University of Southern California Dissertations and Theses

A genetic investigation of chimpanzee distribution and dispersal in the fragmented Budongo-Bugoma Corridor Landscape, Uganda

(USC Thesis Other)

A genetic investigation of chimpanzee distribution and dispersal in the fragmented Budongo-Bugoma Corridor Landscape, Uganda

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

A GENETIC INVESTIGATION OF CHIMPANZEE DISTRIBUTION AND
DISPERSAL IN THE FRAGMENTED BUDONGO – BUGOMA CORRIDOR
LANDSCAPE, UGANDA

by

Maureen Sophia McCarthy

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(INTEGRATIVE AND EVOLUTIONARY BIOLOGY)

May 2016

Copyright 2016 Maureen Sophia McCarthy
ii

Epigraph

“When the music changes, so does the dance.”
-Hausa proverb

© Irene Goede Illustraties
iii

Dedication

In honor of my mother Diane, who supported my endeavors with unconditional love and
provided the best example of grace and courage.

iv

Acknowledgements
Thank you to Craig Stanford for your support and guidance as an advisor. Thank
you to Linda Vigilant for generously welcoming me into your lab and for supporting me
along the way with your guidance and insights. Thank you to Jill McNitt-Gray and Gary
Seaman for being on my dissertation committee and providing valuable feedback. For
thoughtful input over the course of my research, I thank Roberto Delgado, Matt
McLennan, Mimi Arandjelovic, Kevin Langergraber, Deborah Moore, and Jessica
Junker. Thanks to Roger Mundry for assistance with statistical analyses, and to Eric
Howe for assistance with chimpanzee density and abundance estimates. For immensely
valuable assistance during field data collection, I thank Henry Irumba, Nicholas Rugadya,
Tom Sabiiti, Moses Ssemahunge, and Emily Stewart. For assistance with laboratory
work, I thank Anette Abraham, Amy Heilman, and Veronika Staedele. Thanks to Janette
Wallis and the Kasokwa Forest Project for support during data collection in the Kasokwa
Forest. I also thank Zinta Zommers, Richard Wrangham, and Martin Muller for
contributing genotypes used in kinship analyses.
For allowing me to conduct research in Uganda, I thank the Uganda National
Council for Science and Technology, the Uganda Wildlife Authority, and the National
Forestry Authority. I also thank the residents of Hoima and Masindi Districts, Uganda for
their kindness and generosity during the course of this research project. This research
would not have been possible without their cooperation and support.
This research was only possible with the financial support of the University of
Southern California Dornsife College of Letters, Arts, and Sciences and Jane Goodall
v

Research Center, the Max Planck Society, the German Academic Exchange Service
(DAAD), the American Society of Primatologists, the Nacey Maggioncalda Foundation,
and Primate Conservation, Inc.
Thank you to my family for their constant support over the years. My mother
provided constant love and encouraged me to pursue my dreams. My father instilled in
me a reverence for nature at a young age and continues to support the work I love to do.
My brother’s compassionate and patient insistence that we have an ethical imperative to
try has inspired me to keep doing just that.
I also thank the colleagues and friends who helped me during the course of my
research. Thanks to Jess Hartel for guidance, friendship, and inspiration over the course
of my entire graduate student career. Thanks to Nancy Merrick for your friendship and
your dedication to chimpanzees. Thank you to my fellow graduate students at the
University of Southern California and Max Planck Institute for Evolutionary
Anthropology for the much-needed advice and humor.
Finally, a huge thank you to Jack Lester for your selfless support over the years.
Your contributions to this research are too numerous to describe fully, but include taking
the leap with me and volunteering as a field assistant in Uganda years ago, then
encouraging me to pursue graduate research, helping me collect samples, making repairs
to our field vehicle, performing countless PCRs in the lab, reading manuscript drafts, and
listening to practice talks. I will always be indebted to you for this immeasurable support,
but I look forward to doing my best to repay you in the years to come.

vi

Table of Contents
EPIGRAPH ii
DEDICATION iii
ACKNOWLEDGEMENTS iv
LIST OF TABLES ix
LIST OF FIGURES x
LIST OF ABBREVIATIONS xi
ABSTRACT xii

CHAPTER 1: INTRODUCTION 1
1.1. Primates in the Anthropocene: methodological approaches
in changing habitats 1
1.2. Thesis aims and structure 4
1.3. Behavioral flexibility in the Anthropocene 5
1.4. Chimpanzee habitat use and dispersal in human-modified landscapes 9
1.5. Chimpanzees in the Budongo-Bugoma Corridor Landscape, Uganda 10

CHAPTER 2: METHODS 12
2.1. Study area 12
2.2. Study subjects 17
2.3. Genetic census methods 18
2.4. Research permissions 22
2.5. DNA extraction and amplification 22
2.6. Determination of Y-chromosome haplotypes 27
2.7. Discriminating chimpanzee genotypes 29
2.8. Assignment of putative communities and Y-chromosome haplotype
distributions 30
2.9. Abundance estimation 30
2.10. Parentage analysis 32
2.10.1. False positive error rate estimation in parentage assignments 32
2.10.2. Parentage assignments in the corridor chimpanzees 36
2.10.3. Parentage assignments including Busingiro 39
2.10.4. Spatial analysis of repeatedly genotyped chimpanzees 41

vii

CHAPTER 3: A GENETIC INVESTIGATION OF CHIMPANZEE
DISTRIBUTION AND ABUNDANCE IN A FRAGMENTED FOREST
LANDSCAPE 42
3.1. Introduction 42
3.2. Methods 48
3.2.1. Study area and subjects 48
3.2.2. Genetic census methods 48
3.2.3. DNA extraction and amplification 51
3.2.4. Determination of Y-chromosome haplotypes 52
3.2.5. Discriminating chimpanzee genotypes 53
3.2.6. Assignment of putative chimpanzee communities and
Y-chromosome haplotype distributions 53
3.2.7. Abundance estimation 54
3.3. Results 56
3.3.1. Genetic sampling and discriminating individual chimpanzees 56
3.3.2. Putative chimpanzee communities and Y-chromosome
haplotype distributions 58
3.3.3. Abundance estimation 59
3.4. Discussion 61
3.4.1. Abundance estimation 61
3.4.2. Chimpanzee density in the corridor region 66
3.4.3. Putative communities and Y-chromosome haplotypes 66
3.4.4. Conservation implications 68
3.5. Conclusions 70

CHAPTER 4: GENETIC EVIDENCE FOR FEMALE-BIASED
CHIMPANZEE TRANSFER IN A HUMAN-DOMINATED HABITAT 72
4.1. Introduction 72
4.2. Methods 77
4.2.1. Study area and subjects 77
4.2.2. Noninvasive fecal sample collection methods 78
4.2.3. DNA extraction and amplification 80
4.2.4. Discriminating chimpanzee genotypes and putative communities 81
4.2.5. Parentage analysis 82
4.2.5.1. False positive error rate estimation in parentage
assignments 82
4.2.5.2. Parentage assignments in the corridor chimpanzees 86
4.2.5.3. Parentage assignments including Busingiro 89
4.2.5.4. Spatial analysis of repeatedly genotyped chimpanzees 91
4.3. Results 91
4.3.1. Distinct chimpanzee genotypes and community affiliations 91
4.3.2. False positive error rate estimation in parentage assignments 93
4.3.3. Parentage assignments in the corridor chimpanzees 96
4.3.4. Parentage assignments including Busingiro 99
viii

4.3.5. Spatial analysis of repeatedly genotyped chimpanzees 99
4.4. Discussion 102
4.5. Conclusions 109

CHAPTER 5: DISCUSSION 110

BIBLIOGRAPHY 119

APPENDICES 139
Appendix A: Variability and Amplification Success for Microsatellite
Loci in this Research 139
Appendix B: Primer Sequences and Annealing Temperatures for
Autosomal and Y-Chromosome Microsatellite Loci in this Research 140
Appendix C: Alleles and Allele Ranges for Autosomal Microsatellite
Loci in this Research 143
Appendix D: Allele Frequency Data for Corridor Genotypes 146
Appendix E: Y-Chromosome Haplotypes and Their Occurrences 147
Appendix F: Subadult and Adult Composition of Long-Term
Chimpanzee Study Communities 148

ix

List of Tables

Table 2.1. Allelic dropout rates by locus. 27

Table 3.1. Community-specific capwire estimates. 59

Table 4.1. Allele frequency data comparison for corridor and Kibale
genotypes. 93

Table 4.2. False positive error rates for parent-offspring assignments
using KinGroup. 94

Table 4.3. False positive error rates for parent-offspring assignments
using CERVUS. 94

Table 4.4. False positive error rates for mother-father-offspring trio
assignments using CERVUS. 95

x

List of Figures

Figure 2.1. Map of the study area in Uganda. 13

Figure 2.2. The farm-forest mosaic landscape, Hoima District, Uganda. 13

Figure 2.3. Habitat conversion from forest to farmland in the farm-forest
mosaic landscape, Hoima District, Uganda. 16

Figure 2.4. Forest has been converted for tobacco farming, Hoima District,
Uganda. 16

Figure 2.5. Chimpanzees cross a road from a forest to a sugar cane field
as local residents watch, Masindi District, Uganda. 18

Figure 2.6. Map of search effort over the study area. 20

Figure 2.7. Example peaks produced from microsatellite data. 24

Figure 3.1. Genotyped sample collection locations across the study area. 57

Figure 3.2. Putative chimpanzee communities (a) and associated
Y-chromosome haplotypes (b). 60

Figure 4.1. MCPs for genotyped samples found in association. 92

Figure 4.2. Parent-offspring dyads detected among the corridor
chimpanzees using CERVUS. 98

Figure 4.3. Parent-offspring dyads detected among the corridor and
Busingiro chimpanzees using CERVUS. 100

Figure 4.4. Locations of genotyped samples collected for female C73. 101

xi

List of Abbreviations

BSA Bovine serum albumen

capwire Capture with replacement

CI Confidence interval

CV Coefficient of variation

ECM Even capture model

LOD Log-likelihood ratio

MCP Minimum convex polygon

PCR Polymerase chain reaction

SD Standard deviation

SE Standard error

SECR Spatially explicit capture-recapture

TIRM Two innate rates model

xii

Abstract

Habitat loss and fragmentation pose growing challenges to wildlife globally.
Great apes, our closest living relatives, are threatened with extinction, with habitat loss
and fragmentation acting as key drivers in their decline. Growing proportions of great ape
populations live outside protected areas, in habitats that are often fragmented. It is
essential to understand how these species respond to habitat changes to better predict
future impacts and devise appropriate conservation strategies. Despite the survival risks
posed by fragmented habitats, such environments can also offer resources and act as
corridors linking continuous habitat. Therefore, a realistic approach to research and
conservation in the Anthropocene must take into account the value and potential of such
habitats to great apes and other wildlife.
The fragmented forest habitat between the Budongo and Bugoma Forests in
Uganda has undergone substantial changes in recent decades, with the conversion of this
riparian forest-grassland mosaic to accommodate the needs of growing human
population. Though previous research confirmed chimpanzee presence in this region, the
size and distribution of this population has remained little understood. Further, data have
been lacking to examine whether this habitat acts as a corridor for chimpanzees between
the Budongo and Bugoma Forests, each of which contains hundreds of chimpanzees.
This research employed a genetic approach to investigate chimpanzee distribution
and dispersal in this fragmented forest landscape. I noninvasively collected chimpanzee
fecal samples to study this population’s size and distribution using mark-recapture
methods. I also investigated the distribution of co-sampled genotypes to establish the
xiii

locations of putative chimpanzee communities in the study region. I further examined the
distribution of distinct Y-chromosome haplotypes among male chimpanzees in the study
area, since these can be used as markers of community affiliation in territorial, male-
philopatric chimpanzees. I found a minimum chimpanzee population of 182 individuals
based on the number of unique genotypes, and estimated the total population size at
approximately 260 – 320 individuals, depending on the estimator used.
A widely distributed chimpanzee population in this corridor habitat is not
necessarily sufficient to ensure gene flow, however. If habitat fragmentation isolates
remaining communities, female dispersal—a typical inbreeding avoidance mechanism in
chimpanzees—may be inhibited. I used genetic data on parent-offspring relationships and
community affiliations to examine dispersal patterns in this landscape. I found a pattern
of widespread dispersal characterized by mother-daughter dyads affiliated with distinct
communities. Father-son dyads, in contrast, were almost always affiliated with the same
community, confirming a pattern of male philopatry as would be expected for
chimpanzees in continuous habitats. I additionally found evidence of a female dispersal
event during the study period.
Together, these findings suggest chimpanzees maintain species-typical social
structure and dispersal patterns in a highly human-modified landscape, at least in the
short term. This underscores the conservation potential of corridor habitats and the
potential benefits of corridor restoration and enhancement. Conservation measures are
urgently needed, however, given the rapid rate of habitat change and the likelihood of
chimpanzee extinction despite marked adaptability to habitat alterations.
1

Chapter 1
Introduction

1.1. Primates in the Anthropocene: methodological approaches in changing habitats
Many organisms are undergoing rapid and drastic habitat changes in the
Anthropocene, a term coined to describe our modern epoch of human-impacted global
change (Crutzen and Stoermer, 2000; Corlett, 2015). Approximately 40% of known
species are currently threatened with extinction (IUCN, 2014), and habitat loss and
fragmentation are key drivers of extinction risk in many taxa (WWF, 2014). Most forests
worldwide are now highly fragmented, with the majority lying within 1 km of a forest
edge (Haddad et al., 2015). Ecological and behavioral research therefore must
2

increasingly account for the role of anthropogenic factors in the organisms under
consideration. By incorporating measures of anthropogenic influence, researchers can
better elucidate behavioral patterns and predict the likelihood of future survival with
continued habitat changes. Among primates, nearly half are threatened with extinction
(IUCN, 2014), yet relatively little remains known regarding how threatened primates
respond to anthropogenic habitat modification.
Noninvasive approaches are becoming increasingly common for studies of
primates in changing environments. Traditional methods for studying primate behavior
have relied on the habituation of animals to researcher observation. Though highly
valuable for collecting detailed behavioral data, habituation is a lengthy process which is
necessarily restricted to small numbers of individuals, and may not be ethically
appropriate or logistically feasible for many populations (McLennan and Hill, 2010;
Gruen et al., 2013). Given the need to quickly acquire data on remaining primate
populations for both basic research and conservation, effective noninvasive approaches
that do not require habituation are necessary to supplement long-term behavioral studies.
To date, noninvasive approaches for studying unhabituated primates have proven
promising as well as challenging. For baseline research relating to species distribution
and abundance, transect survey methods may be used (Kühl et al., 2008; Ross and Reeve,
2011). For studies of great apes, nest count surveys can be used to estimate the
distribution and abundance of unhabituated populations. These survey methods can
provide valuable data, particularly for determining species presence, though they may be
inaccurate and lack the precision necessary to determine trends in population size
3

(Plumptre and Reynolds, 1996; Devos et al., 2008; Boyko and Marshall, 2010). Such
studies are also arduous to carry out, as commonly used nest count methods rely on data
regarding nest decay rates and nest building and re-use rates, which can be highly
variable and are often unknown locally (Plumptre and Reynolds, 1996; Kouakou et al.,
2009; Boyko and Marshall, 2010; Buckland et al., 2010). Camera trapping and passive
acoustic monitoring have also been utilized for primate censusing as well as behavioral
data collection (Head et al., 2013; Heinicke et al., 2015; Kalan et al., 2015). Though still
in their infancy, such methods can incorporate automated detection to enable the efficient
processing and analysis of large amounts of data (Kühl and Burghardt, 2013; Heinicke et
al., 2015). Radio tracking methods, though sometimes logistically and ethically
challenging, offer promising opportunities for the study and monitoring of some species
(Honess and MacDonald, 2011). The use of drones for research and conservation are also
increasingly popular and offer the potential for noninvasive species detection (Koh and
Wich, 2012; van Andel et al., 2015). Thus, depending on the research questions of
interest, these approaches offer promising potential for future studies, though many are
still under development and difficult to implement at present.
The need for efficient and feasible noninvasive methods has led to the widespread
application of genetic approaches to the study of primate populations. Noninvasively
collected fecal samples can be used to extract DNA, and genetic data can be acquired
from markers such as microsatellite loci, which are selectively neutral and polymorphic
(Selkoe and Toonen, 2006). Microsatellite genotyping can be used to survey and monitor
populations, for example using capture-mark-recapture approaches. The minimum
4

number of individuals using the surveyed area is determined by the number of unique
profiles or genotypes, and resampling frequency can be used to estimate the number of
animals that went undetected (Lukacs and Burnham, 2005; Petit and Valiere, 2006).
Microsatellite genotype data can also be used to examine aspects of kinship, as well as to
study questions relating to population structure, genetic diversity, and reproductive
fitness (Selkoe and Toonen, 2006; Goossens et al., 2009). Given the feasibility of
noninvasive data collection and the statistical power of microsatellite data, this approach
has become highly standardized and is widely used in studies of behavioral and
conservation genetics.
This noninvasive genetic approach therefore carries with it several advantages.
First, the methods used are highly standardized and efficient. Second, the approach
allows researchers to answer a variety of questions relating to the presence and
distribution as well as the behavior and relatedness of the animals under study. Third,
these methods are entirely noninvasive and rely only on fecal samples, allowing their
widespread use in a variety of contexts, including environments heavily impacted by
humans, for which habituation for behavioral data collection may be either infeasible or
inappropriate.
1.2. Thesis aims and structure
This thesis focuses on the use of noninvasive genetic sampling to study the
distribution and dispersal patterns of chimpanzees (Pan troglodytes schweinfurthii) in a
human-dominated landscape in western Uganda as means of 1) elucidating chimpanzee
population parameters in a fragmented corridor habitat and 2) better understanding
5

chimpanzee behavioral flexibility in the Anthropocene. In Chapter 2, I describe the
methods used for field data collection, laboratory work, and analysis for this research. My
first aim was to characterize the size and distribution of this chimpanzee population,
which was previously poorly understood, to determine the degree to which chimpanzees
persist in this human-modified landscape. In Chapter 3, I present the results of this
genetic census. Even the presence of robust chimpanzee population does not, however,
ensure gene flow in such a habitat. The maintenance of species-typical dispersal patterns,
as well as the social structure underlying them, would further indicate flexibility to
changing environmental conditions and provide insights into the potential viability of
chimpanzee populations in western Uganda. My second aim, therefore, was to
characterize patterns of dispersal by chimpanzees in this habitat. In Chapter 4, I present
my findings regarding female dispersal among these chimpanzee communities. In
Chapter 5, I synthesize these findings by describing my conclusions and their
implications for great ape conservation in western Uganda as well as other human-
modified landscapes.
1.3. Behavioral flexibility in the Anthropocene
Behavioral flexibility is a form of phenotypic plasticity that can influence an
organism’s fitness by allowing shifts in behavioral responses across varied environmental
conditions (van Schaik, 2013). Flexible behavioral responses are evident in a wide array
of taxa and can be particularly adaptive for organisms with long life histories and for
those that range broadly across a number of habitat types. For such organisms, a rigidly
6

canalized relationship between genotype and environment can rarely be achieved, leading
toward a selection pressure favoring flexible responses (Stearns, 1989; van Schaik, 2013).
The Primate Order is generally characterized by a relatively high degree of
cognitive complexity and behavioral flexibility (Roth and Dicke, 2005). Primates, in
particular great apes, have long life histories as compared with other mammals (Charnov
and Berrigan, 1993), thereby making behavioral flexibility advantageous across the
lifespan. They also have large brains relative to body size (Roth and Dicke, 2005), which
are associated with a higher degree of behavioral flexibility (van Schaik, 2013). Great
apes often experience large shifts in seasonal resource availability (Knott, 2005), range
over broad areas and across numerous habitat types (Stumpf, 2011), and show evidence
of social learning (Whiten et al., 1999; Whiten, 2000), all of which suggest behavioral
flexibility is adaptive and aids survival and reproductive success.
The study of behavioral flexibility in great apes is important in several regards.
First, it provides insight into the adaptations shared across the family Hominidae—which
includes great apes, modern humans, and their ancestors—for flexible responses to
variations in environmental and social conditions. Such flexibility may have played an
important role in the evolutionary history of members of this clade. Second, it can also
shed light on our evolutionary distinctiveness, since a particularly high level of
behavioral flexibility may have been critical to hominin evolution (Potts, 1998).
Additionally, by better understanding the behavioral flexibility of great apes, we can
better predict their responses to changing environments in the Anthropocene and
therefore devise appropriate conservation and conflict mitigation strategies (Hockings et
7

al., 2015). All extant great apes—including chimpanzees (Pan troglodytes), bonobos
(Pan paniscus), western (Gorilla gorilla) and eastern lowland gorillas (Gorilla beringei),
Bornean orangutans (Pongo pygmaeus), and Sumatran orangutans (Pongo abelii)—are
threatened with extinction (Ancrenaz et al., 2008; Fruth et al., 2008; Oates et al., 2008;
Robbins and Williamson, 2008; Singleton et al., 2008; Walsh et al., 2008). In recent
years, there has been a steep decline in remaining suitable habitat for African great apes
(Junker et al., 2012). Deforestation in the African tropics, which serve as habitat for
African great apes, accounts for nearly a quarter of global forest loss between 1990 and
2009 (Houghton, 2012). Anthropogenic habitat loss and fragmentation, along with
hunting and disease, are key threats to great ape survival (Walsh et al., 2003; Junker et
al., 2012). Up to 81% of chimpanzees in West Africa live outside protected areas, often
in fragmented and degraded forests. Similarly, 78% of Bornean orangutans are estimated
to live outside protected areas, with large-scale commercial agriculture such as oil palm
production posing a major threat to orangutan survival (Meijaard et al., 2010; Wich et al.,
2012).
Great apes have responded to changes in environmental conditions by shifting
their behavior in numerous domains. For example, chimpanzees in human-modified
habitats alter their diets by feeding on a variety of cultivated species at a number of study
sites (Hockings and McLennan, 2012). Mountain gorillas at Bwindi Impenetrable
National Park, Uganda feed on the bark of eucalyptus trees in plantations of this
introduced species (Rothman et al., 2006), while Bornean orangutans feed on mature oil
palm fruits in plantations (Ancrenaz et al., 2014). Great apes also show evidence of
8

altering their activity budgets in human-modified habitats. At Bossou, Guinea,
chimpanzees rest less on crop feeding days (Hockings et al., 2012), while at Sebitoli in
Kibale National Park, Uganda they forage on crops at night, likely to avoid detection by
humans (Krief et al., 2014). Chimpanzees may also alter their social behavior,
demonstrating greater social cohesion in response to risky situations like crop feeding
(Hockings et al., 2012). They may modify their vocal behavior, making fewer
vocalizations in the presence of human impact (Hicks et al., 2013). Finally, they may also
modify their nesting patterns, such as altering nest height in areas of human impact,
perhaps as a means to avoid encounters with humans (Hicks, 2010; Last and Muh, 2013;
Tagg et al., 2013). Great apes also use non-native species for nesting in human-modified
habitats. For example, Bornean orangutans nest in oil palm plantations (Meijaard et al.,
2010; Ancrenaz et al., 2014). Studies such as these collectively demonstrate a trend
toward flexible behavioral responses by great apes in human-modified landscapes.
Behavioral flexibility in the immediate short term does not necessarily predict
long-term survival under anthropogenic impacts, however. The degree to which
individuals to exploit usable habitat so as to maintain adequate population density and
reproductive rates, intact social structure, and species-typical dispersal patterns may help
predict the long-term persistence of great apes despite anthropogenic threats.
Nonetheless, little remains known regarding how great ape habitat use, social structure,
and dispersal patterns are impacted in human-modified landscapes.

9

1.4. Chimpanzee habitat use and dispersal in human-modified landscapes
Chimpanzees, along with bonobos, are our closest living relatives (Mikkelsen et
al., 2005; Prüfer et al., 2012). They comprise four subspecies—eastern (Pan troglodytes
schweinfurthii), central (P. t. troglodytes), Nigeria-Cameroon (P. t. ellioti), and western
(P. t. verus)—all of which are endangered (IUCN, 2014). Among the great apes, they are
also the most numerous, coming into frequent contact with humans, and have been
studied across a variety of habitat types including human-impacted landscapes (Caldecott
and Miles, 2005). The flexibility of their behavioral responses to habitat fragmentation
and anthropogenic disturbance is therefore highly relevant from both evolutionary and
conservation perspectives.
Though the effects of anthropogenic disturbance and habitat fragmentation on
chimpanzees are not well understood, chimpanzees have been well-studied over decades
of research at long-term field sites with relatively lower levels of anthropogenic impact.
In less human-impacted environments, chimpanzees have a fission-fusion social structure
characterized by male philopatry and female dispersal, with a multi-male, multi-female
polygynous mating system (Goodall, 1986). Though chimpanzee habitat use is largely
associated with resource availability (Wrangham, 1977; Doran, 1997; Mitani et al.,
2002), other environmental and social variables can also influence its patterns. For
example, individual sex (Wrangham, 1979) and preferred core areas (Williams et al.,
2002) can influence habitat use, while the presence of estrus females can influence party
composition and, accordingly, patterns of habitat use (Anderson et al., 2002). In
fragmented and human-modified landscapes, these decisions may additionally be
10

influenced by habitat variables such as the risks associated with moving through the
habitat (Hockings et al., 2006; Cibot et al., 2015).
Dispersal patterns in chimpanzees may also be influenced by a number of factors.
Although dispersal is a key mechanism for inbreeding avoidance, the decision about
whether to disperse is not fixed and is influenced by numerous factors in chimpanzees as
well as in a variety of other taxa (Stumpf et al., 2009; Chaine and Clobert, 2012). In less
disturbed habitats, the likelihood that a chimpanzee female will disperse has been
associated with variables such as social rank, resource availability, and affiliative
relationships (Williams et al., 2002; Stumpf et al., 2009; Wroblewski et al., 2015). In
fragmented landscapes, community isolation may limit opportunities for dispersal and
reduce its likelihood (Sugiyama, 1999). Little is currently known, however, regarding the
effects of habitat fragmentation and human disturbance on chimpanzee dispersal patterns.
1.5. Chimpanzees in the Budongo-Bugoma Corridor Landscape, Uganda
Eastern chimpanzees in Uganda often live in human-impacted habitats, thereby
offering apt opportunities to study great ape responses to anthropogenic disturbances. The
total eastern chimpanzee population numbers between 76,400 and 119,600 individuals
and spans a geographic range of seven countries including Uganda (Plumptre et al.,
2010). Uganda’s estimated chimpanzee population size is 5,000 individuals, and this
population is confined to the western part of the country along the Northern Albertine
Rift (Plumptre et al., 2003). The government-owned Budongo and Bugoma Forest
Reserves are each inhabited by over 600 chimpanzees, together composing
approximately one-quarter of this total chimpanzee population in Uganda (Plumptre et
11

al., 2003). The corridor between these forest blocks is a human-dominated landscape
comprising mosaic riparian forest with villages, agricultural lands, and natural grasslands
(McLennan, 2008). Most forests in this habitat are privately owned, but a few small
government-owned forest reserves are present. The small forests in this region have been
targeted for potential corridor enhancement given the vital role they may play for gene
flow among species throughout this region (Nangendo et al., 2010). Chapter 2 provides
further details regarding this habitat.
Although the conservation potential of this region is high, few studies have
examined the population size and distribution of its chimpanzees. A nationwide
chimpanzee census used a nest count survey of forest fragments near the Bugoma Forest
to extrapolate an estimate of ~70 chimpanzees in the corridor region (Plumptre et al.,
2003). Later, McLennan (2008) found evidence of chimpanzees throughout the corridor
habitat and estimated a total regional population of up to 260 individuals, an
extrapolation derived from the estimated density of one chimpanzee community (Bulindi)
in the corridor area (McLennan, 2008). The variability in these estimates underscores the
need to survey such populations using thorough and accurate approaches. Further, the
degree of gene flow among chimpanzees in such habitats can provide critical information
to inform evaluations of population viability and conservation potential. As described in
Section 1.2, this thesis examined these critical issues as a means of better understanding
this chimpanzee population and others like it in increasingly human-dominated
landscapes across Equatorial Africa. In the following sections, I describe this research as
well as conclusions and implications for great ape conservation in the Anthropocene.
12

Chapter 2
Methods

2.1. Study area
I collected data in Hoima and Masindi Districts, Uganda in the corridor region
between the Budongo and Bugoma Forests (1°37' – 1°68'N and 31°1' - 31°6'E; Figure
2.1) from October through December 2011 and October 2012 through September 2013.
This area occurs in the northern section of the Albertine Rift, a region that spans
numerous East African countries and harbors a high level of biodiversity and endemism
(Plumptre et al., 2007). The Budongo and Bugoma Forests are classified as medium-
altitude, moist semi-deciduous forests (Eggeling, 1947; Langdale-Brown et al., 1964).
The Budongo Forest covers 428 km
2
, while the Bugoma Forest measures 411 km
2
(Plumptre et al., 2003; Reynolds, 2005). These forests are the largest in Uganda, and both
are protected reserves under the management of the Uganda National Forestry Authority
(Twongyirwe et al., 2015). The region between these forests, which broadly measures
approximately 40 km long by 30 km wide, is a mosaic habitat comprising agricultural
land, villages, riparian forest fragments, and grasslands (Figure 2.2). These riparian
13

Figure 2.1. Map of the study area in Uganda. The inset map displays the
landscape’s location within Uganda. Green indicates forest cover during
the study period, provided by Hansen et al. (2013).

