
infant died. 
** . Little Bee was st~ll dependent, travelling for most of her time with 
*** 
her mother Madam Bee. However, Little Bee is listed here because she 
was suffi cient ly independent to "leave" the Southern community and 
"join" the Northern community temporarily during 1972. 
Gi lka was at different times a "member" of both communities: she 
returned to her original community (Northern) during 1973. 
2.7 
varied in the extent of their habituation (Bygott 1974), and care was 
taken not to disturb them. Observation normally occurred in the 
company of a field assistant: the subjects appeared to disregard us. 
ii. Selection of targets 
As described in Chapter 1 only males were selected as target 
individuals, in order to reduce variance due to changing reproductive 
condition. Sixteen independent habituated males were available: two 
(SH and SF) were discarded because they were not full - grown. Of the 
other fourteen, eight were Northern and six Southern males (Table 2.2). 
Regular sampling was impossible because of the difficulties 
of finding particular individuals when required. In practice targets 
were selected largely according to availability. Methods of finding a 
target included listening, waiting ln camp, or starting at dawn where 
another observer had left a party the preVlOUS night. All-day 
observati9ns on each male were completed at least once. Table 2.3 
shows the total number of observation minutes with each male as target. 
Table 2.3 
Observation time on target individuals 
An observation period (top) was a continuous period, between night - nests, 
of observing or trying to observe the target. The number of minutes of 
good observation on each target is shown. 
a) Northern males 
b) 
Periods 
Minutes 
Southern 
Periods 
Minutes 
HG 
18 
9493 
males 
GO 
5 
1468 
MK 
10 
3888 
HR 
5 
2118 
FB EV 
9 24 
HM 
10 
3378 4089 11744 
CH WW DE 
6 4 3 
3034 1210 1664 
UNIVBtSI1'T, 
LIBRARY 
I CAMB"I"'GE 
~ 
FG 
14 
5499 
Cl 
10 
5263 
ST 
11 
4904 
JJ 
14 
3822 
Total 
110 
46817 
Total 
33 
14757 
780 hrs 
246 hrs 
b 
2.8 
Table 2.3 shows that some individuals were observed more than 
others. This was partly necessity (e.g. the Southern males were harder 
to find) and partly choice (e.g. three continuous 4-day observations 
were made on EV) . 
iii. Data recorded 
Descriptive notes were taken incidentally. Time-point sampling 
was used to record the behaviour of the target individual every minute, 
and of all individuals in the party every five minutes. The following 
were recorded: activity (walk, climb, sit/stand, lie, make nest, play, 
display, feed); food eaten; height above ground; grooming; habitat type; 
other species seen within 200 m; and, for party members, distance from 
target individual. 
Some data were recorded for the target individual whenever they 
occurred. They were: all calls given and heard; foods eaten during 
minute if not eaten on the minute point; and overt social interactions. 
Notes were taken of the weather and the physical condition of 
all individuals seen. 
iv. Terminology 
Different observers often obtain different results in recording 
even simple measures of activity from a single population (e.g. Jolly 
1972a). A number of terms used in this study are therefore defined here. 
Feed: have food ~n mouth or be in the process of picking or 
transferring it to the mouth. 
Feeding site: position on ground or in tree at which a chimpanzee 
could or does collect food without moving his feet. 
Food: anything known to be eaten (ingested). When a "new" food was 
eaten the individual was scored only after eating had been seen. 
Food source: volume containing feeding sites sufficiently close for a 
chimpanzee to move from one to another without ceasing to feed. 
2.9 
Climb: move horizontally or vertically, all limbs off the ground. 
Walk: move horizontally, at least one limb touching the ground. 
Travel: walk or climb. 
Lie: immobile, side or back of head resting on substrate. 
Party: all individuals within 100 m of the target individual. In the 
dry season months this was fairly simple to record, since 
visibility was good and parties tended to be discrete. In the 
rainy months individuals could be missed. However, it was often 
probable that if the observer was ignorant of their presence the 
target individual was too. 
4. Data from other sources 
~. Camp records 
The study benefited from access to various sources of data 
collected since 1968 by members of the Gombe Stream Research Centre. 
These were (1) accounts of special events, e.g. predations, for which 
recording methods were not standardised; (2) data recorded in the 
artificial feeding area; (3) Group Travel charts (below). The 
reliability of data from the first two sources is discussed in the 
relevant sections. Group Travel charts are described below. 
~~. Group Travel charts 
The regulations of the Tanzania National Parks required that 
observers work in pairs. The second observer was normally one of ten 
field assistants working permanently at the research centre. 
The field assistants were trained to collect simple standardised 
data on party composition and the behaviour of the target individual. 
This provided a pool of standardised data (Group Travel charts) from all 
observations made outside the artificial feeding area. Since January 
1972 about 400 observation hours per month have been recorded on Group 
Travel charts, providing a much larger sample than could be achieved by 
n 
any single observer. Some of these data were analysed in this study: 
in particular, the descriptions of ranging behaviour (Chapter 5) are 
based on data provided by many different observers. 
2.10 
The following ' data were recorded continuously: party 
composition; major feeding bouts and food species of target individual; 
and travel route of target individual. All field assistants were 
trained until judged reliable. Frequent discussion occurred between the 
two observers, and the data were regularly checked. The following 
problems were noted. 
a) Party composition. No formal definition of a party was 
agreed until near the end of the study. In practice the individuals 
recorded as present were those that were visible at least every few 
minutes. This allowed considerable latitude for different interpretation 
by different observers, and probably contributed to an exaggeration of 
the size of parties on occasion. Normally, however, parties were 
clearly discrete entities. 
b) Food species and feeding. The observers were thoroughly 
familiar with the main food spec~es. Uncommon foods, and those that 
were eaten for less than 15 minutes, were not always recorded. 
c) Travel path. Observers drew on printed maps (ca. 1: 11560) 
the path of the target individual, and recorded his position every 15 
or 30 minutes. The maps were drawn by members of the research centre: 
no others were available. Details were added as the area became better 
known. 
A relatively objective test of reliability was to compare 
nesting areas recorded by different observers between one evening and 
the next morning. In a sample of thirty, the median distance between 
the two records of the same nesting area was 70 metres. Since the 
analysis of ranging data (Chapter 5) used squares with sides of 500 m 
the data were considered acceptably reliable. 
2.11 
The scale of the map was estimated from the grid system. The 
overlaid grid in Fig. 5.4 is a graphical extension of a block of 138 
smaller squares cut and measured on the ground by Clutton-Brock (1972). 
The squares were initi'ally aligned by pacing lOO-yard sides. 
Subsequently Clutton-Brock measured 34 squares: the mean length of 
the sides was 92.4 m, S.D. 9.1 m. 
The sides of the small squares have here been treated as being 
100 m long for arithmetic simplicity. This is therefore an 8.2% over-
estimate of Clutton-Brock's measurements. Some of this slack is taken 
up by varied angles of slope, since Clutton-Brock's measurements (with 
a range-finder) did not take them fully into account. Given this and 
the inaccuracies imposed by extending the grid system graphically on 
maps drawn without survey equipment, the scale is considered an 
acceptable approximation to reality. 
5. Sampling problems 
Practical difficulties in finding and keeping up with chosen 
targets resulted in observation hours being distributed unequally 
across individuals and months. However, it was important to consider 
the effect of individual variation when analysing, for instance, the 
change of behaviour over time. Whenever possible different conditions 
(e.g. seasons) were compared using balanced samples of observation from 
the same individuals. Sample sizes were consequently small. 
In selecting targets and deciding when to stop observation the 
problem of bias was always present. positive attempts were made to 
reduce this. For instance, the decision to stop observing was always 
made at least an hour beforehand, in case I was tempted to continue 
observation merely to include a rare event in the sample, or to stop to 
avoid a rainstorm. Observation hours were evenly distributed across the 
L 
2.12 
day (Table 2.4). (Low values at dawn and dusk ~n Table 2.4 represent 
variation in the time when individuals left or entered their night-nests.) 
Table 2.4 
Distribution of observation hours between night-nests 
The percentage of observation hours recorded at each time of day is shown. 
N = 1046 hours. 
Time % obsvn. hours 
5-6 0.0 
6-7 5.6 
7-8 7.4 
8-9 7.8 
9-10 8.6 
10-11 8.5 
11-12 8.6 
12-13 8.0 
13-14 7.5 
14-15 7.3 
15":'16 7.4 
16-17 7.4 
17-18 7.0 
18-19 5.8 
19-20 3.0 
20-21 0.1 
21-22 0.0 
6. Statistics 
Non-parametric statistics (Siege1 1956, Zar 1974) are used 
throughout. All probabilities are two-tailed, and only the 0.05, 0.01 
and 0.001 levels are given. 
I. Introduction 
Chapter 3 
FEEDING BEHAVIOUR 
3.1 
Selected aspects of feeding behaviour are described in this 
chapter. Predatory behaviour,is a special form of feeding which raises 
different questions from those asked here. It is treated separately 
~n Chapter 4. 
There are three specific a~ms in describing feeding behaviour. 
(a) To find out whether it maximised feeding efficiency. 
(b) To identify seasonal changes in environmental variables. 
(c) To describe and interpret seasonal changes in feeding 
behaviour and social structure. 
(a) Maximisation of feeding efficiency 
"Food" is used here to mean "that which ~s eaten". "Feeding 
efficiency" means "the rate at which food resources (i.e. that which 
could be eaten) are converted into physiological benefit". An individual 
therefore max~m~ses his feeding efficiency by maintaining peak physical 
condition while using the least time and energy (and incurring the least 
risk of damage). 
The forms of feeding behaviour and social structure are flexible. 
Yet they are evidently constrained within certain limits, of two kinds. 
First, there are independent constraints set by anatomy. A chimpanzee, 
for instance, cannot survive by eating grass, as an elephant or a 
ternrite can. Constraints on social structure are less obvious, since 
it is not always clear that "species- specific" characteristics, e. g. 
high rates of aggression between males, are inevitable in all 
environments (Hinde 1974). However, observations of the maintenance of 
a particular form of social structure even in an extreme environment, 
b 
3.2 
such as a zoo, are an indication of the importance of genetic influences. 
When hamadryas baboons were confined together they fought to the death 
(Zuckerman 1932); in so doing they established the form of social' 
structure seen in the 'wild (Kummer 1968). 
The second constraint on their forms is that feeding behaviour 
and social structure must be mutually compatible. They form an 
adaptive complex whose function is both to promote survival and to 
max~m~se the individual's genetic contribution to subsequent generations. 
This makes it hard to see how there may ever be different 
constraints on feeding efficiency or reproductive effort. However, s~nce 
individuals who may b'reed several times ga~n by surviving long enough to 
do so, the advantages of a high reproductive effort may often be outweighed 
by the disadvantages of decreased survival probability (Lack 1954, 
Williams 1966, Goodman 1974). Individuals thus face the problem of how 
to apportion matter and energy to somatic or reproductive tissues (Pianka 
1970), or in more relevant terms, when to max~~se feeding efficiency at 
the expense of reproductive effort. 
If survival depends on feeding efficiently, and if a high 
reproductive effort is incompatible with efficient feeding, feeding 
behaviour may be said to constrain reproductive effort. On the other 
hand, efficient feeding may not be at a premium. If food is abundant 
and peak physical condition can be maintained even by behaviour which 
does not maximise feeding efficiency, then the best feeding strategy is 
that which is compatible with a maximal reproductive effort. The form 
of social structure may then be said to influence feeding behaviour. 
In exam~n~ng the relationship between feeding behaviour and 
social structure, it ~s therefore important to find out ,whether feeding 
behaviour was ever directed to maximising feeding efficiency. Relevant 
data are presented throughout the chapter . 
b 
3.3 
(b) Difficulties in estimating a change in food availability 
were described above (I/3/i). Direct measurement was therefore not 
attempted. Instead, two dry seasons in different years were compared. 
There were no apparent -differences in habitat structure and climate 
between years. Differences in food availability were indicated by a 
variety of dependent measures, summarised in section VII. 
Behaviour was also sampled during three wet season months. 
Since there were changes in habitat structure, cli-mate and food types 
between wet and dry seasons, the relative influence of the different 
environmental variables on behaviour could not be estimated. However, 
data from the wet season are compared with those from the two dry 
seasons, to provide a partial test of the proposed relationships between 
feeding and social behaviour. 
Predation on chimpanzees has not been recorded in Tanzania. Man 
appears to be the most important predator in some countries, and it ~s 
conceivable that leopards may occasionally kill chimpanzees (Kortlandt 
1972), though there is no clear evidence to date. However, we may 
safely ignore the influence of predators on chimpanzee behaviour in 
Gombe if we are concerned with understanding changes ~n social structure 
between seasons. 
(c) Seasonal changes in behaviour are des cribed. The 
significance of these changes is interpreted ~n terms of changing 
food availability. 
11. Time Spent in Different Activities 
1. Introduction 
Variations in activity budgets and feeding time are here 
described as a background for the consideration of specific problems 
of feeding behaviour. 
3.4 
2. Activity budgets 
l. Seasonal variation 
Since activity was markedly affected by time of day only all-
day observations were 'analysed. Activities recorded on succeSSlve 
half-hour points were combined across days and individuals. (Statistical 
dependence was tested at l-min., 5-min. and 30-min. intervals for 
travelling, resting and feeding on ten separate days: dependence 
occurred (p < 0.05) only when the data were sampled at succeSSlve 
l-min. intervals.) 
Table 3.1 shows the data from 54 all-day observations. 
Table 3.1 
Seasonal variation In activity budgets of adult males 
Per cent of 30-min. points spent In different activities lS shown. 
Observations at the feeding area (N = 60) are omitted. Eating could 
be combined with grooming or travelling, so that percentages can 
total more than 100. Mean times of leaving and entering night-nests 
are shown from all observations (N = 75). 
Activity Jul-Sep '72 Nov-Jan '73 Jul-Sep '73 May , 72-Sep '73 DRY WET DRY ALL 
Eat fruit 33.8 ) 32.9 ) 35.8 ) 33.1 ) 
Eat leaf 12.4 ) 60.2 8.4 ) 46.5 11.1 ) 56.7 11.8 ) 55.7 
Eat other food 14.0 ) 5.2 ) 9.8 ) 10.8 ) 
Travel 8.4 14.5 20.0 13.8 
Allo-groom 5.5 ) 8.8 ) 4.1 ) 6.2 ) 
Sit/stand inactive 17.4 ) 29.8 22.9 ) 39.0 16.5 ) 24.9 17.8 ) 30.3 
Lie (not night-nest) 6.9 ) 7.3 ) 4.3 ) 6.3 ) 
Not observed 1.6 0.8 0.8 1.2 
Total 100.0 100.8 102.4 101.0 
N (30-min. points) 379 249 369 1197 
Mean time of 
leaving night-nest 0650 hrs 0628 hrs 0647 hrs 0642 hrs 
Mean time of 
entering night-nest 1834 hrs 1836 hrs 1906 hrs 1843 hrs 
Mean time between 
night-nests (min) 704 726 739 721 
Activity budgets were fairly similar across seasons. The 
clearest difference in feeding time was between the wet and the two dry 
3.5 
seasons, but these were not significant. (All-day observations were 
compared between seasons by matching target individuals, wilcoxon test: 
Dry 1972 - Wet: N = 8, T = 15.5, n~s.; Wet - Dry 1973: N = la, 
T = 15.0, n.s.; Dry 1972 - Dry 1973: N = la, T = 16.5, n.s.) 
Table 3.1 also indicates that travelling time was low in the 1972 and 
high ~n the 1973 dry season. Seasonal variation in ranging behaviour 
will be treated in Chapter 5. 
Differences in activity budgets between wet and dry seasons may 
have been due to a number of environmental factors whose influence 
cannot be controlled (food abundance and distribution, climate, shade, 
visibility, etc.). Van Lawick-Goodall (1968) has described some effects 
on behaviour of heavy rain, including the occurrence of violent displays 
and the tendency to sit inactive. Grooming occurred very rarely during 
rain. 
In the wet season individuals tended to rest in trees more than 
~n either dry season: the difference was tested by comparing the amount 
of time when the target was more than 2 m. above the ground, spent 
resting, self-grooming or grooming (data from all-day observations 
matched for target individual). The difference was significant between 
the wet and dry 1972 seasons (Wilcoxon T = 0, N = 8, P < 0.01), the wet 
and dry 1973 seasons (T = 3, N = la, P < 0.01), but not between the two 
dry seasons (T = 20, N = 11, n.s.). Previous observers have noted that 
chimpanzees do not like to sit on wet ground (e.g. van Lawick-Goodall 1968). 
Though the difference between the two dry seasons was not 
significant, it was considerable. In 1973, on 10.2% of 5-min. points 
(N = 13 all-day observations) spent ~n trees the target individual was 
inactive, compared to 2.1% (N = 13) in 1972. 
> 
3.6 
LL. Diurnal variation 
Activities were markedly affected by time of day., as has been 
described by van Lawick-Goodall (1968) and Nishida (1974): there were 
feeding peaks in the morning and afternoon, and individuals te~ded to rest 
at mid-day. Diurnal variation in travelling will be treated in Chapter 6, 
and in feeding on particular foods later in this chapter. 
Because diurnal variation w~ considerable, care is taken where 
possible to control for time of day. This is normally done by comparing 
all-day observations, so that all times of day are equally represented in 
the sample. 
UL. Activities in different habitats 
The major categories of vegetation (semi-evergreen forest, 
deciduous forest, grassland with or without trees) evidently differed Ln 
characteristics important to chimpanzees (e. g. types of food availab le and 
visibility). To test possible effects of these differences, activities 
, 
recorded in 54 all-day ob servations (as above) were separated by 
(Table 3.2). The artificial feeding . area (camp) was scored as a fourth 
habhat type. 
Table 3.2 
, 
Activity budgets of adult males in different habitats 
Data as from Table 3.1, but with observations at the artificial feeding 
area included. 
Semi-evergreen Deciduous Grass land Activity with Camp forest .forest or 
without trees 
Eat fruit 37.9) 34.6) 21. 2) 6.9) 
Eat leaf 13 .5) 53.9 10.5) 56.7 10.4) 56.1 0.0) 6.9 
Eat other food 2.5) 11.6) 24.5) 0.0) 
Travel 13.5 13.2 15.8 6.9 
Allo-groom 5.8) 5.6) 7.2) 32.8) 
Sit/stand inactive 16.6) 30.7 18.1) 29.7 16.9) 28.8 44.8) 86.2 
Lie (riot night-nest) 8.3) 6.0) 4.7) 8.6) 
Not observed 2.1 1.1 0.0 0.0 
Total 100.2 100.7 100.7 100.0 
N (30-min. points) 326 619 278 58 
, j 
3.7 
The major categories of activity were similar between the 
natural habitats and differences within the major categories are readily 
explicable~ Foods other than fruit and leaves, were eaten least in semi -
evergreen forest and mbst in grassland: these were mostly wind- dispersed 
seeds, which were more abundant in grassland. The per cent of active 
time spent lying showed ' the opposite trend, and this was presumably 
related to the decreasing availability of suitable trees (important 
during the wet season, when chimpanzees rarely lay on the ground) from 
semi-evergreen forest to grassland. 
In camp the normal activity budget was not seen. The fruit 
eaten came from E. guineensis, trees, and bananas were sometimes 
available. Otherwise, there was little for chimpanzees to eat, and they 
spent most of their time inactive or grooming. The proportion of 
inactive, time spent grooming was higher in camp (38.1%) than in other 
habitats (20.3%). This may have been related to the larger party SLzes 
in camp (VI/4/iii). 
The similarity of activity budgets in the different natural 
habitats allows data to be lumped across habitats. Owing to the clearly 
abnormal influence of camp on activities, however, data collected in the 
artificiql feeding area have been omitted from most of the analysis. 
LV. Variation between days 
Chimpanzees may do very different things on different days, 
depending on who they happen to meet, whether they hear strangers, and 
so on. Their activity budgets are consequently variable. Examples of 
variations in feeding times are shown in Table 3.3. 
3.8 
Table 3.3 
Variation between days in per cent of minutes spent feeding (adult males) 
Data from time of leaving night-nest to time of lying in night-nest. 
Month Target % minutes spent feeding on different days Range (%) 
June 1972 Hugo 55.1, 65.4, 78.9 23.8 
August 1972 Godi 52.8, 79.4 26.6 
July 1973 Evered 50.0, 61.1, 66.2, 72.3 22.3 
Feeding times depend partly on the nature of the foods. which 
doubtless vary in the nutritive gain per time feeding. Whatever the 
cause, variation in feeding time will affect the time spent in other 
activities. This considerable potential for variation between days 
makes comparisons between individuals difficult without large samples. 
The data are not sufficient to compare activities on succeSSLve 
days satisfactorily. It was my impression that if an individual 
travelling with a large party had fed less than usual, he might on the 
next day 1e-ave his party early and have a fairly solitary day, spending 
much time feeding. The consequences of one day's activity on the next 
LS a problem which has been little investigated in animals, but which 
may give insights into other factors affecting activity patterns. 
v. Variation between age-sex classes 
Since the target individuals Ln this study were all adult males, 
there are no data for the comparison of age-sex classes. It was 
apparent, however, that some differences existed. For instance, oestrous 
females travelling with a party of males frequently fed and nested late, 
sometimes in the dark. This may have been the consequence of feeding 
time being reduced by frequent sexual activity. These and other age-sex 
differences need to be investigated. 
b 
3. Time spent eating different foods 
1. Variation between months 
In every month there were differences 1n the availability of 
particular foods, reflected in the percentage of feeding time spent 
eating anyone food. Appendix 3.1 shows the variation in per cent of 
feeding time eating each food, ranked within months. The data were 
3.9 
taken from all observations (target individual only), and are not sampled 
equally between different times of day or individuals. Statistical 
comparisons between months are therefore omitted: the data show the 
extent of variety between months. Previous studies of chimpanzees in 
similar habitats (van Lawick-Goodall 1968, Suzuki 1969, Nishida 1974) 
found a comparable series of seasonal changes: in general chimpanzees 
find food in forest more in the wet than the middle of the dry season, 
and in grassland more in the dry than the wet. 
The most important changes 1n feeding time on different foods 
are summarized in Fig.3.l, showing monthly changes in feeding time on 
all foods which were eaten for more than 15% of feeding time in at least 
one month. From May 1972 to January 1973 a ser1es of different foods 
became favourites, and the data from July to September 1973 show that 
there was some similarity between the successive dry seasons. However, 
there was one major difference between 1972 and 1973: in the latter 
year P~ curatellifolia fruit were eaten for up to 41% of feeding time, 
whereas in the previous year this food was hardly eaten at all. This 
species appears to have a fairly regular two-year cycle in Gombe. In 
1972 very few of the trees produced any fruit, but they were in flo,,,er; 
and in 1973 they fruited but did not flower: in 1974 there was hardly 
any fruit again (Riss, pers. comm.). 
In 1973 there were no obvious major differences from 1972 1n 
the availability of foods other than P. curatellifolia, which was clearly 
Fig. 3.1 
Seasonal change 1n diet 
The ten foods shown were all those eaten on more than 15% of 
observation minutes (by target individuals) in any month. 
Data are taken from Appendix 3.1. 
SEASONAL CHANGE IN DIET : ALL FOODS 
EATEN > 15 °, FEEDING TIME IN ANY MONTH 
Per cent of 
total feeding 
minutes 
spent eating: 
1~72/3 
1973 
fruit 
leaf 
G 9all 
40 
° 2:j 
50 
° 
3
0
0] 
50 
° 
M J J A S ONO J 
L 
MonaOlholaljs ~Qtj 
Parjnari curatellifolja 
" 
______ ~ ______ Qllwohora e~celsa 
M J J A S 0 
Ficus vallis-choudae 
F 
A""'·,~,, '";'" 
N 0 J 1
1
, 
3.10 
a preferred food. This suggests that food availability Ln the 1972 may 
have been lower than in the 1973 dry season. 
LL. Variation between age-sex classes 
Age-sex differences Ln feeding could not be satisfactorily 
investigated in this study. McGrew (in press,a.) has described differences 
in the amount of time spent by males and females in eating termites, ants 
and meat. It LS certainly possible that similar differences exist in the 
eating of plant foods, though probably on a subtle level. For instance, 
since males tend to travel further than females at some times of year 
(Chapter 5), they may spend more time eating locally abundant foods with 
limited distributions than some females, according to whether the 
female's range includes the food. 
Van Lawick-Goodall (1968, 1973) has described how infants are at 
first unable to eat some foods which requLre special skills or strength 
to obtain. Whether infants eat more of some foods than adults is 
unclear. Kummer (1971) noted that female hamadryas baboons (Papio 
hamadryas) may forage further out in the slender branches of trees than 
males, and thus avoid competition. Adolescent colobus monkeys (Colobus 
badius) frequently occupy feeding sites on smaller stems than adults, 
but adults appear able to reach all food items by stretching; this 
difference in preferred feeding sites was therefore considered 
unimportant in intraspecific competition (Clutton-Brock 1973). Adult 
chimpanzees could normally bend or break branches towards themselves if 
they were unable to sit or hang on the edge of a tree (Plate 3.1). 
Differential physical abilities in occupation of feeding sites were 
therefore probably unimportant in mediating competition between age-sex 
classes. 
Plate 3.1 
Adult male feeding in outer branches 
Figan eats Pterocarpus angolensis flowers. By using arms and legs, 
as here, even adult males could normally feed from the slender outer 
branches of trees. An alternative strategy was to break thin branches 
and carry them to a more convenient feeding site. 
3.11 
Ill. Food Selection 
1. Introduction 
Several models have been proposed which emphasise the importance 
of food type (distribution, density and predictability) to primate 
social structure (e.g. Denham 1971, Eisenberg et al. 1972, Crook 1970a). 
Studies of closely related species have confirmed the value of this 
approach (e.g. Clutton-Brock 1974). 
The wide variety of food types eaten by many primates, living, 
it ~s often suggested, in habitats with superabundant food (e.g. Teleki 
1973), may give the impression that a variety of feeding strategies 
would yield equal nutritional benefits. If this were so, social 
structure would be less constrained than if only one feeding strategy 
were optimal. 
The factors determining food selection are therefore of crucial 
importance, ~but have been little studied in primates. Hladik et al. 
(1971) analysed the diet of sympatric species and showed that different 
items supplied different nutrients, but the extent to which food 
selection was nutritionally optimal was unclear. 
In the absence of nutritional analysis, one method of testing 
for optimisation of food selection is to compare observed variations ~n 
food selection with the predictions from optimisation models having 
different underlying assumptions. The goodness of fit may indicate 
whether or not food selection is optimal. Factors affecting food 
selection are therefore described in this section. 
Only food items recorded by dir.ect observation are included 
here. Plants were identified by reference to previous collections and 
by C. Kabuye and J.B. Gillett of the East African Herbarium. I am also 
grateful to Dr. T. Nishida for showing me the food plants of chimpanzees 
living in the Mahali Mountains. 
3.12 
2. Variation in diet 
~. Between populations 
Appendix 3.2 shows all food types recorded for chimpanzees in 
Gombe National Park, both during this study and at other times. The 
data are summarised in Table 3.4. Leaves which were eaten ("wadged") as 
absorbents with soft or wet foods (e.g. meat, eggs, Strychnos spp. fruit, 
honey) are not included in the list. Some wadging leaves were eaten on 
their own (e.g. Parinari curatellifolia) but most were not (e.g. Annona 
senegalensis, Combretum molle, Grewia mollis, etc.). Since these leaves 
appeared to be chosen for their structural characteristics (large, 
absorbent) rather than their nutritional content, they are omitted from 
the -food list. 
Table 3.4 
Number of food types recorded in Gombe (summary of Appendix 3.2) 
Different PClrts of the same plant species are scored as different food types. 
PLANTS ANIMALS etc. 
Studies which recorded Seen ~n F S B L P R C G Total I V M once 
only 
This study and van Lawick- 34 4 3 13 5 1 1 2 63 0 10 7 4 Goodall (1968) 
This study only 26 4 10 24 8 2 3 0 77 14 1 0 5 
Other observers only 12 0 3 6 1 0 2 0 23 13 0 7 4 
van Lawick-Goodall (1968) 14 6 3 11 1 
only 1 1 0 37 ? 2 1 0 
Total 86 14 19 54 15 4 7 2 201 27 13 15 13 
Key: F fruit, S seed, B blossom, L leaves, P pith, R res~n, C cambium 
or bark, G gall , I insects, V vertebrates, M miscellaneous. 
The 201 recorded plant foods include 17 from van Lawick-Goodall 
(1968) which may have been misidentified (van Lawick-Goodall, pers. comm.). 
The remaining 184 foods were from 141 species of plants, out of about 500 
so far identified from Gombe National Park (Clutton-Brock & Simpson, in 
prep.). Some foods are probably relatively unimportant to the diet, 
however; of the 163 confirmed records, 27 (17%) were seen eaten only 
once. Fifteen of these were commonly available, and their rare 
occurrence in the diet calls for an explanation. 
One possibility is that individuals were experimenting by 
3.13 
eating items not previously used as foods. Intermittent experimentation 
is doubtless important in the diversity of the tropical forest: 
Livingstone (1967) considered that long-term climatic changes in 
Africa were responsible for changing the composition as well as the 
extension of tropical forest; and chimpanzees occupy a range of habitats 
from primary rain forest (e.g. Jones & Sabater Pi, 1971) to savanna 
(miombo) woodland (e.g. Suzuki 1969). Kortlandt (1967) and van Lawick-
Goodall (1973) found wild chimpanzees to reject novel foods offered to 
them but many animals will try new foods only after becoming familiar 
with them, e.g. fish (Ware 1971), birds (Morrell & Turner 1970) and man 
(Campbell ~ Cuthbertson 1963). Japanese macaques (Macaca fuscata) are 
the only primate in which the acquisition of new foods has been 
observed, and there the initiators tend to be juveniles (Kawai 1965, 
Azuma 1973). Of the 15 common items eaten only once by chimpanzees, 
five were eaten by infants or juveniles and ten by adults. 
Only one of these foods is known to be regularly eaten elsewhere 
by chimpanzees: mango fruit (Mangifera indica) are eaten by West 
African chimpanzees (Kortlandt 1967), and there are many mature trees 
in Gombe. A mango was eaten by a juvenile (Goblin) but ignored at the 
time by adults of both sexes. On the other hand, when an infant picked 
up a seed of Garcinia huillensis (which are invariably spat out or 
swallowed without chewing) his mother flicked it away from between his 
teeth. Since the nutritional content of the other rarely eaten foods 
is unknown, the significance of their being tried cannot be determined. 
3.14 
It is possible that these foods may have become temporarily 
suitable because of subtle changes in the condition of the item or the 
eater: this would imply a very fine attunement of the animals to their 
environment. It is also conceivable that food choice may be affected 
by emotional state. The only time when the leaves ofPterocarpus 
angolensis (the fifth most abundant tree in Gombe; Clutton-Brock 1972) 
were eaten was by Jomeo, after having a colobus monkey stolen from him 
before he had started to eat it. 
Diet breadth in Gombe was found to be similar to elsewhere when 
variations in observation time were taken into account. Table 3.5 shows 
data from published studies: the number of food types recorded is 
positively correlated with the number of observation months (Spearman 
r 0.79, N = 9, P < 0.01). Nishida (1974) and this study record just 
s 
over 200 food types: over several years chimpanzees therefore ~robably 
eat between 200 and 300 food types. The ability to use such a large 
range of food items is probably important in permitting adaptation to a 
wide geographical range (Kortlandt & van Zon 1969), incorporating diverse 
and seasonally variable habitats. Bourne (1971) considered that diet 
breadth was important in ensuring an adequate supply of all nutrients, 
but this is still speculative. 
Differences in the diet items of different chimpanzee 
populations appear largely related to availability: some food types ln 
published lists from all the East African and some of the West African 
sites are eaten in Gombe. Similarly, the observed differences between 
the Northern and Southern communities in Gombe could be explained on the 
basis of availability: Mortis lactea, for example, was not seen to be 
eaten by individuals in the Northern community, nor were any trees of 
this species found in the Northern community range. 
t 
3.15 
Table 3.5 
Number of plant food types of chimpanzees in relation to length of study 
All studies listing more than 20 food types are included. 
Study area 
Guinea 
Guinea 
Uganda 
Tanzania 
Gabon 
Tanzania 
Tanzania 
Tanzania 
Tanzania 
No. of months 
of observation 
4 
4 
9 
10 
12 
15 
15 
60? 
78 
No. of plant 
food types 
34 
40 
35 
38 
141 
62 
78 
205 
80 
Reference 
Nissen (1931) 
de Bournonvi11e (1967) 
Reyno1ds & Reyno1ds (1965) 
Azuma & Toyoshima (1962) 
'1< 
H1adik (1973) 
Goodall (1963) 
Suzuki (1969) ~'c 
* Nishida (1974) 
van Lawick-Gooda11 (1968) 
* Feeding behaviour was the major aim of the study. 
However, there is some evidence that availability is not 
sufficient to explain differences in food selection. The use of tools 
may make otherwise inaccessible foods avai1ab1e~ Beatty (1951) observed 
Liberian chimpanzees eating the kernels of E. guineensis fruit after 
smashing them open with rocks. This has not been seen anywhere else, 
but Struhsaker & Hunke1er (1971) and Rahm (1972) found evidence that 
chimpanzees in the Ivory Coast smash fruit of two other species. 
Tool use aside, some species appear to be eaten in some areas 
but not others. In Nishida's list (1974) of food items from the Maha1i 
Mountains, Tanzania, 35 items appear which are known to occur in GoIDbe 
but have not been seen in the diet. This may partly be accounted for 
by difference of definition (Nishida may have included "wadging" leaves 
~n the diet) or by chance omissions from a complete list. However, 
there is still a real indication here of difference between the two 
popu1ations. Similar geographical variations in food selection have 
been found between popu1ations of Japanese macaques (Kawamura 1959), 
gorillas, Gorilla gorilla beringei (Schaller 1965) and ptarmigan, 
Lagopus mutus (Gardarsson & Moss 1970). 
3.16 
Provided there were no geographical differences in food quality, 
these variations in die't appear to suggest that the diet in one area 
was suboptimal or that different diets could be equally beneficial: in 
the latter case the differences could be called cultural in the sense 
used by Kummer (1971). However, a further explanation is that the act 
of selection of an optimal diet involves the periodic testing of "new" 
foods (Westoby 1974, Freeland & Janzen 1974). This will be discussed 
below (VIII/l/ii). 
ii. Betweenseasons 
The number of food items recorded per month was related to 
observation time (Fig.3.2). The asymptotic form of the curve suggests 
that about 60 items were eaten in typical months, and the smoothness of 
the curve indicates that chimpanzees eat a similar variety in all months. 
A comparison of the dry seasons in 1972 and 1973 shows that the 
actual foods eaten most were different (Table 3.6a). However, if foods 
are ranked according to feeding time, the cumulative per cent of feeding 
time spent up to each rank was similar between seasons (Table 3.6b, 
Kolmogorov-Smirnov ~ = 1, N = 10, n.s.). 
It is very unlikely that the relative abundance of the ten 
foods eaten most were as similar as the time spent eating them. The 
similarity of the results between the two seasons suggests that 
chimpanzees selected food items for both their nutritional qualities 
and in order to maintain a similar variety in their diet. 
t 
Fig. 3.2 
Diet variety in relation to observation time 
The data are taken from Appendix 3.1: all food types (plant, 
animal and miscellaneo~s) are included. 
60 
Total 50 
food 
types 40 
seen In 
30 month 
20 , 
10 
0 
0 1 2 3 4 5 6 7 8 9 
Observation minutes (x 1000) 
Table 3.6 
Differences, between successive dry seasons, in feeding time on 
the ten foods eaten for most minutes 
Data from July to Sept~mber 1972 and 1973; nine days of observation 
in each year, matched for target individuals. 
(a) Most eaten foods 
Rank of food 1972 1973 
1 
2 
3 
4 
5 
6 
7 
8 
9 
10 
B. bussei (S) 
E. guineensis (F) 
M. poggei (F) 
S. florida (F) 
P. curatellifolia (F) 
C. excelsa (G) 
G.platyclada (F) 
F .vallis-choudae (F) 
M. lobulata (1) 
G. huillensis (F) 
P. curatellifolia (F) 
E. guineensis (F) 
U. ni tida (F) 
D. condylocarpon (S) 
E. guineensis (P) 
C.excelsa (G) 
Ficus spp. (1) 
P. tinttorius (1) 
F ~ vallis-choudae (F) 
P. ~urate1lifolia (1) 
Key: F fruit, S seed, 1 leaf, P pith, G gall. 
(b) Per cent of feeding minutes spent eating ranked foods 
Rank of 
food 
1 
2 
3 
4 
5 
6 
7 
8 
9 
10 
% of feeding minutes ~n 
which food was eaten 
1972 1973 
18.2 21.9 
17.0 17.5 
7.7 5.0 
6.2 5.0 
5.1 4.0 
4.0 3.9 
3.4 3.7 
3.4 3.5 
3.4 3.5 
3.2 3.4 
iii. Between days 
Cumulative % of feeding 
minutes in which ranked 
foods were eaten 
1972 1973 
18.2 21.9 
35.2 39.4 
42.9 44.4 
49.1 49.4 
54.2 53.4 
58.2 57.3 
61.6 61.0 
65.0 64.5 
68.4 68.0 
71.6 71.4 
The number of plant food types eaten in one day was also 
fairly consistent (Table 3.7). 
3.17 
Table 3.7 
Number of plant foods eaten in one day 
Data from adult males, nest-to-nest observations, May 1972 -
September 1973. 
Northern males Southern males 
Number of ) Mean l3.3 10.5 
plant food ) Minimum 7 6 
types eaten ) Maximum 21 15 per day ) 
S.D. 3.8 2.5 
N (days) 35 l3 
No seasonal differences were apparent across males. However, 
from July to September 1972 Hugo ate a significantly wider variety of 
3.18 
plant foods per day (mean 17.8, N = 6) than other males in the Northern 
community (mean 11.5, N = 12; data lumped for 6 males) (Mann- Whitney 
u = 1, P < 0.001). Hugo's broad food choice was apparently related to 
the ranging ,behaviour at that time: males tended to stick to small 
areas within the community range. Hugo's range included the beach area 
of lower Kakombe valley, where there grew a wider variety of small 
edible leaves (mostly Convolvulaceae) than elsewhere. In the 1973 dry 
season there was no difference in daily diet breadth bet\veen Hugo and 
other Northern males (Mann-Whitney U = 17, n l = 3, n2 = 12, n.s.): all 
males tended to travel widely over the community range, but when Hugo 
or other males spent half the day in. lower Kakombe, they ate as many 
spec~es as Hugo had done regularly the previous year. 
The absence of overall seasonal variation in daily diet breadth 
suggests that chimpanzees maintain some variety in their choice, 
regardless of the abundance of particular foods. Chimpanzees frequently 
left food sources which still contained items apparently similar to 
those they had eaten: often the remaining food items may have been of 
3.19 
poor quality, but it was occasionally clear that chimpanzees left good 
food. For instance, on two successive days three males ate Brachystegia 
~r 
bussei seeds/more than eight hours from the same tree. No seeds were 
seen to be rejected: they were so abundant, and there was such little 
selectivity, that for 211 minutes on the second day one male ate without 
moving his feet. B.bussei seeds are highly nutritious, having very 
similar constituents to soya beans (Nishida, pers. comm.). Yet on both 
days the party left the tree in the late afternoon and ate a variety of 
food types in the evening. Another example of leaving good food is 
given below (IV/2/iv). Chimpanzees left sources with fruit remaining 
less frequently than with leaves remaining. 
There was no evidence that one day's food choice affected the 
next, except in so far as individuals learned the location of new foods. 
3. Factors affecting food selection 
i. Time of day 
Some foods tended to be selected at particular times of day. 
The clearest example was the leaf of Aspilia spp.: the midpoints of 
all recorded feeding bouts (on nine days from this study, eleven from 
Group-Travel charts) were before 0800 hours: chimpanzees ignored 
Aspilia when they saw it later in the day. In contrast, the twelve 
records of F; exasperata leaf being eaten (this study) occurred after 
1500 hours. Thorington (1967) found that squirrel monkeys (Saimiri spp.) 
ate more fruit in the morning and more insects in the afternoon, and 
Chivers (in press) has evidence of fruit being eaten earlier and leaves 
later in the day by siamang (Symphalangus syndactylus): Gombe 
chimpanzees were probably generally similar to siamang in this respect. 
Little work has been done on diurnal variation in food 
selection, and its significance remains unclear. Two examples from this 
3.20 
study may suggest explanations. 
First, Aspilia spp. leaves were eaten only when young. The 
plants were common in grassland, and up to I! m. tall. The leaves were 
eaten in an unusual manner. Normally the distal half was clipped off, 
and taken into the mouth on the tongue: instead of chewing it, the 
chimpanzee would leave his lower jaw slack while apparently rubbing the 
leaf against his upper palate, before swallowing it. Only one leaf, 
between 2 and 10 cm. long, was e'aten at a time, and much more slowly 
than those of any other species: in one bout the mean time to eat each 
leaf was 24.9 secs. (N = 40), compared to, for instance, 1.5 secs. 
(N = 36) for M. lobulata leaves, which are twice the size. Individual 
leaves were carefully selected: chimpanzees sometimes closed their 
lips over a leaf up to six times before tearing it off. Another unique 
feature of eating Aspilia leaves was that chimps sometimes wrinkled 
their faces above the nose when swallowing. 
This curious series of feeding techniques suggests that young 
Aspilia leaves contained an unusual component, which was perhaps 
responsible for their selection as food. Robinson (1974) gave three 
examples of plants in which alkaloid levels fluctuate markedly during 
the day: these compounds are known to affect food selection by animals 
(Freeland & Janzen 1974). Traditional folk-lore prescribes certain 
times of day as best for the collection of remedial herbs (Robinson 
loco cit.): the selection of Aspilia at dawn shows that this ancient 
knowledge may have been rooted in pre-human traditions. 
In the second case, the fruit of E. guineensis, there was no 
evidence of diurnal variation in quality, since they were normally 
eaten throughout the day: during all-day observations in both the wet 
season (Nov- Jan) and the 1973 dry season (Jul - Sep), 54% of feeding 
bouts (mid- points) (N = 16 and 42 respectively) were before and 46% after 
I 
\ 
3.21 
1230 hours. In the 1972 dry season, however, the fruit were eaten 
significantly more often in the morning, 81% of bouts occurring before 
2 1230 hours (X = 5.61, N = 32, P < 0.05). Comparing the dry seasons 
in 1972 and 1973, diurnal variations in food quality and climatic 
conditions were apparently similar. A major difference between the 
seasons, however, was- in the frequency of competition with baboons when 
eating E.guineensis fruit (Vn/2/i). Since the animals which eat this 
fruit in Gombe are mostly diurnal (primates and the palm-nut -vulture 
(Gypohierax angolensis), though nocturnal rodents and the palm civet 
(Nandinia binotata) may take some fruit), the time when most fruit are 
likely to be available ~s ~n the morning. This suggests that in the 
1972 dry season it was more important for chimpanzees to eat as many 
palm nuts as possible, which they did by eating in the morn~ng. The 
argument is speculative, but demonstrates that the factors affecting 
diurnal variation in food selection may be complex. 
~~. Individual differences 
Evidence of significant individual differences in food 
selection was not obtained for plant foods. Individual differences ~n 
meat - eating are discussed in the next chapter. An indication of 
differential preference for eating weaver ants (Oecophylla longinoda) 
was that on three occasions when a party of Southern males passed 
within 50 m. of a tree which regularly had ant-nests, Godi was the only 
male to go and eat some. 
Though individuals ~n a party commonly ate different foods this 
was more easily explained by lack of food sites at preferred foods or 
temporary preferences for var~ous reasons (satiation on one food, 
laziness in climbing, proximity to a particular individual, etc.). 
3.22 
iii. Social factors 
Thorpe (1963) classified three types of imitation. First, 
social facilitation is the performance of a behaviour already in an 
individual's repertoire, as a consequence of the performance of the same 
behaviour by other individuals. Second, local enhancement describes an 
increased tendency to respond to part of the environment as a consequence 
of another individual's response to it. Since feeding involves a 
response to the inanimate environment, imitative feeding behaviour must 
be classed as local enhancement. This certainly occurred sometimes. 
For instance, on 25 July 1972 Charlie walked while wadging Harungana 
madagascariensis fruit. Arriving at two other males who sat chewing 
P. curatellifolia fruit, Charlie looked at them, then on the ground, 
found and picked up some P. curatellifolia fruit, and ate them after 
spitting out his previous food. 
Further examples of local enhancement are g~ven below (IV/3). 
Though responsiveness to the food selection of party members appeared to 
be important in finding food, there was no indication that when two food 
types were freely available one individual's food choice affected 
another's. However, when one chimpanzee ate mud from a termite mound, 
others often copied him (seen 23 times). Since the mud sometimes came 
from a different mound, its selection appeared more related to the 
behaviour of the companion than to local variations ~n quality. 
Thorpe's third type is true imitation, which is the "copying 
of an otherwise improbable act or utterance". This is clearly an 
important long-term factor ~n the selection of foods eaten with the 
help of tools (e.g. termites, van Lawick-Goodall 1968), and perhaps 
hard-shelled fruits. Yerkes (1943) induced chimpanzees to chew and 
swallow filter paper by his own example: imitation may play a part 
in learning about new foods, but appears unimportant from day to day. 
3.23 
IV. How Foods were Found 
1. Introduction 
In describing t~e methods by which chimpanzees found _food, two 
problems are examined in this section. First, what do these methods 
tell us about the cognitive skills of chimpanzees? Second, what is the 
relationship between methods of finding food and social structure? 
Potential foods were distributed unevenly over a wide area; 
every day some sources were exhausted and others became available. 
Chimpanzees could rarely identify food sources from a distance, as some 
animals can (e.g. griffon vultures, Gyps spp., Houston 1974; lion, 
Panthera leo, Schaller 1972) because of the frequently thick vegetation 
and the small food items. This section shows the extent to which 
chimpanzees are able to avoid merely searching at random, as they 
would appear to have to do, in order to find food. 
The - strategies by which different species find food have often 
been considered important determinants of social organisation. For 
instance, Horn (1968), Murton (1971) and Ward & Zahavi (1973) considered 
that the adaptive significance of colonial nesting lies in the 
increased efficiency with which individual birds find food. The 
methods used by chimpanzees have hardly been described, but Reynolds & 
Reynolds (1965) implied their importance in suggesting that survival 
of the community depends on individuals communicating information about 
food sources. This idea appears to involve group selection, which 
appears to be a very weak selective force (Williams 1966), but anecdotal 
evidence suggests that some calls do convey information about the 
location of food sources (Reynolds & Reynolds 1965, Sugiyama 1968, 
van Lawick-Goodall 1968). The context and consequences of 
cOll@unication about food sources are therefore examined, and are later 
considered in relation to social structure. 
3.24 
Methods by which chimpanzees found food are g~ven below. When 
the subject arrived alone at a vacant food source (i.e. when no 
communication had occurred), it was often difficult to estimate the 
relative importance of· chance or memory in bringing him there: and 
when chimpanzees were in a party it was difficult to be sure which 
individual determined the route. Quantitative data on the importance 
of different methods of finding food are therefore impractical: 
evidence of their occurrence is given for each case by four examples. 
Methods of finding food are ordered according to the apparently 
decreasing importance of chance. 
Instances when a chimpanzee gave long- distance calls on 
arrival at a food source could be counted. Variation in the frequency 
of such communication ~s analysed in relation to different foods and 
different seasons. 
2. Foods found without communication 
~. By chance 
No habitat could be so predictable that foods could not 
sometimes be found by chance, as the examples below indicate. However, 
convincing examples of chimpanzees happening on a food were infrequent. 
(a) Monartthotaxis poggei forms a continuous shrub layer under 
forest canopy and normally produces its bright orange fruit from May 
to July. The fruit are then eaten in large numbers, and chimpanzees 
often ignored patches with few fruit. Out of season (e . g. December) 
single fruit occasionally occurred. Three times individuals walking 
through forest took such fruit en passant. 
(b) P. curatellifolia fruit were chewed and then spat out, 
sometimes up to a kilometre away from their source. Chewed fruit were 
twice found and eaten far from P. curatellifolia trees. 
3.25 
(c) When trees begin to produce a food whose early stages are 
hidden chimpanzees may learn about it by chance. On 11 October 1972 
Charlie was leading two males in forest ·when a colobus troop gave alarm 
calls ahead of them. They watched the colobus for three minutes, but 
while his companions continued to look up, Charlie looked down at the 
ground, picked up and smelled some fallen Morus lactea fruit, looked 
again into the 25 m. canopy, then climbed. The others followed when 
they saw Charlie eating. 
(d) On 30 July 1972 a male walking alone suddenly stopped and 
gave a soft grunt: he was looking at a patch of Asystasia gangetica 
leaves to his right. He changed his direction to eat the leaves, then 
returned to his path and continued as before. 
~~. By knowing where to look 
While walking, chimpanzees normally look forward at eye level. 
In the following examples individuals looked in unusual directions 
before having seen the food source itself. 
(a) In example (c) above, a chimpanzee looked up after finding 
fruit on the ground. This frequently occurred, but only in situations 
where the food items could be expected to have fallen. This was shown 
experimentally. A chimpanzee finding palm-nuts under a palm tree 
inspected above him: when another found bananas under the same tree 
he did not inspect the tree. 
(b) Uvaria angolensis fruit are found in the lower canopy of 
forest, growing on a woody vine similar to the many others in which it 
is entangled. Iu· 1972 U ~ angolensis fruited abundantly and chimpanzees 
would look up and scan the canopy when walking in forest. For example, 
three males who had walked through wooded grassland for nine minutes, 
without inspecting, entered forest and began to inspect the canopy as 
they walked. After six minutes, during which he looked up for at least 
5 seconds every other minute, Evered saw some fruit, which he ate for 
42 minutes. 
(c) On 6 July 1972 Jomeo left camp alone at 1025: he had 
arrived at 0820 to find Hugo eating bananas, but Jomeo got none. At 
1721 Jomeo climb~d a tree and looked at camp, 0.5 km. away: s~nce he 
left it he had been alone, and there had been no calls since 1052. 
Jomeo was apparently checking to see if any bananas were being eaten, 
~n which case he could possibly get some. 
(d) In open grassland chimpanzees did not inspect at ground 
level except in January 1973: at this time pith of the grassPanicum 
max~mum was eaten. As in the case of U. angolensis fruit (above) 
individuals appeared to look for the food when in suitable habitat, 
but before actually seeing the food-bearing species. 
iii. By travelling to a productive area 
In these cases chimpanzees appeared not to know the exact 
location of food sources, but visited areas where food sources were 
likely to occur. 
3.26 
(a) Aspilia spp. leaves were eaten exclusively in the early 
morning. On 5 January 1973 Mike left his nest at dawn and walked for 
one minute due east into wooded grassland, where he found an A. rudis 
plant 0.4 m. high. He could not have seen the plant from his nest or 
at any time on 4 January. After eating A. rudis leaves Mike left the 
way he had come, and did not return all day. There were three further 
examples where individuals appeared to make special journeys to wooded 
grassland at dawn, looked for and ate A.rudis leaves, and then changed 
direction. 
(b) Brachystegia bussei trees occurred only in upper miombo 
woodland. On 15 August 1972 Godi left riverine forest and walked 
through semi - deciduous forest. His choice of path seemed odd, since 
3.27 
there was no food in this habitat: he did not inspect the canopy. 
After 21 minutes he reached upper miombo woodland and his behaviour 
changed: walking along the contour instead of ups lope, he looked into 
B. bussei trees. He ate the seeds from three trees (mean bout length 
9 minutes) before reaching a tree where he ate for 215 minutes. There 
were two further examples of journeys apparently made only in order to 
reach upper miombo, where B. bussei trees were then inspected. 
(c) Uapaca nitida trees also occurred only in upper miombo. On 
15 August 1973 Godi walked for 15 minutes through semi-deciduous forest, 
emerging into upper miornbo. His walk immediately became slower and he 
inspected in U. nitida trees. In the next 82 minutes he ate fruit from 
seven different trees, before leaving back towards semi-deciduous forest. 
There was one other occasion when an individual travelled to a U. nitida 
patch, without any apparent external stimuli, and after eating left at a 
different angle. 
(d) , Mellera lobulata leaves were eaten most frequently in the 
evening in the 1972 dry season, and normally occurred in gullies in 
forest. Eight times chimpanzees reached the top of a gully in the late 
afternoon, changed their direction of travel to walk down the gully, and 
ate M~ lobulata leaves from a number of patches found en route. 
Similarly there were two occasions when individuals had not drunk water 
all day, but came to it after changing their path to walk down a gully. 
1V. By returning to known food sources 
As in the other examples, those here are selected because of the 
absence of any long- distance orienting stimuli. Unusual or coincidental 
behaviour indicated that the individual knew the whereabouts of the 
food source. 
(a) On 15 August 1973 Godi stopped and looked at a Pterocarpus 
tinctorius tree above him: he had just finished eating leaves from 
3.28 
another P. tinctorius. Nine hours later, after walking straight 
through closed canopy forest for five minutes, he returned to the tree 
he had examined and fed in it for 46 minutes. At no time after leaving 
it could Godi have seert the tree until he was a few metres from it. 
His route was through forest without obvious paths. It appears, 
therefore, that Godi not only remembered that the tree had food, but 
was als.o able to return directly to it, from a new angle, without being 
able to see the tree or any known paths. 
(b) On 19 January 1973 Charlie left a party and walked rapidly 
east for seven minutes in forest, turned to walk over a termite mound, 
and climbed via an access tree into an E. guineensis where he fed for 
25 minutes. The termite mound and access tree blocked any view of the 
food tree or fallen fruit. After eating he climbed down and walked back 
west to rej oin his party. The only obvious explanation of Charlie IS 
behaviour was that he knew the location of the palm tree, and that it 
had ripe fruit. 
(c) On 14 August 1973 at 0958 Godi approached within 4 m. of 
anE. guineensis under which there were fallen fruit. He suddenly 
stopped, listened, and walked back: buffalo were heard runn~ng away in 
the direction in which Godi was looking. He circled away to the south 
and sat, self-grooming intermittently, alerting quickly to any noise 
such as the rustle of birds in leaf-litter. After 45 minutes Godi 
appeared more relaxed, and he travelled and fed. He returned northwards 
in 1321, and at 1326 he showed signs of nervousness (walking slowly, 
glancing around, orienting to noises normally ignored). At 1333 he 
arrived at the palm he had approached at 0958. The forst vegetation 
was thick, preventing any distant views, and Godi had arrived from a 
totally new direction. His cautious approach showed that he realised 
he was in the same place as when he had heard the buffalo. He fed in 
3.29 
the palm for 49 minutes. 
(d) During the study period Southern males were never seen 
north of Kakombe stream unless they had visited camp: it can be assumed 
that when they did not visit camp they were south of Kakombe stream. On 
15 May 1972 I found a party of Southern males in Mkenke. As happened 
another time in 1972, the observer seemed to be a conditioned stimulus 
for bananas: the party walked north and reached camp three hours later. 
During their stay in Kakombe the Southern males twice walked very 
directly to trees where they ate ripe figs: they reached the second 
fig in the dark after leaving their nests before dawn, and having fed, 
the chimpanzees changed their direction of travel. Charlie, who was 
clearly leading the party, was the most recent visitor to camp, having 
been there five days earlier. He apparently knew the location of the 
figs, and that they had food, the latter from his recent visit. 
An example similar to those above was cited by van Lawick-Goodall 
(1968) and seen in this study: individuals sometimes picked pieces of 
grass before they could see a termite mound, would arrive at a mound 
after a minute or two, and use the grass to catch termites. Since 
feeding rarely required such preparatory activities, evidence of this 
type was seldom available. 
One indication of the importance of memory in finding food 
was the development of my skill at doing so. It was clear that the 
longer I observed in any month the better I became at finding food 
sources and predicting the food sources of the target. Although Gombe 
chimpanzees foraged in a large and complex habitat, they appeared to 
learn intimately a temporary pattern of food distribution. The ability 
to remember the location of many food sources, all with changing amounts 
of food available, was apparently a very important skill. 
The discovery of food sources whose location was communicated 
3.30 
by other chimpanzees is discussed in the next section, and the process 
of learning about new food sources without communication is considered 
below (Feeding methods: inspection). 
3. Communication about foods within parties 
L. Incidental communication 
In this category are observ.ations when one individual perceived 
what another was doing, and thereby learned the location and/or quality 
of a food source. 
(a) On 22 Septembe1! .1972 Charlie led Godi and Rugh in extended 
line when Rugh stopped and inspected a palm tree which the others had 
passed. Two minutes later Godi looked behind him: Rugh was gone. 
Godi walked back and saw Rugh eating in the palm tree, watched him 
briefly, then joined him after a two-minute walk. Charlie could not 
see Rugh, but saw Godi's movements, and followed him. They all fed 
in the tree~ for 40 minutes. 
(b) On 26 August 1972 Faben had a wound, and travelled little 
and slowly. After lying on the ground for 135 minutes he watched Jomeo 
walk away, climb into a Ficus vallis-choudae and eat fruit. Only after 
Jomeo had stretched for his second fruit did Faben stop watching: he 
approached and climbed into the same tree, where he fed. 
(c) On 29 August 1973 Satan ate weaver ants when he heard the 
sound of wood being torn apart: behind him Evered was breaking a 
rotten palm. Satan left his ants and joined Evered, and together they 
opened up a bee nest and ate honey. Similar examples occurred when 
individuals heard others smash Strychnos spp. or Taberuaemontana holstii 
fruit. 
(d) On 4 August 1973 Evered led Rugo up a slope in woodland. 
While Evered walked, Rugo collected a branch of S. quirtqueloba pods, and 
3.31 
followed carrying it. Evered turned and saw Rugo's find, looked up and 
saw pods on another S. quinqueloba tree above him, climbed in and fed. 
In these examples the communicants were adult males: s~nce 
there appear to be few ' differences in food preference among adult males, 
the fact that one was eating indicated that the food was suitable for the 
other. Observations indicated that the identity of the first feeder was 
important. Thus the juvenile Goblin, who regularly ate unripe foods, 
was three times ignored by adult males when he appeared to have 
discovered a new food source. A clear example of the importance of who 
was feeding occurred on 23 December 1972. Mike was travelling with n~ne 
others when he heard and then saw palm nuts dropping from a tree 5 m. 
high. Re walked to the tree and had climbed to I! m. when he looked 
above him: feeding in the tree were a male and female baboon but no 
chimpanzees. Mike climbed down, inspected the fruit lying on the 
ground, climbed to 3 metres and inspected the tree, then climbed right 
up and fed for 21 minutes. 
~~. Vocal communication 
Individuals may grunt softly when feeding, the frequency and 
intensity of grunting varying with the quality of the food, the 
individual, and other factors: no detailed studies have been made. 
(a) On 15 May 1972 Charlie walked away from two grooming males 
and disappeared behind thick vegetation. The groomers ignored his 
departure, but looked up when they heard Charlie grunt. One minute 
later Charlie grunted again, and the groomers rose and walked straight 
to Charlie, who was eating U. angolertsis fruit: his companions looked 
and found some too. 
This type of observation was seen fairly commonly. It 
indicates that a grunt tells others that the caller is feeding. In the 
following examples the call appeared to convey more specific information. 
3.32 
(b) On 22 July 1972 Rumphrey sat eating leaves when he heard 
a scream 100 m. away. Re immediately ran 1n that direction, and when 
I arrived I found him with Sparrow and Rugo, the latter clutching a 
dead colobus. The quality of the scream appeared to have prompted 
Rumphrey's unusual reaction: on arrival he tried hard to get some meat 
by grabbing it. Sparrow had probably called after finding .Rugo with 
the meat. 
F. Plooij (pers. comm.) saw Passion and Pom embrace when they 
heard grunts given by nearby males: they travelled quickly, found two 
males eating a bird, and tried to get some meat themselves. Again, the 
unusual quality of reaction to the grunt indicates that the hearers 
learned not merely that the caller was eating, but that he was eating 
something special. 
(c) Calls given by an individual who had found honey also 
elicited unusual reactions (seen three times). For instance, on 19 July 
1972 S1X chimpanzees were spread out and engaged in different activities 
in lower miombo woodland, when at 1737 Figan gave a low grunt. Rugo, 
who could not see the caller, immediately ran towards him, followed by 
Satan. Both watched Figan and then lost interest: there was little 
honey available, and Figan left after one minute. 
(d) Occasionally grunts appeared not merely to inform about 
the location and quality of food, but also about the intentions of the 
caller. For instance, on 26 January 1973 Goliath and Willy-Wally were 
travelling through forest where D.lucida fruit were available but not 
abundant. Goliath appeared to be directing the travel patterns of the 
pair, and at 1800 my field notes recorded: "Goliath not prepared to 
get settled unti 1 he finds a big patch of fruit". At 1806 Goliath was 
over 15 m. from Willy-Wally, who looked at a good patch of fruit above 
him, grunted, and looked at his companion. Goliath saw Willy- Wally 
look at the fruit, looked up himself, and grunted: Willy-Wally grunted 
again, and they both climbed up and fed for 70 minutes. The initial 
grunt drew Goliath's attention to the food: only when Goliath had 
shown his interest did Willy-Wallyclimb up to feed. 
3.33 
These examples ' have been selected because it was fairly clear 
that the second individual learned something about the food from the 
first. Indirect communication undoubtedly occurs repeatedly when 
individuals with different experience of current food distribution 
travel together. The mechanism, however, may often be more subtle than 
in the examples given. Menzel (e.g., 1971) has shown in a group of 
captive chimpanzees that if only one individual knows the position of a 
food source he is followed to it by others. (In Menzel's case 
information transfer was clearly goal-directed: the "leader" tried to 
make the others follow. However, this may have been an anomaly of the 
captive situation, where the chimpanzees did not like to be alone.) In 
the wild, of course, all individuals in a party have some information 
about location of foods. In a second experiment Menzel showed that 
when two chimpanzees had information, each about different piles of 
food, they tended to be followed in numbers which corresponded 
app.roximately with the size of the food pile they were going to (though 
the number of followers was also affected by which individuals were 
"leaders") . 
Observation ~n the wild suggested that intention movements 
could communicate information about foods. When Faben saw Evered climb 
rapidly into the canopy, Faben climbed too: but he could not have seen 
any food, because there was none where he climbed: Evered had taken the 
only bunch of U. angolensis fruit available. Evidence that information 
about the discovery of new foods is spread by recombination of parties 
comes from the original establishment of the artificial feeding area: 
the first chimpanzee to be given bananas was David, and all subsequent 
3.34 
arrivals came either with him or previous visitors (van Lawick-Goodall, 
pers. comm.). The significance of this form of communication ~s 
discussed in Section VIII. 
4. Communication about foods between parties 
1. Food-calling and its effect 
Long-distance calls (pant-hoots) were g~ven ~n many different 
contexts (van Lawick-Goodall 1968, Bygott 1974), one of which was 
arrival at food. Pant-hoots were variable in nature, forming a graded 
series among which observers could not accurately classify different 
types by ear: however, extreme forms were distinguishable, and one, 
the "food-pant-hoot" was given only on arrival at a food (H. Plooij, 
in prep.). Since pant-hoots which could not be classified were also 
given at food, food-calls have here been defined as pant-hoots given by 
an individual within two minutes of starting to eat. Individuals called 
either as they were approaching the food or shortly after beginning to 
eat. 