Figure 2.2. The farm-forest mosaic landscape, Hoima District, Uganda. Photo by
Jack Lester.
14

forests occur mainly along the Waki, Hoima, and Rwamatonga Rivers and their
tributaries (McLennan, 2008). Pollen and climatic data indicate the Budongo Forest has
been a standalone forest block for thousands of years, so the region to its south likely
existed as a natural mosaic habitat throughout that time (Paterson, 1991). In recent
decades, however, human populations have grown substantially, leading to the extensive
conversion of unprotected riparian forests for commercial and subsistence agriculture
(Mwavu and Witkowski, 2008; FAO, 2010). In addition, 90% of Ugandans rely on wood
fuel as a main energy source, and the vast majority of households gather firewood to
sustain this fuel requirement (MWLE, 2002). These land use pressures have resulted in
the loss of an estimated 134 km
2
of forest cover between 1985 and 2014 (Twongyirwe et
al., 2015). Deforestation in Uganda primarily occurs outside protected areas (MWLE,
2002), with 99% of forest loss near the Budongo and Bugoma Forests in the past three
decades taking place outside the reserves in unprotected forests (Twongyirwe et al.,
2015). Such habitat changes are widespread and ongoing in this largely unprotected
corridor landscape (Figures 2.3 and 2.4), though evidence suggests the rate of
deforestation is slowing due to the diminishing availability of forest resources
(Twongyirwe et al., 2015).
Common tree species in these riparian forests include Phoenix reclinata,
Macaranga schweinfurthii, Pseudospondias microcarpa, and members of the Moraceae
family (McLennan and Plumptre, 2012). Annual rainfall exhibits a bimodal pattern, with
two rainy seasons from March to May and September to November and a mean annual
rainfall of 1600 mm (Eggeling, 1947; Reynolds, 2005). There is one major dry season
15

from December through February, with an average rainfall of < 50 mm (Reynolds, 2005).
The average temperature is 21 °C, ranging between 19 °C and 32 °C (Tweheyo and Lye,
2003; Reynolds, 2005).
Chimpanzees are broadly distributed throughout this corridor region (McLennan,
2008). Additionally, six other diurnal primate species range throughout limited or broad
areas of the corridor, including Papio anubis, Colobus guereza occidentalis,
Cercopithecus mitis stuhlmanni, Chlorocebus tantalus budetti, Lophocebus ugandae, and
Cercopithecus ascanius schmidti (Plumptre et al., 2011). Periodicticus potto has also
been observed (pers. obs.). Other species in the corridor region include red duikers
(Cephalophus callipygus weynsii), bushbucks (Tragelaphus scriptus), crested porcupines
(Hystrix cristata), as well as a variety of bird species (Plumptre et al., 2011).
Hoima and Masindi Districts have human population densities of 156.6 and 113.3
inhabitants/km
2
, respectively (UBOS, 2014). Their total populations number 573,903 and
292,951 (UBOS, 2014). Local residents primarily belong to the Banyoro ethnic group
(Mubiru and Kristjanson, 2012). Small-scale subsistence and commercial agriculture
comprise the primary incomes of most households (CIA, 2013). Uganda has a total
population of approximately 35 million inhabitants with an annual growth rate of 3%
(UBOS, 2014).

16

Figure 2.3. Habitat conversion from forest to farmland in the farm-forest mosaic
landscape, Hoima District, Uganda. Photo by Maureen McCarthy.

Figure 2.4. Forest has been converted for tobacco farming, Hoima District,
Uganda. Photo by Jack Lester.
17

2.2. Study subjects
Though data have been lacking regarding chimpanzee population parameters in
this region, some research has shed light on their behavioral ecology, as well as threats to
their survival. Numerous studies have focused on the chimpanzees in Bulindi, a farm-
forest mosaic in this region. The Bulindi chimpanzees have become semi-habituated over
intermittent behavioral observations since 2007 (McLennan, 2010). Their diet is highly
frugivorous and largely composed of forest fruit but supplemented with cultivars,
particularly during periods of low forest fruit availability (McLennan, 2013).
Chimpanzees in the nearby Kasokwa Forest Reserve also feed on a combination of forest
foods and cultivars (Reynolds et al., 2003). Though chimpanzees are typically not hunted
in this region or elsewhere in Uganda, they often encounter other threats. Close co-
residence of chimpanzees and humans, particularly with the increasing conversion of
natural forests to farmland in recent decades (Mwavu and Witkowski, 2008), has led to
frequent interactions (Figure 2.5), with chimpanzees incurring serious injuries and death
resulting from man-traps, snares, and fires (Reynolds, 2005; McLennan, 2008; McLennan
et al., 2012). Local residents have also been seriously injured and killed as a result of
aggressive interactions with chimpanzees (Reynolds, 2005). Attitudes of local residents
have correspondingly deteriorated in recent years, given that chimpanzees are sometimes
perceived as pests due to cultivar feeding and as risks to human safety (McLennan and
Hill, 2012).

18

Figure 2.5. Chimpanzees cross a road from a forest to a sugar cane field as local
residents watch, Masindi District, Uganda. Photo by Maureen McCarthy.

2.3. Genetic census methods
I collected chimpanzee fecal samples noninvasively throughout the study area
from October through December 2011 and October 2012 through September 2013. I
collected samples opportunistically throughout the region, with a focus on searching
riparian forest fragments for evidence of chimpanzee presence. Information on
chimpanzee presence was also provided by McLennan (2008) and by informal discussion
with local inhabitants. It was not practical to employ strictly systematic survey methods
in this human-dominated habitat comprising mainly privately owned farms and villages.
19

In addition, line transect cutting through forest fragments would have been ethically
inappropriate, as it would have made these small forest fragments more accessible to
further habitat destruction, including logging and illegal snare hunting. Instead, I centered
search effort in forest fragments around village boundaries, which typically encompass
settlements, farmland, and privately owned forests. In accordance with local customs,
prior to searching a forest fragment I first gained permission from the chairperson of the
village in which the forest fragment was located, and from individuals who identified
themselves as landowners of the forest fragment. I used satellite imagery to identify the
forest fragments located within the boundaries of a given village, and visited accessible
and permitted forest fragments within the boundaries of that village. I divided the study
area into a grid of 1 sq. km cells and recorded when any part of each cell was searched
(Figure 2.6).
Chimpanzee fecal samples were typically easy to identify because of 1) their
locations under chimpanzee nests and along trails, 2) their characteristic size, shape, and
odor, and 3) the absence of other sympatric large-bodied nonhuman primates. Olive
baboons (Papio anubis) produce dungs that can superficially resemble those of
chimpanzees (pers. obs.), but have largely been eradicated from the study area. In cases
where I suspected that a fecal sample may have been produced by a sympatric baboon, I
collected a small portion of the dung for genetic analysis, then collected the remainder
separately and washed it later that day in a 1 mm mesh sieve. Fecal samples of baboons
and chimpanzees were easily distinguished from one another by their differing odor and
dietary components when washed through a sieve (Okecha and Newton-Fisher, 2006).
20

Figure 2.6. Map of search effort over the study area. One-km
2
grid cells are overlaid over
the corridor region between the Budongo and Bugoma Forests. Gray shading indicates
relative search effort in each cell, with the number of search occasions (days) binned.
Search effort was not available in the Bulindi area, where samples were collected during
concurrent long-term research. Green indicates forest cover during the study period, as
provided by Hansen et al. (2013).

Any sample suspected to originate from a baboon rather than a chimpanzee was thus
discarded following washing (n = 5).
I determined target sample sizes by roughly estimating the spatial area of a
putative chimpanzee home range, based on direct and indirect evidence of chimpanzee
presence, then multiplying by the known density of chimpanzees in a single studied
community within the corridor region [0.66 chimpanzees per km
2
; (McLennan, 2008)].
This estimate was then tripled to determine a target number of samples to be collected in
21

that area, since at least three times the number of samples as expected individuals has
been recommended to achieve a narrow confidence interval for population size estimates
using genetic mark-recapture approaches (Miller et al., 2005; Petit and Valiere, 2006;
Arandjelovic et al., 2010). Because additional information on chimpanzee presence was
gained over the course of the study period, target sample sizes were adjusted as
necessary. To help achieve this sampling goal and to ensure adequate resampling across
fission-fusion chimpanzee communities, I searched forests seasonally in all cases except
where research permission was granted only for a limited time period.
I collected samples under nests and opportunistically along chimpanzee trails and
at feeding sites. For each sample collected, I recorded a GPS waypoint with a Garmin
GPSMap® 60CSx. I recorded samples with unique identification numbers corresponding
to GPS waypoints, and with party association data when applicable. I recorded samples
as belonging to a party when two or more same-age samples were collected within 30 m
of each other. I determined distances using GPS data and, when necessary, a Bresser®
laser Range Finder 800 to ensure accuracy. I avoided collecting two samples under the
same nest or in close proximity on trails, due to the likelihood of collecting redundant
samples from the same individual and the possibility that closely deposited samples may
have cross-contaminated each other. I collected and stored samples according to the two-
step ethanol-silica method described in Nsubuga et al. (2004). In brief, I collected
approximately 5 g of chimpanzee dung from the outer surface of the bolus using a sterile
tongue depressor. I deposited the sample into a 50-mL sterile collection tube containing
30-mL of 96% ethanol. I mixed the tube by inversion and stored it for 24 – 36 hours
22

before transferring it into a second sterile 50-mL tube containing approximately 20 g of
silica gel beads (Sigma S7625), where I then stored the sample until DNA extraction.
2.4. Research permissions
I carried out data collection with the permission of the Uganda National Council
for Science and Technology, the Uganda Wildlife Authority, and the National Forestry
Authority of Uganda. Additional permissions were granted by local landowners where
applicable, as described above in Section 2.3. Because fecal sample and nest data
collection were entirely noninvasive and required no contact with the chimpanzees,
ethical consent was not necessary for this project.
2.5. DNA extraction and amplification
I stored samples in the field for up to 6 months prior to arrival at Max Planck
Institute for Evolutionary Anthropology, where they were then stored at 4 °C prior to
extraction. I extracted DNA using either the GeneMATRIX Stool DNA Purification Kit
(Roboklon) or the QIAmp Stool kit (QIAGEN) with minor procedural adjustments
(Nsubuga et al., 2004). I prepared fecal samples for extraction under a dedicated
extraction hood that was UV irradiated before and after use to avoid contamination. I
performed all further steps of the extraction in a dedicated extraction room. I included
one to two negative controls per extraction, which followed exactly the protocol for other
extraction tubes except for the addition of the fecal sample. Negative control extracts
were included in an initial polymerase chain reaction (PCR) test (as described below) to
ensure that no contamination occurred during the extraction process.
23

I used autosomal microsatellite loci to determine individual chimpanzee
genotypes. To do so, I first evaluated each DNA extract by simultaneously amplifying
three autosomal microsatellite loci (D1s1622, D12s66, and D18s536), along with an X-Y
homologous segment of the amelogenin gene, used for sex determination (Bradley et al.,
2001), in a one-step multiplex PCR. Appendix A provides a full list of these and other
loci used in this research, including details regarding their performance and variability.
For each reaction, I used 0.5 µL 2x Type-It Multiplex PCR Master Mix (QIAGEN) and 2
µL template DNA with the following optimized concentrations of each forward labeled
and nested reverse primer (Römpler et al., 2006; Arandjelovic et al., 2009): 0.03 mM
amelogenin, 0.15 mM D18s536, 0.32 mM D12s66, and 0.30 mM D1s1622 in a total 10-
µL reaction volume. Each PCR consisted of DNA extracts, as well one to two negative
controls from each extraction, in four independent reactions. In addition, to monitor for
consistency and possible contamination as is prudent when working with low
concentration DNA derived from noninvasive samples, each PCR included one positive
control from a chimpanzee extract with a known genotype and seven negative controls,
which consisted of purified H
2
0 instead of DNA. I centrifuged the plate at a minimum of
1,400 rpm, and thermocycled it using a PTC-225 Thermal Cycler (MJ Research) as
follows: denaturation for 5 min at 95 °C; 45 cycles of 30 sec at 95 °C, 90 sec at 58 °C,
and 30 sec at 72 °C; and a final extension for 30 min at 72 °C, followed by incubation at
10 °C until removal from the thermocycler. Appendix B provides further details
regarding the primer sequences and annealing temperatures used for each microsatellite
locus in this research. Following thermocycling, I centrifuged the PCR product again,
24

then diluted it 1:30 with purified H
2
0 and added 27.4 µL of a 1:135 dilution of ROX
labeled GENESCAN 400HD (Applied Biosystems) and H
2
0 to size alleles relative to an
internal standard. I then denatured the dilution at 90 °C for four to six min, and following
that, I incubated it on ice for five min. PCR products from all four loci were then
electrophoresed using an ABI PRISM 3100 Genetic Analyser. I used GeneMapper
version 3.7 (Applied Biosystems) to analyze the data. Figure 2.7 displays sample peaks
produced and analyzed using GeneMapper. Appendix C provides the alleles and allele
ranges observed for each microsatellite locus used in this research.

Figure 2.7. Example peaks produced from microsatellite data. Peaks were produced
using the ABI PRISM 3100 Genetic Analyser for capillary electrophoresis, in
conjunction with GeneMapper version 3.7 for peak analysis. Red channel peaks indicate
size standard data. Data from two loci are visible in the blue channel, while data from a
single locus are visible in the green and black channels, respectively.

DNA extracts that reliably amplified at a minimum of 3 of the 4 loci in at least 3
independent amplifications were then genotyped at an additional 11 autosomal
microsatellite loci (D1s1656, D2s1326, D3s2459, D3s3038, D4s1627, D5s1457,
25

D5s1470, D7s817, D7s2204, D10s676, D11s2002). Extracts that failed to meet these
criteria were not amplified further. I amplified the additional 11 loci in a two-step
multiplex-singleplex PCR procedure (Arandjelovic et al., 2009). In the first step, I
included 19 forward unlabeled and unnested reverse primers in a single PCR reaction. I
amplified DNA in 20 µL / well reactions using the following solution: 8.56 µL purified
H
2
0, 1.0 µL dNTPs, 1.4 µL MgCl
2
, 0.8 µL bovine serum albumen (BSA), 2 µL 10x PCR
Buffer, 0.1 µL SuperTaq premixed with TaqStart antibody, and 1.14 µL of each forward
and reverse primer, for a total 15-µL solution. I added 5 µL DNA to each well for a total
20-µL volume. Each 96-well plate included three independent amplifications for each
DNA extract, along with one positive control from a chimpanzee extract known to
amplify and eight negative controls that included purified H
2
0 instead of DNA extract. I
centrifuged the 96-well plate and thermocycled it as follows: denaturation for 9 min at 94
°C; 29 cycles of 20 sec at 94 °C, 30 sec at 57 °C, and 30 sec at 72 °C; and a final
extension for 30 min at 72 °C, followed by incubation at 10 °C until removal from the
thermocycler.
In the second step, I performed a singleplex PCR for each locus using the PCR
product from the initial multiplex PCR (Arandjelovic et al., 2009). For this step, each
well contained a total volume of 10 µL, consisting of 4.71 µL purified H
2
0, 0.5 µL
dNTPs, 0.35 µL MgCl
2
, 10x PCR Buffer, 0.4 µL BSA, 0.04 µL SuperTaq premixed with
TaqStart antibody, and 0.25 mM of a single HEX, FAM, or NED labeled forward and
nested reverse primer. I added 2.5 µL of 1:99 diluted multiplex PCR product to each well
for a total 10-µL reaction volume. I centrifuged each singleplex 96-well plate and
26

thermocycled it as follows: denaturation for 9 min at 94 °C; 29 cycles of 20 sec at 94 °C,
30 sec at 55 – 62 °C (optimized by primer), and 30 sec at 72 °C; and a final extension for
30 min at 72 °C, followed by incubation at 10 °C until removal from the thermocycler.
Appendix B provides the primer sequences and annealing temperatures for each
microsatellite locus used in this study. I then centrifuged singleplex PCR products and
combined 3 – 4 products with optimized volumes of 1 – 3 µL per well with 8 – 12 µL
H
2
0 and 27.4 µL of a 1:135 dilution of ROX labeled GENESCAN 400HD (Applied
Biosystems) and H
2
0. I then denatured and electrophoresed PCR products as described
above.
At each locus, I confirmed heterozygous genotypes by observation in at least two
independent reactions (Morin et al., 2001; Arandjelovic et al., 2009). I confirmed
homozygous genotypes when observed in a minimum of three independent reactions.
Individual loci that failed to meet these criteria were instead coded with asterisks and
were excluded from analyses. To further ensure that apparent homozygotes were not the
result of allelic dropout, I calculated allelic dropout rates by locus after recording all
alleles and confirmed that a maximum of two replicates was required at any locus to
confirm homozygosity with 99% certainty (Table 2.1) (Morin et al., 2001; Broquet and
Petit, 2004). Thus, I exceeded this threshold and ensured minimal allelic dropout by
confirming homozygotes only when alleles were observed consistently in three reactions.

27

Table 2.1. Allelic dropout rates by locus.
Locus Dropout rate
Dropout rate
^ 2 replicates
No. replicates
needed*
D1s1622** 0.0732 0.0054 2
D18s536** 0.0702 0.0049 2
D12s66** 0.0632 0.0040 2
D3s3038 0.0251 0.0006 2
D7s2204 0.0250 0.0006 2
D4s1627 0.0229 0.0005 2
D11s2002 0.0180 0.0003 2
D2s1326 0.0156 0.0002 2
D3s2459 0.0153 0.0002 2
D7s817 0.0118 0.0001 2
D10s676 0.0109 0.0001 2
D5s1470 0.0091 0.0001 1
D1s1656 0.0080 0.0001 1
D5s1457 0.0073 0.0001 1

Note. * Refers to the number of replicates needed to achieve 99% certainty regarding
homozygous genotypes. Calculated as described in Arandjelovic et al. (2009). ** Refers
to loci used in the screening PCR and therefore includes data from all extracts
irrespective of quality. Higher dropout rates at these loci reflect the inclusion of some
extracts that were excluded from subsequent genotyping.

2.6. Determination of Y-chromosome haplotypes
Additionally, I determined Y-chromosome haplotypes using eight Y-chromosome
microsatellite loci (Moore and Vigilant, 2014; Langergraber et al., 2014b). To do so, I
performed an initial two-step multiplex PCR to assess the variability of 13 human-
derived Y-chromosome microsatellite loci (DYs392, DYs439, DYs469, DYs502,
DYs510, DYs517, DYs520, DYs533, DYs588, DYs562, DYs612, DYs630, DYs632) in
a test set of 29 male individuals (Erler et al., 2004; Langergraber et al., 2007b). Eight loci
were polymorphic, with at least two alleles present. Thus, I typed the remaining 47 males
at only these eight variable loci, which is similar to the number of variable Y-
28

chromosome microsatellite loci found in various other studies of chimpanzees
(Langergraber et al., 2007b; Arandjelovic et al., 2011; Moore and Vigilant, 2014;
Langergraber et al., 2014b), bonobos (Eriksson et al., 2006), western lowland gorillas
(Douadi et al., 2007; Inoue et al., 2013), and humans (Oota et al., 2001; Kumar et al.,
2006; Kayser et al., 2007). I amplified DNA in 20-µL reactions including 8.92 µL
purified H
2
0, 2.0 µL 10x PCR Buffer, 2.5mM each dNTP, 0.80 µL BSA, 25mM MgCl
2
,
0.15mM each forward unlabeled and unnested reverse primer,and 0.10 µL SuperTaq
premixed with TaqStart antibody. I added 5 µL DNA extract to each well for a total well
volume of 20 µL. Each 96-well plate included three independent amplifications for each
DNA extract, plus one positive control from a chimpanzee extract known to amplify and
a minimum of five negative controls, in which purified H
2
0 was added instead of DNA
extract. I centrifuged each 96-well plate following preparation and thermocycled it as
follows: denaturation for 9 min at 94 °C; 29 cycles of 20 sec at 94 °C, 30 sec at 58 °C,
and 30 sec at 72 °C; and a final extension of 30 min of 72 °C, followed by incubation at
10 °C until removal from the thermocycler.
Following thermocycling, I centrifuged well plates. Next, I used the multiplex
PCR product for subsequent singleplex PCRs at each locus. In the singleplex PCR, I
amplified DNA in a 10-µL reaction which included 4.7 µL purified H
2
0, 2.5mM each
dNTP, 25 mM MgCl
2
, 10mM each forward labeled and reverse nested primer, 1 µL of
10x PCR Buffer, 0.4 µL BSA, and 0.04 µL SuperTaq premixed with TaqStart antibody. I
added 2.5 µL of 1:99 diluted multiplex PCR product to each well for a total 10-µL
reaction volume. I centrifuged each singleplex 96-well plate and thermocycled as
29

follows: denaturation for 9 min at 94 °C; 29 cycles of 20 sec at 94 °C, 30 sec at 54 – 61
°C (optimized by primer), and 30 sec at 72 °C; and a final extension at 4 min at 72 °C,
followed by incubation at 10 °C until removal from the thermocycler. I then centrifuged
singleplex PCR products and combined 3 – 4 products in optimized volumes of 0.75 –
1.25 µL with up to 12 µL H
2
0 and 27.4 µL of a 1:135 dilution of ROX labeled
GENESCAN 400HD and H
2
0 to size alleles, then denatured and electrophoresed it, as
described above. Appendix B provides the primer sequences and annealing temperatures
for each Y-chromosome microsatellite locus used in this study.
2.7. Discriminating chimpanzee genotypes
I distinguished individual chimpanzee genotypes using an identity analysis in
CERVUS 3.0.7 software (Kalinowski et al., 2007). Using the allele frequencies of the
study population, I determined the minimum number of loci necessary to achieve a P
IDsib

<0.001, which would allow sufficient power to distinguish among genotypes and
determine with statistical confidence that two matching genotypes from different samples
originate from the same chimpanzee rather than from full siblings. Matching genotypes
were assigned a consensus name and composite genotype data. I matched genotypes
using a minimum of nine matching loci with no mismatches. Up to four mismatches were
permitted to flag potential matches despite genotyping errors. Any mismatch was
therefore either resolved as a true match with corrected errors or as a true mismatch
comprising distinct genotypes. For rare instances in which genotypes matched with P
IDsib

>0.001, I eliminated the less complete of the two genotypes from further analysis. I
30

included extracts typed at as few as four loci if their genotype was unique and did not
match any other genotype.
2.8. Assignment of putative communities and Y-chromosome haplotype
distributions
I defined putative chimpanzee communities according to the spatial clustering of
co-sampled genotypes. In other words, genotypes found in association with other
genotypes, e.g., as part of the same nest group, were assumed to belong to members of
the same community. Using spatial data from these genotype clusters, I constructed 100%
minimum convex polygons (MCPs) using the Minimum Convex Polygon plugin for
QGIS version 2.4.0 software (Quantum GIS, 2014) to represent to the minimum home
ranges of communities based on genotypes found in association. Additional genotypes
found within these polygons were also assumed to originate from members of the same
community, since spatial overlap among communities generally is not expected.
I analyzed Y-chromosome haplotype distributions using a median joining network
constructed in Network 4.6.1.3 Software (Fluxus Technology Ltd), and mapped them
according to putative community distributions to determine whether spatial clustering of
Y-chromosome haplotypes occurred in agreement with putative community distributions.
2.9. Abundance estimation

I estimated total and community-specific population sizes using capwire models
(Miller et al., 2005). I used a likelihood ratio test to evaluate whether the “even capture”
model (ECM), which assumes all individuals have an equal likelihood of capture, or the
“two innate rates” model (TIRM), which allows for individual heterogeneity, provided a
31

better fit to each data set. I expected capture probabilities to vary among individuals due
to spatially and temporally variable search effort and possibly other factors, so I selected
the TIRM when the P-value for the test was <0.10. Where the TIRM was selected I tested
whether partitioning the data into three groups further improved the fit. The test statistic
used was the ratio of multinomial log likelihoods for a two-class vs. a three-class
multinomial distribution of the capture counts (Pennell et al., 2012), and was evaluated at
an alpha level of 0.05. Confidence intervals were estimated by parametric bootstrap
(Miller et al., 2005).
I also estimated chimpanzee density and population size using spatially explicit
capture-recapture (SECR) models for area searches (Efford, 2011). Search area polygons
were defined as the perimeter of aggregations of adjacent, searched grid cells, or as
individual cells if no adjacent cells were searched. I defined a contiguous region of
integration as a 3-km buffer around these polygons, and verified that using a larger region
did not affect estimates of model parameters. I defined two different integration meshes
or “habitat masks” within this region in order to estimate densities both across the
fragmented landscape and within the forest fragments. One mask treated the entire region
of integration as suitable habitat where individuals’ activity centers could occur; for the
other, I used spatial data describing forest cover (Hansen et al., 2013) to exclude
deforested areas from the mask. Multiple detections of the same individual were modeled
as counts during a single sample (Efford et al., 2009a). Temporal variation in search
effort was modeled as the average number of visits to the grid cells included in each
search area polygon (Efford and Fewster, 2013). I assumed detectability declined with
32

distance according to a half normal detection function, and that home range center
locations were Poisson-distributed. I estimated detection parameters by maximizing the
conditional likelihood for area searches, and density as a derived parameter from the
fitted model (Borchers and Efford, 2008; Efford et al., 2009a; Efford, 2011). I estimated
population size by extrapolating the estimated density within forest fragments across
forested habitat within the region of integration (Efford and Fewster, 2013).
All models assumed that (1) the population was demographically closed during
sampling, (2) detections were independent events, and (3) individuals were correctly
identified. Capwire models further assumed (4) geographic closure, and (5) that all
individuals in the population of interest were at risk of detection. SECR estimates did not
rely on assumptions 4 or 5 above, but assumed (6) that animals occupied approximately
circular home ranges, the central location of which was fixed during sampling (Efford,
2004).
Analyses were performed in R version 3.1.2 (R Core Team, 2014) employing
functions implemented in the “capwire” (Pennell et al., 2012), “secr” (Efford, 2015), and
dependent R packages.
2.10. Parentage analysis
2.10.1. False positive error rate estimation in parentage assignments
Microsatellite genotyping has been used to examine parent-offspring relationships
in numerous taxa, and comparisons among genotypes can help determine dispersal
patterns by revealing whether individuals reside in different social groups than their
parents (Arandjelovic et al., 2014). Genotyping can be used to study parentage in two
33

primary ways. The first approach relies on knowing the identity of one parent to
genetically determine the other. For example, when mother-offspring relationships are
known, the alleles shared between them can be identified and candidate fathers can be
tested against the offspring’s second alleles at each locus to identify a matching father
using an exclusion method or a likelihood-based approach (Whittingham, 2004; Lyke et
al., 2013; Vigilant et al., 2015). Second, likelihood-based methods can be used to assign
parentage with statistical confidence in the absence of data on mother-offspring
relationships (Marshall et al., 1998). In this research, I used the second approach, since
mother-offspring relationships were not known among the unhabituated chimpanzees in
the corridor habitat.
Despite the informative potential of this analysis approach, assignment errors are
possible, particularly if assigning one parent in the absence of a genotype from the
second parent (Blouin, 2003; Csilléry et al., 2006; Van Horn et al., 2008). Erroneous
parentage assignments may include either false positive assignments (Type I Error), in
which individuals are erroneously identified as a parent-offspring pair, or false negative
assignments (Type II Error), in which individuals comprising a true parent-offspring pair
fail to be classified as such (Marshall et al., 1998). Although either error should be
avoided, false positive assignments are especially problematic when studying emigration,
as they can lead to erroneous conclusions regarding the occurrence of dispersal events.
To reduce this potential for false positive assignments, I used genetic data to first
approximate the false positive assignment rate based on 217 genotypes from the Ngogo
and Kanyawara chimpanzee communities in Kibale National Park, Uganda, a study site
34

approximately 130 km away from the Budongo-Bugoma corridor habitat and for which
data are available regarding demography and parent-offspring relationships. Kibale
genotypes were derived from noninvasively collected fecal samples and were published
previously (Langergraber et al., 2012). I conducted parentage analyses using KinGroup
v2 (Konovalov et al., 2004) and CERVUS 3.0.7 software (Kalinowski et al., 2007). Both
CERVUS and KinGroup are likelihood-based analysis methods that can identify parent-
offspring dyads and assign a confidence level (p-value) associated with the assignment’s
likelihood (Konovalov et al., 2004; Kalinowski et al., 2007). By using two approaches I
could 1) compare the false positive error rates obtained in each using the Kibale data and
2) find greater support for the results obtained in the corridor data set. I used the same
analysis approaches I planned for the Budongo-Bugoma corridor data, including using all
individuals as putative parents and offspring and ignoring data regarding the age class
and parentage status of each individual, since this information is not available for the
Budongo-Bugoma corridor population. I then used the frequency of false positive
assignments to choose parentage analysis parameters that would lead to low error rates
and ensure high accuracy in assignments among the corridor chimpanzees.
In KinGroup, I used likelihood tests with a primary hypothesis of parent-offspring
kinship and a null hypothesis of unrelated individuals (Goodnight and Queller, 1999) and
used 1,000,000 simulated pairs to perform significance tests. For CERVUS parent-
offspring analyses, I first conducted simulation of parentage analyses by simulating
10,000 offspring. Because this simulation relies on an estimate of the total number of
parents—both present as well as dead and missing—in the population, I estimated 300
35

total candidate parents in the two Kibale communities, a 20% increase over the total
estimated size of these communities, resulting in a proportion of candidate parents
sampled of 0.7233. I set the minimum number of loci typed as 10, the proportion of loci
typed as 0.9888, and I assumed a proportion of mistyped loci of 0.01. For mother-father-
offspring trios, I assumed 162 candidate mothers and 135 candidate fathers in the
population, which reflected the sex proportion for genotyped individuals in the Kibale
data set and produced a good fit between proportions of expected and observed
assignments.
I checked the resulting parentage assignments against parentage data available for
this population, which were derived from pedigree data and genetic kinship analyses
using a more comprehensive data set from 44 autosomal microsatellite loci, X-linked
microsatellite loci, Y-linked microsatellite loci, and mtDNA (Langergraber et al., 2007a).
False positive error rates reflect the percentages of parentage assignments that contain an
error. These were calculated as a range of values, where the lower limit of the range
indicates known false positive assignments and the upper limit indicates the maximum
possible false positive assignments, since not all parentage assignments were known and
I could not determine whether some putative assignments were correct.
I calculated the percentage of false positive assignments when using 80, 95, and
99% confidence levels to examine the impact of confidence levels on error rates. I also
compared the percentage of false positive assignments when using the same 14 autosomal
microsatellite loci typed in the current data set to that when using the 19 available
microsatellite loci in the multiplex PCR (Section 2.5) to determine whether genotyping
36

individuals at 5 additional loci would result in lower false positive error rates in parentage
assignments. I also examined whether false positive assignments were reduced by strictly
excluding dyads and trios that contained 1) mismatching loci, 2) a second-best match for
one parent, given the genotype of the other parent, and 3) sons in parent-offspring trios,
given that closely related males may be associated with more false parentage assignments
and are not of interest, since only daughters are expected to disperse.
2.10.2. Parentage assignments in the corridor chimpanzees
After using the Kibale genotypes to determine which parameters would result in
the lowest false positive error assignments, I conducted parentage analyses with the
Budongo-Bugoma corridor chimpanzee genotypes using these criteria. Specifically, I
only accepted parentage assignments with a confidence level above 95% and only
accepted parent-offspring dyads and trios with zero mismatching loci and no higher
likelihood parent matches identified. I included 176 genotypes from chimpanzees in the
Budongo-Bugoma corridor, a selection restricted to genotypes from chimpanzees in
fragmented forest habitat, and only those individuals genotyped at a minimum of 10 loci.
I included all individuals as potential parents and as potential offspring, since age data
were not available for most individuals and therefore I could not determine which
individuals should be included as candidate parents and offspring, respectively.
For KinGroup parentage analyses, I used likelihood tests with a primary
hypothesis of parent-offspring kinship and a null hypothesis of unrelated individuals
(Goodnight and Queller, 1999) and used 1,000,000 simulated pairs to perform
significance tests. For CERVUS simulation of parentage analyses, I simulated 10,000
37

offspring and tested whether varying the number of candidate parents would improve the
fits of observed and expected parentage assignment proportions. To do so, I relied on the
following population size estimates, as reported in McCarthy et al. (2015) and in Chapter
3: 1) 246, the lower limit of the total confidence interval for the population size, 2) 256,
the point estimate using the capwire TIRM model, 3) 319, the point estimate using the
SECR model, and 4) 357, the upper limit of the total confidence interval for the
population size. However, the number of candidate parents should exceed those
accounted for in the population size estimates, since the former should account for
parents of all individuals currently in the population, including missing and dead parents
that are excluded from population estimates. Therefore, I added 40% and 50% more
individuals to these population size estimates to approximate the total number of
candidate parents, and tested which estimate better improved the fit between observed
and expected assignment proportions. These percentage increases in the number of
candidate parents are greater than the 20% increase used in the Kibale data set due to the
fact that a lower proportion of the overall population was likely sampled in the corridor
data set, thereby requiring a greater compensation for missing individuals. The minimum
number of loci typed was 10, the proportion of loci typed was 0.9724, and I assumed a
proportion of loci mistyped as 0.01. Despite varying these parameters, the number of
assignments at each confidence level was robust, with only minor variations in
confidence level and associated log-likelihood ratio (LOD) scores. Nonetheless, the final
results included were those with the most conservative confidence assignments and LOD
scores.
38

I assessed parent-offspring kinship and associated community affiliations using
several measures. First, I examined occurrences in which individuals comprising dyads
and trios were attributed to two or more distinct communities, as determined based on the
locations of genotyped samples for each individual, as described in Section 2.8 above. I
used these occurrences, hereafter termed ‘mixed community dyads’ or ‘mixed
community affiliation,’ as indicators of likely dispersal events, since individuals are
generally not expected to be found in association, either spatially or socially, with a
community other than their own. I analyzed father-son dyads and mother-daughter dyads
separately to determine whether mother-daughter dyads were more likely than father-son
dyads to contain mixed community affiliation, thereby indicating a dispersal pattern
characteristic of chimpanzees in continuous habitats. I also assessed parent-offspring
mixed sex dyads to determine the proportion with mixed community affiliation. Because
I lacked age data to correspond to each individual genotype, however, I could not
ascertain which individual in each dyad was the parent and which was the offspring, and
therefore could not determine the likely directionality associated with dispersal events.
To overcome this limitation in directionality data, I also analyzed the genotype data for
mother-father-offspring trios, which provide the highest likelihood mother and father for
each potential offspring. I examined trios with high LOD scores (those associated with
>95% confidence) to identify mixed community affiliation, which would indicate both
dispersal and its likely directionality, given that daughters are expected to disperse from
the natal community in which their parents reside.
39

To statistically test whether the sex differences in the proportions of mixed
community affiliation among parent-offspring dyads were statistically significant, I
conducted a permutation test programmed in R. I randomized the assignment of
individuals to sex and restricted the permutations such that both members of any given
parent-offspring pair had identical sexes throughout all permutations. I conducted 1,000
permutations of sex assignment into which I included the original data as one
permutation. As a test statistic I used the chi-square value obtained from the cross-
tabulation of sex of the parent-offspring dyad with whether they co-resided in the same
group or not.
For parent-offspring dyads and trios with mixed community affiliation, I
measured the distance between the sampling locations of each individual to estimate an
approximate mean distance traveled by dispersing individuals. I did so by measuring the
distance between waypoints using the measuring tool in ESRI® ArcMap™ 9.2
(Redlands, CA). For chimpanzees sampled more than once, I calculated a geographical
mean location of the UTM coordinates of all sampling locations for that individual. I also
calculated geographical mean locations representing a central area for each putative
chimpanzee community based on the MCPs for co-sampled genotypes (McCarthy et al.,
2015).
2.10.3. Parentage assignments including Busingiro
I conducted a second parentage analysis that included genotypes from the
protected Budongo Forest Reserve to look for evidence of transfer events between the
corridor habitat and the continuous forest. Dispersal events between these habitats would
40

demonstrate the potential of the corridor for gene flow among these larger regional
chimpanzee populations. This analysis employed a second data set which included the
same 176 corridor genotypes used in the first analysis as well as 53 genotypes from
Busingiro, a chimpanzee community in the southern area of the Budongo Forest (Figure
2.1). Of these 53 additional genotypes, 14 originated from samples collected during the
current study from Siiba Forest Reserve (see Chapter 3 for further details) while 39 were
collected and genotyped previously (Langergraber et al., 2011). I ran this analysis
separately from that of the corridor genotypes alone, given the possibility that factors
influencing dispersal likelihood in continuous forest habitat differ from those for the
fragmented forest habitat and that lumping all genotypes together may have obscured
these differences.
As with the corridor data set, I included all individuals as candidate parents and
offspring given the lack of age data for each genotype. I conducted KinGroup analyses
exactly as described above for the corridor data set. I also conducted CERVUS
simulation of parentage analyses as described above, with the exception of adding 100
candidate parents, given an approximate estimate of the Busingiro community’s size
(n ≈ 70; Reynolds, 2005) and an estimate of 40% more unsampled candidate parents. As
above, I used a minimum of 10 loci typed and assumed a proportion of mistyped loci of
0.01. The proportion of loci typed was 0.9663 for this data set. I also conducted a
permutation test with the parent-offspring dyads produced from this second data set, and
used the same methods described above.

41

2.10.4. Spatial analysis of repeatedly genotyped chimpanzees
Finally, I compared the spatial locations of samples collected from the same
chimpanzee across multiple data collection days over the study period to determine
whether all samples from each individual fell within 1) a single MCP and were therefore
attributed to a single community or 2) multiple MCPs, indicating a likely dispersal event
during the study period. Because chimpanzees are territorial and typically associate with
members of just one community, they are not generally expected to be found within the
MCP associated with another community unless due to a dispersal event (Nishida, 1979;
Goodall, 1986; Herbinger et al., 2001).

42

Chapter 3
A genetic investigation of chimpanzee distribution and abundance in a
fragmented forest landscape

3.1. Introduction
Habitat loss and fragmentation are key threats to the survival of many species
(Wiens, 1996), with global deforestation resulting in the majority of remaining forest
lying within 1 km of a forest edge (Haddad et al., 2015). Fragmentation can isolate
populations, thereby reducing genetic diversity and population viability, which may result
43

in local extinctions (Stratford and Stouffer, 1999; Gerlach and Musolf, 2000; Keller and
Largiader, 2003). As wildlife populations face increasing anthropogenic threats, there is
growing urgency to better understand how species respond to environmental
disturbances. Although degraded habitats are often thought to have limited conservation
value, many threatened species inhabit such environments (Sheil and Meijaard, 2010).
Riparian forest fragments in particular can offer suitable habitat, providing dense
resources to support wildlife (Gautier-Hion and Brugiere, 2005; McLennan and Plumptre,
2012). In addition, fragmented forests can sustain connectivity by linking larger
populations, thereby enhancing gene flow and population viability (Ranta et al., 1998;
McShea et al., 2009; Bergl et al., 2012). Therefore, the potential of fragmented habitats to
support viable populations must be carefully considered alongside the peril they pose to
wildlife.
Large-bodied, wide-ranging mammals such as great apes are among the taxa most
affected by growing habitat fragmentation. These species often live in unprotected areas,
which are particularly vulnerable to forest loss and fragmentation (Gaveau et al., 2007;
Brncic et al., 2010). In East Africa, deforestation has led to increasing habitat
fragmentation and poses a primary threat to the survival of eastern chimpanzees, Pan
troglodytes schweinfurthii (Plumptre et al., 2010). Eastern chimpanzees inhabit lowland
and montane forest, woodland, savanna, and swamp forest habitats throughout various
parts of East Africa, with much of their current range occurring outside protected areas
(Plumptre et al., 2010). Three-quarters of chimpanzees in Tanzania are estimated to live
outside national parks (Moyer et al., 2006). In Uganda, logging has led to a 37%
44

reduction in forest cover between 1990 and 2010 (Wiens, 1996; FAO, 2010), and much
of this deforestation occurred outside protected areas, leaving chimpanzees in such
habitats vulnerable to local extinction (MWLE, 2002; Haddad et al., 2015). Similar
patterns have also been reported for chimpanzees in West Africa (Stratford and Stouffer,
1999; Gerlach and Musolf, 2000; Keller and Largiader, 2003; Kormos et al., 2003;
Brncic et al., 2010).
Because chimpanzees are an endangered species (Oates et al., 2008), it is essential
to better understand their ability to persist in fragmented and degraded habitats.
Moreover, precise estimates of the sizes and distributions of remaining populations are
needed in order to establish research priorities and conservation management strategies.
Such estimates can be challenging to obtain, however. Chimpanzee habituation allows for
direct monitoring and hence precise censuses, but is a lengthy process which is
necessarily restricted to small numbers of individuals, and may not be ethically
appropriate or logistically feasible for many populations (Gautier-Hion and Brugiere,
2005; McLennan and Hill, 2010; McLennan and Plumptre, 2012; Gruen et al., 2013).
Nest count surveys can be used to estimate the distribution and abundance of
unhabituated chimpanzee populations. However, these survey methods may be inaccurate
and lack the precision necessary to determine trends in population size (Plumptre and
Reynolds, 1996; Ranta et al., 1998; Devos et al., 2008; McShea et al., 2009; Boyko and
Marshall, 2010; Bergl et al., 2012). Such studies are also arduous to carry out, as
commonly used nest count methods rely on data regarding nest decay rates and nest
building and re-use rates, which can be highly variable and are often unknown locally
45

(Plumptre and Reynolds, 1996; Gaveau et al., 2007; Kouakou et al., 2009; Boyko and
Marshall, 2010; Brncic et al., 2010; Buckland et al., 2010). Recently, camera trapping
and passive acoustic monitoring have also been utilized to census apes (Plumptre et al.,
2010; Head et al., 2013; Heinicke et al., 2015; Kalan et al., 2015). However, these
techniques are still in their infancy, while methods for efficiently automating individual
identification are still in development (Moyer et al., 2006; Kühl and Burghardt, 2013).
The challenges of accurately and precisely enumerating chimpanzee populations
are similar to those posed by surveys of other rare and elusive mammal populations,
including bears (Howe et al., 2013), gorillas (Guschanski et al., 2009; Arandjelovic et al.,
2010; 2015), African elephants (Junker et al., 2008), Eurasian otters (Arrendal et al.,
2007), and giant pandas (Zhan et al., 2006). These challenges have led to the widespread
implementation of genetic censusing (e.g., in chimpanzees (Arandjelovic et al., 2011;
Chancellor et al., 2012; Moore and Vigilant, 2013)), which relies on the characterization
of individual DNA profiles derived from noninvasively collected samples (Taberlet et al.,
1999). The minimum number of individuals using the surveyed area is determined by the
number of unique profiles, and resampling frequency can be used to estimate the number
of animals that went undetected (Lukacs and Burnham, 2005; Petit and Valiere, 2006).
Standard approaches for genetic censusing have relied upon accumulation curves
and Bayesian estimators, along with more recent capwire models (Pritchard et al., 2000;
Miller et al., 2005; Petit and Valiere, 2006). However, the population size estimates these
methods provide cannot be converted to density estimates except by collecting ancillary
data or making restrictive assumptions (Soisalo and Cavalcanti, 2006; Obbard et al.,
46

2010). Density is generally a valuable parameter because it can be compared across
populations of varying size and geographic scope, and used as an indicator for behavioral
ecology and conservation questions relating to, for example, resource density, group
structure and dynamics, and hunting pressure (Butynski, 1990; Howe et al., 2013; Imong
et al., 2014). Recently developed SECR models allow the density of geographically open
populations to be estimated directly from spatially-referenced detections of individuals,
by modeling probability of detection as a (usually decreasing) function of the distance
between detectors or areas searched and individuals’ centers of activity (Efford, 2004;
Borchers and Efford, 2008; Efford et al., 2009a; Efford, 2011). SECR models are robust
to spatial gaps in data collection (Borchers and Efford, 2008; Efford, 2011), which are
common when sampling elusive species in degraded or mixed habitats.
In western Uganda, the approximately 1200-km² landscape of the Northern
Albertine Rift separating the Budongo and Bugoma Forests illustrates such a degraded
mosaic habitat. The government-owned Budongo and Bugoma Forest Reserves are each
inhabited by over 600 chimpanzees, together composing approximately one-quarter of
the estimated total chimpanzee population in Uganda [5,000 individuals (Plumptre et al.,
2003)]. The corridor between these forest blocks is a human-dominated landscape
comprising mosaic riparian forest with villages, agricultural lands, and natural grasslands
(McLennan, 2008). Most forests in this habitat are privately owned, but a few small
government-owned forest reserves are present. The small forests in this region have been
targeted for potential corridor enhancement given the vital role they may play for gene
flow in numerous species throughout this region (Nangendo et al., 2010).
47

Despite the conservation potential of this habitat, few studies have examined the
population size and distribution of its chimpanzees. A nationwide chimpanzee census
used a nest count survey of forest fragments near the Bugoma Forest to extrapolate an
estimate of ~70 chimpanzees in the corridor region (Plumptre et al., 2003). Later,
McLennan (2008) found evidence of chimpanzees throughout the corridor habitat and
estimated a total regional population of up to 260 individuals, an extrapolation derived
from the estimated density of one chimpanzee community (Bulindi) in the corridor area
(McLennan, 2008). Given the potentially vital role of this chimpanzee population in
maintaining gene flow among chimpanzees of the Northern Albertine Rift, it is important
to better understand the size and distribution of this population. The goal of this study
was to use genetic censusing techniques to estimate the population size and distribution
of this corridor population of chimpanzees in western Uganda. To do so, we estimated
chimpanzee density using a spatially explicit model, as well as estimating abundance
using both capwire and spatially explicit models. We further examined the number and
spatial distribution of putative chimpanzee communities by analyzing the clustering of
co-sampled genotypes. Additionally, because chimpanzees typically exhibit male
philopatry and female dispersal, we examined the clustering of Y-chromosome
haplotypes, which are paternally inherited and therefore can be used to reveal community
affiliations (Langergraber et al., 2007b; Arandjelovic et al., 2011; Moore and Vigilant,
2013).

48

3.2. Methods
3.2.1. Study area and subjects
I collected data in Hoima and Masindi Districts, Uganda, in the corridor region
between the Budongo and Bugoma Forests (1°37' – 1°68'N and 31°1' - 31°6'E; Figure
2.1) from October to December 2011 and October 2012 to September 2013. The region
between these forests measures approximately 40 km long by 30 km wide and is a mosaic
habitat composed of agricultural land, villages, riparian forest fragments, and grasslands.
This mosaic landscape served as the study area for this research. Further details about the
Budongo and Bugoma Forests, as well as the mosaic habitat between them, are provided
in Chapter 2. Chimpanzee presence in this habitat had previously been confirmed
(Plumptre et al., 2003; McLennan, 2008), though little has been known regarding the
population size and distribution. Further details regarding prior research on this
chimpanzee population are also provided in Chapter 2.
3.2.2. Genetic census methods
I collected fecal samples noninvasively throughout the study area, with particular
focus on searching riparian forest fragments for evidence of chimpanzees. I also acquired
information on chimpanzee presence in McLennan (2008) and via informal discussion
with local inhabitants. As described in Chapter 2, strictly systematic survey methods were
neither feasible nor ethically appropriate in this habitat. Instead, I centered search effort
in forest fragments around village boundaries, which typically encompass settlements,
farmland, and privately owned forests. In accordance with local customs, prior to
searching a forest fragment I first gained permission from the chairperson of the village
49

in which the forest fragment was located, and from individuals who identified themselves
as landowners of the forest fragment. I used satellite imagery to identify the forest
fragments located within the boundaries of a given village, and visited accessible and
permitted forest fragments within the boundaries of that village. I divided the study area
into a grid of 1 sq. km cells and recorded when any part of each cell was searched (Figure
2.6).
Chimpanzee fecal samples were typically easy to identify because of 1) their
locations under chimpanzee nests and along trails, 2) their characteristic size, shape, and
odor, and 3) the absence of other sympatric large-bodied nonhuman primates. Although
olive baboons (Papio anubis) produce dungs that can superficially resemble those of
chimpanzees (pers. obs.), they have been eradicated from many parts of the study area.
For any sample of uncertain species origin, however, I confirmed the species prior to
sample storage and analysis by macroscopic analysis as described in Chapter 2.
To achieve adequate resampling, I estimated a target sample size for each region
with chimpanzee presence across the study area. These target sample sizes were
determined as described in detail in Chapter 2, with the goal of collecting at least three
times the number of samples as expected individuals in order to obtain a narrow
confidence interval for population size estimates using mark-recapture methods (Miller et
al., 2005; Petit and Valiere, 2006; Arandjelovic et al., 2010). Because additional
information on chimpanzee presence was gained over the course of the study period,
these target sample sizes were adjusted as necessary. To help achieve this sampling goal
and to ensure adequate resampling across fission-fusion chimpanzee communities, I
50

attempted to search forests a minimum of once every three months, except where local
research permissions were granted only for a limited time period.
I collected samples under nests and opportunistically along chimpanzee trails and
at feeding sites. For each sample, I recorded a GPS waypoint with a Garmin GPSMap®
60CSx. I recorded samples with unique identification numbers corresponding to GPS
waypoints, and with party association data when applicable. I recorded samples as
belonging to a party when two or more same-age samples were collected within 30 m of
each other. Distances were determined using GPS data and, when necessary, a Bresser®
laser Range Finder 800 to ensure accuracy. I avoided collecting two samples under the
same nest or in close proximity on trails, due to the likelihood of collecting redundant
samples from the same individual and the possibility that closely deposited samples may
have cross-contaminated each other. I collected samples and stored them according to the
two-step ethanol-silica method described in Nsubuga et al. (2004) and detailed in Chapter
2.
I carried out data collection with the permission of the Uganda National Council
for Science and Technology, the Uganda Wildlife Authority, and the National Forestry
Authority of Uganda. Additional permissions were granted by local landowners where
applicable, as described above. Because fecal sample collection was entirely noninvasive
and required no contact with the chimpanzees, ethical consent was not necessary for this
project.

51

3.2.3. DNA extraction and amplification
I stored samples at room temperature in silica (as described in detail in Ch. 2) in
the field for up to 6 months before shipping them to the Max Planck Institute for
Evolutionary Anthropology, Leipzig, Germany, where they were then stored at 4 °C prior
to extraction. I extracted DNA using either the GeneMATRIX Stool DNA Purification
Kit (Roboklon) according to manufacturer’s instructions or the QIAmp Stool kit
(QIAGEN) with minor procedural adjustments (Nsubuga et al., 2004).
I used autosomal microsatellite loci to determine individual chimpanzee
genotypes. To do so, each DNA extract was first evaluated by simultaneously amplifying
three autosomal microsatellite loci (listed in Chapter 2 and Appendix A), along with an
X-Y homologous segment of the amelogenin gene, used for sex determination (Bradley
et al., 2001), in a one-step multiplex PCR. Detailed information regarding this PCR
procedure is provided in Chapter 2. PCR products were electrophoresed using an ABI
PRISM 3100 Genetic Analyser. I used GeneMapper version 3.7 (Applied Biosystems) to
analyze the data.
DNA extracts that reliably amplified at a minimum of 3 of the 4 loci in at least 3
independent amplifications were then genotyped in triplicate at an additional 11
autosomal microsatellite loci (Appendix A). Extracts that failed to meet these criteria
were not amplified further. The additional 11 loci were amplified in a two-step multiplex
PCR procedure as described in detail in Arandjelovic et al. (2009) and in Chapter 2.
At each locus, heterozygous genotypes were confirmed by observation in at least
two independent reactions (Morin et al., 2001; Arandjelovic et al., 2009). Homozygous
52

genotypes were confirmed when observed in a minimum of three independent reactions.
Individual loci that failed to meet these criteria were instead coded with asterisks and
were excluded from analyses. To further ensure that apparent homozygotes were not the
result of allelic dropout, I calculated allelic dropout rates by locus after recording all
alleles and confirmed that a maximum of two replicates was required at any locus to
confirm homozygosity with 99% certainty (Morin et al., 2001; Broquet and Petit, 2004).
Thus, I exceeded this threshold and ensured minimal allelic dropout by confirming
homozygotes only when alleles were observed consistently in three reactions. Table 2.1
presents data regarding the dropout rate for each microsatellite locus used in this
research.
3.2.4. Determination of Y-chromosome haplotypes
To determine Y-chromosome haplotypes, I first used a two-step multiplex PCR to
assess the variability of 13 human-derived Y-chromosome microsatellite loci in a test set
of 29 male individuals (Erler et al., 2004; Langergraber et al., 2007b). Chapter 2 provides
detailed information regarding the PCR procedure used to determine Y-chromosome
haplotypes. Appendix A displays the Y-chromosome microsatellite loci used in this
research. Eight loci were polymorphic, with at least two alleles present. Thus, the
remaining 47 males were typed at only these eight variable loci, which is similar to the
number of variable Y-chromosome microsatellite loci found in various other studies of
chimpanzees (Langergraber et al., 2007b; Arandjelovic et al., 2011; Moore and Vigilant,
2014; Langergraber et al., 2014b), bonobos (Eriksson et al., 2006), western lowland
53

gorillas (Douadi et al., 2007; Inoue et al., 2013), and humans (Oota et al., 2001; Kumar et
al., 2006; Kayser et al., 2007).
3.2.5. Discriminating chimpanzee genotypes
I distinguished individual chimpanzee genotypes using an identity analysis in
CERVUS 3.0.7 software (Kalinowski et al., 2007). Using the allele frequencies of the
study population, I determined the minimum number of loci necessary to achieve a P
IDsib

<0.001, which would allow me sufficient power to distinguish among genotypes and
determine with statistical confidence that two matching genotypes from different samples
originate from the same chimpanzee rather than from full siblings. Matching genotypes
were assigned a consensus name and composite genotype data. Up to four mismatches
were permitted to flag potential matches despite genotyping errors. Any mismatch was
therefore either resolved as a true match with corrected errors or as a true mismatch
comprising distinct genotypes. For rare instances in which genotypes matched with P
IDsib

>0.001, the less complete of the two genotypes was eliminated from further analysis.
3.2.6. Assignment of putative communities and Y-chromosome haplotype distributions
Putative chimpanzee communities were defined according to the spatial clustering
of co-sampled genotypes. In other words, genotypes found in association with other
genotypes, e.g., as part of the same nest group, were assumed to belong to members of
the same community. Further, additional lone samples from those individuals, such as
samples found singly on chimpanzee trails, were inferred to lie within the home range of
that individual’s community (Arandjelovic et al., 2011). Using spatial data from these
genotype clusters, I constructed 100% MCPs using the Minimum Convex Polygon Plugin
54

for QGIS version 2.4.0 software (Quantum GIS, 2014) to represent the minimum home
ranges of communities based on genotypes found in association. Additional genotypes
found within these polygons were also assumed to originate from members of the same
community, since extensive spatial overlap among territories is generally not expected
(Nishida, 1979; Goodall, 1986; Herbinger et al., 2001). Y-chromosome haplotype
distributions were analyzed using a median joining network constructed in Network
4.6.1.3 Software (Fluxus Technology Ltd), and were mapped according to putative
community distributions to determine whether spatial clustering of Y-chromosome
haplotypes occurred in agreement with putative community distributions.
3.2.7. Abundance estimation
I estimated total and community-specific population sizes using capwire models
(Miller et al., 2005). I used a likelihood ratio test to evaluate whether the ECM model,
which assumes all individuals have an equal likelihood of capture, or the TIRM model,
which allows for individual heterogeneity, provided a better fit to each data set. I
expected capture probabilities to vary among individuals due to spatially and temporally
variable search effort and possibly other factors, so I selected the TIRM when the P-value
for the test was <0.10. Where the TIRM was selected I tested whether partitioning the
data into three groups further improved the fit. The test statistic used was the ratio of
multinomial log likelihoods for a two-class vs. a three-class multinomial distribution of
the capture counts (Pennell et al., 2012; Stansbury, 2012), and was evaluated at an alpha
level of 0.05. Confidence intervals were estimated by parametric bootstrap (Miller et al.,
2005).
55

I also estimated chimpanzee density and population size using SECR models for
area searches (Efford, 2011). Search area polygons were defined as the perimeter of
aggregations of adjacent, searched grid cells, or as individual cells if no adjacent cells
were searched. I defined a contiguous region of integration as a 3-km buffer around these
polygons, and verified that using a larger region did not affect estimates of model
parameters. I defined two different integration meshes or “habitat masks” within this
region in order to estimate densities both across the fragmented landscape and within the
forest fragments. One mask treated the entire region of integration as suitable habitat
where individuals’ activity centers could occur; for the other, I used spatial data
describing forest cover (Hansen et al., 2013) to exclude deforested areas from the mask.
Multiple detections of the same individual were modeled as counts during a single
sample (Efford et al., 2009b). Temporal variation in search effort was modeled as the
average number of visits to the grid cells included in each search area polygon (Efford et
al., 2013). I assumed detectability declined with distance according to a half normal
detection function, and that home range center locations were Poisson-distributed. I
estimated detection parameters by maximizing the conditional likelihood for area
searches, and density as a derived parameter from the fitted model (Borchers and Efford,
2008; Efford et al., 2009a; Efford, 2011). I estimated population size by extrapolating the
estimated density within forest fragments across forested habitat within the region of
integration (Efford and Fewster, 2013).
All models assumed that (1) the population was demographically closed during
sampling, (2) detections were independent events, and (3) individuals were correctly
56

identified. Capwire models further assumed (4) geographic closure, and (5) that all
individuals in the population of interest were at risk of detection. SECR estimates did not
rely on assumptions 4 or 5 above, but assumed (6) that animals occupied approximately
circular home ranges, the central location of which was fixed during sampling (Efford,
2004).
Analyses were performed in R version 3.1.2 (R Core Team, 2014) employing
functions implemented in the “capwire” (Pennell et al., 2012), “secr” (Efford, 2015), and
dependent R packages.
3.3. Results
3.3.1. Genetic sampling and discriminating individual chimpanzees
I collected a total of 865 fecal samples over 633 km
2
from October to December
2011 and October 2012 to September 2013 (Figure 3.1). Of these, 662 (76%) amplified
reliably at a minimum of three of four test loci and were thus genotyped at an additional
11 loci. Based on the allele frequencies, I calculated that comparison at a minimum of
nine loci was necessary to obtain a P
IDsib
< 0.001 and thus confidently determine that
identical genotypes originated from the same individual rather than two different
individuals, including for example full siblings. Appendix D provides allele frequency
data associated with this population. Of the 662 genotypes, 459 matched exactly to one or
more other genotypes and were merged to create consensus genotypes. An additional five
genotypes were removed from analysis because they matched other genotypes with a
P
IDsib
> 0.001.
57

Figure 3.1. Genotyped sample collection locations across the study area. Not all samples
are visible due to map scaling. The black line indicates the region of integration used in
the SECR model. Samples outside the region of integration were collected in Siiba Forest
Reserve and were excluded from analysis. Green indicates forest cover during the study
period, as provided by Hansen et al. (2013).