Examples of the effects of food-calls are not g~ven since the 
data were quantifiable. All feeding bouts at which the target gave a 
food-call during the 1972 dry season were listed and matched for food 
type and length of bout when food-calls were not g~ven. Not all bouts 
initiated with a food-call could be matched, but a sample of 26 were 
paired, with a maximum discrepancy of 10 minutes between bouts with 
and without food-calls. Twenty-six bouts were similarly matched from 
the 1973 dry season. The number of bouts during which individuals 
arrived was scored. The data are shown in Table 3.8 and summarised 
~n Fig.3.3. 
Table 3.8 shows that out of 52 bouts when a food-call was 
given, individuals arrived during 17, compared to during 3 bouts when 
Fig. 3.3 
Aggregation probability in relation to food-calling 
The data are summarised from Table 3.8. Individuals were 
significantly more likely to arrive at a target who had started 
his feeding bout with a food-call than at one who had not. 
0/0 food-t::ouS 
in which 
aggregation 
occurred 
40 
20 
o 
o 
No food-
calls 
>1 food-
call 
For each 
for length 
sample, N = 26 food-bouts, 
and food-type . 
matched 
_ Jul - Sep '72 
o Jul - Sep '73 
P < 0·01 
wi th in seasons 
J 
3.35 
no food-calls were glven: the difference is significant (X2 = 9.8, 
d.f. = 1, P < 0.01). A larger sample was considered unnecessary 1n 
view of the clarity of the result. The difference in the probability 
of individuals arriving during the target's feeding bout, according to 
whether or not he gave a food-call, was similar between seasons 
(Fig.3.3). There was no indication that the effect of food-calling 
during the wet season was different from during the dry seasons. 
Table 3.8 
Effect of food - calling on number of individuals arriving 
Only food-calls given by the target are scored. Within seasons feeding 
bouts were matched for food type and as closely as possible for bout 
length. From the 1972 data all recorded food-calls were included 
provided that bouts could be matched; from the 1973 data 26 food-calls 
were selected according to how closely they could be matched. 
Total number of 
bouts samp led 
Sum of minutes 
in all bouts 
Number of target 
males sampled 
Number of bouts 
during which at 
least one 
individual arrived 
Jul - Sep 1972 
No Food-call 
food-call given 
26 26 
499 495 
10 8 
1 9 
Jul - Sep 1973 
No Food- call 
food-call given 
26 26 
687 667 
7 7 
2 8 
The individuals attracted by food-calls were evenly distributed 
between sexes: in the 17 bouts initiated by food-calls during which 
individuals arrived (Table 3.8) 23 of the adults arriving were male, 
and 18 were female. 
11. Individual variation 1n the frequency of food-calling 
From observations of males arriving at food sources it appeared 
that females rarely gave food-calls: this impression was confirmed by 
3.36 
Tutin's (pers. comm.) direct observation of females. Food-calling can 
safely be regarded as an adult male behaviour. 
Within adult males, there appeared to be considerable individual 
variation in the frequency of food-calling. The sample of all-day 
observations, however, was too small to allow a rank order to be 
determined, because differences between seasons and immediate contexts 
(e.g. party size, type of food source, food-calling tendency of 
companions) outweighed individual differences. Individual differences 
in the frequency and consequences of food-calling are an important 
problem whose solution may reveal the function of this behaviour more 
clearly. The only clear difference among Northern males was that in 
all seasons Evered consistently gave more food-calls per bout (mean 
0.32, N = 9) than Rugo (mean 0.03, N = 6): Faben, Figan, Satan and 
Jomeo were intermediate. Data on Rum~hrey and Mike were inadequate. 
~~~. Frequency of food-calling ~n relation to food quality 
Mean bout length is used here as a crude measure of the 
quantity of food at a source. Table 3.9 shows the relationship between 
bout length and the frequency of food-calling. 
Table 3.9 shows that the percentage of bouts at which food-calls 
were given was significantly correlated with mean bout length (Spearman 
r 0.68, N = 11, P < 0.05). This suggests that a factor affecting the 
s 
frequency of food-calling was the amount of food available. When bout 
length was compared within food types between times when a food-call 
was or was not given, a similar relationship was found: mean bout 
lengths were significantly longer following a food-call (Sign test, 
N = 16, x = 3, P < 0.05: data from 14 all-day observations, July to 
September 1973; all foods at which a food-call was given were scored.) 
Table 3.9 
Frequency of food-calling ~n relation to mean bout length 
(between food types) 
Data are from 14 all-day observations (target individual only), July 
to September 1973. All foods observed eaten in more than 10 bouts 
are included. 
3.37 
Food type N Mean bout length % bouts with (mins.) food-call 
Water 23 0.3 4.3 
Clerodendrum schweinfurthii (G) 12 1.2 0.0 
Ficus urceolaris (L) 25 2.4 8.0 
Mellera lobulata (L) 21 4.2 4.8 
Asystasia gangetica (L) 24 4.8 4.2 
Parinari curatellifolia (L) 11 11.0 45.5 
Pterocarpus tinctorius (L) 32 l3.l 18.8 
Diplorhyncus condy locarpon (S) 16 15.9 50.0 
17.1 13.0 Elaeis . . (F) 46 gu~neens~s 
Ficus vallis-choudae (F) 16 18.1 37.5 
Parinari curatellifolia (F) 60 27.5 18.3 
Key: N number of bouts, G gall, L leaf, F fruit. 
The data fit my sUbjective impression that chimpanzees were 
more likely to call if they saw that food was abundant at the source. 
The alternative explanation is that the effect of the call was to make 
an individual eat longer: this seemed unlikely. 
~v. Seasonal variation ~n the frequency of food-calling 
Since the frequency of calling varied with the time of day and 
between individuals, seasons were compared by matching all-day 
observations of target individuals. The median number of food-calls 
given per food-bout was 0.0 in the 1972 dry season, 0.08 in the 1973 
dry season, and 0.13 in the wet season. In testing the significance of 
these differences, the number of food-calls given was scored in relation 
to the total number of calls (pant-hoots) given, in order to allow for ~~ 
3.38 
higher probability of food calls occurrLng by chance when more 
calling occurred overall. Results are given in Table 3.10. 
Table 3.10 
Significance of seasonal differences in frequency of fOdd~calling 
All- day observations of target individuals are matched between seasons: 
comparison is of the number of food- calls per food- bout, divided by the 
total number of calls (target individual only). All tests are Wilcoxon. 
Seasons compared 
Dry '72 & Wet 
Wet & Dry '73 
Dry , n & Dry , 73 
T 
2 
25 
o 
N 
7 
10 
8 
P 
< 0.05 
n.s. 
< 0.01 
There were significant ly fewer food-calls in the 1972 dry than 
Ln the other seasons. The significance of this result will be discussed 
below· (Section VIII) . 
V. Feeding Methods 
1. Introduction 
Feeding methods of chimpanzees have been described by van 
Lawick- Goodall (1963, 1968) and Hladik (1973), so there is no need for 
a general account here. The purpose of this section LS to examLne 
behaviour relevant both to the finding and identification of food and 
to feeding efficiency, and to document observations which add to 
previous reports. 
2. Inspection 
L. Methods of inspection 
Potential food items were often carefully examined by sight, 
smell, feel, taste or a combination of these. Experiments have shown 
that chimpanzees are able to discriminate finely between different 
3.39 
visual cues (e.g. Menze1 & Draper 1965). Inspection is considered 
separately for foods whose eating was beneficial to the plant and those 
for which eating was destructive. 
Colour changes occurred ~n many fruit at the time when they 
became more palatable to me, and chimpanzees tended to select the reddest 
fruit at any food source. However, they were very sensitive to bad 
fruit: many D. lucida fruit were dropped immediately after being picked, 
and these invariably had insect larvae and frass inside them, the only 
visible indication being a tiny hole on the outside of the fruit. Only 
fruit which were nearly r~pe were examined carefully: green fruit which 
would later become brightly coloured (e.g. Pycrtanthus ango1ensis) were 
ignored. It was clear that chimpanzees were good botanists, ab le to 
judge the condition of many, especially arboreal, fruit by sight. This 
was an important skill whose consequence was the saving of energy and time. 
Fruit were normally picked very selectively: s~ze, colour and 
softness of fruit varied on the same tree, and often only a small 
percentage of the total fruit available was eaten. Fruit without 
parasites were rarely picked if not subsequently eaten. It seemed very 
likely that on most occas~ons when chimpanzees left a tree the fruit which 
remained were comparatively unripe: this was confirmed for those I could 
reach. However, in all seasons chimpanzees frequently stopped eating 
leaves even when sources appeared to be abundant with leaves of identical 
age. 
Standards of palatability did not appear to be constant. Green 
fruit of, for instance, Ficus grtapha1ocarpa, F.capensis and D. lucida 
were sometimes eaten and at other times ignored. Four times individuals 
selected green ~ucida fruit when both red and green were available, 
each time being after a long bout of eating only red fruit. These green 
fruit were larger than most. It was not clear what factors affected 
variation in acceptability, but they appeared to operate in the 
chimpanzees rather than their foods. 
3.40 
Selectivity was' less marked for foods other than fruit. When 
seeds were eaten, individuals tended to eat from one site and then 
either leave immediately or spend a very long time at the tree, eating 
all the seeds from each site visited : this occurred at B. bussei, 
S. quinqueloba and D.condylocarpon trees Ce.g. IU/2/iii). Visual 
inspection of pods appeared inadequate for judging seed quality 
effectively once the pods . were full-sized, and seeds were normally 
tasted. Leaf quality was apparently generally judged by size, young 
leaves being smaller and mo~e palatable. A remarkable demonstration 
of their botanical expertise was given by chimpanzees finding the 
young shoots of Dioscorea schimperiana: the stem of this climber, 
superficially similar to many other vines, was several times pulled at 
ground level to bring down a shoot which had been hidden in dense 
vegetation 3 m. or more above the chimpanzee. In this case the feeder 
did not have to see the shoot to know it was there: its stem would not 
be vertical if the shoot had been eaten, and so long as the stem was 
not lignified the shoot ~vould be edible. 
~~. Timing of inspection 
Inspection occurred during travel, on arrival at a food source, 
and while feeding. When one individual in a party stopped and looked 
at a tree, his companions frequently did so too. It was clear that 
individuals could not only learn from inspection that food was 
available, but also remember the knowledge. An example was given 
C!V/2/iv/a) of a chimpanzee returning directly to a food source which 
he had inspected but not tried: when he first saw it he had already 
just been eating the same food in a different source, which may explain 
3.41 
why he did not eat immediately from the tree he inspected. Similar 
examples were often seen. It is possible, as MacKinnon (1974) suggested 
for orang-utans (Pongo pygmaeus), that chimpanzees may be able to assess 
when a food source will -be suitable some days before it is, but clear 
evidence of this was not obtained. 
There were times when individuals concentrated on inspecting 
particular species. For instance, on 15 August 1972 Godi inspected 
many palm trees during three journeys between palms where he fed. 
After leaving the last, he did not inspect palms for the next nine 
hours. On this occasion he stopped inspecting after finding and eating 
the food. , However, on 14 August 1973 Godi inspected many Strychnos spp. 
before nesting without finding any good fruit. The next morning he 
walked through a similar habitat without eXamlnLng the Strychnos trees. 
An individual who was not feeding occasionally approached and 
peered at one who was (Plate 3.2). This happened most often on arrival 
at a feeding party. In 30 examples of peering, 22 different foods were 
being eaten, and they were ordinary types. No reason suggested itself 
why an individual should have been able to find out more about the food 
source in this than other ways. Peering may therefore not have been 
important in food inspection. The possibility that its significance 
was more as a form of greeting behaviour is indicated by the frequency 
of peers given by and to different age-sex classes (Table 3.11). 
Peers were given mostly by young and received mostly by adult 
males (~ test is illegitimate with females included separately because 
expected values are too low: if females and males are taken together, 
2 X = 28.72, d.f. = 1, P < 0.001). It was clearly a socially structured 
behaviour, and the evidence suggests it was a form of greeting. Similar 
peering behaviour is shown by baboons and vervets when approaching an 
individual with food in his mouth (A1tmann & A1tmann 1970). 
Plate 3.2 
Peering 
Sparrow (adolescent female) peers at Figan (adult male) who eats 
Parinari curatellifolia, fruit collected from the ground. More 
ripe fruit were available, but Sparrow approached Figan before 
eating. 
3.42 
Table 3.11 
Peering in relation to age-sex class 
All observations are included (except for mother-offspring) when one 
individual who was not , feeding approached one who was, and watched him 
feeding from a few centimetres away. Numbers of independent 
observations are shown. 
Age-sex classes 
Adult males Adult females Infants to adolescents 
Peers given by 
Peers given to 
3. Foodmanipulation 
3 
28 
i. Variation with food items 
6 
1 
21 
1 
Differences 1n the size, shape and structure of food items 
imposed corresponding differences in th~ir manipulation. 
Many fruit, after detachment or collection with hands, lips 
or occasionally feet, were squeezed and rubbed in the mouth, but not 
chewed. In this way seeds could be spat out or swallowed without 
biting into them: the seeds of juicy fruit invariably tasted bad to 
me. Chimpanzees have muscular lips and a heavily ridged palate which 
are probably a specialisation for squeezing and crushing fruit: 
fruit-bats have similar features (Brambell 1972). 
Not only did each species tend to be eaten in an idiosyncratic 
manner, but the specific technique was very constant. When eating 
Garcinia hui11ensis fruit, for instance, males placed only one in the 
mouth, squeezed it (tongue against the palate?), took out the 
compressed skin and held it between the lips or thumb and forefinger 
. 
while crushing the extruded flesh inside the mouth. A few seconds 
later the skin would be returned to the mouth, squeezed and removed 
3.43 
agaln. Baboons, on the other hand, invariably peeled the skin off 
with their fingers before eating C. huillensis fruit. The extent to 
which feeding technique was shaped by the interaction between fruit 
morphology and mouth structure was indicated by giving C. huillertsis 
fruit to . a tame infant chimpanzee (Kobi). A whole fruit was initially 
rejected, so half a fruit was given. Kobi tasted it, ate a portion 
after partly peeling the skin away, then spat out the skin. Hith his 
next bite he discovered that a seed tasted nasty, and spat it out. He 
was then given a whole fruit, which he put in his mouth and broke open 
in the "proper" manner, subsequently holding the skin in the same way as 
wild chimpanzees. 
Apart from fruit and Aspilia spp. leaves (III/3/i), plant foods 
were chewed with molars. Techniques of eating again varied with the 
speCles. Only when eating F. urceolaris, for instance, did individuals 
sometimes fold a number of leaves together (after picking them 
individually) ~ and eat them as a single wad. 
Canines and incisors were used together to open some large 
fruit. The importance of these teeth was indicated by the fact that 
Hugo, an old male with a broken set, was the only adult who normally 
smashed the fruit of Tabernaemontana holstii while others bit it open. 
This physical disability must have been a nUlsance: once he hit a 
frui t against a root 63 times before it cracked. 
Use of a previously undescribed tool was seen once in eating 
plant foods, when Evered wiped out the shell of a Strychnos fruit with 
a dead leaf, and then sucked and licked the wet flesh from it. Similar 
behaviour has been seen by chimpanzees wiping the brains out of mammal 
skulls (Teleki 1973). 
Crushed leaves were twice seen. to be used as a sponge to obtain 
water, as described by van Lawick-Coodall (1973). Strangely, one of 
1 
3.44 
these times was when Hugo was sitting on an island in a stream: all he 
had to do to drink was to bend down. 
ll. Individual variation 
Observations by van Lawick-Goodall (1968) of maJ or variation 
In feeding techniques were considered to be due to individual differences. 
This idea was not supported in this study, In which variations appeared 
related to the structure of the food. One kind of food could be eaten 
in various ways by the same individual according to its water content, 
ripeness, etc. 
Small individual differences were indicated, however. The 
method of eating G. huillensis fruit was described above: on three 
occaSlons when Satan and Evered were feeding in the same tree, Satan 
held the skin as often in his lips as In his hand, whereas Evered held 
it In his hand only once for every seven times he held it in his lips. 
No individual differences were detected in major techniques. 
There was clear evidence that some individuals had hand 
preferences for tasks involving strength plus precision. In opening the 
hard-shelled fruits of Strychnos spp., T .holstii and D. condylocarpon, 
only once was an individual seen to use different hands for the same 
task: in this occasion Godi banged a Strychnos fruit against a rock 
with his left hand, immediately shifted position and then used his 
right hand for 20 more blows. D. condylocarpon pods were placed, 
upside down, between the incisors after being cracked with the molars: 
most individuals used only one, always the same, hand to pull the pod 
open, but Humphrey invariably used both hands. The data on hand 
preferences is given in Table 3.12. 
One individual (Figan) out of 11 had a variable preference. In 
simpler tasks (e.g. manipulating termite tools) individuals varied much 
3.45 
more, but ag~in some consistency was noted: Hugo once used his right 
hand for 150 consecutive insertions at a termite mound. There is 
evidence of consistent left- or right-handedness in cats and rhesus 
monkeys (Warren et al • . 1967). However, these differences do not appear 
to be associated with asymmetrical brain function or cognitive abilities, 
unlike man (Warren et al. 1969). 
Table 3.12 
Hand preference ~n open~ng hard-shelled fruit 
Individual 
Evered 
Hugo 
Satan 
Godi 
Jomeo 
Sherry 
Figan 
Char lie 
Goblin 
Hugh 
Ml~ 
Left hand 
No. of food 
s6urces 
0 
0 
0 
1 
0 
0 
1 
8 
10 
2 
2 
No. of times 
hand used 
0 
0 
0 
1 
0 
0 
8 
17 
112 
27 
24 
~~~. Rate of eating 
Right hand 
No. of food No. of times 
sources hand used 
17 35 
12 156 
4 4 
3 48 
2 10 
2 2 
2 6 
0 0 
0 0 
0 0 
0 0 
Items which were eaten one at a time could be counted when 
visibility was good. Rates of eating appeared to be fairly standard for 
foods for which there was a small range of tolerance for quality. For 
instance, between picking new G. huillensis fruit the mean number of 
seconds was 18.7, S.D. 7.4 (N = 167 fruit, 5 males, 23 bouts); for 
D. condylocarpon pods, mean time was 22.6 seconds, S.D. 9.4 (N = 61 pods, 
3 males, 11 bouts); for B. bussei pods, mean time was 54.7 seconds, 
S.D. = 3.9 (N = 71 pods, 3 males, 7 bouts). The variance within foods 
could be amply explained by variations in item quality: not all fruit or 
pods were identical in size, number of seeds, etc. There was no 
indication of individual differences in feeding rate when two males were 
4 
3.46 
eating the same food at an abundant source. 
The quality of the food item clearly affected feeding rate. It 
was stated earlier that bouts of eating L. lucida sometimes ended with an 
individual selecting u~ripe fruit. These were eaten much more slowly 
than ripe fruit, because of the greater difficulty in removing the seeds. 
For example, Goliath ate 24 ripe fruit in 3 minutes, then one unripe 
fruit in 3~ minutes. 
Some food items were collected and then eaten together, e.g. 
P . curatellifolia fruit. Overall feeding rates were difficult to 
estimate for these. In general, however, it appeared that feeding rate 
did not change significantly during feeding bouts: once feeding rate 
began to fall, the individual stopped altogether. The consistency of 
feeding rates at particular sources indicates that chimpanzees fed at a 
maximal rate given the food available. Feeding time has therefore been 
used as a measure of feeding efficiency when individuals fed together 
(be low) • 
VI. Feeding in Parties 
1. Introduction 
The suggested relationship (e.g. Azuma & Toyoshima 1962, Nishida 
1974) between food abundance and party size of chimpanzees implies that 
when food is scarce individuals suffer a disadvantage by associating 
with others. This point has not previously been examined. The 
relationship between party size and feeding 'efficiency is therefore 
described here: feeding time is used as a measure of efficiency, as 
exp lained above. 
3.47 
2. Feeding sites of different individuals 
i. Spacing 
Individuals feeding together at a single source spaced themselves 
evenly, generally maintaining a distance of more than one arm's reach 
away from each other. However, the structure of the food source affected 
spacing: only when eating E. guineensis fruit from the tree, for 
instance, did individuals commonly sit within an arm's reach of each 
other. The maintenance of personal space sometimes appeared responsible 
for individuals finishing their available food at different times. For 
instance, Goliath and Sniff were eating S. quinqueloba seeds in a tree 
with only two small branches. Goliath took the branch with more pods, 
and was still eating when Sniff had finished all those on his branch. 
Sniff climbed towards Goliath and sat close to him in a position from 
which he might have been able to reach for some pods. Goliath responded 
by taking hold of Sniff's foot for 10 seconds: Goliath's hair was 
partially erec t, and Sniff squeaked with a low closed grin. Moving 30 cm. 
towards Sniff, Goliath was able to cut off Sniff's access to almost all 
the remaining pods: he fed for a further 78 minutes, while Sniff fed 
for only 4 more minutes, and then waited for Goliath. 
At all food sources one individual frequently waited while 
another finished eating the only remaining food items. Different food 
sites evidently differed in important characteristics, such as the 
number of access routes and the amount of food which could be reached 
without moving. This probably explains why individuals travelling 
together tended to rush to a food source as they approached it: they 
may have been racing for the best food sites, which, once attained, were 
theirs by the right of possession. Displacements from feeding sites were 
infrequent (below). 
In addition to the structure of the food source, another factor 
3.48 
which appeared to affect individuals' tolerance for proximity was the 
quality of their relationship. For instance, when Goblin, eating pith 
of E. guineensis, was approached by Figan, he left and watched Figan eat 
for 8 minutes . After Figan had left, Goblin returned with his mother, 
Melissa, and they ate pith together side by side . Similarly, after 48 
hours of travelling together, Rugo and Gi1ka arrived at a S. qtiiIiqtie10ba 
tree. Rugo climbed in and fed, and Gi1ka approached slowly, giving soft 
grunts. Rugo reached down to touch his companion (Plate 3.3), and she, 
apparently reassured, climbed in and fed close to him. On two other 
occasions Rugo appeared to prevent females from feeding close to him ~n 
this species: his strong association with Gi1ka may have been 
responsible for his tolerance on this occas~on. Tolerance of proximity 
while feeding may be an important component of affi1iative r elationships, 
with varying functional significance according to the type of relationship 
(mother- offspring, male-female etc.): but this problem has received 
little attention to date. 
~~. Intra-specific competition for feeding sites 
Em1en (1973) defined interspecific competition as occurr~ng 
when two or more species experience "depressed fitness attributable to 
their mutual presence in an area. Intraspecific competition may be 
said to occur when two or more individuals experience reduced feeding 
efficiency attributable to their mutual presence in an area. In the 
example above, competition between Go1iath and Sniff can therefore be 
inferred. 
The criteria for inferring competition from direct observation 
fell into four categories. In the following characterisations the 
presence of chimpanzee A appears to have lowered the feeding efficiency 
of chimpanzee B (through loss of feeding time or reduced access to 
preferred food) . Competition over animal foods is treated separately 
(Chapter 4). 
I I 
Plate 3.3 
Reassurance as a result of competition for feeding sites 
Rugo sits in the crutch of a Sterculia quinqueloba at the only 
feeding site. Gilka squats beneath him grunting softly: Rugo 
reaches d0W11 to touch her extended hand. (Gilka then climbed 
past him to share his feeding site.) Sharing of feeding sites 
(sitting within arm's reach) was rare at most foods. Rugo and 
Gilka had heen travelling together for two days, and Rugo may 
have become more tolerant than usual as a result. 
(a) Supplanting. 
feeding site. 
B eats, A arrives and takes over B's 
Example 1: Goblin eats E. guineensis fruit in palm. Mike arrives, 
3.49 
pushes Goblin away from his feeding site, and eats. Goblin climbs round 
the tree and eats in a new site. 
Example 2: Jomeo carries a Ficus sp. branch and eats leaves from it. 
Humphrey approaches Jomeo, who pant-grunts and drops his branch. Humphrey 
eats the Ficus sp. leaves for one minute, then joins another chimpanzee 
calling in a tree. Jomeo returns to his food- branch and finishes the 
leaves after 4 minutes. 
Cb) Following into site. A eats with B nearby. When A leaves, 
B takes over A's feeding site. 
Example 1: Hugo eats E; guineensis fruit ln palm. Figan arrives and 
approaches Hugo, who does not move. Both eat. After 16 minutes Hugo 
climbs down, Figan climbs 60 cm. to Hugo's site, eats for 3 minutes, then 
climbs down fast and follows Hugo. 
Example 2: Leading Godi in grassland, Char1ie comes to a sapling of 
P; curatellifolia, climbs up and eats leaves. Godi sits on the ground 
under Char1ie. When Char1ie leaves, Godi climbs to the vacant feeding 
site and eats. 
Nishida (1973) observed similar behaviour by chimpanzees eating ants 
(Camponotus sp.) when few feeding sites were available. 
(c) Protecting food site. B approaches a new food site, 
A prevents B from reaching it. 
Example 1: Go1iath-Sniff interaction (above). 
Example 2: Evered reaches for a branch of Harungana madagascariensis 
fruit, Faben sees and pulls the branch towards himself before Evered 
touches it. Evered watches Faben eat, then climbs down. 
(d) Occupying preferred site. A eats in good and B ln poor site. 
In these cases there was independent evidence that chimpanzees preferred a 
\ r" 
11 
3.50 
particular site. For example, lone individuals eating Canthium crassum 
collected fruit from the ground, not the tree, and the ground feeding 
site was seen to be protected when a party fed. However, it is 
conceivable that the "poor" site was actually preferred, if there were 
differences in food preference between individuals. In general this 
seems unlikely: females .were less. selective than males when eating 
P. curatellifolia fruit (sometimes eating fruit which had been bitten 
and rejected by males), but appe.ar more selective in the example below. 
Example 1: In a Ficusingens, two males sit eating in single sites while 
two females climb around selecting fruit from all over the tree. 
Example 2: Hugo eats C. crassum fruit from the ground, Figan eats from 
a tree. 
The frequency of observation of these competitive events is 
shown in Table 3.13. 
Table 3.13 
Instances of competition for feeding sites 
Chimp A won or kept the feeding site from chimp B. 
Form of competition Events A dominant A older 
Supp lanting 17 17 16 
Following into site 7 5 6 
Protecting site 14 9 14 
Occupying preferred site 11 10 10 
Total 49 41 49 
% 100.0 83.4 93.9 
Forty-nine clear instances of competition for feeding sites were 
seen (in 1002 hours, i.e. about one per 20 hours). This is probably a 
considerable underestimate of how often a chimpanzee actually experiences 
such competition, for three reasons. First, most competitive situations 
were probably resolved without any overt interactions: second, only ~n 
food sources where feeding sites could be clearly differentiated was it 
possible to interpret competition: and third, this behaviour was not 
3.51 
consistently looked for, but only recorded when obvious. 
Table 3.13 shows that chimp A (the "winner") was the dominant 
individual in 41 out of 49 cases. In cases where A was not dominant he 
was older than B, and the data indicate that relative age predicts the 
outcome of a competitive event slightly better than relative dominance 
status. When both chimp If.. and B were adult males, age was more clearly 
important than dominance status: A was more dominant in 69%, but older 
in 90% of the events (~ = 19). Older chimps never displaced those 
dominant to them, but they appear to have had an advantage when already 
at a site. Females or young chimpanzees were never seen to win or keep 
a food site at the expense of a male. 
The importance of competition for feeding sites is difficult to 
investigate by observation, nnce the number of available feeding sites 
~s difficult to define. An attempt can be made by comparing the number 
of individuals feeding, when in similarly large parties, on different 
foods. During the wet season there are two species, Dictyophleba lucida 
and Gatcinia huillensis, which tend to fruit abundantly if they fruit at 
all, with numerous feeding sites. Parties with more than six individuals 
were compared when the target was feeding ort G. huillensis fruit with 
when he was eating any other food except D. lucida. Only one score was 
taken from each feeding bout. The percentage of party members feeding 
when the target ate G. huillensis (median = 74.7%, N la) was 
significant ly higher than when the target was eating other foods 
(median = 33.3%, N = 13) (Median test, X2 = 4.43, d.L = 1, P < 0.05). 
In this small sample twice as many individuals fed when G. huillensis 
fruit as opposed to other foods were available. The simplest explanation 
for this ~s that more feeding sites were available at G. huillensis. 
Individuals therefore appear to have experienced less competition when 
eating this species than when eating other foods. 
Parties were seen to split up at food sources with limited 
feeding sites. For example, Evered was leading Faben and Figan in 
3.52 
forest when he saw a bunch of Uvaria' angolensis fruit In the canopy. Re 
climbed and began to ea't the fruit, while his companions searched for 
more. Finding none, they walked on, leaving Evered alone. Finishing the 
fruit in 42 minutes, Evered followed the path taken by Faben and Figan, 
but did not meet them in the subsequent two hours. 
These data suggest that when chimpanzees were feeding together 
there was some competition for use of the best feeding sites. Such a 
conclusion is not immediately obvious to an observer confronted with a 
few animals feeding in a tree full of fruit. It suggests that variation 
in food quality and quantity at different sites of a single food source 
may be an important determinant of feeding success. Since relative 
dominance status, and age within males, was important in determining 
access to feedi'ng sites, females and young subordinate males should be 
reluctant to' feed with older, dominant males if feeding efficiency is 
important. 
Competition was twice seen for sites which had no food. 
Carrying palm nuts, Rugo displaced Mustard from a dry patch under a palm 
tree during heavy rain; simi larly, Faben disp laced Gob lin from the only 
available shady spot on a hot afternoon, and ate D.condylocarpon seeds 
he had carried with him. 
Interruption of feeding by social interactions was a further 
important factor which is discussed below. 
3. Social behaviour affecting time spent feeding 
l. Greeting interactions 
When two parties coalesced interactions frequently occurred 
between the meeting individuals (Bygott 1974 , Bauer in prep.). Females 
--------
3.53 
might leave their feeding sites to approach and pant-grunt to males 
(Plate 3.4), or were sometimes attacked before they had the opportunity 
to do so: the same was sometimes true of males, especially if their 
relationship was uncertain. Evered's relationship with Figan was often 
tense, and twice during this study Evered stopped eating when Figan 
arrived, sat for over an hour watching him carefully, and only started 
eating after being attacked or when Figan had left. Attacks and fights 
between two individuals disturbed others in the party as well. Bauer 
(in prep.) has shown that feeding time was less in the 15 minutes after 
individuals first met than in the preceding period. 
ii. Sex 
Parties tended to be larger when oestrous females were present 
(Tutin in prep.). Males lost feeding time both when trying to copulate 
and when overt competition occurred over mating rights: the possess~ve 
male often left his feeding site to attack an individual who approached 
--.. 
"his" female. The females may also have lost feeding time (II/2/v). 
Jarman & Jarman (1973) found that a ter.ritodal male impala (Aepyceros 
melampus) fed for 55% of the day when alone, but only 33% of the day when 
with females. 
iii. Territorial behaviour 
Bygott (1974) has shown that territorial behaviour is shown more 
by large than by small parties, and during this study it was clear that 
if a large party developed there was an increased probability of the 
individuals travelling to the edge of the community range. If they did 
so they were more likely to interact with strangers, during which 
feeding would of course be interrupted. When such a party returned 
towards the centre of the community range it tended to fragment: this 
may have been because individuals had not been able to eat properly, and 
would have a greater chance of finding sufficient food and being able to 
Plate 3.4 
Interruption of feeding by greeting interaction 
(a) Jomeo arrives and sits under a palm tree E. guineensis 
where Gilka has been feeding for ten minutes. Gilka 
pant-grunts and climbs down: she stops above Jomeo, who 
reaches out to touch her. 
(b) Gilka then climbs to the ground and approaches Jomeo, who 
sits with shoulders hunched, hair slightly erect and chin 
extended. (After presenting to Jomeo, Gilka followed him 
to a new food source.) 
3.54 
eat undisturbed in a small party. 
~v. Affiliative behaviour 
Individuals might not always want to groom or play at the same 
time, and although most individuals appeared to do so only when they 
had voluntarily left their food, this was not always so. For instance, 
Charlie and Hugh finished eating B. bussei seeds, from branches which 
they had carried, at different times. While Charlie was still eating, 
Hugh approached him, lifted his food- branch away, and started to groom 
him. Charlie returned the grooming for 12 minutes, and meanwhile 
edged around Hugh, until he had moved far enough to reach his stolen 
food-br~nch. He slowly stretched for it, and for the next 8 minutes ate 
seeds while High continued to groom him. An observer rarely has good 
evidence that one individual sacrifices the performance of a preferred 
activity to "please" another individual. In this case Charlie tolerated 
interruption of feeding by an individual who was not dominant to him: 
that he wanted to feed is indicated by the fact that he subsequently 
continued eating B. bussei seeds for 239 minutes. 
Similarly clear instances of individuals allowing their behaviour 
to be moulded by others who were lower or equal in dominance status were 
apparent in t r avelling. For instance, after arriving at a palm tree 
with Mike, Figan ate palm nuts for 30 minutes, and then sat under the 
tree while Mike ate for a further 37 minutes. As soon as Mike started 
to climb down from the palm, Figan walked and Mike followed. Similar 
instances were frequent: occasionally (seen five times) a male shook a 
branch at another when he wanted him to follow. 
Individuals must inevitably want to behave in mutually 
incompatible ways at times. There are two possible resolutions of such 
conflicts of interest: the first is that they split up, and the second 
that one or both accept inhibition of his activity. In chimpanzees the 
3.55 
pressures to be associative or dissociative are clearly finely balanced, 
and different factors can swing the overall tendency in different 
directions. Party size may vary according to individual benefits ~n 
feeding, sex, community ,defence etc., but it ~s important to note that 
associations and interactions over the short term may take place without 
any obvious a~m except the maintenance of affiliation. If the 
maintenance of affiliation is an end in itself, we can expect a less 
clear relationship between party s~ze and independent variables than if 
individuals associate. merely according to immediately available benefits. 
The important point here is that individuals are frequently 
likely to be disturbed from their preferred activity when ~n a party. 
This is a form of intraspecific interference competition: the extent to 
which individuals tolerate it may depend on a balance between the 
relative importance of disturbances to affiliative and self-maintenance 
behaviour. Evidence of seasonal change in the importance of avoiding 
constraints in maintenance activities will be discussed later. 
4. Party size and food availability 
~. Party size in relation to banana feeding 
From August 1967 to June 1968 bananas were g~ven to all 
chimpanzees in the artificial feeding area on some days (P days) and 
to none on others (No P days) (Wrangham 1974). The difference in party 
(aggregation) size between P days and No P days is shown in Fig. 3.4. 
On both P days and No P days the median aggregation size was highest in 
the early morning: earlier arrivals were more likely to get bananas on 
feeding days. For the rest of the day there were few visitors on No P 
days. The gradual decline in aggregation size on P days was very 
similar to the probability of bananas still being available (Fig. 3.4, 
top and bottom): this emphasises that aggregation sizes corresponded 
• 
Fig. 3.4 
Diurnal variation in aggregation Slze In relation to 
availability of bananas 
Aggregation size was defined as the number of individuals present 
at the feeding area for at least part of the ~ -hr period. All 
independent individuals were scored . Top: median aggregation size 
for each ~ -hr period on all P days (li! = 143 days). Middle: median 
aggregation sinze for each ~ -hr period on all No P days (N = 152 days) . 
Bottom : per cent of P days on which bananas were still available at 
the beginning on each half-hour (N = 91 days). 
The data were taken from records of attendance at the artificial 
f eeding area between August 1967 and June 1968 . P days: bananas given 
to all chimpanzees visiting the feeding area; No P days: no bananas 
given . 
14-
Median 12-
aggregn. 
size 10-
on P days 
8-
6-
4-
2 
0 
7 
6-
Median 
aggregn. 4-
size on 
No P days 2-
0 
7 
100 
Per cent 
days with 80 
bananas 
available 60 
at time 
shown 40 
20 
0 
8 9 10 11 
8 9 10 11 
8 
FEEDING AREA: Aug '67 - Jun '68 
N = No. of days 
N=143 
12 13 14 15 16 
Time of day 
N= 152 
12 13 14 15 16 
Time of day 
14 15 16 
Time of day 
J 
3.56 
closely to the availability of bananas. 
To find out if individuals tended to travel together after 
feeding together, records of arrival and departure at the feeding area 
were analysed: if two individuals left within the same IS - minute period 
by the same path, they were scored as leaving together. Individuals who 
regularly visited camp were listed in a random order (names pulled out 
of a hat). Attendance records were examined by date order until the 
individual highest on the list was found to be in camp. If during his 
time in camp another individual, who had arrived separately, was present, 
a score was entered for one of the four categories shown in Table 3.14. 
If he met more than one individual in camp, the companion to be scored 
was chosen at random. 
The results are shown in Table 3.14. When individuals arriving 
separately were fed together they were significantly more likely to 
leave 
of 54 
together (13 out of 15 times) than when they were not fed (29 out 
- 2 times) (X = 5.4, d.f. = 1, P < 0.05). 
Table 3.14 
Effect of feeding together on tendency to travel together 
Only individuals arriving separately are scored: individuals are sampled 
at random (see text). Data from camp attendance records, January to 
September 1971. The cells show the number of times each combination 
occurred in the sample. 
Individuals leave 
together 
Individuals leave 
separately 
Both individuals 
given bananas 
13 
2 
Neither individual 
g~ven bananas 
29 
25 
The evidence from the artificial feeding area therefore shows 
that parties were larger when more bananas were available, and that 
3 . 57 
individuals who fed together tended to travel together, for a short 
time at least. 
~~. Variation of party s~ze at different foods 
Party sizes at different foods must clearly be compar ed at a 
single time of year. However, the large variety of foods eaten means 
that satisfactory sample sizes are difficult to obtain: only in extreme 
cases will the effect of the different kinds of food sources outweigh 
the variance in party size from other factors. 
Observations during this study suggested that during July 1972 
party sizes were higher when the target was eating Harungana 
rnadagascariensis fruit than when eating E ~ guirteensis fruit. It would 
have been illegitimate to test the difference from my data, s~nce they 
were responsible for indicating that these were suitable foods to compare. 
Group- Travel charts of other observers were therefore analysed. The 
results are shown in Table 3.15. 
Table 3.15 
Variation of party size with different foods of target 
Data from Group-Travel charts (other than those completed for this study) 
in July 1972. Party size (independent individuals only) was recorded at 
the mid- point of all feeding bouts. Cells show number of bouts. 
Party s~ze One or two 
More than two 
Food of target 
H~ madagascariensis fruit E. guineensis fruit 
3 
6 
15 
3 
Party s~zes were significantly higher when the target was eating 
H. madagascariensis fruit (X2 = 6.75, d.f. = 1, P < 0.01). The result 
should be interpreted with caution: target individuals were not matched, 
and the foods are different both in structure and in distribution, 
H. madagascariensis tending to occur further from the lake than 
E. guineensis. Nevertheless, the data support the thesis that party 
size varies with different foods, and in the predicted direction: 
3.58 
E. guineensis trees have few feeding sites compared to H. madagascariensis. 
~~~. Seasonal variation ~n party size 
A number of studies (Azuma & Toyoshima 1962, Reynolds & Reynolds 
1965, Sugiyama 1973, Nishida 1974) have suggested that party size of 
chimpanzees changes according to seasonal changes in food availability. 
There are three main difficulties with this conclusion. The first is 
that there has not been proper sampling of party sizes. Nishida (1974) 
presents the best evidence for changes in party sizes, with mean party 
sizes varying between 4.5 and 13 in different months. However, unless 
different individuals are sampled equally there is the possibility of 
confounding individual differences in aggregation tendency with seasonal 
changes. Care is therefore taken here to sample different individuals 
equally. 
The second difficulty is that these studies have g~ven no evidence 
of changes in food availability. Such estimates are extremely difficult 
to obtain in the large and complex habitats of chimpanzees, especially 
since the diet is so broad. The argument that seasonal changes in party 
s~ze were related to changes in food availability will be developed in 
Section VII . 
The third difficulty is that previous studies have confounded 
possible changes in food availability with changes ~n other environmental 
variables, such as climate and habitat structure. There is evidence that 
vegetation density affects observed party size: in this study party 
s~zes were larger in grassland than in semi-evergreen forest in wet and 
dry seasons (Table 3.16). Though the data in Table 3.16 are not sampled 
3.59 
evenly across individuals, and though they are not controlled for 
different food availability between habitats, they suggest that care 
must be taken before ascribing seasonal changes in party size to changes 
in food availability. The, increase in observed party size with more 
open habitats may be due to observer error or to real differences: 
although we cannot tell which explanation is more important, they could 
both operate between different seasons. 
Table 3.16 
Variation ln party size with habitat type 
Data from 45 all-day observations, observations taken on 3D- min. points. 
Only 'independent individuals are scored. Cells show median number of 
individuals within 200 m. of the target individual. N = number of 
observations. 
Habitat type 
Semi-evergreen Deciduous Grassland with or N forest forest without trees 
Jul to Sep 1972 0.0 0.1 0.6 387 
Nov 1972 to Jan 1973 1.3 1.5 3.9 251 
Jul to Sep 1972 1.2 1.6 2.3 300 
N 240 476 222 938 
The data collected in this study alone were not sufficient for an 
analysis of seasonal ch'anges in party size. Group-Travel charts provided 
enough data on different individuals to obtain matched pair samples of 
males and females between three seasons. 
As a precaution before using the Group -Travel charts, the 
influence of camp on party size outside it was examined (Fig. 3.5). 
When either a male or female was target there was an immediate rise ln 
party Slze on entering camp, and an immediate drop on leaving. The 
party sizes of male targets were indistinguishable before and after 
entering camp: those of female targets were slightly higher after than 
before entering camp. However for both sexes, party sizes were 
3.60 
significantly different before from after entering camp, and before from 
after leaving camp: but were not significantly different before 
entering from after leaving. Significance levels are shown in Table 3.17. 
Tab le 3.17 
Significance of differences ~n party s~zes before entering, while in 
and after leaving camp 
The data are from Fig. 3.5. Sign tests used throughout. N 
targets whose party sizes had changed. 
number of 
Times compared Sex of N x target z P 
One hour before entering and ) 
two hours after leaving ) 
One hour before and ) 
immediately after entering ) 
Immediately before and two ) 
hours after leaving ) 
Male 17 
Female 16 
Male 30 
Female 24 
Male 29 
Female 26 
7 
6 
9 
3 
5 
7 
2.01 
3 . 33 
2.16 
n.s. 
n.s. 
0.05 
0.001 
0 .001 
0.05 
The conclusion from these data is that a visit to camp did not 
have a signific ant short-term effect on party size outside it. 
Observations during which the target visited camp were therefore used ~n 
the analysis of seasonal variation of party size, but as a precaution 
observations within two hours of leaving camp were excluded. 
Seasonal variation ~n party s~ze was tested by controlling for 
target individual and time of day. Party sizes were scored at 1230 hours, 
and listed by target individual. After listing, scores were matched by 
order. The rigour of the procedure reduces the amount of available data 
but imposes precision on the results . Statistical significances are 
shown in Table 3.18. 
The results clearly show that party sizes were significantly 
larger in the wet and the 1973 dry than in the 1972 dry season. The 
most interesting result is the comparison of the two dry seasons: 
climate and habitate structure were similar between these times, and 
1 
Fig. 3.5 
Effect of camp on party slze 
Data from July to December 1972. All records were used in which the 
target individual visited camp during observation, except (a) those 
when he arrived before 0800 or left after 1700 hours, and (b) those 
which overlapped in time with a record previously included. All 
individuals were scored except those permanently travelling with their 
mothers: party size was scored as the half- hour began. For males 
N = 39 records, females N = 38. 
4 
Mean 
party 
3 
size 
2 
before after 
entering leaving 
- 2 -1 0 = 0 0 = 0 +1 +2 +3 +4 
-
No. of half hours 
In Camp 
Male target 
Female " 
4 
3 
2 
3.61 
cannot therefore have been responsible for the change in grouping 
tendency. This result will be discussed below. An indication of 
actual numbers in parties was given in Table 3.16. 
Table 3.18 
Significance of seasonal variations Ln party size 
Data are from Group-Travel charts, party sizes scored at 1230 from all 
records except where target had left camp less than two hours 
previously. Party sizes are matched by target individual. All 
differences tested by Wilcoxon. Seasons: Dry July to September, 
Wet = November 1972 to January 1973. Numbers of individuals as targets 
are shown in brackets. 
Seasons compared Target individuals N T P Season with 
sampled z larger parties 
-------
Dry '72 & Dry '73 All availab le (17) 38 2.85 <.01 Dry '73 
Northern males (8) 17 27 <.05 Dry '73 
Northern females (6) 17 33.5 <.05 Dry '73 
Wet & Dry '73 All avai lab le (19) 51 0.65 n.s. 
Northern males (8) 26 0.51 n.8. 
Northern females (8) 21 64.5 n.s. 
Dry '72 & Wet All available (17) 30 2.89 <.01 Wet 
Northern males (9) 26 2.34 <.05 Wet 
Northern females (5) l3 40.5 n.s. 
5. Party SLze and time spent feeding 
Data have been presented to show that there was competition for 
feeding sites: that when more food was available at a food source 
parties were larger: and that more individuals within the parties fed 
at sources with more feeding sites. The results suggest that one of the 
factors regulating party size was the extent of competition. Since it 
has also been shown that there were significant seasonal differences in 
party size, there may have been corresponding seasonal differences Ln 
the extent and importance of competition. This possibility is now 
examined. 
3.62 
From all - day observations of Northern males in the three seasons 
sampled, the number of individuals present, and the number feeding, 
counted on each 30-min. point. The overall per cent of individuals 
feeding in parties of each' size was calculated: the data are shown 
Tab le 3.19, and graphed in Fig. 3.6 (top). 
Table 3.19 
Seasonal variation in per cent of individuals feeding ~n 
different party sizes 
were 
~n 
Data from 28 all - day observations of Northern males. All individuals 
other than permanent dependents are scored. N = number of parties 
observed, P = per cent of individuals observed feeding. Party size and 
number of individuals feeding were scored on success~ve half-hour points. 
Party 
s~ze 
1 
2 
3 
4 
5 
6 
7 
8 
9 
10 
11 
12 
13 
14 
15 
16 
17 
Jul - Sep '72 
N P 
118 70.1 
54 53.7 
49 55.1 
18 41. 7 
22 48.2 
10 31. 7 
4 3.6 
2 75.0 
5 62.2 
o 0 
o 0 
o 0 
o 0 
o 0 
o 0 
o 0 
o 0 
Nov '72 - Jan '73 
N P 
80 55.0 
36 50.0 
27 23.5 
9 61.1 
10 60.0 
7 40.5 
14 44.9 
11 46.6 
8 56.9 
2 60.0 
6 25.8 
2 70.8 
1 0 
1 0 
5 13.3 
1 43.8 
1 52.9 
Jul - Sep '73 
N P 
25 72.0 
103 67.0 
57 69.6 
28 38.4 
28 62.1 
23 58.7 
14 56.1 
7 25.0 
8 56.9 
4 60.0 
4 47.7 
2 54.2 
o 0 
o 0 
o 0 
o 0 
o 0 
Fig. 3.6 (top) shows considerable scattering of points. The 
relationship is clarified by plotting the cumulative number of 
individuals feeding as a percentage of all individuals seen (Fig. 3.6, 
bottom). In all seasons there was a decrease in per cent individuals 
feeding as increasingly large parties were included. The actual numbers 
feeding are not important, representing only the difference in feeding 
time across seasons: the important seasonal difference is that in the 
t 
Fig. 3.6 
Feeding probability and party size 
Data from Table 3.18. Top: per cent of individuals feeding in each 
party size. Bottom: per cent of individuals feeding in relation to 
cumulative party size. The graph shows that in the period Jul- Sep 
1972 the probability of an individual feeding fell faster with 
increasing party size than at other times. 
80 
Per cent ~. ' 'i>- - -4 ~" 
60 
individuals 
40 
feeding .~ 
20 
' .. 
0 ii 
3 5 7 9 11 13 15 17 
PARTY SIZE 
[J- .... ·IJ Jul - Sep 1972 
----
Nov 1972 - Jan 1973 
~-..:. Jul - Sep 1973 
80 
Per cent t.~ '6- -..b.. 
60 ~ .... ~ .:'6--~-~-6_._._._._..:. 
' {J. .. " 
irdividJals 
40 
feeding 
20 
0 
3 5 7 9 11 13 15 17 
ALL PARTY SIZES up to and including 
3.63 
1972 dry season the proportion of individuals feeding fell faster with 
increasing party size than in either of the other seasons. 
Within each season, the relationship between party Slze and 
proportion of individuals' feeding was tested by comparison of all parties 
up to with all greater than each party size. The results are shown in 
Table 3.20. 
Table 3.20 
Significance of the null hypothesis that per cent of individuals 
feeding did not vary with party size 
The data are taken from Table 3.19, and tested (2 x 2 X2: d.f. = 1) by 
comparing numbers feeding in all parties up to each party size with 
those in larger parties. Dependent individuals were not scored. 
Probabilities are two-tailed, and inspection shows that the probability 
of feeding was greater in smaller parties ln all cases. 
Party Slzes compared 
1 
1-2 
1-3 
1-4 
1- 5 
1- 6 
1-7 
1- 8 
1- 9 
1- 10 
l-ll 
1- 12 
1- 13 
1-14 
1- 15 
1- 16 
& ' >2 
& >3 
& >4 
& >5 
& >6 
& >7 
& >8 
& >9 
& >10 
& >11 
& >12 
& >13 
& >14 
& > 15 
& >16 
& >17 
Jul-Sep 1972 
X2 P 
9.01 
6.97 
8.89 
5.21 
4.81 
0.65 
3.13 
1.01 
0.01 
0.01 
0.01 
0.05 
0.05 
n.s. 
n.s. 
n.s. 
Nov '72 - Jan '73 
i P 
5.82 
7.30 
0.48 
2.38 
6.74 
6.51 
9.31 
5.53 
13.60 
17.72 
10.83 
20.17 
14.88 
9.70 
0.30 
0.45 
0.05 
0.01 
n.s. 
n.s. 
0.01 
0.05 
0.01 
0.05 
0.001 
0.001 
0.001 
0.001 
0.001 
0.01 
n.s. 
n.s. 
Jul-Sep 1973 
X2 P 
0.86 
4.47 
1l.45 
2.34 
3.96 
4.96 
5.31 
0.27 
0.39 
0.81 
0.07 
n.s. 
0.05 
0.001 
n.s. 
0.05 
0.05 
0.05 
n. s. 
n. s. 
n.s. 
n. s • 
In 22 out of the 35 comparisons in Table 3.20, the percentage of 
individuals feeding in smaller parties was significantly higher than ln 
larger parties. In all seasons significance was lost for the final 
samples, where sample size was small. Only ln the wet and 1973 dry 
seasons were there small party sizes (up to 1, 3 and 4) which were not 
3.64 
significantly different from larger parties. These results confirm that 
in all seasons sampled individuals were less likely to feed as party 
S1ze increased. Bygott (1974) found similarly that individuals spent 
more time feeding when alone than when they had companions. 
The hypothesis that individual feeding time fell faster with 
increasing party size in the 1972 dry than 1n other seasons could not be 
adequately tested, oW1ng to differences in overall time spent feeding 
between seasons. 
The significance of these results 1S discussed below. 
VII. Seasonal Changes 
1. Introduction 
In the preceding sections seasonal changes in various aspects 
of feeding behaviour have been described. These changes are now 
summarised, and seasonal differences in related topics are described. 
For simplicity the three seasons sampled during this study are referred 
to as: Dl : July to September 1972 (dry); W: November 1972 to January 
1973 (wet); D2 : July to September 1973 (dry). 
2. Evidence of changes in overall food availability 
1. Changes in the diet 
Seasonal differences 1n diet generally reflected changes in the 
abundance of particular food items. The significance of these changes 
cannot be interpreted with certainty without nutritional analysis. It 
was noted that when Parinari curatellifolia fruit was available it was 
the most eaten food (II/3/i): its absence in Dl suggested that less 
food was available overall than in D2 , when other foods appeared 
approximately equal in abundance. Further evidence on changes in food 
availability comes from a comparison of foods which were eaten in all 
seasons. Two foods were analysed: palm nuts (E. guineensis fruit) 
and salt. 
3.65 
Individual oil palm trees flower and fruit in regular cycles, but 
each individual has its own cycle. In a group of trees there are always 
likely to be some in fruit. The number and weight of fruit is affected 
by a complex interplay of factors including dry-season rainfall two years 
previously, time in relation to the individual's five-year high-yield 
cycle, and physical factors such as light and mineral availability 
(Hartley 1967). Overall seasonal correlations with fruit production are 
therefore absent. 
In three measures feeding behaviour on E~ guirteensis was similar 
between Wand D2 , and differed between these seasons and Dl . 
First, it has already been shown that the concentration of 
feeding bouts in the morning was more marked in Dl than in D2 or W. 
Second, the frequency of competition between baboons and 
chimpanzees feeding at E. guineensis changed. Competition occurred 
when baboons and chimpanzees fed in the same palm: only interactions 
where one or more animals were prevented or inhibited from eating were 
scored. Adult male baboons competed successfully with female or 
juvenile chimpanzees (three times), but all classes of baboon lost to 
adult male chimpanzees (14 times). Interactions included baboons 
waiting for chimpanzees to leave, being . displaced, or fighting: Hugo 
once punched a lone male baboon on the jaw and in the belly before the 
intruder climbed down. 
Table 3.21 shows seasonal changes in the frequency of baboon-
chimpanzee competition. 
Table 3.21 
Frequency of occurrence of overt competition between chimpanzees 
and baboons at E. guineensis fruit 
Data are from all-day observations and only the target individual 
is scored. 
Season 
Dl : Jul-Sep 1972 
W: Nov 1972 - Jan 1973 
D2 : Jul - Sep 1973 
Total mins. 
eating palm nuts 
2235 
940 
592 
Number of 
competitive events 
11 
1 
0 
3.66 
Events 
per hour 
0.30 
0.06 
0.00 
The number of events was too few to test seasonal differences by 
2 
matching all-day observations, and X tests are illegitimate (expected 
values less than 5). However, two conclusions may be tentatively drawn 
from the data. First, the rate of events was more similar between the 
Wand D2 seasons than to Dl : and second, the direction of the difference 
indicates that competition for food was more severe in Dl . 
The third measure taken on feeding behaviour at palms was mean 
bout length. The mean length of all bouts of eating E.guineensis was 
25.4 mins. in Dl (N = 87) compared to 17.4 mins. in Wand D2 (N = 35, 54 
respectively). Again, feeding behaviour in Dl differed from the other 
two seasons. 
Independent evidence on the relative S1zes of the fruit crop in 
Dl and D2 is not available. If more palm nuts were available in Dl 
than D2 , it is reasonable that bout length should have been longer, but 
not that competition should be greater, as indicated by the data on 
interactions with baboons and diurnality of feeding bouts. If less were 
available in Dl than D2 , on the other hand, the greater competition 1S 
understandab le but not the longer bouts. An explanation which fits 
better than a supposed change in the amount of E. guineensis available 
is that the crop size was similar in Dl and D2 but that in Dl it was a 
3.67 
relatively more important crop in relation to all available foods: 
this would explain why individuals were prepared to feed for longer 
and compete more strongly. Longer feeding bouts may thus have been due 
to a tolerance of less good fruit when overall food availability was 
reduced. 
A second food available all year was salt at the artificial 
feeding area. Minerals were not deliberately provided, but since 1969, 
when a cobra skin was prepared with sodium chloride and left out on a 
concrete slab, chimpanzees have intermittently licked the spot. In 1970 
the slab was coated with grease, but chimpanzees were not deterred. 
The number of chimpanzees who licked varied seasonally 
(Table 3.22). It seems unlikely that they got much salt, and many 
individuals licked for only a few seconds: if it was not a permanent 
source of food, the slab may have been a permanent hope of it. 
Table 3.22 
Seasonal variation ~n numbers of individuals seen to lick salt 
at least6nce 
Recording began ~n May 1971. Data are from the artificial feeding area. 
Period Number of individuals Rank licking at least once 
Apr-Jun 1971 29 10 
Jul- Sep 1971 20 7 
Oct-Dec 1971 0 1 
Jan-Mar 1972 18 5 
Apr-Jun 1972 23 8.5 
Jul-Sep 1972 23 8.5 
Oct-Dec 1972 6 2 
Jan-Mar 1973 17 4 
Apr- Jun 1973 19 6 
Jul-Sep 1973 7 3 
When period were ranked by date order, there was little 
correlation with salt- licking tendency (Spearman r .· = -0.43, N = 10, 
s 
n.s . ). Thus salt dilution does not explain the variation in the number 
3.68 
of lickers. Increased sodium requirements are known to be caused by a 
variety of metabolic stresses, including low glucose levels (Forbes 
1962: laboratory rats). Variation in the numbers licking salt suggests 
that the population experienced seasonal variation in the extent of 
metabolic stress. (Salt requirements may be expected to vary with salt 
expenditure, a major cause of which is sweating. However, the climate 
was similar in Dl and D2, and individuals actually travelled further 
in D2 : variations in sweat output do not explain the greater amount of 
salt-licking which occurred in Dl .) 
11. Weight changes 
If an individual was alone at the feeding area he could be 
weighed by luring him up a rope hung on a spring balance: while he ate 
a banana at the top the scale was read with binoculars. Weights were not 
taken regularly: some individuals refused to be tempted up the rope, and 
others climbed away to eat the banana. A reading was only taken when the 
chimpanzee rema ined still for several seconds. 
Data on weights of adult males Slnce July 1970 are shown in 
Table 3.23, separated into three-month periods. Females with small 
infants were inevitably weighed with them, and therefore could not be 
used in this analysis. No Southern male other than Goli~th was weighed 
more than twice. 
The significance of differences between periods was tested by 
calculating a Kendall coefficient of concordance. The available matrix 
with the highest number of complete cells included weights of MK, HM, FG, 
JJ and ST in periods A, D, H, I and J: weight differences between 
periods were significantly similar across individuals (W = 0.56, s 137.6, 
k = 5, N = 5, P < 0.05). Lest it be argued that FG, JJ and ST had 
similar variations because they may still have been growing, differences 
were also tested using only older males: for MK, HM, EV and GO in periods 
3.69 
A, C, D and J, weight differences between periods were aga~n ranked in a 
significantly similar way (W = 0.83 ;' s = 64.5, k = 4, N = 4, P < 0.01). 
Table 3.23 
Mean weights of adult males between August 1970 and September 1973 
Weights in kilograms: ~ sample size. There was no data for Jan-Mar 1972. 
Period 
Ju1-Sep 
1970 (A) 
Oct- Dec 
1970 (B) 
Jan- Mar 
1971 (C) 
Apr - Jun 
1971 (D) 
Ju1- Sep 
1971 (E) 
Oct- Dec 
1971 (F) 
Apr-Jun 
1972 (G) 
Ju1- Sep 
1972 (H) 
Oct-Dec 
1972 (I) 
Jan-Mar 
1973 (J) 
Apr- Jun 
1973 (K) 
Ju1- Aug 
1973 (L) 
MK 
kg N 
HG 
kg N 
HM 
kg N 
EV 
kg N 
FG 
kg N 
JJ 
kg N 
ST 
kg N 
GO 
kg N 
35.4 1 44.5 1 34.1 1 33.6 1 39.5 1 39.5 1 31.8 1 
36.3 1 38.6 1 41.8 1 
37.7 2 39.0 4 45.5 2 39.5 2 37.7 2 
33.6 2 35.9 2 . 41.4 4 36.8 5 36.8 4 45.5 4 42.3 5 35.5 4 
35.9 2 45.0 2 47.3 1 
38.2 2 46.4 1 48.6 2 44.1 2 38.6 1 
38.2 1 39.5 2 45.9 1 38.6 1 47.7 1 45.9 1 
36.4 3 39.1 2 44.1 1 38.2 3 49.1 1 44.5 2 37.7 1 
36.8 3 40.0 3 45.9 3 40.0 2 36.4 4 48.2 3 47.3 4 
37.7 4 44.5 1 39 . 1 6 37.7 1 49.1 5 45.5 4 40.5 1 
37.7 1 39.1 1 48.2 2 
39.1 1 38.6 1 50.0 1 
The scale was not calibrated against a standard measure. It was 
conceivable, therefore, that the observed changes were the result of 
inaccuracies in the spring balance. This is unlikely since wei ghts did 
not become hi gher or lower w'ith time, or vary with wet and dry seasons, 
as might be expected with a mechanical fault: and the scale was 
sufficiently sensitive to show expected weight changes of maturing 
3.70 
juveniles and pregnant females. 
It seems probable, therefore, that the observed weight changes 
were real. In order to compare the sampled periods, the data from 
Table 3.23 were transformed into a "win-lose" matrix. For each period 
the coefficient (W;L) was calculated: W is the number of periods in 
which the majority of chimpanzees were lighter than in that period (wins), 
L is the number of periods when the majority of chimpanzees were heavier 
(losses), and T 1S the total number of periods. The result is shown in 
Table 3.24. 
Table 3.24 
Dyadic comparison of variation 1n weights between seasons 
Data are taken from Table 3.23, and ordered according to the value of 
W-L W-L (-r-). Chimpanzees were heavier during periods with high (-r-) scores. 
Period L FIG J . C H B K E A D (W;L x 100) 
Ju1-Sep 1973 L + + + + + + + + + +82 
Oct-Dec 1971 F + + - + + + + + + + +64 
Oct-Dec 1972 I + + + + + + + + +45 
Apr-Jun 1972 G + + + + = + + + +45 
Jan-Mar 1973 J - + + + + + + +36 
Jan-'Mar 1971 C + + + + + + +27 
Ju1-Sep 1972 H = + + + + -18 
Oct-Dec 1970 B ? + + + -20 
Apr-Jun 1973 K + + + - 30 
Ju1-Sep 1971 E + + -64 
Ju1-Sep 1970 A -91 
Apr- Jun 1971 D -91 
+ Win to row period (majority of individuals heavier in row period 
than column period) 
Win to column period (majority of individuals heavier in column 
period than row period) 
Equal numbers of individuals heavier, in row period than column period, 
as vice-versa (or all individuals weighed the same in both periods) 
? No individual was weighed in both periods 
The order of high -weight to low-weight seasons (L to D, Table 
3.24), was not closely related to time or climate: the latter was 
tested by labelling Apr- Jun and Ju1-Sep "dry" and Oct-Dec and Jan-Mar 
~~--------------------------------______ b 
3.71 
"wet": differences in rank of the two groups weye not significant (Mann-
Whitney U = 9, n l = 5, n2 = 7, n.s.). However, wet season weights were 
generally higher than dry season. If weight changes reflected changes 
in food avai labi li ty, this is to be expected. The wet season is 
generally a productive time, but not necessarily. Internal rhythms of 
plants, pollination success, etc. may cause failure of crops at any 
time: in 1965-66, for instance, D. lucida failed altogether (shown by 
records of faecal analysis), and van Lawick-Goodall (1968) gave other 
examples. 
111. Comparison between changes 1n weight and 
numbers licking salt 
Seasonal changes in weight and the number of individuals licking 
salt in camp have been described. These are two measures which may be 
expected to change in relation to metabolic stress. If both were 
affected by the same factor, their oscillations should be correlated. 