The final genotype list consisted of 128 individuals identified in multiple samples (range:
2-12) and 68 individuals genotyped once. For the analyses presented in Chapter 3, I
removed 16 genotypes representing 14 individuals from a chimpanzee community in
Siiba Forest Reserve, a continuous forest located to the south of the Budongo Forest
(Figure 3.1). Since these genotypes originated from few samples in an under-searched
area of continuous forest habitat, they were not informative or representative of the study
58

population. The remaining genotypes represented 182 individuals, of which 111 (61%)
were identified as female and 71 (39%) as male. Consensus genotypes for these
individuals were 95% complete, with 134 individuals typed at all 14 loci. Nine
individuals were genotyped at fewer than nine loci, but their genotypes did not match any
others and thus were retained in the data set. The mean number of captures per genotyped
individual was 3.5.
3.3.2. Putative chimpanzee communities and Y-chromosome haplotype distributions
By grouping genotypes from samples found together I identified 10 spatial
clusters that were geographically distinct from one another, thus suggesting the presence
of at least nine potential communities in the study area, along with one additional cluster,
Kiraira. Community-specific population sizes estimated using capwire ranged from 5 to
48, and totaled 244 (Table 3.1). Data were insufficient to evaluate the fit of different
models to data from Kiraira, and the upper confidence limit under the ECM was equal to
the maximum population size I provided when fitting the model, indicating estimation
problems. Figure 3.2 displays the distribution of putative communities. From 76 total
males (including those from Siiba) I found 14 Y-chromosome haplotypes, and these were
99% complete. Ten of these haplotypes were observed respectively only in single
putative communities, thereby supporting community association data from genotype
clusters. However, four haplotypes were shared among more than one putative
community (Haplotypes B, G, I, and M). Overall, haplotypes shared a high degree of
similarity as shown by their proximity in a median joining network (Figure 3.2).
Appendix E provides detailed information regarding each Y-chromosome haplotype.
59

Table 3.1. Community-specific capwire estimates.
Group n N (groups) 95% CI Monitoring estimate
Bulindi 17 19 (2) 17 – 21 19
Kasokwa 8 8 (1) 8 – 9 15
Kasongoire 28 38 (3) 31 – 56 34
Katanga 26 48 (1) 31 – 83
Kiraira 5 5 (1) 5 – 200
Kiryangobe 13 15 (2) 13 – 20
Kityedo 16 18 (2) 16 – 21
Kyamuchumba 11 13 (1) 11 – 19
Mukihani 25 46 (2) 36 – 70
Wagaisa 33 34 (2) 33 – 38
Group-Specific Total

Overall Total
182

182
244

256 (3)

246 -321

Note. Numbers of unique individuals genotyped (n) and population sizes (N) are shown
with 95% confidence intervals (CI) for each putative chimpanzee community in the study
area. The numbers of groups of chimpanzees with different probabilities of detection
included in the estimate model appear in parentheses following the abundance estimate.
Monitoring estimates refer to the number of chimpanzees reported during the study
period for communities monitored for research or conservation (provided via pers. comm.
as follows: Bulindi, Matthew McLennan; Kasokwa, Janette Wallis; Kasongoire, Geoffrey
Muhanguzi). The sum of group-specific estimates, and the estimate of total population
size obtained by pooling data from all communities for analysis, appear at the bottom.

3.3.3. Abundance estimation
A likelihood ratio test supported the capwire TIRM model over the ECM model when fit
to the full data set (ratio = 132.4, P < 0.01). Partitioning into three groups was also
supported (P < 0.01). I obtained a population size estimate of 256 (95% confidence
interval [CI] = 246–321). The SECR estimate of average density across the fragmented
landscape was 0.404 chimpanzees per km
2
(standard error [SE] = 0.033, 95% CI = 0.34–
0.47). The SECR density within forest fragments was 2.13 chimpanzees per km
2

60

Figure 3.2. Putative chimpanzee communities (a) and associated Y-chromosome
haplotypes (b). (a) MCPs for genotyped samples found in association. Names of putative
chimpanzee communities correspond to nearest villages and are listed below the MCP,
with Y-chromosome haplotypes found in that putative community listed in parentheses.
Underlined names indicate researched communities with preexisting data on approximate
community sizes and home range extents. Each community is represented by a unique
color. (b) Median joining network for the 14 Y-chromosome haplotypes. The relative
similarity of haplotypes is represented by the lengths of branches, and the relative
frequency of occurrence of each haplotype is indicated by the sizes of circles. Colors in
haplotype circles correspond to putative communities in (a) exhibiting that haplotype.
Gray indicates forest cover during the study period, as provided by Hansen et al. (2013).

61

(SE = 0.17, 95% CI = 1.8–2.5). The associated estimate of population size was 319 (SE =
17.6, 95% CI = 288–357). The precision of the population size estimates, calculated as
the CI width divided by the estimate, was 29% and 22% for the capwire and SECR
estimates, respectively. The coefficient of variation (CV) of the SECR population
estimate, measured as SE divided by the estimate, was 0.055.
3.4. Discussion
3.4.1. Abundance estimation
I employed two established estimators to determine the abundance of
chimpanzees in a human-dominated landscape composed of small fragmented forests
amid agricultural land. While a previous census estimated a population of ~70
chimpanzees in the study region (Plumptre et al., 2003), I obtained population size
estimates of 256 and 319, more than tripling this previous estimate. These substantially
higher estimates likely reflect the advantages of this approach over indirect abundance
estimates, which can lack accuracy if little is known regarding habitat suitability and
species distribution (Plumptre and Reynolds, 1996; Boyko and Marshall, 2010). Indeed,
my estimates more closely resemble those of McLennan (2008), who extrapolated
chimpanzee density in the studied Bulindi community to similarly suitable habitat across
the corridor region. One could alternatively explain the higher estimates as evidence of
substantial population growth since the time of the previous census. However, given the
slow interbirth interval of chimpanzees and the high rate of habitat loss throughout the
region over the intervening years between surveys, this explanation seems highly
improbable.
62

In addition to the improved accuracy of our estimates, the high recapture rate for
chimpanzee genotypes across the study area also resulted in a relatively high degree of
precision. Though adequate sampling is necessary to achieve precise estimates using
mark-recapture methods (Miller et al., 2005; Petit and Valiere, 2006), this has proved
challenging in numerous prior studies of great apes (Arandjelovic et al., 2011; Moore and
Vigilant, 2013; Roy et al., 2014). My relatively high rate of resampling was aided by
habitat heterogeneity, which led to a clustering of samples in confined areas of suitable
forested habitat despite the large size of the total study area. I also directed search efforts
based on reports from local residents who live near the chimpanzees, which further
benefited my sampling success rate.
Despite their relative precision, I found differences in the population size
estimates provided by the capwire and SECR estimators, which may be an artefact of the
differences in the specific quantities estimated by the models and their applications to a
population with a heterogeneous distribution over a large spatial area. Capwire assumes
all individuals were at risk of being detected. However, this may not have been the case,
given the presence of spatial gaps in sampling and chimpanzees’ fission–fusion social
structure, which could have caused resampling of parties of similar composition while
failing to detect some community members, particularly where search effort was low.
This may have caused underestimation of overall and group-specific population sizes
when using capwire. To examine this possibility, one can assess the relative accuracy of
the group-specific capwire estimates by comparing them with community size estimates
based on observational data from communities being monitored for research or
63

conservation. Of three such communities, two (Kasongoire and Bulindi) resulted in
monitoring estimates that fall within the 95% confidence interval of my capwire
estimates (Table 3.1). For the third community, Kasokwa, the TIRM estimate I obtained
was substantially lower than the monitoring estimate. Spatial search effort in this region
was relatively light, which may have resulted in identification of fewer genotypes from
chimpanzees there and a corresponding underestimate as compared to Kasongoire, for
which available search effort data reflect a broader spatial area search (Figure 2.6).
Therefore, where search effort was greater and more broadly distributed, the TIRM
estimate appears to be highly accurate, while in undersearched areas the TIRM estimate
may fall short.
In contrast, by modeling detection probability as a function of distance between
animals’ activity centers and areas searched, SECR models allow for the presence of
additional individuals whose probability of detection is negligible because they spend
most or all of their time outside the areas searched. However, this also means that the
SECR model could have slightly overestimated population size if forest fragments far
from the areas searched were, in fact, not occupied. I also note that the SECR region of
integration included small portions of contiguous forest in the Katanga area (near Siiba
Forest Reserve; Figure 3.1), such that my SECR model slightly overestimates the number
of animals that rely exclusively on small forest fragments (between forest reserves).
Despite the differences between estimators, the capwire and SECR estimates were
qualitatively similar, with overlapping confidence intervals. Perhaps most importantly,
64

the 182 distinct genotypes alone confirm a minimum corridor population size far
exceeding that estimated in the previous nationwide census of chimpanzees in Uganda.
Additionally, the estimates presented here can be considered conservative if
applied to the entire study area. The search area did not include some southern sections of
the corridor, and I refrained from extrapolating density estimates to these areas since little
is known regarding the current distribution of chimpanzees there (Figure 2.6).
Chimpanzees have, however, been reported to inhabit forest fragments to the south and
east of Wambabya Forest Reserve near the villages of Bugambe, Munteme, and Buhimba
in Hoima District (JGI-UWA, 2002; Plumptre et al., 2003; 2011). Additionally,
Wambabya Forest has an estimated chimpanzee population of 136 individuals (Plumptre
et al., 2003). My searches of the northern part of this forest yielded no evidence of
chimpanzees, though relatively few searches could be allocated to this region. One
additional chimpanzee community may also inhabit Rwensama Forest Reserve, just south
of the Budongo Forest, but little is known regarding the size or range of this putative
community. Future censuses in these areas may help clarify chimpanzee population size
and distribution in Rwensama Forest Reserve, Wambabya Forest Reserve, and
neighboring fragments of riparian forest.
My estimates may also be conservative given that genetic censuses of great ape
population size may tend to under-sample infants and juveniles due to the difficulty of
finding their fecal samples. Based on a review of published demographic data from
habituated chimpanzee communities, as detailed in Appendix F, an average of 39% of a
chimpanzee community is typically composed of infants and juveniles. If none of these
65

individuals are sampled and are effectively at zero risk of detection, then the total size of
a community or population will be underestimated. However, given the efforts to
exhaustively search areas with evidence of chimpanzee presence in this study, as well as
my data indicating the small bolus size of some samples, I have reason to believe some
infants and juveniles were sampled in my study population. If so, their detection risk
would be elevated and the estimates should have adjusted accordingly to accommodate
them.
Despite the advantages of these abundance estimators, potential model
assumption violations should still be noted. Given the timescale of this study (15 months
of sample collection over a total period of two total years), it is possible that I violated the
assumption of demographic closure. However, given the slow life history traits of
chimpanzees, whose average interbirth interval is more than five years (Emery Thompson
et al., 2007), this is unlikely since relatively few deaths, births, or migrations into or
outside of the corridor area would be expected to occur during this time. In addition,
Arandjelovic et al. (2010) found similar TIRM estimates when one longer-term (3 years)
and two shorter-term (<1 year) sampling periods were compared for the same population
of western lowland gorillas, suggesting the sampling period used in this study should not
have strongly impacted abundance estimates. Community transfers would violate the
assumption of fixed activity centers, but given the relative infrequency of female
transfers in eastern chimpanzees (Goodall, 1986; Nishida et al., 2003; Langergraber et al.,
2014a), few instances are expected during the study period.

66

3.4.2. Chimpanzee density in the corridor region
I used SECR models to estimate chimpanzee density both across the entire
fragmented study area and within the forest fragments, obtaining estimates of 0.40 and
2.13 per km
2
, respectively. Estimated densities for chimpanzees in the Budongo and
Bugoma Forests are approximately 1.3 and 2 chimpanzees per km
2
, respectively
(Plumptre et al., 2003; Plumptre and Cox, 2005). Therefore, it appears that while the
overall density of chimpanzees in the corridor region is relatively low, the density within
forest habitat is much higher and may exceed that in continuous forest nearby. Chancellor
et al. (2012) found similarly high chimpanzee density (~2.1 individuals/km
2
) for eastern
chimpanzees in a forest fragment of western Rwanda despite lower densities in montane
rainforest nearby. Such findings may 1) indicate a crowding effect, whereby chimpanzee
density is particularly high in small remaining areas of suitable habitat, 2) reflect the
expected distribution of chimpanzees in a mosaic habitat with clumped resources, or 3)
result from a combination of these factors. Previous estimates, however, have employed
various non-genetic survey methods, thereby limiting our ability to draw conclusions by
comparing densities across fragmented and continuous forests.
3.4.3. Putative communities and Y-chromosome haplotypes
The spatial clustering of genotypes suggests the presence of at least nine different
chimpanzee communities in the study area, in a non-overlapping distribution similar to
that seen elsewhere among studied chimpanzees (Goodall, 1986; Herbinger et al., 2001).
Overall, Y-chromosome haplotypes show a structuring across putative communities, but
4 of 14 haplotypes are shared among more than one putative community. This overlap
67

could indicate 1) remnants of older diversity from precursor groups in the region that
eventually fissioned into different chimpanzee communities, 2) transfer events in which
parous females with sons emigrated to new communities, thereby bringing with them
new Y-chromosome haplotypes, 3) instances of extra-group copulations resulting in male
offspring of different communities sharing the same Y-chromosome haplotype, or 4)
mutations at microsatellite loci that caused closely related Y-chromosome haplotypes to
converge into a single haplotype as defined using our markers. The reasons for its
occurrence in this study cannot yet be determined but may result from one or a
combination of these factors. A less plausible explanation is that shared Y-chromosome
haplotypes indicate adult male dispersal. However, given that eastern chimpanzee males
display a high degree of territoriality and intercommunity aggression (Mitani et al., 2010;
Wilson et al., 2014), this explanation seems unlikely, even in a degraded habitat. One
additional possibility is that putative communities sharing a single haplotype are actually
a single community. However, this explanation also seems unlikely given the high
average recapture rate in this study, which often led to individuals being sampled among
different party associations, as well as the large distances between some sampling
clusters sharing a haplotype. For example, if one considers the maximum distance
between sampling points for two males sharing the same haplotype (~34 km), and
conservatively assumes those points demarcate the outer edges of a single community
home range, their circular home range would measure more than 900 km
2
in size. The
sharing of Y-chromosome haplotypes among multiple chimpanzee communities has also
been seen elsewhere (Schubert et al., 2011; Moore and Vigilant, 2014; Langergraber et
68

al., 2014b). Future studies may better clarify the distribution of male philopatric
chimpanzee communities across this region. Nonetheless, these results indicate likely
conservatism in male philopatric territorial community structure despite substantial
habitat degradation, a pattern that appears to hold for chimpanzees across numerous
habitat types (Moore et al., 2015). These findings support the behavioral data collected
for chimpanzee communities in the region such as Bulindi, where fission-fusion
community structure within defined territories appears intact despite widespread
anthropogenic habitat destruction (McLennan, 2008).
3.4.4. Conservation implications
The results of this study suggest chimpanzees are both numerous and widespread
in the human-dominated landscape between the Budongo and Bugoma Forests. This is
perhaps surprising, given the paucity of forest habitat and the high human population
density of 157 residents per km
2
in this region (UBOS, 2014). However, chimpanzees in
this area are known to utilize home ranges encompassing numerous forest fragments
while feeding on a combination of natural and cultivated food resources (Reynolds, 2005;
McLennan, 2013). These forest fragments, which are largely riparian, are additionally
known to harbor relatively high fruit tree density (McLennan and Plumptre, 2012).
Indeed, riparian forest fragments in Central Africa have been noted for having high
conservation value for chimpanzees and other species (Gautier-Hion and Brugiere, 2005;
Fleury-Brugiere and Brugiere, 2010). In addition, chimpanzee survival under
anthropogenic pressure is likely aided by their behavioral flexibility (Junker et al., 2012;
Hockings et al., 2015). Though their behavioral strategies in such habitats remain little
69

understood, they include incorporating new (often human-cultivated) foods into their
diets and adopting more aggressive or cryptic behaviors to mitigate human threats
(McLennan and Hill, 2010; Hockings et al., 2012; Hicks et al., 2013; McLennan, 2013;
Tagg et al., 2013). In western Uganda, their persistence is also attributable to relatively
low hunting pressure, since Ugandans traditionally have not hunted chimpanzees for meat
as in some other countries. However, customs are changing and chimpanzees are
sometimes hunted for meat or killed as pests in Uganda, thereby making anthropogenic
activities a threat to chimpanzee survival there (McLennan, 2008; McLennan et al.,
2012).
Despite anthropogenic pressures, these findings underscore the importance of
greater investment in chimpanzee conservation in this region. A targeted solution such as
translocating individual chimpanzee communities, as has been discussed (Reynolds et al.,
2003; Reynolds, 2005; McLennan and Hill, 2012), appears impractical given the large
and broadly distributed population documented in our study. In contrast, these results
suggest the potential may be high for a corridor enhancement project to benefit
chimpanzees in this region (Nangendo et al., 2010), given that an increase in functional
connectivity to the chimpanzee populations in the Budongo and Bugoma Forests would
collectively impact 30% of Uganda’s total chimpanzee population (including the
chimpanzee populations of both forests and the region between them). Any such project
must be considered carefully, however. The need for firewood, building materials, and
agricultural land are often cited as reasons for deforestation of the region’s unprotected
forests (Akwetaireho et al., 2011). Humans and chimpanzees also have a history of
70

conflict interactions, given their close co-residence in this area (Reynolds, 2005;
McLennan and Hill, 2012). Stakeholder needs such as these must be taken into
consideration to ensure the effectiveness of any conservation initiative. However, riparian
forests play a key role in protecting rivers and the agricultural needs they support, so their
conservation may increasingly be recognized as vital to the futures of both humans and
other species locally. Additionally, habitat corridors may protect wildlife against the
detrimental effects of climate change, thereby enhancing their value even further (Smith,
1997).
These findings point to the value of conservation planning for unprotected areas
with great potential to enhance gene flow and population viability among endangered
wildlife populations. In this region as with many others like it, however, conservation
action is urgently required. An estimated 134 km
2
of forest is estimated to have been lost
between the Budongo and Bugoma Forests from 1985 to 2014 (Twongyirwe et al., 2015).
Given the human population growth rate, this trend is likely only to change if concerted
efforts are made to slow the rate of deforestation in the region. Though chimpanzees have
proven surprisingly resilient to date in this habitat, their ability to withstand continued
habitat losses, along with other threats to their survival, is highly uncertain.
3.5. Conclusions
Using genetic censusing, I found a surprisingly large population of chimpanzees
inhabiting largely unprotected forest fragments in western Uganda. These findings
confirm the presence of over 200 hundred chimpanzees, distributed in at least 9
communities across a human-dominated landscape. The large size and widespread
71

distribution of this population suggests it serves as a vital link between larger populations
in the neighboring Budongo and Bugoma Forests. These results demonstrate the potential
for forest fragments to serve as wildlife corridors, and for animal populations to be
widely distributed in degraded habitats. Despite this potential, however, the habitat is
rapidly being altered, and its capacity to support chimpanzees and other species may not
persist unless the rate of habitat change is slowed considerably.

72

Chapter 4
Genetic evidence for female-biased chimpanzee transfer in a human-
dominated habitat

4.1. Introduction
Dispersal, the movement from one reproductive unit to another, is a key
mechanism for inbreeding avoidance in numerous taxa (Pusey, 1987; Clutton-Brock,
1989). Despite its biological significance and widespread occurrence, however, dispersal
is not a rigid, invariable behavioral pattern. Instead, dispersal may be expected to occur
under conditions associated with enhanced reproductive success for the disperser.
Environmental conditions such as season, resource availability, and habitat quality, and
73

demographic factors such as population density can affect the reproductive outcomes
associated with dispersal and therefore its likelihood of occurrence (Aars and Ims, 2000;
Cooper and Walters, 2002; Baglione et al., 2006; Bowler and Benton, 2009; Ims and
Hjermann, 2012). These conditions, in turn, influence physiological states including
reproductive status, energy reserves, and overall health, which also play a role in the
timing and likelihood of dispersal (Pusey, 1980; Lens and Dhondt, 1994; Nunes et al.,
1997). For dispersal to occur, its expected benefits should outweigh the costs associated
with leaving the home area, traveling, and settling in a new home area (Chaine and
Clobert, 2012).
Habitat fragmentation can alter the costs and benefits associated with dispersal in
a number of taxa including birds (Cooper and Walters, 2002; Baglione et al., 2006),
invertebrates (Baguette et al., 2003; Keller and Largiader, 2003) and mammals (Gerlach
and Musolf, 2000; Ripperger et al., 2012; Fietz et al., 2014; Li et al., 2014; McManus et
al., 2014). Limitations in dispersal ability can lead to inbreeding depression and therefore
limit population viability and heighten extinction risk (Dudash and Fenster, 2000).
Species with small home ranges and niche specializations are often at greater risk of local
extinction due to habitat fragmentation (Debinski and Holt, 2000). In contrast, large-
bodied mammals may be buffered against the effects of habitat fragmentation because
they are more mobile over large ranges, especially if they are dietary generalists
(Debinski and Holt, 2000; Flagstad et al., 2012). However, they also tend to be
conspicuous, occur at low densities, and have long interbirth intervals, all of which make
74

them more susceptible to population declines in fragmented habitats (Charnov and
Berrigan, 1993; Debinski and Holt, 2000; Kosydar et al., 2014).
Anthropogenic habitat loss and fragmentation, along with hunting and disease, are
key threats to great ape survival (Walsh et al., 2003; Junker et al., 2012; Tranquilli et al.,
2014). In the past two decades, there has been a marked decline in remaining suitable
habitat for African great apes (Junker et al., 2012). Up to 81% of chimpanzees (Pan
troglodytes) in West Africa live outside protected areas, often in fragmented and
degraded forests (Kormos et al., 2003). Similarly, an estimated 78% of Bornean
orangutans (Pongo pygmaeus) live outside protected areas, with large-scale commercial
agriculture such as oil palm production posing a major threat to orangutan survival
(Meijaard et al., 2010; Wich et al., 2012). A number of recent studies have demonstrated
that great apes alter their diets (Hockings and McLennan, 2012), activity budgets
(Hockings et al., 2012; Krief et al., 2014), social behavior (Hockings et al., 2012),
vocalizations (Hicks et al., 2013), and nesting patterns (Hicks, 2010; Meijaard et al.,
2010; Last and Muh, 2013; Tagg et al., 2013; Ancrenaz et al., 2014) under anthropogenic
pressure. Data are currently lacking regarding the effects of habitat fragmentation and
loss on great ape dispersal patterns, however.
Chimpanzees studied at long-term research sites display a territorial fission-fusion
social structure characterized by male philopatry and female dispersal, with a multi-male,
multi-female polygynous mating system (Goodall, 1986). Females typically disperse
from the natal community upon reaching sexual maturity, though there is considerable
variability in the likelihood of female dispersal. Reported transfer rates vary between half
75

and nearly all females in most studied communities (Nishida, 1990; Boesch and Boesch-
Achermann, 2000; Reynolds, 2005; Wroblewski et al., 2015). However, at Bossou,
Guinea, where chimpanzee community isolation is high, female transfer has not been
observed (Sugiyama, 1999). Dispersal decisions can impact a female’s reproductive
success (Williams et al., 2002), so transfer should be expected under environmental and
social conditions that favor doing so to enhance this success. Factors affecting female
dispersal timing and likelihood include resource availability, social rank, within-group
competition, and affiliative relationships (Williams et al., 2002; Stumpf et al., 2009;
Wroblewski et al., 2015). In fragmented and human-disturbed landscapes, the costs of
dispersal may be higher due to greater community isolation and more aggression toward
immigrant females if resources are limited. Simultaneously, potential benefits such as the
availability of unrelated mates may be decreased if communities are small and isolated in
fragmented landscapes. Few studies have considered ecological limitations on
chimpanzee dispersal, however.
In western Uganda, chimpanzees inhabit a fragmented forest corridor between the
Budongo and Bugoma Forests, presenting an opportunity to investigate dispersal patterns
in a fragmented landscape. The Budongo and Bugoma Forests are inhabited by ~600
chimpanzees each (Plumptre et al., 2003), and genetic evidence suggests that the
approximately 1,200-km
2
region between these forests is inhabited by an estimated 260
to 320 chimpanzees (see Chapter 3 and McCarthy et al., 2015). The substantial size and
widespread distribution of this chimpanzee population indicate that this region may serve
as a functional corridor for chimpanzees by aiding gene flow between the Budongo and
76

Bugoma Forest populations. However, habitat degradation could lead to the fissioning of
communities into smaller and more isolated social groups that subsequently lack the
signatures of male philopatry, female dispersal, and territoriality as observed in
chimpanzee communities in less disturbed environments.
As described in Chapter 3 and McCarthy et al. (2015), I used the clustering of
repeatedly sampled genotypes to identify a minimum of nine putative chimpanzee
communities. Community affiliations among male chimpanzees corresponded largely to
the distribution of Y-chromosome haplotypes, which can be used as indicators of
chimpanzee community affiliations (Langergraber et al., 2007b; Arandjelovic et al.,
2011; Moore et al., 2015). Together, these findings suggest that male philopatric
territorial community structure likely remains intact despite habitat fragmentation in this
region. By examining dispersal patterns among these chimpanzees, we can better
understand the effects of habitat fragmentation on chimpanzee social structure and
whether this region functions as a corridor for gene flow among intact communities.
The purpose of this study was to examine evidence of chimpanzee dispersal in a
human-dominated landscape mosaic. I did so by analyzing parent-offspring relationships
and the community affiliations attributed to each individual, as inferred by the spatial
clustering of co-sampled genotypes. To do so, I first used a chimpanzee population with
known parent-offspring relationships to estimate error rates in parentage assignment and
ensure a low rate of false positive assignments. I conducted parentage analyses using both
CERVUS and KinGroup software and compared the rates of assignment error in each. I
then applied parameters that yielded low error rates in parent-offspring assignments to the
77

study population in this fragmented forest landscape in Uganda, using both analysis
approaches to obtain robust results. In particular, I examined whether parent-offspring
dyads were sometimes attributed to different communities. I expected that if female
dispersal remains common in the fragmented habitat, then members of mother-daughter
dyads will be more frequently found in two different communities, while father-son
dyads will be found in the same community. I also examined whether mother-father-
offspring trios were attributed to the same community or whether offspring were
sometimes attributed to different communities, further indicating a pattern of chimpanzee
dispersal. I repeated these analyses with a data set that included genotypes from both the
corridor habitat and the protected Budongo Forest Reserve to examine evidence for
dispersal between them, which would confirm the functional connectivity of the corridor
habitat to nearby continuous forest. Finally, I examined the spatial locations of genotypes
from repeatedly sampled chimpanzees to look for evidence of dispersal events occurring
during the study period.
4.2. Methods
4.2.1. Study area and subjects
As described in detail in Chapter 2, I collected data in Hoima and Masindi
Districts, Uganda in the corridor region between the Budongo and Bugoma Forests (1°37'
– 1°68'N and 31°1' - 31°6'E; Figure 2.1) from October through December 2011 and
October 2012 through September 2013. Chapter 2 summarizes previously published
information regarding the chimpanzee population inhabiting this region. Chapter 3
78

further describes the size and distribution of this population during the course of this
study period.
4.2.2. Noninvasive fecal sample collection methods
I collected fecal samples noninvasively throughout the study area, with particular
focus on searching riparian forest fragments for evidence of chimpanzees. I also acquired
information on chimpanzee presence in McLennan (2008) and via informal discussion
with local inhabitants. As described in Chapter 2, strictly systematic survey methods were
neither feasible nor ethically appropriate in this habitat. Instead, I centered search effort
in forest fragments around village boundaries, which typically encompass settlements,
farmland, and privately owned forests. In accordance with local customs, prior to
searching a forest fragment I first gained permission from the chairperson of the village
in which the forest fragment was located, and from individuals who identified themselves
as landowners of the forest fragment. I used satellite imagery to identify the forest
fragments located within the boundaries of a given village, and visited accessible and
permitted forest fragments within the boundaries of that village. I divided the study area
into a grid of 1 sq. km cells and recorded when any part of each cell was searched (Figure
2.6).
I collected samples under nests and opportunistically along chimpanzee trails and
at feeding sites. For each sample, I recorded a GPS waypoint with a Garmin GPSMap®
60CSx. I recorded samples with unique identification numbers corresponding to GPS
waypoints, and with party association data when applicable. I recorded samples as
belonging to a party when two or more same-age samples were collected within 30 m of
79

each other. Distances were determined using GPS data and, when necessary, a Bresser®
laser Range Finder 800 to ensure accuracy. I avoided collecting two samples under the
same nest or in close proximity on trails, due to the likelihood of collecting redundant
samples from the same individual and the possibility that closely deposited samples may
have cross-contaminated each other. I collected samples and stored them according to the
two-step ethanol-silica method described in Nsubuga et al. (2004) and detailed in Chapter
2.
Chimpanzee fecal samples were typically easy to identify because of 1) their
locations under chimpanzee nests and along trails, 2) their characteristic size, shape, and
odor, and 3) the absence of other sympatric large-bodied nonhuman primates. Although
olive baboons (Papio anubis) produce dungs that can superficially resemble those of
chimpanzees (pers. obs.), they have been eradicated from many parts of the study area.
For any sample of uncertain species origin, however, I confirmed the species prior to
sample storage and analysis by macroscopic analysis as described in Chapter 2.
To achieve adequate resampling, I estimated a target sample size for each region
with chimpanzee presence across the study area. These target sample sizes were
determined as described in detail in Chapter 2, with the goal of collecting at least three
times the number of samples as expected individuals in order to obtain a narrow
confidence interval for population size estimates using mark-recapture methods (Miller et
al., 2005; Petit and Valiere, 2006; Arandjelovic et al., 2010). Because additional
information on chimpanzee presence was gained over the course of the study period,
these target sample sizes were adjusted as necessary. To help achieve this sampling goal
80

and to ensure adequate resampling across fission-fusion chimpanzee communities, I
attempted to search forests a minimum of once every three months, except where local
research permissions were granted only for a limited time period.
I carried out data collection with the permission of the Uganda National Council
for Science and Technology, the Uganda Wildlife Authority, and the National Forestry
Authority of Uganda. Additional permissions were granted by local landowners where
applicable, as described above. Because fecal sample collection was entirely noninvasive
and required no contact with the chimpanzees, ethical consent was not necessary for this
project.
4.2.3. DNA extraction and amplification
Samples were preserved at room temperature in silica (as described in detail in
Ch. 2) in the field for up to 6 months prior to arrival at Max Planck Institute for
Evolutionary Anthropology, Leipzig, Germany, where they were then stored at 4 °C prior
to extraction. DNA was extracted using either the GeneMATRIX Stool DNA Purification
Kit (Roboklon) according to manufacturer’s instructions or the QIAmp Stool kit
(QIAGEN) with minor procedural adjustments (Nsubuga et al., 2004).
I used autosomal microsatellite loci to determine individual chimpanzee
genotypes. Each DNA extract was first evaluated by simultaneously amplifying three
autosomal microsatellite loci, along with an X-Y homologous segment of the amelogenin
gene, used for sex determination (Bradley et al., 2001), in a one-step multiplex PCR, as
described in detail in Chapter 2 and McCarthy et al. (2015). DNA extracts that reliably
amplified at a minimum of 3 of the 4 loci in at least 3 independent amplifications were
81

then genotyped in triplicate at an additional 11 autosomal microsatellite loci (McCarthy
et al., 2015). Extracts that failed to meet these criteria were not amplified further. The
additional 11 loci were amplified in a two-step multiplex PCR procedure as described in
detail in Arandjelovic et al. (2009).
At each locus, heterozygous genotypes were confirmed by observation in at least
two independent reactions (Morin et al., 2001; Arandjelovic et al., 2009). Homozygous
genotypes were confirmed when observed in a minimum of three independent reactions.
Individual loci that failed to meet these criteria were instead coded with asterisks and
were excluded from analyses. To further ensure that apparent homozygotes were not the
result of allelic dropout, I calculated allelic dropout rates by locus after recording all
alleles and confirmed that a maximum of two replicates was required at any locus to
confirm homozygosity with 99% certainty (Morin et al., 2001; Broquet and Petit, 2004).
Table 2.1 further details the allelic dropout rates by locus. Thus, I exceeded this threshold
and ensured minimal allelic dropout by confirming homozygotes only when alleles were
observed consistently in three reactions.
4.2.4. Discriminating chimpanzee genotypes and putative communities
I distinguished individual chimpanzee genotypes using an identity analysis in
CERVUS 3.0.7 software (Kalinowski et al., 2007). Using the allele frequencies of the
study population, I determined the minimum number of loci necessary to achieve a P
IDsib

<0.001, which would allow sufficient power to distinguish among genotypes and
determine with statistical confidence that two matching genotypes from different samples
originate from the same chimpanzee rather than from full siblings. Matching genotypes
82

were assigned a consensus name and composite genotype data. I matched genotypes
using a minimum of nine matching loci with no mismatches. Up to four mismatches were
permitted to flag potential matches despite genotyping errors. Any mismatch was
therefore either resolved as a true match with corrected errors or as a true mismatch
comprising distinct genotypes. For rare instances in which genotypes matched with P
IDsib

>0.001, I eliminated the less complete of the two genotypes from further analysis. I
included extracts typed at as few as four loci if their genotype was unique and did not
match any other genotype.
I defined putative chimpanzee communities according to the spatial clustering of
co-sampled genotypes. In other words, genotypes found in association with other
genotypes, e.g., from samples collected under nests comprising a single same-age nest
group, were assumed to belong to members of the same community. Using spatial data
from these genotype clusters, I constructed 100% MCPs using the Minimum Convex
Polygon plugin for QGIS version 2.4.0 software (Quantum GIS, 2014) to represent to the
minimum home ranges of communities based on genotypes found in association.
Additional genotypes found within these polygons were also assumed to originate from
members of the same community, since spatial overlap among communities generally is
not expected (Nishida, 1979; Goodall, 1986; Herbinger et al., 2001).
4.2.5. Parentage analysis
4.2.5.1. False positive error rate estimation in parentage assignments
Microsatellite genotyping has been used to examine parent-offspring relationships
in numerous taxa, and comparisons among genotypes can help determine dispersal
83

patterns by revealing whether individuals reside in different social groups than their
parents (Arandjelovic et al., 2014). Genotyping can be used to study parentage in two
primary ways. The first approach relies on knowing the identity of one parent to
genetically determine the other. For example, when mother-offspring relationships are
known, the alleles shared between them can be identified and candidate fathers can be
tested against the offspring’s second alleles at each locus to identify a matching father
using an exclusion method or a likelihood-based approach (Whittingham, 2004; Lyke et
al., 2013; Vigilant et al., 2015). Second, likelihood-based methods can be used to assign
parentage with statistical confidence in the absence of data on mother-offspring
relationships (Marshall et al., 1998). In this study, I used the second approach, since
mother-offspring relationships were not known among the unhabituated chimpanzees in
the corridor habitat.
Despite the informative potential of this analysis approach, assignment errors are
possible, particularly if assigning one parent in the absence of a genotype from the
second parent (Marshall et al., 1998; Blouin, 2003; Van Horn et al., 2008). Erroneous
parentage assignments may include either false positive assignments (Type I Error), in
which individuals are erroneously identified as a parent-offspring pair, or false negative
assignments (Type II Error), in which individuals comprising a true parent-offspring pair
fail to be classified as such (Marshall et al., 1998). Although either error should be
avoided, false positive assignments are especially problematic when studying emigration,
as they can lead to erroneous conclusions regarding the occurrence of dispersal events.
To reduce this potential for false positive assignments, I used genetic data to first
84

approximate the false positive assignment rate based on 217 genotypes from the Ngogo
and Kanyawara chimpanzee communities in Kibale National Park, Uganda, a study site
approximately 130 km away from the Budongo-Bugoma corridor habitat and for which
data are available regarding demography and parent-offspring relationships. Kibale
genotypes were derived from noninvasively collected fecal samples and were published
previously (Langergraber et al., 2012). I conducted parentage analyses using KinGroup
v2 (Konovalov et al., 2004) and CERVUS 3.0.7 software (Kalinowski et al., 2007). Both
CERVUS and KinGroup employ likelihood-based methods to identify parent-offspring
dyads and assign a confidence level (p-value) associated with the assignment’s likelihood
(Konovalov et al., 2004; Kalinowski et al., 2007). By using two approaches I could 1)
compare the false positive error rates obtained in each using the Kibale data and 2) find
greater support for the results obtained in the corridor data set. I used the same analysis
approaches planned for the Budongo-Bugoma corridor data when conducting false
positive assessments with the Kibale data, including using all individuals as putative
parents and offspring and ignoring data regarding the age class and parentage status of
each individual, since this information was not available for the Budongo-Bugoma
corridor population. I then used the frequency of false positive assignments to choose
parentage analysis parameters that would lead to low error rates and ensure accuracy in
assignments among the corridor chimpanzees.
In KinGroup, I used likelihood tests with a primary hypothesis of parent-offspring
kinship and a null hypothesis of unrelated individuals (Goodnight and Queller, 1999) and
used 1,000,000 simulated pairs to perform significance tests. For CERVUS parent-
85

offspring analyses, I first conducted simulation of parentage analyses by simulating
10,000 offspring. Because this simulation relies on an estimate of the total number of
parents—both present as well as dead and missing—in the population, I estimated 300
total candidate parents in the two Kibale communities, a 20% increase over the total
estimated size of these communities, resulting in a proportion of candidate parents
sampled of 0.7233. I set the minimum number of loci typed as 10, the proportion of loci
typed as 0.9888, and I assumed a proportion of mistyped loci of 0.01. For mother-father-
offspring trios, I assumed 162 candidate mothers and 135 candidate fathers in the
population, which reflected the sex proportion for genotyped individuals in the Kibale
data set and produced a good fit between proportions of expected and observed
assignments.
I checked the resulting parentage assignments against parentage data available for
this population, which were derived from pedigree data and genetic kinship analyses
using a more comprehensive data set from 44 autosomal microsatellite loci, X-linked
microsatellite loci, Y-linked microsatellite loci, and mtDNA (Langergraber et al., 2007a).
False positive error rates reflect the percentages of parentage assignments that contain an
error. These were calculated as a range of values, where the lower limit of the range
indicates known false positive assignments and the upper limit indicates the maximum
possible false positive assignments, since not all parentage assignments were known and
I could not determine whether some putative assignments were correct.
I calculated the percentage of false positive assignments when using 80, 95, and
99% confidence levels to examine the impact of confidence levels on error rates. I also
86

compared the percentage of false positive assignments when using the same 14 autosomal
microsatellite loci typed in the current data set to that when using the 19 available
microsatellite loci in the multiplex PCR (detailed in Chapter 2) to determine whether
genotyping individuals at 5 additional loci would result in lower false positive error rates
in parentage assignments. I also examined whether false positive assignments were
reduced by strictly excluding dyads and trios that contained 1) mismatching loci, 2) a
second-best match for one parent, given the genotype of the other parent, and 3) sons in
parent-offspring trios, given that closely related males may be associated with more false
parentage assignments and are not of interest, since only daughters are expected to
disperse.
4.2.5.2. Parentage assignments in the corridor chimpanzees
After using the Kibale genotypes to determine which parameters would result in
the lowest false positive error assignments, I conducted parentage analyses with the
Budongo-Bugoma corridor chimpanzee genotypes using these criteria. I included 176
genotypes from chimpanzees in the Budongo-Bugoma corridor, a selection restricted to
genotypes from chimpanzees in fragmented forest habitat, and only those individuals
genotyped at a minimum of 10 loci. I included all individuals as potential parents and as
potential offspring, since age data were not available for most individuals and therefore I
could not determine which individuals should be included as candidate parents and
offspring, respectively.
For KinGroup parentage analyses, I used likelihood tests with a primary
hypothesis of parent-offspring kinship and a null hypothesis of unrelated individuals
87

(Goodnight and Queller, 1999) and used 1,000,000 simulated pairs to perform
significance tests. For CERVUS simulation of parentage analyses, I simulated 10,000
offspring and tested whether varying the number of candidate parents would improve the
fits of observed and expected parentage assignment proportions. To do so, I relied on the
following population size estimates, as described in Chapter 3 and in McCarthy et al.
(2015): 1) 246, the lower limit of the total confidence interval for the population size,
2) 256, the point estimate using the capwire TIRM model, 3) 319, the point estimate
using the SECR model, and 4) 357, the upper limit of the total confidence interval for the
population size. However, the number of candidate parents should exceed those
accounted for in the population size estimates, since the former should account for
parents of all individuals currently in the population, including missing and dead parents
that are excluded from population estimates. Therefore, I added 40% and 50% more
individuals to these population size estimates to approximate the total number of
candidate parents, and tested which estimate better improved the fit between observed
and expected assignment proportions. These percentage increases in the number of
candidate parents were greater than the 20% increase used in the Kibale data set due to
the fact that a lower proportion of the overall population was likely sampled in the
corridor data set, thereby requiring a greater compensation for missing individuals. The
minimum number of loci typed was 10, the proportion of loci typed was 0.9724, and I
assumed a proportion of loci mistyped as 0.01. Despite varying these parameters, the
number of assignments at each confidence level was robust, with only minor variations in
88

confidence level and associated LOD scores. Nonetheless, the final results included were
those with the most conservative confidence assignments and LOD scores.
I assessed parent-offspring kinship and associated community affiliations using
several measures. First, I examined occurrences in which individuals comprising dyads
and trios were attributed to two or more distinct communities, as determined based on the
locations of genotyped samples for each individual (see Section 4.2.4 above). I used these
occurrences, hereafter termed ‘mixed community dyads’ or ‘mixed community
affiliation,’ as indicators of likely dispersal events, since individuals are generally not
expected to be found in association, either spatially or socially, with a community other
than their own. I analyzed father-son dyads and mother-daughter dyads separately to
determine whether mother-daughter dyads were more likely than father-son dyads to
contain mixed community affiliation, thereby indicating a dispersal pattern characteristic
of chimpanzees in continuous habitats. I also assessed parent-offspring mixed sex dyads
to determine the proportion with mixed community affiliation. Because I lacked age data
to correspond to each individual genotype, however, I could not ascertain which
individual in each dyad was the parent and which was the offspring, and therefore could
not determine the likely directionality associated with dispersal events. To overcome this
limitation in directionality data, I also analyzed the genotype data for mother-father-
offspring trios, which provide the highest likelihood mother and father for each potential
offspring. I examined high likelihood trios for instances of mixed community affiliation,
which would indicate both dispersal and its likely directionality, given that daughters are
expected to disperse from the natal community in which their parents reside.
89

To test whether the observed sex differences in the proportions of mixed
community affiliation among parent-offspring dyads were statistically significant, I
conducted a permutation test programmed in R. I randomized the assignment of
individuals to sex and restricted the permutations such that both members of any given
parent-offspring pair had identical sexes throughout all permutations. I conducted 1,000
permutations of sex assignment into which I included the original data as one
permutation. As a test statistic I used the chi-square value obtained from the cross-
tabulation of sex of the parent-offspring dyad with whether they co-resided in the same
group or not.
For parent-offspring dyads and trios with mixed community affiliation, I
measured the distance between the sampling locations of each individual to estimate an
approximate mean distance traveled by dispersing individuals. I did so by measuring the
distance between waypoints using the measuring tool in ESRI® ArcMap™ 9.2
(Redlands, CA). For chimpanzees sampled more than once, I calculated a geographical
mean location of the UTM coordinates of all sampling locations for that individual. I also
calculated geographical mean locations representing a central area for each putative
chimpanzee community based on the MCPs for co-sampled genotypes (McCarthy et al.,
2015).
4.2.5.3. Parentage assignments including Busingiro
I conducted a second parentage analysis that included genotypes from the
protected Budongo Forest Reserve to look for evidence of transfer events between the
corridor habitat and the continuous forest. Dispersal events between these habitats would
90

demonstrate the potential of the corridor habitat for gene flow among these larger
regional chimpanzee populations. This analysis employed a second data set which
included the same 176 corridor genotypes used in the first analysis as well as 53
genotypes from Busingiro, a chimpanzee community in the southern area of the Budongo
Forest (Figure 2.1). Of these 53 additional genotypes, 14 originated from samples
collected during the current study from Siiba Forest Reserve (see Chapter 3 for further
details) while 39 were collected and genotyped previously (Langergraber et al., 2011). I
ran this analysis separately from that of the corridor genotypes alone, given the
possibility that factors influencing dispersal likelihood in continuous forest habitat differ
from those for the fragmented forest habitat and that lumping all genotypes together may
have obscured these differences.
As with the corridor data set, I included all individuals as candidate parents and
offspring given the lack of age data for each genotype. I conducted KinGroup analyses
exactly as described above for the corridor data set. I also conducted CERVUS
simulation of parentage analyses as described above, with the exception of adding 100
candidate parents, given an approximate estimate of the Busingiro community’s size
(n ≈ 70; Reynolds, 2005) and an estimate of 40% more unsampled candidate parents. As
above, I used a minimum of 10 loci typed and assumed a proportion of mistyped loci of
0.01. The proportion of loci typed was 0.9663 for this data set. I also conducted a
permutation test with the parent-offspring dyads produced from this second data set, and
used the same methods described above.

91

4.2.5.4. Spatial analysis of repeatedly genotyped chimpanzees
Finally, I compared the spatial locations of samples collected from the same
chimpanzee across multiple data collection days over the study period to determine
whether all samples from each individual fell within 1) a single MCP and were therefore
attributed to a single community or 2) multiple MCPs, indicating a likely dispersal event
during the study period. Because chimpanzees are territorial and typically associate with
members of just one community, they are not generally expected to be found within the
MCP associated with another community unless due to a dispersal event (Nishida, 1979;
Goodall, 1986; Herbinger et al., 2001).
4.3. Results
4.3.1. Distinct chimpanzee genotypes and community affiliations
I collected a total of 865 fecal samples over 633 km
2
from October to December
2011 and October 2012 to September 2013 (Figure 3.1). Of these, 662 (76%) amplified
reliably at a minimum of three of four test loci and were thus genotyped at an additional
11 loci. Based on the allele frequencies, I calculated that comparison at a minimum of
nine loci was necessary to obtain a P
IDsib
< 0.001 and thus confidently determine that
identical genotypes originated from the same individual rather than two different
individuals, including for example full siblings. Appendix D provides allele frequency
data associated with this population. Of the 662 genotypes, 459 matched exactly to one or
more other genotypes and were merged to create consensus genotypes. An additional five
genotypes were removed from analysis because they matched other genotypes with a
P
IDsib
> 0.001. Chapter 3 provides further details regarding how I discriminated among
92

these genotypes. The final genotype list consisted of 196 distinct genotypes, of which 187
were complete at a minimum of 9 loci and could therefore be included in subsequent
parentage analyses.
By grouping genotypes from samples found together I identified 10 spatial
clusters that were geographically distinct from one another, thus suggesting the presence
of at least nine potential communities in the study area, along with one additional cluster,
Kiraira (Figure 4.1).

Figure 4.1. MCPs for genotyped samples found in association. Each MCP is labeled with
the name I assigned to the chimpanzee community, based on the name of a nearby
village. MCPs for all communities include a lightly shaded one-km buffer to indicate
likely minimum home range sizes extending beyond sample collection locations. Gray
indicates forest cover during the study period and was provided by Hansen et al. (2013).
93

Chapter 3 provides further details on the distribution and sizes of these distinct
chimpanzee communities.
4.3.2. False positive error rate estimation in parentage assignments
Characteristics of the genotype data for the Kibale and corridor data sets were
similar, including similar heterozygosity, a similar number of alleles per locus, and
similar completeness of genotypes. Details of these allele frequencies are provided in
Table 4.1. False positive error rates found using the Kibale data set are detailed in Tables
4.2 – 4.4 and are expected to approximate those for the corridor data set. I obtained the
lowest error rates by accepting parentage assignments with a confidence level above 95%
and only accepting parent-offspring dyads and trios with zero mismatching loci and no
higher likelihood parent matches identified.

Table 4.1. Allele frequency data comparison for corridor and Kibale genotypes.
Allele frequency measure Corridor Kibale
Number of individuals 176 217
Number of loci 14 14
Mean number of alleles per locus 8.2140 8.7860
Mean proportion of loci typed 0.9724 0.9888
Mean expected heterozygosity 0.7407 0.7789
Mean polymorphic information content (PIC) 0.7045 0.7459
Combined non-exclusion probability (sibling identity) 0.0000031 0.0000014

94

Table 4.2. False positive error rates for parent-offspring assignments using
KinGroup.

p-value False positive range (%)
Female dyads Male dyads
< 1 3.8 - 5.8 7.9 - 17.5
< 0.05 3.8 - 5.8 7.9 - 17.5
< 0.01 3.8 - 5.8 7.9 - 17.5
< 0.001 3.8 - 5.8 7.9 - 17.5
< 0.0001 2.2 - 4.3 7.1 - 11.9

Note. Data are based on 217 genotypes from Kibale National Park, Uganda. The range of
values at each confidence level indicates the lower and upper limit of possible false
positive assignments, given that not all parentage assignments were known and I could
not determine whether some putative assignments were correct.

Table 4.3. False positive error rates for parent-offspring assignments using
CERVUS.

Confidence False positive range (%)
Female dyads Male dyads Mixed sex dyads
>99% 0.0 - 3.1 7.7 - 11.5 0.0 - 1.9
>95% 2.3 - 4.7 6.9 - 10.3 0.0 - 4.6
>80% 2.3 - 4.7 6.5 - 16.1 0.0 - 4.5

Note. Data are based on 217 genotypes from Kibale National Park, Uganda. The range of
values at each confidence level indicates the lower and upper limit of possible false
positive assignments, given that not all parentage assignments were known and I could
not determine whether some putative assignments were correct.

95

Table 4.4. False positive error rates for mother-father-offspring trio assignments
using CERVUS.

Confidence False positive range (%)
All trios Strict exclusion
>99% 15.2 - 23.8 0.0 - 3.3
>95% 24.8 - 33.1 0.0 - 3.3
>80% 32.0 - 44.9

Note. Data are based on 217 genotypes from Kibale National Park, Uganda. The range of
values at each confidence level indicates the lower and upper limit of possible false
positive assignments, given that not all parentage assignments were known and I could
not determine whether some putative assignments were correct.

Using the known relationships of the chimpanzees of the Kibale Forest, I found
that about 5% of the mother-daughter dyads were falsely identified in blind parentage
analyses (3.8 – 5.8% for KinGroup mother-daughter dyads with p < 0.05 and 2.3 – 4.7%
for CERVUS dyads with >95% confidence and no mismatching loci). Father-son dyads
were associated with higher error rates than mother-daughter dyads (7.9 – 17.5% in
KinGroup with p < 0.05 and 6.9 – 10.3% in CERVUS with >95% confidence and no
mismatching loci). False positive error rates for mixed sex dyads were 0 – 4.6% with
>95% confidence and no mismatching loci. Among familial trios, those with ‘strict
exclusion’ applied—that is, 1) exclusively female offspring, 2) no mismatching loci
between either parent and the offspring, and 3) no higher probability matches identified
for either parent in the absence of the other—resulted in lower false positive error rates,
with a maximum 3.3% false positive error. The higher error rate among father-son dyads
and trios with male offspring is likely due to the presence of more closely related males
than females in male philopatric chimpanzee communities such as those in Kibale
96

National Park. This error rate among father-son dyads does not pose a problem, however,
since dispersal events are detected through mother-daughter dyads, and false positive
assignments among father-son dyads have little impact on the results. The inclusion of
data from five additional microsatellite loci, thereby bringing the total number of loci to
19, generally did not substantially decrease the error rate in parentage assignments (data
not shown).
4.3.3. Parentage assignments in the corridor chimpanzees
The CERVUS and KinGroup father-son parentage analyses with >95%
confidence yielded no father-son mixed community dyads (n = 12 CERVUS dyads; n =
29 KinGroup dyads). In contrast, the CERVUS parentage analysis yielded 5 of 25 (20%)
mother-daughter mixed community dyads. The KinGroup parentage analysis produced a
similar pattern, but identified additional mother-daughter dyads with high confidence (p <
0.05), with 21 of 54 (39%) dyads containing mixed community affiliation. The observed
proportions of mixed community dyads among father-son and mother-daughter dyads
differed significantly (χ
2
= 13.11, p
perm
= 0.001, Figure 4.2). I additionally found that 2 of
30 (7%) mixed sex parent-offspring CERVUS dyads contained mixed community
affiliation, while 15 of 81 mixed sex dyads (19%) in KinGroup contained mixed
community affiliation. Because the available data do not allow the discernment of the
ages of sampled individuals, I cannot distinguish whether these dyads comprise father-
daughter or mother-son pairs, however. The mean distance between mean sampling
locations for mother-daughter dyads with mixed community affiliation was 12.8 km
(standard deviation [SD] = 5.0 km).
97

Among trios with ‘strict exclusion’ and >95% confidence, 2 of 11 contained
mixed community affiliation. In one of these trios, the daughter was attributed to a
different community than that of her two parents, with a distance of 15 km between the
mean sampling locations for these two communities. In the second mixed community
trio, however, the mother was attributed to a different community than the father and
daughter, with a distance of 15 km again between the communities attributed to the
mother and the daughter and father. The confidence associated with this trio was very
high (trio LOD score = 23.97, trio confidence score >99%). Nonetheless, one cannot rule
out the possibility that this trio contains a false positive assignment. If a true parent-
offspring trio, however, this could indicate the occurrence of a secondary or adult
dispersal event in which the mother emigrated to a different community following sexual
maturity and reproduction in a prior community. No trios with male offspring contained
mixed community affiliation.

98

Figure 4.2. Parent-offspring dyads detected among the corridor chimpanzees using
KinGroup. Black circles represent putative chimpanzee communities, arranged according
to their relative spatial locations in the corridor. Colored dots indicate genotyped
individuals, and lines connecting the dots indicate KinGroup parent-offspring
assignments among dyads. The upper figure (part a) indicates father-son dyads, while the
lower figure (part b) indicates mother-daughter dyads.
99

4.3.4. Parentage assignments including Busingiro
I found similar results when including genotypes from the Busingiro community
in order to explicitly test for dispersal between the corridor and the protected Budongo
Forest. In addition to the dyads previously identified among the corridor chimpanzees, I
identified 5 additional KinGroup mother-daughter dyads containing mixed community
affiliation between Busingiro and a corridor community, as well as 5 KinGroup father-
son dyads and 11 mixed sex parent-offspring dyads with mixed community affiliation
involving Busingiro and a corridor community. The permutation test once again showed
a significant difference between sexes (χ
2
= 10.99, p
perm
= 0.005, Figure 4.3).
4.3.5. Spatial analysis of repeatedly genotyped chimpanzees
By examining the spatial locations of samples from individuals who were
genotyped on multiple occasions, I identified the occurrence of one putative dispersal
event during the study period. A female chimpanzee assigned the consensus ID ‘C73’
was sampled in Katanga, a farm-forest mosaic near the southern border of the Budongo
and Siiba Forest Reserves, in November 2011. The location of this sample fell within the
MCP of co-sampled genotypes for the Katanga chimpanzee community. She was then
sampled three times during June 2013 at Kasongoire, an area with riparian forest
fragments surrounded by sugar cane plantations located approximately 13 km from the
sampling location at Katanga (Figure 4.4). These three samples were all located within
the MCP associated with the Kasongoire chimpanzee community. In three of four total
sampling occasions, her fecal sample was found in association with other nearby
similarly decayed samples, indicating her likely affiliation with other chimpanzees in that
100

Figure 4.3. Parent-offspring dyads detected among the corridor and Busingiro
chimpanzees using KinGroup. Black circles represent putative chimpanzee communities,
arranged according to their relative spatial locations. All communities fall in the corridor
with the exception of Busingiro, which is in the Budongo Forest and labeled for
clarification. Colored dots indicate genotyped individuals, and lines connecting the dots
indicate KinGroup parent-offspring assignments among dyads. The upper figure (part a)
indicates father-son dyads, while the lower figure (part b) indicates mother-daughter
dyads.
101

community. On two of those three occasions, once in each respective community, the
samples in association indicate C73’s likely inclusion in a mixed sex party. On the third
occasion, her sample was associated with one other female in Kasongoire.

Figure 4.4. Locations of genotyped samples collected for female C73. A sample from
this female was collected first in Nov. 2011 in the boundaries of the Katanga MCP, then
three times in June 2013 in the Kasongoire MCP. MCPs for all communities include a
lightly shaded one-km buffer to indicate likely minimum home range sizes extending
beyond sample collection locations. Green indicates forest cover during the study period,
as provided by Hansen et al. (2013). Chimpanzee illustration © Irene Goede Illustraties,
used with permission.