Comparison o~ Tables 3.22 and 3.24 shows that the number of individuals 
licking salt at least once was negatively correlated with high weight 
values (W-L x 100) (Spearman r . = -0.83, N = 9, P < 0.01). This result T s 
argues strongly in favour of the suggestion that changes in weight and 
the frequency of salt-licking were affected by the same independent 
variable: the likely variable was food availability. 
The measures agree in the relative ranking of Dl and D2 : the 
data are consistent with the hypothesis that food was more abundant 
in D2 . The wet season (W) does not correspond exactly to the periods 
sampled for weight and salt-licking, but appears to fall between Dl 
and D2 . 
3.72 
3. Changes in social behaviour 
L. Feeding behaviour in relation to parties 
It was shown in Dl , Wand D2 that the number of individuals 
feeding fell as party size ' increased, but that in Dl the decrease Ln the 
proportion of feeders fell faster with increasing party size than Ln W 
or D2 • This suggests that it was more of a disadvantage to be in a 
party of given SLze in Dl than in W or D2 • Consistent with this 
hypothesis was the finding that significantly fewer food-calls were 
given in Dl than in W or D2 : since it was shown that food-calls tended 
to attract individuals, this suggests that in Dl individuals were 
reluctant to encourage an increase in party SLze. Party size was found 
to be significantly smaller in Dl than in W or D2 , for parties both with 
at least one male and with at least one female 
LL. Individual relationships 
The seasonal changes in party SLze meant that individuals spent 
different amounts of time with each other. "An example of the way Ln 
which association patterns changed between Dl and D2 is shown in Fig.3.7. 
Although changes in association patterns and party SLze were 
correlated with changes in feeding behaviour, it is possible that the 
relationship was not cause and effect: for instance, developmental 
changes in ongoing relationships might have led to tensions among 
individuals, which would yield an overall change in party size. 
Dominance relationships among high-ranking adult males, and the 
availability of oestrous females, are two factors which may affect 
association patterns. 
Among the Northern males, a stable hierarchy suffered a violent 
change in September 1972 when Figan attacked the previous dominant, 
Humphrey (Halperin, in prep.). Subsequently, the hierarchy was unstable, 
Fig. 3.7 
Seasonal change in association patterns 
Data from all-day observations only, five Northern males. Records 
of association (two indiv~duals within 100 m.) were scored on 
successive half- hour points, and each individual contributed 20% of 
total observation time as target (20% x 5 = 100%). 
Top: dyadic association patterns among the five males sampled. Per 
cent time together increased for all pairs between the 1972 and 1973 
dry seasons (EV- FB increased from 10.8% to 15.6%). 
Bottom: time spent by each male with all females. Total half -hours 
with each female present were scored and summed across females. All 
males spent more time with females in 1973 than in 1972 dry season 
(JJ: 15 "female-~-hrs" per day in 1972, 18 per day in 1973). 
Sample size: N = 10 all-day observations in 1972, N = 20 in 1973. 
(Five observations were by field assistants, others by RWW.) 
JUL - SEP 1972 
~ 
~~ ~ 
JUL - SEP 1973 
Omitted 
KEYS 
"10 of samples 
in proximity 
o -9 
10 -19 
20 -29 
30 -39 
40 -49 
N f ,,()1-h" 0.0 +-2 rs. 
per day 
Omitted 0 - 9 
10 -19 
20 - 29 
etc. 
3.73 
and coalitions became important in dominance interactions. It could be 
argued that party size was primarily a function of the stability of the 
hierarchy. Various reasons make this a weak argument. 
First, the fall of the alpha-male occurred nine months before D2 : 
by July 1973 the hierarchy had become fairly stable once more. Second, 
although paucity of data prevented adequate testing, party size among the 
Southern community appeared to change in the same way as among the 
Northern, without any change in the alpha-male. Third, there is no 
obvious reason why changes in party size should cause the observed 
changes in feeding behaviour, whereas the observed changes in party s~ze 
follow logically from the greater competition for food in Dl . Indeed, 
if food availability was similar in Dl and D2 , and individuals aggregated 
more in D2 for social reasons, one would expect more, not less, 
competition in that season. 
Adult males often aggregated around oestrous females. Riss 
(pers. comm.) found that the mean party size with Figan present was 3.5 
on 15 days when Gigi was anoestrous, compared to 5.5 on 13 days when Gigi 
was fully oestrous (data from 1974). Tutin (in prep.) found that party 
size varied between different oestrous females. 
The number of Northern females in oestrous was similar between 
the seasons sampled. Data are shown in Table 3.25. 
Table 3.25 
Seasonal variation ~n number of oestrous females 
Northern females only. 
Females oestrous in Females oestrous in Season all three months one or two months 
Dl : Jul-Sep 1972 AT, GG, GK, SW 
~,~ 
W: Nov 1972 - Jan 1973 ~, GG, SW, ~, NP GK 
D2 : Jul-Sep 1973 PL, GG, GK, SW 
3.74 
The concept of availability is complicated by the fact that 
oestrous females may travel in different community ranges. For instance, 
Gilka left the Northern community temporarily during the wet season, and 
Little Bee, a Southern female not included above, was an adolescent who 
occasionally visited the Northern community. However, it is clear from 
Table 3.25 that the number of oestrous females was similar in the sampled 
seasons: in Dl , when there were smaller party sizes, there were slightly 
more oestrous females available than in other seasons. The data show no 
evidence in favour of the hypothesis that seasonal change in party size 
was caused primarily by change in the availability of oestrous females. 
The main factors affecting female reproductive condition appear 
to be rhythms of growth, birth and lactation in the individual (van 
Lawick-Goodall, in eress). There is little evidence for an overall seasonal 
rhythm in the timing of births (Tutin, in prep.). 
VIII. Discussion 
1. Maxirnisation of feeding efficiency 
It was argued earlier Oil) that the extent to which feeding 
behaviour influences social structure depends on whether feeding 
efficiency is maximised at the expense of reproductive effort. Evidence 
of maxirnisation of feeding efficiency is therefore examined here. 
1. Item selection at a food source 
The data showed that chimpanzees competed for feeding sites (vI/2), 
and that they were able to judge the condition of fruit and accordingly 
select what appeared to be the best (V/2). Furthermore, methods of 
manipulation of food items, and the rate of intake, were consistent 
within food types and across individuals (V/3). These data are compatible 
.with the hypothesis that behaviour was directed to maximising feeding 
11 
efficiency when chimpanzees were feeding. Differences in feeding 
behaviour shown at different sources can largely be ascribed to 
differences in the distribution and structure of the food items. 
3.75 
It was noted that 'chimpanzees fed more selectively on fruit than 
on other items (particularly leaves and seeds) (In/2/iii). This is 
probably because plants have evolved signals indicating the palatability 
of their fruits: a plant benefits if its fruit are eaten, provided that 
the seed is not destroyed (Stebbins 1971, Janzen 1971). In Gombe various 
mechanisms made unripe fruit unpalatable such as bad taste (Monanthotaxis 
poggei), husk hardness (Pycnanthus angolensis) and gritty particles 
(Grewia platyclada). These characteristics disappeared when the seed 
hardened, when it was more likely to endure passage through the gut. 
Colour changes tended to occur simultaneously. These were beneficial for 
the plant because animals could then identify palatab le fruits without 
experimenting on and destroying immature ones. Chimpanzees benefited by 
saving time an~ energy which would be used in climbing to test unr~pe 
fruit. 
Seeds and leaves, on the other hand, are useless to the plant 
once digested. Signals indicating ripeness have therefore not been 
evolved, making it harder for chimpanzees to feed selectively, and 
therefore efficiently. 
~~. Diet selection 
The optimal diet means that which maximises feeding efficiency 
(defined in 3/1). The difficulty in demonstrating optimisation ~s that 
only exhaustive nutritional analysis would show that available foods 
could be selected with greater advantage than actually occurs. Proof 
of optimisation is therefore impracticable. However, it might be 
possible to show that the diet is suboptimal. 
3.76 
Two lines of evidence appear to be relevant. First, it was 
found that foods eaten in one area were not eaten in another (III/2/i). 
Second, the observed patterns of food selection do not conform to the 
predictions of some optimisation models: chimpanzees maintained a 
similarly varied diet between days and between seasons (III/2/ii & iii); 
they spent similar amounts of time feeding in different seasons (II/2/i); 
and they often left food sources, especially of leaves, with apparently 
good food remaining (III/2/iii). Emlen (1966), MacArthur & Pianka (1966), 
Schoener (1971) and Pulliam (1974) have developed models of optimal diet 
based on different assumptions, but all predict that the number of food 
types in the diet should decrease when food becomes abundant. If food 
abundance varied between seasons, as seems probable (the evidence ~s 
reviewed below) the models suggest that chimpanzee food selection was 
nutritionally suboptimal. 
The objection to these models is that their measurement of 
profitability ' of different strategies ~s merely in terms of energy gained 
or time saved (Krebs 1973). Though these may be valuable assumptions in 
predator- prey systems, where food types vary little in nutritional 
composition, their drawbacks for herbivores may be considerable, as 
demonstrated by Westoby (1974) and Freeland & Janzen (1974). 
Westoby noted that a balanced diet demands selection of some 
foods which may have low energy content but supply specific nutrients. 
Since there is a limit to how much bulk an animal can ingest, the 
optimal diet is one which obtains the best nutritional balance for a 
fixed volume. The result is that individual nutrients cannot be 
selected optimally, and the diet is broader than predicted by previous 
models. 
A similar prediction is made by Freeland & Janzen" (1974) from 
considering the importance of plant toxins to mammals. Different plants 
~-----------____ """"""'I 
3.77 
produce different toxins, which eaters must be able to destroy. The 
effects of particular toxins are cumulative, so mammals must eat a wide 
variety of foods, in smaller amounts than would otherwise be expected. 
Since the value of a new food cannot be predicted from smell or taste 
(Rozin & Ka1at 1971: laboratory rats), the optimal diet can only be 
achieved by learning the consequences of eating particular items. 
Continuous but cautious sampling is therefore necessary. In effect, a 
theoretically optimal diet is impossible to achieve: the act of 
discovering what is best to eat broadens the diet beyond the best 
solution. 
The models of Westoby and Free1and & Janzen cogently demonstrate 
the weaknesses of assuming that the diet of mammalian herbivores is 
determined solely by energetic profitability. The apparently protean 
food choice of chimpanzees, their maintenance of a varied diet and their 
occasional disregard for abundant foods may be largely explained by the 
need to obtain different nutrients from different foods and to avoid 
large amounts of particular toxins. 
The significance of the observed geographical variations ln food 
selection cannot be fully realised until quantitative data are available 
from different areas. (Variations due to special skills, such as 
smashing palm nuts with rocks or "fishing" for termites with grass stems 
are not relevant here: limitations in insight or learning abilities, 
rather than in the need for optimal food selection, are clearly 
responsible.) However the predicted need for continuous sampling 
suggests that the differences are the products of the process of 
optimisation: different foods would be sampled at different times. 
This may also explain some of the differences between the diets of 
Gombe chimpanzees recorded before and after 1968 (III/2/i). 
It was also noted that the food choice of an individual did not 
appear to be affected by what his companions ate, except In drawing 
attention to what was available (IU/3/iii). 
3.78 
Thus although chimpanzees may eat a diet which is less 
beneficial than theoretic'ally possible, the data are consistent with 
the Vlew that their diet selection lS optimal In relation to practical 
alternatives. Like other mammals, they could select food more 
efficiently if they had better methods of estimating the value of a 
potential food before eating it. 
iii. Methods of finding food 
Smith (1974) pointed out that optimally adapted food-searching 
techniques should not be expected in view of the unpredictability of 
the environment. This is a similar conclusion to that made above in the 
case of diet selection. Furthermore, there is no basis for deciding 
what an optimal food-searching strategy would be. The issue in 
examining food-searching is to find out how far feeding behaviour lS 
constrained by social behaviour. Two questions are therefore relevant. 
First, do lone individuals show adaptive food-searching behaviour? 
Second, is such behaviour shown by individuals even when with others? 
The descriptions of methods of finding food clearly 
demonstrated the ability of individuals to return to known food sources, 
to find sources by searching in suitable habitats, and to interpret the 
significance of subtle environmental signals of the presence of food 
(IV/2). Changes in the frequency of inspecting in different habitats 
appeared to be adaptive responses to the type of food source the habitat 
could be expected to contain. Similar results showing a capacity for 
detailed spatial memory have been found by Menzel (1973) for captive 
chimpanzees, which were able to judge the value of a food source and 
remember its location, after seeing it once. 
3.79 
That such behaviour may be shown by an individual with others 
was illustrated by an example of a male leaving his party, travelling 
to and eating at a known food source, and subsequently rejoining his 
companions (IV /2/iv). This, type of behaviour was comparatively rare, 
however: normally when an individual left a party to feed he did not 
return immediately. It ~s therefore important to see whether individuals 
travelling together were able to make use of their knowledge of food 
availability in particular habitats or sources. 
Evidence on this point is difficult to obtain without 
experiment. It was noted that intention movements appeared to 
communicate information about what the actor expected to find (IV /3/i) • 
However, only Menzel's (1971) experiments clearly demonstrated that 
chimpanzees without information tend to follow others who know the 
location of food. 
This proven ability shows that the efficiency of locating food 
sources may actually be increased by the formation of parties. If 
chimpanzee A knows only poor food sources, and B knows the location of 
a good one, the inference from Menzel's experiments is that A will tend 
to follow B to his good food source, and thus directly find a food 
source which he might otherwise have missed. Of course, this may be 
disadvantageous for B if the source contains only enough food for one. 
The question of competition for food in different party sizes is a 
separate issue however, and will be treated below: here the concern is 
to note that the ability to locate a food source may be increased by 
the presence of other individuals. 
In the hypothetical example A finds a food source by following B. 
At any time an individual is in potenFial contact with different pools 
of information according to who his companions are. If the tendency 
to follow is increased when an individual is ignorant, and the tendency 
3.80 
to lead is increased when he knows of good food sources, party travel 
(in relation to food-searching) would be a simple function of the "best" 
information available. 
Undoubtedly the situation ~s much more complicated than this. 
The timing and direction of an individual's travel pattern is affected, 
for example, by how sated he is, his social independence and his 
familiarity with the particular place where he happens to be. His 
movements may also be affected by his expectations of his companions' 
response. It is possible, for instance, that if an individual knows of 
a food source, but also knows that by walking towards it in a normal 
manner he will attract too many companions, he may be able to disguise 
his behaviour to modify his companions' response. 
Clear proof of the ability to deceive was cited by van Lawick-
Goodall (1971). After unscrewing the bolts of a banana box, Figan would 
sit on the box without opening it or in any way showing excitement, as 
~ 
individuals normally do, at having achieved access to bananas: only 
when more high- ranking individuals had left camp did he open the box and 
eat the bananas. Sometimes Figan even led other individuals out of 
camp, left them, and subsequently returned to be fed. (At that time Figan 
was only fed when adult males were absent.) His behaviour in the first 
instance suggests the ability to hide an expectation about food, and in 
the second to act as if he knew of an attraction which did not in fact 
exist. These abilities appear to be individual skills: Evered also 
learned how to open banana boxes, but did not wait for high- ranking 
individuals to leave: as a result he invariably lost more of his 
bananas to others. 
It was my impression that similar behaviour occurred in finding 
natural foods. Sometimes while a groom~ng party broke up, an individual 
would wait watching his companions move off, and only when the last was 
3.81 
out of sight would he turn and head in another direction. 
The value of a food source to an individual may fall as more 
individuals come to it: if feeding sites are limited, competition 
occurs and the individual who knew the food source suffers by having 
been followed to it. Chimpanzees are capable of deception, but only to 
a limited extent: Evered did not learn Figan's trick in camp. Therefore 
although increased party size may facilitate the spread of information 
about food sources, it may actually decrease overall feeding efficiency. 
Whether or not it does so depends on the relationship between party size 
and amount of food available at the food sources. This will be discussed 
below. 
2 . Seasonal changes in environmental variables 
~. Habitat structure 
The only important change was in the amount of leaf, which ~s 
greater in almo&t all species during the wet season (Clutton-Brock 1972), 
causing reduced visibility compared with the dry season (McGinnis 1973). 
No attempt was made to measure seasonal changes in habitat structure. 
There were no expected or discernible differences between the two dry 
seasons. 
~~. Climate 
No significant climatic differences were apparent between Dl 
and D2 . In the wet season rain was frequent and humidity was higher than 
in the dry seasons. 
~~~. Food availability 
There were several indications of a change in overall food 
availability between Dl and D2 • First, the preferred food in D2 failed 
almost completely in Dl (II/3/i). Second, weights of adult males were 
lower, and more individuals licked for salt in camp in Dl than D2 : 
these measures are both indicators of metabolic stress, and were 
significantly correlated across nine 3-month periods (VII/2). Third, 
feeding behaviour on E. guineensis fruit indicated that competition 
3.82 
(both inter- and intra-specific) was more severe in Dl than D2 (VII/2/i). 
~v. Food distribution 
No independent measures of food distribution were made. Seasonal 
differences in the distribution of food sources used by chimpanzees are 
discussed in Chapter 5. 
The data indicate that food availability was lower in Dl than D2, 
while other important environmental variables were constant. Changes ~n 
habitat structure, climate and food availability occurred between wet and 
dry seasons; the fact that weights, frequency of salt-licking and 
behaviour while eating E.guineensis fruit were more similar between W 
and DZ than be~ween Wand Dl indicates that food availability was higher 
in W than Dl . 
3. The relationship between food availability artd social structure 
In Dl food avai labi lity appeared lower (VIII/Z) and party sue 
was smaller than in D2 or W (VI/4/iii). The change in mean party s~ze 
did not appear to be the consequence of changes in dominance 
relationships or the availability of oestrous females (VII/3/iii); nor 
did it seem to be due to changes in habitat or climate (VIII/2). It 
may therefore have been the consequence of changing food availability. 
This suggestion is supported by evidence showing a relationship between 
party size and competition for food. 
Data collected in the artificial feeding area showed that party 
sue was correlated with the number of bananas available (VI/4/i). 
Furthermore, individuals meeting in camp were more likely to travel 
; 
3.83 
together if given bananas. Though food availability could not be 
accurately measured at natural food sources, there was evidence that 
party size was greater when more feeding sites were available (VI/4/ii). 
At individual food sources, therefore, party size appears to have been 
influenced by the probability of each individual getting food. This 
implies that the relationship between food availability and party size 
was mediated by the extent of competition. 
Overt competition for feeding sites was uncommon at most natural 
food sources: it was observed about once every 20 hours (VI/2/ii). It 
appeared, however, that individuals frequently experienced competition 
in the sense of a depressed feeding efficiency due to the presence of 
other individuals. Important factors were (1) occupation by one 
individual of a preferred or the only available feeding site (VI/2), and 
(2) interruption of feeding by social interactions (VI/3). In large 
parties at a food source with many feeding sites, relatively more 
individuals fed than when few feeding sites were available (VI/2/ii). 
The data therefore suggest that competition occurred because the number 
of adequate feeding sites available at a food source was often less than 
the number of chimpanzees in the party. 
If the seasonal change in party SLze was caused by a change in 
the extent of competition (as a result of a change in food availability), 
then the data should reveal evidence of greater competition for a given 
party size in Dl than in D2 or W. In all seasons there was evidence of 
competition increasing with party size (VI/S). However, the rate of 
decline of feeding probability with increasing party size was higher Ln 
Dl than Ln D2 or W (VI/S). This result is in the predicted direction, of 
greater competition Ln Dl . (As party size increased the proportion of 
individuals feeding fell. It may be argued that this was due to the more 
likely presence of individuals, adult females perhaps, who fed less than 
3. 84 
the target regardless of party Slze. The data shown here do not 
contradict this hypothesis. However, (1) Bygott (1974) has shown that 
the feeding time of the target individual is greater when alone than 
when with others; (2) there was no indication that more females were 
included in parties in Dl than D2; (3) in species with little size 
difference between the sexes, females have longer feeding time than males 
(Colobus badius, Clutton-Brock 1972; Macaca rtemestrina, Bernstein 1970; 
waterbuckKobusdefassa, Spinage 1968; impala Aepyceros melampus, Jarman 
& Jarman 1973); only if males are much larger than females are they known 
to feed for longer (gorillas Gorilla gdrilla beringei, Harcourt pers. 
comm.). It lS therefore most unlikely that the greater apparent 
competition In Dl than D2 was caused by differences In party sex ratios 
between seasons.) 
We may therefore conclude that low food availability caused an 
increase in competition, to which individuals responded by dispersion. 
Increased scattering of individuals within a social group was 
predicted by Crook (1970a) to be a general response of primates to food 
scarcity. However, there have been few studies in which a change in 
food availability has been convincingly demonstrated. Hall (1963) found 
that Papio ursirtus dispersed into small independent parties when the 
rising waters of Lake Kariba removed land which was previously available. 
Similarly when part of the home range of Lernur catta was eroded by river 
action, Jolly (1972b) observed an increase in inter-individual distances. 
In other studies wet and dry seasons have been compared (Therdpithecus 
gelada, Crook 1966~; Cercopithecus aethiops, de Moor & Steffens 1972; 
chimpanzees, Nishida 1974), and it has been assumed that more food was 
available in the wet than the dry season. Although this may often be 
true, it needs to be shown. In the absence of evidence to the contrary, 
an observed change in dispersion tendency could be due to differences in 
r 1 
3.85 
food type rather than food abundance, as shown in Cercopithecus mitis 
(A1drich-B1ake 1970) and Presbytis entellus (Rip1ey 1970). Furthermore, 
differences in vegetation structure are known to affect the degree of 
scattering (Crook 1970a, Wi1son 1972). 
Although there are few hard data, it seems likely that most 
primates disperse more within their social groups when food is scarce. 
In many other vertebrates there is evidence that the state of dispersion 
~s affected by the level of food supply (Watson & Moss 1970). However, 
the direction of the relationship is not always the same as suggested for 
primates. For instance, Cody (1974) found that finches (Emberizidae and 
Fringillidae) gathered in larger flocks at a time when food was hard to 
find. In this case the response appeared to be mediated by the need for 
each individual to know that the food source to which he was travelling 
had not recently been visited by a large flock: the flocks acted as 
"return time regulators 11, ensuring that visits to food sources coincided 
with replenishment of food items. This example demonstrates that the 
relationship between food availability and social structure may vary 
between species as a consequence of differences in feeding behaviour. 
If the methods by which individuals find food involve cooperation, it is 
possible that during a time of food scarcity the benefits of increased 
cooperative behaviour may outweigh the costs of increased competition by 
feeding in groups. Since this appears not to be the case with chimpanzees, 
the significance of apparently cooperative behaviour in finding food is 
examined below. 
The data indicate that low food abundance caused individuals to 
disperse in order to avoid competition for food. It is also important 
to find out why individuals should congregate when food is abundant. 
This question will be discussed below. 
3.86 
4. The significance of apparently cooperative behaviour In finding food 
i. Food-calling 
There are two alternative functional explanations of food-calling. 
The first lS that it is cooperative, In the sense of being "collaborative 
behaviour of two or more individuals In the production of some common 
behavioural effect" (Crook 1970b). The second is that although the 
individual who gives and responds to the food-call both obtain benefit, 
the benefits to each are of different kinds. These explanations are 
examined separately. 
(a) Cooperative explanations. If the benefit to an individual 
who glves a food- call comes from an increased chance of finding food, he 
relies on another individual giving a food-call in the future. While he 
lncurs immediate costs, therefore (in sharing a food source), he obtains 
only the expectation of benefit. Two explanations may be suggested 
whereby a future benefit may be sufficiently likely to occur to permit 
food-calling to be advantageous. 
First, Trivers (1971) suggested that if the costs to the giver 
of an altruistic act are less than the benefits to the recipient, 
altruism may evolve through the expectation of reciprocity. However, 
this is not satisfactory. It is not clear why an individual's best 
strategy is not to cheat, by responding to but not giving altruistic 
acts: there is no evidence of a mechanism which punishes individuals 
who do not give food-calls. Reciprocal altruism has not actually been 
shown to occur in animals. In the case of the warnlng calls of birds, 
for instance, Trivers listed five possible advantages to the caller 
other than the expectation of a reciprocal act. 
Second, altruistic behaviour may evolve through kin selection 
(Hamilton 1964). There are no data yet on coefficients of relationship 
among chimpanzees, but there is evidence of a partially closed breeding 
, 
3.87 
system (Tutin, in prep.) which may allow kin selection to occur within 
communities, as Bertram (in prep.) has argued for lions (Partthera leo). 
However, kin selection would not explain why adult males are the only 
individuals who regularly give food-calls. 
There is therefore no clear explanation of how cooperative 
behaviour could be maintained . Furthermore, the selective advantage, by 
whatever mechanism, appears to be very small, since food- calling was 
uncommon at the time when it would have given the greatest benefits (D l , 
when food was scarce: IV/4/iv). This suggests that food-calling is not 
primarily a cooperative behaviour: the benefit to the caller appears to 
be different from the benefit to the individual who arrives to share his 
food source. Chimpanzees gave food-calls when least likely to incur 
costs from competition: the food sources had many feeding sites (lV/4/iii) 
and the caller was already present in a good feeding site when his 
beneficiary arrived. It may be true that as food becomes scarcer the 
costs to the caller rise faster than the benefit to others, and hence 
faster than the probability of a reciprocal act. While this may explain 
why there is less food- calling when food is scarce, it does not explain 
why, if food-calling occurs only at a time when it is easy to find food, 
there should be any advantage in terms of reciprocity. 
(b) Alternative explanations. It is fairly clear that one 
benefit to an individual who responds to a food-call by approaching the 
caller is to get food. The important question is what benefit the caller 
obtains. Two possibilities are considered here. 
First, affiliative behaviour may be important ~n establishing a 
social unit. Crook (1972) argued that a social unit functions to 
protect offspring by guaranteeing them access to sufficient environmental 
resources: ~n some species it is also clearly important in providing 
protection for offspring from predators . In chimpanz.ees, however, it ~s 
. ~ 
If 
.J 
3.88 
possible that the protection of resources ~s a less important consequence 
of affiliative behaviour than competition between the males of different 
communities for access to females. This will be examined in Chapters 5 
and 6. 
Second, s~nce the consequence of food-calling is an increased 
probability of both males and females arriving, the caller may obtain 
social and thus sexual benefits. Occasionally he may attract an oestrous 
female, with obvious advantages. More often the females who arrive will 
be anoestrous. In this case, the male may improve the quality of his 
relationship with the female, who must sometime select a mate. It is to 
the female's advantage to select the best male possible: his ability to 
find food and willingness to share a source may be a valuable measure of 
physical and psychological condition. Interactions following food-calling 
may thus have important consequences on relationships within the 
community in terms of intra- sexual competition. 
This suggestion can best be tested by companng the frequency of 
food-calling of different individuals with their sexual success. 
Differences between males in the frequency of food-calling were unclear 
(IV/4/ii) but Evered was noted to be a frequent caller. Tutin (in prep.) 
found that Evered copulated less often than most males. These results 
indicate a poor correlation between food-calling and sexual success, and 
tend to argue against the suggestion that an important consequence of 
food-calling was on reproductive competition within the community. 
However, more data are clearly needed. 
~~. Party leadership 
It was argued earlier (VITI/l/iii) that individuals may find 
food as a consequence of following others. Increased efficiency ~n 
locating food sources ~s though to be the function of some bird 
aggregations (Horn 1968, Murton 1971, Ward & Zahavi 1973), and 
3.89 
Reynolds & Reynolds (1965) considered party recombination to be necessary 
for chimpanzees to find their discretely distributed food sources. 
However, care is needed for such an interpretation. If the function of a 
behaviour consists of the consequences through which natural selection 
acts (Hinde in press,b) a system is needed for deciding what those 
consequences are. 
Simple elimination is sometimes used. For instance, Goss-Custard 
et al. (1970) listed the possible advantages of flocking in wading birds. 
Increased feeding efficiency did not explain why birds gathered in flocks: 
consequently it was concluded that defence against predators was the most 
important result of flocking. Conversely, Murton (1971) decided that the 
function of flocking in rooks (Co~vus frugilevus) was to increase feeding 
efficiency only after accepting the views of Zahavi (1971) that flocking 
did not result in decreased predation pressure. The most obvious 
disadvantage of this system 1S that it relies largely on disproof of a 
limited list o~ alternatives. There may in fact be important benefits of 
flocking other than from food-searching and defence against predators. 
For instance, a male may increase his reproductive success by competing 
with as many other males as possible. 
If there are several beneficial results of a structure or 
behaviour, there is no universal principle for distinguishing between 
their relative importances in natural selection (Williams 1966). Lazarus 
(1972) argued that to speak of relative importance in these circumstances 
was meaningless since the selective pressures influencing behaviour are 
permanently interacting. However, one circumstance in which this may be 
meaningful is where the benefits associated with a behaviour vary in the 
extent to which they are appropriate in different contexts. 
Party size was largest when food was most abundant (VI/4/iii). 
From the point of view of an individual needing to locate food, therefore, 
3.90 
information obtained by following others was most available when it 
appeared least important. As in the . case of food-calling, the important 
benefit of joining a party appeared not to be the gaining of information 
about food sources. 
It seems more likely that the most important consequence of party 
formation is on individual relationships, as argued above. Because 
feeding efficiency is reduced by increasing party size, individuals can 
only afford to associate when feeding efficiency does not need to be 
maximised, or at food sources which have so many feeding sites that 
competition is slight. No special signals were seen to be given by 
"leaders", and deception sometimes occurred (VIII/l/iii). The ability 
to interpret intention movements may be important ~n enabling an 
individual to decide whether to stay with others: the evidence does not 
suggest that finding food is the consequence which promotes the tendency 
to congregate. Social behaviour appears to reduce feeding efficiency: 
the benefits of party formation must therefore be sought in other areas 
than feeding behaviour (Chapter 6) . 
I. Introduction 
Chapter 4 
PREDATION 
4.1 
Since the first report by Gooda11 (1963) of predation on mammals, 
chimpanzees have been seen to eat fifteen species of mammals ranging in 
s~ze from mice (H1adik 1973) to adult red co10bus monkeys (van Lawick-
Gooda11 1968): these include eight primate , three ungulate, two 
rodent and one carnivore species (van Lawick- Goodall 1968, Suzuki 1971, 
Hladik 1973, Nishida 1974). Cannibalism has also been seen twice 
(Suzuki 1971, Bygott 1972). Predation on mammals has been seen in 
habitats from miombo woodland (Kawabe 1966) to primary forest (Suzuki 
1971), and over the extremes of the geographic range of chimpanzees from 
Western Tanzania (van Lawick-Goodall 1968) to Senegal (Brewer, pers. corum.: 
rehabilitated chimpanzees). It has been found in studies with (Teleki 
1973) and without (Suzuki 1971) artificial feeding. In many field studies 
predation has not been seen (Nissen 1931, Azuma & Toyoshima 1962, 
Kort1andt 1962, Reyno1ds & Reynolds 1965, Izawa & Itani 1966, de 
Bournonvi1le 1967, Dupuy 1970, Izawa 1970, Jones & Sabater pi 1971, 
Albrecht & Dunnett 1971, Kano 1971): ~n most of these there was little 
direct observation of chimpanzees ~n the wild, and predation has since 
been reported ~n several of these study areas. There is therefore no good 
evidence of a chimpanzee population which does not occasionally eat 
mammals. Though predation appears to be uncommon wherever it occurs, it 
is evident that it is a widespread phenomenon. 
Despite the numerous reports of predation by chimpanzees, there 
have been few discussions of its ecological and behavioural significance: 
because of its infrequency systematic data are hard to get. The long-term 
studies at Gombe have provided the most detailed descriptions to date 
(van Lawick- Goodall 1968, Teleki 1973), but the data in these studies were 
4.2 
mostly collected near camp, at a time when chimpanzee behaviour was 
influenced extensively by large - scale artificial feeding. Since 1970 
observations on Gombe chimpanzees have been made by following target 
individuals wherever they ~o, and the scale of artificial feeding has 
been enormously reduced since 1968 (Wrangham 1974). Consequently there 
now exists a pool of data on predations in a relatively natural situation. 
These data are analysed here in an attempt to understand the significance 
of predation both to chimpanzees and to their prey populations. 
Wherever predations have been seen chimpanzees have shared the 
meat, as first described by Goodall (1963). No explanations have been 
given as to why this occurs, other than Kortlandt's (1972) idea that it 
is "a combination of displacement and reassurance behaviour". To label 
meat-sharing as altruism (Hinde 1974, Cowgill 1972) is not explanatory 
in an evolutionary sense if the behaviour is considered adaptive in the 
context in which it LS seen: the theory of natural selection demands 
that the assoc~ated benefits be greater than the costs. Meat-eating is 
therefore examined here with the aim of understanding why chimpanzees 
share their prey. 
A detailed description and review of predation by chimpanzees 
has been given by Teleki (1973): the new data presented here are 
selected to solve the problems stated above. Much of the information 
comes from the observations of numerous students at Gombe, mostly from 
August 1970 to August 1973. However, where it was desirable to use data 
extending over a longer period, this has been done. The predation data 
had to be used selectively: only "well observed" predations were 
analysed, but those which were "well observed" varied according to the 
type of information taken from them. Reliability between observers was 
high because the data were mostly simple (e.g. records of which 
individuals were present) and behavioural descriptions were standardised 
to an agreed glossary of terms (e.g. for components of agonistic 
behaviour). For all detailed descriptions of unusual behaviour the 
observer is cited, except for my observations. 
4.3 
Predation on a chimpanzee has been seen once ~n Gombe, and little 
interest was shown in the meat (Bygott 1972). Because cannibalism clearly 
has special ecological and adaptive consequences it is not included among 
the predations analysed in this chapter. 
11. Predation on Animals other than Mammals 
1. Irttroduction 
The significance of mammals as a source of food can only be 
understood when data are available on other animals as prey. For instance, 
Kortlandt (1972) suggested that mammals could hardly be regarded as being 
nutritionally valuable to Gombe chimpanzees since they appeared not to eat 
more easily obtained animals. Data on the eating of all animal foods ~s 
therefore reviewed here, and recent new observations are described. 
2. Invertebrates 
~. Types eaten 
Invertebrate foods recorded in Gombe are listed in Appendix 3.2. 
All are insects: the most important in the diet (in terms of total intake) 
was the termite Macrdterrnes bellicosus, as described by van Lawick-Goodall 
(1968). Some insects (in galls and figs) were eaten incidentally with 
plant foods, but others were searched for when seasonally available. 
During the "termite season"for instance, termite mounds were visited ~n 
the same way as plant food sources: individuals clearly knew the 
location of mounds and sometimes travelled for an hour or more inspecting 
and eating at a regular series. 
I ' 
1 
I 
I 
4.4 
Van Lawick- Goodall (1968) has shown that insects may be eaten in 
all months of the year. Since figs appear to be eaten in all months it 
seems likely that chimpanzees have a regular, if sometimes small, supply 
of animal food. 
Termites have been recorded in chimpanzees' diet in several areas 
(Itani & Suzuki 1966, Suzuki 1969, Nishida 1974 in Tanzania; Jones & 
Sabater pi 1971, Rio Muni; Hladik 1973, Gabon; Brewer pers. comm., 
Gambia; Struhsaker & Hunkeler 1971, indirect evidence in Cameroon): 
ants are also eaten in Tanzania (Suzuki 1969, Nishida 1974) and Gabon 
(Hladik 1973). Records of other insects as food are sporadic. Probably 
chimpanzees can only afford to eat colonial species regularly, because 
of the time involved in finding others. The two caterpillars which are 
eaten in Gombe are found in large numbers on their food- plants. 
The only study in which chimpanzees clearly ate many more 
insects than in Gombe was of individuals reintroduced onto an island In 
Gabon (Hladik 1973). At least twenty species of insect and one scorplon 
were eaten, and these formed the major part of the chimpanzees' animal 
food. 
Species which were not eaten were generally ignored: arboreal 
termites were often seen but never eaten. Butterflies and grasshoppers 
were seen to be killed after being played with, but were not eaten. 
ll. Sex differences 
McGrew (in prep.) has shown that there are clear sex differences 
in eating insects: females eat Macrotermes bellicosus, Dorylus nigricans 
and Oecophylla longinoda more frequently than males. The sex difference 
was considered to be related to differences in ranging patterns 
(Chapter 5), since if males travel more widely they cannot return as 
frequently as females to known sources of these species. It is also 
4.5 
possible that males like them less than females: this was indicated by 
a few observations when both sexes were together at these foods (and at 
Crematogaster sp.), but further data are required. 
3. Vertebrates 
1. Fish 
The only fish which chimpanzees encountered were dead "dagaa" 
(Stolothrissa sp.) found drying on sand. These were regularly eaten by 
baboons but never by chimpanzees. 
11. Amphibians 
Though various frogs and toads occurred in the Park they were 
nocturnal and rarely seen. There 1S no evidence that they are eaten by 
chimpanzees. 
iii. Repti les 
No rept~ les have been seen to be eaten. Monitor lizards <CVaranus 
niloticus and V. exanthomaticus) were occasionally met and normally 
ignored: Sherry once approached and flicked the tail of a monitor but 
made no attempt to catch it. Skinks (mostly Mabuya maculilabris) were 
ignored 1n camp, though a juvenile male played with and mutilated two 
spec1mens. Reactions to snakes have been described by van Lawick-Goodall 
(1968) . 
1V. Birds 
At least nine speC1es of birds have been seen to be eaten 
(Appendix 3.2). Individuals varied in how often they ate them: out of 
14 birds or nest contents eaten from April 1971 to September 1973, 7 were 
found and eaten by the juvenile female Pom. Pom's success was due to her 
habit of inspecting weaver nests and perhaps her sensitivity to other 
stimuli such as the calls of young francolin. The other seven birds were 
eaten by adult males (5 times) and females (twice). The only healthy 
adults seen to be caught were weavers in their nests. 
4.6 
There was normally little meat available when a bird was 
captured, and the eater was often ignored by other chimpanzees. Begging 
occurred for large prey, as in the case of mammals. 
Ill. Factors Affecting Predation on Mammals 
1. Introduction 
Predation by chimpanzees has been observed so rarely that no 
clear picture has yet emerged of why it occurs. Te1eki (1973) noted 
that the prey tend to be small (less than 9 kg.), and considered that 
some of the factors affecting predation included the frequency of 
encounter with prey species, the time of day, and the "disposition" of 
the chimpanzees: few data were available to test these ideas. Although 
it was found to be affected neither by convenience nor hunger, Te1eki 
noted some seasonal variation in the frequency of predation, and 
speculated that this was related to seasonal changes in the availability 
of other foods: van Lawick-Gooda11 (1971) speculated that there might 
be "meat-eating crazes" lasting one or two months. Finally, Te1eki's 
data indicated that the frequency of predations may have been affected 
by the methods of providing bananas at the artificial feeding area. 
These ideas are tested in the light of recent observations, and other 
factors possibly affecting predation are examined. 
Different prey were caught in different situations. Detailed 
analysis of situations in which they were caught is therefore made 
separately for each species. 
4.7 
2. Factors Affecting Predation on All Species 
i. Age-sex class of chimpanzees 
Between August 1970 and August 1973 the identity of 18 killers 
was seen: all were adult males. Though evidently less often than 
adult males, females may sometimes kill: three (one with a 2~-year old 
infant) have taken young bushpigs. Adolescents of both sexes have often 
been seen to take part in hunts, and males have twice been seen to kill 
(van Lawick-Goodall 1968, Busse & Riss in prep.) • 
11. Individual differences 
The best sample within an age-sex class with approximately equal 
observation hours is of Northern males. 1he number of kills recorded 
for each in the study period is shown in Table 4.1. 
Table 4.1 
Number of kills made by Northern males 
Data from August 1970 to August 1973. Only direct observations are 
included. 
Number of observed kills 
RG 
2 
MK 
2 
RM 
2 
Indi viduals 
FB 
2 
EV 
o 
FG 
3 
ST 
1 
JJ 
3 
The even distribution of kills across individuals indicates that 
there are no marked differences in killing tendency. In December 1973 
Evered was seen to kill for the first time (Tutin, pers. comm.). No 
differences in hunting tendency were perceptible, and it is remarkable 
that even Faben, with a paralysed arm, has caught a bushpig and a redtail, 
in addition to a baboon in 1968 (Teleki 1973): he caught the young red-
t ail by chas ing when it came to the ground in fores t. wi lly-Wally, wi th 
a partially paralysed leg, has also caught a colobus monkey. 
I ~ 
4.8 
~~~. Time of day 
Teleki (1973) found that predations were most frequent early in 
the day: this result appears to have been an artefact. The relationship 
of diurnal variation in predation frequency to that in observation hours 
~s shown in Fig. 4.1. Predations occurred throughout the day, but Fig. 4.1 
shows that they were slightly less frequent in the middle of the day in 
l ' b ' '(1' 2 re at~on to 0 servat~on t~me Ko mogorov-Sm~rnov D = 0.71, X = 0.9, 
max 
d.f. = 2, n.s.). Teleki's sample of 30 predations all occurred before 
1500. The difference is probably the result of sampling methods: 
Teleki's data were collected by observers based at the artificial feeding 
area, which chimpanzees were more likely to visit in the morning (Fig. 3.4). 
In the natural situation predations appear to occur during the periods of 
greatest activity: diurnal variation in travelling patterns is described 
in Chapter 5. 
~v. Time of year 
Monthly variations ~n the numbers of predations recorded over 
six years are shown in Fig. 4.2. The peaks in predation frequency are 
located differently ~n the two three-year samples, and there is little 
evidence of an overall seasonal variation. (For predations occurr~ng ~n 
2 
successive three-month blocks, X = 7.4, d.f. = 3, n.s.). 
This does not necessarily mean that ecological factors did not 
affect the frequency of predation. It was shown earlier that ecological 
conditions can vary between the same months in different years. To 
compare seasons in different years, predation frequency must be related 
to observation hours, within a group of individuals with approximately 
equal predation tendency. This is done in Table 4.2. 
I ~ 
Fig.4.1 Diurnal variation ~n frequency of predation 
Predations seen during observation of a target individual are scored, 
provided that the time of capture of the prey was recorded. Diurnal 
variation in observation time was sampled from every fifth observation 
period in 1970, 1971 and 1972. 
300 6 
Obsvn. 5 No. of 
200 4 
hours 3 predations 
0 100 2 ~ 1 
(N=',288) 0 0 (N= 25) 
4 6 8 10 12 14 16 18 20 22 
Time of day 
Fig. 4.2 Monthly variation in predation frequency 
All predations observed between 1968 and 1973 are shown. The figure 
indicates that there was no consistent cycle of predations over the 
year. However, evidence of seasonal variation within years is shown 
in Table 4.2. (Variations In observation time are not taken into 
account.) 
NUMBER II OF 
PREDATIONS 8 
6 
RECORDED 4 
2 
o 
J FMAMJ J ASON D 
MONTH 
~ baboon 
_ colobus 
D other species 
1968 - 1970 
1971 - 1973 
Table 4.2 
Seasonal variation in rate of predation by Nbrthern males 
Data from all observers. Only predations attended by a Northern 
male as target are included. 
Period Number of Number of Predations per predations obsvn. hrs. 100 obsvn. hrs. 
Jan-Mar 1972 1 246 0.41 
Apr-Jan 1972 0 224 0.00 
Ju1-Sep 1972 3 464 0.65 
Oct-Dec 1972 3 374 0.80 
Jan-Mar 1973 1 327 0.31 
Apr-Jun 1973 6 548 1.09 
Ju1-Aug 1973 4 310 1.29 
If food availability affected the probability of predations, 
4.9 
their seasonal variation (Table 4.2) might be expected to correlate with 
variations in weight and salt-licking (Tables 3.24 and 3.22 respectively), 
but this was not significant (for high- weight periods, Spearman r = 0.42, 
s 
N = 7, n.s.; for low frequency of salt-licking, Spearman r = 0.19, N = 6, 
s 
n.s.). The ecological factors affecting predation were therefore 
apparently not food availability alone. Although chimpanzees only 
preyed on the young of bushpig, bushbuck and baboons, there was no 
indication of regular seasonal changes in the frequency of particular 
species as prey. 
Despite the lack of overall seasonal variation, it ~s possible 
that the rate of predation might have been non-randomly high ~n 
particular months: if so, this would indicate variations in prey 
availability or predation tendency. This was tested by comparing the 
frequency distribution of the number of predations seen per month with 
the expected number if predations were independent events • . The result 
is shown in Fig. 4.3. Predations were clumped within certain months 
only slightly more than expected (for 0 - 2 predations per month, 
Fig. 4.3 Clumping of predations in months 
Data are taken from all observed predations (N = 71), seen from 
March 1968 to January 1974. Expected frequency is calculated from 
Poisson distribution, which assumes that all events occur at a 
fixed probability, regardless of the occurrence of other events. 
The similarity of the two curves indicates that the underlying 
assumption is acceptable. 
Frequency 
20 
(months) 
10 
o 1 
(l)served l 
Expected I 
2 3 4 5 6 
Predations per month 
If 
4.10 
2 X = 1.32, d.f. = 2, n.s.: for 3 - 6 predations per month, expected 
values are too small to test). Since observation time was highly 
variable between different months, there is no evidence that predations 
occurred together more than expected by chance. 
This is slightly surprising Slnce it sometimes appeared that 
chimpanzees hunted in patches. For instance, on three days out of four 
In August 1974 Godi met and carefully watched a redtail troop in one 
place: similar observations have been made several times (Busse & Riss, 
in prep.). Repeat encounters could of course have been the chance 
result of predator and prey species both returning to a temporarily 
abundant food area. Other factors than the frequency of encounter, 
however, may affect the probability of a successful predation (below). 
If so, even hunting in patches would not lead to a clumping of 
predations in time. 
v. Location 
The locations of 33 predations recorded between 1971 and 1973 
are shown in Fig. 4.4. The number of predations recorded In the major 
valleys was distributed similarly to the observation time In each of 
them. Although predations have been seen most frequently In areas of 
forest-grassland mosaic (Gombe; Mahali, Nishida 1974) all predations In 
Gornbe have been seen in or close to forest, where prey species were 
more often encountered. 
Predations were not clumped around the artificial feeding area. 
This is not surprising since prey animals were distributed all through 
the study area. 
Vl. Size of prey 
As found by Teleki (1973), only blue, redtail and colobus 
monkeys were eaten as adults. Predatory interest in adult bushpig, 
Fig. 4.4 Frequency of predation in different parts of the study area 
All predations whose location was recorded between 1971 and 1973 are 
shown (excluding one in Rutanga Valley, in which vegetation was not 
mapped). Only the major areas of forest are shown. The artificial 
feeding area is to the low~r left of centre (where a baboon was killed, 
just in grassland). Most predations occurred in forest (at least one 
continuous tree canopy layer, no grass). 
Scale (km) 0 2 3 4 5 I I I , 
Predation sites 
1971 - 1973 
Colobus 
2 Bushpig 
3 Bushbuck 
4 Baboon 
5 Redtail 
6 Blue 
• Grasslard 0 Forest 
: I 
I 
I 
I 
--L1 
4.11 
bushbuck or baboons has not been seen: an adult female baboon was once 
held on the ground to prevent her interfering while her infant was eaten, 
but no attempt was made to harm her. 
V11. Social stress 
From the literature on Gombe predations (van Lawick-Goodall 1967, 
1968), Kortlandt (1972) suggested that one factor contributing to 
predatory behaviour was social stress among a group of chimpanzees. 
Kortlandt's idea was based largely on an artist's impression of the 
killing of a juvenile baboon (in van Lawick-Goodall 1967): the killer 
and his companions were indeed portrayed with faces and gestures seen in 
the context of intra-specific aggression. On the four occasions when I 
could clearly see the behaviour of a chimpanzee who was killing prey he 
appeared to me "expressionless": Bygott (pers. cornrn.) agrees that the 
picture is atypical. When obvious signs of excitement occurred they 
appeared after the prey had been sighted: observations provide no 
support for the s~ggestion that social stress encouraged predatory 
attempts. 
3. Vulnerability of different prey species 
1. Red colobus monkey 
Chimpanzees frequently met colobus without showing any interest 
1n them: the circumstances in which they watched or chased them seemed 
highly specific, and I found at the end of the study that I could predict 
well the probability of predatory interest. The major determinant was 
th-e continui ty of the canopy layer. In continuous canopy the monkeys 
were normally ignored: in broken canopy they were not. One particular 
tree, Albizia spp., appeared to be particularly dangerous: from 14 good 
descriptions of colobus predations, the prey was first caught in trees 
4.12 
emerging from the canopy in 9 cases, on the ground after having left 
such trees in 3 cases, and in open canopy 'vood1and once. Six of the 
emergent trees were A1bizia spp., which are shown in Plate 4.1. 
Significant features are the ·strong uncluttered boughs, which chimpanzees 
can climb fast, the long distances between the boughs, and the fine 
tapering of the outer branches. In discontinuous canopy co1obus normally 
use a strong branch as a springboard: many of the end branches in 
Plate 4.1 would merely sag as the co1obus ran up them. These trees 
therefore act as a trap from which co1obus can only escape by dashing 
to the ground or by dropping into the next canopy layer. Twice when I 
saw chimpanzees chase co1obus on straight branches they gained fast: 
one was caught and the other jumped down. Chimpanzees sometimes followed 
co1obus on aerial leaps (one misjudged the gap and fell a clear 15 m. to 
the shrub 1ayer~), but more often the danger seemed to come from others 
waiting in the surrounding vegetation. Van Lawick-Gooda11 (1968) has 
given convincing 4escriptions of coordinated hunting techniques, and 
these have often been seen ln the context of cutting off escape routes. 
The importance of the continuity of the canopy layer is 
illustrated by the fact that co1obus may sometimes actually attack 
chimpanzees. On 27 July 1973 five chimpanzees arrived under 
P. curate11ifo1ia trees and collected fallen fruit. Twenty metres 
from the nearest male, and ignored by all the party, a troop of co1obus 
watched them. Leaving the body of the troop, an adult male climbed to 
the end of a branch 6 m. off the ground and watched Jomeo beneath him, 
then leaped off and landed on Jomeo's head: Jomeo, completely 
surprised, gave a "waa-bark", waved his arm, and sprang after the 
monkey, but the latter was already safe, back ln the continuous canopy. 
Jomeo sat agaln and ate his fruit. Two hours later, however, when he 
found the same troop in broken canopy, Jomeo chased and killed a sub-
Plate 4.1 Albizia ? gummifera emerging from forest canopy 
Red colobus monkeys were frequently caught by chimpanzees in Albizia sp. 
Chimpanzees could run faster than colobus up the main boughs, and the 
only escape route was often a weak jump to the low canopy, where other 
chimpanzees might be waiting. 
4.13 
adult. Brown (1971) saw Sykes monkeys (Cercopithecus mitis) mobbing 
crowned eagles (Stephanoaetus coronatus), which eat them then they have 
the advantage: one monkey regularly approached and snatched at the 
eagle's feet. Kruuk (1972) gives instances of a number of ungulate 
spec~es approaching potentially dangerous carnivores. The reactions of 
co1obus therefore seem fairly typical of prey species. Mobbing reactions 
have been suggested to be adaptive because the prey species keeps the 
predator in sight: if the predator realises this and knows that stealth 
can no longer help it, it will leave and the prey species can return to 
feeding more quickly (Smythe 1970~. 
Defence during attempted predations ~s uncommon, however, perhaps 
because it is unwise. Esi10m & Ruwongeza (pers. comm.) saw Sniff snatch 
an infant co1obus off its mother: while the baby was torn apart the 
mother stayed close and threatened the predators, one of whom caught 
and killed her a minute later. Busse & Riss (in prep.) made a similar 
observation. The normal reaction to the presence of chimpanzees is 
vigorous calling and moving away (though this observation is uncertain 
because the presence of man causes the same response). 
C1utton- Brock (1972) noted that in most areas of evergreen forest 
~n Gombe the canopy was badly broken, and tentatively attributed this to 
human influence (burning and clearing) before the area was scheduled as 
a reserve in 1943. Several typically emergent trees were co1obus food 
species (A1bizia?glaberrima, Newtonia buchananii, Ficus val1is-choudae) 
and were common (ranked 2nd, 11th and 33rd respectively in Mkenke and 
Kahama valleys) (C1utton-Brock 1972). Thus the high predation rate ~n 
Gombe may be partly the result of a high probability of chimpanzees 
finding co1obus in trees from which there are limited escape routes. 
Further evidence of the importance of vulnerability to the 
probability of predation comes from analysis of the age-sex class 
4.14 
distribution of prey. Fig. 4.5 shows that more adults have been caught 
when alone than when in a troop. The sample size when observers clearly 
recorded whether or not the prey was alone is small, but the trend is 
supported by the distribution of age- sex classes from unrecorded groups, 
in which adults were the commonest prey: an observer was only likely not 
to mention troop behaviour if the prey was in fact alone. Since there 
were certainly more adults in troops than there were alone (Clutton-Brock, 
pers. comm.) the data indicate that lone adults suffered more predation 
than those in troops. Schaller (1972) similarly found that lions had 
greater hunting success when attacking lone rather than grouped ungulates. 
ll. Bushpig 
While travelling, chimpanzees would stop and search if they 
heard plgS. They were evidently looking for piglets (the largest prey 
taken were half-size, and commonly smaller) and interest fell if a group 
could be seen moving off without young. On three occasions I saw adult 
males search an area where I had heard no pigs: they appeared to have 
smelled fresh scent, as I did. Hugo once spent twenty seconds rummaging 
through an old nest Ca pile of matted grass one metre in diameter) but 
found nothing. Another time Mike heard a sow run off and then searched 
the nearby ground: I lost Mike but found three piglets lying motionless 
and unprotected. The sow did not return to her young for over eight 
hours, and then moved them. 
Piglets were normally snatched from a group of plgS running in 
confusion: the adults often subsequently charged any chimpanzees or 
people on the ground. A form of defence seen once was for three adults 
to stand with a piglet between them. The surrounding chimpanzees, 
observed by Plooij, did not approach though clearly excited by the sight 
of the potential prey, until Hugo threw a rock which hit one of the 
Fig. 4.5 Age/sex classes of colobus killed by chimpanzees 
All available records are shown, August 1968 to August 1973 . The Slze 
of prey was not always clearly described by observers, and the sub-adult/ 
j uveni le and j uveni le/infant categories are therefore merely approximate 
indications: infants were those carried by their mothers. Where the 
observer failed to record whether the prey was found alone or in a troop 
it is more like ly to have been alone. Lone adults appear to be more 
vulnerable than troop adults. 
Troop Lone Unrecorded 
6 
Number 5 
killed 4 
(1968- 3 ~ 0 2 1973) fIB 
A s/j j/i A s/j Yi A slj jl i 
KEY: A Adult ~ Male 
s/j Sub-adult I juvenile 11 Female 
j/i Juvenile I infant D Sex unknown 
4.15 
adult p~gs: the pigs ran and the chimpanzees chased, Rugo making the 
kill. This is the only occasion when a tool has been seen to be used 
while chimpanzees were hunting. 
The main factor affecting bushpig predation therefore appears 
to have been the chance of chimpanzees finding a group with young. 
~~~. Bushbuck 
Between 1969 and 1973 bushbuck were seen to be eaten la times, 
but the kill was only seen once. Tutin was with six males and an 
oestrous female when a young bushbuck bleated nearby. The chimpanzees 
screamed and pant-hooted (Tutin described these as food-calls) with hair 
erect, and stood looking at the prey which lay motionless. After several 
seconds Rugo ran forward from the party, grabbed the fawn by a leg and 
flailed it once: the screaming males, who were touching and embracing 
each other 5 m. away, ran forward and grabbed at the prey while Rugo 
chewed its foot. 
Two observations were only of individuals found eating the prey. 
Of the 7 other cases, there was evidence in 6 that chimpanzees had 
stolen the carcass from baboons. The relevant data are shown in 
Table 4.3. 
The evidence from Table 4.3 ~s that on 6 occasions chimpanzees 
stole prey which had been killed by baboons: two further cases ~n 1974 
(Busse & Riss in prep.) included observation of the kill by baboons and 
the subsequent piracy by chimpanzees. These observations extend the 
number of known ways in which chimpanzees obtain prey, which they were 
forme~ly thought to eat only after killing it themselves (Teleki 1973). 
Of course, chimpanzees regularly eat meat which has been killed 
by other individuals. Several studies have shown that predators may 
obtain meat by piracy (carnivores and man, Schaller & Lowther 1969; 
4.16 
Table 4.3 
Observations of chimpanzees eating bushbuck 
All observations from 1969 to 1973 are included, except two where there 
was no evidence of how the prey was caught, and one where the prey was 
observed to be killed by a chimpanzee. 
Date 
Baboon calls 
Observer heard before 
Baboons 
attend 
chimps arrive predation 
31. 3.69 Teleki 
5. 8.70 McGinnis 
7. 8.72 Pusey & 
Packer 
11. 9.72 Bauer 
21.12.72 Wrangham 
17. 6.73 Riss 
18. 7.73 Riss 
No 
Yes 
? 
Yes 
Yes 
? 
No 
-----
Briefly 
1 hour, 
and eat 
meat 
1 hour 
l~ hours 
n hours 
2 hours 
No 
Reports 
suspects 
stealing 
by chimps 
Yes 
Yes 
Yes 
Yes 
Yes 
Yes 
No 
Points of interest 
Much calling by 
baboons as the first 
chimps arr~ve. Did 
chimps see the kill? 
Baboons threaten RM, 
who is found 
flailing the prey 
Chimps seen without 
prey, charge into 
baboon troop, emerge 
with prey 
Baboons threaten 
chimps with prey for 
more than 5 min. 
Baboons stay though 
not habituated to man 
Some baboons have 
blood on fur. 
Many aggressive 
interactions between 
baboons and chimps 
with meat 
A chimpanzee got 
prey after being out 
of sight for 
30-60 secs. 
lull 
4.17 
raptors, Meinertzhagen 1959). It is therefore not surprising that 
chimpanzees should steal meat from a weaker predator. Baboons are known 
to be opportunistic predators of young antelope and other species both 
~n Gombe (van Lawick-Goodall 1967) and elsewhere (Altmann & Altmann 1970, 
Harding 1973). In the 8 good observations the meat was apparently taken 
from baboons 6 times: baboons are more likely to find the hidden fawns 
because of their higher density than chimpanzees. 
As with colobus and bushpig, the stimuli for getting meat were 
clearly availability of the prey. On 21 December 1972 I was with a 
party travelling south when there were calls of baboons to the north-
west: their volume and variability clearly indicated that the baboons 
were very aroused. The chimpanzees immediately ran in the direction of 
the calls, and when I found them again, over a kilometre away, several 
were eating parts of a bushbuck, while baboons walked around among them. 
Other observations similarly indicate that chimpanzees were attracted by 
the sight or sound of a predation by baboons: there are no observations 
of chimpanzees ignoring these stimuli. 
~v. Redtail and blue monkeys 
There ~s little information on the capture of these species . 
They are less common than colobus in Gombe, and occur in smaller troops. 
The only observed kill was by a male who rapidly gained on a juvenile 
redtail running on the ground. Predatory interest has been seen often, 
and the monkeys typically call persistently, watching the chimpanzees 
carefully: at one predation where I arrived within minutes of the kill, 
an adult male and female remained for over 90 minutes, displaying 
within 15 m. whenever the chimpanzee with the carcass moved. It seems 
likely that these species are vulnerable when caught in broken canopy, 
but they appeared to spend less time in emergent trees than colobus. 
4.18 
v. Baboons 
The factors affecting the vulnerability of baboons appeared to 
be more subtle than for other prey species. At all times adult male 
baboons are capable of being dangerous opponents (even to lion, Schaller 
1972), and as elsewhere (e.g. Sto1tz & Saayman 1970) they defended troop 
members seen to be attacked by predators. There is one observation of 
an open wound being inflicted on a hunting chimpanzee by a baboon. 
From the many (more than twenty?) troops of baboons in the study 
area, only once has a member of a troop not visiting camp been seen to be 
eaten. Two troops have learned to come to the feeding area for bananas 
(Wrangham 1974), but these have suffered at least 23 predations between 
them. The disproportionate rates of predation on baboons according to 
whether or not they visited camp strongly indicates that artificial 
feeding had an important effect, though observational biasses may be 
partly responsible for the difference. 
Teleki (1973) noted that predatory attempts and successes were 
more frequent on days when bananas were given than when they were 
withheld. He also found that the rate of predation dropped after a 
change in the artificial feeding method in July 1968: selective feeding 
was then introduced, so that for the first time observers distributed 
bananas only to chosen animals (Wrangham 1974). The low predation rate 
on baboons since 1968 has been maintained, as shown in Fig. 4.6. These 
facts confirm that the high rate of predation on baboons was closely 
linked to the manner of artificial feeding. 
From before 1967 to July 1968, baboons and chimpanzees spent an 
increasing amount of time in close proximity at the feeding area, 
interacting both peacefully and aggressively (van Lawick-Goodall 1971, 
Wrangham 1974). Teleki (1973) noted that this may have increased the 
opportunities for predation, but the only stimulus he suggested for 
Fig. 4.6 Annual number of recorded baboon predations 
All recorded predations on baboons by chimpanzees are shown, March 1968 
to August 1973. Fifteen prey were from troops which regularly visited 
the artificial feeding area, where they became habituated both to man 
and chimpanzees . 
10 
8 
Number 
6 
I of knOM1 
kills 4 
2 
o 
Art~icial feeding: 
selective method 
introduced In 
July 1968 
'68 '69 70"71 72'73 
Year 
D Habituated 
Unhabituated 
4.19 
predation was the distress call of an infant or juvenile baboon. This 
appeared to be an important orienting signal, but was not a necessary 
or sufficient condition. 
An indication of the importance of subtle contextual details 
came from an observation of the capture of an infant baboon by Jomeo and 
Sherry. On 9 June 1972 these two climbed into a F. gnaphalocarpa tree 
where two adult baboons (male Z and female M) were feeding: within arm's 
reach of M her infant lay immobile with a dislocated leg. 
En route to a feeding site Jomeo approached within 2! m. of M, 
who threatened him with a "gek" call and retracted lips. The chimpanzees 
settled and fed: twelve minutes later Jomeo stared at the infant for 
several seconds, then continued feeding. Jomeo's behaviour was unusual 
because he stayed in the open sun: this indicated that his choice of 
feeding site was affected by staying close to the infant. Over the next 
five minutes M several times left her infant, but neither chimpanzee 
made any move towards it. It was not until Sherry left his feeding site 
and came between Z and Jomeo that the latter swung down and snatched the 
infant off M. Previously any move by Jomeo could have been blocked by a 
quick reaction from Z, the adult male. By the time Z and other baboons 
had approached and attacked Jomeo, the infant was dead. 
This observation indicated that chimpanzees would only attempt a 
predation when they felt reasonably safe from retaliation by adult males. 
Analysis of the eleven best observed predations on baboons supports this 
hypothesis (Table 4.4). Although many baboons are normally ~n sight of 
each other, six predations occurred when the prey was alone or with only 
one other. Of the other five, the prey was twice lost after attacks by 
troop members. 
4.20 
Table 4.4 
Interspecific aggression at baboon kills 
Data from all good observations since 1968, including those described by 
Teleki (1973). 