102

4.4. Discussion
This study represents a broad-scale effort to examine dispersal patterns across
multiple chimpanzee communities over more than 600 km
2
of fragmented forest habitat.
These findings support a pattern of species-typical male philopatry in a degraded habitat,
with father-son dyads in the fragmented forest habitat invariably exhibiting matching
community assignment. Further, these results demonstrate that species-typical female
dispersal has occurred in recent decades despite substantial habitat loss and
fragmentation, with members of 39% of KinGroup mother-daughter dyads containing
mixed community affiliation. I also found genetic evidence of a female transfer event
during the course of the study period. Collectively, these findings demonstrate that
chimpanzees in a degraded habitat display a similar social structure and dispersal
mechanism to those in more intact habitats. They also suggest this habitat can act as a
corridor for gene flow between the nearby Budongo and Bugoma Forests, each of which
contain populations of hundreds of chimpanzees.
To better compare these findings to what has been observed among chimpanzees
in less degraded habitats, one would need comparable data for both fragmented and intact
habitats. Unfortunately, data on dispersal patterns have often been difficult to obtain,
even among habituated chimpanzee communities in protected forests, since they typically
rely on long-term observations to obtain adequate life history data and these are only just
becoming available at some sites. Even where chimpanzees have been habituated,
females can be more elusive than males, making inferences regarding putative transfers
potentially erroneous (Langergraber et al., 2014a). Transfer rates can also vary
103

substantially over time within communities and among nearby communities, given the
influence of locally varying social and environmental factors that play a role in dispersal
likelihood (Nishida et al., 1985; Rudicell et al., 2010). In addition, the dispersal rate for
individual communities in this corridor cannot be inferred based on the available data
from mixed community dyads due to the lack of data on directionality of dispersal events
and life histories of individual females in each community. Despite the limited ability to
compare dispersal rates among communities, however, this study presents dispersal
patterns across a broad geographic scale in an increasingly common habitat type for
chimpanzees and other great apes, thereby providing a valuable opportunity for future
research in this region as well as an opportunity for future studies with comparative data
sets like this one.
Although these results demonstrate evidence of female dispersal in this human-
disturbed habitat, several limitations should be recognized for their interpretation. First,
these findings are necessarily limited by sampling scheme. Other unsampled
communities, either in the corridor habitat itself or in the Budongo and Bugoma Forests,
may have been associated with transfer events among these communities. Where
sampling ended, however, so did the feasibility of detecting such events. Nonetheless,
results from the data set including Busingiro indicate transfer between this community in
the Budongo Forest and the forest fragments to the south, suggesting gene flow between
the corridor region and the larger chimpanzee population in the continuous forest block.
The current study lacks genotype data for chimpanzees in the Bugoma Forest, but future
104

research may help identify the occurrence of transfer events between Bugoma and the
corridor habitat as well.
A second sampling limitation is that individuals I failed to successfully sample
and genotype in the corridor communities could not be included in the analysis, which
likely led to some missed parent-offspring dyads both within and among communities.
Given that there was a mean of 3.5 genotyped samples per individual in the corridor
habitat, however, sampling coverage was relatively strong. Finally, I used stringent
confidence levels for parentage assignments to minimize false positive errors despite the
potential for increased false negative errors. That is, I likely failed to detect some true
parent-offspring dyads in an effort to exclude false positive dyads. Nonetheless, I found a
significant sex difference in the proportions of mixed community affiliation despite this
potentially limited power. Therefore, the results presented here are strong given the
likelihood that these findings represent only the minimum level of dispersal, and that
some true parent-offspring dyads were likely not sampled or not detected using these
conservative analyses.
Though false positive assignments cannot be completely eliminated, their impact
on these data was likely minimal. If one considers the false positive error rates for the
Kibale data set and applies them to the corridor data set, one should expect 2 - 3 false
positive mother-daughter dyads in KinGroup and a maximum of 1 in CERVUS. Among
‘strict exclusion’ trios, one should expect a maximum of 1 trio containing a false positive
error. In general, parent-offspring dyads and trios containing sons had higher error rates,
but this should have little impact on conclusions regarding female dispersal. Though each
105

parentage assignment carries with it a modest degree of uncertainty, the pattern of results
remained similar regardless of the analysis approaches and parameters used, suggesting
these findings are robust.
Caution is also warranted when interpreting these results given that the timescale
of dispersal events in relation to recent habitat degradation is unknown. Considering that
female chimpanzees have been reported to survive to age 55 (Hill et al., 2001), and that
mean dispersal age is approximately 13 (Stumpf et al., 2009), some dispersal events
could have occurred over 40 years before the study period. However, given that females
survive to a mean age of 30 (Hill et al., 2001), most dispersal events are likely to have
occurred within the past two decades. Further, one would expect a steep attenuation in
recent dispersal frequency if landscape changes were either 1) drastic enough to inhibit
movement through the habitat substantially or 2) drastic enough to suppress female
reproductive success, leading to a ‘top-heavy’ demographic structure comprising mainly
adults. Though habitat loss has been severe in recent decades, there is still substantial
behavioral and genetic evidence within communities suggesting chimpanzees move
broadly among forest fragments in this habitat (McLennan, 2008; McCarthy et al., 2015).
Thus, while the impact of ongoing habitat fragmentation on chimpanzee mobility is
largely unknown, chimpanzee movement is clearly not restricted within the confined
boundaries of forest fragments. Regarding the overall demographic structure of these
communities, there is also evidence to suggest this has not been severely impacted by
forest fragmentation to date. Based on my review of published demographic data from
long-term study communities, a mean 39% of chimpanzee communities is composed of
106

infants and juveniles (Appendix F). In comparison, based on data from Bulindi, a
chimpanzee community in the corridor for which demographic composition is known
from years of behavioral observation, infants and juveniles comprised 42% of the
community during the study period (M. McLennan, pers. comm.). Thus, the demographic
stratification of this community does not appear to deviate from the norm despite
substantial habitat fragmentation in this region. This structure is likely not unique to
Bulindi among communities in this region, given that infants and/or juveniles were
visually observed in every chimpanzee community for which I had observations during
the course of data collection, which includes all but one putative community (M.
McCarthy, unpublished data). In addition, I documented genetic evidence of a dispersal
event during the study period, suggesting dispersal has not been entirely inhibited in
recent years despite habitat changes. The likelihood of future dispersal events is difficult
to predict, however, given ongoing habitat destruction and human-wildlife conflict in this
region.
Even if female chimpanzee dispersal is robust to a degree of habitat
fragmentation, other effects on its likelihood and timing are still possible. Though rarely
reported, secondary transfer and the delayed transfer of adult parous females sometimes
occur in chimpanzees and have been associated with severe disturbances in group
structure and dynamics (Nishida et al., 1985; Rudicell et al., 2010). Strong ecological
pressure from habitat loss could disrupt the structure of a community or constrain its size,
thereby leading to unusual dispersal events in response. These analyses identified two
mother-father-offspring mixed community trios. However, in contrast to the expectation
107

that the daughters these trios would be found in a different community from that of their
parents, the mother in one of these trios was attributed to a different community than the
father and daughter. Despite the high confidence score attributed to this trio, one must
interpret this finding with caution given the possibility of a false positive parentage
assignment. However, this may indicate a secondary transfer event or the delayed transfer
of a parous female who had reproduced in her natal community. Delayed or secondary
transfer may be advantageous to females if resources or unrelated mates are particularly
limited in the resident community, as may be the case under extreme situations of habitat
alteration or social upheaval.
In addition to its potential impacts on the timing of dispersal events, habitat
fragmentation could also lead to higher costs associated with female dispersal. High
human population density is associated with higher risks for chimpanzees traversing the
landscape, since traveling outside the forest can be associated with road crossings,
harassment from local residents, or other threats leading to chimpanzee injuries or death
(Hockings et al., 2006; McLennan et al., 2012; Cibot et al., 2015). The increasing
isolation of forest fragments may also make it more challenging for females to locate
habitat outside their natal community with remaining resident chimpanzee communities.
Limited resource availability may potentially lead to higher within-group competition for
stranger females arriving in a new community, which can also lead to increased risks of
aggression from resident females and may limit reproductive success (Williams et al.,
2002; Kahlenberg et al., 2008). More data are required to better understand how habitat
108

changes impact the likelihood and timing of chimpanzee dispersal, as well as the risks
associated with it.
The findings presented here have meaningful implications for the conservation of
chimpanzees in this region and others like it. The potential for ongoing gene flow is a
strong incentive to conserve remaining riparian forests and restore the corridor. This
would likely enhance the gene flow potential for other species as well, since the corridor
habitat is reported to harbor an additional six diurnal primate species, red duikers
(Cephalophus callipygus weynsii), bushbucks (Tragelaphus scriptus), and crested
porcupines (Hystrix cristata), as well as over 200 bird species, representing a higher
avian species richness than that found in the Bugoma Forest nearby (Plumptre et al.,
2011). Translocation of individual chimpanzee communities, as has been suggested
previously (Reynolds et al., 2003; Reynolds, 2005; McLennan and Hill, 2012), may be
both unwarranted and detrimental to chimpanzees in this region. First, it may be
unnecessary as it relies on the erroneous assumption that chimpanzees in this region are
highly isolated within individual forest fragments. Second, translocation may actually
disrupt gene flow since it would eliminate links among remaining chimpanzee
communities, thereby further isolating them. Such conservation strategies have been
suggested in part due to their potential to reduce conflict between chimpanzees in humans
in this region, which can be quite frequent (McLennan and Hill, 2012). However, given
the potentially detrimental consequences of such a strategy, conflict reduction measures
should be favored over translocation strategies whenever possible. Such measures may
include implementing education programs to promote less aggressive responses toward
109

chimpanzee presence, incentivizing tree planting and conservation in community forests,
and promoting the cultivation of crops that are less likely to cause negative interactions.
Finally, these findings demonstrate the conservation potential of unprotected
forests. Though protected areas are typically given conservation priority, fragmented and
unprotected habitats are becoming very common and may become highly valuable with
their increasing potential to act as both habitable areas and corridors for numerous
species (Turner and Corlett, 1996; McLennan and Plumptre, 2012; Ancrenaz et al., 2014;
McCarthy et al., 2015). The potential for this corridor and others like it can likely only be
realized if conservation action is taken quickly, however. Given the ongoing rate of
habitat destruction, the continued ability of chimpanzees and other species to inhabit and
disperse in this region remains highly uncertain.
4.5. Conclusions
Chimpanzees in the Budongo-Bugoma corridor landscape displayed evidence of
species-typical female dispersal in spite of substantial habitat fragmentation in recent
decades. The distribution of mother-daughter dyads between different communities was
widespread in this region. In contrast, father-son dyads were located in the same
community, as is typically observed among male philopatric chimpanzees in less
degraded habitats. Together, these findings suggest chimpanzees respond flexibly to
habitat fragmentation and that gene flow is not severely inhibited, at least in the short
term. However, the degree to which species-typical dispersal can be maintained despite
ongoing anthropogenic habitat modifications is currently unknown.

110

Chapter 5
Discussion

This research used a genetic approach to identify an unexpectedly large and
broadly distributed chimpanzee population inhabiting a fragmented and degraded forest
habitat. This population consisted of at least nine distinct communities, as identified
based on the spatial clustering of repeatedly sampled genotypes. These data also indicate
that chimpanzees in this habitat have had a species-typical social structure in recent
decades, with male philopatry and female dispersal, despite ongoing habitat loss and
111

fragmentation. These findings also suggest the occurrence of a female chimpanzee
dispersal event during the study period.
The use of a genetic approach to study chimpanzees in a fragmented forest habitat
allowed for the investigation of questions that would have been extremely challenging or
impossible using other methods. Data on chimpanzee density and abundance have
sometimes resulted from direct observations, achievable only through habituation of
study subjects to researcher presence. This process commonly requires years of work
before individuals can be identified reliably (Gruen et al., 2013). Because females can be
particularly elusive, detecting them can be challenging (Pepper et al., 1999; Bertolani and
Boesch, 2008). In addition, habituation is only feasible for single communities and
therefore cannot be used to estimate density and abundance at the multi-community
population level. Beyond these logistical challenges, habituation may also be fraught with
ethical dilemmas. Though researcher presence can help protect great apes (Campbell et
al., 2011), it is also associated with potentially increased risks of disease transmission,
stress to study subjects, and impacts on behavior, thereby calling into question the need
and value of habituating new primate populations (Wallis and Lee, 1999; Fedigan, 2010;
Gruen et al., 2013; Nekaris and Nijman, 2013). The particularly close co-residence of
chimpanzees and humans in this corridor landscape increases the potential for heightened
disease transmission risk, stress, and conflict through a habituation approach (McLennan,
2008; McLennan and Huffman, 2012). Given these considerations, habituation would not
have been an appropriate or feasible means of study.
112

Nest count surveys are the most common and well-established alternative to
habituation for gathering baseline data on great ape density and abundance. However,
their use in this research would not have allowed me to examine the movement of
individual chimpanzees or to estimate the number and locations of chimpanzee
communities in the region. Moreover, nest count surveys could not have been used to
examine the occurrence of chimpanzee dispersal in this region. Finally, as discussed in
Chapter 3, a previous small-scale nest count survey in this region led to an underestimate
of the number of chimpanzees (Plumptre et al., 2003). Therefore, the genetic approach
used in this research allowed a more precise and accurate means to measure chimpanzee
abundance and the only feasible means to study dispersal and individual movement.
Despite the potential associated with microsatellite genotyping, there are
drawbacks worth noting. First, microsatellite genotyping is a relatively time-consuming
approach as compared to other methods such as nest count surveys for abundance
estimation. It can take years to acquire and analyze the data necessary to answer
questions such as those posed in this research project. Hundreds of fecal samples are
typically necessary to achieve narrow confidence intervals in abundance estimation using
capture-mark-recapture methods. It has been suggested that three times the number of
samples as individuals are necessary to obtain narrow confidence intervals using standard
capture-mark-recapture methods (Miller et al., 2005; Petit and Valiere, 2006;
Arandjelovic et al., 2010). In this study I collected more than 800 fecal samples and
obtained a mean of 3.5 captures (genotyped samples) per chimpanzee. This allowed a
moderate degree of precision, though more captures would have been necessary,
113

particularly in lightly sampled areas, to obtain a highly precise estimate with a very
narrow confidence interval. Precision is particularly important if one hopes to monitor
population changes over time. The samples in this study were collected over a 15-month
period and required nearly a year of genotyping. Data analyses to estimate chimpanzee
abundance, distribution, and dispersal patterns took many additional months, requiring an
investment of several years for total project completion. This timeline is typical and
would be expected for other projects with similar spatial scales and similar target sample
sizes. Given the rate of habitat change and the risk severity for some endangered
populations, this time investment would not be appropriate for all potential projects.
Projects involving microsatellite genotyping are also quite costly and require the
investment of thousands of dollars in research funds for both data collection and genetics
lab work. Therefore, the considerable time and monetary investment needed must be
weighed against the potential benefits of this approach. Nonetheless, if appropriate given
the time and monetary costs required, the benefits can make this approach highly
advantageous for answering questions that would otherwise be impossible to answer
using other approaches currently available to researchers.
In addition to the methodological advantages it affords, this genetic approach also
provides important information to help guide conservation efforts. The findings presented
in Chapter 3 demonstrate that chimpanzees persist in a fragmented forest habitat more
than would have previously been expected. Though research and conservation investment
is more often focused on protected areas, these findings demonstrate the value of
investing research and conservation effort in degraded habitats. Further, these findings
114

demonstrate that that this habitat acts as a corridor for gene flow among chimpanzee
communities in unprotected forests and, importantly, into and out of the larger forest
blocks.
This conservation potential can be translated into action in a number of ways. The
first involves the enforcement of existing laws. Most of the forests in this region are
privately owned, with 70% of forests in Uganda on private lands (Turyahabwe and
Banana, 2008). Although this may seem to suggest their use is governed by the discretion
of local landowners, current regulations should prevent over-exploitation. River banks,
defined as the 100-m buffer at the edges of rivers, are protected from destruction in
Uganda given their vital role in protecting waterways and preventing erosion (MWLE,
2000). Further, the use of chainsaws for timber extraction is restricted (MWLE, 2002).
These laws are not well enforced and deforestation in riparian forests is widespread
(McLennan and Plumptre, 2012; Twongyirwe et al., 2015). The enforcement of these
laws would protect the riparian forests and waterways that provide vital habitat to
chimpanzees and critical ecosystem services for local residents.
Second, plans should be implemented to restore and enhance this corridor region
strategically. Previous research has identified specific corridors that could be enhanced
among these forest fragments to benefit the ranging and dispersal of key taxa (Nangendo
et al., 2010). However, such plans are costly to implement and have not been executed to
date. The research presented here suggests this effort would be highly worthwhile given
its potential to enhance gene flow among the remaining chimpanzee population in this
region, and would like benefit many other species as well.
115

In addition, conflict mitigation strategies are needed to lessen the frequency of
negative interactions between chimpanzees and human residents. The Banyoro people
have not traditionally hunted chimpanzees. Customs have been changing in recent years,
however, and reports have emerged citing instances of chimpanzee hunting in Hoima and
Masindi Districts (McLennan, 2008). Land use changes and habitat loss have also
increased the frequency of interactions and led to heightened conflict in recent decades
(McLennan and Hill, 2012). Chimpanzees are increasingly at risk of being snared, caught
in man-traps, speared, or otherwise injured or killed by humans (Reynolds, 2005;
McLennan, 2008; McLennan et al., 2012). In severe instances, children have also been
injured or killed by chimpanzees (Wrangham et al., 2000; Kamenya, 2002; Reynolds,
2005; McLennan and Hill, 2012). The planting of unpalatable buffer crops along riparian
forest edges may help attenuate such conflict. Additionally, the harassment of
chimpanzees, particularly when they feed on crops, has been cited as a potential reason
for instances of their aggressive behavior, and avoiding such provocation may help lessen
conflict (McLennan and Hill, 2012). Finally, educational programs, particularly those
targeting young people, can help shift attitudes and increase understanding of
chimpanzees and, more generally, of the value of ecosystem services (Borchers et al.,
2014).
These measures are urgently needed. A recent spatial analysis demonstrated a
slowing of forest loss in unprotected forests surrounding the Budongo and Bugoma
Reserves from 2010 to 2014 (Twongyirwe et al., 2015). Though this finding appears
promising superficially, Twongyirwe and colleagues conclude that this result is actually
116

driven by a steep decline in available remaining forest for exploitation. Concurrently, the
overall chimpanzee population is likely on the decline throughout this region. In 2007 –
2008, the Bulindi chimpanzee community consisted of ~34 individuals (McLennan,
2010). Six years later, during the data collection period for this research, there were 19
chimpanzees in the community (M. McLennan, pers. comm.). The steep decline in
community size over such a short period reflects the ongoing threats faced by this
community and more broadly, by this population as a whole.
To slow the loss of remaining forests in Uganda, however, one must address the
need for forest resources. Approximately 90% of Ugandans depend on wood fuels as a
primary energy source (MWLE, 2002), with more than 95% of households gathering
firewood from local sources for use in cooking (UBOS, 2007). In a survey of forest use in
this region, 84% of respondents said they acquired house construction materials from
these riparian forests (Akwetaireho et al., 2011). Local residents also depend strongly on
the rivers in these forests, with 53% of households locally citing these waterways as their
main domestic water source (Akwetaireho et al., 2011). Agriculture and livestock are
critically dependent on the waterways as well.
How, then, can the need for these resources be reconciled with conservation
goals? First, it is critical to recognize that these interests are not mutually exclusive.
Riparian forests help protect waterways and buffer the effects of climate change (Smith,
1997). Therefore, investments in their protection are investments in the futures of human
and non-human residents alike. Second, more sustainable practices could help ensure that
the needs of local residents are met while also protecting forests and wildlife. The
117

clearing of riparian forests to plant commercial crops like tobacco is commonplace
regionally and should be actively prohibited as it is both largely unnecessary and illegal
(see Figure 2.4). High efficiency cookstoves, in combination with the use of plantation
forest woods that do not compromise natural forests, could help reduce reliance on
limited natural forest resources. These alternatives, though promising, are not yet broadly
available and affordable for residents of Hoima and Masindi Districts.
Considering the challenges inherent in conservation strategies that must balance
the needs of local residents with those of wildlife, they may seem like a futile endeavor.
Indeed, the future prospects for this chimpanzee population seem bleak given its current
trajectory. Nonetheless, the critical value of this population and the broader ecosystem
dictates that efforts must be made to conserve them. Chimpanzees are among our
evolutionary next of kin and can teach us a great deal about ourselves and our
evolutionary past. Beyond that, however, they are also inherently valuable beings,
exhibiting cultural traditions and complex cognition as well as sharing numerous
behavioral and emotional similarities with us. To allow their local or global extinction
would be a great detriment not only to the ecosystems they support but also to our
biological heritage.
As species worldwide inhabit increasingly degraded landscapes, the need to study
and conserve them in these habitats will also increase. This research has demonstrated
remarkable behavioral flexibility by chimpanzees in a human-dominated and fragmented
environment. Despite substantial habitat changes in recent decades, they have persisted,
maintaining a widespread distribution with intact social structure and dispersal. These
118

findings underscore not only the behavioral flexibility inherent in chimpanzees but also
the value of conserving species in unprotected landscapes. Increasingly in the
Anthropocene, research in habitats like this may overturn the ‘tainted-nature delusion’, in
which nature is valued only when it seems sufficiently pure for our liking (Sheil and
Meijaard, 2010). Given the current reality, in which great apes and many other species
live primarily outside protected areas, we should give greater value and emphasis to
studying and protecting populations such as these.

119

Bibliography
Aars J, Ims RA. 2000. Population dynamic and genetic consequences of spatial density-
dependent dispersal in patchy populations. Am Nat 155:252–265.
Akwetaireho S, Akugizibwe T, Plumptre A. 2011. Socioeconomic values of corridor
forests in the Albertine Rift forests of the Murchison-Semliki Landscape. Kampala:
Wildlife Conservation Society.
Ancrenaz M, Marshall AJ, Goossens B, van Schaik CP, Sugardjito J, Gumal M, Wich
SA. 2008. Pongo pygmaeus. The IUCN Red List of Threatened Species. Version
2015.2. Available from: http://www.iucnredlist.org
Ancrenaz M, Oram F, Ambu L, Lackman I, Ahmad E, Elahan H, Kler H, Abram NK,
Meijaard E. 2014. Of Pongo, palms and perceptions: a multidisciplinary assessment
of Bornean orang-utans Pongo pygmaeus in an oil palm context. Oryx 49:465–472.
Anderson DP, Nordheim EV, Boesch C, Moermond TC. 2002. Factors influencing
fission-fusion grouping in chimpanzees in the Taï National Park, Côte d'Ivoire. In:
Boesch C, Hohmann G, Marchant LF, editors. Behavioural diversity in chimpanzees
and bonobos. Cambridge: Cambridge University Press. p 90–101.
Arandjelovic M, Bergl RA, Ikfuingei R, Jameson C, Parker M, Vigilant L. 2015.
Detection dog efficacy for collecting faecal samples from the critically endangered
Cross River gorilla (Gorilla gorilla diehli) for genetic censusing. R Soc Open Sci
2:140423.
Arandjelovic M, Guschanski K, Schubert G, Harris TR, Thalmann O, Siedel H, Vigilant
L. 2009. Two-step multiplex polymerase chain reaction improves the speed and
accuracy of genotyping using DNA from noninvasive and museum samples. Mol
Ecol Resour 9:28–36.
Arandjelovic M, Head J, Boesch C, Robbins MM, Vigilant L. 2014. Genetic inference of
group dynamics and female kin structure in a western lowland gorilla population
(Gorilla gorilla gorilla). Primate Biol 1:29–38.
Arandjelovic M, Head J, Kühl HS, Boesch C, Robbins MM, Maisels F, Vigilant L. 2010.
Effective non-invasive genetic monitoring of multiple wild western gorilla groups.
Biol Conserv 143:1780–1791.
Arandjelovic M, Head J, Rabanal LI, Schubert G, Mettke E, Boesch C, Robbins MM,
Vigilant L. 2011. Non-invasive genetic monitoring of wild central chimpanzees.
PLoS One 6:e14761.

120

Arrendal J, Vilà C, Björklund M. 2007. Reliability of noninvasive genetic census of
otters compared to field censuses. Conserv Genet 8:1097–1107.
Baglione V, Canestrari D, Marcos JM, Ekman J. 2006. Experimentally increased food
resources in the natal territory promote offspring philopatry and helping in
cooperatively breeding carrion crows. Proc Biol Sci 273:1529–1535.
Baguette M, Mennechez G, Petit S, Schtickzelle N. 2003. Effect of habitat fragmentation
on dispersal in the butterfly Proclossiana eunomia. C R Biol 326:200–209.
Bergl RA, Warren Y, Nicholas A, Dunn A, Imong I, Sunderland-Groves JL, Oates JF.
2012. Remote sensing analysis reveals habitat, dispersal corridors and expanded
distribution for the Critically Endangered Cross River gorilla Gorilla gorilla diehli.
Oryx 46:278–289.
Bertolani P, Boesch C. 2008. Habituation of wild chimpanzees (Pan troglodytes) of the
South Group at Taï Forest, Côte d'Ivoire: empirical measure of progress. Folia
Primatol 79:162–171.
Blouin MS. 2003. DNA-based methods for pedigree reconstruction and kinship analysis
in natural populations. Trends Ecol Evol 18:503–511.
Boesch C, Boesch-Achermann H. 2000. The Chimpanzees of the Taï Forest. New York:
Oxford University Press.
Borchers C, Boesch C, Riedel J, Guilahoux H, Ouattara D, Randler C. 2014.
Environmental education in Côte d'Ivoire/West Africa: extra-curricular primary
school teaching shows positive impact on environmental knowledge and attitudes. Int
J Sci Educ, Part B 4:240–259.
Borchers DL, Efford MG. 2008. Spatially explicit maximum likelihood methods for
capture-recapture studies. Biometrics 64:377–385.
Bowler DE, Benton TG. 2009. Variation in dispersal mortality and dispersal propensity
among individuals: the effects of age, sex and resource availability. J Anim Ecol
78:1234–1241.
Boyko RH, Marshall AJ. 2010. Using simulation models to evaluate ape nest survey
techniques. PLoS One 5:e10754.
Bradley BJ, Chambers KE, Vigilant L. 2001. Accurate DNA-based sex identification of
apes using non-invasive samples. Conserv Genet 2:179–181.
Brncic TM, Amarasekaran B, McKenna A. 2010. Sierra Leone national chimpanzee
census. Freetown: Tacugama Chimpanzee Sanctuary.
121

Broquet T, Petit E. 2004. Quantifying genotyping errors in noninvasive population
genetics. Mol Ecol 13:3601–3608.
Buckland ST, Plumptre AJ, Thomas L, Rexstad EA. 2010. Design and analysis of line
transect surveys for primates. Int J Primatol 31:833–847.
Butynski TM. 1990. Comparative ecology of blue monkeys (Cercopithecus mitis) in
high- and low-density subpopulations. Ecol Monogr 60:1–26.
Caldecott JO, Miles L. 2005. World atlas of great apes and their conservation. Berkeley,
CA: University of California Press.
Campbell G, Kühl HS, Diarrassouba A, N'Goran PK, Boesch C. 2011. Long-term
research sites as refugia for threatened and over-harvested species. Biol Lett 7:723–
726.
Chaine AS, Clobert J. 2012. Dispersal. In: Candolin U, Wong BBM, editors. Behavioural
responses to a changing world. Oxford: Oxford University Press. p 63–79.
Chancellor RL, Langergraber K, Ramirez S, Rundus AS, Vigilant L. 2012. Genetic
sampling of unhabituated chimpanzees (Pan troglodytes schweinfurthii) in Gishwati
Forest Reserve, an isolated forest fragment in western Rwanda. Int J Primatol
33:479–488.
Charnov EL, Berrigan D. 1993. Why do female primates have such long lifespans and so
few babies? Or life in the slow lane. Evol Anthropol 1:191–194.
CIA ed. 2013. The World Factbook 2013-14. Available from:
http://www.cia.gov/library/publications/the-world-factbook/index.html
Cibot M, Bortolamiol S, Seguya A, Krief S. 2015. Chimpanzees facing a dangerous
situation: a high-traffic asphalted road in the Sebitoli area of Kibale National Park,
Uganda. Am J Primatol 77:890–900.
Clutton-Brock TH. 1989. Female transfer and inbreeding avoidance in social mammals.
Nature 337:70–72.
Cooper CB, Walters JR. 2002. Experimental evidence of disrupted dispersal causing
decline of an Australian passerine in fragmented habitat. Conserv Biol 16:471–478.
Corlett RT. 2015. The Anthropocene concept in ecology and conservation. Trends Ecol
Evol 30:36–41.
Crutzen PJ, Stoermer EF. 2000. The Anthropocene. International Geosphere Biosphere
Programme Newsletter 41:17–18.
122

Csilléry K, Johnson T, Beraldi D, Clutton-Brock T, Coltman D, Hansson B, Spong G,
Pemberton JM. 2006. Performance of marker-based relatedness estimators in natural
populations of outbred vertebrates. Genet 173:2091–2101.
Debinski DM, Holt RD. 2000. A survey and overview of habitat fragmentation
experiments. Conserv Biol 14:342–355.
Devos C, Sanz C, Morgan D, Onononga JR, Laporte N, Huynen M-C. 2008. Comparing
ape densities and habitats in northern Congo: surveys of sympatric gorillas and
chimpanzees in the Odzala and Ndoki regions. Am J Primatol 70:439–451.
Doran D. 1997. Influence of seasonality on activity patterns, feeding behavior, ranging,
and grouping patterns in Taï chimpanzees. Int J Primatol 18:183–206.
Douadi MI, Gatti S, Levrero F, Duhamel G, Bermejo M, Vallet D, Menard N, Petit E.
2007. Sex-biased dispersal in western lowland gorillas (Gorilla gorilla gorilla). Mol
Ecol 16:2247–2259.
Dudash MR, Fenster CB. 2000. Inbreeding and outbreeding depression in fragmented
populations. In: Young AG, Clarke GM, editors. Genetics, demography and viability
of fragmented populations. Cambridge: Cambridge University Press. p 35–54.
Efford M. 2004. Density estimation in live-trapping studies. Oikos 106:598–610.
Efford MG, Borchers DL, Byrom AE. 2009a. Density estimation by spatially explicit
capture-recapture: likelihood-based methods. In: Thompson DL, Cooch EG, Conroy
MJ, editors. Modeling demographic processes in marked populations. New York:
Springer. p 255–269.
Efford MG, Borchers DL, Mowat G. 2013. Varying effort in capture-recapture studies.
Methods Ecol Evol 4:629–636.
Efford MG, Dawson DK, Borchers DL. 2009b. Population density estimated from
locations of individuals on a passive detector array. Ecol 90:2676–2682.
Efford MG, Fewster RM. 2013. Estimating population size by spatially explicit capture-
recapture. Oikos 122:918–928.
Efford MG. 2011. Estimation of population density by spatially explicit capture-recapture
analysis of data from area searches. Ecol 92:2202–2207.
Efford MG. 2015. SECR: spatially explicit capture-recapture models. R package version
2.9.3. Available from: http://CRAN.R-project.org/package=secr
Eggeling WJ. 1947. Observations on the ecology of the Budongo Rain Forest, Uganda. J
Ecol 34:20–87.
123

Emery Thompson M, Jones JH, Pusey AE, Brewer-Marsden S, Goodall J, Marsden D,
Matsuzawa T, Nishida T, Reynolds V, Sugiyama Y, Wrangham RW. 2007. Aging
and fertility patterns in wild chimpanzees provide insights into the evolution of
menopause. Curr Biol 17:2150–2156.
Eriksson J, Siedel H, Lukas D, Kayser M, Erler A, Hashimoto C, Hohmann G, Boesch C,
Vigilant L. 2006. Y-chromosome analysis confirms highly sex-biased dispersal and
suggests a low male effective population size in bonobos (Pan paniscus). Mol Ecol
15:939–949.
Erler A, Stoneking M, Kayser M. 2004. Development of Y-chromosomal microsatellite
markers for nonhuman primates. Mol Ecol 13:2921–2930.
FAO. 2010. Global forest resources assessment 2010: country report, Uganda. Rome:
Food and Agriculture Organization of the United Nations.
Fedigan LM. 2010. Ethical issues faced by field primatologists: asking the relevant
questions. Am J Primatol 72:754–771.
Fietz J, Tomiuk J, Loeschcke V, Weis-Dootz T, Segelbacher G. 2014. Genetic
consequences of forest fragmentation for a highly specialized arboreal mammal - the
edible dormouse. PLoS One 9:e88092.
Flagstad Ø, Pradhan NMB, Kvernstuen LG, Wegge P. 2012. Conserving small and
fragmented populations of large mammals: non-invasive genetic sampling in an
isolated population of Asian elephants in Nepal. J Nat Conserv 20:181–190.
Fleury-Brugiere M-C, Brugiere D. 2010. High population density of Pan troglodytes
verus in the Haut Niger National Park, Republic of Guinea: implications for local and
regional conservation. Int J Primatol 31:383–392.
Fruth B, Benishay JM, Bila-Isia I, Coxe S, Dupain J, Furuichi T, Hart J, Hart T,
Hashimoto C, Hohmann G, Hurley M, Ilambu O, Mulavwa MN, Ndunda M,
Omasombo V, Reinartz G, Scherlis J, Steel L, Thompson J. 2008. Pan paniscus. The
IUCN Red List of Threatened Species. Version 2015.2. Available from:
http://www.iucnredlist.org
Gautier-Hion A, Brugiere D. 2005. Significance of riparian forests for the conservation of
Central African primates. Int J Primatol 26:515–523.
Gaveau D, Wandono H, Setiabudi F. 2007. Three decades of deforestation in southwest
Sumatra: have protected areas halted forest loss and logging, and promoted re-
growth? Biol Conserv 134:495–504.

124

Gerlach G, Musolf K. 2000. Fragmentation of landscape as a cause for genetic
subdivision in bank voles. Conserv Biol 14:1066–1074.
Goodall J. 1986. The chimpanzees of Gombe: patterns of behaviour. Cambridge, MA:
Harvard University Press.
Goodnight KF, Queller DC. 1999. Computer software for performing likelihood tests of
pedigree relationship using genetic markers. Mol Ecol 8:1231–1234.
Goossens B, Anthony N, Jeffery K, Johnson-Bawe M, Bruford MW. 2009. Collection,
storage and analysis of non-invasive genetic material in primate biology. In: Setchell
JM, Curtis DJ, editors. Field and laboratory methods in primatology: a practical
guide. 2nd ed. Cambridge: Cambridge University Press. p 371–386.
Gruen L, Fultz A, Pruetz J. 2013. Ethical issues in African great ape field studies. ILAR J
54:24–32.
Guschanski K, Vigilant L, McNeilage A, Gray M, Kagoda E, Robbins MM. 2009.
Counting elusive animals: comparing field and genetic census of the entire mountain
gorilla population of Bwindi Impenetrable National Park, Uganda. Biol Conserv
142:290–300.
Haddad NM, Brudvig LA, Clobert J, Davies KF, Gonzalez A, Holt RD, Lovejoy TE,
Sexton JO, Austin MP, Collins CD, Cook WM, Damschen EI, Ewers RM, Foster BL,
Jenkins CN, King AJ, Laurance WF, Levey DJ, Margules CR, Melbourne BA,
Nicholls AO, Orrock JL, Song DX, Townshend JR. 2015. Habitat fragmentation and
its lasting impact on Earth's ecosystems. Sci Adv 1:e1500052.
Hansen MC, Potapov PV, Moore R, Hancher M, Turubanova SA, Tyukavina A, Thau D,
Stehman SV, Goetz SJ, Loveland TR, Kommareddy A, Egorov A, Chini L, Justice
CO, Townshend JRG. 2013. High-resolution global maps of 21st-century forest
cover change. Science 342:850–853.
Head JS, Boesch C, Robbins MM, Rabanal LI, Makaga L, Kühl HS. 2013. Effective
sociodemographic population assessment of elusive species in ecology and
conservation management. Ecol Evol 3:2903–2916.
Heinicke S, Kalan AK, Wagner OJJ, Mundry R, Lukashevich H, Kühl HS. 2015.
Assessing the performance of a semi-automated acoustic monitoring system for
primates. Methods Ecol Evol 6:753–763.
Herbinger I, Boesch C, Rothe H. 2001. Territory characteristics among three neighboring
chimpanzee communities in the Taï National Park, Côte d'Ivoire. Int J Primatol
22:143–167.
125

Hicks TC, Roessingh P, Menken SBJ. 2013. Impact of humans on long-distance
communication behaviour of eastern chimpanzees (Pan troglodytes schweinfurthii) in
the northern Democratic Republic of the Congo. Folia Primatol 84:135–156.
Hicks TC. 2010. A chimpanzee mega-culture? Exploring behavioral continuity in Pan
troglodytes schweinfurthii across northern DR Congo. Doctoral dissertation,
University of Amsterdam, The Netherlands.
Hill K, Boesch C, Goodall J, Pusey A, Williams J, Wrangham R. 2001. Mortality rates
among wild chimpanzees. J Hum Evol 40:437–450.
Hockings K, Anderson JR, Matsuzawa T. 2006. Road crossing in chimpanzees: a risky
business. Curr Biol 16:R668–R670.
Hockings KJ, Anderson JR, Matsuzawa T. 2012. Socioecological adaptations by
chimpanzees, Pan troglodytes verus, inhabiting an anthropogenically impacted
habitat. Anim Behav 83:801–810.
Hockings KJ, McLennan MR, Carvalho S, Ancrenaz M, Bobe R, Byrne RW, Dunbar
RIM, Matsuzawa T, McGrew WC, Williamson EA, Wilson ML, Wood B,
Wrangham RW, Hill CM. 2015. Apes in the Anthropocene: flexibility and survival.
Trends Ecol Evol 30:215–222.
Hockings KJ, McLennan MR. 2012. From forest to farm: systematic review of cultivar
feeding by chimpanzees – management implications for wildlife in anthropogenic
landscapes. PLoS One 7:e33391.
Honess PE, MacDonald DW. 2011. Marking and radio-tracking primates. In: Setchell
JM, Curtis DJ, editors. Field and laboratory methods in primatology: a practical
guide. 2nd ed. Cambridge: Cambridge University Press. p 189–206.
Houghton RA. 2012. Carbon emissions and the drivers of deforestation and forest
degradation in the tropics. Curr Opin Environ Sustain 4:597–603.
Howe EJ, Obbard ME, Kyle CJ. 2013. Combining data from 43 standardized surveys to
estimate densities of female American black bears by spatially explicit capture–
recapture. Popul Ecol 55:595–607.
Imong I, Robbins MM, Mundry R, Bergl R, Kühl HS. 2014. Distinguishing ecological
constraints from human activity in species range fragmentation: the case of Cross
River gorillas. Anim Conserv 17:323–331.
Ims RA, Hjermann DO. 2012. Condition-dependent dispersal. In: Clobert J, Danchin E,
Dhondt AA, Nichols JD, editors. Dispersal. New York: Oxford University Press.
126

Inoue E, Akomo-Okoue EF, Ando C, Iwata Y, Judai M, Fujita S, Hongo S, Nze-Nkogue
C, Inoue-Murayama M, Yamagiwa J. 2013. Male genetic structure and paternity in
western lowland gorillas (Gorilla gorilla gorilla). Am J Phys Anthropol 151:583–
588.
IUCN ed. 2014. IUCN Red List of Threatened Species. Available from:
http://www.iucnredlist.org
JGI-UWA. 2002. Investigation into the human-chimpanzee conflict in Hoima District.
Entebbe: Jane Goodall Institute (JGI), Uganda Wildlife Authority (UWA).
Junker J, Blake S, Boesch C, Campbell G, Toit LD, Duvall C, Ekobo A, Etoga G, Galat-
Luong A, Gamys J, Ganas-Swaray J, Gatti S, Ghiurghi A, Granier N, Hart J, Head J,
Herbinger I, Hicks TC, Huijbregts B, Imong IS, Kuempel N, Lahm S, Lindsell J,
Maisels F, McLennan M, Martinez L, Morgan B, Morgan D, Mulindahabi F, Mundry
R, N'Goran KP, Normand E, Ntongho A, Okon DT, Petre C-A, Plumptre A, Rainey
H, Regnaut S, Sanz C, Stokes E, Tondossama A, Tranquilli S, Sunderland-Groves
JL, Walsh P, Warren Y, Williamson EA, Kühl HS. 2012. Recent decline in suitable
environmental conditions for African great apes. Divers Distrib 18:1077–1091.
Junker J, van Aarde RJ, Ferreira SM. 2008. Temporal trends in elephant Loxodonta
africana numbers and densities in northern Botswana: is the population really
increasing? Oryx 42:58–65.
Kahlenberg SM, Thompson ME, Muller MN, Wrangham RW. 2008. Immigration costs
for female chimpanzees and male protection as an immigrant counterstrategy to
intrasexual aggression. Anim Behav 76:1497–1509.
Kalan AK, Mundry R, Wagner OJJ, Heinicke S, Boesch C, Kühl HS. 2015. Towards the
automated detection and occupancy estimation of primates using passive acoustic
monitoring. Ecol Indic 54:217–226.
Kalinowski ST, Taper ML, Marshall TC. 2007. Revising how the computer program
CERVUS accommodates genotyping error increases success in paternity assignment.
Mol Ecol 16:1099–1106.
Kamenya S. 2002. Human baby killed by Gombe chimpanzee. Pan Africa News 9:26.
Kayser M, Brauer S, Weiss G, Schiefenhovel W, Underhill P, Shen P, Oefner P,
Tommaseo-Ponzetta M. 2007. Reduced Y-chromosome, but not mitochondrial DNA,
diversity in human populations from West New Guinea. Am J Hum Genet 72:281–
302.

127

Keller I, Largiader CR. 2003. Recent habitat fragmentation caused by major roads leads
to reduction of gene flow and loss of genetic variability in ground beetles. Proc R Soc
Lond B Biol Sci 270:417–423.
Knott CD. 2005. Energetic responses to food availability in the great apes: implications
for hominin evolution. In: Brockman DK, van Schaik CP, editors. Seasonality in
primates. Cambridge: Cambridge University Press. p 351–378.
Koh LP, Wich SA. 2012. Dawn of drone ecology: low-cost autonomous aerial vehicles
for conservation. Trop Conserv Sci 5:121–132.
Konovalov DA, Manning C, Henshaw MT. 2004. Kingroup: a program for pedigree
relationship reconstruction and kin group assignments using genetic markers. Mol
Ecol Notes 4:779–782.
Kormos R, Boesch C, Bakarr M, Butynski TM. 2003. Status survey and conservation
action plan: West African chimpanzees. Gland, Switzerland: IUCN/SSC Primate
Specialist Group.
Kosydar AJ, Conquest LL, Tewksbury JJ. 2014. Can life histories predict the effects of
habitat fragmentation? A meta-analysis with terrestrial mammals. Appl Ecol Env Res
12:505–521.
Kouakou CY, Boesch C, Kuehl H. 2009. Estimating chimpanzee population size with
nest counts: validating methods in Taï National Park. Am J Primatol 71:447–457.
Krief S, Cibot M, Bortolamiol S, Seguya A, Krief J-M, Masi S. 2014. Wild chimpanzees
on the edge: nocturnal activities in croplands. PLoS One 9:e109925.
Kumar V, Langstieh BT, Madhavi KV, Naidu VM, Singh HP, Biswas S, Thangaraj K,
Singh L, Reddy BM. 2006. Global patterns in human mitochondrial DNA and Y-
chromosome variation caused by spatial instability of the local cultural processes.
PLoS Genet 2:e53.
Kühl HS, Burghardt T. 2013. Animal biometrics: quantifying and detecting phenotypic
appearance. Trends Ecol Evol 28:432–441.
Kühl HS, Maisels F, Williamson MAEA. 2008. Best practice guidelines for surveys and
monitoring of great ape populations. Gland, Switzerland: IUCN/SSC Primate
Specialist Group.
Langdale-Brown I, Osmaston HA, Wilson JG. 1964. The vegetation of Uganda and its
bearing on land use. Entebbe: Government Printer.

128

Langergraber K, Schubert G, Rowney C, Wrangham R, Zommers Z, Vigilant L. 2011.
Genetic differentiation and the evolution of cooperation in chimpanzees and humans.
Proc R Soc Lond B Biol Sci 278:2546–2552.
Langergraber KE, Mitani JC, Vigilant L. 2007a. The limited impact of kinship on
cooperation in wild chimpanzees. Proc Natl Acad Sci USA 104:7786–7790.
Langergraber KE, Prüfer K, Rowney C, Boesch C, Crockford C, Fawcett K, Inoue E,
Inoue-Murayama M, Mitani JC, Muller MN, Robbins MM, Schubert G, Stoinski TS,
Viola B, Watts DP, Wittig R, Wrangham RW, Zuberbuhler K, Paabo S, Vigilant L.
2012. Generation times in wild chimpanzees and gorillas suggest earlier divergence
times in great ape and human evolution. Proc Natl Acad Sci USA 109:15716–15721.
Langergraber KE, Rowney C, Crockford C, Wittig R, Zuberbuhler K, Vigilant L. 2014a.
Genetic analyses suggest no immigration of adult females and their offspring into the
Sonso community of chimpanzees in the Budongo Forest Reserve, Uganda. Am J
Primatol 76:640–648.
Langergraber KE, Rowney C, Schubert G, Crockford C, Hobaiter C, Wittig R,
Wrangham RW, Zuberbuhler K, Vigilant L. 2014b. How old are chimpanzee
communities? Time to the most recent common ancestor of the Y-chromosome in
highly patrilocal societies. J Hum Evol 69:1–7.
Langergraber KE, Siedel H, Mitani JC, Wrangham RW, Reynolds V, Hunt K, Vigilant L.
2007b. The genetic signature of sex-biased migration in patrilocal chimpanzees and
humans. PLoS One 2:e973.
Last C, Muh B. 2013. Effects of human presence on chimpanzee nest location in the
Lebialem-Mone Forest Landscape, Southwest Region, Cameroon. Folia Primatol
84:51–63.
Lens L, Dhondt AA. 1994. Effects of habitat fragmentation on the timing of crested tit
Parus cristatus natal dispersal. Ibis 136:147–152.
Li Y, Lancaster ML, Cooper SJB, Taylor AC, Carthew SM. 2014. Population structure
and gene flow in the endangered southern brown bandicoot (Isoodon obesulus
obesulus) across a fragmented landscape. Conserv Genet 16:331–345.
Lukacs PM, Burnham KP. 2005. Review of capture-recapture methods applicable to
noninvasive genetic sampling. Mol Ecol 14:3909–3919.
Lyke MM, Dubach J, Briggs MB. 2013. A molecular analysis of African lion (Panthera
leo) mating structure and extra-group paternity in Etosha National Park. Mol Ecol
22:2787–2796.
129

Marshall TC, Slate J, Kruuk LE, Pemberton JM. 1998. Statistical confidence for
likelihood-based paternity inference in natural populations. Mol Ecol 7:639–655.
McCarthy MS, Lester JD, Howe EJ, Arandjelovic M, Stanford CB, Vigilant L. 2015.
Genetic censusing identifies an unexpectedly sizeable population of an endangered
large mammal in a fragmented forest landscape. BMC Ecol 15:21.
McLennan MR, Hill CM. 2010. Chimpanzee responses to researchers in a disturbed
forest-farm mosaic at Bulindi, western Uganda. Am J Primatol 72:907–918.
McLennan MR, Hill CM. 2012. Troublesome neighbors: changing attitudes towards
chimpanzees (Pan troglodytes) in a human-dominated landscape in Uganda. J Nat
Conserv 20:219–227.
McLennan MR, Huffman MA. 2012. High frequency of leaf swallowing and its
relationship to intestinal parasite expulsion in “village” chimpanzees at Bulindi,
Uganda. Am J Primatol 74:642–650.
McLennan MR, Hyeroba D, Asiimwe C, Reynolds V, Wallis J. 2012. Chimpanzees in
mantraps: lethal crop protection and conservation in Uganda. Oryx 46:598–603.
McLennan MR, Plumptre AJ. 2012. Protected apes, unprotected forest: composition,
structure and diversity of riverine forest fragments and their conservation value in
Uganda. Trop Conserv Sci 5:79–103.
McLennan MR. 2008. Beleaguered chimpanzees in the agricultural district of Hoima,
western Uganda. Primate Conserv 23:45–54.
McLennan MR. 2010. Chimpanzee ecology and interactions with people in an
unprotected human-dominated landscape at Bulindi, western Uganda. Doctoral
dissertation, Oxford Brookes University, UK.
McLennan MR. 2013. Diet and feeding ecology of chimpanzees (Pan troglodytes) in
Bulindi, Uganda: foraging strategies at the forest–farm interface. Int J Primatol
34:585–614.
McManus JS, Dalton DL, Kotzé A, Smuts B, Dickman A, Marshal JP, Keith M. 2014.
Gene flow and population structure of a solitary top carnivore in a human-dominated
landscape. Ecol Evol 5:335–344.
McShea WJ, Stewart C, Peterson L, Erb P, Stuebing R, Giman B. 2009. The importance
of secondary forest blocks for terrestrial mammals within an Acacia/secondary forest
matrix in Sarawak, Malaysia. Biol Conserv 142:3108–3119.

130

Meijaard E, Albar G, Nardiyono, Rayadin Y, Ancrenaz M, Spehar S. 2010. Unexpected
ecological resilience in Bornean orangutans and implications for pulp and paper
plantation management. PLoS One 5:e12813.
Mikkelsen TS, Hillier LW, Eichler EE, Zody MC, Jaffe DB, Yang S-P, Enard W,
Hellmann I, Lindblad-Toh K, Altheide TK, Archidiacono N, Bork P, Butler J, Chang
JL, Cheng Z, Chinwalla AT, deJong P, Delehaunty KD, Fronick C, Fulton LA, Gilad
Y, Glusman G, Gnerre S, Graves TA, Hayakawa T, Hayden KE, Huang X, Ji H, Kent
WJ, King M-C, Kulbokas EJ III, Lee MK, Liu G, Lopez-Otín C, Makova KD, Man
O, Mardis ER, Mauceli E, Miner TL, Nash WE, Nelson JO, Paabo S, Patterson N,
Pohl C, Pollard KS, Prüfer K, Puente XS, Reich D, Rocchi M, Rosenbloom K,
Ruvolo M, Richter DJ, Schaffner SF, Smit AFA, Smith SM, Suyama M, Taylor J,
Torrents D, Tuzun E, Varki A, Velasco G, Ventura M, Wallis JW, Wendi MC,
Wilson RK, Lander ES, Waterston RH. 2005. Initial sequence of the chimpanzee
genome and comparison with the human genome. Nature 437:69–87.
Miller CR, Joyce P, Waits LP. 2005. A new method for estimating the size of small
populations from genetic mark-recapture data. Mol Ecol 14:1991–2005.
Mitani JC, Watts DP, Amsler SJ. 2010. Lethal intergroup aggression leads to territorial
expansion in wild chimpanzees. Curr Biol 20:R507–R508.
Mitani JC, Watts DP, Lwanga JS. 2002. Ecological and social correlates of chimpanzee
party size and composition. In: Boesch C, Hohmann G, Marchant LF, editors.
Behavioural diversity in chimpanzees and bonobos. Cambridge: Cambridge
University Press. p 102–111.
Moore DL, Langergraber KE, Vigilant L. 2015. Genetic analyses suggest male philopatry
and territoriality in savanna-woodland chimpanzees (Pan troglodytes schweinfurthii)
of Ugalla, Tanzania. Int J Primatol 36:377–397.
Moore DL, Vigilant L. 2013. A population estimate of chimpanzees (Pan troglodytes
schweinfurthii) in the Ugalla region using standard and spatially explicit genetic
capture-recapture methods. Am J Primatol 76:335–346.
Moore DL, Vigilant L. 2014. Genetic diversity at the edge: comparative assessment of Y-
chromosome and autosomal diversity in eastern chimpanzees (Pan troglodytes
schweinfurthii) of Ugalla, Tanzania. Conserv Genet 15:495–507.
Morin PA, Chambers KE, Boesch C, Vigilant L. 2001. Quantitative polymerase chain
reaction analysis of DNA from noninvasive samples for accurate microsatellite
genotyping of wild chimpanzees (Pan troglodytes verus). Mol Ecol 10:1835–1844.

131

Moyer D, Plumptre AJ, Kamenya S, Athumani S, Sikombe S. 2006. Surveys east of
Mahale Mountains National Park. In: Moyer D, Plumptre AJ, Pintea L, Hernandez-
Aguilar A, Moore J, Stewart FA, Davenport TRB, Piel A, Kamenya S, Mugabe H,
Mpunga N, Mwangoka M, editors. Surveys of chimpanzees and other biodiversity in
Western Tanzania. Wildlife Conservation Society, Jane Goodall Institute, University
of California San Diego. p 20–23.
Mubiru DN, Kristjanson P. 2012. Summary of baseline household survey results: Hoima
District, West Central Uganda. Copenhagen, Denmark: CGIAR Research Program on
Climate Change, Agriculture, and Food Security (CCAFS). Available from:
http://ccafs.cgiar.org/resources/baseline-surveys
Mwavu EN, Witkowski ETF. 2008. Land-use and cover changes (1988-2002) around
Budongo Forest Reserve, NW Uganda: implications for forest and woodland
sustainability. Land Degrad Dev 19:606–622.
MWLE. 2000. The national environment (wetlands, river banks, and lake shores
management) regulations, No. 3/2000. Kampala: Ministry of Water, Lands, and
Environment.
MWLE. 2002. The National Forest Plan. Kampala: Ministry of Water, Lands, and
Environment.
Nangendo G, Plumptre AJ, Akwetaireho S. 2010. Identifying potential corridors for
conservation in the Murchison-Semliki Landscape. Kampala: Wildlife Conservation
Society.
Nekaris KAI, Nijman V. 2013. The ethics of conducting field research. In: MacClancy J,
Fuentes A, editors. Ethics in the field: contemporary challenges. Oxford: Berghahn
Books. p 108–123.
Nishida T ed. 1990. The chimpanzees of the Mahale Mountains: sexual and life history
strategies. Tokyo: University of Tokyo Press.
Nishida T, Corp N, Hamai M, Hasegawa T, Hiraiwa-Hasegawa M, Hosaka K, Hunt K,
Itoh N, Kawanaka K, Matsumoto-Oda A, Mitani JC, Nakamura M, Norikoshi K,
Sakamaki T, Turner L, Uehara S, Zamma K. 2003. Demography, female life history,
and reproductive profiles among the chimpanzees of Mahale. Am J Primatol 59:99–
121.
Nishida T, Hiraiwa-Hasegawa M, Hasegawa T, Takahata Y. 1985. Group extinction and
female transfer in wild chimpanzees in the Mahale National Park, Tanzania.
Zeitschrift Fur Tierpsychologie 67:284–301.

132

Nishida T. 1979. The social structure of chimpanzees in the Mahale Mountains. In:
Hamburg DA, McCown ER, editors. The great apes. Menlo Park, CA:
Benjamin/Cummings.
Nsubuga A, Robbins MM, Roeder AD, Morin PA, Boesch C, Vigilant L. 2004. Factors
affecting the amount of genomic DNA extracted from ape faeces and the
identification of an improved sample storage method. Mol Ecol 13:2089–2094.
Nunes S, Zugger PA, Engh AL, Reinhart KO, Holekamp KE. 1997. Why do female
Belding's ground squirrels disperse away from food resources? Behav Ecol Sociobiol
40:199–207.
Oates JF, Tutin CEG, Humle T, Wilson ML, Baillie J, Balmforth Z, Blom A, Boesch C,
Cox D, Davenport TRB, Dunn A, Dupain J, Duvall C, Ellis CM, Farmer KH, Gatti S,
Greengrass E, Hart J, Herbinger I, Hicks TC, Hunt K, Kamenya S, Maisels F, Mitani
JC, Moore J, Morgan BJ, Morgan DB, Nakamura M, Nixon S, Plumptre AJ,
Reynolds V, Stokes EJ, Walsh PD. 2008. Pan troglodytes. The IUCN Red List of
Threatened Species. Version 2015.1. Available from: http://www.iucnredlist.org
Obbard ME, Howe EJ, Kyle CJ. 2010. Empirical comparison of density estimators for
large carnivores. J Appl Ecol 47:76–84.
Okecha A, Newton-Fisher NE. 2006. The diet of olive baboons (Papio anubis) in the
Budongo Forest Reserve, Uganda. In: Newton-Fisher NE, Notman H, Paterson JD,
Reynolds V, editors. Primates of western Uganda. New York: Springer. p 61–74.
Oota H, Settheetham-Ishida W, Tiwawech D, Ishida T, Stoneking M. 2001. Human
mtDNA and Y-chromosome variation is correlated with matrilocal versus patrilocal
residence. Nat Genet 29:20–21.
Paterson JD. 1991. The ecology and history of Uganda's Budongo Forest. For Conserv
Hist 35:179–187.
Pennell MW, Stansbury CR, Waits LP, Miller CR. 2012. Capwire: a R package for
estimating population census size from non-invasive genetic sampling. Mol Ecol Res
13:154–157.
Pepper JW, Mitani JC, Watts DP. 1999. General gregariousness and specific social
preferences among wild chimpanzees. Int J Primatol 20:613–632.
Petit E, Valiere N. 2006. Estimating population size with noninvasive capture-mark-
recapture data. Conserv Biol 20:1062–1073.

133

Plumptre AJ, Akwetaireho S, Leal M, Mutungire N, Kyamanywa J, Tumuhamye D,
Bamutura B, Ayebale J, Ronald M. 2011. Biodiversity surveys of the corridor forests
east of Bugoma Forest Reserve up to Budongo Forest Reserve. Kampala: Wildlife
Conservation Society.
Plumptre AJ, Cox D, Mugume S. 2003. The status of chimpanzees in Uganda. Kampala:
Wildlife Conservation Society.
Plumptre AJ, Cox D. 2005. Counting primates for conservation: primate surveys in
Uganda. Primates 47:65–73.
Plumptre AJ, Davenport T, Behangana M, Kityo R, Eilu G, Ssegawa P, Ewango C,
Meirte D, Kahindo C, Herremans M. 2007. The biodiversity of the Albertine Rift.
Biol Conserv 134:178–194.
Plumptre AJ, Reynolds V. 1996. Censusing chimpanzees in the Budongo Forest, Uganda.
Int J Primatol 17:85–99.
Plumptre AJ, Rose R, Nangendo G, Williamson EA, Didier K, Hart J, Mulindahabi F,
Hicks TC, Griffin B, Ogawa H, Nixon S, Pintea L, Vosper A, McLennan M, Amsini
F, McNeilage A, Makana JR, Kanamori M, Hernandez A, Piel A, Stewart FA, Moore
J, Zamma K, Nakamura M, Kamenya S, Idani G, Sakamaki T, Yoshikawa M, Greer
D, Tranquilli S, Beyers R, Hashimoto C, Furuichi T, Bennett E. 2010. Eastern
chimpanzee (Pan troglodytes schweinfurthii): status survey and conservation action
plan 2010-2020. Gland, Switzerland: IUCN/SSC Primate Specialist Group.
Potts R. 1998. Variability selection in hominid evolution. Evol Anthropol 7:81–96.
Pritchard JK, Stephens M, Donnelly P. 2000. Inference of population structure using
multilocus genotype data. Genet 155:945–959.
Prüfer K, Munch K, Hellmann I, Akagi K, Miller JR, Walenz B, Koren S, Sutton G,
Kodira C, Winer R, Knight JR, Mullikin JC, Meader SJ, Ponting CP, Lunter G,
Higashino S, Hobolth A, Dutheil J, Karakoç E, Alkan C, Sajjadian S, Catacchio CR,
Ventura M, Marques-Bonet T, Eichler EE, André C, Atencia R, Mugisha L, Junhold
J, Patterson N, Siebauer M, Good JM, Fischer A, Ptak SE, Lachmann M, Symer DE,
Mailund T, Schierup MH, Andrés AM, Kelso J, Paabo S. 2012. The bonobo genome
compared with the chimpanzee and human genomes. Nature 486:527–531.
Pusey AE. 1980. Inbreeding avoidance in chimpanzees. Anim Behav 28:543–552.
Pusey AE. 1987. Sex-biased dispersal and inbreeding avoidance in birds and mammals.
Trends Ecol Evol 2:295–299.

134

Quantum GIS. 2014. Quantum GIS Geographic Information System. Open Source
Geospatial Foundation Project. Available from: http://qgis.osgeo.org
R Core Team. 2014. R: a language and environment for statistical computing. Vienna,
Austria: R Foundation for Statistical Computing. Available from: http://www.R-
project.org
Ranta P, Blom T, Niemela J, Joensuu E, Siitonen M. 1998. The fragmented Atlantic rain
forest of Brazil: size, shape and distribution of forest fragments. Biodivers Conserv
7:385–403.
Reynolds V, Wallis J, Kyamanywa R. 2003. Fragments, sugar, and chimpanzees in
Masindi District, western Uganda. In: Marsh L, editor. Primates in fragments. New
York: Kluwer. p 309–320.
Reynolds V. 2005. Chimpanzees of the Budongo Forest. New York: Oxford University
Press.
Ripperger SP, Tschapka M, Kalko EKV, Rodriguez-Herrera B, Mayer F. 2012. Life in a
mosaic landscape: anthropogenic habitat fragmentation affects genetic population
structure in a frugivorous bat species. Conserv Genet 14:925–934.
Robbins MM, Williamson L. 2008. Gorilla beringei. The IUCN Red List of Threatened
Species. Version 2015.2. Available from: http://www.iucnredlist.org
Ross C, Reeve N. 2011. Survey and census methods: population distribution and density.
In: Setchell JM, Curtis DJ, editors. Field and laboratory methods in primatology: a
practical guide. 2nd ed. Cambridge: Cambridge University Press. p 111–132.
Roth G, Dicke U. 2005. Evolution of the brain and intelligence. Trends Cogn Sci 9:250–
257.
Rothman JM, Van Soest PJ, Pell AN. 2006. Decaying wood is a sodium source for
mountain gorillas. Biol Lett 2:321–324.
Roy J, Vigilant L, Gray M, Wright E, Kato R, Kabano P, Basabose A, Tibenda E, Kühl
HS, Robbins MM. 2014. Challenges in the use of genetic mark-recapture to estimate
the population size of Bwindi mountain gorillas (Gorilla beringei beringei). Biol
Conserv 180:249–261.
Römpler H, Dear PH, Krause J, Meyer M, Rohland N, Schöneberg T, Spriggs H, Stiller
M, Hofreiter M. 2006. Multiplex amplification of ancient DNA. Nat Protoc 1:720–
728.