Circumstances l.n 
which baboon 
captured 
Defence by other 
baboons Result 
Number of 
times seen 
Many others present 
Many others present 
Many adult males 
attack chimps 
Many adult males 
attack chimps 
Chimps lose prey 
Chimps keep prey 
Taken off adult male: The male threatened Chimps keep prey 
no others present but did not attack 
Alone apart from 
mother 
Alone 
Mother threatens (1) Chimps keep prey 
or held down by 
chimps (1) 
None Chimps keep prey 
One mechanism by which artificial feeding increased the 
predation rate is therefore clear: the fact that chimpanzees and 
2 
3 
3 
2 
1 
baboons spent much more time together than normal meant that the chance 
of a baboon being in a vulnerable position, and of a chimpanzee 
recognising it, increased accordingly. The extent to which chimpanzees 
learned to recognise or create such situations is unknown. 
It is therefore unnecessary to suggest the existence of 
"specific hungers" whose strength varied l.n time (Teleki 1973) or that 
predatory motivation varied with degrees of social stress (Kortlandt 
1972). It is certainly possible that such factors might have affected 
whether an individual chimpanzee made a predatory attempt on baboons, 
but until these hypotheses have supporting evidence, the most important 
determinant is best regarded as the vulnerability of the prey. The 
numbers of each species present, the three-dimensional structure of the 
environment, and the positions of individuals in space would all 
4.21 
contribute to determining how probably a chimpanzee could se~ze a young 
baboon and escape without injury. 
v~. Other mammals 
Species which occur in Gombe but have not been recorded as prey 
are listed in Table 4.5. 
The rarity or nocturnal habits of most of these spec~es are 
adequate explanations of their absence ~n the list of recorded prey. 
Although bushpig are mostly nocturnal, they stay above ground during the 
day, and are therefore vulnerable: porcupines, on the other hand, had 
holes in rocks, and genets probably stayed in tree holes. (On the only 
occasion I saw a porcupine during the day, a chimpanzee chased it until 
it escaped into a small cave: on another occasion Hilali saw three 
porcupines chase Jomeo up a tree, where he stayed until they went away.) 
The following notes deal with the mammals on which predation might have 
been expected from the frequency of encounter. 
Elephant shrew. Though commonly encountered in the dry season, 
I did not see chimpanzees sufficiently close to attempt a predation. 
Elephant shrews are fast and wary animals, and attempts to catch them 
would probably be spurious. 
Fruit bat. Thorndahl saw Pom find a fruit bat in a weaver nest. 
After playing with the bat for 6 minutes she left it, still alive 
though with one wing m~ss~ng. By contrast baboons have been seen to eat 
bats found in nests: perhaps they taste nasty to chimpanzees. 
Man. Van Lawick-Goodall (1968) gives evidence of attacks by 
chimpanzees on children. Children were not left alone in the Park, but 
on visits to the feeding area they have been frightened by chimpanzees, 
and observers suspected predatory interest. 
4.22 
Table 4.5 
Mammals of Gombe National Park; except prey species of chimpanzees 
English name Species Notes 
Elephant shrew Rhyncocyon eirnei Common In forest ; diurnal 
Fruit bat Eidolon helvum Nocturnal, found in nests 
of weaver birds (Ploceinae) 
Needle- nailed galago Euoticus sp. Uncommon, nocturnal 
Vervet monkey Cercopithecus aethiops Small population confined 
to parts of lake-shore. 
Rarely encountered 
Man Homo sapiens Common in study area, 
migrant population elsewhere 
Genet Genetta tigrina Nocturnal, rarely encountered 
Palm civet Nandinia binotata Nocturnal, rarely encountered 
Civet Civettictis civetta Nocturnal, rarely encountered 
Banded mongoose Mungus mungo Not seen wild since 1968. See 
text for chimpanzee inter -
actions with tame mongoose 
White-tailed mongoose Ichneumia albicauda 
Great grey mongoose Herpestes ichneumon 
Black-tipped mongoose Herpestes sanguineus 
Honey badger 
Serval 
Leopard 
Duiker 
Buffalo 
Porcupine 
Cane rat 
Giant rat 
Tree squirrel 
Ground pangolin 
Mellivora capenSlS 
Felis serval 
Panthera pardus 
Sylvicapra grimmia 
Syncerus caffer 
Hystrix galeata 
Thryonomys sp. 
Cricetomys gambianus 
Furtisciurus?sp. 
Manis temmincki 
Mostly nocturnal, rarely 
encountered 
Seen once during the day 
Occasionally seen during 
the day 
Seen once during the day 
Seen once during the day 
Nocturnal, rarely encountered 
Confined to grassland 
About 5 animals in study 
area, infrequently 
encountered 
Nocturnal, common 
Mainly nocturnal~ seen once 
Common, strictly nocturnal 
Diurnal, common 
Seen once during the day 
Small bats, rodents and shrews also occured, but few have yet been 
identified. 
4.23 
Banded mongoose. A tame male wandered freely in Gombe for 
several months in 1973: though he often encountered chimpanzees at the 
artificial feeding area he was never attacked. Adults tended to ignore 
him, while juveniles and adplescents sometimes approached and played 
with or displayed at him: they were so rough that he often escaped to 
the safety of human contact. Baboons also ignored him as prey: 
indeed, they seemed to treat him as a troop member, for three times when 
I picked the mongoose up from among a troop, females and juveniles ran 
to threaten and slap me. It is surprising that chimpanzees did not 
regard him as prey: a white-tailed mongoose has been seen to be eaten 
by chimpanzees in the Mahali Mountains (Nishida 1974). 
Cane rat. The only encounter seen occurred when Humphrey 
crossed a reed-bed: he suddenly reached out and flicked a cane rat 1 m. 
up into the air with his left hand. When it landed the rat stood on its 
hind legs threatening Humphrey, who flicked it away and walked on, 
throwing a rock ,as he did so. Humphrey tended to throw rocks when he 
was followed by an inexperienced and therefore clumsy human observer: 
it seemed to be a sign of annoyance at a disturbing companion. There 
was no indication that he regarded the rat as potential prey, though it 
is bigger than many birds which are eaten, and said to be good meat for 
human consumption. It might be difficult for a chimpanzee to kill it 
without being bitten. 
Tree squirrel. Though occasionally seen in the same trees as 
chimpanzees, squirrels were never chased. Only in unusual circumstances 
would they be caught, however. Nishida (1974) records the squirrel 
Protoxerus stangeri as prey of chimpanzees in Mahali. 
The evidence from encounters of all mammals suggests that the 
opportunity to kill is the overwhelmingly important factor in determining 
that an attempt be made. The s~x prey spec~es were all more frequently 
4.24 
encountered than any other mammal, and behavioural, ecological and 
morphological differences between them are sufficient to explain the 
differences in their rate of predation per encounter. Of the non-prey 
species only one bat, one ' mongoose and one cane rat have been seen to 
be vulnerable. It is not clear why these were not killed and eaten, but 
it is clearly possible that a learned response is involved. Perhaps 
chimpanzees show as much caution with little known animal foods as they 
do with plant foods (Chapter 3). However, mice, bushbabies (Galago 
alIeni (Hladik 1973), squirrels and mongooses (Nishida 1974) have all 
been recorded as prey in other areas, and may yet be seen to be eaten 
by Gombe chimpanzees. 
IV. Frequency of Predation on Mammals 
1. Introduction 
Calculation of the frequency of predation is of interest for 
two reasons, for its significance both to the predators and to the prey. 
There are few data available at present. Teleki (1973) found that an 
average of 10 predations a year was seen in Gombe between 1960 and 1970, 
but was unable to estimate how many had occurred unobserved. The 
frequency of predation has not been estimated in other areas. 
2. Principles of estimation of predation rate 
L. Omission of habituated baboons as prey species 
It was shown above that the frequency of predation on habituated 
baboons (i.e. those visiting the feeding area) was closely related to 
the manner of artificial feeding. Inclusion of predations on habituated 
baboons in the data used to calculate overall predation rate would 
therefore cause an artificial bias in the results. There is no reason 
to think that chimpanzees preyed on unhabituated baboons less as a 
I 
~
4.25 
result of eating one at the artificial feeding area. In the following 
calculations predations on habituated baboons have therefore been 
omitted from the sample. 
~~. Calculation method 
The rate of predation is calculated in two steps: first, the 
annual number of predations occurring in the joint range of the Northern 
and Southern communities is estimated from the frequency of attendance 
at predations by target individuals: second, the relative frequency with 
which each species was prey is taken from four independent samples in 
Gombe. 
The first part of the calculation involves a number of assumptions, 
as follows: 
(a) It is assumed that within community- sex classes the rate of 
attendance at predations did not vary between different months, times of 
day or individuals. In Appendix 4.1 it is shown that these assumptions 
were not significantly violated. Furthermore, where sufficient data are 
available the predation rate is calculated without these assumptions: 
the effect of treating months, times of day and individuals as a uniform 
group is shown to have caused the annual number of predations to be over-
estimated by 14.7% (months) and 5.6% (times of day) and underestimated 
by 7.5% (individuals) (Appendix 4.1). These are considered acceptable 
marg~ns of error. 
(b) It is assumed that the probability of an individual attending 
a predation did not affect, and was not affected by, his being a target 
individual. Observers may sometimes have lost targets who ran to join 
predations: this would mean that the calculated frequency of predations 
was underestimated. On the other hand, observers may occasionally have 
selected a target because a predation had already occurred. Care was 
taken to discard such records, but it is possible that some were 
4.26 
accidentally included. 
(c) It is assumed that predations were equally likely to occur in 
any part of the range. It was found (Fig. 4.4) that predations occurred 
in all valleys but were more likely to occur ~n forest. Observations 
were concentrated around Kakombe valley, and there was no indication 
that the frequency of predations was different between valleys. 
(d) It is assumed that the validity of the assumptions is equally 
true for all prey species. Only when more data are available can this 
assumption be tested satisfactorily: however, it did not appear to be 
violated. 
3. Estimation of predation rates 
i. Annual frequency of predations, regardless of prey species 
Individuals were assigned to one of four community- sex classes 
(Northern male or female, Southern male or female) within which the rate 
of attendance at predations could be expected to be fairly uniform: 
dependents were scored with their mothers. From August 1970 to August 
1973 the number of observation hours on each individual as target, and 
the number of predations attended by each individual when target, were 
summed within community- sex classes. From these data the mean number of 
observation hours expended between predations was calculated: by 
allowing for the number of predations not attended by each individual 
an estimate is obtained of the mean number of predations per year 
attended within each community range. The data and calculation method 
are shown in Table 4.6. 
By using separate data for males and females, two partially 
independent estimates of the number of predations occurring in each 
community range are obtained. In the Northern community range, the 
estimates are 39.2 (males) or 21.3 (females) predations per year 
4.27 
Table 4.6 
Calculation of annual number of predations (regardless of prey species) 
occurring in the study area 
Predations on habituated baboons and a chimpanzee are omitted. Data 
from all observers, August 1970 to August 1973. Dependent offspring 
are included with their mothers. SH is included with Northern males. 
Target individual 
N male N female S male S female 
A 
B 
Total predations attended 
by target individual 
Total hours observed 
(as target) 
C Mean number of observation 
hours per attendance at a 
predation (= B / A) 
D Number of predations per year 
attended by "average" target 
individual (1 day = 13 hrs) 
(= 13x365/C) 
E = Mean number of individuals In 
each community- sex class 
attending predations when at 
least one individual in that 
class attended. Data from 
completely ~observed predations 
only. Sample size (number of 
predations) in brackets. 
F 
G 
H 
I 
Number of individuals in each 
community- sex class (average 
from August 1970 to August 1973) 
Estimated number of predations 
per year attended by at least 
one individual in that 
community- sex class (= DxF/E) 
Proportion of completely 
observed predations attended 
only by the other sex class 
within each community (sample 
size is denominator) 
Estimated number of predations 
per year attended by at least 
one member of either sex within 
each community (= G / (l- H) ) 
22 
3628 
164.9 
28.8 
6.6 
(15) 
9 
39.2 
0/15 
39.2 
7 
5547 
792.4 
6.0 
3.9 
( 13) 
12 
18.5 
2/15 
21. 3 
4 
1012 
253.0 
18.8 
4.6 
(5) 
7 
28.6 
0/5 
28.6 
1 
210 
210.0 
22.6 
2.3 
(4) 
3 
29.5 
1/5 
36.9 
4.28 
(Table 4.6), yielding a mean estimate of 30.3 predations per year. 
The corresponding figures in the Southern community range are 28.6 
(from males) and 36.9 (from females), giving a mean estimate of 32.8 
predations per year. 
If Northern and Southern individuals never met at predations, 
the Northern and Southern estimates could be added to give an estimate 
of the number of predations occurring annually in the joint ranges of 
the two communities. The situation is complicated by the changing 
relationship of the communities, which became gradually separated and 
finally split in 1971 (Bygott 1974): June 1971 was the last month when 
parties from both communities were seen together at predations. Since 
then, the only contacts were a brief visit by the old Southern male, 
Go1iath, to a Northern predation, and four predations where adolescent 
and one adult female were with individuals of the other community. 
Despite the partial overlapping of Northern and Southern 
communities, th~ir independent estimates of annual number of predations 
are summed here to give the estimate for the whole study area. The 
error incurred by failing to calculate a correction factor is probably 
slight for two reasons. First, only four out of 34 predations attended 
by target individuals included Northern and Southern parties, and there 
was only one target individual at each of these. Second, even before 
the communities had stopped forming joint parties it was clear that 
individuals ranged in different areas (Bygott 1974): they were most 
likely to associate near what is now the boundary area of the community 
ranges (Chapter 5). Even ~n 1970, therefore, the communities were 
partially independent. 
The mean estimate for the joint range of the two communities ~s 
therefore 30.3 + 32.8 = 63.1 predations occurring per year. Maximum and 
minimum estimates are obtained by combination of estimates from each 
4.29 
community (Table 4.6): the maximum LS 39.2 + 36.9 = 76.1, and the 
minimum is 21.3 + 28.6 = 49.9 predations per year. Use of these three 
estimates of predation frequency allows the results to be treated with 
reasonable confidence. 
ii. Relative frequency of each speCLes as prey 
All observations of meat-eating Ln Gombe are used to give an 
overall estimate of the relative frequency of each species as prey 
(Table 4.7). 
Table 4.7 
Relative frequency of different species as prey 
Data are from the sources given. At some of the colobus and bushpig 
predations more than one animal was killed, but these are counted as 
single records of predation. N = number of records of predation. 
Only baboons not habituated to the feeding area are included. 
Colobus Bushpig 
Period Source N % N % 
Bush-
buck 
N % 
Red-
tail 
N % 
Blue 
N % 
Baboon 
N % 
All 
species 
N % 
1960-1967 van Lawick- 10 38.5 7 26.9 7 26.9 1 3.8 1 3.8 0 0.0 26 99.9 
Goodall 
1968-1970 Teleki 1973 4 40.0 3 30.0 2 20.0 o 0.0 o 0.0 1 10.0 10 100.0 
Aug 1970- Attended 
Aug 1973 by target 16 57.1 6 21.4 4 14.3 1 3.6 1 3.6 0 0.0 28 100.0 
individuals 
Aug 1970- Not attended 
Aug 1973 by target 6 33.3 5 27.8 5 27.8 1 5.6 1 5.6 0 0.0 18 100.1 
individuals 
1960-1973 All Gombe 36 43.9 21 25.6 18 21.9 3 3.7 3 3.7 1 1.2 82 100.0 data 
Maximum 57.1 30.0 27.8 5.6 5.6 10.0 
estimates 
Minimum 33.3 21.4 14.3 0.0 0.0 0.0 
estimates 
4.30 
The four samples in Table 4.7 rank the prey specLes Ln a 
significantly similar way (Kendall coefficient of concordance W 0.95, 
s = 248, P < 0.01). Differences between samples are explicable by 
differences in data collection. For instance, the relative frequency 
of colobus was high during observations of target individuals. This 
sample gives an estimate of the frequency of attendance at predations on 
different species, rather than of the frequency of predations themselves. 
A larger prey takes longer to eat, and individuals are more likely to 
arrive before it is finished: several adult colobus were eaten, and the 
amount of meat available may therefore explain the differences between 
the samples. Given the small sample sizes, the observed differences are 
unimportant. 
Table 4.7 shows the maXLmum and minimum estimates of the relative 
frequency of each species as prey, and an overall estimate from 
combination of the four samples. These figures are used in the 
calculation of the annual predation rate on each species. 
iii. Annual predation rate on each species 
Combination of the estimates of the annual number of predations 
Ln the study area with the relative frequency of each species as prey 
allows calculation of the estimated number of prey animals eaten each 
year. Data from Tables 4.6 and 4.7 are used in the calculation 
(Table 4.8). The only additional data in Table 4.8 are the numbers of 
animals killed at each predation: between August 1968 and September 1973, 
24 colobus were eaten in 20 well observed predations (mean 1.2 prey per 
predation) and 15 bushpig in 11 well observed predations (mean 1.4 per 
predation). 
Table 4.8 shows the results. The mean estimate was of 34 
colobus eaten per year, 23 bushpig, 14 bushbuck and small numbers of 
redtail and blue monkeys and baboons. 
4.31 
Table 4.8 
Calculation of numbers of each prey species eaten annually 
Estimates of the annual number of predations and the relative frequency 
of prey species ~n the diet are taken from Table 4.6 and 4.7 respectively. 
Number of %, of total Mean number Number of 
Species Estimate predations predations, on of prey eaten prey eaten 
per year this species per predation annually 
Maximum 76.1 57.1 1.2 52.1 
Colobus Mean 63.1 43.9 " 33.2 
Minimum 49.9 33.3 " 19.9 
Maximum 76.1 30.0 1.4 31.9 
Bushpig Mean 63.1 25.6 " 22.6 
Minimum 49.9 21.4 " 14.9 
Maximum 76.1 27.8 1.0 21.2 
Bushbuck Mean 63.1 21.9 " 13.8 
Minimum 49.9 14.3 " 7.1 
Maximum 76.1 5.6 1.0 4.3 
Redtail Mean 63.1 3.7 " 2.3 
Minimum 49.9 0.0 " 0.0 
Maximum 76.1 5.6 1.0 4.3 
Blue Mean 63.1 3.7 " 2.3 
Minimum 49.9 0.0 " 0.0 
Maximum 76.1 10.2 1.0 7.8 
Baboon Mean 63.1 
~ 
1.2 " 0.8 
Minimum 49.9 0.0 " 0.0 
Method of calculation A B C AxBxC 100 
Variation in the estimated number of predations per year was not 
closely related to the frequency of encounter between chimpanzees and 
the prey species (Fig. 4.7): this is expected in view of the differences 
in vulnerability (above) and group sizes. From the point of view of prey 
individuals, however, a more important aspect of predation is its 
frequency of occurrence in relation to the population size of the prey 
species. 
The only formal estimates of population s~zes in Gombe are given 
by C1utton- Brock (1972), for primates, and these were not based on formal 
censuses. The study area covered about forty per cent of Gombe National 
Fig. 4.7 Frequency of predations in relation to time spent 
with ·prey ·species 
Estimates of the number of prey eaten per year are tru(en from 
Table 4.8. "In contact with" means prey were seen or heard within 
200 m. Data on time spent with prey species were from 11,942 5-min 
points during observation of target males. Habituated baboons are 
not included. 
Data are from Northern and Southern males combined. There was no 
apparent difference between communities in the amount of time spent 
with prey species. 
60 
50 
Number of 
prey eaten 
30 TBusrpg 
20 T I per year 
10 l,BushbuCk 
Colobus 
lO ,.5 
% 5 - min. points "in contact with" prey species 
4.32 
Park, for which Clutton-Brock's estimates were made, so the latter 
were reduced proportionately. Ihese figures were treated as minimum 
estimates, and maxima were obtained by doubling. The resulting maximum 
and minimum estimates are shown in Fig. 4.7. 
Bushbuck and bushpig numbers have not been counted. In well 
watered habitat in Chobe, Botswana, densities as high as 20 bushbuck per 
sq. km. have been recorded (C.D. Simpson, pers. comm.): other estimates 
available are 1.6/sq. km. (N. Morris, pers. comm.) and la/sq. km. 
(R. Taylor, pers. comm.) both in Southern Africa. Bushbuck density Ln 
Gombe was guessed at between 3 and 6 animals per sq. km., partly on the 
basis of the observed range of a fairly tame adult, and partly by 
comparison with the quoted figures: bushbuck were certainly much less 
numerous than in Chobe (personal observation). 
Bushpig numbers were estimated by comparison with baboons: 
although difficult to count because of their nocturnal habits and 
variable group sizes, bushpig were considered less abundant than baboons. 
It seems likely that if the estimate of between 400 and 800 animals Ln 
the study area LS wrong, it is too large rather than too small. 
EStimates of population density have been related to estimates 
of the number of each prey species eaten per year (Fig. 4.8). On the 
graph the maximum and minimum estimates of each parameter combine to 
form a rectangle for each species: the rectangles are areas of 
uncertainty within which it may be concluded that the real predation 
rate (per cent of population eaten annually by chimpanzees) should fall. 
The small population of blue monkeys means that there is a wide range 
of estimates of annual mortality due to chimpanzees, but for the other 
species the order of magnitude is clearly indicated. The best estimate 
is probably for colobus monkeys, which are the only prey species to have 
been studied in detail (Clutton-Brock 1972) and for which the biggest 
4.32 
Park, for which Clutton-Brock's estimates were made, so the latter 
were reduced proportionately. These figures were treated as minimum 
estimates, and maxima were obtained by doubling. The resulting maximum 
and minimum estimates are shown in Fig. 4.7. 
Bushbuck and bushpig numbers have not been counted. In well 
watered habitat in Chobe, Botswana, densities as high as 20 bushbuck per 
sq. km. have been recorded (C.D. Simpson, pers. comm.): other estimates 
available are 1.6/sq. km. (N. Morris, pers. comm.) and la/sq. km. 
(R . Taylor, pers. comm.) both in Southern Africa. Bushbuck density ~n 
Go'mbe was guessed at between 3 and 6 animals per sq. km., partly on the 
basis of the observed range of a fairly tame adult, and partly by 
comparison with the quoted figures: bushbuck were certainly much less 
numerous than in Chobe (personal observation). 
Bushpig numbers were estimated by comparison with baboons: 
although difficult to count because of their nocturnal habits and 
variable group sizes, bushpig were considered less abundant than baboons. 
It seems likely that if the estimate of between 400 and 800 animals ~n 
the study area ~s wrong, it is too large rather than too small. 
Estimates of population density have been related to estimates 
of the number of each prey species eaten per year (Fig. 4.8). On the 
graph the maximum and minimum estimates of each parameter combine to 
form a rectangle for each species: the rectangles are areas of 
uncertainty within which it may be concluded that the real predation 
rate (per cent of population eaten annually by chimpanzees) should fall. 
The small population of blue monkeys means that there is a wide range 
of estimates of annual mortality due to chimpanzees, but for the other 
species the order of magnitude is clearly indicated. The best estimate 
is probably for colobus monkeys, which are the only prey species to have 
been studied in detail (Clutton-Brock 1972) and for which the biggest 
Fig. 4.8 Frequency of predations in relation to numbers of prey species 
Estimates of the number of prey eaten per year are taken from Table 4.8. 
Estimated minimum prey densities are half the estimated maxima (numbers 
shown are for joint ranges of Northern and Southern communities). The 
rectangle for each species lS an area of uncertainty within \vhich the 
annual per cent of each 'population eaten by chimpanzees is expected to 
fall. 
60 
50 
Number of 
40 
prey eaten 
30 
per year 
5% 
Bushpig 2Hlo 
1% 
Baboon 
600 700 800 900 lCXD 
Population size 
KEY: % popn. eaten per year: 1"10,2'5%,5"10,10%, 15 % lines are sho 
Area bounded by max. and min. estimates 
- - - - - Mean estimate 
4.33 
sample of predations was available. Different prey spec~es suffered 
different rates of predation, and the factors contributing to this have 
been discussed earlier: in relation to population size, the fact that 
some species were not eaten ·when' adult is also important. Although 
bushbuck suffered the highest rates of predation ~n relation to numbers, 
it should be remembered that many of those eaten by chimpanzees \..;rere 
first killed by baboons (in perhaps 6 out of 8 observations, = 75%, see 
Table 4.3). 
There are few comparative data available for primates. Some 
examples of predation rates for mammals are given in Table 4.9. 
Table 4.9 
Examples of predation rates on large mammals 
Specific names not mentioned elsewhere: zebra Equus burchelli, 
wi ldebeest Connochaetes taurinus, hyo..en~ CrOCLAUt CKoW,itt. 
Site Pr~dator Prey % Population Source killed/yr 
Chobe, Baboon Bushbuck 4 C.D. Simpson Botswana (pers • comm. ) 
Ngorongoro, Hyaena Wildebeest 11 Kruuk 1972 Tanzania (adults only) 
Ngorongoro, Hyaena Zebra 9 Kruuk 1972 Tanzania 
Serengeti, Hyaena Wildebeest 1.6-2.6 Kruuk 1972 Tanzania (adults only) 
Serengeti, Hyaena Zebra 0.8-1.0 Kruuk 1972 Tanzania 
Serengeti, Lion Wildebeest 2.2-3.3 Schaller 1972 Tanzania 
Serengeti, Lion Zebra 5.9-7.2 Schaller 1972 Tanzania 
Comparison of Fig. 4.8 and Table 4.9 shows that the annual 
mortality of colobus and bushpig due to chimpanzees, and of bushbuck 
due to chimpanzees and baboons, are similar to those found in other 
4.34 
mammalian predator systems. The significance of this result for the 
regulation of the prey populations cannot be determined without long-
term studies of population numbers. 
~v. A~ount of meat eaten annually 
In the absence of proper data on the weights of prey animals 
estimation of total meat intake must be highly speculative. It ~s 
nevertheless of interest to obtain an order of magnitude. This has been 
done using the mean estimates for the number of prey animals eaten 
annually: the result is shown in Table 4.10. 
Tab le 4.10 
Estimation of total weight of meat eaten annually in study area 
Estimates of the number of each species eaten are from Table 4.8. 
Mean weights of prey animals were guessed. 
Species Number eaten Mean Total weight per year weight (kg.) eaten per year (kg. ) 
Colobus 33 S l6S 
Bushpig 23 3 69 
Bushbuck 14 S 70 
Redtail 2 3 6 
Blue monkey 2 4 8 
Baboon 1 3 3 
Total 7S 321 
The estimate calculated in Table 4.10 is that 7S prey animals 
supply over 300 kg. of meat a year. In the study area there were an 
average of 31 adult chimpanzees for the period (August 1970 - August 
1973) from which the data were taken, but individuals varied considerably 
in how much meat they ate (below). Since an even distribution of meat 
would have provided each individual (or family) with about 10 kg. of 
meat ~n a year, it is probabl~ that some individuals actually got much 
more than this. Thus though meat formed a comparatively small part of 
4.35 
the diet, it was not insignificant. The only other primates which 
regularly eat mammals are baboons and man. Harding (1973) found that a 
troop of baboons ate 47 prey in 1000 observation hours (about 200 
animals a year), which is much the highest rate of predation reported 
for non-human primates. Lee (1968) found that !Kung Bushmen ate about 
100 kg. of meat a year, ten times the average for Gombe chimpanzees. 
V. Predatory Behaviour 
1. Introduction 
Descriptions of meat-eating behaviour Ln Gombe have been given 
by van Lawick-Goodall (1968, 1971) and Teleki (1973). After prey has 
been killed, chimpanzees tend to aggregate around individuals with 
meat; they obtain portions themselves by finding scraps on the ground 
or by taking meat from others, who occasionally hand it over when 
another gives begging gestures. Similar behaviour has been seen in 
other areas (Suzuki 1971, Hladik 1973). 
Teleki (1973) found that "competition occurred very rarely over 
meat" ("competition" appeared to be used to mean "aggression"). This 
is surprising since it means that either chimpanzees use time and energy 
and incur danger to obtain something (meat) which is of little value, or 
there is a mechanism which stops individuals competing for a valuable 
item even when it is in short supply. This problem has been side-stepped 
by suggestions that the significance of predation lies not in obtaining 
meat, but (a) in the opportunity for individuals to increase social 
status by "gaining a temporary control role" (Teleki 1973); (b) in re-
direction of aggression between chimpanzees on to individuals of another 
species (Kortlandt 1972); and (c) in controlling the numbers of species 
which compete with chimpanzees for food (Kortlandt 1972). These 
4.36 
suggestions imply that meat is nutritionally unimportant: the hypothesis 
is apparently supported by the fact that predation by chimpanzees occurs 
very rarely or not at all in some environments (Kortlandt 1972), and 
Teleki's finding that individuals "compete" little for meat. 
In order to exam~ne these hypotheses selected aspects of meat-
eating behaviour are described. Unusual behaviour which has not 
previously been reported ~s also documented. 
2. Behaviour before eating 
1. Killing prey 
Methods of killing prey have been described by Teleki (1973): 
animals were bitten on the neck, torn apart, or hit (flailed) on the 
ground or a branch. Additionally prey were sometimes eaten until they 
died, perhaps from loss of blood: this occurred on 7 out of 16 
occasions between March 1968 and August 1973 when the kill method was 
seen. Killing was often inefficiently performed, taking several (up 
to 20) minutes. Baboons and hyaenas also kill by eating their prey 
(Harding 1973, Kruuk 1972). 
The most difficult prey to kill were evidently adult monkeys. 
One colobus became limp after being throttled by Humphrey, one had its 
guts and legs torn off while held on the ground, and the capture of a 
third was seen in detail by Kakuru. Mike and Sherry caught a lone male 
colobus in a tree, and threw it off the branch: it landed on a lower 
one, and the process was repeated. On the third time of catching it 
Mike threw the monkey to the ground, which it reached after bouncing 
off a buttress: by this time it was hardly capable of mov~ng. Mike 
led Sherry rapidly down to the prey, and between them they broke all 
its legs, the prey screaming all the time "in a low voice". The monkey 
finally died after a head-bite from Mike, who then dragged and carried 
4.37 
the carcass up the tree to eat his prey. 
11. Behaviour immediately after a kill 
Any individual seen to make a kill was immediately surrounded 
by others: the killer tended to climb to a relatively inaccessible 
branch, apparently so as to avoid the attentions of others so far as 
possible. The suggestion that individuals with meat tried to avoid 
others is supported by other observations. When Satan caught and killed 
two bushpig on his own, he made no calls and ate his prey alone (except 
for the company of baboons, eating fallen scraps) (Collins, pers. comm.); 
similarly when Jomeo and Sherry killed a baboon on their own near camp 
(where other chimpanzees were likely to hear calls of the attacking 
baboons) they ran at top speed to the next valley where I lost them, 
and there were no calls for over two hours . An example was also given 
(2/IV/3/ii) of individuals aggregating around meat when the killer 
appeared to have met, by accident, another chimpanzee, who then 
"announced" the predation. 
Individuals who obtained meat from another also tended to climb 
to a high branch with their possession, especially when the carcass was 
first divided (see also van Lawick-Goodall 1971). The importance of 
trees was indicated when Rugo, with meat, was being attacked by 
Rumphrey, without. Rugo ran along looking up at the trees he passed, 
but climbed none of them: vines in the canopy prevented any branch from 
being isolated. Eventually Rugo stayed on the ground with his back 
agains tat runk • 
3. Meat-eating behaviour 
1. Evidence that meat 1S a preferred food 
Because predations tended to be noisy, rapidly attracting 
aggregations, there are few data on meat-eating 1n the absence of social 
I' 
excitement. However, three observations clearly indicate that 
individuals may eat meat 1n such a situation. 
First, Collins saw a lone male (Satan) kill two bushpigs and 
continue eating them until the observation ended. 
Second, Bauer and Halperin saw Mike examine his faeces, pick 
out some bones, and chew them. 
4.38 
Third, Halperin and Meic saw two chimpanzees arrive at the 
feeding area where they dropped the partially eaten head of a monkey. 
One of them played with it idly before leaving it. Later Humphrey came 
into camp, found the head and ate it. 
11. Overt aggression over meat 
To obtain a standard measure of the frequency of aggress1on, 
behaviour 1n the first half-hour of well observed predations was 
analysed. Attacks were either protracted interactions involving 
components such as charging, grappling, biting, hitting, gripping and 
slamming, or brief physical contact together with a chase. 
There were 19 good accounts of predations with more than one 
chimpanzee present from August 1970 to August 1973. The frequency of 
aggress1ve interactions is shown in Table 4.11. 
Out of 20 attacks on a possessor the aggressor obtained meat 
only three times (Table 4.11). One of these was the only time a female 
was attacked by a male (all the other attacks were by and on males): 
the second ended in grappling between Mike and Hugo, and the latter 
eventually succeeded in obtaining the prey. The third, on 25 September 
1972, was a long series of aggressive incidents between Mike and Hugo 
which illustrate well the strength of interest 1n meat: excerpts from 
my field notes are given. (Attacks between Mike and Hugo accounted for 
4.39 
Tab le 4.11 
Frequency of aggressive interactions at predations 
Predations included are all (19) from August 1970 to August 1973 in which 
the first half-hour following the kill (or seizure of prey from baboons) 
was well observed. A predation where only one individual was present is 
omi tted. 
Behaviour Number of Number of predations times seen at which behaviour seen 
Attack on possessor: 3 3 
aggressor gets meat 
Attack on possessor: 
aggressor fails to get meat 17 11 
Attack on non-possessor 21 9 
s~x of the twenty involving a possessor (Table 4.11), and all but two 
involved one of these pair.) 
"1116: I arrive a minute after a bushpig has been caught. 
Rurnphrey, Faben and Gigi have bits of meat, Mike has the body. Screams. 
~ 
Rugo jumps onto Mike, hitting MLke with his right hand, Mike holds the 
p~g ~n his right hand, trying to eat it. 
"1117: Faben thrusts on Mike who holds p~g and eats from it. 
Faben holds Mike by the neck, Rugo next to Faben holding his scrotum. 
Screams continue. 
"1118: Rugo chases female, who screams. Returns to Mike, hangs 
~n front of him kicking him, Mike no response. 
"1120: Rugo still not got any meat .•• most others have had 
scraps. Swings in front of Mike, puts his hands over meat, Mike guards 
it by crouching over it. Rugo begs with extended hand, Mike ignores. 
"1122: Faben comes and pulls at the meat, harasses and pulls 
Mike, maybe gets a little. This exposes Mike to Rugo, who gets both 
hands on the meat, pulls, it's a tug-of-war. Mike holds a branch with 
one hand, meat with other, gnaws at the meat and maybe at Rugo's fingers 
1 1 
(to make him let go?). Ruga lets go with one hand, Mike now leaning 
right over the prey, all protected by arms and legs. 
4.40 
"1123: Mike runs down tree, Ruga displays at him, Mike lies, 
clutches meat under his body ; Ruga and Faben display on top of him, 
embrace while squatting on his back, Faben with mouth on Ruga's scrotum. 
"1124: Ruga gets a hand on the meat, Mike pushes Ruga's hand 
away. Ruga disp lays aga~n, Mike rushes away, runs bipeda11y holding pig 
in both hands. Ruga catches up, falls on Mike, bites Mike's scrotum, 
Faben approaches and thrusts his bottom against Ruga's. 
"1127: Ruga charges at Mike, pulls him against a tree with both 
hands, Mike screams, extends his hand to Faben, holds Faben's scrotum, 
Faben thrusts against Mike's head, holds his chin, Ruga quietens and sits 
behind Mike. 
"1129: Ruga approaches as if to attack Mike, Mike avoids round 
the tree, Ruga still approaches and climbs on top of Mike, then sits by 
him and begs with, extended hand. 
"1130: Ruga and Mike fight, Faben jumps on top of them, Mike 
screaming with glottal crambs, Ruga grappling to get at meat, Faben and 
Figan dance round them. 
"1131: Mike eludes Ruga, runs 2 m., Ruga rugger-tack1es him 
from behind, they grapple again. Ruga thrusts against Mike's head, begs 
by getting lips and hands on meat and pUlling. 
"1132: Ruga becomes unrestrained, pulls hard, it becomes a fight. 
Mike pushes Ruga over, slips round a tree and thus avoids Ruga. Ruga 
re turns to beg. 
"1133: Ruga has both hands on hind part of prey, carcass still 
held by Mike. Ruga eats for the first time. 
"1134: Ruga and Mike have a tug-of-war again, then Ruga attacks 
Mike, pulls him downs1ope 4 m., Mike runs back ups1ope. Ruga joins him 
and begs again, not pulling hard. Gets a piece of skin, big enough to 
4.41 
eat for several minutes." 
The ma~n interactions included in this edited account were 
between Mike (holding the prey) and Hugo, and they show how determined 
each individual was to keep or get at the meat. The striking feature at 
this predation was that a number of individuals who did not attack Mike 
(at least two males and two females) succeeded in getting meat either by 
gently taking meat while Mike was eating it, or by picking up scraps. 
Similarly on 22 July 1972, when Hugo held the prey, almost the only 
in~vidual present not to get any meat was Humphrey, who was also the 
only one to attack Hugo. 
It was clear from these observations that if an individual wanted 
to keep his meat he was physically capable of doing so, provided he saw 
the attack coming. By lying face down he could huddle over the prey and 
thus completely hide it. Even if others jumped on top of him or kicked 
and hit his back, he was fairly safe from harm: the soft organs are 
, 
protected by a combination of ribs and ischium. 
However, the disadvantage of protecting the meat ~s that it was 
then impossible to eat it. As soon as the possessor sat holding the meat 
by his mouth others could reach and pull at it. Only the intestines were 
normally easily detachable, and in most predations these were eaten first 
(as found by Teleki 1973), either by the individual holding the prey or 
by those clustering around him. 
The fact that an individual allows others to share his meat may 
therefore be partly explained by the impossibility of effectively 
protecting the prey while eating it. 
Individuals sometimes lost their prey without being attacked. 
Immediately after killing a colobus Jomeo was approached by Gigi 
followed by four males. Gigi took the prey from Jomeo, and then lost 
it to the males. Many similar observations have been made (see also 
van Lawick- Gooda11 1971, Teleki 1973) (sometimes the prey was pulled 
4.42 
from the hands of possessor, and at other times it was snatched at a run). 
Individuals who made little or no attempt to defend their prey were low-
ranking but evidently wanted meat, for their subsequent begging patterns 
were typical of others. It seems likely that some low-ranking 
individuals did not have the confidence to withstand requests for meat 
which could turn into attacks. 
These observations suggest that an individual trying to get meat 
has a choice of two strategies. He can attempt to take all there is, in 
which case he may get it if the possessor lacks confidence, but may be 
thwarted if the other responds by defence. Alternatively he can try to 
take pieces from the possessor, by a variety of begging gestures 
(described by van Lawick-Goodall 1968 and Teleki 1973), in which case he 
~s likely to be successful more often but to get less meat. The choice 
of strategy app~ars to vary with the two individuals involved and the 
amount of meat available: these factors will be examined below. 
~~~. Individual success in getting meat 
It was earlier shown that there was no evidence that adult 
males varied in the frequency of killing prey. However, there was 
considerable variation in the amount of meat obtained by individuals 
attending the same predations. At well observed predations individuals 
were scored as obtaining no meat, some meat, or a major portion (not 
less than a third of the carcass): individuals were compared in pairs 
for their relative success in obtaining meat according to these three 
criteria (Table 4.12). 
4.43 
Tab le 4.12 
Relative success of Northern males at obtaining meat 
From 30 well observed predations between August 1970 and January 1974 
W-L 
the coefficient (~x 100) w~s calculated for each pair of individuals 
(A and B): N 
predation. W 
422 occasions when two individuals attended the same 
number of predations at which A got more meat than B, 
L = number of predations at which A got less meat than B, T = total 
predations (attended by both A and B). Cells show values of (W-L x 100) . 
T 
RG 
RG 
MK 13 
FG 47 
Individual 8 FB 69 
HM 64 
EV 75 
JJ 100 
ST 91 
MK FG 
69 
47 5 
50 
41 
78 
87 
19 
38 
33 
50 
Individual A 
FB 
14 
29 
40 
50 
RM EV 
33 
54 
70 
31 
40 
JJ ST 
o 
A linear hierarchy was found for success at getting meat 
(Table 4.12). Not only did each individual get less meat than those 
above him in the hierarchy and more than those below, but each individual 
ranked the others in a significantly similar order (Kendall coefficient 
of concordance W = 0.76, s = 1354, P < 0.01). The order of individuals 
in success at getting meat was significantly correlated with age (from 
Table 4.13, Spearrnan r = 0.83, N = 8, P < 0.01), but was less closely 
s 
related to dominance status: Rugo and Mike were both low-ranking 
throughout most of the period sampled (Bygott 1974). Statistical 
correlation with dominance status is impossible because of the difficulty 
of assigning ranks of dominance to individuals whose status changed 
during the period. 
I 
I 
4.44 
Table 4.13 
Relation of age to success at getting meat 
Relative ages are taken from estimates in Bygott (1974) (Rugo oldest). 
The rank order of success a~ getting meat is from Table 4.12. 
Individual Rank of Rank of success 
age at getting meat 
RG 1 1 
MK 2 2 
RM 3 5 
FB 4 4 
EV 5 6 
FG 6 3 
JJ 7 7.5 
ST 8 7.5 
Fewer data were available for females than for males. Although 
some Northern females could not be paired for relative success at 
predations, Table 4.14 shows that there was an approximately linear 
hierarchy when all independent Northern individuals were compared. Only 
two males, the brothers Jomeo and Sherry, were less successful at 
getting meat than females. 
The data on females are too few to allow confident correlations 
to be made with other factors. Inspection of Table 4.14 indicates that 
relative age was less important than for males (e.g. Sparrow (sw) was 
one of the two youngest females, the oldest being Passion (ps), Nope (np) 
and Melissa (ml». As suggested by Te1eki (1973), kin relations may be 
important: Fifi (ff) was the sister of Figan and Faben, and the most 
successful female, but Gilka (gk), Evered's sister, obtained little meat. 
It is also likely that oestrous state was important (Teleki 1973), but 
s~nce most of the females were oestrous for some time during the period, 
the data in Table 4.14 cannot resolve this. 
Tab le 4.14 
Relative success at getting meat: all independent males and females 
of the Northern community 
4.45 
The matrix shows relative success of two chimps when attending the same 
predation (data from 30 predations, November 1970 to January 1974). 
Each chimp was scored as having no meat, some meat or a major portion, 
and a "win" by one chimp over another was recorded when he clearly had 
more meat, according to the three criteria. For each dyad, the matrix 
shows: 
HC 
MK 
FC 
FB 
HM 
EV 
ST 
ff 
ps 
gg 
sw 
at 
wk 
np 
pl 
ml 
SH 
gk 
JJ 
mf 
+ Column chimp had more wins than row chimp 
Column chimp and row chimp had equal number of w~ns 
HC MK FC FB HM EV ST ff ps gg swat wk np pl ml SH gk JJ mf 
+ 
+ + 
+ + + 
+ + + + 
+ + + + + 
+ + + + + + 
+ + + + + 
+ + + + + + + 
+ + + + + + 
+ + + + + + + 
+ + + + + + 
+ + + + + + + 
+ 
+ + + + + + + 
+ + + + + + 
+ + + 
+ + + + + + + 
+ + + + + + + + + + 
+ + + + + + + + + 
+ = 
+ 
+ 
+ = + + 
+ + 
+ + + + + + 
+ + + + + + 
+ + + + + + 
+ + + 
+ 
+ + 
+ 
+ + 
HC MK FC FB HM EV ST ff ps gg swat wk np pl ml SH gk JJ mf 
HC 
MK 
FC 
FB 
HM 
EV 
ST 
ff 
ps 
gg 
sw 
at 
wk 
np 
pl 
ml 
SH 
gk 
JJ 
mf 
4.46 
~v. Methods of eating meat 
The methods of eating meat have been described by van Lawick-
Goodall (1968) and Teleki (1973). Meat was often eaten together with 
leaves, which are also used for other wet foods such as bananas and 
Strychnos spp. fruit. (In Gabon, Hladik (1973) found that chimpanzees 
preferred to chew meat with bark.) Generally the whole carcass is 
completely consumed, and even drops of blood are licked from leaves. 
However, meat may be left uneaten. The adolescent male Goblin was seen 
to kill an adult female colobus and then abandon the body: other 
chimpanzees also ignored it, though some were eating her infant which 
had been killed shortly earlier (Busse & Riss, in prep.). 
Teleki (1973) and van Lawick-Goodall (1971) considered that 
brain tissue was a preferred part, less likely to be shared than the 
rest of the prey. Observations since 1970 do not support this hypothesis. 
The brain was several times one of the last parts to be eaten, and was 
seen to be shared in the typical manner. 
Predations are time-consuming both because individuals spend time 
trying to get meat and also because meat ~s eaten slowly (van Lawick-
Goodall197l). For instance, when a juvenile colobus, weighing about 
4 kg., was shared between nine chimpanzees on 27 July 1972, individuals 
were scored as eating meat on 115 5-min. points, and I estimated that a 
further 24 5-min.-meat-eating points were missed when individuals were 
out of sight. This means that the carcass took more than ten chimpanzee-
hours to consume. By contrast, Harding (1973) found that baboons took 
between 15 minutes and one hour to eat infant antelope or adult hares. 
The difference in the size of prey does not account for this discrepancy, 
which indicates clearly that chimpanzees eat meat more slow"ly than 
baboons. Baboons have stronger jaws than chimpanzees: only the former 
can crack palm nuts open. 
4.47 
v. Relation to baboons 
Evidence was presented earlier (Table 4.3) that chimpanzees 
sometimes stole meat from baboons, who subsequently tended to stay near 
the meat. Baboons were twice seen to eat scraps of bushbuck meat dropped 
by chimpanzees, and when Satan ate a bushpig without other chimpanzees 
present, baboons approached and watched Satan from less than a metre away: 
an adult male ate quite a large piece which fell to the ground (Collins, 
pers. comm.). At three predations on colobus, baboons approached the 
predators and appeared to be interested in finding meat. Such behaviour 
was never seen when a baboon was prey. 
VI. Discussion 
1. The significance of predation to chimpanzees 
In discussing predation by chimpanzees, Kortlandt (1972) and 
Teleki (1973) were reluctant to conclude that its functional significance 
was nutritional. Problems to regarding it as such are stated below, and 
solutions are offered in the light of new data. 
Ca) Plant food appears to be abundant, and animal foods are 
therefore unnecessary (Teleki 1973). There are two 1ssues here. First, 
it was shown (Chapter 3) that plant foods may not in fact be sufficient 
at all times. Second, qualitative differences between plants and animals 
may mean that animal foods are an important, even if infrequent, part of 
the diet. Furthermore, even if chimpanzees could subsist without 
predation, they may be able to live at a higher popUlation density if 
animal foods are eaten. 
(b) Other animals appear to provide more easily obtainable food 
than medium-sized mammals (Kortlandt 1972). There is little evidence for 
this in Gombe. Invertebrates which were not eaten but available were 
L 
1III 
i; I 
I ~, I' 
'I 
! 
,! 
11', 
.'i 
'Ill 
I I 
I 
I 
1 
! 
:1 
i" ,I 
! I 
r :: I 
I' 1 
I' I 
1,1 
I 
III 
I 
1 
'liiLtjl ~ 
4.48 
probably unpalatable. Few encounters have been seen between chimpanzees 
and vertebrates other than their prey, and in only three cases (bat, 
cane rat and mongoose) was it clear that chimpanzees could have eaten 
mammals. The rarity of encounter may have meant that these were treated 
as novel items. However, it is not yet clear why males should eat 
insects less than females do: this 1S an important problem. 
(c) If mammals are an important source of food, then females 
should eat as much meat as males (Kortlandt 1972). Since it 1S now known 
that females eat insects more frequently than males (McGrew, 1n press, a), 
the sexes may differ merely in which animals they eat most, rather than 
in how much animal food they consume. 
(d) Meat is eaten in a leisurely manner despite the high energy 
expenditure in obtaining it, and its low calorific value. The term 
"leisurely" (Teleki 1973) is patently inadequate for much of the 
behaviour at predations (IV/2): there was evidence of intense competition 
, 
for meat. The fact that a dead prey was once not eaten has yet to be 
explained, but it should be noted that carnivores which rely solely on 
meat for food sometimes kill more meat than they eat, and may abandon 
whole carcasses (Schaller 1972). The fact that meat has low calorific 
value does not mean it has low food value. 
(e) Meat is apparently only eaten in parties, and the social 
functions of predation may therefore be more important than the 
nutritional benefits (Kortlandt 1972). Instances of individuals eating 
meat alone and from their faeces were described, so this generalisation 
is no longer valid. 
(f) Predations apparently occur as an outcome of social stress, 
and are therefore not primarily directed to obtaining food (Kortlandt 1972). 
Although one of the predations observed by Teleki (1973) did occur while 
chimpanzees were apparently frustrated at not getting bananas, there is 
4.49 
no evidence that this was other than coincidental. Analysis of the 
situations in which chimpanzees became aware of and caught their prey 
(111/3) showed that the primary determinant of hunting success was the 
vulnerability of the prey iri the particular situation in which they were 
encountered. 
(g) Not all chimpanzee popu1ations eat meat, which is therefore 
unnecessary nutritionally. It was shown in the Introduction that there 
have been no long field studies which have failed to observe meat - eating. 
Even in Gombe, where the particular nature of the vegetation appears to 
favour high rates of predation (111/3), predations were seen only once 
in about two hundred observation hours (Table 4.6). It ~s quite possible 
that all chimpanzee popu1ations do eat meat: any which do not might ga~n 
if they were able to do so. 
(h) If meat is a valuable food, why did chimpanzees not kill 
baboons more often when they mingled with them in the artificial feeding 
area (Te1eki 1973)? Evidence was given (III/3/v) that when chimpanzees 
killed baboons they selected occasions when adult male baboons could not 
defend very effectively. Though it is possible that such situations arose 
more often than predations were attempted, evidence of this would be 
difficult to obtain. The suggestion that chimpanzees preyed as often as 
they could is supported by the fact that unsuccessful hunting attempts 
were more frequent than successful predations: 18 out of 28 attempts to 
catch young baboons failed (Te1eki 1973). Inhibitions against killing 
other species can apparently occur in primates: DeVore & Washburn (1963) 
saw a baboon catch a young vervet monkey in its mouth, and then release 
it unharmed, even though vervets were eaten by baboons. However, there 
is no clear evidence that such inhibitions were important here. 
The objections to the theory that predation by chimpanzees is 
primarily a food-getting behaviour are not substantial . In support of 
4.50 
the theory, chimpanzees commonly eat meat in captivity (Kort1andt & Kooij 
1963). Many chiefly herbivorous mammals are also occasionally carnivorous, 
especially those feeding selectively on browse leaves and fruit (primates: 
Jones 1972, Lorenz 1971; ungulates: Jarman 1974). When it is known which 
nutrients are most important the suggestion that the functional consequences 
of predation lie in obtaining nutritional benefit will be more acceptable. 
Salt and amino-acids uncommon in plants (Brambe11 1972) are two obvious 
candidates. 
Meat was evidently a very unreliable source of food: it was 
obtained when prey could be caught, but the opportunities arose rarely. 
Like the !Kung Bushmen, who "eat as much vegetab le food as they need, 
and as much meat as they can" (Lee 1968), chimpanzees appear to use meat 
as a desirable extra. The ability to do so derives from their mobility 
on ground and in the trees (Te1eki 1973) and the lack of adequate defence 
mechanisms in the prey species. In Gombe the frequency of predation by 
chimpanzees appear s to be higher than elsewhere (though no figures are 
available for comparison). It was suggested that this was due partly to 
the common presence of broken canopy, in which co10bus monkeys were 
particularly vulnerable: regular predation is possible because of the 
low density of chimpanzees in relation to their prey. Other factors 
were evidently the high density of bushpigs and the fact that baboons 
killed bushbuck which could then be stolen. Doubtless the experience of 
individuals in recogn~s~ng that prey are available and may be captured 
is important: however, it seems that if we are to regard the hunting and 
killing of mama1s as a culturally influenced behaviour (van Lawick- Gooda11 
1973), we must also remember that the potential for its development is 
subject to a number of ecological variables. 
4.51 
2. The significance of predation to the prey species 
The calculation of predation rate used data from over 9000 hours 
of direct observation to yield a range of estimates of the number of 
predations occurring per year: the relative frequency of each species 
as prey was found from data extending over 13 years. Data were taken 
both from the Northern community, which includes the artificial feeding 
area in its range, and from the Southern community, which does not. The 
large sample of observation hours allows the results to be treated with 
confidence. 
Colobus and bushpig were found to suffer a significant annual 
mortality (estimates between 2.5% and 15%, Fig. 4.8) from chimpanzees, 
and bushbuck from the joint attentions of chimpanzees and baboons. To 
individuals of these species chimpanzees were as dangerous, statistically, 
as plains carnivores to ungulates (Kruuk 1972, Schaller 1972). Thus 
although to chimpanzees predation may be a relatively uncommon event, 
their density in relation to the prey ~s sufficiently high to be a very 
important influence on the prey species. The species which was least 
affected numerically was baboons, apparently because adult males are 
effective defenders. 
It seems likely that such high predation rates are not 
widespread. An indication of the fact that colobus monkeys are not used 
to being preyed on by chimpanzees is that on two occasions mothers have 
been killed while threatening chimpanzees who had killed their infants: 
such behaviour is clearly maladaptive. The evidence that special 
conditions in Gombe favoured the opportunities for predation has 
already been given. 
Their predatory capacity means, nevertheless, that chimpanzees 
may compete more effectively with their prey spec~es than if they were 
4.52 
strictly vegetarian. Kortlandt's suggestion (1972) that one of the 
functions of predation 1S to reduce competition now seems unwarranted: 
the three prey speC1es which appear to be their most important competitors 
(blue and redtail monkeys qnd baboons) were eaten least. However, if 
chimpanzees guarantee for themselves undisputed access to any food 
sources they visit they may thereby be more effective competitors. It 
was shown that active competition occurred at E. guineertsis between 
chimpanzees and adult male baboons when food was scarce in 1972 (Table 
3.21): had they been able to displace baboon males as they could other 
competitors, time, energy and available food would have been saved. 
The occurrence of such high rates of predation means that 
chimpanzees may exert important selective forces on prey populations 1n 
certain circumstances. This has hardly been suspected until recently. 
For instance, in discussing the adaptive significance of the large troop 
sizes of red colobus, Clutton-Brock (1972) ignored the possible influence 
of predation: the suggestion that lone colobus were more vulnerable than 
those in troops (Fig. 4.5) suggests that it could affect group 
organisation (Goss - Custard et al. 1972). (Data on predation are 
difficult to obtain by study of the prey: observers can watch few animals 
at a time, and the chance of seeing predation is therefore low: further-
more, chimpanzees are usually afraid of seeing man.) Even if chimpanzees 
prey rarely in forest habitats, as seems to be the case (Kortlandt 1972), 
this may be partly because potential prey species have effective defence 
strategies which if relaxed would allow predation to occur. Man, crowned 
eagles and leopards have been suggested to be the most important predators 
of arboreal monkeys (Haddow 1952), but the threat of chimpanzees should 
not be ignored: not only may they be the most important predator in 
some circumstances, but they will also exert different selective pressures 
from those of other predators. 
4.53 
3. Why do chimpanzees share meat? 
1. Why do individuals try to get meat from others? 
This is simply answered. The evidence above indicated that meat 
1S a valuable food which chimpanzees want to eat: but prey 1S hard to 
catch. However this raises further questions. It is clear that individual 
behaviour varied at predations, from merely watching others (Teleki 1973) 
to a variety of increasingly intense attempts at meat-getting including 
prolonged attack. Why did some individuals try harder to get meat than 
others? 
On the principle of Occam's razor, the best answer 1S that which 
involves the simplest assumptions. The assumptions made here are that 
the only benefit is the food value of the meat, and that a given amount 
of meat is equally beneficial to all adults. This is probably simplistic: 
pregnant females, for instance, might derive more benefit from meat than 
prime males. However, the advantage of this approach is that more 
complex factors affecting meat-sharing need only be considered when the 
simpler assumptions are shown to be inadequate. 
If all adults obtain equal benefit from eating meat, but behave 
differently in trying to get it, they must experience different costs: 
otherwise they would all try equally hard to get meat. Cost analysis 
may therefore explain the observed differences in meat-getting behaviour. 
The potential costs of trying to get meat are suggested to be: 
(a) Time is wasted in trying to get meat unsuccessfully. 
(b) The nearer the individual comes to the meat-eating clusters, the 
more likely he is to be involved in aggression. It was shown that 
attacks were frequent, even on non-possessors (Table 4.14) (see also 
van Lawick-Goodall 1971 and Teleki 1973). Aggression may cause physical 
damage. 
(c) Aggression may also cause psychological damage. All interactions 
4.54 
affect the nature of the relationship between individuals. The cost of 
being involved in overt interactions may vary according to how much 
bluff an individual normally uses, and according to how he can expect 
his future relationships to . develop. For instance, an old male who has 
lost dominance, as typically occurs (Bygott 1974), cannot expect to 
regain it, and therefore has little to lose in challenging. By contrast, 
a young male who loses confidence as a result of attacking unsuccessfully 
or of being attacked may subsequently find it harder to rise in the 
dominance hierarchy: he therefore incurs costs. (The term "expect" 
need not imply any consciousness by the individual: the only requirement 
here is that the expectation is sufficiently probable that its effects 
can be subject to natural selection.) Dominance is certainly not the 
only relationship which may be involved, though the most obvious: the 
cost of aggression by a male towards a female may be a reduced chance 
of mating. 
(d) Energy may be wasted in trying to get meat. In the examples 
gLven (V/3/ii) considerable energy was spent in attacking a possessor, 
with little or no meat gained as a result. 
Teleki (1973) showed that the harder an individual tried to get 
meat the more he tended to get. The hierarchy of success at getting 
meat (Tables 4.12 and 4.14) is therefore equivalent to a hierarchy of 
how hard an individual tried to get meat. Since the potential benefits 
for all individuals were assumed to be equal, the hierarchy of meat-
getting should therefore also be an order of decreasing value of costs 
incurred in trying to get meat. 
Costs per unit effort presumably decrease with increasing age, 
since damage to an animal affects less of his reproductive potential as 
his age Lncreases (Trivers 1971, Parker 1974). The hierarchy of meat-
getting is therefore expected to be correlated with age. A significant 
4.55 
correlation between age and success at getting meat was found 
(Table 4.13). Differences between males in their success at getting 
meat can therefore be explained largely in terms of competition, without 
the need for considering the' benefits to the giver: old males can afford 
to compete more effectively because they have less reproductive potential 
to lose. The two males who obtained more meat than expected on this 
hypothesis (Figan and Faben, Table 4.13) were high- ranking brothers. 
Cooperation within the family may have increased their success. Figan 
was high- ranking but this does not seem very important on its own: 
Humphrey was dominant to Figan until September 1972, but was lower ~n 
the meat - getting hierarchy. 
The relationship of cost per effort for males to that for females 
~s less clear than within a sex. Clearly mothers with young have high 
costs, since damage to the offspring is damage to her reproductive 
investment. All females, however, may have higher costs per effort than 
males because physical damage may be more disastrous. A clear example of 
this is a comparison of Faben and Madam Bee. Both lost the use of one 
arm from a disease (poliomyelitis?) in December 1966. Madam Bee has 
since given birth twice, but both infants died after a few months: her 
deformity appeared responsible for poor motherhood, and the loss of her 
reproductive potential. Faben, on the other hand, appears to be 
sexually more successful than most other males (Tutin, in prep.). 
Table 4.14 shows that all females were less successful at obtaining 
meat than all but one of the adult males. 
Among females factors affecting costs per effort may be affected 
by age and number of young. Teleki (1973) found that the aged female 
Flo obtained more meat and that oestrous females were more persistent ~n 
their efforts to obtain meat than others: both these results are 
consistent with the expectations of the hypothesis. 
4.56 
Although it was found that in general individuals got more meat 
as they tried harder (Teleki 1973) (from finding scraps, to begging, 
to pulling meat from the possessor), it was shown (Table 4.11) that 
aggressors normally failed to' get meat by attacking a possessor. This 
apparent anomaly is partly a semantic problem: attacks were defined 
(V/3/ii) by the extent and nature of physical contact, but many potential 
attacks did not occur because the possessor immediately relinquished his 
meat. The fact of holding a resource is expected to increase the chance 
of one of a particular pair of individuals winning a fight (Parker 1974): 
and it was shown that individuals were physically capable of keeping 
meat while under attack. Nevertheless, if individuals differed much in 
age (or were different sexes), the possession of a resource may not be 
enough to balance the difference in the costs of aggression. The two 
oldest males, Hugo and Mike, were involved in 18 out of 20 attacks, and 
fought together 6 times. These data are therefore consistent with the 
hypothesis that su~cess at getting meat is explicable by the costs of 
competition and that all adults benefited equally from eating meat. 
Individuals with low costs try hardest and get most meat. 
ii. Why do possessors allow other individuals to take meat? 
If the hypothesis that the benefit of meat-eating varies with 
the amount of meat consumed is correct, then an individual should not be 
prepared to share meat until the cost of holding it reaches a certain 
critical point: this point should be determined by the probability of 
incurring damage to his reproductive potential. Thus the cost-benefit 
relationship in keeping meat from another should be similar to the cost-
benefit relationship in trying to get meat from another. This appears 
to have been so: for instance, males lost meat more frequently to 
older, but less dominant, males than to females. 
However, the real point of interest here is why the critical 
I 11 
i 
4.57 
point is ever reached. If the benefit is equally great to the possessor 
and those without, and if the fact of possession alone confers a 
competitive advantage (Parker 1974), then the hypothesis predicts that 
the oldest male should never. share his meat: in fact, a male should 
never allow younger ones to take any. This expectation is not fulfilled 
(Teleki 1973), and the hypothesis is therefore wrong. 
The obvious fallacy in the hypothesis lies in the assumption 
that benefits increase directly with amount of meat consumed. It was 
clear that individuals tended to lose interest in meat the more they ate 
(see also van Lawick-Goodall 1971 and Teleki 1973), and observations were 
cited of individuals voluntarily leaving meat. There are probably two 
factors involved. First, although the important nutritional components 
of meat are unkn'own, it seems likely that, as with any food, it is 
possible to have too much of a good thing: variety is essential. 
Second, evidence was given that chimpanzees eat meat very slowly: time 
spent eating meat with unnutritious leaves may, after a certain amount 
has been eaten, be better used in eating plant foods. 
An individual with a large amount of meat may therefore possess 
a resource only part of which is valuable to him. However, he may use 
the parts which he does not need to reduce the costs of possession, by 
"paying" other individuals to go away. The tendency to avoid others 
when an individual has meat, and the chance of being attacked, clearly 
indicate that the persistent attentions of those without meat are a cost. 
Once an individual succeeds in getting meat he moves away, sometimes 
followed "by others (V/2/ii and van Lawick-Goodall 1971). The dispersive 
effect of permitting others to take meat may therefore be a benefit 
which outweighs the cost of losing some, especially where th~ possessor 
either has consumed or still holds a good portion. Meat is very 
occasionally actually handed to a begging individual, and this can be 
,.... 
4.58 
regarded as an extreme case (sometimes elicited by extreme, but non-
aggressive, persistence, such as "throwing a temper tantrum" (van 
Lawick-Goodall197l». In an experimental situation, Nissen & Crawford 
(1936) found that chimpanzees were more willing to share less valuable 
items. 
The fact that individuals are ever prepared to share, and the I 
differences between individuals in how hard they work to get meat, can 
therefore be explained by assuming that individuals incur the same 
I I 
benefits but different costs. Giving by one individual is not considered 
altruistic, even in the sense used by Trivers (1971) of being "apparently 
detrimental": rather, it is the least costly consequence of the 
! 11 
competitive behaviour of others. 
Finer analysis is hardly possible at present. High rates of 
I I 
I 
meat-getting by kin or oestrous females (Teleki 1973) may be due to 11 
I 
increased benefits to the giver, who can choose which individuals may 
i I 
I, , 
I 
take meat: by allowing a female to take meat he may rid himself of 11 
::11 
other competitors as well, since some of them may follow her (van Lawick- I 
Goodall 1971). However, it is also possible that an oestrous female 
would gain more than most individuals from a fixed amount of meat, and 
I'l' 
I 
li 
that she can therefore afford more costs. Though the importance of 
11 
mating rights are therefore questionable, it lS reasonable to assume 
, 
that kin selection favours selective release of meat within a family. 
Thus it seems likely that there are cooperative elements to the sharing 
of meat, but that they affect which individuals get meat, rather than I 
whether it is shared. 
Food-sharing at predations is therefore considered to occur 
because meat is in limited supply, and lS difficult to eat without 
exposing it to others, who can thereby impose costs on the possessor. 
4.59 
At plant foods competition normally takes the alternative form of 
obtaining the best feeding site and eating as fast as possible (3/VI/2 
and 3/V/3). However, where feeding sites were limited, old males 
appeared to compete more successfully than younger, though more 
dominant, males. Like chimpanzees at most food sources, hyaenas compete 
at large carcasses by eating fast (Kruuk 1972). Only at a small carcass 
can hyaenas, like chimpanzees, physically prevent others from eating 
meat: some individuals can therefore compete more successfully than 
others. Whereas hyaenas eating large prey and chimpanzees eating plant 
foods share food more evenly than chimpanzees eating meat, baboons share 
meat even less. Harding (1973) found that adult males competed for prey 
but that meat-sharing was only tolerated when the possessor was satisfied. 
Since baboons never took more than one hour to eat the prey, all the 
meat was beneficial: in this case, unlike for chimpanzees, the critical 
point at which the benefits of possess~on are outweighed by the costs ~s 
apparently reach~d only when another baboon attacks. The difference 
between baboon and hyaena meat - eating behaviour is apparently due to 
differences in the size of the prey, and hence in the value to the 
possessor of the whole carcass. Baboon prey are normally much smaller 
than those of hyaenas: when hyaenas compete around small pieces of 
meat, one individual may prevent others from eating, like baboons 
(Kruuk 1972) and chimpanzees eating small birds. Differences in the 
form of competition between species and foods are therefore attributable 
to differences in individual competitive abilities. Though superficially 
non-competitive (Teleki 1973), meat - sharing by chimpanzees now appears, 
paradoxically, to be the result of differential competitive ability for 
a food 'Ilhose value decreases as more is eaten. 
4.60 
4. Sharing of other foods 
Transfer of natural plant food was seen only five times 1n this 
study (except between mother and offspring). Cardboard, a canvas bag 
and a rush mat were also shared. Individuals twice attacked the 
possessor, but normally took food uS1ng the begging gestures described 
by van Lawick-Goodall (1968). The pieces allowed to be taken were 
generally of low value to the possessor (plants: small fragments or 
chewed fruit; miscellaneous: part or whole when possessor had lost 
interest) . 
Strychnos fruit were shared four times. These were the Slze 
of an orange and took 5- 10 minutes to eat. Like meat and the 
miscellaneous foods, a possessor could grasp and carry the entire item, 
and only with his compliance could another individual approach and take 
food. The other plant food shared between adults was a bunch of 
Sterculia quinqueloba pods: these grew on small trees where one 
individual could ' prevent access by others (3/VI/2). The factors 
responsib le for begging and sharing of these foods thus appear to be 
similar to those involved in meat distribution. 
However this 1S not true of sharing bananas. Unlike meat, 
Strychnos fruit or a canvas bag, bananas are easily torn 1n pieces; 
nor can chimpanzees climb nimbly to an inaccessible perch carrying a 
bunch of them. Normally when observers opened a box or boxes the most 
high-ranking individuals present took the bananas: occasionally 
subordinates found the bananas first, and they would be attacked and 
lose their food (van Lawick-Goodall 1968). In contrast to meat, no 
cases are known to me of subordinate individuals violently attacking 
or defending against dominants in competition for bananas. (No attacks 
over other plant foods were seen. However, old but subordinate males 
competed successfully with their younger dominants for feeding sites, 
!! 
I 
I 
I 
i 
IJ 
I: 
11 
I; 
[I 
I 
, I 
" 
1 
[I 
! 
I 
i 
I 
I 
11 
I: 
1 ~ 
3/VI/2/ii) • 
Individual success ~n getting bananas thus appears to have 
been more closely related to rank than in getting other foods. This 
may have been because if an individual with bananas was attacked by a 
determined aggressor he was likely to lose them: bananas need 
4.61 
delicate handling. If so, he could afford to retaliate. A series of 
attacks could ensue, during which neither individual could eat or even 
hold his bananas; meanwhile other individuals could retrieve them. 
Differences in rank exist and are predictive of the outcome of attacks 
not in the context of food (Bygott 1974). If an originally competitive 
attack or dispute tended thus to lose its banana context, the 
subordinate would be punished by losing the fight. This would explain 
why subordinates should have learned that it is not worth their while 
to compete for bananas. 
Cases of individuals begging successfully for bananas were 
analysed by McGrew (in press, b). The majority (86%) were offspring 
obtaining food from their mothers. McGrew related this to Trivers' 
(1972) model of conflict between parent and offspring over the amount 
of parental investment: by sharing bananas (and other plant foods) the 
mother may bring benefit both to her infant (faster development etc.) 
and her future reproductive potential (breeding sooner because of 
earlier independence of present infant). Those obtaining bananas from 
"unrelated" individuals were mostly adult females begging from adult 
males: individual differences appeared related to other socio- sexual 
behaviour patterns, and thus mating rights may have been important. 
Successful beggers normally obtained bananas from an individual with a 
large pile. 
Chimpanzees are the only primate seen to share (transfer) food 
~n the wild as adults (reviewed by Dare 1974). Sharing between mother 
4.62 
and young, however, 1S known in several species (Kavanagh 1972, Dare 
1974). If maternal food transfer is merely an extension of lactation, 
we may expect it to occur 1n many species, wherever the mother, but not 
her offspring, can obtain or process a food which the latter can digest. 
It is not obvious, however, why chimpanzees should be the only species 
which transfer food, or allow it to be taken, between adults. 
The foods which chimpanzees share are ones they eat slowly: 
they have no shearing teeth to cut meat, nor cheek pouches to store 
bulky plant foods. While eating meat a chimpanzee rema1ns 1n 
possession of a visible food which others want. During weaning a 
chimpanzee doubtless learns that by pestering his mother he may get 
what he wants (milk or solid food): perhaps this experience promotes 
the tendency for adults to beg from each other. It is interesting in 
this respect that a temper tantrum, seen to be given three times by 
adults trying to get meat, is an extreme form of pestering by an infant. 
Unusual characteristics of the food may thus contribute to its 
transfer by begging. These may include the following. First, the food 
item tm~es a long time to be consumed and is in short supply: 
individuals are therefore 1n competition. Second, the food can be held 
or protected during attack: dominance rank is then not predictive of 
the outcome (4/VI/3 and Parker 1974). Third, the food item decreases 
in value to the eater as more is eaten: and if it is divisible only 
with difficulty, as meat is, aggression 1S unlikely to be rewarding. 
Consideration of these features appears to explain fairly well 
the differences in the extent of aggression over and transfer of food 
items by "unrelated" chimpanzees. In other primates food items are 
normally small and can be eaten rapidly: cheek pouches allow many 
species to store food before chewing it. A notable exception is meat-
eating by baboons. This may be similar to banana- eating by chimpanzees. 
If in both cases an adult cannot effectively defend his food from 
another without releasing it, final possession is predicted by 
dominance rank. "Sharing" occurs through one individual taking over 
the food. 
4.63 
Strum (in press) recently described limited meat-sharing by the 
baboons previously studied by Harding (1973). An adult male and an 
adult female were seen to eat from the same piece, and one even moved 
to allow access by the other. Mothers and offspring shared in a similar 
way. Baboons therefore behaved more like chimpanzees than previously 
reported to do. Strum found that the baboons preyed more often than 
when observed by Harding, but that sharing still normally occurred 
only when one individual supplanted another or when the possessor lost 
interest. Within age-sex classes there were substantial differences 
in meat-getting, and experience appeared to be one factor affecting the 
interest of juveniles in meat. 
The increased sharing of meat observed by Strum may have been 
related to the higher predation rates. Possibly the latter meant that 
the nutritional value of meat was lowered, at any rate for those 
individuals who ate the most. As in the case of other species, the 
explanations suggested here can be tested by further data. 
Reviewing studies of rats, dogs, porpoises and captive 
chimpanzees, Krebs (1971) found that acts which have been described as 
altruistic have equally good alternative explanations. For instance 
Nissen & Crawford (1936) concluded from their study of food-sharing by 
captive chimpanzees that "responsiveness to begging may be (1) a 
response, probably related to reactions given ~n answer to calls for 
help and appeals for sympathy, which gets rid of the noxious stimulus 
of begging by the partner, or (2) an exchange with delayed reward •.. ". 
This discussion similarly rejects the need for labelling the sharing of 
b 
4.64 
meat or other foods as altruistic. 
Chimpanzees are closely related to man, and are the only 
primate in which adults are known to transfer food. The coincidence 
has led to speculation that the roots of human moral behaviour may be 
seen ~n chimpanzee meat- eating (van Lawick-Gooda11 1971). Individual 
chimpanzees and men can both monopolise and defend large indivisible 
food items more effectively than most primates: strong hands grasp 
better than mouths. Of course food-sharing between adults is known ~n 
a variety of non-primate spec~es ~n courtship displays (e.g. Lack 1966, 
Nisbet 1973) and in males bringing food to females rearing young (e.g. 
Schaller 1972). Clearly food transfer is not in itself a precursor of 
morality. Nevertheless the particular circumstances which promote it 
in chimpanzees may be similar to those which have contributed to the 
complexity of human attitudes to object possession. 
4.64 
meat or other foods as altruistic. 
Chimpanzees are closely related to man, and are the only 
primate ~n which adults are known to transfer food. The coincidence 
has led to speculation that the roots of human moral behaviour may be 
seen in chimpanzee meat-eating (van Lawick-Goodall 1971). Individual 
chimpanzees and men can both monopolise and defend large indivisible 
food items more effectively than most primates: strong hands grasp 
better than mouths. Of course food-sharing between adults is known ~n 
a variety of non-primate species ~n courtship displays (e.g. Lack 1966, 
Nisbet 1973) and in males bringing food to females rearing young (e.g. 
Schaller 1972). Clearly food transfer is not in itself a precursor of 
morality. Nevertheless the particular circumstances which promote it 
in chimpanzees may be similar to those which have contributed to the 
complexity of human attitudes to object possession. 
I. Introduction 
Chapter 5 
RANGING BEHAVIOUR 
5.1 
One of the problems in understanding the adaptive significance 
of chimpanzee social structure has been the difficulty in describing its 
nature. Few data are available on the ranging behaviour of individual 
chimpanzees, yet until we understand how individuals are dispersed it 
will be difficult to speculate on the function of their social structure. 
In many animals, such as most primates, individuals travel in semi-
permanent groups, and their ranging patterns may be relatively easily 
described. The instability of chimpanzee parties creates problems: 
individuals must be treated separately, and this demands considerable 
observation time. 
To solve this, a system was developed for pooling ranging data 
from all observe rs. Since January 1972 about 400 observation hours per 
month have been recorded on standardised Group Travel charts (Chapter 2). 
Their reliability will be discussed in the relevant sections. 
One of the a~ms of this chapter is to use the data to g~ve an 
overall view of chimpanzee ranging behaviour in Gombe. This may be 
useful not only for the socio- ecological problems posed here, but also 
for specific studies concerned with individual relationships. Grateful 
acknowledgement is made to the many observers who contributed their data. 
The second a~m is to understand how rang~ng behaviour is related 
both to ecological pressures and to social structure. 
Ranges over three time periods (one day, four days and one year) 
are analysed. These data indicate the existence of differences ~n 
ranging behaviour between sexes and seasons. These are further 
investigated by analysing the directions by which individuals entered 
the artificial feeding area. 
5.2 
The data also show that separate communities of males may be 
defined on the basis of ranging patterns. Interactions between members 
of different communities are described. 
The final data section briefly describes factors affecting the 
timing and direction of travel. 
11. One-Day Range 
1. Methods 
Collection of these data was time-consuming: observations were 
aborted when the target individual was lost during a follow. Group 
Travel charts were analysed to supplement the 54 one- day ranges collected 
during this study. Their overall reliability was discussed in Chapter 2: 
in addition it ~as important that the concept of travel was standardised. 
Although this was not formally agreed by all observers until 1973, it is 
generally safe to say that movements around a point (e.g. food source, 
grooming party) were ignored unless more than about 30 m. Uniformity 
across observers was maintained throughout the study by regular checking 
and discussion. 
Path length was measured by counting the number of lines crossed 
on a superimposed grid (100 m. by 100 m.). When a hodometer became 
available the "lines crossed" measure was checked against the hodometer 
readings on twelve male and twelve female records of different lengths. 
The two measures were highly correlated (Pearson R = 0.995, N = 24). A 
regression line was calculated and all readings were converted to 
kilometres by the equation Y = (0.7X - 0.3) x 0.115 , where Y distance 
in kilomet~es and X = number of lines crossed. Caution in using the 
5.3 
ki10metric distances was advised in Chapter 2. 
Records selected for inclusion were all those completed by 
competent observers following between night-nests, from January 1972 to 
September 1973. In addition four observations begun before 0715 hrs 
were included. 
Where possib le, a standard method is used below to compare 
distances travelled in a day by different individuals. As an example, 
the test for differences in one-day ranges of Northern and Southern males 
was as follows. Each available record of a Northern male was matched 
with the nearest Southern male record by date, provided (a) it was taken 
within 14 days, (b) it had not been previously matched in this test and 
(c) it was not observed during a consortship. The matched pairs so 
produced were compared by a Wi1coxon test. 
Day ranges were assigned only to the target individual, except 
in the case of juveniles, whose records were assigned to the mother when 
they travelled together. The data are summarised in Table 5.1. 
Table 5.1 Distances travelled ~n one day 
Distribution of one-day ranges from January 1972 to September 1973, with 
target individuals classed by age, sex and community. Consort pa~rs are 
shown separately. The table takes no account of variations in sampling 
across months or individuals within classes. NR = number of records ~n 
sample. NI = number of individuals (or consort pairs) in sample. 
Class 
N male 
S male 
Females 
N juvenile 
Consort pair 
;~ 
Distance travelled between night-nests 
(kilometres) 
--'--------
0-1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10 
1 1 13 24 13 14 5 8 2 2 
,/, ;~ 
0 1 6 6 2 3 3 0 2 0 
4 14 15 13 9 3 3 0 0 0 
0 2 1 1 2 0 0 2 0 0 
0 4 3 5 3 0 0 0 0 0 
One record in each of these cells started near camp. 
NR NI Median (km.) 
83 8 4.2 
23 7 3.8 
61 10 2.8 
8 1 4.0 
15 3 3.2 
'I I . 
I 
11 
I) 
I' 
., I' 
:, 
11 
/! 
1 
I 
I 
"IJ, '1 -iliI~ 
5.4 
.1 
'I . I! 
2. Factors affecting length of range , 
~. Sex 
Northern males had longer one-day ranges than Northern females 
(z = 3.62, N = 29, P < 0.001). Most of the females observed were , i 
anoestrous. 
Males and females did travel together for complete days 
occasionally, the reasons being sexual (males consorting or accompanying 
oestrous females), familial (e.g. FG with FL) or merely friendly (e.g. 
HG with GK). Omitting consort pairs, differences between one-day ranges 
of males with and not with a female all day were not significant (T = 26, 
N = 12, n.s.). However, females travelling with males had longer ranges 
than females not with males (T = 2, N = 6, P < 0.05). This suggests that 
heterosexual association caused females to change their travel patterns 
more than males, except when consorting. Tutin (in prep.) found that 
females travelled further when oestrous than when anoestrous. 
i i. Communi ty 
The two Southern male one-day ranges which started near camp 11 
!' I 
(Table 5.1) were omitted from this comparison, since their length was 
'\; 
apparently influenced by the starting- point (v). 
One- day ranges of Northern males were longer than those of \"1 
Southern males (T = 50.5, N = 20, P < 0.05). It is not clear what was 
responsible for this difference, which may have been related to r! 
differences in range size over four days and one year (Ill & IV). 
11 
H~ • Individual 
Differences between individuals were tested witruncommunities, 
I 
"I 
age - sex classes and seasons. The data are shown in Table 5.2. ~ 
I 
I 
I I I 
I I , J~ \ )l! 
5.5 
Table 5.2 
Individual differences in length of one-day range 
Indi vidua1s are compared only where both have at least three records. 
Records are matched within co?UUunities, age-sex classes and seasons. 
Dry 1972 Wet 1972/3 Dry 1973 
Old N male S male Prime N male Prime N male N female 
HG MK CH GI FB FG ST EV FG ML PS 
2.1 1.9 3.1 2.1 2.4 3.5 2.4 2.7 4.0 2.2 2.8 
2.5 3.1 3.2 2.8 3.5 3.5 3.5 4.6 4.2 2.4 3.3 
2.9 3.9 4.4 3.6 3.8 4.2 6.3 4.6 7.0 2.5 4.1 
3.9 4.9 8.3 5.9 5.0 5.0 7.2 2.8 4.2 
4.1 6.2 5.3 8.3 3.0 4.8 
5.0 7.5 5.3 8.4 3.3 6.0 
5.6 6.0 3.8 6.6 
6.5 3.9 
7.1 4.6 
7.3 4.9 
7.3 
9.2 
9.9 
Mean: 
4.3 3.4 4.7 3.6 3.2 5.0 4.0 6 . 2 6.5 3.3 4.5 
The only pair ~n Table 5.2 which were significantly different 
were FG and FB (Mann-Whitney U = 3, n l = 3, n2 = 6, P < 0.05). One 
result at the 0.05 level in a sample of seven does not argue strongly ~n 
favour of individual differences. Possibly FB was affected by the 
disability of his "polio arm" (VI/2/i), but he was certainly capable of 
long-distance travel. For instance, he travelled more than 7 km. in a 
day three times when other individuals were targets in the 1973 dry 
season. Thus there is no strong evidence of individual differences ~n 
one-day ranges within the stated classes. 
iv. Age 
Old males beyond their physical prime were eas~er to follow than 
others because they were slower. HG, WZ, GO and HH were compared with 
younger males within their respective communities, and had significantly 
shorter ranges (z = 1. 70, N = 26, P < 0.05). However, they were all 
capable of travelling a long distance (e.g. HG, 18 Sep 1973, 8.3 km) . 
5.6 
v. Camp 
The two Southern male ranges which started near camp were long 
(Table 5.1). A further record on HH, omitted because it started at 
0800 hrs, also began near camp and was the longest one-day range recorded 
(10.7 km). Though this was probably the effect of starting on the 
boundary of the community range, it was possible that one-day ranges of 
Northern males might be different between days when they did or did not 
visit camp. A comparison revealed no difference (within age-classes: 
T = 104, N = 21, n.s.). Distance travelled in a day by Northern males 
was therefore apparently not influenced by visiting the artificial 
feeding area. 
v~. Season 
The data did not allow compar~sons between Dl , Wand D2 (3/VII) 
if individuals were to be matched. The periods compared here are 
therefore May to October (dry seasons) and November to April (wet season). 
Because of the variable number of observations of each individual, means 
were taken of individuals' range length during each time period (Table 5.3). 
Table 5.3 
Mean lengths of one-day ranges (arranged by target individual and seaSon) 
Dry season: May-October. Wet season: November-April. Only individuals 
with records in more than one season are included. Distances ~n km. 
N males Juv N females S males Season HG MK HM FB EV FG ST JJ GB ML PS FF WIZ GO CH GI 
Dry Mean 3.5 4.2 3.5 4.4 3.9 4.4 3.7 3.2 3.4 3.0 2.5 3.1 3.5 4.8 3.8 
'72 N 9 4 2 2'1~ 1 7 1 2 1 2 1 l2'~ 2 4 4 
Wet Mean 3.8 4.1 2.5 3.2 4.3 5.0 4.4 4.4 2.8 2.0 2.0 1.9 2.8 4.8 
'72/3 N 4 2 1 3 1 6 3 1 1 3 2 1 1 1 
Dry Mean 3.8 - ' 7.7 5.5 6.2 6.5 7.5 4.9 3.3 4.5 2.5 4.4 3.3 
'73 N 4 1 1 13 6 1 5 10 7 2 2 2 
,'c 
One record for FB (28 Aug 72) and three for WIZ (20-22 Oct 72) were 
omitted because their ranges appeared 'to have been affected mostly by 
their physical condition (injury and giving birth respectively). 
i I 
, 
, 
. 
'I I 
:: 
li 
1 
, 
11 
11 
I, 
11 
Ili 
I", 
I,!; 
! 
1 
11 
5.7 
The three seasons were tested for differences by matching means 
of each individual. The results are shown in Table 5.4. 
Table 5.4 
Significance of seasonal di f fe rences 1n lengths of one- day ranges 
Data from Table 5.2. 
I 111'1 II 1 
, I 
Seasons compared Indi viduals N Wilcoxon T P Season with longer ranges 
Dry '72 & Wet '72/3 All 13 41 n.s. 
Wet '72/3 & Dry '73 All 10 2.5 <'02 Dry '73 
Dry '72 & Dry '73 All 11 12.5 n.s. I ' 
Dry '72 & Wet '72/3 Northern only 12 29.5 n. s. 
Wet '72/3 & Dry '73 Northern only 8 0 <'01 Dry '73 
Dry '72 & Dry '73 Northern only 8 1 <.02 Dry '73 11 
'I 
The results were calculated both for all matched pa1rs available 
and for matched pairs of individuals within the Northern community. 
Considering only Northern individuals, Table 5.4 clearly shows that the 
1973 dry season differed from the other two seasons: r~nges were 
, 
significantly longer. 
Inclusion of the one-day ranges of Southern males causes the ~ : 
difference between the two dry seasons to be just less than significant. , 
This suggests that different ecological factors were operating in the 
Northern and Southern community ranges in the dry seasons of 1972 and 
1973. This was so: P. curatellifolia fruited extensively in the 
Northern range in 1973, but there were few trees in the Southern range. 
It is possible that the local abundant fruiting of P .curatellifolia was 
a cause of the relatively longer one-day ranges ln the Northern range 
during 1973. 
V11. Miscellaneous 
Factors influencing the direction and timing of travel and 
discussed below (VII). The occurrence of stimuli such as the calls of 
non- community males doubtless affected lengths of one- day ranges: 
habitat type and land form ~n particular areas may also have been 
important. 
5.8 
There were two examples of individuals in unusual physical 
condition travelling short distances. When FB's hand was wounded he 
travelled only 0.6 km in a day. WK covered a mean of 0.4 km per day ~n 
three days when she gave birth (she gave birth after the first "short" 
day: Plooij, in prep.). 
Another influence may have been the length of the previous day 
range. On the day after he travelled 10.7 km, RH walked lethargically 
and covered only 2.4 km. It was my impression that after travelling 
widely for three or four days an individual tended to leave a large 
party and spend a day or two travelling alone in a small area. This was 
not confirmed. 
3. Comparison with other studies 
In this study the distance travelled ~n a day varied from 0.2 km 
(pregnant female) to 10.7 km, with different means for different seasons, 
sexes and communities (Tables 5.1, 5.2, 5.3): a reasonable average was 
4 km for males and 3 km for females. By observing parties and estimating 
their travel paths from calls, Nishida (1974) found one- day ranges 
between 0.2 and 4.5 km, also varying seasonally. However, Nishida's data 
are not exactly equivalent since it was not certain that any individual 
was present throughout the observation. Jones & Sabater pi (1971) 
estimated a one-day range of chimpanzees in Rio Muni as being 4.1 km long. 
The only primates known to have longer day ranges than chimpanzees 
are all terrestrial: these are geladas (6.4 km: Crook 1966(1), patas 
monkeys (0.7 - 11.8 km: Hall 1965) and baboons Papio spp. (4.7-13.2 km: 
reviewed by Altmann 1974). P. anubis living in forest have shorter ranges 
(2.4 km: Rowell 1966) than ~n savanna. Other monkeys which are largely 
5.9 
or partly terrestrial also tend to have longer day-ranges than their 
arboreal relatives (reviewed by Clutton-Brock 1972). Apes appear to 
have a similar relationship. Orang-utans followed by Mackinnon (1974) 
had day-ranges averaging les 'S than 0.5 km, and siamang average 0.9 km 
per day (Chivers 1973). 
However, factors affecting day range clearly include factors 
other than terrestriality. Mountain gorillas, for instance, have median 
day-ranges of about 0.35 km (Harcourt, pers. comm.). One important 
factor is the density of food. Baldwin & Bald\"in (1972) considered that 
the relatively long ranges of squirrel monkeys Saimiri oerstedi were due 
to a low density of food; Clutton-Brock (1972) found a similar 
relationship in red colobus. Food distribution is also undoubtedly 
important (Altmann 1974), and changes in it may explain why chimpanzees 
travelled further ~n D2 than W. Thus although there is some correlation 
between day- range and ecological "grade" (Jolly 1972), it is clearly 
embedded in an adaptive complex of factors including locomotion costs, 
density and distribution of environmental resources and, in some spec~es, 
differential sexual strategies (e.g. orang-utans, MacKinnon 1974). The 
significance of the length of day ranges ~s therefore likely to be found 
by detailed studies within particular species. 
4. Circularity 
Individuals sometimes crossed their own path within a day. 
Differences in the frequency of doing so would reflect differences in 
ranging strategy which might have important ecological consequences. 
For instance, an individual was likely to cross his own path at a 
potential food source (VII/3/ii): the rate of doing so would therefore 
affect the probability of the source having more food than when he 
left it. 
I . 
I I 
5.10 
The number of times an individual crossed his own path after an 
interval of at least two hours was found for Northern males and Northern 
females. The data were divided into two periods, during which 
individuals tended to have s,hort (dry season 1972 and wet season) or 
long (dry season 1973) one-day ranges (Table 5.5). 
Table 5.5 
Frequency of cross~ng own path during one day 
Northern community only. Only paths crossed after at least two hours 
were scored. Kolmogorov-Smirnov two-tailed test. 
Period Sex Number of times path crossed N D P 
max 
0 1 2 3 4 5 
May '72-Apr '73 Male 29 24 7 3 0 0 63 0.16 Female 25 8 6 2 0 1 42 n. s. 
May '73-Sep '73 Male 15 7 2 1 0 0 25 ? n. s .i'~ Female 16 5 2 0 0 0 23 
Total Male 44 31 9 4 0 0 88 0.13 Female 41 13 8 2 0 1 65 n.s. 
* These data canngt be tested but the lack of significance is clear 
from inspection. 
The data are lumped across individuals: there is no reason ID 
thinl that individual differences would outweigh differences between 
sexes. Table 5.5 shows that there was no significant difference between 
sexes in the frequency of crossing their path in a day. It is also 
clear that there were no differences between time periods, within males 
or females. Season and sex, therefore, appeared not to affect the 
frequency of returns to food sources within one day. 
Ill. Four-Day Range 
1. Methods 
Group Travel charts completed between January 1972 and 
September 1973 were analysed. Observer reliability was discussed ~n 
11 
5.11 
Chapter 2. Travel paths of all individuals followed for four successive 
days were mapped: these individuals were normally target individuals, 
but some four-day ranges were extracted for individuals "Tho happened to 
have been with targets for four ' days. The maps are shown in Figs. 5.1 
to 5.3, and reference maps are shown in Fig. 5.4. The number of squares 
known to have been occupied during four days was taken as the size of 
the four-day range. Only squares greater than half-size were scored. 
Individuals and seasons were not sampled evenly: the choice of 
four days as a suitable time period was made in July 1973. Only limited 
statistical analysis can therefore be made. I am particularly grateful 
to M. Thorndahl for permission to use seven of the maps shown in Figs. 
5.1 and 5.2, and to C. Tutin, D. Riss and other observers for their 
valuable contributions. 
2. Factors affecting the size of four-day range 
~. Sex 
Northern male and female four-day ranges from Figs. 5.1 and 5.2 
were matched as closely as possible by date (Table 5.6): four-day 
ranges were significan~larger for males than for females (all females 
were anoestrous) (wilcoxon T = 0, N = 7, P < .02). 
Table 5.6 
Four-day ranges of Northern males and females 
Ranges were paired if recorded less than two months apart. 
MALES FEMALES 
Date Squares used Date Squares used 
FB 1-4 Apr '72 16 ML 23-26 Mar '72 9 
FG 21-24 Aug '72 21 WK 25-28 Sep '72 12 
FG 10-13 Nov '72 19 WK 14-17 Oct '72 12 
EV 10-13 Jul '73 26 ML 31 Jul-3 Aug '73 8 
EV 19-22 Jul '73 19 ML 4-7 Aug '73 10 
EV 5-8 Aug '73 26 ML 8-11 Aug '73 13 
HG 18-21 Sep '73 19 PS 8-11 Sep '73 14 
11 
III 
Fig. 5.1Four~day ranges of Northern males 
Individuals were targets except where shown in brackets. Maps are of 
adult males except for GB (xi), a juvenile/adolescent (nine years old) 
temporarily travelling without his mother. Reference maps and key are 
shown in Fig. 5.4. 
10-13 Jun 1972 vi (MK.FG.ST.JJ~ 
21-24 Jun 1973 
25-28 May1973 (~K. FB. FG. EV. ST) 
" _-:->/:--I:~2: 11-1;~ Jun 1973 -- ~ 
,I '- ...... 
'" ~ { 
EV 10-13 Jul 1973 
F S(+FG) 11 -14 Sep 1973 
v I ," ' 
L 
II 
Fig. 5.2 Four- day ranges of Southern males and Northern females 
Format as in Fig. 5.1. Maps i - iii are Southern males, rest Northern 
females. 
CH(+SF) 7-10 Apr 1973 
1'/ 
GI 13 -16 Aug 1972 
VII PS 20-23 Dec 1972 
< I' -i~, 
IX FT (+ FL, FG) 
14-17 Mar 1972 
III GI 13 -16 Aug 1973 
/ 
X GB (+ML) 23-26 Mar 1972 
iv WK 25-28 Sep 1972 
/ ''- I / lot 
GB (+ML) 
XII 31 Jut - 3 Aug 1973 
I ' . 
V WK 9-12 Oct 1972 
j~ 
VI WK 14-17 Oct 1972 
iii GB (+ MU 4 -7 Aug 1973 
~., .. ' .~ ... ) ..... ". 
xiv GB (+~L) 8-11 Aug 1973 
b 
Fig. 5.3 Four-day ranges of consort pairs, pregnant female and 
woUnded male 
Format as in Fig. 5.1. Maps i-vii show consort pairs (only on last day 
of Map vii). WK gave birth on 22 Oct 1972. FB's hand was wounded 
throughout the observation in Map ix. 
FG 22-25 Jan 1972 vii 
. - -
11 FG (+GG) . 26-29 Jan 1972 
/ 
III FG (+GG) 30 Jan - 2Feb 1972 
~ VIII WK 19 -22 Qct 1972 ~ . ./ 
iv PL (+ FB) 27 -30 Jun 1973 
~
V PL (+FB) "-4 Jul 1973 
~ IX 
vi PL (+ FB) ·5-8 · Ju I 1973 
'I 
I , 
I 
Fig. 5 . 4 
The reference maps show the valley systems (top) and habitat 
distribution in the most used part of the study area (bottom). All 
maps in other figures have the same grid lines as here, though some 
are shown at a different enlargement. 
REFERENCE MAPS 
-
Lake Ihore • Feeding Area (Camp) Y Stream Of gully 
0 2 1 N~ I I 
SCALE (Km .) 
KEYS TO DATA MAPS 
FOUR-DAY MAPS YEAR RANGES 
Known nest site In Rank 010 obsvn. points 
6 Nest not observed sq..Klre recorded in squ:re 
... ~ .. " Route on day 1 
__ day 2 
" day 3 
.............." "day 4 
:-:-:' Route not observed 
• Feeding area (Camp) 
Blank 
0 
• 
• 
• 
0 0 
1 ·01-1·0 
2 lO-2·5 
3 2·5- 5·0 
4 > 5·0 
I I 
---- -
5.12 
The comparison involves only four males and three females and 
may ·therefore be biassed in favour of particular individual tendencies. 
However, there were only two recorded four-day ranges of males which 
were smaller than those of a female. HG (Fig. 5.1, viii) had a small 
four - day range in the dry season 1972: not only were his one-day ranges 
small, but he tended to use the same paths on successive days. In this 
respect old males (HG, MK and GO) appeared sometimes to range more like 
anoestrous females than prime males: but if oestrous females were 
available, or if food was abundant, they appeared to behave more like 
prime males (e.g. HG, Fig. 5.1, ix). 
The second instance of a small male four- day range (other than 
for consortship or health reasons) was when FG accompanied his mother FL 
(Fig. 5.2, ix). FL died from old age five months after this record, and 
was already. very slow in her movements. This observation demonstrates 
the flexibility of male ranging patterns, and emphasises the importance 
of kinship ties into adulthood. Consort pairs had small four- day ranges 
(Fig. 5.3, i-vii), more similar to those of anoestrous females than males. 
(In Fig. 5.3, vii, only the final day was spent consorting.) 
1.1.. Conununity 
Only three four - day ranges of Southern males were recorded 
(Fig. 5.2, i-iii), and none of Southern females. The Southern male 
ranges were all smaller (10, 15, 12, respectively) than most Northern 
ranges, but the data are too few for statistical comparison. 
iii. Individual 
Four- day ranges 1.n Figs. 5.1 and 5.2 are grouped by individual. 
There were no clear differences, except between sexes, in the sizes of 
the ranges of different individuals; records taken at different times of 
year on the same individuals (e.g. Fig. 5.1, i-ii - iii and vii-viii) 
11 
5.13 
suggest that variance in four-day range s~ze was greater within than 
between individuals. 
~v. Season 
The only comparison with sufficient data for testing was between 
dry seasons in 1972 and 1973 for Northern males. FB's four-day range 
ending in a consortship on the final day is included (Fig. 5.3, vii). 
The data are shown in Table 5.7. 
Table 5.7 
Comparison of four-day ranges of Northern males between dry seasons 
June is included as a dry season month. 
1972 1973 
Male Date Squares used Male Date Squares used 
FG 10- 13 Jun 14 MK etc 11-14 Jun 25 
HG 24-27 Jun 11 FB FG 21-24 Jun 17 
FB 11-14 Aug 16 EV 10-13 Jul 26 
FG 21-24 Aug 21 EV 19-22 Jul 19 
EV 5-8 Aug 26 
FB FG 11-14 Sep 18 
HG 18- 21 Sep 19 
Ranges were significantly larger ~n 1973 than ~n 1972 (Mann-
Whitney U = 4, n l = 4, n2 = 7, P < .05). Though this finding ~s 
weakened by the small sample of males, it agrees with my impression 
during observation that not only were one-day ranges longer in 1973, 
but also the males travelled to more parts of the Park. For instance, I 
followed males across Linda Stream many times in 1973, but only once 
in 1972. 
v. Physical condition 
Fig. 5.3 shows examples of small four-day ranges due to unusual 
conditions of the target individual. FB's tiny range from 26-29 Aug 1972 
(Fig. 5.3, vii) was apparently due to an ~nJury to a finger on his right 
I .. L:..UJI __ 
5.14 
hand: he carried his paralysed right arm in his left throughout the 
observation. Fig. 5.3 (ix) shows WK's range around the time when she 
gave birth to her first infant, first seen on the morning of October 22. 
3. Centres of activity 
1. Males (not consorting) 
Fig. 5.1 shows that over four days males sometimes visited a 
particular place every day. In both his recorded four-day ranges, for 
instance, HG visited the artificial feeding area every day (Fig. 5.1, 
viii & ix). Similarly during two four - day periods in July 1973 Evered 
visited the same food source (Parinari curatellifolia, Linda Valley) 
every day (Fig. 5.1, xii & xiii), and in three days out of four in 
August (Fig. 5.1, xiv). Again, a Southern male, GI , visited the same 
place in Kahama Valley every day from 13- 16 August 1972, and one year 
later visited the same place on three days out of four (Fig. 5.2, ii & 
iii). Godi was ret~rning to a spot in Kahama stream- bed where a fig was 
in fresh leaf at the two periods sampled, and also where there were many 
trees suitable for nesting 1n. 
The maps show that chimpanzees made regular returns to 
temporarily important locations. During the study period the Northern 
males visited the feeding area more regularly than any other place. The 
abnormality of this behaviour appears to be in the length of time over 
which visits were made rather than in the principle of making regular 
returns. 
11. Females (not consorting) 
The four-day ranges of ML and GB, mostly collected by 
M. Thorndahl, show remarkable consistency (Fig. 5.2, x-xiv). In all 
five maps a similar area is covered, both in size and location, centred 
on camp but ballooning to the south. PS's range was similar (Fig. 5.2, 
5.15 
vii & viii), but she ventured further east. The maps on WK, organised 
by F. Plooij in order to observe the early hours of WK's first infant, 
are centred more completely on camp. These maps suggest that individual 
females may spend much of their .time travelling in the same fairly 
confined area. However, there are no four-day ranges available on 
females who did not regularly visit camp. 
LLL. Consort paLrs 
The data on FG and GG (Fig. 5.3, i - iii) and FB and PL (Fig. 5.3, 
iv- vi) were organised by H. Bauer and C. Tutin respectively. They 
clearly show that a consorting pair may travel to the edge of the most 
used area and stay there for several days. These two consortships were 
both in the north. FB also consorted with AT to the east of the most 
used area (Fig. 5.3, vii). A detailed description of consorting 
behaviour is in preparation by C. Tutin. 
4. Four- day range in relation to year range 
The four- day ranges of Northern males (Fig. 5.1) show that 
individuals often turned back towards the centre of the most used area 
when near its boundary. Furthermore, paths taken on concurrent days 
were often in different directions. A one-day range is clearly a poor 
approximation to the year range, but with four points to the compass 
the range over four days may predict the year range more fully. For 
instance, the northern, eastern, southern and western boundaries to 
EV's three four-day ranges in July and August 1973 were similar to each 
other. Four days may be a less adequate sample of the year range when 
individuals have short one- day ranges, as males did in the 1972 dry 
season (II/2/v). At that time the small one - day and f our- day ranges 
suggest that males tended to travel Ln small areas within the year 
range. The sample on anoestrous females (Fig. 5.3) also suggests that 
ln four days a female was likely to stay ln a small part of the year 
range (cf. Fig. 5.5). 
IV. Year Range 
1. Introduction 
5.16 
Few studies of chimpanzee ranglng are based on regular 
observations of known individuals: none have examined how the range of 
a community of chimpanzees is related to the ranges of the individuals 
within it. This problem is examined here by finding the year ranges of 
all independent individuals. 
The concept of home range lS an abstraction. The cumulative 
area occupied by an animal may be expected to increase with time, due 
both to its leaving one area in favour of another and to occasional 
extensions of its range in all directions. Since "home range" must be 
related to time, the term year range is used here to mean the total area 
used by an individual in a given year. 
A practical problem is that an observer cannot hope to record 
all areas visited by a chimpanzee even within a given time span. The 
observed year ranges are therefore necessarily smaller than the real 
year ranges. The observed year ranges include all the points \l7here 
each individual was seen, . and are an approximation to the real year 
ranges. 
2. Methods 
1. Analysis 
Group Travel charts were analysed. Locations were scored at 
30-minute intervals for all individuals present with a target. Although 
initially referred to a superimposed grid of 100 m. by 100 m. squares, 
locations were later lumped within 500 m. by 500 m. squares , i n order to 
5.17 
reduce the data. For comparison between species the grid squares used 
in analysis of range should ideally be a standard proportion of the 
cumulative range. Currently there is still considerable variation ~n 
the number of squares per range in different studies (e.g. 23 for 
siamang, Chivers, 1972, 138 for red colobus, Clutton-Brock 1972). The 
grid used here falls within these limits (e.g. 53 for Northern males). 
The data on map locations and individuals present were analysed 
with a computer program kindly written by H.C.M. Meekings. 
~~. Sampling problems 
The data were collected by twenty-seven different observers 
engaged in nine different studies: selection of target individuals, 
length of record and method of finding a party varied idiosyncratically 
between studies. Significant overall biasses were that the frequency of 
selection as target varied between individuals, that observation 
periods tended to , be longer on males than on females, and that 
observations tended to begin in camp. 
The importance of the frequent selection of certain individuals 
as targets ~s here reduced by us~ng the map records of all individuals, 
regardless of whether they were the target: the influence of 
observation time is examined in relation to the results: and ~n order 
to reduce the bias in favour of observations near camp, . only those 
records which were begun outside the feeding area have been analysed. 
Even so, records tended to start near the feeding area: observers 
looking for chimpanzees normally set off from the feeding area, or 
sometimes waited there until they heard calls. Though limited account 
can therefore be taken of the sampling problems, they impose considerable 
restrictions on the strengths of the conclusions to be drawn from these 
data. 
5.18 
~11. Statistical problems 
The data for all individuals are presented in Fig. 5.5 in a 
standard format of percentiles (see Fig. 5.4). Statistical comparison 
of percentiles is illegitimate because of varying observation times, 
but this is not possible anyway, for other reasons: the data are 
dependent because (a) successive observation points (every 30 minutes) 
are likely to be close and (b) a single observation of N individuals is 
counted N times, once for each individual. Consequently no statistical 
tests have been used. The maps are intended to show the existence and 
form of differences between individuals, rather than precisely what 
those differences were. 
3. Independent males 
1. Size of year range 
Sm~ll differences in size are apparent between the observed 
year ranges of the sixteen males (Fig. 5.5). The importance of 
observation time is shown by relating it to the number of squares 
occupied by each individual (Fig. 5.6, top). We expect an asymptotic 
curve if males share the same range, and Fig. 5.6 (top) suggests this 
may be true within each community. 
Factors affecting the observed Slze of range include: 
a. Sampling bias. The Southern males were recorded whenever 
they came to the northern edge of their range (around the feeding area: 
this explains why they had such high values for time spent in camp). 
The Northern males, on the other hand, had no boundary to their ranges 
where their chances of being observed was higher than in the centre. 
This suggests that the Southern male ranges would reach an asymptote 
earlier than the Northern males, since the data were biassed towards 
inclusion of some boundary squares. 
b. Manner of range use. Some squares used by one male were 
Fig. 5.5 Year ranges 
Year ranges are shown for all independent individuals ~n 
Northern and Southern communities. Reference maps and key 
are shown in Fig. 5.4. N = No. of observation points; 
the percentage falling in camp is shown. The red dot marks 
the artificial feeding area. Data source and explanation 
are given ~n text. 
The ranges are in five groups, within \vhich they are ordered 
(approximately) from more northern to more southern locations. 
Independent Northern males: left-hand column, ~-~x. 
Partially independent Northern males: left-hand column, x-x~. 
Independent Southern males: centre column, i-vii. 
Independent Southern females: centre column, viii-xii. 
Independent Northern females: right-hand column, i-xiii. 
vi 
viii 
ix 
x 
xi 
xii 
Fig. 5.6 Observed year range size in relation to number of 
observationpoirtts 
Squares had sides of 0.5 km. Observation points were scored at least 
30 minutes apart. The dashed line in the lower graph is illustrative 
only: all females who were oestrous (while not pregnant) at some time 
during the year lie above it. ' 
Data are from Fig. 5.5. 
50 FB • 
ST I-M EV • FG 40 • A • • 1-13 
.A • 
",¥:,E 
"'~'" JJ~ SH 30 SF GI 
20 '" HH 
..c"O ww .~ <IJ 10 MALES 
..c"O (.=N,"'=S) ~ 0 00 u 2 4 6 8 10 12 14 .~ <IJ 
'-
(/) ~ ~ 
0 ~ 50 ::) GG. 0'"- 40 sw • .- .-(/) 0 AT 
::) LBO, .~ 
,- ~ 
'0 "0 30 " PI. N~ • -ML :~ OGK" FF 20 .-
-0 "0 ~B ,- PS C o .- FEMALES z 10 o MD ... NV DO,," PL FL (-=N,o=S) 
0 
.-
0 2 4 6 8 10 12 14 
No. of observation points (x 100) 
Fig. 5.6 Observed year range size ln relation to number of 
observationpoirtts 
Squares had sides of 0.5 km. Observation points were scored at least 
30 minutes apart. The dashed line in the lower graph is illustrative 
only: all females who were oestrous (while not pregnant) at some time 
during the year lie above it. 
Data are from Fig. 5.5. 
50 FB A 
ST i-f,1 EV A FG 40 A A A A 1-13 AA A 
A~E A~A JJ"f< 
9-j 
30 SF Gi 
20 A HH 
..c ww 
"'0 MALES u (]) 10 '£. "'0 (A=N,A=S) ~ \.... 0 00 u 2 4 6 8 10 12 14 c (]) \.... 
~ ~ 
0 3 50 
:J GG. 0"- 40 If) 0 AT sw. ;' ;' 
:J LBO .ft'F ,,- fiM<. 
15 "'0 30 
, 
PI. NP. • ·ML :~ OGK, • FF 20 ;' 0 "'0 ~B ,,- PS C o ;' FEMALES z 10 o MD .... NV DO," PL FL (.=N,o=S) 
0 
;' 
0 2 4 6 8 10 12 14 
No. of observation points (x 100) 
11 j 
5.19 
not used by others. Examples are shown by FG (in the north) and FB (east) 
(Fig. 5.5). The areas uniquely recorded for each of these males were 
visited while consorting with adult females (GG and AT respectively, 
Fig. 5.3). The tendency for consort pairs to travel outside the normal 
range of either member means that the size of observed range is likely 
to increase suddenly, rather than gradually, with observation time (after 
reaching an initial asymptote). Of course, individuals may also travel 
outside their normal ranges for other reasons. GO, for instance, was 
seen ~n three squares (north of the feeding area) not visited by other 
Southern males: he associated more with Northern males than they did 
(Bygott 1974). 
The year ranges observed here were clearly not complete. 
However, their sizes appear to have approached an asymptote in relation 
to observation time. The largest sample of observation points was on 
Figan, who was seen in 48 squares (12 sq.km.). The full year range of 
a male ~s therefore likely to have been close to 12 sq.km. in area in 
1972. In 1973, however, the year range was larger: Northern males 
visited several squares in the north, not scored in their 1972 ranges. 
ii. Location of year range 
Bygott (1974) traced the relationships between the Northern and 
Southern males and showed from behavioural data that they could be 
regarded as belonging to separate sub- communities during 1971. In 1972 
almost all interactions between Northern and Southern males were 
agonistic, and they could be called separate communities. This result 
is confirmed from the geographical data (Fig. 5.5). Within each 
community the boundaries of the areas used by each male were fairly 
consistent. Furthermore, Northern males were mostly centred on 
Kakombe valley and Southern males on Kahama valley (cf. Fig. 5.4). 
The maps suggest that within each community the males all visited 
5.20 
essentially the same area. The area which includes all observations 
of any male within a community is henceforth called the "community 
range". It is also clear from Fig. 5.5 that there was a partial over-
lapping of the ranges of Nortl;lern and Southern males. 
The partial overlap may be the consequence either of animals 
having different central areas or of their having different peripheries 
to a single central area. The ranges of males within communities have 
been combined in Fig. 5.7, which shows that as the intensity of use of 
squares increased, the percentage of squares used by both communities 
fell. The Northern and Southern communities therefore had different 
central areas. Behaviour tending to maintain this difference is 
described below (VII). 
The establishment of the artificial feeding area ~n 1962 
apparently caused individuals to spend more time in a small part of 
Kakombe valley than they would otherwise have done. Its effect on the 
location of the community range may have been important during the 
years of unlimited banana feeding (up to 1967, Wrangham 1974), and may 
continue to be so. However, it now appears that a community of males, 
within which individuals share a community range, exists independently 
of the central location of an artificial feeding area. 
Map points recorded ~n camp were not assigned to a grid square, 
and are therefore not shown ~n Fig. 5.7. Northern and Southern males 
were both recorded in the feeding area on 7.6% of observation points 
(median score for each community). The observed extent of overlap in 
camp was exaggerated by sampling errors: although the data suggests 
that Southern males spent between 2.1. and 10.4% of their time in camp, 
the median number of days when a Southern male visited camp during the 
year was only 9. However, it remains true that the frequency of 
encounter was higher in camp than elsewhere. This shows that a 
Fig. 5.7 Male community ranges 
Key to maps (including rank of squares) is shown in Fig. 5.4. The 
artificial feeding area is indicated by the red dot. Median ranks of 
use of each square were taken within male communities (independent 
males: 9 Northern, 7 Souther n). 
Nand S MALES: Relation of OVERLAP OF 
YEAR RANGES to MEDIAN RANKS OF SQUARES 
Key: 
I;; N males' 
15"10 
+ S males 
5.21 
particularly attractive place may be responsible for a higher frequency 
of contact between communities than would be expected if the range were 
homogeneous. Similar phenomena occur in primates (e.g. patas monkeys at 
waterho1es, Struhsaker & Gart1art 1970, barbary macaques at wells, Deag & 
Crook 1971) and other animals (e.g. mockingbirds, Michener 1951). 
~. Independent females 
~. Size of range 
For several of the females the number of observation points was 
clearly too small to cover their full ranges (Fig. 5.5). Real differences 
~n s~ze of year ranges may be discerned, however, by examination of 
Fig. 5.6. The higher variance of number of squares recorded suggests 
that individual differences in range size were greater among females 
than among males. Factors contributing to this variance were: 
a. Location of range Females tended to occupy different 
parts of the community range (see below). Those who were most often ~n 
the most heavily sampled area were less likely to be recorded on their 
periphery than those which rarely visited camp. 
b. Oestrous state The dashed line on Fig. 5.6 is 
illustrative only, but happens to separate females who were oestrous 
during the year (above) from those who were not (below): the only 
exception was WK (below dashed line), but her oestrous swellings all 
occurred while she was pregnant. Females tend to travel more widely 
when oestrous than when anoestrous (McGinnis 1973, Tutin in prep.) . 
The differences between these two groups, in number of squares occupied 
in relation to number of observation points, were significant whether 
WK was scored as oestrous or anoestrous (Mann-Whitney U = 4 or 0, 
P < 0.001). (This measure ignores the expected flattening of the 
curve: but since curve shape should reduce differences ~n the measure, 
the fact that a clear difference between the groups exists makes its 
negligence irrelevant.) 
5.22 
There were three maLn ways in which females increased their 
observed year range as a result of being oestrous. They covered larger 
areas within the community range (e.g. NP, SW), they visited areas on 
the boundary of the community range when consorting (e.g. GG, AT) and 
they travelled in both community ranges. Temporary transfer of young 
females between community ranges has been recorded in Goffibe (Tutin, in 
prep; Pusey, in prep.) and Kasoge (Nishida & Kawanaka 1972): this LS 
seen here for LB, DO and GK (the latter being a Northern female 
originally) (Fig. 5.5). 
LL. Location of range 
As for the males, the year ranges of the .females have been 
ordered Ln Fig. 5.5, according to the northern or southern location of 
most used areas. The differences in location within each community 
range are clearer for females than for males. Sampling errors affected 
these differences, as they did for males. 
a. Dependence due to successive observation points meant that 
observation points were likely to be near each other. 
b. Many observation hours in one place (lower Kakombe) would 
reduce the impression of differences between ranges. 
c. Females who were rarely observed as targets were scored 
when with others. At this time (e.g. NP, when oestrous) she might be 
outside the area she used most when alone. 
These errors suggest that the observed differences may reflect 
the existence of real differences, but not their magnitude. That MF, 
for instance, occupied a more northern range than PS is based on a good 
observation sample: the difference between PI (east) and FL (west) is 
also large enough to be meaningful. 
5.23 
The best observation samples on anoestrous females were on 
those whose range was centred near camp; observers selected the females 
who were easiest to find. The year ranges of sometimes oestrous females 
were enlarged during their periods of oestrous (Tutin, in prep.). The 
data are compatible with the hypothesis that while anoestrous, females 
tended to occupy small ranges in different areas from each other. 
Furthermore, these areas are not always concentrated more in the range 
of one community than another. PS and NV, for example, had ranges 
concentrated in the area of overlap between the Northern and Southern 
community ranges. Females to the north and east of the Northern 
community range may well have similarly overlapped between community 
ranges, but this could not be observed because other communities were 
not habituated. The significance of this result will be discussed below 
(VI/3 and VIII) • 
S. Juveniles 
l. Size of range 
The only juveniles to travel far from their mothers during the 
observation year were FT and GB, both male. Their mothers (FL and ML) 
had small ranges, neither being oestrous during the year. The juveniles 
both had larger ranges than their mothers (Fig. 5.5) and GB's range 
approximated the Northern community range. FT left his mother less 
often than GB (Thorndah1, in prep.). 
Until they first start travelling away from their mothers, 
juvenile ranges are probably determined by their moth.ers' ranges. 
Though her year range may be expected to be only part of the community 
range during the offspring's first three to five years, she will then 
come into oestrous and travel more widely. The juvenile is normally 
still dependent at this time (van Lawick-Gooda11 1974). He is thus 
• I 
5.24 
likely to have had experience of the community range in the company of 
his mother shortly before he sets out on his own. 
LL. Location of range 
The core areas of FT and GB were similar to those of their 
mothers. The maps suggest that they tended to travel further, rather 
than have a dramatic shift of range, as they became independent. 
6. Comparison with other studies 
.It was pointed out that the "home range", meanLng the area 
which includes all points where an individual was observed (Brown & 
Orians 1970), must be related to time. This is especially important 
when describing animals as mobile as chimpanzees. Previous estimates 
of home range size and density are related to the number of observation 
hours in Table 5.8. 
Table 5.8 
Estimated size of "home range" (community range) and density in 
relation to observation time 
Densities are approximate, and Ln some cases are calculated from 
data given by the author. 
Type of observation 
Much indirect 
observation 
Few known 
individuals 
Mostly direct 
observation 
At least some 
known individuals 
Source 
Azuma & Toyoshima 1962 
de Bournonvi11e 1967 
Suzuki 1969 
Izawa 1970 
Kano 1971 
Jones & Sabater pi 1971 
Reyno1ds & Reyno1ds 1965 
A van Lawick-Gooda11 1968 
Sugiyama 1968 
A A1brecht & Dunnett 1971 
Suzuki 1971 
A Nishida & Kawanaka 1972 
A This study 
No. of 
obsvn. 
months 
10 
4 
15 
11 
11 
18 
9 
78 
6 
3 
17 
31 
12** 
*Low density is for savanna, high density for forest 
**Community range data were analysed over one year 
A Studies using artificial feeding 
Density 
( /km2) 
0.1 
0.05 
0.5-lc- 7.2* 
0.5 
0.2 
0.3 
4 
2 
5 
10 
4 
2.5-5 
2.5 
Ran~e 
(km) 
? 
? 
25-200 
120 
150-400 
? 
15-25 
39 
7.5 
5 
20 
10-17 
12 
! I 
11 
5 . 25 
Table 5.8 shows that the studies of home range and density of 
wild chimpanzees may be divided into two distinct groups, with low 
densities and large home ranges in the first, and high densities and 
small home ranges in the se'cond. (Home ranges here are equivalent to 
community ranges, which have . not previously been distinguished from 
individual ranges.) Furthermore, most of the studies in which large 
home ranges were found involved much indirect observation. It seems 
possible, therefore, that observational error may be involved. However, 
it is equally possible that the chimpanzees in these studies were indeed 
living at low densities, and that consequently observers found it hard 
to recognise and habituate individuals. Except for the work of Jones & 
Sabater pi (1971) (lowland forest, Rio Muni) these studies were of 
chimpanzees living on the edge of the geographical range in highly 
seasonal habitats, where low densities might be expected. 
The study in this group with the most observation of recognised 
individuals was -by Izawa (1970). ' Even here, however, movements were 
monitored for much of the time by listening for the calls of large 
parties, and when no calls were heard the chimpanzees were thought to 
have moved. This may not be true, as pointed out by Kano (1971), since 
it is possible that the absence of calling was the result of dispersion. 
Thus the evidence for very large home ranges of chimpanzees remains 
equivocal. 
Studies In the second group indicate community ranges of 
between 5 and 40 sq km, averaging about 12 sq km. Four out of the seven 
studies used artificial feeding, which could possibly have influenced 
the size of the comunity range. No controlled studies are available to 
test this point, but range size was similar in studies without 
artificial feeding where chimpanzees were regularly identified (all in 
the Budongo forest, Uganda). Furthermore it is clear that individuals 
, ' 
5.26 
may range independently of an artificial feeding area even after they 
have become habituated. For instance, Nishida & Kawanaka (1972) studied 
a community which migrated away from an artificial feeding area 
regularly every November, returning ~n February. Again, the Southern 
males in Gombe received bananas regularly since 1963, but have now 
stopped visiting camp, even though they received bananas more often per 
visit than Northern males (Bygott 1974). Artificial feeding may 
therefore be a small influence on the size of the community range. 
Table 5.9 suggests that the typical community range size of chimpanzees 
may be between 10 and 20 sq km. 
It is interesting to note that this home range s~ze accords 
with the predictions of McNab (1963). McNab found that in mammals home 
range size was related to body weight by the equation (for frugivores 
and carnivores) R = 0.051 x (WO. 7l ) , where R = range size (sq km) and 
W = weight (kg). For an individual chimpanzee weighing 40 kg, R is 
predicted to be 0.69 kg, and for 20 chimpanzees sharing a community 
range, as ~n Gombe, the community range is predicted to be 
20 x 0.69 = 13.8 sq km. 
studies in Table 5.9. 
This is close to the results from some 
V. Directions of Entry into Camp 
1. Methods 
~. Data source 
The data on year ranges suggest that individuals may be 
dispersed within community ranges. The artificial feeding area (camp) 
was located near the centre of the Northern community range (Fig. 5.7); 
the direction by which individuals entered camp might therefore reflect 
differences in range use within the Northern community. This is 
examined here. 
.' I: A 
~27 
The artificial feeding area was roughly circular, with a radius 
of about twenty metres. Within this area the vegetation was regularly 
cut in order to facilitate observation. There was therefore a fairly 
clear boundary to most of the area. 
The directions by which all chimpanzees entered or left camp 
were recorded by observers as part of the long- term programme of the 
Gombe Stream Research Centre. The boundary was considered to be 
divided into eight sections of approximately equal length: these were 
labelled North, North-East, etc. In each section there was at least one 
well -worn trail. However, these were not equally favoured by 
chimpanzees. The east path, for instance, was large and straight, and 
allowed individuals a view of camp before they actually entered. This 
was perhaps why individuals approaching camp from the north- east or 
south-east sometimes went out of their way to enter by the east path. 
The simplicity of the data meant that observations were highly 
reliable. Observers were trained before recording alone, and were 
regularly checked. A possible source of error was confusion about the 
exact location of the boundaries between compass point sections, but 
such discrepancies were small. 
Individuals selected for analysis were all those who were 
independent and entered camp at least 5 times in all periods sampled. 
LL. Analysis 
Matrices were prepared showing how many tim~ each independent 
individual used each direction for his first entry of the day. Only 
the first entry of the day was scored since this was most likely to 
reveal differences between individuals. The matrices were constructed 
separately for males and females. Two analyses were made. 
First, for each matrix a mean direction of entry was found, 
together with a measure (r) of how concentrated the scores were ~n 
that direction. Methods of dealing with circular distributions were 
given by Zar (1974). The mean direction (;) of a number of angles a. 
~ 
(i = 1 to N) ~s calculated from 
N 
Sum (cos a.) 
i=l ~ y 
tan a. = where X X ' N and 
N 
Sum (sin a. ) 
.i =l ~ Y N 
The measure of directedness, or concentration, 'rI, is g~ven by 
r = J x2 + y2. A high value of 'r' shows that scores tended to be 
clumped around the mean angle. I;' and 'r' were calculated from the 
5.28 
mean percentages of entries made in each direction by each individual. 
Second, the data were "normalised" in order to reveal 
differences between individuals. For each individual in the matrix 
the percentage of total entries made in each direction was calculated, 
and subtracted from the mean percentage in that direction. The 
difference was added to 50 in order to give positive scores ~n all 
directions. This "normalised" matrix has no mean angle and no 
directedness. However, each individual in the "normalised" matri x has 
scores which allow calculation of his mean angle and directedness. 
These were calculated by the method above. 
~~~. Interpretation of results 
The following problems arise from the nature of the data and 
the method of analysis. 
a. Because the directions of entry were not all equally 
favoured, the value of 'r' was affected by the mean angle. For 
instance, the \vest section was used little at all times. When the 
overall tendency was for individuals to enter from the east, overall 
directedness would be relatively high. When the overall direction was 
for individuals to enter from the west, but in fact tended to use the 
north- west or south- west paths, 'r' would be relatively low. However, 
I 
11 
5.29 
since there was considerable spread at all times ~n the directions used, 
this was probably unimportant. 
b. The value of 'r' for individuals in a "normalised" matrix 
~s affected by the number of individuals. In the comparison of males 
and females, the sample size was 9 and 11 respectively. In order to 
test the effect of sample size the calculations whose results are shown 
were repeated after o~ss~on of two females (GG and SW). The effect 
was to increase the value of 'r' for some females and decreas e it for 
others. Statistical comparisons of 'r' values between seasons and sexes 
were made both with and without the data for GG and SW. There was no 
difference in the results. Only the results using the full matrix for 
females are g~ven below. 
c. The' "normalised" mean angle for an individual depends on 
the mean angles of the other individuals in the same matrix. If one 
individual changes his mean angle, this change affects all others. A 
circumstance in which this might have occurred ~s when a mother became 
oeastrous. Even if one female did not change her mean direction of 
entry over time, she could appear to have done so as a result of a 
change by another. Inspection of the results indicates that this was 
not an important effect. 
d. The value of 'r' for individuals in a "normalised" matrix 
is affected by the overall value of 'r'. When all individuals tend to 
come in from one' direction, the overall value of 'r' is high. 
Differences between individuals are then relatively low, and individual 
values of 'r' are low. The more "normalisation" that occurs, the 1mver 
the individual values of 'r'. 
In these data a high overall value of 'r' means that many 
individuals tended to enter from a certain direction. This may mean 
that camp was not then in the centre of the foraging area: it was then 
I I 
a less suitable location for sampling differences between individuals 
than when overall directedness was low. However, the observed 
individual values of 'r' may be compared between seasons by taking 
account of overall values of '.r'. 
2. Overall directions of entry 
5.30 
Data were selected to correspond to Dl , Wand D2 • At the time 
of analysis the data from only two months in each period were available 
(Jul-Aug 1972, Dec 1972 & Jan 1973, Jul-Aug 1973). 
The data were analysed as described above, and the results are 
shown in Fig. 5.8. Only the large black arrows outside the circles are 
relevant. 
The mean angle of entry is shown by the direction of the large 
arrow. In the two dry seasons the overall tendency was for individuals 
to enter from the eastern sector. It seems likely that this represents 
the direction in which most food was found. Several of the dry season 
foods (e.g. P. curatellifolia, B. bussei) were largely confined to the 
upper valleys, generally east of camp. 
In the wet season the females again showed an overall tendency 
to come in from the east. However, this was less marked than ~n the 
dry seasons: the mean angle was not highly directed towards the east 
(low 'r' value is shown by the shortness of the arrow). Males also 
showed more angular dispersion than in the dry seasons, and an overall 
tendency to enter from the north-west. The results suggest that in the 
wet season camp was more nearly in the centre of the range of males and 
females than in the dry seasons. Many foods were common in the lower 
valleys (near camp) in the wet season (e.g. A. senegalensis, 
B. capparidifolia). 
11 
Fig. 5.8 Directions of entry into camp: seasonal differences 
"Normalised" mean angles of entry are shown within the circles (north 
to the top): angles are consistent across seasons for many individuals. 
The heavy arrow outside the circles shows the overall mean entry 
direction for each season and sex. Circles are visual aids only. 
The lengths of all lines are proportional to 'rI, a measure of the 
concentration of scores about the mean angle. See text for further 
explanation. 
Females who were oestrous during the observation period are marked with 
an asterisk. 
Number of entries: Males 350 (D 1), 283 (W), 368 (D2); Females 351 (D1), 
293 (W), 347 (D2). 
Jul-
Aug 
1972 
Dec 
1'iJl2-
Jan 
1973 
Jul-
Aug 
1973 
FEMALES 
Fig. 5.9 Direction of entry into camp: seasons combined 
Format and data from Fig. 5.B. The figure shows that females were 
more directed (higher values of 'r') than males. 
MA~S FEMALES 
5.31 
3. Mean angles of individuals 
~. Males 
The "normalised" mean angles of each individual were calculated 
as described and the results' are shown in Figs.5.8 and 5.9. In Fig.5.8 
the three periods are treated separately. 
Individual males show considerable consistency ~n their 
"normalised" mean entry directions. From all three periods the 
directions of six males (MK, HG, HM, EV, FG, ST) fall within 90 0 
(though each male is in a different direction). Those whose directions 
show less consistency (FB, JJ, SH) had at least one period in which 
directedness was very low. No male showed a strong tendency to enter 
by two dissimilar directions. 
The data therefore show that different males tended to enter 
camp by particular directions which are stable over at least one year. 
This could mean that different males tended to use different parts of 
the community range, or alternatively that they tended to enter camp 
from a particular direction even though they used all areas equally. 
Three lines of evidence are relevant. 
First, the year ranges (Fig. 5.5) may be compared with Fig. 5.8. 
Although the boundaries of the year ranges of different Northern males 
were similar, variations in use of particular parts are discernible. 
For instance, FG's year range was more concentrated to the west than 
any other male's (Fig. 5.5). Similarly, Figs. 5.8 and 5.9 show that 
FG's "normalised" mean entry direction was western. Again, HG shows a 
more south-western tendency than other males both in his year range 
(Fig. 5.5) and his entry directions (Figs. 5.8, 5.9). It is 
particularly instructive to compare the year ranges (February 1972 to 
January 1973) with entry directions in the period Jul-Aug 1973, since 
the data were collected in different times (Fig. 5.10). SH is the only 
Fig. 5.10 Comparison of entry directions and year ranges: Northern males 
Entry angles and directedness are taken from Fig. 5.8, Jul-Aug 1973. Data 
on grid square occupation from Fig. 5.5 (only squares of rank 4 are used), 
Feb 1972 - Jan 1973 (N.B. that here north is to the top). Although the 
data are from different periods and different sources, some correlation 
is apparent. The angles of entry for all males except SH point more ~n 
the direction of their most used grid squares than they would if 
different by 180 0 • The angles shown therefore appear to represent real 
differences in range use between males within the community range . 
~. 
•• SH 
: 
•• 
..... FG 
• HG 
~. 
• 
• 
• 
/ 
• ~ .. 
MK 
JJ 
•• fe· 
,..  
ST. 
EV , •• 
•• 
HM • ~ 
Grid squares (500 m x 500 m) having 
50/0 of obsvn. pis. (Feb '72 - Jan '73) 
Cclrrp 
Mean "normalised" direction of entry 
into camp (Jul - Aug '73) 
S.32 
individual whose direction of entry does not appear to conform with 
the most used area of his year range. He was also the only male who 
was not a frequent target for observation. The consistency shown by 
other individuals is a strong argument in favour of the suggestion 
that "normalised" entry directions represent the direction from camp 
of the area most used by each individual. 
Secondly, during the 1972 dry season there were two recorded 
four-day ranges of males which were smaller than those recorded at any 
other time. These four-day ranges (FG, Fig.S.1(i), HG, Fig.S.1(viii» 
were to the west and south of camp respectively. Fig. 5.9 shows that 
FG's overall "normalised" mean angle was to the west, and HG's to the 
south-west of camp. The direction of entry into camp conforms to the 
area used by a male when he confined his travel to a small part of the 
community range. 
Thirdly, my experience of encountering males travelling alone 
indicated that some ~a1es were most likely to be found alone in a 
particular part of the community range. The clearest example was MK. 
On six occasions when I met MK while observing other males, he was in 
Upper Kaseke1a or Linda Valley. On three separate days of observation 
when mostly alone, he travelled in these valleys. I therefore 
characterised ~{ as having a preferred range in the northern part of 
the community range. The data conform with this suggestion (Fig.S.10). 
We may conclude that the "normalised" mean angles of entry 
direction represent real differences in ranging tendency of males 
within the Northern community. This is the first evidence that within 
a community different males have different pre"ferred areas. 
1.1.. Females 
The "normalised" mean angles of females show even stronger 
consistency of direction between periods than those of males. Only one 
5.33 
female (AT) shows an angular spread across three periods greater 
than 90 0 • The probability of obtaining an angle less than or equal 
to 900 from two lines at random is 0.5. Only one angle out of a 
possible 33 (11 females in 3 se'asons) exceeds 90 0 : individual females 
were therefore clearly consistent in their direction of entry. 
Furthermore, as in the case of males, entry directions appeared to 
represent the direction of the area most used by each female. 
First, the year ranges of females reveal differences in ranging 
tendency which generally conform to the differences in entry direction. 
Fig. 5.5 shows that the four Northern females with the most southerly 
ranges were ML, PL, PS and NV: Figs. 5.8 and 5.9 show the same. There 
are similar correlations in other directions. The only female whose 
directions of entry did not conform to the recorded year range is NP. 
This is particularly surprising since NP's entry directions were both 
consistent (to the north) and highly directed. Observation problems 
appear responsible for this anomaly. NP is a shy female who was 
difficult to observe when alone. Most of the observations were 
therefore ' made when she was with large parties after visiting camp or 
in the period between being seen at or near camp and when she lost the 
observer as she travelled north or east. 
Second, there were four-day ranges recorded on three females 
whose entry directions are shown. Fig. 5.2 shows that ML's and PS's 
four-day ranges were concentrated to the south of camp: their entry 
directions were also southern (Figs.5.8, 5.9). WIZ's four - day ranges 
(Fig.5.2) were centred on camp but show no clear concentration in any 
direction. WK's "normalised" mean entry direction was north-western 
(Fig.5.9), but more importantly she showed less overall directedness 
than any other female. This is expected if her most used area is 
centred on camp. 
I. 
5.34 
Third, females encountered alone during the study period were 
~n the areas expected from Fig. 5.9. For instance, NP, MF and GG were 
never encountered alone south of Kakombe stream, and most often in 
Upper Kasekela or Linda Valley. Similarly ML, PL, PS and NV were 
typically encountered alone in Lower Kakombe or Mkenke valleys. 
As in the case of males, the data on "normalised" mean entry 
directions indicate consistency of direction over time, and the 
directions of individual females conform to the areas they used most. 
Change in oestrous state between periods was not a strong 
influence on entry direction (Fig.5.8, NP, AT, MF, PL). This is 
expected, even if, when oestrous, females travelled more widely in the 
community range. When travelling more widely they would be expected 
to enter camp more equally from all directions, but in the two-month 
sample they would also have been anoestrous. The mean angle should 
therefore be stable between periods, whether or not females were 
oestrous or anoestrous. 
iii. Sex differences 
Statistical techniques for comparison of angular deviations 
are not available. However, there ~s some suggestion that males were 
less consistent than females ~n their entry directions. Six out of 
o 27 angles for males (3 seasons, 9 males) were greater than 90 , 
compared to lout of 33 for females. Between the two dry seasons the 
angular deviations of males were greater than those of females 
(Mann- Whitney U = 19, n l = 9, n2 = 11, P < 0.05). There was no 
difference between the wet season and either dry season. It is not 
clear how this difference may have been affected by the number of 
individuals contributing data to the "normalised" matrix. Although 
females appear more consistent than males, further analysis is 
required. 
4. Seasonal variation 1n directedness 
1. Males 
In Figs. 5.8 and 5.9 the value of 'r' (directedness) for each 
individual is given by the length of the line. Thus in Jul-Aug 1972 
HM was the most directed male, and JJ and SH were the least directed. 
There were no significant differences between any pairs of 
seasons in values of 'r' (Wilcoxon, T = 14 (Dl-W), T = 16 (D l -D2), 
5.35 
T = 22 (W-D2), N = 9, n.s.). However, directedness was higher in Dl 
than in D2 or W. This is not expected on the basis of overall direction 
tendency, which was higher 1n Dl than in the other seasons. Since the 
process of "normalisation" tends to decrease 'r' values when overall 
direction tendency is greater (V/l/iii), the data indicate that males 
were relatively more directed in Dl than in D2 or W. 
Older males appeared to be more directed than younger males. 
This was significant only in Dl (Spearman rs = 0.88, N = 9, P < 0.01). 
This may be a chance result, but it is possible that young males were 
then travelling more widely about the community range than old males. 
11. Females 
Females were more highly directed in W than Dl (Wilcoxon 
T = 4, N = 11, P < 0.01). Differences were not significant between 
other seasons (r = 13 (D l -D2), T = 12 (W-D2), N = 11, n.s.). The 
relative directedness between seasons 1S as expected from the overall 
direction tendencies. Overall directedness was highest in Dl and 
lowest in W, and the directedness of individuals was lowest in Dl and 
highest in W: high overall directedness imposed a reduction of 
differences due to the "normalisation" process (V /l/iii). Thus it 
seems that females were equally highly dispersed in all seasons, but 
that when the position of camp was relatively further from the main 
• I 
5.35 
4. Seasonal variation 1.n directedness 
1.. Males 
In Figs. 5.8 and 5.9 the value of 'r' (directedness) for each 
individual is given by the length of the line. Thus in Ju1-Aug 1972 
HM was the most directed male, and JJ and SH were the least directed. 
There were no significant differences between any pairs of 
seasons in values of 'r' (Wi1coxon, T = 14 (D1-W), T = 16 (D1-D2), 
T = 22 (W-D2), N = 9, n.s.). However, directedness was higher in D1 
than in D2 or W. This is not expected on the basis of overall direction 
tendency, which was higher 1.n D1 than in the other seasons. Since the 
process of "normalisation" tends to decrease 'r' values when overall 
direction tendency is greater (V/1/iii), the data indicate that males 
were relatively more directed in D1 than in D2 or W. 
Older males appeared to be more directed than younger males. 
This was significant only in D1 (Spearman r . = 0.88, N = 9, P ( 0.01). 
, s 
This may be a chance result, but it is possible that young males were 
then travelling more widely about the community range than old males. 
1.1.. Females 
Females were more highly directed in W than D1 (Wi1coxon 
T = 4, N = 11, P < 0.01). Differences were not significant between 
other seasons (T = 13 (D1-D2), T = 12 (W- D2), N = 11, n.s.). The 
relative directedness between seasons 1.S as expected from the overall 
direction tendencies. Overall directedness was highest in D1 and 
lowest in W, and the directedness of individuals was lowest in D1 and 
highest in W: high overall directedness imposed a reduction of 
differences due to the "normalisation" process (V/1/iii). Thus it 
seems that females were equally highly dispersed in all se.asons, but 
that when the position of camp was relatively further from the main 
+ 
5.35 
4. Seasonal variation 1n directedness 
1. Males 
In Figs. 5.8 and 5.9 the value of 'r' (directedness) for each 
individual is given by the length of the line. Thus in Ju1- Aug 1972 
HM was the most di~ected male, and JJ and SH were the least directed. 
There were no significant differences between any pairs of 
seasons in values of 'r' (wi1coxon, T = 14 (D1-W), T = 16 (D1-D2), 
T = 22 (W-D2), N = 9, n.s.). However, directedness was higher in D1 
than in D2 or W. This is not expected on the basis of overall direction 
tendency, which was higher 1n D1 than in the other seasons. Since the 
process of "normalisation" tends to decrease 'r' values when overall 
direction tendency is greater (V/1/iii), the data indicate that males 
were relatively more directed in D1 than in D2 or W. 
Older males appeared to be more directed than younger males. 
This was significant only in D1 (Spearman rs = 0.88, N = 9, P < 0.01). 
This may be a chance result, but it is possible that young males were 
then travelling more widely about the community range than old males. 
11. Females 
Females were more highly directed in W than D1 (wi1coxon 
T = 4, N = 11, P < 0.01). Differences were not significant between 
other seasons (T = 13 (D1-D2), T = 12 (W-D2), N = 11, n.s.). The 
relative directedness between seasons 1S as expected from the overall 
direction tendencies. Overall directedness was highest in D1 and 
lowest in W, and the directedness of individuals was lowest in D1 and 
highest in W: high overall directedness imposed a reduction of 
differences due to the "normalisation" process (V/1/iii). Thus it 
seems that females were equally highly dispersed in all seasons, but 
that when the position of camp was relatively further from the main 
1,..;..1 
+ 
5.36 
food sources, females showed lower directedness on entering camp. 
lll. Sex differences 
In Dl there was no difference in the directedness of individual 
males and females (Mann- Whitney U = 47, n l = 9, n 2 = 11, n.s.). In 
both the other seasons (Fig.5.8) and in the data lumped across seasons 
(Fig.5.9), females were significantly more directed than males (W: U 6, 
P < 0.001 ; D2 : U = 19, P < 0.05; lumped seasons: U = 18, P < 0.05). 
The higher directedness of females indicates that they were 
more likely to come in from a particular direction than males. It was 
argued that the tendency to enter from a particular direction reflects 
a real tendency for an individual to use a particular area. The data 
therefore suggest that in Wand D2 males travelled more widely than 
females over the community range. This was not true in Dl . 
In Dl food appeared to be less abundant than in W or D2 
(Chapter 3) anft individuals travelled in smaller parties. Their greater 
dispersion in Dl is indicated by the result that this was the only 
season when individual females were not significantly more directed 
than males. The data therefore suggest that the formation of smaller 
parties was correlated with a greater tendency for individuals to adhere 
to particular parts of the community range. This is discussed below 
(VIII/2). 
VI. Interactions between Individuals of Different Communities 
1. Introduction 
From data collected up to December 1971 Bygott (1974), like 
Izawa (1970) and Nishida (1974), found evidence of avoidance between 
males of different sub-communities and communities. However, there was 
some peaceful association between males and females. Bygott's results 
are confirmed and extended here. 
5.37 
2. Between males 
~. Types of interaction 
The variety of interactions seen between Northern, Southern 
and unhabituated males during this study was similar to that reported 
by Bygott (1974), and included peaceful association, avoidance, chasing 
and attacking. The following notes add to previous observations. Where 
parties were heard calling it was assumed that they included males. 
a. Monitoring behaviour On fourteen occas~ons parties 
travelled to ridges on the edge of the current community range, and sat 
without feeding: they listened into the valleys beyond (Plate 5.1). 
If they heard calls their reaction appeared to depend on the size of 
the 'stranger' party, as judged by the calls. Twice, after hearing a 
single call, parties advanced showing signs of excitement, only to 
retreat in confusion a few seconds later when a further volley of calls 
showed that a large party was present. On both occasions the observed 
party retreated well into the community range and split up. Three times 
when several Northern males heard a call of a single Southern male they 
advanced and chased or attacked the 'stranger'. Similar behaviour was 
seen when a Northern or Southern party heard a 'stranger' without 
having sat and listened. The fact that parties sometimes travelled to 
ridges without eating en route, and then returned after listening for 
30 minutes or more, suggested that they were monitoring the boundaries. 
Prime males were invariably involved. 
Another form of monitoring appeared to be the investigation of 
nests found on the edge of the community range. This was seen twice 
during the study, and a third time subsequently (Riss, pers. comm.). 
In the first observation Evered was travelling alone when he saw two 
nests high in Kakombe valley. He climbed rapidly up to one, smelled it, 
and then began a branching display which ended with his standing in the 
nest, flailing his arms through and thus destroying it. He also smelled 
Plate 5.1 Monitoring behaviour 
A Northern party on the edge of its community range (Rutanga valley, 
18 September 1973). The trees have no food, and were apparently used 
only to provide a view. Figan, the most high-ranking male, led the 
party and sat at the highest point. Though no calls had been heard, the 
chimps sat and scanned north for ten minutes, then heard calls of 
'strangers' . 
S.38 
inside the other. On the second occasion Humphrey and Evered also 
found nests high in Kakombe (Fig.S.ll), and they climbed up and each 
spent several seconds bending into the nest and smelling. Observers 
had seen two unidentified chimpanzees near this point on the previous 
day. The unusual reactions to nests found on the edge of the range 
suggests that chimpanzees were able to recognise each other's smell and 
hence learn about the movements of individuals of other communities. 
Monitoring behaviour between groups ~s shown by some other 
primates (Barbary macaques Macaca sylvanus, Deag 1973) but not all 
(rhesus macaque M; mulatta, Vessey 1968). Like chimpanzees, spotted 
hyaenas form variable parties but only within a certain set ('clan') of 
individuals. Hyaenas show monitoring behaviour similar to chimpanzees, 
but, as in the case of many other carnivores, they are able to mark 
the ground with glandular secretion (Kruuk 1972). Wolves appear to 
behave similarly (Mech 1970). Monitoring or patrolling behaviour is 
probably more common in species which mark, since it is more likely 
to be rewarding. 
b. Retreating, chasing or attacking Observations supported 
Bygott's conclusion that the tendency to retreat or advance when one 
party heard strangers depended on the relative size of the parties. It 
was also noted that prime males were the most important component of 
the party. There were three good observations of the response of a 
large bisexual party to the calls of a smaller party, and in all the 
first to arrive at the 'strangers' site were the prime males. 
Although this may partly reflect motivational differences 
between individuals, physical condition was also important. Thus Faben, 
with a disabled arm, was always the last of the prime males to arrive 
after a party ran together. The oldest Northern males, Hugo and Mike, 
were also slow, and tended to arrive at the interaction at the same time 
5.39 
as females, after the 'strangers' had retreated. 
c. Peaceful associations Since January 1972 the only 
peaceful interactions observed between males of different communities 
have been between the old Northern male, Rugo, with Southern males, and 
between the old Southern male, Goliath, with Northern males. (An 
exception is that Southern males occasionally came to camp during 1972, 
and associated briefly with Northern males after an aggressive 
encounter.) In these interactions the old male joined a feeding party 
of the other community, but was not attacked. Possibly the interactions 
were peaceful because of the long-term relationships involved (Bygott 
1974). An alternative explanation is that a patently old male poses 
less threat than a prime male as a reproductive competitor. 
Bygott (1974) described an observation by G. Teleki ~n which 
prime 'stranger' males were seen in the company of Southern males. No 
similar examples havereen seen, but the observation below is suggestive. 
On 18 September 1973 a party of 7 males and 4 females (two 
with young) heard calls of 'strangers' in Rutanga. All listened but 
did not reply. An hour later the female GG left the party and walked 
in the direction of the calls. This was not remarkable: females may 
associate with more than one community (below). Curiously, however, 
she was followed not only by the other females but also by Faben and 
Evered. Watching them go, the remaining males started to walk in the 
opposite direction, back towards the centre of the community range. 
Figan was the last to leave. 
During his first ' minute of walking Figan stopped and turned 7 
tim~ to look at the party walking towards the strangers: he seemed as 
surprised as I was. During the third minute of walking, while standing 
bipedally and looking back . again, Figan saw Faben stop, turn and walk 
back from the females: Faben joined Figan, they "hooed" to each other, 
Fig. 5.11 Four-day range (detail) 
This example is shown also ~n Fig. 5.1 (xii). The key is shown on 
the figure: 'p' and 'e' were food sources accounting for 
approximately 60% of feeding time. Interactions apparently involving 
members of other communities (A ,and B) occurred on the boundaries of 
the four-day range, which approximately covered the community range. 
FOUR - DAY RANGE : PRIME MALE (Evered) 
10 - 13 July 1973 
Oay 1 
2 
3 - - -
4 .......... . 
Nests A 
A Smelled nests 
8 Heard 'S1 rangers' 
p~ 
e ~ 
........ ~ 
\;( 
1::::::::," 
N E 
Scale: taR:roxfflate) 
o 05km 
I I, I I I 
,. - . .". ii-' _.- ..... 
(i 
10km 
I 
Plate 5.2 Reassurance behaviour on the edge of the 
con'unurtity range 
(a) In Rutanga valley Evered mounts Figan after hearing 
the calls of 'strangers'. The calls were very distant, 
and the chimps listened intently and showed further 
reassurance behaviour (e.g. hand-holding) for 8 minutes. 
They heard nothing more, and continued feeding. 
August 1973. 
(b) In the same p lace, and again eating P ~ curatellifo1ia 
fruit, Figan grasps Evered's erect penis after hearing 
a stick snap. They continued to listen, heard nothing 
more, and resumed eating. July 1973. 
(a) 
(b) 
5.40 
and then continued back towards the centre of the community range. 
Evered, however, had stayed with the females. His behaviour ~vas unusual 
in two ways: first, it was normally only in a party with several males 
that a male went to the edge of the community range, yet by the time he 
was out of sight he was already further north than any party I had seen 
in three months; second, after hearing the calls of strangers a party 
normally advanced only if it appeared to contain more males than the 
stranger party. 
Evered's behaviour on this occaSlon seemed not to be consistent 
with a model of permanent aggression between males of different 
communities. Evered had "disappeared" at times. For instance, he was 
not seen from 21 August 1972 to 2 October 1972. Evidence of males 
transferring temporarily between communities may yet be obtained. It 
lS possibly relevant that Evered's mother, who died in 1969, appeared 
to have a Northern range. If Evered grew up on the boundary of two 
communities it is conceivable that he could still associate with both 
of them. 
ii. Location of interaction 
Bygott (1974) found that 'stranger' interactions tended to 
occur on the edge of the community range, and this result was confirmed 
(Fig.5.ll). Additionally it was noted that the response to calls or 
other noises depended on where they were heard. When a party well 
within its community range heard 'stranger' calls the reaction was 
violent and positive: they immediately called and ran in that 
direction, some individuals stopping briefly to touch or embrace others. 
On the other hand, parties in an area used more by another than their 
own community showed passive and timid responses. Plate 5.2 shows the 
response of Northern males In upper Rutanga valley to hearing the 
distant calls of strangers or to hearing a stick breaking. 
1 -
5.41 
Similarly when a party in lower Rutanga heard the footsteps 
of an animal they 'froze' for 6 minutes. If any of them had moved they 
would have been audible because the ground was covered in dead leaves. 
After two minutes Figan very s'lowly crept forward by treading on stumps 
and grasping saplings 20 cm off the ground. Eventually he saw that the 
mystery animal was a lone baboon, and the others relaxed. It is 
interesting that the one male who was not worried by the noise was the 
old male Mike: he continued to eat and shift position, rustling leaves 
as he did so. Within the most used part of the community range 
individuals normally ignored the sound of snapping sticks or footsteps 
of other animals. 
iii. Frequency of interaction 
When four - day ranges tended to include visits to several 
boundaries of the community range, interactions were common. Fig. 5.10 
is an example of a four-day range in which a male visited three 
boundaries. At one he smelled nests, at another he joined others ~n 
apparently chasing a Southern male; at the third his party heard calls 
of an unhabituated community and retreated. 
The frequency of interactions in three seasons ~s shown in 
Table 5.9, which includes interactions with females. 
'Stranger' interactions occurred in all seasons sampled: the 
data suggest they were more frequent in D2 than at other times. This 
is expected from the larger community range (and one-day and four - day 
ranges) at this time, when individuals of neighbouring communities 
,,,ere more likely to be encountered. 
, -".,.,..--,......... .... 
11 
5.42 
Table 5.9 
Interactions between communities 
Data from all-day observations only, Northern and Southern males. The 
number (N) and percentage of days on which each type of interaction was 
seen are shown. Some data were collected by field assistants in D2 . 
Type of interaction 
Hear or see a party, react by calling and 
advancing or retreating 
Meet male(s), attack or be attacked, 
followed by retreat 
Travel to ridge on the edge of community 
range, listen beyond, 'walk on after 
hearing no ca 11s 
Travel to ridge on the edge of community 
range, listen beyond, then return 
towards centre of community range 
Meet a mother of a different community 
(females who changed their range when 
oestrous are not scored here) 
Number of days observation 
Number of target m~les 
% of days on which at least one event 
occurred 
* Data from 19 days observation only 
3. Between males and females 
Jul-Sep 
1972 
N % 
o 0 
2 10 
1 5 
3 15 
2 10 
20 
10 
35 
act '72-
Jan '73 
N % 
2 10 
1 5 
1 5 
2 10 
2 10 
19 
11 
37 
Jul-Sep 
1973 
N % 
6 20 
o 0 
2* 10 
9 30 
30 
9 
67 
Since observations were only made on male targets, the reaction 
of lone females to hearing the calls of 'stranger' males was not seen. 
a. Monitoring behaviour Females accompanied males travelling 
to the boundary of the community range, but sometimes behaved differently 
when there. Twice when males were silent after hearing 'stranger' calls, 
females in the party called. On both occasions their calls were given 
when other community members joined the party, as frequently happens. 
Females sometimes groomed or played with their infants while males 
listened. It was my impression that monitoring was essenti ally a male 
I 
5.43 
behaviour. 
b. Chasing and attacking Females unfamiliar to observers 
have been seen to be attacked on a number of occasions (Bygott 1974). 
This may partly be the consequence of the presence of man: if a female 
fears the observers, her fearful behaviour might elicit an attack from 
the males who meet her. Nevertheless, attacks sometimes appeared to 
start before the female could see the observer (e.g. Bygott 1972). 
Attacks by males on females may take place within a community 
(Bygott 1974). The importance of the unfamiliarity of the female is 
therefore not clear. However, there is indirect evidence that males 
may even kill females. 
On 13 August 1973 Yasini noted at 1825 that a party of Southern 
males was travelling slowly "as if they had seen another animal". Two 
minutes later, in thick vegetation, they approached the dead body of a 
young adult female, and one male smelled it and groomed it briefly. 
Examination showea the body to be fresh, probably less than two days 
old. Plate 5.3, taken in situ, shows the female lying on a slope with 
her left arm stretched and holding on to vegetation above her: it 
seems that she was being pulled downs lope when she died. The most 
obvious damage was a number of puncture wounds on her back, which could 
have been due to chimpanzee canLnes. Clearly it is possible that a 
baboon or leopard may have been responsible, though it seems unlikely: 
the body was not eaten at all. She could not have fallen into the 
position where she died. The female was unfamiliar to all observers. 
The observation suggests that 'stranger' females may be very 
violently attacked by males. 
As was described earlier, when a mixed party of males and 
females responded to the calls of 'strangers' by advancing and chasing 
or attacking, females were not seen to play an active part. 
b 
Plate 5.3 Dead female 'stranger' 
The photograph shows the body as it was found. Her left arm is 
fully stretched, fingers curled round vegetation. The left shoulder 
rests on a springy sapling which she must have pushed down as she 
collapsed. The discomfort of her position indicates a violent death. 
5.44 
c. Peaceful interactions It is useful to distinguish three 
categories of females who may associate peacefully with more than one 
community. 
First, adolescent females are known to 'transfer' temporarily 
or permanently between communities (Nishida & Kawanaka 1972, Tutin in 
prep.). Adolescent females do not invariably 'transfer' permanently: 
for instance, FF remained in the Northern community after adolescence, 
though she probably 'transferred' temporarily. 
Second, there is evidence that older females may also 'transfer' 
temporarily when they become oestrous (Nishida & Kawanaka 1972). Clear 
evidence of this has not been obtained in Gombe. 
Third, anoestrous mothers may associate with more than one 
community. During the study period, for instance, three females (PS, 
NV, ME) whose ranges overlapped the Northern and Southern community 
ranges were seen to associate peacefully with parties of both communities. 
Though there are, of course, no good observations of unhabituated 
parties, there is evidence ·of similar peaceful association with a female 
(Sprout) who appears to 'belong' to a community north of the Northern 
community. Sprout has been seen occasionally since 1966, most often ln 
Linda valley, but a few times in Kasekela and twice in Kakombe. In 
1971 she was seen too rarely with members of the Northern community to 
justify being included within it (Bygott 1974). However, during D2 the 
frequency and type of interaction between Sprout and the Northern 
community justified her being labelled a Northern female. She was 
encountered on 5 of the 20 days sampled in Table 5.9, and this is an 
artificially low figure; four potential all-day observations were 
aborted when Sprout, still shy and unhabituated, ran away from the 
observers together with her habituated companions. Interactions with 
Sprout were the same as with other mothers: sometimes males displayed 
5.45 
on arrival, while at other times she joined a feeding party. 
There is no evidence that Sprout changed her range in D2 . 
However, it was clear that Northern males were using Linda valley much 
more in D2 than at other times in the study period. The change in the 
frequency of interaction with Sprout appeared to be due to a northern 
expansion of the Northern community range. 
The evidence that anoestrous mothers may associate peacefully 
with more than one community suggests that the community concept may be 
of limited use for describing the association patterns of females. 
This problem will be discussed in Section VIII. 
4. Between females 
There are no detailed studies of female social interactions. 
Females occasionally attack each other but interactions are normally 
peaceful (van Lawick-Goodall 1968). Since females appear to be dispersed 
within a community (Sections IV and V) aggressive behaviour may be 
expected in the establishment of their ranges. There is no evidence of 
this to date. 
VII. Timing and Direction of Travel 
1. Circadian rhythms 
Group Travel charts were analysed ~n order to increase the 
number of data. The number of grid lines crossed per half-hour was 
taken as a measure of distance travelled: the two measures were 
closely correlated (11/1). The results are shown in Fig. 5.12. 
Males and females both travelled further in the early morn~ng 
and late evening than in the rest of the day. Similar travel patterns 
have been seen in orang-utans (MacKinnon 1974), baboons (Altmann & 
Fig. 5.12 Circadian rhythms in travelling, party stability 
and calling 
(a) Travelling Data from Group Travel charts, January 1972 
to September 1973. All-day observations only. N = 72 
(males), 60 (females). The number of grid lines crossed 
was shown to be directly proportional to distance 
travelled (n/l). 
(b) Party stability Data from Group Travel charts, January 
to June 1973. N = 31 all-day observations (19 on males). 
A party was scored as having changed its composition if it 
was different at the end of the half-hour period. 
(c) Calling Data from July to September, 1972 and 1973. 
All-day observations on males; N = 40. 
(a) 
(b) 
(c) 
3·0 
60 
0/0 parties 
In which 40 
composition 
changes 
20 
Qver ~-hr. 
Mean no. 
of calls 
6 
4 
2 
I 0°1 ---- 99 
6 7 8 9 10 11 12 13 14 15 16 17 18 19 
Time of day 
6 8 10 12 14 16 
Time of day 
All calls heard 
Calls given by target 
",,---,- -.,"' ..... _---- .... __ .... 
O+-~~--~~~~--~.--.-.--.-.--.-.,-.-
6 8 10 12 14 16 18 20 
Time of day 
5.46 
Altmann 1970), talapoin monkeys Miopithecus talapoin (Gautier-Hion 1971), 
vervet monkeys (De Moor & Steffens 1972) and red colobus (Clutton- Brock 
1972). Siamung, however, travel further in the middle of the day than 
in the morning or evening (Chivers 1973). Factors affecting the 
circadian rhythm may include diurnal variation in temperature (Stoltz 
& Saayman 1970: baboons) and activity rhythms imposed by digestive 
processes, in addition to stimuli considered below. 
The morning peak ln distance travelled was also shown ln the 
frequency of calling and ln party stability (Fig. 5.10). The total 
number of calls heard while observing target males during two dry 
seasons was highest between 0700 and 0930 hours, and fell fairly 
steadily for the rest of the day. Dawn or dusk choruses are exhibited 
by a wide variety of vertebrates and insects (Wynne-Edwards 1962). 
The total number of calls heard was not significantly 
correlated with distance travelled in each half-hour (Spearman 
r = 0.30, N = 27, n.s.). However, the percentage of parties having 
s 
different membership at the beginning from the end of the half- hour 
was significantly correlated both with distance travelled (r = 0.36, 
s 
N 28, P < 0.05) and with the number of calls heard (r = 0.49, N = 28, 
s 
P < 0.01). Precise interpretation of these relationships is impossible 
since the data are from different periods, and are sampled from 
different individuals. However, the fact that the correlations are 
significant despite the different sources of data indicate their 
reality. A possible explanation is that as an individual travelled 
further he was more likely to enter or leave the areas most used by 
other individuals: the probability of party composition changing was 
then higher. Bygott (1974) showed that the frequency of calling 
increased when parties coalesced. 
I' 
5.47 
2. Factors influencing direction of travel 
1. Location of food sources 
Individuals were shown to be capable of returning precisely 
to food sources which they had previously visited (3/IV/2/iv). Table 5.10 
shows that when a male crossed his own path he was more likely to do so 
at a food source if he returned within three days than after a month 
2 (X = 16.8, d.f. = 1, P < 0.001). These data confirm that the direction 
of travel was influenced by the location of known food sources. 
Table 5.10 
Activities at points where a male crossed his own path 
Travel paths on different days were superimposed, and activity scored 
where an individual crossed his own path (after at least four hours) • 
Cells show number of events: 42 path-crossings were scored for each 
time period. 
Time elapsed before crossing path 
Activity Four hours to More than 30 
three days days 
Feed 24 6 
Sit in camp 10 10 
Travel along ridge 2 2 
Return to nest site 2 0 
No feeding, and no apparent 4 24 
reason for return 
11. Avai 1abi 1i ty of "s 1eeping trees" 
Although trees are abundant in Gombe National Park, chimpanzees 
sometimes appeared to make special journeys to a suitable nesting site. 
During the dry season, particularly, suitable trees are few. At this 
time there was a greater tendency to use old nests (Table 5.11): 
sometimes, and probably normally, a nest used for a second time was 
originally built by the same individual. 
5.48 
Table 5.ll 
Frequency of use of old nests 
Data on all individuals seen to make night-nests. 
Months No. of No. ·of No. of % of nests 
new nests old nests indi vi duals used twice 
Jun-Sep 1972 26 14 15 53.8 
Oct 1972 - Jan 1973 54 0 23 0.0 
Jul - Sep 1973 32 12 17 27.3 
Use of old nests suggests that it was hard to find trees in 
which to make new nests. This idea is supported by the data on use of 
palms for nesting In. Palm trees were in full leaf all year but were 
used more during the dry than the wet season (Table 5.12), both for new 
and old nests (39.5% of old nests were in palm trees). Many other 
trees were leafless from July to September. 
Table 5.12 
Frequency of nesting ln palm trees 
Data on all individuals seen to make night-nests. 
Months 
Jun-Sep 1972 
Oct 1972 - Jan 1973 
Jul- Sep 1973 
No. of nests in 
trees other than 
palm trees 
26 
52 
28 
No. of nests 
ln palm trees 
14 
2 
16 
% of nests in 
palm trees 
53.8 
3.7 
36.4 
Changes in the frequency of nesting ln palm trees were 
suggested by van Lawick-Coodall (1968) to be a 'fashion'. These data 
indicate, however, that the major influence is environmental. 
Selectivity for other trees occurred even during the wet season. For 
instance, when Combretum molle came into fresh leaf in November 1972 
target individuals nested in it on 5 out of 10 nights: I had seen no 
5.49 
nests Ln this speCLes since the previous wet season. 
Chimpanzees thus clearly selected certain species for nesting 
Ln. In addition to comfort, the height of a tree above the ground was 
clearly important (van Lawick- Coodall 1968). Van Lawick- Goodall observed 
that nests were uncommon in food trees. This seems to be because the 
food trees were often unsuitable. On six occasions I saw nests made Ln 
trees after eating in them, and twice individuals ate some of their 
nest material. 
Behavioural evidence that individuals selected particular 
nesting areas came from calls given within a party. On thirteen 
occasions at late evening some or all individuals in a party gave low-
pitched exhaled grunts, sometimes Ln serLes. I did not hear these calls 
Ln other contexts, and therefore called them nest-grunts. S. Brewer 
(pers. comm.) has heard similar calls in the same context in Senegal 
(rehabilitated chimpanzees). 
Nest- grunts seemed to be used by individuals who had found a 
potential nest area. Others then looked where the caller was looking. 
Sometimes the party nested there, but twice the party went on. For 
instance, on 25 July 1972 Charlie nest - grunted at 1906, and his 
companions looked into the trees above him. Hugh turned and walked on, 
and Charlie followed. Four minutes later Charlie climbed a tree, and 
Hugh then nest-grunted. The party all made their nests at that spot. 
Nest-grunts thus appeared to allow coordination in the choice of a 
nesting area. Three times they were given only after all individuals 
in a party had made their nests. 
The availability of suitable nesting trees therefore appeared 
to influence travel patterns, particularly during the dry season. 
5.50 
~~~. Topography 
Certain routes were favoured for crossing streams, ridges and 
patches of dense vegetation. On the steep upper slopes chimpanzees 
tended to walk along paths, which were probably made by baboons, and thus 
save energy by remaining on a contour. Similarly in Fig. 5.11 the paths 
north-east of camp on days 1, 2 and 3 converged for 200 m. along a grassy 
ridge, where walking was easier. MacKinnon (1974) found similarly that 
orang-utans favoured particular routes through the trees. 
Although the structure of the landscape clearly influenced 
travel routes, there were no geographical barriers other than Lake 
Tanganyika. 
~v. Long-distance calls 
Chimpanzees could clearly identify individual pant-hoots, and 
often changed their route after hearing calls. The effect of food-calls 
on party size was described (3/IV/4/i). Occasionally individuals 
appeared to avoid taking a direction in which others had been heard. 
Calls from non- community members invariably influenced travel 
patterns (VI/2). 
When calls were frequent, e.g. D2 compared to Dl , it was 
difficult to predict a party's travel path, since it might be affected 
at any time by the calls of others. In Dl , however, routes appeared to 
be more strongly influenced by ecological factors, such as the location 
of food sources and nesting sites. It was then possible to predict an 
individual's travel path fairly accurately after observing him for one 
or two days, us~ng knowledge of food sources, nesting sites and 
favoured routes. 
bz 
5.51 
v. Behaviour of other individuals 
The influence of one individual on others in travelling was 
discussed earlier (3/VIII/1/iii), in relation to knowledge of food 
sources. Bygott (1974) showed that some individuals were more likely 
to be first ~n a travel line than others. However, this did not 
necessarily mean that they were influencing the direction of travel. 
On 13 occasions when I saw a male take a different direction from his 
companion ahead of him, the previous leader changed direction to walk 
~n front aga~n. Both in forest and grassland parties sometimes walked 
~n file, with females normally towards the rear (Plate 5.4). In these 
situations it was clear that some individuals were merely following 
others rather than positively selecting a path. 
Individuals normally left each other without any special 
interactions. Bygott (1974) noted that males occasionally coerced 
others into following them. On two occasions a party became larger 
after a pr~me ma1e ~had called: he appeared to coerce his companions 
into following, and the party visited the boundary of the community 
range. 
VIII. Discussion 
1. The concept of the community 
~. Review 
The data presented ~n this chapter give more detail than 
previously available of the ranging behaviour of individual chimpanzees. 
They are therefore used to examine our understanding of the structure of 
chimpanzee society. 
No social unit above the level of the maternal family was 
recognised by early field studies (Reyno1ds & Reyno1ds 1965, Gooda11 
Plate 5.4 Travel line 
A typical formation of chimps travelling in open grassland 
(September 1973). Prime males tend to be at the front of such 
colunms. 
5.52 
1965). However individuals tended to be seen ~n particular areas. 
Fig. 5.l3(i) shows a diagram which would fit those observations. 
Although in this diagram males are shown to have ranges of the same 
size as females, Reynolds & Reynolds (1965) suggested that male ranges 
were larger. 
The concept of the chimpanzee community stemmed from Itani & 
Suzuki's suggestion (1967) that a number of individuals share a common 
home range and travel within it in parties of varying size and 
composition. This is represented in Fig. 5.l3(ii), in which males and 
females are pictured as travelling equally widely within the community 
range. 
It is certain that all individuals within a community may at 
times travel widely. However, it now appears that individual males 
and females have preferred areas within the community range (V/3/iii). 
These will be called 'core areas', meaning the area within which an 
individual spends a certain per cent, K , of his time. When evenly 
sampled data become available, we may be able to define a core area 
more usefully by putting a figure to K. 
Anoestrous mothers appeared to have smaller core areas than 
males or oestrous females. The evidence was that one-day (II/2/ii), 
four-day (III/2/i) and year ranges (IV /3/i) were smaller for anoestrous 
females than for males, and individual females were more directed than 
individual males in their directions of entry into camp (V/4/iii). 
Oestrous females had larger year ranges than anoestrous females (III/3/i) 
and tended to travel, with males, further than when anoestrous (Tutin 
~n prep.). 
There was also evidence that male core areas were smaller than 
the community range and dispersed within it (V/3/i). However, it was 
shown that males often travelled widely over the year range within four 
Fig. 5.l3 Models of chimpanzee dispersion 
The different diagrams are explained in the. text. Circles with thick 
lines represent males, thin lines represent females. Small circles are 
core areas of individuals, large circles are shared ranges of several 
individuals (community ranges). 
In Fig. iv the dashed circle represents the expansion of a male 
community range. 
ii 
iii 
iv 
5.53 
days (111/4), and at such times the core areas of males appeared to 
approximate the year range. There was no such evidence for anoestrous 
females. For the moment males will be treated as sharing the whole 
community range equally. The ·wider travelling of females when oestrous 
will be discussed after a general model of the relationship between the 
core areas of males and anoestrous females has been obtained. 
Fig. 5.13(iii) fits the suggestion that males travel equally 
widely over the community range, but that mothers have separate core 
areas within it. Males and females are represented as 'belonging' to a 
particular community. 
An alternative hypothesis is shown in Fig. 5.13(iv). Mothers 
are shown continuously scattered over the ground, regardless of the 
boundaries of the male communities. The core areas of some mothers ~n 
Fig. 5.13(iv) lie entirely within the range of one community of males. 
Observers of this system might regard these mothers as 'belonging' to a 
community. Many mothers, however, have ranges which overlap two male 
communities, and these could not usefully be labelled as members of a 
particular community. 
There are important conceptual differences between these ideas. 
In the first (Fig.5.13,iii), all individuals have a "group identity" 
and the social system is in this sense similar to that of cercopithecoid 
monkeys and social carnivores, which, like chimpanzees, are sometimes 
regarded as models for early man. In the second (Fig.5.13,iv), some 
females have no firm group identity; males and females have essentially 
separate social systems, in that they disperse themselves in relation to 
members of the same, and not the opposite, sex. One circumstance in 
which this difference would be crucial is if the males extend or move 
their range. If all individuals 'belong' to a community, all would move 
together: if only the males form a community, they would change their 
pattern of association with some females. 
At present it may not be useful to think that either of these 
models is a complete description. Points in favour of each are given 
below. 
5.54 
LL. Evidence Ln favour of the concept of the bisexual community 
a. Males may be aggressive to strange females (VI13/ii). This 
suggests that males distinguish between 'in-group' and 'out-group' 
females. However, the significance of this behaviour is unclear since 
males are sometimes highly aggressive to females 'within' the community. 
b. A bisexual community implies that there is coordination in 
the gross migrations of males and females. There is some evidence of this. 
First, large parties (up to 43 individuals) of both sexes have 
been seen travelling together for short distances (Itani & Suzuki 1967). 
These were suggested to be congregations of the normally dispersed 
individuals in a community. Unfortunately it is not known whether any 
of the observed parties included all the members of a community, or how 
far they travelled. In some cases the parties were seen to disperse 
without migration. 
Second, Kano (1971) suggested that certain communities migrated 
between separate areas of seasonally abundant food. The evidence was 
that at one time of year large parties were frequently heard in one 
area, but were not heard later: when Kano searched further afield, he 
found another area with large parties, and concluded that it was his 
original community. He considered the possibility that parties had 
merely dispersed within his original study area: but he observed few 
nests and heard no calls in the later period, and concluded that "except 
for a few chimpanzees who parted from the main bands, the whole band 
migrated out of the area". Significantly, however, those "few 
5.55 
chimpanzees" remaining were referred to as "comprising females and 
j uveni les". Thus these data in fact appear more compatible with the 
suggestion that males and mothers were dispersed independently. 
Third, the entry directions of males and females into camp were 
most directed towards the east in the same season (D2) (V/2). Within 
the community range it thus appeared that males and females shifted their 
main foraging areas, presumably in relation to a changing pattern of food 
distribution. Although this does not show active coordination of males 
and females, it suggests that at times both males and females have to 
move in order to find food. In some circumstances, therefore, it is 
possible that males and females could show coordinated long- distance 
migrations, as argued by Kano (1971). However, there is still no clear 
evidence of this. 
Fourth, the Southern males and females stopped visiting camp at 
about the same time. During the study period the last visits of males 
were between June 1972 (CH) and January 1973 (GO). Two Southern females 
attended camp intermittently until 1972: MD's last visit was 1n June 
and ME's 1n September 1972. These females often came to camp 1n the 
company of Southern males. The northern journeys of males and females 
thus appear to have been closely coordinated: this evidence supports the 
bisexual community concept. However, the situation becrune more complex 
in 1974, when ME again visited camp, though no Southern males had come 
since January 1973 (van Lawick- Goodall, pers. comm.). Again, therefore, 
there 1S evidence that a female's ranging habits may be independent of 
h I I • those of t e males of her commun1ty. 
111. Evidence in favour of the concept of the male-only community 
a. Although pr1me males showed only aggression or avoidance 
between the habituated communities, three females associated peacefully 
with males of more than one community (vI/2- 3). 
5.56 
b. When the Northern males expanded their range to the north, 
they regularly encountered a mother (Sprout) who was only rarely seen 
at other times. There was no evidence that the mother had moved. 
c. In the Mahali Mountains, Nishida & Kawanaka (1972) studied 
cases of females 'transferring' between communities at a time when one 
community (M-group) appeared to supplant a smaller one (K-group). 
Every November M-group migrated north, returning south in February. 
From November to February at least three females 'belonging' to K- group 
stayed in the normal K-group range, where they joined the intruding 
M-group without being rejected. As in other studies, the community was 
identified by its males, and females seen with the males were considered 
to be members of that community: females seen with more than one 
community were said to have 'transferred'. Although this term implies 
movement on the part of the female, the evidence suggests that the 
changed pattern of associations was due to movement by the males, not 
by the females. A community of males apparently shifted its range, and 
thereby reduced its association with some females, and increased its 
association with others. 
d. Although females have been labelled as Northern or Southern 
~n this study, the process was somewhat arbitrary. Bygott (1974) 
plotted the proximity data on all individuals from the Northern and 
Southern community ranges in the form of a dendrogram in which 
individuals were linked at the highest level of association. Only six 
out of sixteen independent females were linked within the male 
clusters: the remaining ten were described as peripheral. Peripheral 
females were assigned to the community which included their most 
frequently observed companion. Thus on the basis of associations both 
within and between clusters of individuals, the community concept ~s 
more firmly based on male than on female behaviour. 
5.57 
~v. Synthesis 
The evidence for the existence of chimpanzee communities ~s 
based largely on interactions between males. Females seen in the 
company of males have been described as belonging to a community, but 
there is no evidence yet that the females who occupy different parts of 
a male community range associate together more frequently than females 
separated by equivalent distances but 'belonging' to different 
communities. Nor do such females appear to show coordinated reactions 
to 'strangers', as males do. In the cases where males have been 
observed to change their range, either through migration or expans~on, 
their patterns of association with females also changed. The hypothesis 
that mothers are dispersed in relation to each other, rather than to 
males, appears to account for these observations. This suggestion needs 
to be tested by further observation of females, especially those living 
on the border of community ranges. The extent to which it fits some 
current observations is briefly examined. 
a. Adolescent transfer. There is evidence in Gombe (Tutin 
~n prep.) and the Mahali Mountains (Nishida & Kawanaka 1972) of females 
transferring between communities either permanently or temporarily 
shortly before the first pregnancy. If females are dispersed . in 
relation to each other rather than to males, their tendency to move at 
adolescence would depend on the density of females ~n relation to 
environmental resources. Having travelled with their mothers until 
late adolescence, a female must find a core area which she may occupy 
on her own. If her mother dies at this time she may not need to move. 
This may explain why Fifi has appeared to stay in the same core area 
(west of camp) as her mother, Flo, occupied. (Although Flo was 
patently old during Fifi's first pregnancy, she actually died . 15 months 
after the birth of Fifi's first infant.) Normally, however, a female 
will have to seek a new area, and this mayor may not take her into 
5.58 
the range of a different connnunity. 
Access to females is clearly vital for males. However, 
retention of females with whom mating would produce inbreeding may be 
a waste of environmental resources which could be used to maintain 
females of higher reproductive value. We may therefore expect 
mechanisms tending to promote out-breeding, such as affiliative 
behaviour to young 'stranger' females. There ~s some evidence of this 
(Pusey, in prep.). To this extent a female's choice of range at 
adolescence may be influenced by the dispersion of males, and not 
merely by the dispersion of other females. 
b. Incorporation of young males. A mother with a male infant 
may associate with two connnunities if her range is covered by both 
connnunity ranges. As her infant matures, he would perhaps confine his 
association with adult males to one connnunity only. This may in turn 
affect the rang~ng behaviour of his mother, who would, if she followed 
her son, become increasingly associated with one connnunity at the 
expense of the other. This again suggests that females may be dispersed 
according to the distribution of males, and argues in favour of the 
bisexual connnunity. 
c. Large mixed parties. Observations of parties which 
include many males and females may seem difficult to interpret on the 
male-only community hypothesis. However, it is clear that the ranges 
of females overlap very considerably, so that aggregations of females 
may be expected. Furthermore, it is important to distinguish between 
anoestrous and oestrous females. Although we expect, on the male-only 
community model, that anoestrous females are dispersed in relation to 
each other, the ranges of oestrous females should be influenced by the 
distribution of males. In large parties many of the females tend to 
be oestrous (Kortlandt 1962, Reynolds & Reynolds 1965, Izawa 1970) . 
5.59 
We may postulate that the anoestrous females in large parties are ~n 
their normal range,and that as a "sexual group" moves about the 
community range, it will include different anoestrous females as it 
leaves or enters their ranges . . My observations appeared to fit this 
suggestion well. However, it is also possible that anoestrous females 
are more likely to travel far outside their core areas when accompanied 
by familiar males. 
d. Expansion of range. During the 1973 dry season, males 
appeared to travel further north than in the same months in 1972. Thus 
the community range appeared to increase at a time when more food was 
available. In some species territory size increases when food is 
scarce, such as in predatory birds (Watson & Moss 1970). In these 
cases the territory is an exclusive feeding area, and the consequence 
of its expansion is probably an increase ~n food availability. The 
expansion of chimpanzee community ranges when food was abundant only 
makes sense if chimpanzees thereby increased their reproductive 
potential. They could do so if they associated with more females by 
expanding their territory (Fig.5.l2,iv): the data indicate that this 
occurred. This implies that females are expected to be resident ~n an 
area and associate with whatever males are present. 
e. Killing of infants is known by male langurs Presbytis spp. 
(Sugiyama 1966, 1967, Rudran 1973) and lions (Bertram in prep.) when 
they first share the ranges of breeding females. This behaviour may 
be adaptive to the male through its consequence in bringing the 
bereaved mothers into rapid oestrous. Bygott (1972) observed a party 
of Northern males killing and partly eating the infant of a 'strange' 
mother, and we may speculate that the same system is involved. Suzuki 
(1971) also observed the killing of an infant chimpanzee by males, but 
the relationship of its mother to the males was unknown. 
5.60 
This discussion suggests that neither the bisexual nor the 
male- only community model is a complete description. It indicates that 
chimpanzee society may be something which is different in form for 
different individuals. For instance, a female whose core area lies at 
the centre of a stable male community range may experience a bisexual 
group identity; this would not be true of a female whose core area 
overlaps the ranges of two male communities. The difference in 
experience could be important in its consequences on the females' 
response to change in male ranges. 
Females 1n oestrous clearly adjust their travel patterns to 
those of males, and vice-versa. At other times the dispersion of males 
and females appears to be determined by factors which are different for 
the two sexes due to different reproductive strategies. This conclusion 
supports the thesis of Goss-Custard et al. (1972) and Crook (1972): 
the adaptive significance of parameters of social structure should be 
sought less in terms of individual survival than of reproductive success. 
2. A functional explanation .of ranging behaviour 
In Chapter 3 the functional significance of the tendencies to 
form large parties and give food- calls was interpreted according to 
their relevance in different ecological conditions. This process 1S 
repeated here for ranging behaviour, and the same conclusion is 
relevant: namely, that less food was available in Dl (July to 
September 1972) than in D2 (July to September 1973). In W (November 
1972 to January 1973) food abundance was probably similar to that in 
D2 , but changes also occurred in habitat structure, climate and food 
distribution, so it is difficult to compare W with Dl and D2 • 
The ranging data must be regarded as tentative. There was 
considerab le dependence: for instance, although 17 female one- day 
5.61 
ranges ~n D2 were suitable for testing seasonal changes, these came 
from only two individuals (II/2/v). The aim of the following discussion 
is to provide a testable working hypothesis rather than to establish 
firm conclusions. 
There were insufficient data to provide an adequate statistical 
test of differences in one-day range length between Dl and D2 . However, 
males spent twice as long travelling per day in D2 as ~n Dl (3/II/2/i), 
and one-day ranges were significantly longer from May to October in 1973 
than ~n 1972 (III/2/v). Male four - day ranges appeared to be larger from 
June to September ~n 1973 than 1972 (IV/2/iv), and so did the male 
corrnnunity range. 
The data on males therefore suggest that when food was 
relatively scarce (Dl ) individuals travelled less far over one or 
several days. Furthermore, there was evidence that they were then more 
dispersed within the corrnnunity than at other times (V/3/iii) . This 
conclusion fits the results on seasonal change in party size: parties 
were smaller in Dl than in D2 (3/V1/4/iii). 
For convenience all males were represented as travelling 
equally widely over the corrnnunity range in Fig. 5.13. The data on male 
dispersion is interpreted in Fig. 5.14. The diagram suggests that when 
food is sparse (e.g. Dl ) male core areas are small and show least 
overlap: when food is abundant (D2) males travel more widely, and 
their core areas increase in size, so as to approximate the corrnnunity 
range (111/4). Even under these conditions, however, individual males 
were associated with particular parts of the corrnnunity range (V/3/i). 
There was no clear evidence of similar seasonal changes in the 
ranging behaviour of anoestrous females. When males showed least 
dispersion within the corrnnunity range (Wand D2) individual females were 
significantly more directed than males in their angle of entry into 
Fig. 5.14 Models of seasonal change in chimpanzee dispersion 
Each circle represents the core area of an individual: in each condition 
one individual's core area is shown by a thick line. Females are shown 
to be dispersed at all times (unless sexually active), and males are 
similar when food is not abundant. Although males remain slightly 
dispersed when food is abundant, an increase in the size of their core 
areas causes sufficient overlapping to create a community range which is 
almost fully shared. 
(Core area is defined as the area in which an individual spends a certain 
proportion, K , of his time, where K 1S an unknown constant.) 
~: EQQQ:t- MALES MOTHERS 
SPARSE 
ABUNDANT 
5.62 
camp (V/4/iii). The data suggest that females tended to maintain 
small, dispersed core areas regardless of changes in food aburidance. 
This is shown in Fig. 5.14. Female core areas are represented as being 
slightly smaller than the core -areas of males in D2 : Table 5.2 
indicates that at all times females had smaller one-day ranges than 
males had in Dl • 
The data on which Fig. 5.14 is based are consistent with the 
following hypothesis. Under all conditions the ranging behaviour of 
anoestrous mothers is directed to maximising feeding efficiency. Males 
have similar core areas to anoestrous mothers when they also must 
maximise feeding efficiency (when food is scarce). When food is 
abundant males travel more widely: ~n doing so they sacrifice feeding 
efficiency for the sake of increased reproductive effort, the rewards of 
which come from increased frequency of association with females. 
Anoestrous mothers cannot increase their reproductive effort by 
increasing the size of the core area; but when they are oestrous, it ~s 
to their advantage to travel widely ~n order to select a mate from as 
many suitable males as possible. 
This hypothesis is consistent with the present data. Its 
significance for males and females is examined separately. 
a. Males. In all seasons males were capable of showing 
group territoriality in the sense of Noble (1939), in that they 
defended the community range against neighbouring males (although there 
was considerable overlap between male community ranges). Brown & Orians 
(1970) noted that few mammals are territorial, and Wilson (1972) found 
the same for terrestrial primates with large home ranges. We must 
therefore ask why male chimpanzees should defend a territory. 
The data suggested that territorial behaviour was more 
frequent in D2 than in Dl (VI/2/iii). This would be a logical 
5.63 
consequence of territory expansion, since neighbours would then be 
encountered more frequently. The explanation suggested for this 
(VIII/l/iv) was that by expanding their territory, males encountered 
more females, and thus increased their chances of successful mating. 
Even if the "new" females were anoestrous, males might benefit from 
establishing with them the kind of relationship which would increase 
their probability of mating in the future. An additional long-term 
benefit may come from increased familiarity with the area (Hinde 1956). 
Though territoriality clearly has diverse functions (Hinde 1956, 
Brown & Orians 1970), one important consequence appears to be the 
maintenance of access to females and the exclusion of other males from 
them. Even if this is accepted, however, two problems remain. 
First, why should individual males have greater reproductive 
success by sharing a territory than by having lone territories? Smith 
(1968) and Hamilton & Watt (1970) showed that when two or more animals 
combine their feeding territories their energy requirements increase. 
However, this is only true if the size of the shared territory varies 
directly with the number of animals using it . The unpredictable food 
sources of chimpanzees may require each individual to retain potential 
access to an area larger than that which he needs at anyone time of 
year. Individual home ranges may therefore be expected to overlap each 
other considerably. In these circumstances individual feeding 
territories are not expected (Brown & Orians 1970). 
There are two ways in which group territoriality may be selected 
where the function of territory is to retain exclusive rights to females. 
First, if several males can collectively defend a territory larger than 
the total area of individually defended territories, each male would 
gain a reproductive advantage. This would be a strictly cooperative 
system, sensu Crook (1970b), and could evolve through the experience of 
5.64 
first two, and then more males, forming coalitions. However, those 
males would still compete between themselves for reproductive rights, 
and the system would probably therefore be unstable. Second, if one or 
more males can prevent others' from successful mating, he or they may 
tolerate the presence of others in the territory, and may enlist their 
aid in its defence. This relies on the less successful males accepting 
the status quo: this might be their best strategy if they could expect 
to achieve greater reproductive success when older. 
Both of these systems could be important. At present little is 
known of the mating success of different males (but see Tutin, in prep.). 
However, it is interesting that the only males seen to show affiliative 
behaviour with neighbouring communities have been those past prime. 
These males cannot expect to increase their reproductive success within 
the community, and territorial defence is therefore perhaps less 
important for them than for prime males. 
The second problem is why male ranging behaviour should vary 
with food abundance. The solution may be similar to that found for 
feeding behaviour. Maximisation ' of feeding efficiency may inhibit 
reproductive effort. Following the model in Fig. 5.14, males tended to 
behave like females in Dl , travelling in small parties and having small, 
dispersed core areas. Under conditions of greater food abundance males 
could retain peak condition while sacrificing feeding efficiency. 
Locomotion costs could accordingly increase, and hence the ability to 
patrol a larger area. 
This suggestion may contribute to explaining the large party 
slzes in D2 (3/VIII/3). The outcome of territorial encounters appeared 
to depend on the relative size of the two parties (VI/2/i): a method 
of increasing reproductive success may therefore be to congregate into 
parties as large as possible (i.e. compatible with adequate feeding) 
I I 
I 
5.65 
and patrol the boundaries of the territory. The large parties may thus 
be a result of reproductive competition between males both between and, 
as already argued (3/VIII/4/i), within communities. 
b. Females. The model ~n Fig. 5.14 suggests that anoestrous 
mothers tend to stick to small core areas even when food is abundant. 
Further data are clearly needed before this hypothesis can be accepted, 
but its implications may be examined. 
Whereas a male can increase his reproductive success by 
fertilising many females, a female is probably limited in breeding rate 
by physiological constraints. In Gombe the first post-partum oestrous 
coincides roughly with the end of weaning (van Lawick-Goodall i.n Eress). 
Frisch & McArthur (1974) showed that women need to reach a certain 
weight for height ratio before ovulation occurs, and suggested that one 
cause of this was that a minimum level of body fat (22%) is required: 
and in lactating women conception occurs sooner after birth ~n well-
nourished than poorly nourished groups (Bonte et al. 1974). Possibly 
chimpanzees lactating in the wild cannot attain a sufficient level of 
body fat during lactation (which demands 1000 extra calories a day for 
women, Gunther 1971). This suggestion is supported by the fact that 
chimpanzee growth rates are faster in captivity than in the wild, and 
menarche is earlier (Pusey in prep.). If the breeding rate of females 
is held down by the rate of fat deposition it may, unlike males, be 
increased by maximisation of feeding efficiency at all times. 
The ranging behaviour of chimpanzees is thus considered to be 
different in form and function for males and females. The hypothesis 
is consistent with known facts but requires to be tested by further 
data, on two points in particular. 
,First, it is assumed that the most efficient way of exploiting 
the environment in Gombe is for individuals to disperse in core areas 
d 
5.66 
smaller than the community range. There ~s no evidence for this other 
than that small core areas were used, as described above. Testing of 
this point would be difficult but valuable. 
Second, the distribution of food has been assumed to vary less 
importantly than its overall abundance. This was my impression during 
observation in Dl and D2, when most of the foods were similar between 
seasons. However, it is certain that the distribution of food has an 
important effect on ranging patterns (e.g. Nishida 1974), and study of 
the effect of changing food distributions on ranging patterns within a 
community is required. 
> 
I. Introduction 
Chapter 6 
OVERVIEW 
6.1 
The significance of the data on feeding and ranging behaviour 
was discussed in the relevant chapters. Two points are discussed here 
to put the results ~n perspective. 
First, the influence of artificial feeding is examined. This 
was one of a number of limitations to the study. Others included the 
small sample size (of months and individuals) and confinement of the 
study to a single area; the short-comings of these are obvious. 
Second, chimpanzees are compared to apes and other mammals 
having similar social structures. The extent to which their social 
structures may be explained in similar terms to those proposed for 
chimpanzees is examined. 
11. The Influence of Artificial Feeding 
1. Introduction 
Any change to an environment influences the behaviour of 
animals using it. The important question ~s not whether but how much 
the artificial feeding system affected what was observed. 
The only satisfactory control would be to study a wild 
population without interference and subsequently establish an 
artificial feeding area. No detailed study of this type has been made. 
However, the influence of provisioning in Gombe may be examined by 
comparing (1) different methods of artificial feeding, (2) individuals 
who did and did not visit the feeding area and (3) the artificial 
feeding area with natural food sources. 
6.2 
2. Community structure 
i. Community ranges 
It was shown that the artificial feeding area lay on the 
northern boundary of the Southern male community range. Bygott (1974) 
found that the frequency of visits to camp by Southern males fell almost 
continuously in successive six-month blocks from 1968 to 1972, even 
though they received bananas more often per visit than Northern males: 
no Southern male has visited camp Slnce January 1973. Furthermore, 
Northern and Southern males had four-day ranges which were similar in 
form, tending to return to a central point (5/111). For Northern males 
this point was sometimes camp and sometimes, as for Southern males, a 
natural food source. 
It is possible that the tendency for Southern males to share a 
community range, within which their travel centred on temporarily 
important food sources (or nesting areas) was a consequence of the past 
influence of artificial feeding. This seems unlikely. First, it is 
clear .from several field studies (e.g. Reyno1ds & Reyno1ds 1965) that 
unprovisioned chimpanzees may return daily to temporarily abundant food 
sources. Second, if ranging patterns were easily distorted by 
artificial feeding, one would expect them to return equally easily to 
their natural form. 
The similarity of the rang1ng behaviour of Northern and 
Southern males, despite the unimportance of camp to the latter, suggests 
that the size and form of the community range was not greatly influenced 
by the feeding system. 
11. Individual ranges 
The presence of a permanent food source would be expected to 
aggregate individuals using it. Nevertheless individuals were found to 
I I 
I I 
occupy different core areas. This result is opposite to the expected 
effect, and is therefore unlikely to be an artefact. 
6.3 
Detailed study of the overlap of individual ranges within the 
Southern community has not been made. In the Northern range it seems 
likely that the degree of overlap of female core areas was increased 
by artificial feeding. Nope, for instance, was often followed by 
observers when she left the feeding area, and tended to travel rapidly 
to the north-east, feeding little en route. The core areas of a number 
of females may similarly have extended in the direction of camp. 
Natural food sources could also induce normally separated females to 
aggregate: the difference here is that bananas were given all year, 
and some female core areas could thus show more permanent overlap than 
normal. Males could have been affected similarly, but the observed 
seasonal variation in male dispersion cannot be explained in terms of 
the influence of artificial feeding. 
3. Relation to feeding behaviour 
i. Importance of bananas as food 
Wrangham (1974) showed that from 1965 to 1967 the daily output 
of bananas provided about 44000 calories, sufficient for fourteen adult 
chimpanzees at Yerkes Regional Primate Cent er (Bourne 1971). At this 
time the feeding area clearly contributed significantly to the food 
reserves of the local population. However, following a change in the 
feeding method in 1968 many fewer bananas were given. 
During the present study visitors received bananas less than 
once a week (for 9 males and 11 females in Dl , Wand D2 , mean = 9.9 days 
between feeds, range 6.1 - 19.3); the mean number of independents fed 
per day was 2.2 (for Dl , W, D2); the number of bananas given varied 
between one and fifteen. It is doubtful that an average ration of 
-6.4 
bananas supplied as much as 20% of an adult's food requirements for a 
day: the real figure is probably much lower. Given that 10% of the 
Northern community wer.e fed per day the feeding area therefore 
contributed less than 2% to the .tota1 food requirements of the Northern 
community. This shows that the importance of natural food sources was 
not much diminished by the provision of bananas during the study period. 
ii. Predation 
It was shown that the frequency of predation on baboons was 
closely related to the manner and extent of artificial feeding (4'/IV/2/i). 
It could be argued that as a result of the experience of predation in an 
artificial situation (the aggregation of predator and prey species) 
chimpanzees learned to kill more frequently outside camp. However, 
(1) chimpanzees were seen to kill monkeys in Gornbe before being 
provisioned (van Lawick-Gooda11 1968); (2) studies without artificial 
feeding have recorded predations (references in Chapter 4); (3) few 
baboons have been killed by chimpanzees since 1968. The factors 
affecting the probability of a predation occurring appeared to be 
largely independent of the degree of predatory interest (4/IH/3). Thus 
it seems unlikely that use of the artificial feeding area had long-term 
consequences on the frequency of predation. 
Myer (1969) found ~n the laboratory situation that rats exposed 
to mice from weaning were less likely to be mouse- killers than those 
brought up without them. Some young baboons and chimpanzees formed 
friendly relationships in camp (van Lawick-Gooda11 1971): perhaps when 
adult those individuals would be less likely to kill baboons than others. 
However, since baboons are rarely killed in the natural situation this 
is probably hypothetical. 
-iii. The nature of camp as a food source 
Possibly some individuals were able to time their visits to 
the artificial feeding area so as to improve their chances of getting 
bananas. All Northern individu,a1s, however, visited camp many times 
more often than they were fed (for 9 males and 11 females, D1 , W, D2 : 
mean per cent of days attendance when fed was 19.6%, range 10.7 to 
32.3%). Furthermore, the more often an individual attended camp, the 
less often per visit he received bananas. 
If a chimpanzee finds a fruit tree and returns to it daily, 
he will normally find some r1pe fruit available until the crop is 
completely finished. Sometimes he will arr1ve just after a feeding 
party of chimpanzees or other species, and find none. Nevertheless, 
his chance of finding some food is clearly higher than 20% (e.g. see 
Table 5.10). 
6.5 
The low predictability of food in camp means that the 
motivation of a chimpanzee 1n going there may be different from when 
visiting natural sources. In both cases he may want to eat, but he is 
gambling more by coming to camp: he might waste potential feeding time 
on natural foods which would be more predictable even if less 
desirable than bananas. 
Camp was therefore not quite equivalent to natural food sources. 
The extent to which different individuals were prepared to gamble on 
bananas may have varied in a different way from their attendance 
patterns at more predictable food sources. It is also possible that 
part of the attraction of camp for some was the opportunity to form 
parties. Dyadic association and interaction patterns in camp cannot 
therefore be assumed to be representative of those outside it, even 
when allowing for the larger parties. In this study analysis of 
social interactions over food was consequently confined to observations 
outside the feeding area. 
6.6 
lV. Seasonal variation 
It was shown that two hours after a target left camp his party 
size was not different from one hour before entering, but that at both 
times parties were smaller than when in camp (3/VI/4). This suggests 
that the aggregating effect of the feeding system was limited to the 
feeding area. However, it was also shown that individuals who had 
arrived separately were more likely to leave together if they were 
given bananas (3/VI/4). These results appear to conflict: the 
explanation is that the second did not happen very often. 
In identifying seasonal change in environmental variables (3/VII) 
it was assumed that only natural variables were important. However, if 
the manner or frequency of artificial feeding had changed dramatically 
between seasons party size could conceivably have been affected even 
outside camp. This possibility is therefore briefly examined. 
Parties were larger m D2 than in Dl O/VI/4/iii). If this was 
the consequence of a change ln artificial feeding methods, we would 
expect to find that more bananas, creating larger parties, were given 
ln D2 than in Dl • 
For 9 males and 11 females there was a significant decrease ln 
the number of days they received bananas from Dl to D2 (Wilcoxon 
T = 31.5, N = 19, P < 0.01). There were no intended or perceived 
changes in the manner of feeding. The lower frequency of feeding ln 
D2 was perhaps the consequence of larger parties: bananas were glven 
to individuals in small parties in order to avoid competition. 
Since fewer bananas were given when parties were larger (D2) 
it is most unlikely that unnoticed changes in the feeding system were 
responsible for the observed seasonal changes in behaviour. 
6.7 
Ill. Comparisons with Other Species 
1. Introduction 
Meaningful conclusions from comparisons of behaviour and ecology 
between species are elusive. As long as no yardsticks of ecological 
similarity are available arguments relating behavioural to environmental 
similarities are essentially circular: ecological comparisons may be 
made at whatever level is most convenient. Thus Eisenberg et al. (1972) 
classed Ateles spp., Cercopithecus mitis and chimpanzees as fruit-eaters, 
and used this to explain similarities in their social structures. On the 
other hand Crook (1972) explained similarities between open-country and 
forest Cercopithecidae partly in terms of their both experiencing food 
poverty, whose importance was judged to supercede that of differences 
in food type. 
This problem ~s inevitable: the relevant ecological differences 
between species do indeed depend on which species are compared. For 
instance the variables relevant to comparison with chimpanzees could be 
food distribution in the case of gorillas and food density in the case 
of gibbons. The same ~s true of similarities: chimpanzees may be 
similar to wolves and hyaenas because the optimal s~ze of feeding 
parties varies similarly (Mech 1970, Kruuk 1972); but their 
similarities with wallabies (Kaufmann 1974) have to be sought at a 
different level. 
In an attempt to avoid difficulties such as these Altmann (1974) 
derived testable principles about baboon socio- ecology from the 
proposition that "for any set of tolerable ecological conditions, the 
adaptive activities of baboons tend in the long run toward some optimal 
distribution away from which mortality rate is higher, or reproductive 
rate is lower, or both". Again however there was the problem of 
circularity, though in a different form: the principles were derived 
6.8 
using observations of baboons, and then tested against them. 
The search for adaptive principles is based on the hope that 
they may apply more widely than to the species from which they were 
derived. As A1tmann (1974) pointed out, the principles are more 
valuable where their supposed effect has an understandable mechanism. 
In spite of the disadvantages arising from a sUbjective choice of 
variables, therefore, interspecific comparisons may be valuable both 
heuristica11y and sometimes in suggesting or affirming the processes 
responsible for the evolution of particular behaviour patterns. In 
this section a few species are selected for comparison with chimpanzees 
on this basis. 
2. Closely related species 
Explanation of behaviour ~n functional terms means discovering 
the consequences through which natural selection acts. Rodman (1973) 
proposed that the social structure of orang-utans could be best 
explained by applying socio-eco1ogica1 models to the behaviour of 
females. This implies that natural selection acts on the individual 
competitive abilities of females to raise young who are themselves good 
competitors, and follows from Trivers' (1972) emphasis that in mammals 
females are a limiting resource for males and not vice-versa (see also 
Crook 1972 and Goss-Custard et al. 1972). The principle of this 
approach is accepted here and applied to a comparison of the apes. 
Orang-utans (Fongo pygmaeus) live ~n pr~mary ra~n forest and 
are almost entirely solitary, other than mothers travelling with young. 
Their social structure appears to vary somewhat between areas and has 
not yet been fully elucidated. Females occupy dispersed but widely 
overlapping home ranges, while adult males cover larger areas and may 
possibly be territorial. Small aggregations of females and males 
b 
r 
I 
L 
occasionally occur at fruiting trees. It ~s possible that a form of 
group~ng exists at a higher level than so far identified, but 
convincing data are lacking (Rodman 1973, MacKinnon 1974). 
Gorillas occupy pr~mary and secondary ra~n forest, living in 
social groups of two to twenty animals: solitary males also occur. 
6.9 
In the mountain gorilla G. g. beringei groups contain at least one fully 
adult male (silver-back), together with younger males, adult females and 
their offspring. Interactions between groups are highly variable in 
nature, from peaceful association to aggression (Schaller 1963, Fossey 
1974, Harcourt in prep.). 
Gibbons (Hylobates lar) and siamang (Symphalangus syndactylus) 
live in pr~mary rain forest in family groups containing an adult male 
and female with their one to three offspring: some adults are solitary. 
Home ranges are defended as territories against conspecifics (Ellefson 
1968, Chivers 1972). 
Numerical data on the females of these species are summarised 
~n Table 6.1. 
a. Territoriality The advantage of territoriality to a 
female is the protection of valuable resources, and the disadvantage ~s 
the use of time and energy and the danger incurred in its defence. 
Whether the advantages outweigh the dis.advantages will depend partly on 
the importance of competition for resources. For instance if a female 
needs to have access over a year to a larger area than she needs at any 
one time, many of the resources she would protect by being territorial 
would be unimportant at the time they were defended. 
A possible method of estimating the s~ze of the home range in 
relation to the area which a female needs at one time is to calculate 
a predicted home range size on the basis of body weight. McNab (1963) 
6.10 
found a close relationship between body weight and home range s~ze 
in largely non-social species: Table 6.2 shows the predictions from 
McNab's correlation. 
Tab le 6.1 
Comparison of ranging behaviour of female apes 
The data are summarised from Ellefson (1968) (Hylobates lar) , Chivers 
(1972, pers. comm.) (Symphalangus syndactylus), Schaller (1963), 
Fossey (1974), Harcourt (in prep.) (Gorilla gorilla beringei), Rodman 
(1973), MacKinnon (1974) (Pongo pygmaeus) and this study. The figures 
are approximate and do not show the range of variation, but show 
differences in order of magnitude. The table refers to mothers rather 
than sexually active females. 
Gibbon Siamang Orang-utan Chimpanzee Gorilla 
No. of animals/sq.km. 
(all age-sex classes) 
Distance travelled 
in one day (km) 
Size of home range 
(sq.km.) (over one 
year) 
Territorial defence 
of home range 
Weight (kg) 
Diet 
Food distribution 
Table 6.2 
9 
2 
0.4 
Yes 
5 
Fruit 
Patchy 
17 
1 
0.2 
Yes 
10 
Leaf 
Patchy 
2 
0.5 
0.6 
No 
40 
Fruit 
Patchy 
Predicted and observed home range s~ze of female apes 
4 
3 
8 
No 
35 
Fruit 
Patchy 
The data are calculated 
McNab's (1963) equation 
in sq.km. 
from body weights shown in Table 6.1, 
for "hunters" (see 5/IV/5). Area is 
0.5-2 
0.5 
5 
No 
100 
Leaf 
Even 
us~ng 
shown 
Gibbon Siamang Orang-utan Chimpanzee Gori lla 
Observed home range 
Predicted home range 
Ratio of observed to 
predicted home range 
0.40 
0.35 
1.1 
0.20 0.65 
0.41 0.70 
0.5 0.9 
8 5 
0.63 1.32 
12.7 3.4 
6.11 
Chimpanzees and gori 11as have much larger home ranges than 
predicted on the basis of body weights. Although McNab's (1963) 
correlation should be treated with care since most of the species 
considered were comparatively small, a real difference from the other 
apes is indicated. Association into permanent feeding groups causes 
inefficient foraging in animals whose range s~ze must increase in 
proportion to the number in the group (Hamilton & Watt 1970): if 
feeding efficiency ~s the most important factor the larger home ranges 
of chimpanzees and gorillas than predicted can be taken as a cause and 
not an effect of their social structure. 
A possible explanation ~s the need to visit different areas ~n 
different seasons according to changes in food distribution. There 
was evidence for this in chimpanzees (S/V). Fossey (1974) found that 
there was little seasonal variation in diet or use of different parts 
of the range by gorillas: however the year range ~s larger than the 
month range (Harcourt in prep.). Gorillas may need to cover a large 
area at all times in order to allow regeneration of food sources: 
Fossey (1974) found that in the small core area of a lone male much of 
the foliage was destroyed. If chimpanzee and gorilla females both need 
to occupy large home ranges ~n relation to their body s~ze, 
territoriality should not be advantageous. 
Table 6.2 shows that gibbons, siamang and orang-utans all have 
home ranges similar in size to those predicted on the basis of body 
weight. We should therefore expect these species to occupy feeding 
territories. Gibbons and siamang fit this expectation, but orangs 
do not. This is possibly a datum error: MacKinnon (1974) considers 
that females may have larger home ranges than the 0.65 sq.km. found by 
Rodman (1973). However the anomaly of orang-utans may be explained on 
the basis of locomotion costs. Orangs are clearly slower than gibbons 
6.12 
or siamang (MacKinnon 1974), and this is reflected in their smaller 
day ranges in relation to home range s~ze (Table 6.2). If locomotion 
~s much costlier for orang-utans in terms of time and energy than it 
~s for the agile hylobatids ~erritoriality would not be favoured: the 
advantage of a defended area would be high, but the costs would be 
even higher. 
b. Female groups If females have feeding territories they 
should not share them s~nce locomotion costs would then increase without 
advantage, as Smith (1968) argued for squirrels. This "prediction" ~s 
upheld for gibbons and siamang: only one female occupies the 
territory (Ellefson 1968, Chivers 1972). 
How much non-territorial individuals should associate may be 
expected to depend partly on the distribution of the food. In many 
species (e.g. Crook 1965, Jarman 1974) continuously distributed food 
is associated with group living. Schoener (1971) argued that group 
living cannot increase feeding efficiency. However it is possible that 
by coordinating animals which forage in large areas for foods which 
must be allowed to regenerate, group formation reduces the probability 
of individuals returning to a recently visited source (cf. Cody 1974): 
individuals can therefore feed more efficiently in groups than alone. 
The continuous distribution of the food of gorillas could then be held 
responsible for their living in permanent social groups. 
The food sources of chimpanzees and orang- utans are patchily 
distributed and vary considerably in size. Under these conditions 
individuals are not expected to form permanent associations s~nce 
competition may regularly occur at smaller food sources. Competition 
might similarly be expected in gorillas if a lone individual could 
regularly finish the food available at individual sources. This 
possibly happens rarely because of the presence of toxins in their 
3 
foods: chimpanzee parties typically finished the rlpe fruit but not 
the leaves at food sources. 
6.13 
c. Male association with females In some seasonally breeding 
mammals ln which territories are maintained by females the sexes only 
come together during the breeding season (Ewer 1968). The system in 
chimpanzees and orang-utans appears essentially similar, given the year-
round breeding. Individual females are expected to associate with males 
only when sexually receptive. However, Slnce the males cannot predict 
when females will become receptive, they must monitor them continually. 
This picture is consistent with data on chimpanzees and orang-utans. 
The model of chimpanzee ranging ln Fig. S.13(iv) is very 
similar to that proposed by Rodman (1973) for orang- utans. The only 
structural difference is that in orang-utans a single male spreads his 
home range (territory?) over several females, whereas in chimpanzees 
several males occupy a territorial community range. The factors 
permitting or promoting group territoriality in chimpanzees were 
discussed earlier CS/VIII). In comparison to orang-utans two features 
are probably important: first, chimpanzees are faster and probably 
expend less energy in locomotion than orangs; and second it is 
conceivable that feeding competition in parties is less severe in 
chimpanzees. If true, these conditions would make group territoriality 
impossible for orang- utans. 
In gibbons and slamang it would seem advantageous for females 
to exclude males until they were required for breeding. However this 
would mean that males had nowhere to live unless they occupied 
interstitial territories, in which case they would be competing with 
the females for land. An alternative solution would be for males to 
occupy larger territories than females, so that only a part of each 
I 'I 
a 
male's food requirements came from each female's territory. Both of 
these would seem acceptable alternatives unless there were some 
specific advantages to sharing territories. 
6.14 
Two advantages seem to operate In favour of permanent consorting. 
First, although the females give the loud calls which announce the 
location of groups, males take the major burden of territorial defence, 
thus relieving the female of the most important costs. Second, the male 
assists the female in care of the offspring, actually carrying the infant 
for over a year in the case of siamang (Chivers 1972). Bisexual sharing 
of the territory is thus explained by the male cooperating with the 
female: the result may be to increase her reproductive rate above the 
level if she were alone. This would be impossible in chimpanzees and 
orang-utans because territoriality is not a viable strategy for a 
solitary female, let alone one permanently accompanied by a male. 
In gorilla groups the number of adult males and females is 
variable, but on average there are four times as many females as 
silver-backs (fully adult males) (Schal1er 1963). Female acceptance 
(in evolutionary terms) of a male permanently sharing their range may 
be a sensible strategy if a male is needed for mating at unpredictable 
times. If he must monitor them he must share their range to some 
extent: the large food sources and advantage in coordinated returns 
may mean that permanent association is the most economical solution. 
The male's size (up to twice the female's weight) suggests that he has 
additional importance in protecting them: gorillas cannot run as fast 
as chimpanzees or climb trees to escape terrestrial predators, and 
when surprised by man the male gorilla is the only ape which charges 
rather than flees (Schaller 1963). Furthermore when interactions 
between groups are aggressive silver-backs are the main protagonists 
(Fossey 1974). 
~I 
------------------------------....... 
r 
These "explanations" of differences in social structure are 
inevitably partly circular. However they demonstrate that it may be 
possible to explain the observed differences in terms of a common 
theoretical framework based on individual competition within sexes. 
The usefulness of this approach can be determined by further analysis 
within the apes and other animal groups. 
3. Distantly related species 
1.. Introduction 
Ecological and behavioural convergence between two distantly 
related species argues strongly in favour of behaviour having evolved 
6.15 
as a response to a specific environment. For instance, Eisenberg et al. 
(1972) noted convergence between the Old World Miopithecus and New World 
Saimiri: both species eat insects and fruit and have similar grouping 
tendencies. 
The situation 1.S confused, however, by distantly related 
species having similar social structures but no obvious ecological 
similarities (e.g. chimpanzees and wallabies, below). If the behavioural 
similarities in such cases are fortuitous, they could also be so for 
species with similar eco10gies. It is therefore important to examine 
these cases in the hope of finding an explanation for behavioural 
convergence. Three species with different degrees of ecological 
similarity to chimpanzees are selected to see how far this is possible. 
1.1.. Spider monkeys 
The South American spider monkeys Ate1es spp. are arboreal 
fruit-eaters weighing 5-6 kg (Hill 1962). Both in A. geoffroyi 
(Eisenberg & Kuehn 1966) and A. be1zebuth (K1ein & K1ein 1973) 
individuals travel in parties of unstable composition, but the parties 
appear to be made up from a fixed number of animals equivalent to a 
I I 
, I 
These "explanations" of differences in social structure are 
inevitably partly circular. However they demonstrate that it may be 
possible to explain the observed differences in terms of a common 
theoretical framework based on individual competition within sexes. 
The usefulness of this approach can be determined by further analysis 
within the apes and other animal groups. 
3. Distantly related species 
i. Introduction 
Ecological and behavioural convergence between two distantly 
related species argues strongly in favour of behaviour having evolved 
6.15 
as a response to a specific environment. For instance, Eisenberg et al. 
(1972) noted convergence between the Old World Miopithecus and New World 
Saimiri: both species eat insects and fruit and have similar grouping 
tendencies. 
The situation ~s confused, however, by distantly related 
species having similar social structures but no obvious ecological 
similarities (e. g. chimpanzees and wallabies, below). If the behavioural 
similarities in such cases are fortuitous, they could also be so for 
species with similar ecologies. It is therefore important to examine 
these cases in the hope of finding an explanation for behavioural 
convergence. Three species with different degrees of ecological 
similarity to chimpanzees are selected to see how far this is possible. 
~~. Spider monkeys 
The South American spider monkeys Ateles spp. are arboreal 
fruit-eaters weighing 5-6 kg (Hill 1962). Both in A. geoffroyi 
(Eisenberg & Kuehn 1966) and A. belzebuth (Klein & Klein 1973) 
individuals travel in parties of unstable composition, but the parties 
appear to be made up from a fixed number of animals equivalent to a 
rl I 
6.16 
chimpanzee community. Klein & Klein (1973) found that the commonest 
party s~ze (independent individuals only) was two: 15% of parties had 
eight animals or more, and the largest observed had 22. Individual 
A. belzebuth shared the home range of the "community", but there was 
some evidence that individuals tended to be associated with particular 
parts. There was 20 - 30% overlap between neighbouring "community ranges", 
which were 2.5-4 sq. km. in area. 
Eisenberg et al. (1972) suggested that the similarity of spider 
monkeys and chimpanzees was the result of the uneven distribution of 
their food sources, which could best be exploited by foraging units 
searching independently: they can aggregate when one unit finds a fruit 
tree and announces its location to others. Though the correlation may 
be right the proposed mechanism does not explain how the social system 
could have resulted from natural selection of individuals (or kin). 
The description of spider monkey social structure is strikingly 
similar to that of chimpanzees and the functional explanations may 
therefore also be similar. Differences between male and female ranging 
habits of spider monkeys have not been identified, but females were not 
observed to take part in aggressive interactions between parties from 
different "communities" (Klein & Klein 1973). Further data may show 
whether the factors suggested in this study to be responsible for the 
form of chimpanzee society can be applied to spider monkeys. 
An explanation is also required for the differences between 
spider monkeys and gibbons, which are fruit-eaters of similar size. 
The solution is expected to be in terms of the minimum area required 
to support a female over a year: this cannot be tested at present. 
Possibly gibbons have more specialised digestive systems than spider 
monkeys and are consequently able to eat foods of lower nutritional 
value, such as unripe fruit w"hen no ripe ones are available. 
JI 
6.17 
Alternatively the density of fruiting trees may be higher in South-East 
Asia than in South America. 
It may seem surprising In Vlew of the climatic rhythms known to 
occur in tropical forest that generalised frugivores can expect an 
adequate supply of fruit all year even over a large area: the beginning 
of the rains is the best time for vegetative growth and the end of the 
dry season is therefore the best time for fruit production (Baker 1973). 
In fact plants in tropical forests tend to have staggered fruiting 
seasons (Croat 1969, Smythe 1970a, Leck 1972, Clutton-Brock 1972). 
Snow (1964) suggested that this occurred because of intense competition 
both for dispersal agents and between seedlings: he noted that the 
effect is to provide frugivorous animals with a sufficient food supply 
all year. new forest species have wind- dispersed seeds, both because 
seeds must be big and because of the poor chance of dispersion (Ashton 
1969): even forests suffering long dry seasons (up to 5 months) are 
therefore habitable by' sedentary fruit-eaters. The tolerance of the 
animal for poor food items, however, will determine how large an area 
it needs to visit over the year in order to include a sufficient 
number of its seasonally productive food sources. 
iii. Social carnlvores 
Spotted hyaenas (Kruuk 1972) live in social groups (clans) 
occupying a home range within which individuals travel in parties of 
variable size and composition. Wolves Canis lupus (Mech 1970) and 
lions (Schaller 1972) have similar societies. Fission- fusion societies 
are uncommon In animals, and the circumstances promoting it in hyaenas 
could therefore be similar to those important to chimpanzee social 
structure. 
It seems fairly clear that a common feature is temporal 
variation in the relationship bet'veen feeding efficiency and party Slze 
-6.17 
Alternatively the density of fruiting trees may be higher in South- East 
Asia than in South America. 
It may seem surprising ln view of the climatic rhythms known to 
occur ln tropical forest that generalised frugivores can expect an 
adequate supply of fruit all year even over a large area: the beginning 
of the rains is the best time for vegetative growth and the end of the 
dry season is therefore the best time for fruit production (Baker 1973). 
In fact plants in tropical forests tend to have staggered fruiting 
seasons (Croat 1969, Smythe 1970a, Leck 1972, Clutton- Brock 1972). 
Snow (1964) suggested that this occurred because of intense competition 
both for dispersal agents and between seedlings: he noted that the 
effect is to provide frugivorous animals with a sufficient food supply 
all year. few forest species have wind-dispersed seeds, both because 
seeds must be big and because of the poor chance of dispersion (Ashton 
1969): even forests suffering long dry seasons (up to 5 months) are 
therefore habitable by sedentary fruit-eaters. The tolerance of the 
animal for poor food items, however, will determine how large an area 
it needs to visit over the year in order to include a sufficient 
number of its seasonally productive food sources. 
iii. Social carnlvores 
Spotted hyaenas (Kruuk 1972) live in social groups (clans) 
occupying a home range within which individuals travel in parties of 
variable size and composition. Wolves Canis lupus (Mech 1970) and 
lions (Schaller 1972) have similar societies. Fission-fusion societies 
are uncommon ln animals, and the circumstances promoting it in hyaenas 
could therefore be similar to those important to chimpanzee social 
structure. 
It seems fairly clear that a common feature is temporal 
variation in the relationship between feeding efficiency and party Slze 
6.18 
due to the variable SLze of food sources. Kruuk (1972) found that the 
number of hyaenas feeding together varied directly with the weight of 
the prey: chimpanzee party sizes also varied at different food sources 
apparently in relation to food availability (3/VI/4). This suggests 
that competition for food is responsible for variation of party size 
both Ln chimpanzees and hyaenas. 
Unlike chimpanzees male and female hyaenas share the territory 
equally, and both sexes defend it against intruders. This indicates, 
as Kruuk concluded, that territorial defence by hyaenas is related 
more to feeding than to sexual behaviour, which implies that feeding 
efficiency is greater in groups than alone. This LS so: success Ln 
hunting certain prey increased with the number of hunters. If 
aggregation promotes feeding efficiency the consequences may increase 
the tendency to aggregate, as Krebs (1974) found for herons Ardea 
herodias. This was not found for chimpanzees. Thus both the 
similarities and differences between hyaenas and chimpanzees can be 
explained in terms of the relationship between feeding efficiency and 
party size. 
The social structure of human hunter-gatherer populations may 
be explained in similar terms. The camp size of Hadza people is 
determined partly by competition for food, with camps splitting up as 
a result of quarrels when few large animals are killed (Woodburn 1968): 
Tobias (1964) similarly found that Bushman bands broke up when food was 
scarce. The factors responsible for aggregation when food is in good 
supply are less clear however, and are presumably more complex than 
suggested for hyaenas and chimpanzees. 
iv. Wallabies 
The whiptail wallaby Macropus parryL LS probably the most 
social of all the marsupials (Kaufmann 1974). Between 30 and 50 
L 
individuals occupy home ranges which overlap broadly or completely: 
they are of all age-sex classes and form a "mob" equivalent to a 
chimpanzee community. Mob members associate in parties having a mean 
6.19 
of six and a mode of t~vo individuals. Adult males travel more widely 
than females within the mob range, and regularly check the oestrous state 
of females by sniffing. Interactions between mobs occur in a small 
overlap area, and vary from peaceful association to aggression: when 
aggression occurs the males from each mob may act together. Females 
have not been observed to transfer between mobs but there is evidence 
that a mother's range may overlap two different mobs. Males appear to 
transfer to a new mob before adulthood (Kaufmann 1974). 
The social system of whiptail wallabies is thus remarkably 
similar to that of chimpanzees, the major difference being that male 
wallabies may change mobs. However their environment ~s totally 
different. They are strictly grazers, living in woodland and open 
grassland. Kaufmann (1974) considered that various possible ecological 
advantages of group life did not apply (increase in feeding efficiency, 
exploitation of seasonal .environment, or modification of the 
environment). However, he noted that the nature of their food supply 
permits aggregation, and suggested that this is favoured for two ma~n 
reasons: first, individuals may be safer from predators when in 
parties than when alone; second, since there is no breeding system 
males must be able to monitor females all year. 
It seems likely that variations in food supply would be 
responsible for dispersal at times, as happens even in species forming 
permanently associating social groups (e.g. Aldrich-Blake et al. 1971): 
if the distribution and density of food does not set a varying upper 
limit to party size one would expect individuals to travel as a 
permanent group, especially if predation is reduced by association. 
I, 
Thus the social structure of the two species could be similar for the 
same reasons: a female's best feeding strategy ~s to occupy a larger 
home range than it is economical to defend, and males compete for 
access to females at the time they are oestrous. 
6.20 
Ecology is the relationship between behaviour and the 
environment: ecological pressures are the product of that relationship. 
The fact that chimpanzees, hyaenas and wallabies live in different 
environments and eat different foods does not prevent them from 
experiencing similar ecological pressures. At present the suggestion 
that similar pressures have caused these species to have similar social 
structures is a largely circular argument based on selective 
interpretation of the importance of different variables. Nevertheless 
it seems possible that emphasis of the different selective pressures 
acting on males and females as individuals may prove a more successful 
theoretical framework than on the adaptive significance of grouping 
systems per se. 
A. \ 
Appendix 3.1 
Monthly variation in time spent eating different foods 
In each month foods are ranked according to the per cent of 
observation minutes on which they were eaten. Feeding was scored on the 
minute: foods marked X were seen eaten only between minute points. 
Data are for target individuals only. 
Not all foods recorded in Appendix 3.2 as being observed during 
this study are included. Some foods were never seen to be eaten by 
the target chimpanzee; others were identified only to genus in the field. 
Parts eaten are coded as in Appendix 3.2. 
May 1972: 2776 minutes of good observation, 8 observation periods, 
6 targets 
Spec;.ies Part 
eaten Rank 
% feeding minutes on Number of 
Uvaria ango1ensis 
Dip10rhyrtcus condy1ocarpon 
Monanthotaxis poggei 
Ficuspo1ita 
Me11era lobu1ata 
E1aeis guineensis 
F. grtapha10carpa 
Pterocarpus tinctorius 
Musa sp. (banana) 
~a11is-choudae 
Ficus spp. 
F. va11is-choudae 
Afrosersa1isia cerasifera 
F. irtgens 
Apocynaceae 
F. gnapha10carp a 
F. ·capensis 
Whitfie1dia sp. 
Smi1ax kraussiana 
Tric1isia sac1euxii 
Vangueria acuti10ba 
F. brachy1epis 
F. urceo1aris 
F; depauperata 
Ipomoea cairica 
Ampe10cissus cavicau1is 
Termite mud 
Dead leaf 
Entada abyssinica 
F 
S 
F 
F 
F 
F 
F 
L 
F 
F 
L 
L 
F 
F 
L 
L 
F 
L 
P 
F 
F 
F 
L 
L 
L 
F 
L 
R 
1 
2 
3 
4 
5 
6 
7 
8.5 
8.5 
10 
11 
12 
13 
14.5 
14.5 
16 
17 
18 
19.5 
19.5 
23 
23 
23 
23 
23 
26.5 
26.5 
28.5 
28.5 
which food was eaten bouts 
36.1 
11.2 
9.7 
8.2 
6.8 
6.5 
3.1 
2.6 
2.6 
2.0 
1.7 
1.4 
1.3 
1.2 
1.2 
1.1 
1.0 
0.6 
0.3 
0.3 
0.2 
0.2 
0.2 
0.2 
0.2 
0.1 
0.1 
X 
X 
19 
10 
22 
2 
13 
3 
2 
3 
1 
2 
3 
1 
3 
2 
2 
2 
4 
1 
4 
1 
3 
1 
1 
1 
1 
1 
1 
1 
1 
- -
A. 2 
June 1972: 4134 minutes of good observation, 16 observation periods, 
9 targets 
Species 
Monanthotaxis poggei 
Sterculia quinque1oba* 
Dip1othyncuscondy1ocarpon 
E1aeis guineensis 
Harungana madagascariensis 
Ipomoea cairica 
Asystasia gangetica 
Ficus gnapha10carpa 
Parinari curate11ifo1ia 
Canthium crassum 
Uvaria ango1ensis 
Me1lera lobu1ata 
F. gnapha10carpa 
Grewiap1atyc1ada 
I.eriocarpa 
Ash 
F. va11is - choudae 
Larvae on Annona 
Phragmites mauritianus 
P.curate11ifo1ia 
Pi1iostigmathoningii 
F .ottoniifolia 
Tinospora caffra 
Ps eudos portdi as mi crocarp a 
Musa sp. (banana) 
Fteridium aqui1inum 
F. va11is-choudae 
Smi1axkraussiana 
Unidentified 
Vigna ob1ongifo1ia 
Costus afer 
Sterculia tragacantha 
Merremia pterygocau1os 
Aca1yphaornata 
E~ guineensis 
Oecophy11a longinoda 
Meat 
Sabicea orienta1is 
Unidentified 
F • . depauperata 
F. urceo1aris 
Unidentified 
A1bizia gummifera 
C1erodendrum schweinfurthii 
Honey 
Water 
F. ·lukanda 
Baphia capparidifo1ia 
Apocynaceae 
Termite mud 
* 
Part 
eaten 
F 
S 
S 
F 
F 
L 
L 
F 
L 
F 
F 
L 
L 
F 
L 
F 
P 
F 
S 
F 
L 
F 
F 
L 
L 
P 
L 
S 
P 
L 
L 
L 
P 
L 
F 
L 
L 
P 
R 
G 
F 
L 
L 
Rank 
1 
2 
3 
4 
5 
6 
7 
8 
9 
10 
11 
12 
13 
14 
15 
16 
17.5 
17.5 
19 
20.5 
20.5 
23 
23 
23 
27.5 
27.5 
27.5 
27.5 
27.5 
27.5 
32.5 
32.5 
32.5 
32.5 
36.5 
36.5 
36.5 
36.5 
41.0 
41.0 
41.0 
41.0 
41.0 
45.0 
45.0 
45.0 
49.5 
49.5 
49.5 
49.5 
One of these records was S. mhosya 
% feeding minutes on Number of 
which food was eaten bouts 
19.0 
8.2 
7.7 
7.6 
6.8 
5.5 
5.2 
5.0 
4.1 
4.0 
3.7 
2.9 
2.8 
1.9 
1.8 
1.7 
1.6 
1.6 
0.8 
0.6 
0.6 
0.5 
0.5 
0.5 
0.4 
0.4 
0.4 
0.4 
0.4 
0.4 
0.3 
0.3 
0.3 
0.3 
0.2 
0.2 
0.2 
0.2 
0.1 
0.1 
0.1 
0.1 
0.1 
0.0 
0.0 
0.0 
X 
X 
X 
X 
60 
4 
15 
11 
19 
16 
14 
9 
9 
4 
3 
11 
10 
5 
12 
1 
5 
14 
4 
1 
1 
1 
7 
1 
4 
3 
3 
8 
8 
3 
1 
2 
1 
1 
1 
1 
1 
1 
1 
1 
3 
1 
2 
2 
1 
2 
1 
1 
1 
3 
July 1972: 8884 minutes of good observation, 20 observation periods, 
10 targets 
Species 
Brachystegia bussei 
Monanthotaxis poggei 
E1aeis guineensis 
Harungana madagascariensis 
Grewia p1atyc1ada 
Parinari curate11ifo1ia 
Ficus va11is-choudae 
Me11era lobu1ata 
Asystasia gangetica 
Uvaria ango1ensis 
Cordia africana 
Canthium crassum 
Ipomoea eriocarpa 
Ficus ingens 
F. -lukanda 
Ch1orophora exce1sa 
F. gnapha10carpa 
P. curate11ifo1ia 
Ficus sp. 
Unidentified 
Phragmites mauritianus 
Pi1iostigma thoningii 
Aspi1iap1uriseta 
Saba florida 
F .vallis-choudae 
Crota1aria 1achnophora 
Smi1ax kraussiana 
F .kituba1u 
F. urce101aris 
I~ cairica 
Stercu1ia .tragacantha 
Azanza garckeana 
F. capensis 
Water 
Baphia capparidifo1ia 
Macrotermes be11icosus 
Dory1us nigricans 
C1erodendrum schweinfurthii 
Tinospora caffra 
E. guineertsis 
Tric1isia sac1euxii 
Costus afer 
Schrebera a1ata 
Smi1ax kraussiana 
Paper 
Ficus sp. 
Merremia pterygocau1os 
Pteridium aqui1inum 
Termite mud 
A1bizia gummifera 
Combretum mo11e 
F~urceo1aris 
I ~ obscura 
P avoni a urens 
Cloth 
Part 
eaten 
S 
F 
F 
F 
F 
L 
F 
L 
L 
F 
F 
F 
L 
F 
F 
G 
L 
F 
L 
L 
P 
S 
L 
F 
L 
L 
F 
F 
L 
L 
L 
F 
F 
L 
G 
L 
P 
F 
P 
B 
P 
F 
L 
L 
R 
R 
F 
L 
P 
Rank 
1 
2 
3 
4 
5 
6 
7 
8 
9 
10 
11 
12 
l3 
14 
15.5 
15.5 
17.5 
17.5 
19.5 
19.5 
21 
23 
23 
23 
28.5 
28.5 
28.5 
28.5 
28.5 
28.5 
33.5 
33.5 
33.5 
33.5 
36.5 
36.5 
40 
40 
40 
40 
40 
44.5 
44.5 
44.5 
44.5 
51.5 
51.5 
51.5 
51.5 
51.5 
51.5 
51.5 
51.5 
51.5 
51.5 
% feeding minutes on 
which food was eaten 
10.2 
9.5 
9.4 
8.0 
7.5 
6.3 
6.2 
5.9 
5.1 
4.9 
3.1 
2.9 
1.9 
1.8 
1.6 
1.6 
1.2 
1.2 
1.0 
1.0 
0.9 
0.6 
0.6 
0.6 
0.5 
0.5 
0.5 
0.5 
0.5 
0.5 
0.4 
0.4 
0.4 
0.4 
0.3 
0.3 
0.2 
0.2 
0.2 
0.2 
0.2 
0.1 
0.1 
0.1 
0.1 
0.0 
0.0 
0.0 
0.0 
0.0 
0.0 
0.0 
0.0 
0.0 
0.0 
Number of 
bouts 
9 
67 
26 
25 
22 
18 
25 
40 
42 
14 
4 
6 
26 
2 
5 
1 
9 
5 
3 
12 
10 
1 
5 
1 
7 
2 
1 
1 
13 
11 
2 
2 
2 
25 
7 
1 
2 
9 
1 
2 
1 
3 
1 
5 
1 
1 
2 
3 
6 
1 
1 
1 
1 
1 
1 
I' 
August 1972: 7589 minutes of good observation, 17 observation periods, 
9 targets 
Species 
E1aeis guineensis 
Sabaflcirida 
Ptetocatpus tinctorius 
Brachystegia bussei 
Garciniahui11ensis 
Sterculia quinque10ba 
Parinaricurate11ifo1ia 
Ficusva11is-chciudae 
Ch1orophora exce1sa 
Acacia hockii 
Pseudospondias microcarpa 
F. griapha10carp a 
Macrotermes be11icosus 
Asystasia gangetica 
Strychnos spp. 
Antiaris ·toxicaria 
Ficus spp. 
Me11era lobu1ata 
Tabernaemontana ho1stii 
Hymertocardia acida 
Ipomoea cairica 
Ficus ~ngens 
Grewia p1atyc1ada 
F. va11is-choudae 
Phragmitesmauritianus 
E. guineensis 
Dip10rhyncus condy1ocarpon 
A1bizia glaberrima 
Ficus poli ta 
A1bizia gummifera 
F. capensis 
P. curate11ifo1ia 
I. eriocarpa 
Unidentified 
Ch1orophora exce1sa 
F. urce101aris 
C1erodendrum schweinfurthii 
Cloth 
Termite mud 
Dictyoph1eba lucida 
Honey 
F. urce101aris 
Uapaca ni tida 
Oecophy11a longinoda 
Tinospora caffra 
Baphia capparidifo1ia 
Combretum mo11e 
Water 
Unidentified 
Ficus spp. 
Sterculia tragacantha 
Monanthotaxis poggei 
Canthium crassum 
Part 
eaten 
F 
F 
L 
S 
F 
S 
L 
F 
G 
S 
F 
L 
L 
F 
F 
L 
L 
F 
S 
L 
F 
F 
L 
P 
P 
S 
L 
F 
R 
F 
F 
L 
L 
F 
L 
G 
F 
F 
F 
L 
L 
R 
S 
F 
L 
F 
F 
Rank 
1 
2 
3 
4 
5 
6 
7 
8 
9 
10 
11 
12 
14 
14 
14 
17 
17 
17 
19.5 
19.5 
21 
22 
24 
24 
24 
27 
27 
27 
29 
30.5 
30.5 
33 
33 
33 
36.5 
36.5 
36.5 
36.5 
41 
41 
41 
41 
41 
47.5 
47.5 
47.5 
47.5 
47.5 
47.5 
47.5 
47.5 
54 
54 
% feeding minutes on 
which foo·d was eaten 
28.6 
9.1 
7.1 
5.9 
5.7 
5.3 
3.5 
2.9 
2.6 
2.0 
1.8 
1.7 
1.5 
1.5 
1.5 
1.4 
1.4 
1.4 
1.3 
1.3 
1.1 
1.0 
0.9 
0.9 
0.9 
0.7 
0.7 
0.7 
0.6 
0.5 
0.5 
0.4 
0.4 
0.4 
0.3 
0.3 
0.3 
0.3 
0.2 
0.2 
0.2 
0.2 
0.2 
0.1 
0.1 
0.1 
0.1 
0.1 
0.1 
0.1 
0.1 
0.0 
0.0 
Number of 
bouts 
40 
11 
16 
5 
8 
8 
6 
11 
2 
4 
7 
4 
4 
20 
6 
2 
4 
6 
7 
10 
10 
1 
8 
7 
8 
3 
4 
1 
1 
3 
4 
1 
6 
4 
1 
9 
12 
1 
14 
3 
1 
4 
2 
1 
2 
2 
1 
20 
1 
1 
1 
2 
1 
11 
August 1972 (cont.) 
Species 
Mus a sp. (b anana) 
Ash 
Cement lick 
Part Rank 
eaten 
F 54 
54 
54 
% feeding minutes on Number of 
which food was eaten bouts 
0.0 1 
0.0 1 
0.0 1 
September 1972: 4596 minutes of good observation, 12 observation periods, 
9 targets 
Species Part 
eaten 
E1aeis guineensis F 
Pterocarpus tinctorius L 
P . . artgo1ensis B 
Ch1orophora exce1sa G 
Garcirtia hui11ensis F 
Ficus va11is-choudae F 
Morus 1actea F 
Phragmites mauritianus P 
Sabaf10rida F 
Ash 
Ch1orophora exce1sa 
F ~ capensl.s 
F ~ pd1ita 
Mat 
F 
F 
F 
Canthium hispidum F 
Combretum mo11e R 
Sterculia quinque10ba S 
Cloth 
Hymenocardia acida S 
E. guineensis P 
Meat 
F. gnapha10carpa L 
Myrianthus arboreus F 
Unidentified L 
Tric1isia sac1euxii F 
Baphia capparidifo1ia L 
Patirtari curate11ifo1ia F 
Acacia hockii S 
Water 
F .va1lis - choudae L 
Arttiaris toxicaria F 
Smi1ax kraussiana P 
F. urce101aris F 
Ipdmoeaeriocarpa L 
Termite mud 
Anisophy11ea pomifera F 
Crota1aria 1achnophora L 
C1erodendrum schweinfurthii G 
Ipomoea obscura L 
Rank 
1 
2 
3 
4 
5 
6 
7.5 
7.5 
9 
10 
11 
12 
13 
15.5 
15.5 
15.5 
15.5 
18 
20 
20 
20 
22 
23 
24.5 
24.5 
26.5 
26.5 
29 
29 
29 
33 
33 
33 
33 
33 
37.0 
37.0 
37.0 
39.0 
% feeding minutes on 
which food was eaten 
21.3 
17.8 
12.6 
8.8 
7.5 
5.6 
3.3 
3.3 
2.3 
1.7 
1.6 
1.5 
1.2 
1.1 
1.1 
1.1 
1.1 
1.0 
0.8 
0.8 
0.8 
0.6 
0.5 
0.4 
0.4 
0.3 
0.3 
0.2 
0.2 
0 . 2 
0.1 
0.1 
0.1 
0.1 
0.1 
0.0 
0.0 
0.0 
X 
Number of 
bouts 
19 
28 
13 
10 
9 
7 
1 
4 
2 
2 
1 
1 
5 
3 
1 
9 
2 
2 
3 
1 
3 
2 
1 
2 
1 
2 
1 
1 
9 
1 
1 
3 
1 
1 
9 
1 
1 
1 
1 
.I 
A. b 
October 1972: 6042 minutes of good observation, 14 observation periods, 
13 targets 
Species 
Elaeis guineensis 
Pterocarpus tinctorius 
F~ vallis-choudae 
Garcinia huillensis 
Baphia capparidifolia 
Canthium hispidum 
Brachystegia longifolia 
Salacia Sp. 
Ficus Spp. 
Musa sp. (banana) 
TIIiacorafunifera 
Morus lactea 
Dictyophleba kirkii 
Ficus sp. 
F. ·lukanda 
Combretum mo11e 
Phragrnites mauritianus 
Oecophylla longinoda 
Tabernaemontana holstii 
Baphia capparidifolia 
DioScorea schimperiana 
Anisophyllea boehmii 
F; ·ingens 
Unidentified 
F; capens1.S 
Pseudospondias microcarpa 
F. vallis-chouaae 
Pycnanthus angolensis 
Unidentified 
Hibiscus surratensis 
Smilax kraussiana 
Monanthotaxis poggei 
E. guineensis 
Termite mud 
F. urcelolaris 
F. gnaphalocarpa 
Asystasia gangetica 
Water 
D. lucida 
F. gnaphalocarpa 
Merremiapterygocaulos 
At r amomum s p . 
Unidentified 
Ipomoea cairica 
CoStus afer 
Crematogaster sp. 
Rhus vulgaris 
Acalypha ornata 
Soil 
Uapacani tida 
I.eriocarpa 
F. vallis-choudae 
Sterculia tragacantha 
Part 
eaten 
F 
L 
F 
F 
L 
F 
B 
F 
L 
F 
F 
F 
F 
F 
F 
R 
P 
F 
B 
P 
F 
F 
S 
F 
F 
L 
F 
L 
B 
P 
F 
P 
L 
F 
L 
F 
L 
L 
P 
F 
L 
P 
L 
L 
F 
L 
C 
L 
Rank 
1 
2 
3 
4 
5 
6 
7 
8 
9 
10 
11 
12 
l3 
14.5 
14.5 
16 
18 
18 
18 
20.5 
20.5 
22 
23.5 
23.5 
26 
26 
26 
29 
29 
29 
31 
32.5 
32.5 
34 
36.5 
36.5 
36.5 
36.5 
43 
43 
43 
43 
43 
43 
43 
43 
43 
48.5 
48.5 
51.5 
51.5 
51.5 
51.5 
% feeding minutes on 
which food waS eaten 
22.5 
16.9 
10.5 
7.7 
4.4 
4.3 
3.5 
2.7 
1.9 
1.8 
1.7 
1.6 
1.4 
1.3 
1.3 
1.2 
1.1 
1.1 
1.1 
1.0 
1.0 
0.9 
0.8 
0.8 
0.7 
0.7 
0.7 
0.6 
0.6 
0.6 
0.5 
0.4 
0.4 
0.3 
0.2 
0.2 
0.2 
0.2 
0.1 
0.1 
0.1 
0.1 
0.1 
0.1 
0.1 
0.1 
0.1 
0.0 
0.0 
X 
X 
X 
X 
Number of 
bouts 
27 
21 
10 
7 
l3 
12 
2 
7 
3 
5 
4 
1 
3 
3 
3 
6 
7 
2 
1 
3 
13 
2 
2 
6 
1 
2 
2 
2 
2 
1 
9 
1 
2 
8 
4 
2 
2 
13 
1 
2 
1 
1 
2 
1 
1 
1 
1 
1 
2 
1 
2 
1 
1 
I I 
....... ------------------
A.i 
November 1972: 3053 minutes of good observation, 10 observation periods, 
7 targets 
Species 
Garcinia hui11ensis 
Baphia cappatidifo1ia 
E1aeis guineensis 
Pseudosportdias microcarpa 
Mactotermes be11icosus 
Ptetocarpus tirtctorius 
Canthium hispidum 
Dioscotea schimperiana 
Ficus ingens 
Dictyoph1eba lucida 
Ficus sp. 
Parinari curate11if61ia 
F. urceo1aris 
Sa1acia sp. 
Termite mud 
Ptetidium aqui1inum 
F. ·lukanda 
Tinospora caffra 
F. depauperata 
Unidentified 
Steganotaenia ara1iacea 
Unidentified 
F. kituba1u 
Part 
eaten 
F 
L 
F 
F 
L 
F 
P 
F 
F 
L 
L 
L 
F 
L 
F 
L 
L 
L 
P 
P 
L 
Rank 
1 
2 
3 
4 
5 
6 
7.5 
7.5 
9 
10 
11 
12 
14 
14 
14 
16 
18 
18 
18 
21 
21 
21 
23 
% feeding minutes on 
which food was eaten 
21.6 
16.3 
14.5 
12.9 
11. 3 
10.0 
2.4 
2.4 
1.9 
1.8 
1.7 
0.6 
0.5 
0.5 
0.5 
0.3 
0.2 
0.2 
0.2 
0.1 
0.1 
0.1 
X 
Number of 
bouts 
15 
17 
7 
l3 
14 
9 
4 
8 
1 
2 
2 
1 
2 
1 
7 
3 
1 
5 
2 
1 
1 
2 
1 
F 
A, ' 
December 1972: 4137 minutes of good observation, 11 observation periods, 
9 targets 
Species 
Garcinia hui11ensis 
Dictyoph1eba lucida 
E1aeis guineensis 
Meat 
Ficus spp. 
Tinospora caffra 
F. gnapha10carpa 
Macrotermes be11icosus 
Dioscorea schimperiana 
Saba florida 
Aspi1ia rudis 
Baphia capparidifo1ia 
Phy11anthus mue11erianus 
Phragmites mauritianus 
T. caffra 
Sa1acia sp. 
F~ urceo1aris 
F. exasperata 
Mellera lobu1ata 
B. capparidifo1ia 
Musa sp. (banana) 
Aftamomum sp. 
Artnona senega1ensis 
Parinari curate11ifo1ia 
Water 
Ipomoea sp. 
Hos1undia opposita 
Smi1ax kraussiana 
Pavonia urens 
Termite mud 
F. capens~s 
F. depauperata 
Merremia pterygocau1os 
Pteridium aqui1inum 
Canthium hispidum 
Begonia princeae 
D. kirkii 
F. kituba1u 
Unidentified 
Unidentified 
Steganotaenia ara1iacea 
Part 
eaten 
F 
'F 
F 
L 
L 
F 
P 
F 
L 
L 
F 
P 
B 
F 
L 
F 
L 
B 
F 
P 
F 
L 
L 
B 
P 
P 
F 
L 
L 
L 
F 
L 
F 
L 
L 
P 
P 
Rank 
1 
2 
3 
4 
5 
6 
7.5 
7.5 
9 
10 
11 
12 
l3 
14 
15 
16 
17 
18 
19.5 
19.5 
21 
22 
24 
24 
24 
26.5 
26.5 
29 
29 
29 
33.5 
33.5 
33.5 
33.5 
33.5 
33.5 
38.5 
38.5 
38.5 
38.5 
41 
% feeding minutes on 
which food was eaten 
21.0 
19.8 
10.2 
6.0 
5.2 
4.3 
3.3 
3.3 
3 . 1 
3.0 
2.9 
2.7 
2.2 
1.7 
1.5 
1.4 
1.3 
1.1 
0.8 
0.8 
0.7 
0.6 
0.4 
0.4 
0.4 
0.3 
0.3 
0.2 
0.2 
0.2 
0.1 
0.1 
0.1 
0.1 
0.1 
0.1 
0.0 
0.0 
0.0 
0.0 
X 
Number of 
bouts 
22 
32 
14 
1 
3 
9 
6 
2 
26 
la 
5 
2 
5 
3 
2 
3 
3 
1 
2 
1 
1 
5 
6 
1 
5 
1 
1 
2 
1 
5 
1 
1 
2 
2 
1 
1 
1 
1 
1 
1 
1 
p 
A.~ 
January 1973: 5408 minutes of good observation, 12 observation periods, 
la targets 
Species 
Dictyoph1eba lucida 
Annona senega1ensis 
Ficus gnapha10carpa 
E1aeis guineensis 
Me11era lobu1ata 
Saba florida 
Tinospora caffra 
Garcinia hui11ensis 
F. exasperata 
Ficus spp. 
E1aeis guineensis 
Phragmites mauritianus 
Dioscorea schimperiana 
F. urceo1aris 
Larvae on Baphia 
Panicum maximum 
Macrotermes be11icosus 
Aeschynomene baumii 
A1bizia sp. 
Aspilia rudis 
Ipomoea cairica 
Unidentified 
Merremia pterygocau1os 
Baphia capparidifo1ia 
Begonia princeae 
D. kirkii 
Asystasia gangetica 
F .kituba1u 
F. urceo1aris 
A1biziagummifera 
F .gnapha1ocarpa 
1. eriocarpa 
Apocynaceae 
Smi1ax kraussiana 
Termite mud 
Water 
Sabicea orienta1is 
Bidens grantii 
Pteridium aqui1inum 
Pterocarpus tinctorius 
Unidentified 
Part 
eaten 
F 
. F 
F 
F 
L 
F 
L 
F 
F 
L 
P 
P 
P 
L 
P 
L 
L 
L 
L 
F 
L 
L 
L 
F 
L 
L 
F 
R 
L 
L 
L 
P 
F 
L 
L 
L 
L 
Rank 
1 
2 
3 
4 
5 
6 
7 
la 
la 
la 
la 
10 
l3.5 
l3.5 
15 
16.5 
16.5 
18 
20 
20 
20 
22.5 
22.5 
24 
25.5 
25.5 
28.5 
28.5 
28.5 
28.5 
33.5 
33.5 
33.5 
33.5 
33.5 
33.5 
39 
39 
39 
39 
39 
% feeding minutes on 
which food was eaten 
42.3 
11.1 
7.8 
7.0 
5.6 
2.9 
1.7 
1.5 
1.5 
1.5 
1.5 
1.5 
1.4 
1.4 
1.1 
0.9 
0.9 
0.8 
0.7 
0.7 
0.7 
0.5 
0.5 
0.4 
0.3 
0.3 
0.2 
0.2 
0.2 
0.2 
0.1 
0.1 
0.1 
0.1 
0.1 
0.1 
0.0 
0.0 
0.0 
0.0 
0.0 
Number of 
bouts 
18 
91 
3 
14 
25 
8 
14 
5 
3 
3 
2 
5 
13 
12 
6 
7 
4 
2 
3 
2 
7 
1 
l3 
2 
1 
2 
3 
1 
4 
4 
2 
3 
1 
la 
3 
2 
2 
2 
1 
1 
2 
r 
11 
1 
bz 
A. JO 
July 1973: 5597 minutes of good observation, 8 observation periods, 
4 targets 
Species 
Parinari curate11ifo1ia 
E1aeis guineensis 
Dip10rhyncus condy1ocarpon 
P .curatellifolia 
Ficus gnapha10carpa 
Asystasia gangetica 
Canthium crassum 
Uapaca nitida 
Ipomoea eriocarpa 
F. va11is-choudae 
Me11era lobu1ata 
Cordia africana 
Ficus spp. 
Ch1orophora exce1sa 
Grewia p1atyclada 
F. exasperata 
C. exce1sa 
1. cairica 
E. guineensis 
Phragmites mauritianus 
E. guineensis 
Merremia pterygocau1os 
A1bizia gurnmifera 
F. urceo1aris 
Pterocarpus tinctorius 
Sterculia tragacantha 
Clerodendrum schweinfurthii 
F. 1ukanda 
Meat 
Water 
Azanza garckeana 
P. ango1ensis 
Tabernaemontana ho1stii 
F. va11is-choudae 
P. tinctori us 
Dombeya rotundifo1ia 
Smi1ax kraussiana 
F. gnapha10carpa 
Termite mud 
I. obscura 
Monanthotaxis poggei 
Strychnos spp. 
Vangueria acuti10ba 
Entada abyssinica 
Oncinotis inandensis 
Part 
eaten 
F 
F 
S 
L 
F 
L 
F 
F 
L 
F 
L 
F 
L 
G 
F 
L 
F 
L 
B 
P 
P 
L 
R 
L 
B 
B 
G 
F 
F 
L 
F 
L 
L 
B 
P 
L 
L 
F 
F 
F 
R 
P 
Rank 
1 
2 
3 
4 
5 
6 
7.5 
7.5 
9 
10 
11 
12 
13 
14 
15.5 
15.5 
17.5 
17.5 
20.5 
20.5 
20.5 
20.5 
23.5 
23.5 
26 
26 
26 
30 
30 
30 
30 
30 
36.5 
36.5 
36.5 
36.5 
36.5 
36.5 
36.5 
36.5 
43 
43 
43 
43 
43 
% feeding minutes on 
which food was eaten 
41.4 
13.0 
11.0 
5.0 
4.0 
2.9 
1.9 
1.9 
1.8 
1.6 
1.5 
1.4 
1.3 
1.1 
1.0 
1.0 
0.9 
0.9 
0.6 
0.6 
0.6 
0.6 
0.5 
0.5 
0.4 
0.4 
0.4 
0.2 
0.2 
0.2 
0.2 
0.2 
0.1 
0.1 
0.1 
0.1 
0.1 
0.1 
0.1 
0.1 
0.0 
0.0 
0.0 
0.0 
0.0 
Number of 
bouts 
50 
24 
24 
14 
7 
19 
10 
5 
10 
2 
12 
5 
4 
1 
3 
2 
1 
5 
1 
2 
3 
7 
1 
11 
2 
2 
9 
2 
2 
17 
2 
1 
3 
3 
1 
1 
8 
1 
7 
1 
1 
1 
1 
1 
1 
A,)! 
August 1973: 6439 minutes of good observation, 11 observation periods, 
6 targets 
Species Part 
eaten 
Parinari curate11ifo1ia F 
E1aeis guineensis F ' 
Pterocarpustinctorius L 
E1aeis guineensis P 
Uapaca nitida F 
Combretum mo11e R 
Ficus va11is-choudae F 
Ch1orophora exce1sa F 
Ficus spp. F 
Strychnos spp. F 
F. ingens F 
Sterculia quinque10ba S 
Ficus spp. L 
Asystasia gangetica L 
F. urceo1aris L 
Macrotermes be11icosus 
Me11era lobu1ata L 
F. exasperata L 
C. exce1sa G 
P. curate11ifo1ia L 
F. po1ita F 
Cordia africana F 
Hymenocardia acida S 
A1bizia gunnnifera R 
Ipomoea eriocarpa L 
Dip10rhyncus condy1ocarpon S 
Morus 1actea F 
Pseudospondias microcarpa F 
Antiaris toxicaria F 
Tabernaemontana ho1stii F 
Oncinotis inandensis P 
E. guineensis B 
F.urce101aris F 
F. capens~s F 
F~ 1ukanda F 
Water 
Unidentified L 
Honey 
Oecophy11a longinoda 
F. gnapha10carpa F 
Acacia h6ckii S 
C1erodendrum schweinfurthii G 
Termite mud 
Smi1ax kraussiana P 
Entada abyssinica R 
F. gnapha10carpa L 
Aframomum sp. P 
Pavonia urens P 
Pycnarithus ango1ensis F 
Baphia capparidifo1ia L 
Tinospora caffra L 
Phragmites mauritianus P 
Rank 
1 
2 
3 
4 
5 
6 
7 
8 
9 
lO 
11.5 
11.5 
13 
14 
15.5 
15.5 
17 
19 
19 
19 
21 
22 
23 
25.5 
25.5 
25.5 
25.5 
30 
30 
30 
30 
30 
33 
34.5 
34.5 
37.5 
37.5 
37.5 
37.5 
42 
42 
42 
42 
42 
46.5 
46.5 
46.5 
46.5 
50.5 
50.5 
50.5 
50.5 
% feeding minutes on Number of 
which food was eaten bouts 
28.9 
15.5 
lO.l 
3.8 
3.7 
3.6 
3.5 
3.3 
2.3 
1.8 
1.7 
1.7 
1.6 
1.5 
1.4 
1.4 
1.2 
1.1 
1.1 
1.1 
1.0 
0.8 
0.7 
0.6 
0.6 
0.6 
0.6 
0.5 
0.5 
0.5 
0.5 
0.5 
0.4 
0.3 
0.3 
0.2 
0.2 
0.2 
0.2 
0.1 
0.1 
0.1 
0.1 
0.1 
0.0 
0.0 
0.0 
0.0 
X 
X 
X 
X 
32 
30 
24 
5 
4 
17 
12 
3 
5 
4 
3 
4 
4 
9 
13 
2 
9 
2 
2 
3 
2 
1 
4 
2 
4 
1 
2 
4 
2 
10 
1 
2 
9 
3 
1 
21 
1 
2 
2 
1 
1 
4 
9 
3 
1 
1 
1 
1 
1 
1 
1 
1 
,I 
A. )2,. 
September 1973: lS02 minutes of good observation, 3 observation periods, 
3 targets 
Species Part Rank % feeding minutes on Number of 
eaten which food was eaten bouts 
Ch10rophora exce1sa 'G 1 24.2 2 
Ficusva11is-choudae F 2 20.4 3 
Parirtari ' curate11ifo1ia F 3 15.9 7 
Pterocarpus ango1ensis B 4 13.3 4 
Pterocarpus tinctorius L S 10.9 4 
Pseudospondiasmicroc~rpa F 6 3.4 2 
Aspilia rudis L 7 2.7 2 
Pycnanthus ango1ensis F 8 2.4 2 
F; capensis F 9 1.9 4 
Combretummo lle R 10.S 1.4 6 
F. gnapha10carpa L 10.S 1.4 1 
Soil 12 1.3 1 
Monanthotaxis pogge~ F l3 0.3 1 
Uapaca nitida F 14.S 0.2 1 
Termite mud 14.S 0.2 3 
Water 16 X 3 
-A.13 
Appendix 3.2 
Foods recorded ~n Gombe National Park 
Key: 1 Recorded ~n this st~dy and van Lawick-Goodall 1968. 
2 Recorded for the first time ~n this study. 
3 Recorded for the first time by: a: David Bygott; b: Tim 
Clutton-Brock; c: Es lom Mpongo; d: Frans Plooij; e: Mitzi Thorndahl; 
f: R.W.W. (1970-71) ; g: Yasini Selemani; h: David Riss; ~: Caroline 
Tutin; j: general records of Gombe Research Centre. 
4 Recorded by van Lawick-Goodall (1968) but not in this study. 
Food not identified or rarely seen. 
5 Recorded in van Lawick-Goodall (1968) but not ~n this study. 
Food frequently seen available and record is doubtful. 
Underlining: food observed eaten once only. 
Parts eaten: F fruit, S seed, B blossom, L leaf, P pith, R resin, 
C cambium, G gall. 
Foods whose code (Plants: parts eaten; Animals & Miscellaneous: 
observation category) is underlined have only been recorded once. The 
"number of observations" given for Animal and Miscellaneous Foods is by 
all observers in Gombe National Park. 
1. PLANT FOODS 
PTERIDOPHYTA 
Dennstaedtiaceae 
Pteridium aquilinum (L.) Kuhn 
ANGIOSPERMAE: DICOTYLEDONS 
Acanthaceae 
Asystasia gangetica (L.) T. And. 
Mellera lobulata S. Moore 
Whitfieldia sp. 
Anacardiaceae 
Mangifera indica L. (alien) 
!,seudospondias microcarpa (A. Rich.) Engl. 
Rhus vulgaris Meikle 
Part eaten 
FSBLPRCG 
3f 
2 
1 
2 
2 
1 3e* 
3f 2 
A. )'1-
11 
Part eaten 
F S B L P R C G 
Annonaeeae 
Artnona senegalensis Pers. 1 
Mortanthotaxis poggei Engl. & Diels 1 
Uvatia angolensis Oliv. 2 
Apoeynaeeae 
Dietyophleba (Landolphia) kirkii 2 
D. lucida (K. Sehum. ) Piehon 1 
Diplorhyneus eondyloearpon (Muell. Arg.) 
Piehon 1 5 
Oneinotis inandensis Wood & Evans 2 
Saba florida (Benth.) Bullock 1 
Tabernaemontana holstii (K. Sehum. ) Stapf. 1 5 
Araliaeeae 
Cussonia kirkii Seem. var. kirkii 3e 
Aselepiadaeeae 
Pentarhinum sp. 3b 
Begoniaeeae 
Begonia prineeae Gilg. 2 
Boraginaeeae 
Cordia afrieana Lam. 1 3f 
Caes alpini aeeae 
Braehystegia bussei Harms 2 5 5 5 
B. X longifolia Benth. 2 
Piliostigma thoningii (Sehumaeh .) Mi lne-Redh. 1 
Celastraeeae 
Salacia sp. 2 
Combretaeeae 
Combretum molle G. Don 5 5 2 
Composi tae 
Aspilia pluriseta Sehweinf. ssp. gondensis 
(0. Hoffm.) Wild. 2 
A. rudis Oliv. & Hiern spp. rudis 1 
Bidens grantii (Oliv.) Sherff 5 1 
Veronia eolorata (Welld.) Drake 3f 
Connaraeeae 
Byrsoearpus eoeeineus Sehum. & Thonn. 3e 
Convolvulaeeae 
Ipomoea eairiea (L. ) Sweet 2 
A.IS 
Part eaten 
F S B L P R C G 
L eriocarpa R. Br. 1 
1. obscura (L. ) Ker-Gaw1 2 
Merrernia pterygocau1os (Choisy) Hall. L 1 
Ebenaceae 
Diospyros sp. 4 4 
Euphorbiaceae 
Acalypha ornata A. Rich. 5 2 
Arttidesmavenosum Tu1. 1 
Bride1ia micrantha (Hochst. ) Baill. 1 
B~ sc1eroneura Mue11. Arg. 1 
Croton sylvaticus Hochst. 4 
Hymenocardia acida Tu1. 2 2 2 
11 Phyllanthus mue11eriartus (0. Kuntze) Eng1. 1 5 2 
Securinega virosa (Willd.) Pax & K. Hoffm. 3f 
Uapacakirkiana Mue11. Arg. 1 2 
U. nitida Mue1l. Arg. (= U. sans ib ari ca 
Pax ?) 1 
Flacou'rtiaceae 
F1acourtia afra (= F. indica (Birm. L) 
Merr. ?) 4 
Oncoba spinos a Forsk. 2 
Lindackeria kivuensis Bamps 2 
Hyperi caceae 
Harungana madagascariensis Poir. 1 2 
Labiatae 
Hos1undia opposita Vah1. forma 5 2 
Loganiaceae 
Arithoc1eista schweinfurthii Gi1g. 3f 
Strychnos coccu1oides Bale 1 
S ~ . inrtocua De 1. 1 
S. madagascariensis Poir. 2 
Ma1vaceae 11 
Azanza garckeana (F. Hoffm.) Exell & Hillc. 1 
Hibiscus surratensis L. 4 2 4 
Hibiscus sp. 4 
Pavonia urens Cav. var. ~labrescens (U1br. ) 
Brenan 2 
Menispermaceae 
Cissampe10s sp. 4 
A.IS 
Part eaten 
F S B L P R C G 
L .. ' R. Br. 1 en_ocarpa 
1. obscura (L. ) Ker-Gawl 2 
Merremia pterygocaulos (Choisy) Hall. L 1 
Ebenaceae 
Diospyros sp. 4 4 
Euphorbiaceae 
Acalypha ornata A. Rich. 5 2 
Arttidesmavenosum Tul. 1 
Bridelia micrantha (Hochst. ) Baill. 1 
B~ scleroneura Muell. Arg. 1 
Croton sylvaticus Hochst. 4 
Hymenocardia acida Tul. 2 2 2 
Phy 11 an th us muellerianus (0. Kuntze) Engl. 1 5 2 
Securinega virosa (Willd.) Pax & K. Hoffm. 3f 
Uapaca kirkiana Muel1. Arg. 1 2 
U. nitida Muell. Arg. (= U. sans ib ari ca 
Pax ?) 1 
Flacourtiaceae 
Flacourtia afra (= F. indica (Birm. L) 
Merr. ?) 4 
Oncob a sp inos a Forsk. 2 
Lindackeria kivuensis Bamps 2 
Hypericaceae 
Harungana madagascariensis Poir. 1 2 
Labiatae 
Hoslundia opposita Vahl. forma 5 2 
Loganiaceae 
Artthocleista schweinfurthii Gilg. 3f 
Strychrtos cocculoides Bak. 1 
S ~ . inrtocua De 1. 1 
S. madagascariensis Poir. 2 
Malvaceae 
Azanza garckeana (F. Hoffm.) Exell & Hillc. 1 
Hibiscus surratertsis L. 4 2 4 
Hibiscus sp. 4 
Pavonia urens Cav. var. ilabres cens (Ulbr. ) 
Brenan 2 
Menispermaceae 
Cissampelos sp. 4 
Part eaten 
FSBLPRCG 
Stephartia abyssinica (Dill. & A. Rich.) Walp. 5 
Tiliacora funifera (Miers. Oliv.) 2 
Tinospora caffra (Miers.) Troupin 
Triclisia sacleuxii (Pierre) Diels 
Mirnosaceae 
Acacia hockii De Wild. 
A. sieberiana DC. var. verrnoesenii (De Wild.) 
Albizia glaberrirna (Schumach. & Thonn.) Benth. 
\Tar. glabrescens (Oliv.) Brenan 
A. gummifera (J. F. Grnel.) C. A. Srn. 
Entada abyssinica A. Rich. 
Newtonia buchananii (Bak.) Gilb. & Bout. 
Moraceae 
Antiaris toxicaria (Pers.) Lesch. 
Bos<{ueiopsis luj ae De Wi Id. ve 1. sp. aff. ? 
Chlorophora excelsa (Welw.) Benth. & Hook. 
Ficus asperifolia Miq. ? 
F. brachylepis Hiern ? 
F. capensis Thunb. 
F. congensis Engl. 
F. depauperata Sirn. ? 
F. exasperata Vahl. 
F; gnaphalocarpa (Miq.) A. Rich. 
F. ingens Miq. 
F. kitubalu Hutch. (= F. subcalcarata 
Warb. & Schweinf.) 
F. lukanda Ficalho 
F. ottoniifolia (Miq.) Miq. e descr. 
F. polita (Miq.) Vahl. 
F. thoningii Blurne 
F. urceolaris Hiern(= F. storthophylla 
Warb .) 
F. vallis - choudae Del. 
Morus lactea (Sirn.) Mildbr. 
Myrianthus arboreus P. Beauv. 
Myristicaceae 
Pycnanthus angolensis (Welw.) Exell. 
Myrtaceae 
Syzigiurn gu~neense (Willd.) DC 
1 
2 
1 
2 
2 
2 
2 
2 
2 
1 
2 
2 
1 
2 
1 
2 
1 
1 
2 
1 
1 
1 
1 
4 
1 1 
2 
2 
2 
5 
3f 
2 
2 
2 
1 
1 
2 
1 
2 
1 
2 
2 
1 
2 
1 3f 
1 
2 
-Part eaten 
F S B L P R C G 
Stephartia abyssinica (Dill. & A. Rich.) Wa1p. 5 
Ti1iacora funifera (Miers. 01iv.) 2 
Tinospora caffra (Miers.) Troupin 
Tric1isia sac1euxii (Pierre) Die1s 
Mirnosaceae 
Acacia hockii De Wild. 
A. sieberiana DC. var. verrnoesenii (De Wild.) 
A1bizia glaberrirna (Schumach. & Thonn.) Benth. 
var. glabrescens (Oliv.) Brenan 
A. gummifera (J. F. Grne1.) C. A. Srn. 
Entada abyssinica A. Rich. 
Newtonia buchananii (Bak.) Gi1b. & Bout. 
Moraceae 
1 
2 
Antiaris toxicaria (Pers.) Lesch. 1 
Bosqueiopsis 1uj ae De Wi 1d. veL sp. aff.? 2 
Ch10rophora exce1sa (We1w.) Benth. & Hook. 
Ficus asperifo1ia Miq. ? 
F. brachy1epis Hiern ? 
F. capensis Thunb. 
F. congensis Eng1. 
F. depauperata Sirn. ? 
F. exasperata Vah1. 
F; gnapha 10 carp a (Miq.) A. Ri ch. 
F. ingens Miq. 
F. kituba1u Hutch. (= F. subca1carata 
Warb. & Schweinf.) 
F. 1ukanda Fica1ho 
F. ottoniifo1ia (Miq.) Miq. e descr. 
F. po1ita (Miq.) Vah1. 
F. thonirtgii B1urne 
F. urceo1aris Hiern(= F. storthophy11a 
Warb. ) 
F. va11is - choudae Del. 
Morus 1actea (Sirn.) Mi1dbr. 
Myrianthus arboreus P. Beauv. 
Myristicaceae 
Pycnanthus ango1ensis (We1w.) Exe11. 
Myrtaceae 
syzigium gU1.neense (Willd.) DC 
2 
2 
2 
2 
2 
1 
2 
2 
1 
2 
1 
2 
1 
1 
2 
1 
1 
1 
1 
4 
1 1 
2 
2 
2 
5 
3f 
2 
2 
2 
1 
1 
2 
1 
2 
1 
2 
2 
1 
2 
1 3f 
1 
2 
'I 
Oleaceae 
Schrebera alata (Hochst.) Welw. 
Papilionaceae 
Aeschynomene baurnii Harms. 
Afrotmosia angolensis 
Baphia capparidifolia Bak. ssp. multiflora 
(Harms.) Brurnrnitt 
Crotalaria lachnophora A. Rich. 
Dalbergia malangensis E. P. Souza 
Desmodi urn sp. 
Eriosema psoraleoides G. Don. 
~eorautanenia mitis (A. Rich.) Verdc. 
Pterocarpus angolensis DC. 
P. tinctorius Bak. var. tinctorius 
Vigna oblongifolia A. Rich. var. parviflora 
(Bak.) Vdc. 
V. racemosa (G. Don.) Hutch. & Dale 
Polygalaceae 
Carpolobia alba G. Don. 
Protaceae 
Protea welwitschii Engl. 
Rharnnaceae 
Maesopsis em~nl~ Engl. 
Rhizophoraceae 
Anisophyllea boehrnii Engl. & Brehm. 
A. pomifera Engl. & Brehm. 
Rosaceae 
Parinari curatellifolia Benth. 
Rubus pirtnatus Willd. 
Rubiaceae 
Canthium crassum Hiern 
C. hispidum Benth. 
Mussaenda arcuata Poir 
Pavetta olivierana 
Sabicea orientalis Wernham 
Vangueria ? acutiloba Robyns 
Sapindaceae 
Allophylus sp. 
A.17 
Part eaten 
FSBLP RCG 
3d 
3c 
1 
2 
1 
3c 
2 
1 
4 
4 
2 
2 
4 
1 
2 
4 
221 
2 
4 
5 
3a 
4 
5 1 2 
2 2 
2 
4 
3f 2 
1 
3f 
2 
I 
I1 
1 i 
I 
p 
~-
A.18 
Part eaten 
F S B L P R C G 
Deinbollia fulvotomentella Bale f. 3f 
Paullinia pinnata L. 5 
Zanhagolungensis Hiern 4 
Sapotaceae 
Afrosersalisia cerasifera (Welw.) Aubrev. 2 
Pachystela sp. 4 
Sterculiaceae 
Dombeya ? r6tundifolia Mast. 2 
sterculia mhosya Engl. 2 
S. quinqueloba (Garck.) K. Schum. 1 
S. tragacantha Lindl. 2 1 2 
Tiliaceae 
Grewia mollis Juss. 5 
G. P latyclada K. Schum. 1 
Umbelliferae 
Steganotaenia araliacea Hochst. 1 
Verbenaceae 
Clerodendrum schweinfurthii Guerke 1 
Lippia p licata Bak. 4 
Vitex doniana Sweet 2 
V. ferruginea Schumach. & Thonn. 3c 
V. fischeri Guerke 4 
V. madiensis Oliv. var. milanjiensis 
(Bri tt.) Pieper 3c 
Vitaceae 
Ampelocissus cavicaulis (Bak. ) Planch 1 
Cyphostemma sp. 3b 
ANGIOSPERMAE: MONO COTYLE DONS 
Dioscoreaceae 
Dioscorea odoratissima Pax. 2 2 
D. schimperana Kunth. 2 
Gramineae 
P ani cum maximum J acq . 1 
Phragmites mauritianus Kunth. 1 
Palmae 
Elaeis guineensis Jacq. 1 2 1 5 1 
smi lacaceae 
Smilax kraussiana Meisn. 1 2 
Zingiberaceae 
Afromomum sp. 1 
Afromomum sp. 2 
Costus afer Ker 
* The root of P. microcarpa was chewed. 
2 . ANIMAL FOODS 
INSECTS 
Dictyoptera 
Mantidae (unidentified sp.) 
Isoptera 
Macr6termes bellicosus 
Pseudacartthotermes rnilitaris 
Hemiptera 
Paracopium glabricorne (in 
galled flowers of Cl er od end rum 
schweinfurthii) 
Phytolyma lata (in galled leaves 
of Ch16rophora excelsa) 
Lepidoptera 
Moth (unid. sp.) eating leaf of 
Annona senegalensis 
Moth (unid. sp.) eating leaf of 
Baphia capparidifolia 
Hymenoptera 
Blastophaga (various sp~ in 
fructifications of Ficus spp.) 
Crematogaster sp. 
Dorylus (Anornrna) nigricans 
Oecophylla longinoda 
Apis mellifera 
Coleoptera 
Cerambycidae (unidentified sp.) 
A. J~ 
Part eaten 
FSBLPRCG 
Part eaten 
Eggs 
3c 
3c 
Soldiers, 
(workers) , 
reproducti ves 
Reproductives 
Larvae 
Larvae 
Larvae 
Larvae 
All stages 
Soldiers and 
workers? 
Soldiers and 
workers 
Workers etc. 
Honey, (workers 
eaten 
incidentally) 
Larvae 
2 
2 
2 
Observation No. of 
category obsvns. 
4 
1 
1 
1 
1 
1 
2 
1 
1 
1 
1 
1 
4 
? 
Abundant 
<20 
Abundant 
Abundant 
Abundant 
< 10 
Abundant 
<20 
Many 
Many 
Many 
2 
A.20 
Stage eaten Observation No. of 
category obsvns. 
BIRDS 
Gypohierax angolensis Palm-nut 
vulture Eggs, young 3e < 10 
Francolinus sp. Francolin Nestlings 3f < 10 
Numida meleagris Helmeted 
guinea-fowl Nest ling? 3d 1 
Treron australis Green pigeon Eggs la 1 
Picidae ? Woodpecker? Nestling 3i 1 
Motacilla aguimp African pied 
wagtai 1 Injured adult 3f 1 
Pycnonotus barbatus Ye llow- vented 
bulbul Unknown 4 1 
Ploceidae Weavers (at 
least two Nest lings 1 >15 
species) 
MAMMALS 
Potamochoerus porcus Bushpig Young 1 21 
Tragelaphus scriptus Bushbuck Young 1 18 
Colobus badius Red colobus All 1 36 
Cercopithecus ascanius Red- tail Young, adult 
monkey female 1 3 
Cercopithecus mitis Blue monkey All 1 3 
Papio anubis Baboon Young 1 23 
Pan troglodytes Chimpanzee Infant 3a 1 
3. MISCELLANEOUS FOODS 
.1 
Inorganic items 
Ash 2 >10 
Cardboard/paper 2 Many 
Cloth/matting 1 Many 
Dead leaf 2 2 
Rotten wood 3f < 10 
Soil 1 <20 
Termite mud 2 Abundant 
Water 1 Abundant 
A.2\ 
Observation No. of 
category obsvns. 
Organic items 
Blood (own) from wounds 2 2 
Faeces (own) 2 5 
Semen (eaten by Miff after being mated) 3c 1 
Urine (human) 3h 1 
Vomit 3f 
1 
Appendix 4.1 
Assumptions underlying calculation of frequency of attendance at 
predations 
i. "Predations are equally like ly to occur at any time of 
A.22. 
year." It ,,,as shown (Table 4.2) that although seasonal differences ~n 
the frequency of predations were not significant they were 
considerable. If the number of observation hours varied between months, 
variations in predation rate between months would contribute 
disproportionately to the rate calculated by lumping predations and 
observation hours across months. The effect of variations in number of 
observation hours is tested by comparing the predation rate calculated 
treating each month separately or lumping them together. Only for 
Northern males is the sample size large enough for testing: data on 
observation hours per month were only available for the period January 
1972 to August 1973. The data are shown ~n Table 4.15. 
If months are lumped, predation rate = 18/2493 = 0.0072 
predations per hour. Months are treated separately by sampling the 
rates calculated for each month and dividing by the total number of 
months (N = 20): by this method predation rate overall is calculated 
as 0.1256/20 = 0.0063 predations per hour. This result suggests that 
uneven sampling between months has resulted in the predation rate, 
calculated by ignoring monthly differences in predation rate, being 
overestimated by 0.0072 - 0.0063 x 100 = 14 7% 0.0063 • 0 • 
A.23 
Table 4.15 
Monthly variation in numbers of predations and observation hours 
The data are only for Northern males (including Sherry) as targets. 
Month Number of Number of Predations per predations attended observation hours observation hour 
Jan 1972 1 173 0.0058 
Feb " 0 52 0.0 
Mar " 0 21 0.0 
Apr " 0 56 0.0 
May " 0 63 0.0 
Jun " 0 105 0.0 
Jul " 1 202 0.0050 
Aug " 0 203 0.0 
Sep " 2 59 0.0339 
Oct " 1 116 0.0086 
Nov " 0 127 0.0 
Dec " 2 131 0.0153 
Jan 1973 1 120 0.0083 
Feb " 0 60 0.0 
Mar " 0 147 0.0 
,I Apr " 0 111 0.0 
May " 2 185 0.0108 
Jun " 4 252 0.0159 
Jul " 4 182 0.0220 
Aug " 0 128 0.0 
Sum 18 2493 0.1256 
ii. "Predations are equally likely at any time of day." The 
effect of the breru~ing of this assumption is calculated ~n a similar 
way. There is no reason to think that community-sex classes might 
differ in the extent of diurnal variation in the probability of their 
attending a predation. The data are therefore taken from all available 
predations, and are shown in Table 4.16. 
With different times of day lumped, predation rate = 25/1285 = 
0.019 predations per K hours (where K is an unknown constant, incurred 
because the number of observation hours for each hour of the day is a 
sample, not a real figure). With allowance made for variation in the 
nurrher of observation hours with different times of day (N = 14) (as for 
A.2t+ 
months), predation rate = 0.247/14 = 0.018 predations per K hours. The 
effect of lumping times of day is therefore to have overestimated the 
real predation rate by 0.019 - 0.018 0.018 = 5.6% . 
Table 4.16 
Hourly variation ln numbers of predations and observation hours 
The data are taken from Fig. 4.1. Note that the number of observation 
hours is a samp le of the actual numbers: K is an unknown constant. 
Time of day (hrs) Number of Number of Predations / (K predations attended observation hours observation hours) 
0600- 0659 0 45 0.000 
0700- 0759 1 67 0.015 
0800- 0859 2 100 0.020 
0900- 0959 4 129 0.031 
1000-1059 1 139 0.007 
1100- 1159 3 142 0.021 
1200- 1259 1 135 0.007 I' 
1300- 1359 3 107 0.028 
1400- 1459 5 91 0.055 
1500- 1559 1 94 0 . 011 
1600- 1659 3 77 0.039 
1700- 1759 1 76 0.013 
1800- 1859 0 54 0.000 
1900- 1959 0 29 0.000 
Sum 25 1285 0.247 
111. "Predations are equally likely to be attended by any 
individual within each community- sex class". It was shown earlier that 
no differences were apparent between Northern adult males in the 
frequency of killing prey: it is still possible, however, that within 
community- sex classes individuals varied in their rate of attendance at 
predations. For all predations since 1968 the number of predations 
attended by each individual lS shown in Table 4.17. 
For individuals who were present in each community- sex class 
throughout the sampled period (1968- 1973) differences in rates of 
A. 2s" 
Table 4.17 
Individual variation, within community-sex classes, of frequency 
of attendance at predations 
Data are from all observations of predations, March 1968 - August 1973. 
Individuals who were absent for part of this period are not included 
(e. g. FL (died), GK ("transferred" between communities)). 
(a) Northern males 
Individuals: HG MK HM FB EV FG ST JJ SH 
Number of predations attended: 40 37 35 28 32 39 29 32 27 
X2 = 5.53, d. L =8,n.s. 
(b) Northern females 
Individuals: PS ML NP AT MF NV PL FF GG WK 
Number of predations attended: 17 22 13 18 16 9 13 22 21 23 
i = 11. 49, d.L 9, n.s. 
(c) Southern males 
Individuals: GO HH CH DE WW GI SF 
Number of predations attended: 15 16 20 6 12 13 19 
2 X = 9.60, d.L = 6, n.s. 
(d) Southern females 
Individuals: MD MB LB 
Number of predations attended: 8 7 7 
i= 0.09, d.f. = 2, n.s. 
attendance were not significant (Table 4.17). The best sample of 
predations attended by a class is for Northern males: these were therefore 
tested to find how lumping data from different individuals affected the 
calculation of predation rate. The data are shown in Table 4.18. Using 
the same method as above, predation rate calculated overall was 
22/3628 = 0.0061 predations per hour, and calculated separately for 
each individual was 0.0598/9 = 0.0066 predations per hour. Thus the 
effect of lumping individuals was to underes timate pred.ct.. tion rate by 
0.0066 - 0.0061 
0.0061 7.5% • 
- -,-----:- - -_. 
-
r 1I 
, 
'\ 
I; 
I' 
I 
, 
, 
I' 
11 
I; 
!i 
I 
, 
I 
I 
11\ 
I 
Table 4.18 
Individual variation in numbers of predations attended and 
observation hours 
Data are from August 1970 to ~ugust 1973. Only predations attended 
by individuals as targets are included: Sherry is included with the 
Northern males. 
Individual 
RG 
MK 
RM 
FB 
EV 
FG 
ST 
JJ 
SR 
Sum 
Number of 
predatiorts 
1 
1 
4 
2 
4 
3 
3 
4 
0 
22 
attended 
UNIVERSITY 
L1BRAR 
CAM'RIDGE 
Number of Predations per 
observation hours observation hour 
363 0.0028 
252 0.0040 
373 0.0107 
467 0.0043 
604 0.0066 
838 0.0036 
261 0.0115 
246 0.0163 
224 0.0000 
3628 0.0598 
r 
REFERENCES 
Albrecht H. & S.C. Dunnett (1971) "Chimpanzees l.n Western Africa" 
R. Piper & Co., Verlag, Munich. 
J<.,. t 
Aldrich-Blake F.P.G. (1970) "Problems of social structure in forest 
monkeys" In . ' Social behaviour in birds and mammals' (J .H. Crook, 
Ed.): 79-101, Academic Press. 
Aldrich-Blake F.P.G., T.K. Bunn, R.I.M. Dunbar & P.M. Headley (1971) 
"Obs ervations on baboons, P apio anubis, in an arid region in 
Ethiopia" Folia primat. 15: 1-35. 
Altmann S .A. (1974) "Baboons, space, time and energy" Amer ~Zool.14: 
221-248. 
Altmann S.A. & J. Altmann (1970) "Baboon ecology: African field research" 
University of Chicago Press, Chicago. 
Ardrey R. (1967) "The territorial imperative" Anthony Blond, London. 
Ashton P.S. (1969) "Speciation among tropical forest trees: some 
deductions in the light of recent evidence" Biol.J.LimLSoc. 1: 
155-196. 
Azuma S. (1973) "Acquisition and propagation of food habits in a troop 
of Japanese monkeys" In 'Behavioural regulators of behaviour l.n 
primates' (C.R. Carpenter, Ed.) Bucknell University Press. 
Azuma S. & A. Toyoshima (1962) "Progress report of the survey of 
chimpanzees in their natural habitat, Kabogo Point area, 
Tanganyika". Primates 3; 6/ - 70. 
Baker H.G. (1973) "Evolutionary relationships between flowering plants 
and animals in American and African tropical forests" In 
Meggers B.J., E.S. Ayensu & W.D. Duckworth (Ed.) 1973 'Tropical 
forest ecosystems in Africa and South America: a comparative 
review' Smithsonian Institution Press. 
Baldwin J.D. & J. Baldwin (1972) "The ecology and behaviour of squirrel 
monkeys (S&imiri oerstedi) in a natural forest in western 
Panama" Folia primat. 18: 161-184. 
Barash D.P. (1974) "The evolution of marmot societies: a general theory" 
Science 185: 415-420. 
Beatty H. (1951) "A note on the behaviour of the chimpanzee" 
J.Mammal. 32: 118. 
Bernstein 1.S. (1970) "Activity patterns in pigtail monkey groups" 
Folia primat. 12: 187-198. 
Bonte M., E. Akingeneye, M. Gashakamba, E. Mbarutso & M. Nolens (1974) 
"Influence of the socio-economic level on the conception rate 
during lactation" Int.J .Fertil. ~: 97-102. 
Bourne G.H. (1971) "Nutrition and diet of chimpanzees" The Chimpanzee 
4 : 373-400. 
REFERENCES 
A1brecht H. & S .C. Dunnett (1971) "Chimpanzees ~n Western Africa" 
R. Piper & Co., Ver1ag, Munich. 
A1drich- B1ake F.P. G. (1970) "Prob 1 ems of social structure in forest 
monkeys" In 'Social behaviour in birds and marruna1s' (J.H. Crook, 
Ed.): 79-101, Academic Press. 
A1drich-B1ake F.P.G., T.K. Bunn, R.I.M. Dunbar & P.M. Head1ey (1971) 
"Observations on baboons, P apio anubis, in an arid region in 
Ethiopia" Folia primat. 12: 1- 35. 
A1tmann S .A. ( 1974) "Baboons, space, time and energy"Amer .Zoo1.14: 
221- 248. 
A1tmann S.A. & J. A1tmann (1970) "Baboon ecology: African field research" 
University of Chicago Press, Chicago. 
Ardrey R. (1967) "The territorial imperative" Anthony Blond, London. 
Ashton P. S. (1969) "Speciation among tropical forest trees: some 
deductions in the light of recent evidence" Bio1.J.LimLSoc; 1: 
155-196. 
Azuma S. (1973) "Acquisition and propagation of food habits in a troop 
of Japanese monkeys" In 'Behavioural regulators of behaviour ~n 
primates' (C.R. Carpenter, Ed.) Buckne11 University Press. 
Azuma S. & A. Toyoshima (1962) "Progress report of the survey of 
chimpanzees in their natural habitat, Kabogo Point area, 
Tanganyika". 'PrLmates 3: 61 -70. 
Baker H.G. (1973) "Evolutionary relationships between flowering plants 
and animals in American and African tropical forests" In 
Meggers B.J., E.S. Ayensu & W.D. Duckworth (Ed.) 1973 'Tropical 
forest ecosystems in Africa and South America: a comparative 
review' Smithsonian Institution Press. 
Ba1dwin J .D. & J. Ba1dwin (1972) "The ecology and behaviour of squirrel 
monkeys (Saimiri oerstedi) in a natural forest in western 
Panama" Fo li a p rimat. 18: 161-184. 
Barash D.P. (1974) "The evolution of marmot societies: a general theory" 
Science 185: 415-420. 
Beatty H. (1951) "A note on the behaviour of the chimpanzee" 
J.Mamma1. 32: 118. 
Bernstein I.S. (1970) "Activity patterns ~n pigtail monkey groups" 
Folia primat. 12: 187-198. 
Bonte M., E. Akingeneye, M. Gashru(amba, E. Mbarutso & M. Nolens (1974) 
"Influence of the socio- economic level on the conception rate 
during lactation" InLJ .Fertil. 12.: 97- 102. 
Bourne G. H. (1971) "Nutrition and diet of chimpanzees" The Chimpanzee 
4: 373- 400. 
Brambell M. R. (1972) "Mammals: their nutrition and habitat" In 
R.N.T-W- Fiennes (Ed.) 'Biology of Nutrition' 1972, Pergamon 
Press, Oxford. 
Brown J .L. & G.H. Orians (1970) '.'Spacing patterns l.n mobile animals" 
Ann ~Rev.Ecol . Systemat. !: 239- 262. 
Brown L.H. (1971) "The relations of the Crowned Eagle Stephanoaetus 
cotonatus and some of its prey animals" Ibis 113: 240-243. 
Bygott D. (1972) "Cannibalism among wild chimpanzees" Nature 238: 
410- 411. 
Bygott D. (1974) "Agonistic behaviour and dominance in wild chimpanzees" 
Ph.D. thesis, University of Cambridge. 
Campbe11 R.M. & D.P. Cuthbertson (1963) "Factors influencing man's choice 
of foods" In D.P. Cuthbertson (Ed . ) 'Progress innutrition and 
allied sciences'. 
Chivers D.J. (1972) "The siamang and the gibbon in the Malay peninsula" 
In Rumbaugh D.M. (Ed.) 'Gibbon and siamang' vol . l: 103- 135 
(Karger, Basel). 
Chivers D.J. (1973) "An introduction to the socio- ecology of Malayan 
Forest primates" In Michael R.P. & J .H. Crook (Ed . ) 
'Comparative ecology and behaviour of primates': 101- 146, 
Academic Press , London. 
Chivers D.J. (in press) "Daily patterns of ranging and feeding in 
siamang" In Ptae. VthCong. Int .Primat .Soc. , Nagoya (Karger, 
;Basel). 
Clutton- Brock T .H . (1972) "Feeding and ranging behaviour of the red 
colobus monkey" Ph.D. thesis, University of Cambridge. 
Clutton- Brock T.H. (1973) "Feeding levels and feeding sites of red 
colobus (Colobu8 badius tephrosceles) in the Gombe National 
Park" . 
Clutton- Brock T.H. (1974) "Pr imate social organisation and ecology" 
Nature 250: 539- 542. 
Cody M.L. (1974) "Optimisation in ecology" Science 183: 1156 - 1164. 
Cowgi 11 U .M. (1972) "Death in Petodicticus" Primates "13: 251- 256. 
Croat T.B. (1969) "Seasonal flowering behaviour in Central Panama" 
AntLMissouri Bot .Gar d. 56: 295 - 307. 
Cr ook J .H. (1965) "The adaptive significance of avian social 
organizations" Symp.zool.Soc. Lond. ~: 181- 218. 
Crook J .H. (1966a) "Gelada baboon herd structure and movement: a 
comparative report" Symp.zool.Soc.Lond. 1&: 237- 258. 
Crook J.H. (1966b) "Cooperation in primates"The Eugenics Review 58: 
63- 70. 
Crook J .H. (1970a) "The socio-ecology of primates" In Crook J .H. (Ed.) 
'Social behaviour in birds and mammals' 103-166, Academic Press. 
Crook J.H. (1970b) "Sources of cooperation in animals and man" In 
'Man and beast. Comparative social behaviour' (J.F. Eisenberg 
& W. Dillon, Eds.) Smithsonian Annual Ill, Smithsonian 
Institute, WashingtoR, D.C. 
Crook J.H. (1972) "Sexual selection, dimorphism, and social 
organisation in the primates" In B. Campbe 11 (Ed.) 'Sexual 
selection and the descent of man' 231-281, Heinemann, London. 
Crook J .H. & J.S. Gartlan (1966) "Evolution of primate societies" 
Nature 210: 1200-1204. 
Dare R. (1974) "Food- sharing in the free-rangingAte:lesgeoffroyi (red 
spider monkeys) "Laboratory Primate Newsletter 13: 19-21. 
de Bournonville D. (1967) "Contribution a 1 'e'tude du chimpanz{ en 
Republique de Guine'e" Bul1.Inst.Fond.Aft.Noire 29: 1188- 1269. 
de Moor P.P. & F.E. Steffens (1972) "The movements of vervet monkeys 
(Cercopithecus aethiops) within their ranges as revealed by 
radio-tracking" J .AniIil.Ecol. 41: 677-687. 
Deag J .M. (1973) "Intergroup encounters in the wild barbary macaque 
Macaca sylvanus L." In' Comparative Ecology and behaviour of 
primates' (R.P. Michael & J.H. Crook, Ed.): 316-373. 
Deag J.M. & J.H. Crook (1971) "Social behaviour and 'agonistic 
buffering' in the wild barbary macaque Macacasylvana L." 
Folia primat. 15: 183- 200. 
Denham W.W. (1971) "Energy relations and some basic properties of 
primate social organisation" AIil~ Anthropologist 72.: 77-95. 
DeVore 1. (1963) "A comparison of the ecology and behaviour of monkeys 
and apes" In' Classification and human evolution' S .L. Washburn 
(Ed.) Aldine, Chicago. 
DeVore 1. & S.L. Washburn (1963) "Baboon ecology and human evolution" 
In 'African ecology and human evolution' F. Clark Howell & 
F. Bourliere (Ed.) Aldine, Chicago. 
Dunbar R.1.M. (1973) "The social dynamics of the gelada baboon 
Theropithecus gelada" Ph.D. thesis, University of Bristol. 
Dupuy A.R. (1970) "Sur la presence du chimpanze- dans les limites du 
Parc national du Niokolo-Koba (Senegal)" Bull~Inst.fortdartlent. 
Afr~noire A32: 1090-1099. 
Eisenberg J.F. (1963) "The behaviour of heteromyid rodents" University 
of California Publications in Zoology, Vol.69. 
Eisenberg J. F. (1973) "Mammalian social systems: are primate social 
systems unique?"Symp. IVth InLCongr.Prirtlat. 1.: 232-249, 
Karger, Basel. 
Eisenberg J.F. & R.E. Kuehn (1966) "The behaviour of Ate1es geoffroyi 
and related species" Smithson.Misc.Collect. 151: 1-63. 
Eisenberg J.F., N.A. Muckenhirn & R. Rudran (1972) "The relation 
between ecology and social structure in primates" Science 176: 
863- 874. 
Ellefson J .0. (1968) "Territorial behaviour in the connnon white-handed 
gibbon, Hy10bates 1ar Linn." In 'Primates: studies in adaptation 
and variability" P.c. Jay (Ed.) Ho1t, Rinehart & Winston, 
New York. 
Em1en J.M. (1966) "The role of time and energy in food preference" 
Amer~Natur. 100: 611-617. 
Em1en J .M . (1973) "Ecology: an evolutionary approach" Addison- Wes1ey, 
Reading, Massachusetts. 
Ewer R.F. (1968) "Ethology of mannna1s" E1ek Science, London. 
Forbes G.B. (1962) "Sodium" In 'Mineral metabolism, vo1.II: The 
elements (Part B)' Comar C.L. & F. Bronner (Ed.) 67-86, 
Academic Press, London. 
Fossey D. (1974) "Observations on the home range of one group of 
mountain gorillas (Gorilla gorilla beringei)"Anim~Behav. 22: 
568- 581. 
Free1and W.J. & D.H. Janzen (1974) "Strategies in herbivory by mannna1s: 
the role of plant secondary compounds "Am~Nat. ·108: 269-289. 
Frisch R.E. & J.W. McArthur (1974) "Menstrual cycles: fatness as a 
determinant of minimum weight for height necessary for their 
maintenance or onset"Science 185: 949-951. 
Gardarsson A. & R. Moss (1970) "Food selection by Icelandic ptarmigan" 
In Watson A. (Ed.) 'Animal popu1ations in relation to their 
natural resources' Symp~Brit.Eco1.Soc~10: 47-71. 
Gart1an J.S. & C.K. Brain (1968) "Ecology and social variability in 
Cercopithecus aethiops and Cercopithecus mitis" In Jay P. (Ed.) 
'Primates: studies in adaptation and variability' 253- 292, 
Ho1t, New York. 
Gautier- Hion A. (1973) "Social and ecological features of ta1apoin 
monkey - comparisons with sympatric cercopithecines" In 
Michae1 R.P. & J.H. Crook (Ed.): 147-170. 
Goodall J. (1963) "Feeding behaviour of chimpanzees: a pre 1iminary 
report" Symp.ioo1.Soc.Lond. 10: 39 - 48. 
Goodall J. (1965) "Chimpanzees of the Gombe Stream Reserve" In 
I. DeVore (Ed.) 'Primate behaviour' Ho1t, Rinehart & Winston, 
New York. 
Goodman D. (1974) "Natural selection and a cost ceiling on. reproductive 
effort" Am.Nat. 108: 247-268. 
,,--
Goss-Custard J.D. (1970) "Feeding dispersion in some over-wintering 
wading birds" In Crook J .H. (Ed.) 'Social behaviour in birds 
and mammals': 3-35. 
Goss-Custard J.D., R.I.M. Dunbar & F.P.G. Aldrich-Blake (1972) 
"Survival, mating and rearing strategies in the evolution of 
primate social structure" Foliaptirilat.17: 1-19. 
Gunther M. (1971) "Infant feeding" Penguin, Harmondsworth. 
Haddow A.J. (1952) "Field and laboratory studies on an African monkey, 
·Cetcopithecusascanius schmidti Matschie" Ptoc~zooLS6c .Lond. 
122: 297-394. 
Hairston N.G., F. Smith & L.B. Slobodkin (1960) "Community structure, 
population control and competition" Am.Nat~~: 421-425. 
Hall K.R.L. (1963) "Variations in the ecology of the chacma baboon 
(Papio ursirtus) "Syrilp.zoci1.Soc.L6nd.lO: 1-28. 
Hall K.R.L. (1965) "Social organisation of the old world monkeys and 
apes"Syrilp ~zool.Soc~Lond. 14: 265-289. 
Hamilton W.D. (1964) "The genetical evolution of social behaviour" 
J~Theoret.Biol. 7: 1- 52 . 
Hamilton W.J. & K.E.F. Watt (1970) "Refuging" Ann~Rev.Eco1.Systemat. 1: 
263-286. 
Harding R.S.O. (1973) "Predation by a troop of olive baboons (Papio 
anubis)" AIilet~J.Phys.Anthrop. 38: 587-592. 
Hartley c.w.s. (1967) "The oil palm" Longman, London. 
Hill W.G.O. (1962) "Primates. Comparative anatomy and taxonomy. Vo1.V: 
Gebidae. Part B" Edinburgh, University Press. 
Hinde R.A. (1956) "The biological significance of the territories of 
birds" Ibis 98: 340-369. 
Hinde R.A. (1974) "Biological bases of human social behaviour" McGraw-
Hill, New York. 
Hinde R.A. (in press, a) "The nature of social structure" In D.A. Hamburg 
& J. van Lawick-Goodall (Eds.) 'Perspectives on human evolution, 
Vol.IV: the behaviour of great apes' Holt, Rinehart & Winston, 
New York. 
Hinde R.A. (in press, b) "Evolutionary aspects of animal behaviour" In 
G.P. Baerends, C. Beer & A. Manning (Eds.) 'Festschrift for 
N. Tinbergen'. 
Hladik C.M. (1973) "Alimentation et activit{ d'un groupe de chimpanzEfs 
rEfintroduits en for~t gabonaise" La Terre et La Vie 27: 343-413. 
Rladik C.M., A. Hladik, J. Bousset, P. Valdebouze, G. Viroben & 
J. Delort-Laval (1971) "Le r~gime alimentaire des primates de 
1 file de Barro-Colorado (Panama)" Folia primat. 16: 85-122. 
I 
I 
Horn H.S. (1968) liThe adaptive significance of colonial nesting in the 
Brewer's blackbird (Euphagus cyanocephalus)" Ecology 49: 682-694. 
Houston D.C •. (1974) "Food searching in griffon vultures" E.Afr.Wildl.J. 
12: 63-77. 
Itani J. & A. Suzuki (1967) lIThe social unit of chimpanzees" Primates 8: 
355-381. 
Izawa K. (1970) "Unit groups of chimpanzees and their nomadism in the 
savanna woodland" Primates 11: 1-46. 
Izawa K. & J. Itani (1966) "Chimpanzees in Kasakati Basin, Tanganyika. 
(1) Ecological studies in the rainy season, 1963-1964" Kyoto 
Univ.Afr.Stud. 1: 73-156. 
Jarman P.J. (1974) "The social organisation of antelope in relation to 
their ecology" Behaviour 48: 215-267. 
Jarman M.V. & P.J. Jarman (1973) "Daily activity of impala" E.Afr. 
Wildl.J. 11: 75-92. 
Jolly A. (1972a) liThe evolution of primate behaviour" Macmillan, New 
York. 
Jolly A. (1972b) "Troop continuity and troop spacing in Propithecus 
ver:teauxi and Lemur catta at Berenty (Madagascar)" Folia Primat. 
17: 335-362. 
Jones C. (1972) "Natural diets of wild primates" In T-W-Fiennes R.N. 
(Ed.) 'Pathology of Simian primates' I: 58-77 (Karger, Basel). 
Jones C. & J. Sabater pi (1971) "Comparative ecology of Gorilla gorilla 
(Savage and Wyman) andPartt:toglodytes (Blumenbach) in Rio Muni, 
West Africa" Bibliotheca P:timato16gica 13. S. Karger, Basel, 
New York. 
Kano T. (1971) "The chimpanzee of Filabanga, Western Tanzania" Primates 
12(3-4): 229-246. 
Kaufmann J .H. (1974) "Social ethology of the whiptail wallaby, 
'Macropusparryi, in northeastern New South Wales" Anim.Behav. · 
22: 281-369. 
Kavanagh M. (1972) "Food sharing behaviour within a group of douc 
monkeys (Pygathrix nemaeusnemaeus)" Nature 239: 406-407. 
Kawabe M. (1966) "One observed case of hunting behaviour among wi Id 
chimpanzees living in the savanna woodland of Western Tanzania" 
Primates 7: 393-396. 
Kawai M. (1965) "Newly acquired precultural behaviour of the natural 
troop of Japanese monkeys on Koshima islet" Primates 6: 1-30. 
Kawamura S. (1959) "The process of sub-culture propagation among 
Japanese macaques" Primates 2: 43-60. 
!<.,. i 
Kleiman D. & J.F. Eisenberg (1973) "Comparisons of canid and felid 
social systems from an evolutionary perspective" Artim.Behav. 
21 : 637- 659. 
Klein L.L. & D.J. Klein (1973) "Social and ecological contrasts between 
four taxa of neotropical primates (Ateles belzebuth, Alouatta 
seniculus, Saimiti sciureus, Cebus · ape 11 a) " IXth International 
Congress of Anthropological and Ethnological Sciences, Chicago. 
Kortlandt A. (1962) "Chimpanzees in the wild" Sci.Amer. 206: 128-138. 
Kortlandt A. (1967) "Experimentation with chimpanzees in the wild" 
In D. Starck, R. Schneider & H.-J. Kuhn (Eds.) 'Neue ergebnisse 
der primatologie' Fischer, Stuttgart. 
Kortlandt A. (1972) "New perspectives on ape and human evolution" 
Stichting Voor Psychobiologie, Amsterdam. 
Kortlandt A. & M. Kooij (1963) "Protohominid behaviour 1n primates" 
Symp.zoo1.Soc~Lond. 10: 61 - 88. 
Kortlandt A. & J.C.J. van Zon (1969) "The present state of research on 
the dehumanisation hypothesis of African ape evolution" Proc. 
IInd Int.Congr.Primat. Atlanta 1968 2: 10-13. Karger, Basel. 
Krebs D. (1971) "Infrahuman altruism" Psych.Bull. !.i: 411- 414. 
Krebs J .R. (1973) "Behavioural aspects of predation" In P.P .G. Bateson 
& P.H. Klopfer (Eds.) 'Perspectives in ethology' Plenum Press, 
New York. 
Krebs J .R. (1974) "Colonial nesting and social feeding as strategies for 
exploiting food resources in the Great blue heron (Ardea 
hetodias)" Behaviour 51: 99-134 . 
Kruuk H. (1972) "The spotted hyena: a study of predation and social 
behaviour" University of Chicago Press, Chicago and London. 
Kununer H. (1968) "Social organisation of hamadryasbaboons" University 
of Chicago Press, Chicago. 
Kummer H. (1971) "Primate societies: group techniques of ecological 
adaptation" Aldine Atherton, Chicago. 
Lack D. (1954) "The natural regulation of animal numbers" Oxford 
University Press, London. 
Lack D. (1966) "Population studies of birds" Oxford University Press, 
London. 
Lack D. (1968) "Ecological adaptations for breeding 1n birds" Methuen, 
London. 
Lazarus J. (1972) "Natural selection and the functions of flocking in 
birds: a reply to Murton" Ibis 114: 556-558. 
Leck C.F. (1972) "Seasonal changes in feeding pressures of fruit - and 
nectar- eating birds in Panama" The Condor 74: 54- 60. 
,t. 8 
Lee R.B. (1968) "What hunters do for a living, or, How to make out on 
scarce resources" In Lee R.B. & I. DeVore (Eds.) 'Man the 
hunter' 30-48, Aldine-Atherton, Chicago. 
Lind E.M. & M.E.S. Morrison (1974) "East African vegetation" Longman, 
London. 
Livingstone D.A. (1967) "Postglacial vegetation of the Ruwenzori 
Mountains in euqatorial Africa" Ecol ~Monogr. '}2: 25-52. 
Livingstone D.A. (1971) "Speculations on the climatic history of 
mankind" Amer. Scientist 59: 332-337. 
Lorenz K. (1966) "On aggression" Methuen, London. 
Lorenz R. (1971) "Goeldi' s monkey Callimico goeldii Thomas 1904 preyl.ng 
on snakes" Folia primat. ..!2.: 133-142. 
MacArthur R.H. & E.R. Pianka (1966) "On optimal use of a patchy 
environment" Arner.Natur. 100: 603-609. 
MacKinnon J. (1974) "The behaviour and ecology of wild orang-utans 
Pongo pygmaeus" Anim.Behav.22: 3-74. 
McGinnis P .R. (1973) ' "Patterns of sexual b'ehaviour in a connnunity of 
free-living chimpanzees" Ph.D. thesis, Univ~rsity of Cambridge. 
McGrew W.C. (in press, a) "Evolutionary implications of sex differences 
in chimpanzee predation and tool use" In Hamburg D.A. & J. van 
Lawick-Goodall (Eds.) 'Perspectives on human evolution, Vol.IV: 
the behaviour of great apes' Holt, Rinehart & Winston, New York. 
McGrew W.C. (in press, b) "Patterns of plant food sharing by wild 
chimpanzees" In proc. Vth Congr. Int .Primat Soc. Nagoya. 
Karger, Basel. 
McNab B. (1963) "Bioenergetics and the determination of home range size" 
Am.Nat. 97: 133- 140. 
Mech L.D. (1970) "The wolf: the ecology and behaviour of an endangered 
species" Natural History Press, New York. 
Meinertzhagen R. (1959) "Pirates and predators" Oliver & Boyd, London. 
Menzel E. w. (1971) "Communication about the environment in a group of 
young chimpanzees" Folia primat. 12.: 220-232. 
Menzel E.W. (1973) "Chimpanzee spatial memory organisation" Science 182: 
943-945. 
Menzel E.W . & W.A. Draper (1965) "Primate selection of food by size: 
visible versus invisible rewards" J~Cornp.PhysioLPsych. 59: 
231-239. 
Michener J. R. (1951) "Territorial behaviour and age composition in a 
population of mockingbirds at a feeding station" Condor 53: 
276- 283. 
'I 
I 
Morre11 G.M. & J .R.G. Turner (1970) "Experiments on ml.TIUcry. I. The 
response of wild birds to artificial prey" Behaviour 36: 116-130. 
Murton R.K. (1971) "Why do some bird species feed in flocks?" Ibis 113: 
534-536. 
Myer J.S. (1969) "Early exper~ence and the development of mouse killing 
by rats" J ~Comp .PhysioI.PsychoI. 67: 46 - 49. 
Nisbet I.C.T. (1973) "Courtship-feeding, egg-size and breeding success 
in common terns" Nature 241: 141- 142. 
Nishida T. (1970) "Social behaviour and relationships among wild 
chimpanzees of the Mahali mountains"Ptirnates11: 47-87. 
Nishida T. (1973) "The ant-gathering behaviour by the use of tools 
among wi Id chimpanzees of the Mahali mountains" Journal of Human 
Evolution 2: 357-370. 
Nishida T. (1974) "Ecology of wild chimpanzees" In Ohtsuka, Tanaka & 
Nishida (Eds.) 'Human Ecology' Kyoritsu-Shuppan, Tokyo 
(In Japanese) • 
Nishida T. & K. Kawanaka (1972) "Inter unit- group relationship among 
wild chimpanzees of the Mahali mountains" Kyoto Urtiv.Afr.Stud. 
7: 131-169. 
Nissen H.W. (1931) "A field study of the chimpanzee: observations of 
chimpanzee behaviour and environment in western French Guinea" 
Cornp~Psych.Monogr. ~: 1-122. 
Nissen H.W. & M.P. Crawford (1936) "A preliminary study of food-sharing 
behavior in young chimpanzees" J ~C6mp.PsychoI. g: 383-419. 
Noble G.K. (1939) "Dominance in the life of birds" Auk 56: 263-273. 
Parker G.A. (1974) "Assessment strategy and the evolution of fighting 
behaviour" J. theor. BioI. 47: 223-243. 
Pianka E.R. (1970) "On r - and K-selection" Amer.Natur. 104: 592-597. 
Pu11iam H.R. (1974) "On the theory of optimal diets" Am.Nat. 108: 59-74. 
Rahm U. (1972.) "L'emploie d'outils par les chimpanze's de l'ouest de la 
Cote-d'Ivoire" Terre et la Vie 25: 506-509. 
Reynolds V. & F. Reynolds (1965) "Chimpanzees of the Budongo forest" 
In DeVore I. (Ed.) 'Primate behaviour' Holt, Rinehart & Winston, 
New York. 
Richard A. F. (1973) "Social organisation and ecology of Propithecus 
verreauxi Grandidier 1867" Ph.D. thesis, University of London. 
Richards P. W. (1966) "The tropical rain forest" Cambridge University 
Press. 
Ripley S. (1970) "Leaves and leaf-monkeys. The social organisation of 
foraging in gray langurs (Presbytisente11usthetsites)" In 
Napier J.R. & P.H. Napier (Eds.) 'Old world monkeys. Evolution, 
systematics and behaviour' 481- 509, Academic Press , New York. 
Robinson T. (1974) "Metabolism and function of alkaloids ~n plants" 
Science 184: 430-435. 
Rodman P.S. (1973) "Population composition and adaptive organisation 
among orang-utans of the Kutai reserve" In Michae1 R.P. & 
J.H. Crook (Eds.) 'Comparative ecology and behaviour of primates' 
Academic Press. 
Rowell T.E. (1966) "Forest living baboons in Uganda" J.Zoo1. Lond. 149: 
344-364. 
Rudran R (1973) "Adult male replacement in one-male troops of purple-
faced 1angurs (Presbytis senex sertex) and its effect on 
population structure" Foliaprimat. 19: 166 - 192. 
Schaller G.B. (1963) "The mountain gorilla" Chicago University Press, 
Chicago. 
Schal1er G.B. (1965) "The behaviour of the mountain gorilla" In 
DeVore I. (Ed.) 'Primate behaviour: field studies of monkeys 
and apes' Holt, Rinehart & Winston, New York. 
Schaller G.B. (1972) "The Serengeti lion: a study of predator-prey 
relations" Chicago University Press, Chicago. 
Scha11er G.B. & G.R. Lowther (1969) "The relevance of carn:Lvore 
behaviour to the study of early hominids" S.W~Journ.Anthrop. 
25: 307-341. 
Schoener T. W. (1971) "Theory of feeding strategies" Ann. Rev .Eco1.Systemat. 
2: 369-404. 
Sinc1air A.R.E. (1970) "Studies of the ecology of the East African 
buffalo" D.Phi 1. thesis, Oxford University. 
Smith C.C. (1968) "The adaptive nature of the social organisation in the 
genus of tree squirrels, Tamasciurus" Eco1.Monogr. 38: 31-64. 
Smith J.N .M. (1974) "The food searching behaviour of two European 
thrushes. 11: The adaptiveness of the search patterns" Behaviour 
59: 1-61. 
Smythe N. (1970a) "Relationships between fruiting seasons and seed 
dispersal methods in a Neotropica1 forest" Am;Natur. 104: 25-35. 
Smythe N. (1970b) "On the existence of 'pursuit invitation' signals in 
mammals "Ariler .Natur. ,104: 491- 494. 
Snow D.W. (1964) "A possible selective factor in the evolution of 
fruiting seasons in tropical forest" oikos 15: 274-281. 
Spinage C.A. (1968) "A quantitative study of the daily activity of the 
Uganda defassa waterbuck" E .Afr ~Wi ld1.J ~ 6: 89-93. 
Stebbins G.L. (1971) "Adaptive radiation of reproductive characteristics 
~n Angiosperms. II: seeds and seedlings" Anrt.Rev~Ecol.Systemat. 
2: 237-260. 
Stoltz L.P. & G. S. Saayman (1970) "Ecology and behaviour of baboons in 
the Northern Transvaal" Ann. Transvaal Mus. 26: 99-143. 
Struhs aker T. T. (1969) "Correlates of ecology and social organisation 
among African cercopithecines" Folia primat. .!!: 80-118. 
Struhsaker T. & P. Hunkeler (1971) "Evidence of tool-using by 
chimpanzees in the Ivory Coast" Folia primat. 12: 212-219. 
Strum S.C. (in press) . "Primate predation: an interim report on the 
development of a tradition in a troop of olive baboons" 
Science. 
Sugiyama Y. (1966) "An artificial social change in a hanumim langur 
troop (Presbyttisentellus)" Primates J....: 41-72. 
Sugiyama Y. (1967) "Social organisation of hanurnan langurs" In 
Altmann S.A. (Ed.) 'Social communication among primates' 221-236, 
Univ. of Chicago Press. 
Sugiyama Y. (1968) "Socal organisation of chimpanzees ln the Budongo 
Forest, Uganda" Primates 9: 225-258. 
Sugiyama Y. (1973) "The social structure of wild chimpanzees: a review 
of field studies" In Michael R.P. & J.H. Crook (Eds.) 
'Comparative ecology ~nd behaviour of primates' Academic Press, 
London. 
Suzuki A. (1969) "An ecological study of chimpanzees ln a savanna 
woodland" Primates 10: 103-148. 
Suzuki A. (1971) "Carnivority and cannibalism observed among forest-
living chimpanzees" J ~Anthrop ;Soc. Nippon. JJ... 
Teleki G. (1973) "The predatory behaviour of wild chimpanzees" 
Bucknell University Press, Lewisburg. 
Thorington R.W. (1967) "Feeding and activity of Cebus and sairniri in a 
Colombian forest" In Starck D., R. Schneider & H-J. Kuhn (Ed.) 
180-184 'Neue Ergebnisse der Primatologie' Fischer, Stuttgart. 
Tobias P. (1964) "Bushman hunter-gatherers: a study in human ecology" In 
Davis D. (Ed.) 'Ecological studies in southern Africa' 67-86, 
W. Junk, The Hague. 
Trivers R.L. (1971) "The evolution of reciprocal altruism" QuarLRev.Bio1. 
46: 35-57. 
Trivers R.L. (1972) "Parental investment and sexual selection" In 
Campbell B. (Ed.) 'Sexual selection and the descent of man' 
136- 179, Heinemann, London. 
van Lawick-Gooda11 J. (1967) ''My friends the wi ld chimpanzees" 
National Geographic Society, Washington. 
van Lawick-Goodall J. (1968) "The behaviour of free-living chimpanzees 
in the Gombe Stream Reserve" Anim.Behav ~Monogr. 1(3): 165 - 311. 
van Lawick-Goodall J. (1971) "In the shadow of man" Collins, London. 
van Lawick-Goodall J. (1973) "Cultural elements ~n a chimpanzee 
community" Symp. 'IVth Int.Congt.Primat. 1: 144-184. Karger, Basel. 
van Lawick-Goodall J. (in press) "The behaviour of the chimpanzee" In 
Eibl-Eibesfeldt I. .(Ed.) 
Vessey S.H. (1968) "Interactions between free-ranging groups of rhesus 
monkeys" Folia primat. ~: 228-239. ' 
Ward P. & A. Zahavi (1973) "The importance of certain assemblages of 
birds as 'information-centres' for food- finding" Ibis 115: 517-534. 
Warren J.M., J.M. Abplanalp & H.B. Warren (1967). "The development of 
handedness in cats and rhesus monkeys" In Stevenson, H., 
E.H. Hess & H. Rheingold (Eds.) 'Early behaviour: comparative 
developmental approaches' Wiley, New York. 
Warren J .M., P .R. Cornwell & H.B. Warren (1969) "Unilateral frontal 
lesions and learning by rhesus monkeys" J ~ Comp .Physiol.Psychol. 
69: 498-505. 
Ware D.M. (1971) "Predation by rainbow trout (Salmo gairdneri). The 
effect of experience" J.Fish.Res.Bd.Can.28: 1847...;.1852 . 
Watson A. & R. Moss (1970) "Dominance, spacing behaviour and aggression 
in relation to population limitation in vertebrates" In 
Watson A. (Ed.) 'Animal populations in relation to their food 
resources' 167-220. Blackwells, Oxford and Edinburgh. 
Westoby M. (1974) "An analysis of diet selection by large generalist 
herbivores" Am.Nat. 108: 290-304. 
Williams G.C. (1966) "Adaptation and natural selection: a critique of 
some current evolutionary thought" Princeton University Press. 
Wilson C.C. (1972) "Spatial factors and the behaviour of nonhuman 
primates" Folia primate 18: 256-275. 
Wilson E.O. (1971) "The insect societies" Harvard University Press, 
Cambridge, Mass. 
Woodburn J. (1968) "Stability and flexibility in Hadza residential 
groupings" In Lee R. B. & I. DeVore (Eds.) 'Man the hunter', 
Aldine-Atherton, Chicago. 
Wrangham R. w. (1974) "Artificial feeding of chimpanzees and baboons ~n 
their natural habitat" Anim.Behav. 22: 83-93. 
Wynne -Edwards V.C. (1962) "Animal dispersion in relation to social 
behaviour" Oliver & Boyd, Edinburgh and London. 
Yerkes R.M. (1943) "Chimpanzees: a 'laboratory colony" Yale University 
Press, New Haven. 
Zahavi A. (1971) "The function of pre-roost gatherings and communal 
roosts" Ibis 113: 106- 109. 
Zar J .H. (1974) "Biostatistical analysis" Prentice-Hall, Englewood 
cliffs. 
Zuckerman s. (1932) "The social life of monkeys and apes" K. Paul, 
Trench, Trubner & Co., London. 
Articles in preparation 
Bauer H. "Behaviour of chimpanzees at reunions" Ph.D. thesis, University 
of Stanford. 
Bertram B. C. R. "Social factors influencing reproduction in wi Id lions". 
Busse C. & D. Riss "Behaviour of adult male chimpanzees". 
Clutton-Brock T .H. & M.J .A. Simpson "Ecology of Gombe National Park". 
Halperin S.D. "Agonistic and other relationships of male chimpanzees" 
Ph.D. thesis, University of Washington, St. Louis. 
Harcourt A. R. "Social relationships and determinants of ranging 
patterns in mountain gorilla". 
Plooij F. X. "Early behavioural development of chimpanzee infants" 
Ph.D. thesis, University of Groningen. 
Pusey A.E. "Behaviour of chimpanzees around adolescence" Ph.D. thesis, 
University of Stanford. 
Thorndahl M. "Family relationships and chimpanzee social structure" 
Ph.D. thesis, University of Cambridge. 
Tutin C.E. G. "Chimpanzee sexual behaviour" Ph.D. thesis, University of 
Edinburgh. 
U)IIVBtSITY 
U RAR 
CAMBRIDG "