135

Rudicell RS, Holland Jones J, Wroblewski EE, Learn GH, Li Y, Robertson JD,
Greengrass E, Grossmann F, Kamenya S, Pintea L, Mjungu DC, Lonsdorf EV,
Mosser A, Lehman C, Collins DA, Keele BF, Goodall J, Hahn BH, Pusey AE,
Wilson ML. 2010. Impact of simian immunodeficiency virus infection on
chimpanzee population dynamics. PLoS Pathog 6:e1001116.
Schubert G, Stoneking CJ, Arandjelovic M, Boesch C, Eckhardt N, Hohmann G,
Langergraber K, Lukas D, Vigilant L. 2011. Male-mediated gene flow in patrilocal
primates. PLoS One 6:e21514.
Selkoe KA, Toonen RJ. 2006. Microsatellites for ecologists: a practical guide to using
and evaluating microsatellite markers. Ecol Lett 9:615–629.
Sheil D, Meijaard E. 2010. Purity and prejudice: deluding ourselves about biodiversity
conservation. Biotropica 42:566–568.
Singleton I, Wich SA, Griffiths M. 2008. Pongo abelii. The IUCN Red List of
Threatened Species. Version 2015.2. Available from: http://www.iucnredlist.org
Smith JB. 1997. Setting priorities for adapting to climate change. Glob Environ Change
7:251–264.
Soisalo MK, Cavalcanti SMC. 2006. Estimating the density of a jaguar population in the
Brazilian Pantanal using camera-traps and capture–recapture sampling in
combination with GPS radio-telemetry. Biol Conserv 129:487–496.
Stansbury CR. 2012. Monitoring gray wolves (Canis lupus) using noninvasive genetic
sampling at rendezvous sites. Master's thesis, University of Idaho, USA.
Stearns SC. 1989. Evolutionary significance of phenotypic plasticity. BioScience 39:436–
445.
Stratford JA, Stouffer PC. 1999. Local extinctions of terrestrial insectivorous birds in a
fragmented landscape near Manaus, Brazil. Conserv Biol 13:1416–1423.
Stumpf RM, Emery Thompson M, Muller MN, Wrangham RW. 2009. The context of
female dispersal in Kanyawara chimpanzees. Behaviour 146:629–656.
Stumpf RM. 2011. Chimpanzees and bonobos: inter- and intraspecies diversity. In:
Campbell CJ, Fuentes A, MacKinnon KC, Bearder SK, Stumpf RM, editors.
Primates in perspective. Oxford: Oxford University Press. p 340–356.
Sugiyama Y. 1999. Socioecological factors of male chimpanzee migration at Bossou,
Guinea. Primates 40:61–68.

136

Taberlet P, Waits LP, Luikart G. 1999. Noninvasive genetic sampling: look before you
leap. Trends Ecol Evol 14:323–327.
Tagg N, Willie J, Petre C-A, Haggis O. 2013. Ground night nesting in chimpanzees: new
insights from central chimpanzees (Pan troglodytes troglodytes) in south-east
Cameroon. Folia Primatol 84:362–383.
Tranquilli S, Abedi-Lartey M, Abernethy K, Amsini F, Asamoah A, Balangtaa C, Blake
S, Bouanga E, Breuer T, Brncic TM, Campbell G, Chancellor R, Chapman CA,
Davenport TRB, Dunn A, Dupain J, Ekobo A, Eno-Nku M, Etoga G, Furuichi T,
Gatti S, Ghiurghi A, Hashimoto C, Hart JA, Head J, Hega M, Herbinger I, Hicks TC,
Holbech LH, Huijbregts B, Kühl HS, Imong I, Yeno SL-D, Linder J, Marshall P,
Lero PM, Morgan D, Mubalama L, N'Goran PK, Nicholas A, Nixon S, Normand E,
Nziguyimpa L, Nzooh-Dongmo Z, Ofori-Amanfo R, Ogunjemite BG, Petre C-A,
Rainey HJ, Regnaut S, Robinson O, Rundus A, Sanz CM, Okon DT, Todd A, Warren
Y, Sommer V. 2014. Protected areas in tropical Africa: assessing threats and
conservation activities. PLoS One 9:e114154.
Turner IM, Corlett RT. 1996. The conservation value of small, isolated fragments of
lowland tropical rain forest. Trends Ecol Evol 11:330–333.
Turyahabwe N, Banana AY. 2008. An overview of history and development of forest
policy and legislation in Uganda. Int Forest Rev 10:641–656.
Tweheyo M, Lye KA. 2003. Phenology of figs in Budongo Forest Uganda and its
importance for the chimpanzee diet. Afr J Ecol 41:306–316.
Twongyirwe R, Bithell M, Richards KS, Rees WG. 2015. Land Use Policy. Land Use
Policy 49:236–251.
UBOS. 2007. Hoima District 2002 population and housing census analytical report.
Kampala: Uganda Bureau of Statistics.
UBOS. 2014. NPHC 2014 provisional results report. Kampala: Uganda Bureau of
Statistics.
van Andel AC, Wich SA, Boesch C, Koh LP, Robbins MM, Kelly J, Kühl HS. 2015.
Locating chimpanzee nests and identifying fruiting trees with an unmanned aerial
vehicle. Am J Primatol 77:1122–1134.
Van Horn RC, Altmann J, Alberts SC. 2008. Can't get there from here: inferring kinship
from pairwise genetic relatedness. Anim Behav 75:1173–1180.
van Schaik CP. 2013. The costs and benefits of flexibility as an expression of behavioural
plasticity: a primate perspective. Philos Trans R Soc Lond B Biol Sci 368:20120339.
137

Vigilant L, Roy J, Bradley BJ, Stoneking CJ, Robbins MM, Stoinski TS. 2015.
Reproductive competition and inbreeding avoidance in a primate species with
habitual female dispersal. Behav Ecol Sociobiol 69:1163–1172.
Wallis J, Lee DR. 1999. Primate conservation: the prevention of disease transmission. Int
J Primatol 20:803–826.
Walsh PD, Abernethy KA, Bermejo M, Beyers R, De Wachter P, Ella Akou M,
Huijbregts B, Mambounga DI, Toham AK, Kilbourn AM, Lahm SA, Latour S,
Maisels F, Mbina C, Mihindou Y, Obiang SN, Effa EN, Starkey MP, Telfer P,
Thibault M, Tutin CEG, White LJT, Wilkie DS. 2003. Catastrophic ape decline in
western equatorial Africa. Nature 422:611–614.
Walsh PD, Tutin CEG, Oates JF, Baillie J, Maisels F, Stokes EJ, Gatti S, Bergl R,
Sunderland-Groves JL, Dunn A. 2008. Gorilla gorilla. The IUCN Red List of
Threatened Species. Version 2015.2. Available from: http://www.iucnredlist.org
Whiten A, Goodall J, McGrew WC, Nishida T, Reynolds V, Sugiyama Y, Tutin CE,
Wrangham RW, Boesch C. 1999. Cultures in chimpanzees. Nature 399:682–685.
Whiten A. 2000. Primate culture and social learning. Cogn Sci 24:477–508.
Whittingham LA. 2004. Effects of extra-pair and within-pair reproductive success on the
opportunity for selection in birds. Behav Ecol 16:138–144.
Wich SA, Gaveau D, Abram N, Ancrenaz M, Baccini A, Brend S, Curran L, Delgado
RA, Erman A, Fredriksson GM, Goossens B, Husson SJ, Lackman I, Marshall AJ,
Naomi A, Molidena E, Nardiyono, Nurcahyo A, Odom K, Panda A, Purnomo,
Rafiastanto A, Ratnasari D, Santana AH, Sapari I, van Schaik CP, Sihite J, Spehar S,
Santoso E, Suyoko A, Tiju A, Usher G, Atmoko SSU, Willems EP, Meijaard E.
2012. Understanding the impacts of land-use policies on a threatened species: is there
a future for the Bornean orang-utan? PLoS One 7:e49142.
Wiens J. 1996. Wildlife in patchy environments: metapopulations, mosaics, and
management. In: McCullough DR, editor. Metapopulations and wildlife
conservation. Washington, D.C.: Island Press. p 53–84.
Williams JM, Pusey AE, Carlis JV, Farm BP, Goodall J. 2002. Female competition and
male territorial behaviour influence female chimpanzees' ranging patterns. Anim
Behav 63:347–360.

138

Wilson ML, Boesch C, Fruth B, Furuichi T, Gilby IC, Hashimoto C, Hobaiter CL,
Hohmann G, Itoh N, Koops K, Lloyd JN, Matsuzawa T, Mitani JC, Mjungu DC,
Morgan D, Muller MN, Mundry R, Nakamura M, Pruetz J, Pusey AE, Riedel J, Sanz
C, Schel AM, Simmons N, Waller M, Watts DP, White F, Wittig RM, Zuberbuhler
K, Wrangham RW. 2014. Lethal aggression in Pan is better explained by adaptive
strategies than human impacts. Nature 513:414–417.
Wrangham R, Wilson M, Hare B. 2000. Chimpanzee predation and the ecology of
microbial exchange. Microb Ecol in Health and Dis 12:186–188.
Wrangham RW. 1977. Feeding behavior of chimpanzees in Gombe National Park,
Tanzania. In: Clutton-Brock TH, editor. Primate ecology. London: Academic Press. p
503–538.
Wrangham RW. 1979. Sex differences in chimpanzee dispersion. In: Hamburg DA,
McCown ER, editors. The great apes. Menlo Park, CA: Benjamin/Cummings. p 480–
489.
Wroblewski E, Norman PJ, Guethlein LA, Rudicell RS, Ramirez MA, Li Y, Hahn BH,
Pusey AE, Parham P. 2015. Signature patterns of MHC diversity in three Gombe
communities of wild chimpanzees reflect fitness in reproduction and immune defense
against SIVcpz. PLoS Biol 13:e1002144.
WWF. 2014. Living planet report 2014. Gland, Switzerland: WWF International.
Zhan X, Li M, Zhang Z, Goossens B, Chen Y, Wang H, Bruford MW, Wei F. 2006.
Molecular censusing doubles giant panda population estimate in a key nature reserve.
Curr Biol 16:R451–R452.

139

Appendix A
Variability and Amplification Success for Microsatellite Loci in this Research
Microsatellite locus No. alleles No. individuals typed
Autosomal

D1s1622* 8 180
D12s66* 11 181
D18s536* 7 174
D1s1656 14 189
D2s1326 10 182
D3s2459 9 193
D3s3038 9 183
D4s1627 9 179
D5s1457 8 195
D5s1470 9 188
D7s817 9 188
D7s2204 8 179
D10s676 7
188
D11s2002 7
188
Y-chromosome

DYs439 2 74
DYs469 3 76
DYs510 2 74
DYs517 2 76
DYs520 4 76
DYs588 2 76
DYs612 5 76
DYs630 3 76
(DYs392) 1 N/A
(DYs502) 1 N/A
(DYs533) 1 N/A
(DYs562) 1 N/A
(DYs632) 1 N/A

Note. Asterisks indicate loci included in the single-step test multiplex. Y-chromosome
loci indicated in parentheses were tested but were not variable and thus were not used
further.

Appendix B
Primer Sequences and Annealing Temperatures for Autosomal and Y-Chromosome Microsatellite
Loci in this Research

Autosomal Microsatellite Loci
Locus Forward primer 5'-3' Reverse primer 5'-3' Ta, °C
D10s676 F/R GAGAACAGACCCCCAAATCT ATTTCAGTTTTACTATGTGCATGC 55
D11s2002 F/R2 CATGGCCCTTCTTTTCATAG AGTGTGAGCCACCACACCAGC 60
D12s66 F/R TCATTTAAGCATTTGAGGGAA AGACTTCAAAACAGACACTT 55
(D14s306 F/R) AAAGCTACATCCAAATTAGGTAGG TGACAAAGAAACTAAAATGTCCC 55
(D16s2624 F/R) TGAGGCAATTTGTTACAGAGC TAATGTACCTGGTACCAAAAACA 55
D18s536 F/R ATTATCACTGGTGTTAGTCCTCTG CACAGTTGTGTGAGCCAGTC 55
D1s1622 F/R CCCTCTGTCTCCAGCTGTAA TCACCCTCACATGATGCC 55
D1s1656 F/R GTGTTGCTCAAGGGTCAACT GAGAAATAGAATCACTAGGGAACC 55
D2s1326 F/R2 AGACAGTCAAGAATAACTGCCC CTGTGGCTCAAAAGCTGAAT 60
(D2s1329 F2/R4) ACCGTTCTCAAATACCAGGAATC CCTGGGTTCTTAATTTAACCATAATTC 55
140

D3s2459 F/R CTGGTTTGGGTCTGTTATGG AGGGACTTAGAAAGATAGCAGG 60
D3s3038 F/R CATCTTTCTTTTCCTGTTCCC GATACCATATTCAACATGAAGAGG 55
D4s1627 F/R AGCATTAGCATTTGTCCTGG GACTAACCTGACTCCCCCTC 62
D5s1457 F/R TAGGTTCTGGGCATGTCTGT TGCTTGGCACACTTCAGG 55
D5s1470 F/R CATGCACAGTGTGTTTACTGG TAGGATTTTACTATATTCCCCAGG 60
(D6s1056 F/R) ACAAGAACAGCATGGGGTAA CCTGGATCATGAATTGCTAT 56
D7s2204 F/R4 TCATGACAAAACAGAAATTAAGTG AGTAAATGGAATTGCTTGTTACC 55
D7s817 F2/R TTGGGACCTCTTATTTTCCA TAAATCTCTTTATGGCTGACTG 60
(D9s910 F/R) AAGTCAGTTAGCTGAAGGTTGC TATATGAAGTGCTTAGAAAAAGTGC 60
Y-Chromosome Microsatellite Loci
Locus Forward primer 5'-3' Reverse primer 5'-3' Ta, °C
(DYs392) TCATTAATCTAGCTTTTAAAAACAA AGACCCAGTTGATGCAATGT 58
DYs439 TCCTGAATGGTACTTCCTAGGTTT GCCTGGCTTGGAATTCTTTT 58
DYs469 TTTGGGGACTGAATTCAAAA CCCCAGCTGGTAAAATGAGT 58
(DYs502) CAGCAAGCCACCATACCATA TGTGCTTTTGGAGTTTGGAG 58
DYs510 TTTTTCCTCCCTTACCACAGA TCTGGAGAAGACAGAACTTGTCA 60 141

DYs517 TAATCGTCCCATTTTGAGCA TGCAATCCCAAACTCAGAAA 60
DYs520 AACAGCCTGCCCAACATAGT ACCATCATGCCCTGCAATA 58
(DYs533) CATCTAACATCTTTGTCATCTACC TGATCAGTTCTTAACTCAACCA 58
(DYs562) GGGTGTATAAAGAGGGGCATA GGTAAAGGTTATCACGCCATC 60
DYs588 GAATGCAGAACCCTCAAGGA AGCCTGGGTGACAGAAACAC 60
DYs612 CCCCCATGCCAGTAAGAATA TGAGGGAAGGCAAAAGAAAA 58
DYs630 GCCTTTGGACAGAGCAAGAC AGCCATGGAAAGCTGTGAGT 65
(DYs632) GGCCGTTGCAAAATAAACTG TCTGGGCAACAGAAGGCGAC 60
Note. Ta refers to the annealing temperature used in the PCR for that locus, as described in detail in Chapter 2. Parentheses
indicate autosomal microsatellite loci used only in the multiplex PCR (see Section 2.5 above) and Y-chromosome
microsatellite loci used only in an initial test of locus variability (see Section 2.6 above).

142

Appendix C

Alleles and Allele Ranges for Autosomal Microsatellite Loci in this Research

Locus Group 1 (Screening test loci)

Amelogenin D1s1622 D18s536 D12s66
Allele Allele Range Allele Allele Range Allele Allele Range Allele Allele Range
104 103.27 - 105.00 229 227.63 - 229.56 130 129.09 - 130.68 133 131.67 - 134.88
110 109.34 - 110.55 238 236.64 - 238.59 134 133.12 - 135.36 137 136.90 - 138.17

243 241.65 - 243.94 138 137.07 - 138.94 141 139.73 - 142.30

246 244.52 - 247.44 146 145.16 - 146.94 145 144.57 - 146.45

249 248.27 - 249.89 150 149.05 - 150.74 149 148.04 - 149.79

252 251.30 - 253.37 154 153.33 - 154.95 153 152.39 - 154.45

255 254.31 - 255.94 170 170.75 - 171.30 157 155.32 - 157.93

261 260.00 - 261.95

161 160.92 - 161.60

169 169.04 - 169.81

177 175.95 - 178.37

181 181.48 - 182.59

143

Locus Group 2

D5s1457 D10s676 D5s1470 D4s1627
Allele Allele Range Allele Allele Range Allele Allele Range Allele Allele Range
102 99.78 - 101.39 136 136.26 - 137.57 165 165.44 - 165.70 218 216.76 - 217.73
106 104.80 - 106.99 153 152.44 - 154.17 173 171.80 - 172.89 222 220.66 - 222.13
110 108.93 - 109.63 157 156.04 - 158.17 177 176.03 - 178.28 226 224.83 - 225.64
114 112.64 - 113.90 161 159.77 - 162.31 181 178.53 - 182.40 230 228.82 - 229.96
118 117.68 - 119.09 165 164.97 - 167.01 185 184.44 - 185.39 234 232.96 - 234.82
122 121.72 - 123.33 169 168.17 - 170.67 189 188.72 - 189.48 238 237.02 - 238.11
126 125.85 - 127.53 173 172.46 - 174.75 193 191.67 - 193.68 242 241.30 - 242.77
130 129.98 - 131.52

197 195.81 - 197.29 246 245.36 - 245.47

205 204.33 - 204.71 250 249.24 - 251.12

Locus Group 3

D7s817 D11s2002 D3s3038 D3s2459
Allele Allele Range Allele Allele Range Allele Allele Range Allele Allele Range
102 99.03 - 101.40 129 128.11 - 128.26 174 173.49 - 173.95 169 167.70 - 168.29
106 105.11 - 105.69 133 132.24 - 133.48 178 177.67 - 178.29 176 175.18 - 176.13
110 108.73 - 110.14 137 136.28 - 137.67 182 180.67 - 182.53 184 182.27 - 184.92
114 113.71 - 114.35 141 140.73 - 141.76 186 185.69 - 186.58 188 186.49 - 188.53
118 117.99 - 118.72 145 145.14 - 145.83 187 186.92 - 187.85 192 191.94 - 192.19
122 122.41 - 122.85 149 149.18 - 149.71 190 189.67 - 190.57 196 195.66 - 196.21
126 126.70 - 127.00 153 153.29 - 153.75 191 190.90 - 191.56 204 201.86 - 202.44
130 130.36 - 130.85

194 193.87 - 194.59 208 206.79 - 207.43
134 134.49 - 135.52

198 197.85 - 198.56 220 218.25 - 219.18

144

Locus Group 4

D7s2204 D2s1326 D1s1656
Allele Allele Range Allele Allele Range Allele Allele Range
144 143.70 - 144.33 182 181.51 - 182.11 106 104.33 - 105.18
147 147.76 - 148.35 190 189.34 - 191.33 118 117.09 - 117.77
164 164.19 - 164.79 194 193.87 - 194.63 130 129.51 - 130.29
168 168.28 - 169.01 198 198.07 - 198.74 134 133.64 - 134.16
172 172.23 - 173.15 199 199.11 - 199.78 138 137.35 - 138.22
176 176.28 - 177.14 203 202.10 - 204.02 140 139.59 - 140.23
180 179.78 - 181.27 207 206.17 - 208.08 142 141.69 - 142.28
184 184.47 - 185.18 211 211.32 - 212.14 144 143.82 - 144.07

215 215.30 - 216.16 146 145.67 - 146.30

219 218.26 - 220.21 148 147.16 - 147.95

150 149.67 - 150.76

152 151.91 - 153.10

154 154.33 - 154.39

158 158.04 - 158.60

Note. For each autosomal microsatellite locus, alleles and observed allele ranges are provided based on the data set used in this
research. The loci in this appendix are displayed according to how they were grouped in the PCR procedure. Locus groups
were arranged to avoid overlap among HEX, NED, and FAM labeled primers. Loci in Locus Group 1 were used in the one-
step screening PCR described in Chapter 2. Extracts that amplified successfully in Group 1were then amplified in a two-step
multiplex-singleplex PCR at the loci listed in Locus Groups 2-4. Further details of this amplification procedure are provided in
Chapter 2.
145

146

Appendix D
Allele Frequency Data for Corridor Genotypes
Locus k N HObs HExp NE-SI HW F(Null)
D18s536 7 174 0.661 0.725 0.419 NS 0.0413
D12s66 11 181 0.834 0.816 0.358 NS -0.012
D1s1622 8 180 0.511 0.575 0.516 NS 0.0612
D5s1457 8 195 0.744 0.742 0.407 NS -0.0008
D10s676 7 188 0.771 0.801 0.367 NS 0.0171
D5s1470 9 188 0.75 0.783 0.38 NS 0.0196
D4s1627 9 179 0.832 0.838 0.344 NS 0.0002
D7s817 9 188 0.601 0.699 0.438 NS 0.0726
D11s2002 7 188 0.745 0.69 0.443 NS -0.0444
D3s3038 9 183 0.71 0.74 0.409 NS 0.0155
D3s2459 9 193 0.539 0.601 0.496 NS 0.0502
D1s1656 14 189 0.788 0.833 0.346 NS 0.0283
D7s2204 8 179 0.844 0.843 0.341 NS -0.0014
D2s1326 10 182 0.775 0.827 0.35 NS 0.0316

Note. Loci listed are all autosomal microsatellite loci used in this study. K refers to the
number of alleles at each locus. N refers to the number of individuals typed at that locus.
HObs and HExp refer to the observed and expected heterozygosity for each locus. NE-SI
refers to the non-exclusion probability for the identity of siblings, as described in
Marshall et al. (1998). HW refers to Hardy-Weinberg Equilibrium, with “NS” indicating
that the Hardy-Weinberg test was not significant and the locus is in equilibrium. F(Null)
refers to the null allele frequency value.

Appendix E

Y-Chromosome Haplotypes and Their Occurrences

Haplotype DYs630 DYs510 DYs612 DYs517 DYs439 DYs588 DYs469 DYs520
Putative
communities
No.
males
A 158 171 199 191 234 185 213 156 Mukihani 1
B 158 171 205 191 234 185 213 156
Kyamuchumba,
Kiryangobe,
Mukihani
12
C 158 171 205 195 234 185 213 156 Mukihani 1
D 162 171 205 187 234 185 210 156 Kasongoire 7
E 162 171 205 187 234 185 213 156 Katanga 2
F 162 171 205 187 234 185 213 160 Katanga 2
G 162 171 208 187 234 185 210 156 Kiraira, Kasongoire 8
H 162 171 208 187 234 185 213 156 Busingiro 5
I 162 171 208 191 234 185 213 152 Mukihani, Kasokwa 8
J 162 171 208 191 234 185 216 152 Kasokwa 4
K 162 179 192 187 234 175 213 148 Wagaisa 1
L 162 179 196 187 234 175 213 152 Wagaisa 15
M 166 171 205 187 234 185 213 156 Bulindi, Katanga 8
N 166 171 205 191 230 185 213 152 Kityedo 2
Total

76

Note. The 14 Y-chromosome haplotypes found in this research are listed on the left, according to arbitrary designations by
letter. The 8 Y-chromosome microsatellite loci used to determine these haplotypes are listed in the columns, with
corresponding alleles provided. The putative communities and number of males exhibiting a given haplotype are also provided.
147

148

Appendix F

Subadult and Adult Composition of Long-Term Chimpanzee Study
Communities

Chimpanzee
Community, location
Mean
community size
Adults and
subadults (%)
References
M Group,
Mahale Mountains
National Park, Tanzania
79.1 66.2 Nishida et al. (2003)
Bossou, Guinea 19.7 55.8 Sugiyama (2004)
Sonso, Budongo Forest,
Uganda
52.1 60.7 Reynolds (2005)
Kasakela, Gombe
National Park, Tanzania
48.1 64.4 Goodall (1983)
Taї National Park,
Ivory Coast
61.1 57.8
Boesch and Boesch-
Achermann (2000)
MEAN

61.0

Bulindi 19 57.9 M. McLennan, pers. comm.

Note. Community sizes were calculated as a mean value over study years as reported in
the cited literature. For Bulindi, in the Budongo-Bugoma corridor, group structure and
composition is known from behavioral observations since 2007. Bulindi’s demographic
structure during the study period was provided by M. McLennan.

Abstract (if available)

Abstract Habitat loss and fragmentation pose growing challenges to wildlife globally. Great apes, our closest living relatives, are threatened with extinction, with habitat loss and fragmentation acting as key drivers in their decline. Growing proportions of great ape populations live outside protected areas, in habitats that are often fragmented. It is essential to understand how these species respond to habitat changes to better predict future impacts and devise appropriate conservation strategies. Despite the survival risks posed by fragmented habitats, such environments can also offer resources and act as corridors linking continuous habitat. Therefore, a realistic approach to research and conservation in the Anthropocene must take into account the value and potential of such habitats to great apes and other wildlife. ❧ The fragmented forest habitat between the Budongo and Bugoma Forests in Uganda has undergone substantial changes in recent decades, with the conversion of this riparian forest‐grassland mosaic to accommodate the needs of growing human population. Though previous research confirmed chimpanzee presence in this region, the size and distribution of this population has remained little understood. Further, data have been lacking to examine whether this habitat acts as a corridor for chimpanzees between the Budongo and Bugoma Forests, each of which contains hundreds of chimpanzees. ❧ This research employed a genetic approach to investigate chimpanzee distribution and dispersal in this fragmented forest landscape. I noninvasively collected chimpanzee fecal samples to study this population’s size and distribution using mark‐recapture methods. I also investigated the distribution of co‐sampled genotypes to establish the locations of putative chimpanzee communities in the study region. I further examined the distribution of distinct Y‐chromosome haplotypes among male chimpanzees in the study area, since these can be used as markers of community affiliation in territorial, male‐philopatric chimpanzees. I found a minimum chimpanzee population of 182 individuals based on the number of unique genotypes, and estimated the total population size at approximately 260-320 individuals, depending on the estimator used. ❧ A widely distributed chimpanzee population in this corridor habitat is not necessarily sufficient to ensure gene flow, however. If habitat fragmentation isolates remaining communities, female dispersal—a typical inbreeding avoidance mechanism in chimpanzees—may be inhibited. I used genetic data on parent‐offspring relationships and community affiliations to examine dispersal patterns in this landscape. I found a pattern of widespread dispersal characterized by mother‐daughter dyads affiliated with distinct communities. Father‐son dyads, in contrast, were almost always affiliated with the same community, confirming a pattern of male philopatry as would be expected for chimpanzees in continuous habitats. I additionally found evidence of a female dispersal event during the study period. ❧ Together, these findings suggest chimpanzees maintain species‐typical social structure and dispersal patterns in a highly human‐modified landscape, at least in the short term. This underscores the conservation potential of corridor habitats and the potential benefits of corridor restoration and enhancement. Conservation measures are urgently needed, however, given the rapid rate of habitat change and the likelihood of chimpanzee extinction despite marked adaptability to habitat alterations.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Social dynamics of intragroup aggression and conflict resolution in wild chimpanzees (Pan troglodytes) at Kanyawara, Kibale National Park, Uganda

PDF

The many ways of growing old: aging, activity patterns, and function in modern foragers

PDF

Evolution of body size and sexually dimorphic traits in North American Gopherus tortoises

PDF

Behavioral ecology of the Sichuan snub-nosed monkey (Rhinopithecus roxellana) in Shennongjia, China

PDF

A spatio-temporal analysis of the metabolizable energy intake of the Tana River red colobus (Procolobus rufomitratus rufomitratus)

Asset Metadata

Creator McCarthy, Maureen Sophia (author)

Core Title A genetic investigation of chimpanzee distribution and dispersal in the fragmented Budongo-Bugoma Corridor Landscape, Uganda

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Integrative and Evolutionary Biology

Publication Date 02/19/2016

Defense Date 01/14/2016

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag chimpanzee,conservation genetics,habitat fragmentation,OAI-PMH Harvest,Pan troglodytes

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Stanford, Craig (committee chair), McNitt-Gray, Jill (committee member), Seaman, Gary (committee member)

Creator Email m.sophia.mccarthy@gmail.com,msmccart@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-210457

Unique identifier UC11278162

Identifier etd-McCarthyMa-4114.pdf (filename),usctheses-c40-210457 (legacy record id)

Legacy Identifier etd-McCarthyMa-4114.pdf

Dmrecord 210457

Document Type Dissertation

Format application/pdf (imt)

Rights McCarthy, Maureen Sophia

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA