University of Southern California Dissertations and Theses

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

(USC Thesis Other)

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

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

SOCIAL DYNAMICS OF INTRAGROUP AGGRESSION AND CONFLICT
RESOLUTION IN WILD CHIMPANZEES (PAN TROGLODYTES) AT
KANYAWARA, KIBALE NATIONAL PARK, UGANDA

by

Jessica Andrea Hartel

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)

August 2015

Copyright 2015 Jessica Andrea Hartel

ii
Epigraph

“Only if we understand can we care. Only if we care will we help.
Only if we help shall they be saved.”
- Jane Goodall

iii

Dedication

In memory of my mom, Linda Hartel, who always believed in my dreams and me.
Even in your absence, I am still inspired by the woman and mother you were.

For my dad, Lane Hartel, who pushed me to work hard and developed my ambition.

For my sister, Elizabeth Hartel, who sparked an unsaid competition and was always
willing to help when I needed to solve a problem.

For my husband, Matthew Barth, who not only encouraged my dreams, but also willingly
embarked on the journey with me. Thank you for loving chimps, and for putting up with
the good, the bad, and my crazy.

iv
Acknowledgements

Firstly, I would like to thank Dr. Craig Stanford for his understanding and
advisement throughout this process. I would also like to extend my gratitude to Dr.
Martin Muller for his valuable insights, comments on this manuscript, and mentorship.
To Dr. Jill McNitt-Gray and Dr. Justin Wood for their participation on my qualifying and
defense committees and their valuable feedback and support. I would also like to thank
Dr. Richard Wrangham for his support, guidance, and continued mentorship.
Secondly, a huge thank you to Patrick Tusiime for entering some of my data and
to Pawel Fedurek for his companionship in the field as well as the Kibale Chimpanzee
Project field assistants. To my wonderful friend and colleague, Maureen McCarthy, thank
you for all of the support and manuscript comments. To my fellow graduate students,
Rob O’Malley, Angela Garbin, Laura Loyola, and Andy Fogel for their friendship and
advice over the years. To Alex Georgiev for his help with linear hierarchy calculations
and valuable comments.
Thirdly, I would like to thank the Uganda Wildlife Authority and the Uganda
National Council of Science and Technology for their approval and support. Thank you
to the Makerere University Biological Field Station for providing me with a home away
from home and the Kibale Chimpanzee Project for accepting me into their research
family. Thank you to the Wenner-Gren Foundation (Grant # 8249), the International
Society of Primatology, and the University of Southern California for all of the financial
support and project funding throughout my dissertation.

v
Lastly, I would like to thank my life-long support system, my family, who have
always believed in me and supported my lifelong dream. To my Mom for never doubting
me and never letting me doubt myself, which, as a result, has made me into a stronger
person. I aspire to be half the woman that she was. To my Dad for being a great father
that has instilled in me a sense of duty, hard work, and good morals, while always
remaining in good humor. To my sister for always challenging me whether she means to
or not and for never being one step behind, but always right there beside me. And for her
witty sarcasm which always makes me laugh. To my husband, Matthew Barth, for his
encouragement and willingness to embark on many adventures.
Finally, I would first like to acknowledge those that have forever touched my
heart and will continue to influence my life’s path and values henceforth: to the
Kanyawara chimpanzees for allowing me to be a “part of the party” for a short while.
It is, therefore, important for me to emphasize that this thesis is a product of not
one, but many, as all of the above mentioned in one fashion or another have directly
influenced my life and this manuscript, thereby helping me rise to the challenge.
Thank you!

vi

Table of Contents

EPIGRAPH ii

DEDICATION iii

ACKNOWLEDGEMENTS iv

LIST OF TABLES ix

LIST OF FIGURES x

ABSTRACT xi

CHAPTER 1: GENERAL INTRODUCTION 1
General Significance of Study 2
Intragroup Aggression 3
The Relational Model 3
Wild Chimpanzee Intragroup Aggression 5
Conflict Resolution Strategies 7
Reconciliation: Form and Function 8
Consolation: Form and Function 13
Other Third-Party Interactions 19
Measures of Stress and Distress in Primates 23
Self-Directed Behaviors 23
Distress and Recruitment Vocalizations 25
Conflict Resolution in Chimpanzees 27
Behavioral Patterns 28
Project Aims 33

CHAPTER 2: GENERAL METHODS 35
Study Site 36
Study Community and Subjects 38
Behavioral Data Collection 39
Locating Chimpanzees 39
Recording Tools and Equipment 40
Focal (Baseline) Observations 41
Aggression Observations 43
Post-Conflict (PC) Observations 45
Self-Directed Behaviors (SDBs) 46
Data Analysis and Statistical Procedures 48
Proximity Association 50
Contact Affiliation 52

vii
Agonistic Support 53
Aggression Level 54
Standardization of Indices 55
Reconciliation 58
Bystander Affiliation (Consolation and Appeasement) 63
Multivariate Analyses 65

CHAPTER 3: AGONISTIC AND PREFERRED SOCIAL RELATIONS AMONG
KANYAWARA CHIMPANZEES (PAN TROGLODYTES): REVISITED AND
EXPANDED 70
Introduction 71
Cooperation and Valuable Relationships 74
Aggression Between PSPs 76
Methods 77
Results 80
Preferred Social Partners 80
Dominance Relationships 82
Directed Aggression 84
Predictors of Conflict Variation 95
Discussion 105
Preferred Social Partners and Aggression 105
Maximizing the Benefits of Aggression 109
Minimizing the Costs of Aggression 115
Conclusion 119

CHAPTER 4: CONFLICT AND RELATIONSHIP DETERMINANTS OF THE
OCCURRENCE AND VARIABILITY IN RECONCILIATION IN WILD
CHIMPANZEES (PAN TROGLODYTES) 121
Introduction 122
Reconciliation 122
Proximity and Reconciliation 126
Valuable Relationships and Reconciliation 127
Variation in the Components of Reconciliation 129
Methods 131
Study Site and Subjects 131
Data Collection 132
Operational Definitions 133
Statistical Analyses 133
Results 136
Reconciliation 137
Predictors of Reconciliation 139
Reconciliation Variation 143
Predictors of Reconciliation Variation 144
Discussion 152

viii
Predictors and Function of Reconciliation 153
Predictors and Function of Reconciliatory Variation 159
Conclusion 168

BIBLIOGRAPHY 169

APPENDICES 194
Appendix A: Kanyawara Community Demographic Information 194
Appendix B: Behavioral Ethogram 196
Appendix C: Operational Definitions 204

ix

List of Tables

Table 2.1. Kanyawara focal demographic information. 40
Table 3.1. Sex differences in proximity association, agonistic support,
and contact affiliation strength. 81

Table 3.2. Dyads classified as mutual PSPs for proximity association,
agonistic support, and contact affiliation. 82

Table 3.3. Male linear dominance rank. 84
Table 3.4. Victims of directed aggression by sex in adult chimpanzees. 85
Table 3.5. Rates of aggression, indicating number of PSP dyads, by
dyad sex combination. 86

Table 3.6. Frequency of conflicts involving adult females when
accounting for parity, estrous state, and role in aggression. 87

Table 3.7. Context of directed aggression by sex at Kanyawara in 2011
compared to data in 1998 from Muller (2002). 88

Table 3.8. Frequency of conflicts between PSPs for measures of
proximity association, agonistic support, and contact
affiliation by opponent sex combination. 89

Table 3.9. Frequency of conflicts between mutual PSPs based on
overall strength of the opponents’ relationship, categorized
by opponent sex combination. 92

x

List of Figures

Figure 2.1. Map depicting Kibale National Park in Uganda, noting location
of the Kanyawara chimpanzee community (modified from
Chapman et al., 2002). 36

Figure 3.1. Mean rate of aggression relative to opponent PSP type/strength
and sex. 90

Figure 3.2. Mean rate of aggression accounting for overall mutual PSP
strength and opponent dyadic sex combination. 93

Figure 3.3. Influence of conflict type on overall conflict duration. 104

Figure 4.1. Percent of reconciled dyads across 10-min post-conflict period. 137

Figure 4.2. Mean conciliatory tendency (CCT) across sex composition. 138

Figure 4.3. Overall affiliation rates in post-conflict (PC) period compared to
three overall baseline observation rates, varying by distance of
opponents, for reconciled dyads. 139

Figure 4.4. Effect of victim’s sex and estrous state on likelihood of
aggressor or victim approach to reconcile. 146

Figure 4.5. Proportional relationship between the approacher and initiator
of reconciliation. 147

Figure 4.6. Percent of reconciled dyads across duration of reconciliation. 151

xi

Abstract

Peacemaking strategies are often necessary to mitigate aggression and maintain
group stability in social animals. While this topic of behavior has been extensively
studies in captive primate populations, little is known about the post-conflict behaviors of
their wild counterparts. In particular, wild chimpanzees have become well-known for
their high rates of intragroup aggression relative to other animal species and even humans
making them a model candidate in which to study aggression mitigation strategies.
Captive chimpanzees, like humans, use reconciliation and consolation as mechanisms to
mitigate the negative effects of aggression. However, the fission-fusion social structure of
wild chimpanzees makes it difficult to extrapolate captive results to wild populations. In
particular, distancing strategies following aggression may play a key role in wild post-
conflict behavior. This research investigated the post-conflict behaviors of wild
chimpanzees at Kanyawara. In a one-year period, 632 aggression post-conflict
observations between 181 dyads were collected and analyzed.
While the Kanyawara chimpanzees exhibited wide variability in directed
aggression, there were distinct patterns to their choices during the conflict. In general,
male chimpanzees directed aggression most often towards parous females. Parous
females directed their aggression most often towards nulliparous females. Conflicts were
also influenced by social relationships. Male dyad preferred social partners engaged in
more aggression than one-sided and neutral partners, whereas female dyad preferred
social partners engaged in less aggression than neutral partners. Mixed-sex dyad partners
engaged in more aggression only if they were mutual contact preferred social partners. In

xii
general, the chimpanzees appeared to adjust their responses to their opponent based on
preceding behaviors within the conflict bout and intrinsic attributes of their opponent,
quickly weighing the various costs and benefits of each situation to determine which
mechanisms were more effective at managing the ongoing cost-benefit battle.
Furthermore, the potential long-term costs to opponent relationships after aggression
could be further diminished if conciliatory behaviors followed.
After a fight, opponents can chose whether to reconcile, renew aggression,
redirect aggression, or avoid one another. In the course of this study, reconciliation
occurred 123 times between 57 dyads. The Kanyawara chimpanzees had a dyadic
corrected conciliatory tendency of 14.4% and a group corrected conciliatory tendency of
15.7%. Overall, the Kanyawara chimpanzees had a corrected conciliatory tendency
similar to other wild populations (15-19%), but notably lower than their captive
counterparts (22-48%). Male-male dyads had the highest conciliatory tendency (23%),
followed by male-female (14%) and female-female (3%) dyads. Generalized linear mixed
models determined opponent distance was the best predictor of reconciliation. Opponents
within 10 meters after the conflict were more likely to reconcile than those beyond 10
meters (p<0.001), indicating distancing strategies influence reconciliation. The next best
predictor of reconciliation was the social relationship between opponents. Mutual
preferred social partners were significantly more likely to reconcile than neutral partners
(p<0.001). If the aggressor was an overall preferred social partner of the victim, this was
an even better predictor of reconciliation (p<0.001), providing support for the valuable
relationship hypothesis.

1
Chapter 1
General Introduction

2
General Significance of Study
Chimpanzees are well known for their high rates of intragroup aggression. Boehm
(1999) hypothesized that chimpanzees had higher rates of non-lethal intragroup
aggression than most human populations. Wrangham et al. (2006) found support for
Boehm’s hypothesis in chimpanzees from five different communities, including long-
term research sites like Kanyawara (Kibale) and Kasekela (Gombe). The median rates of
non-lethal intragroup aggression for these five chimpanzee populations were 182 times
(for females) and 384 times (for males) higher than among Australian aboriginals (human
data from Burbank, 1992). With such high aggression rates in chimpanzee populations
relative to human, strategies to mitigate aggression and maintain group cohesion should
be fundamentally important. Many studies have shown that humans [and other primates]
frequently use peacemaking behaviors, like reconciliation and consolation, following
aggressive conflicts (Butovskaya et al., 2000; Ljungberg et al., 2005; Sackin & Thelen,
1984; Verbeek & de Waal, 2001). It is logical to assume peacemaking behaviors should
also play an important role in chimpanzee societies.
Many studies have attempted to evaluate these peacekeeping behaviors in
chimpanzees, yet more work needs to be done. Koyama and Palagi (2006) stressed the
importance of non-human primate conflict resolution research and suggested more
research should be conducted to better understand peacekeeping in primates. Kutsukake
and Castles (2004) argue that the wide intra-specific variation in chimpanzee social
behavior (Boesch et al., 2002) calls for more studies on social behavior from various
sites. Wrangham (2008) emphasized the need for continued research in chimpanzee

3
behavioral ecology and stressed the importance of human and non-human primate
comparisons. Over the past decade, behavioral ecology publication numbers have
suffered a decline. The research trend has shifted more toward genetics, cognition, and
physiology (Durham, 2006). While these studies are also valuable, Wrangham (2008)
argues that behavioral ecology studies are imperative for two predominant reasons: 1) to
better understand the evolutionary roots of human social behavior and 2) for the
conservation of the study species. Toward that end, the broad aims of this wild
chimpanzee behavioral ecology study were to better understand: 1) aggression and
conflict resolution behaviors (i.e., reconciliation and consolation) in wild chimpanzees, 2)
what contributes to their peacemaking decisions, 3) the proximate functions of these
behaviors, and 4) the evolutionary significance of the conflict resolution behaviors that
chimpanzees and humans share.
Intragroup Aggression
The Relational Model: Cost and Benefits of Aggression in Gregarious Species
While social living provides many cooperative benefits (i.e., predator protection,
group resource monopolization, etc.), intragroup aggression becomes an inevitable
component of these interactions (Stanford, 2002; van Hooff & van Schaik, 1992; Walters
& Seyfarth, 1987). Group members fight 1) to establish and maintain dominance and 2)
over access to limited and/or valuable resources (e.g., food, mates, social partners, etc.).
While aggression may be necessary and beneficial in terms of resource acquisition or
status, it can be quite costly physically (i.e., injury and energy expenditure; Neat et al.,
1998), hormonally (i.e., stress hormones increase; Crockford et al., 2008; Sapolsky, 1995;

4
Wallner et al., 1999), and also socially (i.e., damaging valuable relationships; de Waal,
1996; Wittig & Boesch, 2003a). These costs disturb levels of tolerance and cohesion
within the group (Cords, 1992; Wittig & Boesch, 2005). To optimize aggression,
opponents must carefully and quickly evaluate each potential conflict in order to
maximize their benefits and minimize their costs.
The Relational Model, introduced by de Waal (1996), and recently tested by
Wittig and Boesch (2003b), addresses the cost-benefit ratio of within group competition
through conflict resolution strategies. When group members share a conflict of interest
(i.e., incompatible goals in a competitive situation; van der Dennen & Falger, 1990),
there are three primary options: fight, tolerate, or avoid. Individuals must undergo a
decision-making process to determine the most appropriate course of action that will
result in the highest net benefit. The model predicts that as the value of the limited
resource increases so do the benefits of taking aggressive action to obtain that resource.
Alternatively, as the value of the relationship increases, taking aggressive action becomes
more costly because the relationship can suffer considerable damage. However, if the
relationship can be repaired through affiliative post-conflict interactions, partners of high
value may still choose to fight over prized commodities (e.g., food, estrous females,
dominance status, etc.). Thus, the likelihood of aggressive escalation is determined not
only by resource and relationship value, but also by the probability of relationship repair
and the risk of injury (de Waal, 2000).
While the Relational Model provides a solid foundation for the cost-benefit
analysis of aggression, Wittig and Boesch (2003b) extended the model to include

5
additional cost-benefit parameters. They classified the likelihood of winning as an
additional benefit, and conflict duration as an additional cost (due to increased energy
expenditure). In accordance with this extended model, Wittig & Boesch found that the
Tai chimpanzees demonstrated flexible decision-making processes before engaging in
conflicts. While individuals sometimes misjudged the situation (e.g., initiators/dominant
individuals losing a conflict), the chimpanzees showed consistent behavioral patterns of
aggression that resulted in an overall net benefit. Wittig and Boesch suggest that more
cost-benefit conflict comparison studies are needed to make inferences about within- and
between-species differences in conflict decisions.
Wild Chimpanzee Intragroup Aggression
According to the ecological constraints model, as group size increases, intragroup
resource (i.e., feeding) competition also increases (Chapman & Chapman, 2000).
Chimpanzees live in fluid fission-fusion societies (Goodall, 1986; Nishida, 1979), which
offer unique temporal and spatial dispersal options contrary to other group-living animals
(i.e., fission-fusion dynamics; Amici et al., 2008). Dispersal flexibility is of particular
importance when resources are patchily distributed and easily monopolized by a few
individuals (Janson & Goldsmith, 1995). By fissioning into smaller subgroups (i.e.,
parties) according to resource availability and location, chimpanzees can potentially
reduce intragroup aggression between community members (Anderson et al., 2002). In
support of this idea, Muller (2002) found the rate of male aggression in chimpanzees is
positively correlated with party size. Thus, fissioning social strategies may have evolved
as a conflict management strategy (i.e., conflict avoidance and escalation); however,

6
reunions involving the fusion of parties often incite aggression (Aureli & Schaffner,
2007; Muller, 2002). Affiliative social relationships between community members
therefore play a key role in promoting tolerance and regulating aggression. Chimpanzees
use affiliative behaviors during reunions for this very purpose (Nishida et al., 1999;
Okamoto et al, 2001). However, these tactics are not always employed or effective in
avoiding aggression.
Chimpanzees in Kibale National Park tend to congregate in large parties when
parous females are in estrous (Wakefield, 2008), large preferred trees are abundantly
fruiting (Wrangham, 2000; Wrangham et al., 1992), hunting for meat (Watts & Mitani,
1999), and patrolling near their territory boundary (Watts & Mitani, 2001). Large parties
can encourage competition between community members and provide more opportunities
for fighting over limited resources (Watts, 1998; Wrangham, 2002). [However when
large parties are engaged in cooperative behaviors where the competition is redirected
towards another species (i.e., during hunting, displacing monkeys from a feeding tree,
etc.) or another chimpanzee community (i.e., patrolling), aggression between group
members often declines (Hartel, unpublished data).] Thus, it is logical to assume as the
number of individuals present in the party, time spent in close proximity to party
members, and intense resource competition increase, the probability of intraparty conflict
also increases.
Due to these factors, chimpanzees, in general, exhibit high rates of intragroup
aggression relative to humans (Wrangham et al., 2006) and other primates. Aggression
rates are however disproportionate between males and females, where males are typically

7
the more aggressive sex (Muller, 2002). The context of aggression is also divided by sex.
Males tend to fight more in social and sexual contexts (Goodall, 1986; Muller, 2002),
while females fight more over food competition (Wittig & Boesch, 2003c) and offspring
protection (Goodall, 1986; Muller, 2002; Nishida, 1989). Because chimpanzees cannot
always avoid conflict and often exhibit high rates of intragroup aggression over prized
resources, conflict resolution strategies become even more important to maintain group
cohesion and repair social bonds (Aureli et al., 2002; de Waal, 2000).
Conflict Resolution Strategies
While other species have been shown to exhibit conciliatory behaviors following
fights (reviewed in Schino, 2000; e.g., hyenas: Hofer & East, 2000; dolphins: Holobinko
& Waring, 2010; meerkats: Kutsukake & Clutton-Brock, 2010; goats: Schino, 1998),
nonhuman primates have been the predominant species studied in conflict resolution
(reviewed in Aureli et al., 2002). This interest is largely driven by our close evolutionary
and genetic relationship with nonhuman primates (Wilson & Sarich, 1969), and how
similar we are behaviorally (Kinzey, 1987).
When the first study of conflict resolution in primates emerged (de Waal &
Roosmalen, 1979), anthropocentric functional terms (i.e., reconciliation, consolation,
appeasement) were assigned to the observed post-conflict behaviors before their
functionalities were empirically tested. The primates’ post-conflict behaviors so closely
mirrored those of humans (Butovskaya et al., 2000; Cords & Killen, 1998; Verbeek & de
Waal, 2001), it seemed appropriate to use the same terms. However, as the field of
conflict resolution has evolved, certain researchers have proposed alternative terms that

8
do not infer function (i.e., peaceful post-conflict interaction (PPCI), signals (PPCS),
contact (PPCC), Silk, 1997, 1998, 2000; third-party affiliation (TPA), etc.). While this
may be theoretically more appropriate, it complicates the discussion by littering text with
long descriptions and numerous acronyms. Therefore for the sake of simplicity, this study
will retain the more functional terms with the understanding that function is not being
implied. Behavioral functionality of each term will instead be discussed below and tested
in subsequent chapters.
Reconciliation: Form and Function
The traditional view of conflict management in social animals, prior to systematic
studies of post-conflict behavior in primates, was that aggression caused group members
to disperse. This hypothesis predicted a decreased probability of contact between former
opponents. However, in the first systematic study of reconciliation, de Waal & van
Roosmalen (1979) tested this ‘dispersal hypothesis,’ and instead found that captive
chimpanzee opponents were actually more often ‘attracted’ to one another. In short, the
opponents often engaged in affiliative behaviors shortly following the conflict rather than
dispersing. The first occurrence of affiliative behavior between opponents was
subsequently termed ‘reconciliation.’ Later studies further refined and contributed
methodological improvements to the study of reconciliation and post-conflict behaviors.
These captive studies now serve as the standard practice for investigating and comparing
post-conflict behaviors (standard methodology: de Waal & Yoshihara, 1983; standard
conciliatory calculations: Veenema et al., 1994). Some wild chimpanzee studies have
adopted these standardized behavioral observation methods (Arnold & Whiten, 2001;

9
Kutsukake & Castles, 2004), but have made slight adjustments as wild observations are
more challenging than those in captivity (Aureli et al., 2002). Alternatively, others have
devised new methods for analyzing wild conciliatory behavior (Wittig & Boesch, 2003a,
2005).
While the operational definition of reconciliation is clear (de Waal & van
Roosmalen, 1979; Cords, 1993), the function is still debated. The term reconciliation is
defined as the restoration of friendly relations between two or more individuals after a
conflict. As previously mentioned, function is explicitly implied in the very definition.
This implied functional mechanism may however be limited and may not account for the
full functionality of reconciliation in both animals and humans. Understanding
reconciliatory functionality is arguably the most interesting component of the behavior
itself. It implies a certain evolutionary basis for the development of complex social
behavior in gregarious species.
To better understand function, the social costs of conflict must be considered.
Aggression can damage social relationships in several ways by 1) increasing the
probability of renewed (Aureli & van Schaik, 1991; Cords, 1992; Koski et al., 2007a;
Kutsukake & Castles, 2001; Silk et al., 1996; Watts, 1995; Wittig & Boesch, 2003a) or
redirected aggression (Aureli et al., 1989; Koski et al., 2007a; Kutsukake & Castles,
2001; Wittig & Boesch, 2003a), 2) reducing tolerance levels of group members and the
likelihood for coalitionary support (Aureli et al., 2002; Watts, 2006), and 3) increasing
anxiety, measured by self-directed behaviors, in victims (Aureli, 1997; Aureli et al.,
1989; Castles & Whiten, 1998; Cooper et al., 2007; Koski et al., 2007b; Kutsukake &

10
Castles, 2001; Majolo et al., 2005; Schino et al., 2007) and even aggressors (Aureli,
1997; Castles & Whiten, 1998; Cooper et al., 2007; Das et al., 1998; Schino et al., 2007).
Reconciliation likely functions to reduce these costs of aggression by repairing
relationships, reducing anxiety, and restoring group cohesion by reducing uncertainty (de
Waal & Yoshihara, 1983). There are currently three predominant hypotheses that address
the functionality and benefits of reconciliatory behavior in primates:
1. The valuable relationships hypothesis (VRH) predicts that opponents sharing
relationships of higher [mutual] value should preferentially reconcile to preserve and
maintain the integrity of those relationships (de Waal & Aureli, 1997). Relationship
quality has been defined by three main components: value, compatibility, and security
(Cords & Aureli, 2000), and standardized to delineate between mutual, one-sided, and
neutral preferred social partnerships (Gilby & Wrangham, 2008). Value is a measure of
the benefits afforded by the relationship (e.g., food sharing, agonistic support, grooming,
etc.). Compatibility is a measure of tolerance and affiliation (e.g., proximity, directed
aggression, etc.). Security is a measure of predictability and consistency (e.g., rates of
self-directed behaviors) between two opponent pairs during baseline observations (Cords
& Aureli, 2000).
Despite the above definitions, the criterion used to assess relationship value has
varied considerably between studies. Some studies have used measures of association or
affiliation (Arnold & Whiten, 2001), while others have used agonistic support and food
sharing (Witting & Boesch, 2003a, 2005), coalitionary support (Preuschoft et al., 2002;
Watts, 2006), proximity (Kutsukake & Castles, 2004), and partner sex combination

11
(Arnold & Whiten, 2001, de Waal, 1986; Koski et al., 2007b) all as indices. Furthermore,
some studies have failed overall to find support for the VRH (Kutsukake & Castles,
2004). It is therefore not surprising that the results from these studies sometimes differ as
much as the aforementioned methods used.
Given these inconsistencies, it is imperative that future research focuses on
comprehensively defining relationship quality in order to properly test the VRH. Various
relationship quality parameters (including affiliative and agonistic measures) should be
tested to determine their candidacy and validity for the population of interest. Subsequent
analysis should determine what factors (e.g., kinship, sex-combination, conflict intensity,
age difference, dominance rank difference, relationship tenure, etc.) result in variation in
each of these parameters, which may indeed vary from one population to the next or
between captive and wild environments. While the VRH is the most widely cited, well-
supported explanation for the occurrence of reconciliation in chimpanzee populations
(Arnold & Whiten, 2001; de Waal & van Roosmalen, 1979; Fraser et al., 2008a; Watts,
2006; Wittig & Boesch, 2003b, 2005), more comprehensive research is still needed to
tease apart the correlation between relationship quality and reconciliation, especially in
wild ape populations.
2. The integrated hypothesis proposed by Aureli (1997) combines the VRH with
the uncertainty-reduction hypothesis (Aureli & van Schaik, 1991), where post-conflict
stress is measured by the opponents’ levels of behavioral anxiety (e.g., self-directed
behaviors). The predictions for this hypothesis are two-fold: 1) Reconciliation will
increase for conflict partners with more valuable relationships. 2) Since conflicts disrupt

12
the value or integrity of a relationship, those with more valuable relationships will exhibit
higher post-conflict anxiety (measured by self-directed behaviors) prior to reconciliation.
Post-conflict anxiety will then decrease following reconciliation.
Some chimpanzee studies have found support for this hypothesis. Arnold &
Whiten (2001) were the first to investigate the relationship between reconciliation and
post-conflict stress in chimpanzees, but were unable to reach a conclusion due to
conflicting results. However, captive chimpanzee populations have reliably shown that 1)
opponents exhibiting higher behavioral anxiety levels reconciled preferentially (Koski et
al., 2007b), and 2) self-directed behaviors were significantly lower following
reconciliation (and also consolation) as compared to other PC periods when
reconciliation did not occur (Fraser, 2008). Given the small number of studies examining
this relationship in chimpanzees, further research is necessary to understand the
functional relationship between self-directed behaviors and reconciliation.
3. The benign intent hypothesis has been proposed as an alternative explanation to
the VRH. It argues that individuals simply reconcile to reduce the uncertainty of the
situation and signal that the conflict is over and the actor’s intentions are peaceful and
benign (Aureli & van Schaik, 1991; Silk, 1996, 2000). This hypothesis suggests
individuals reconcile to gain short-term benefits (e.g., access to resources or preferred
partners) as opposed to reconciling for more long-term objectives like relationship repair
(Silk, 2002a). However, Cords & Aureli (1996) argue this as a more short-term functional
explanation, whereas the VRH is a more appropriate ultimate explanation for
reconciliatory behavior.

13
Although most studies tend to claim support for one hypothesis or another, it
should be questioned whether these various functional explanations of reconciliatory
behavior and conflict resolution are actually mutually exclusive, meaning only one
hypothesis explains why individuals reconcile. Instead, these hypotheses when
‘reconciled’ may each provide possible functional explanations for the complex suite of
behaviors that occur during conflict resolution in primates.
Regardless of its overall functionality, reconciliation does not come without costs.
In order to reconcile, opponents must approach one another shortly after the fight to
engage in an affiliative behavior. The approach is the most risky moment, as opponents
must decide if they are going to reconcile or renew aggression (Aureli & van Schaik,
1991; Aureli et al., 2002). This risk gives pause to many opponents and instead often
results in avoidance. Since renewed aggression can carry some serious costs like causing
further damage to opponents’ relationships and increasing physical injury, energy
expenditure, and stress, the potential costs may sometimes outweigh the potential
benefits. Reconciliation should only be attempted when the known benefits outweigh the
potential risks (e.g., probability of renewed aggression, intensity of conflict, etc.; de Waal
& Aureli, 1997)
Consolation: Form and Function
Consolation, another affiliative post-conflict behavior, is defined as affiliative
contact between the victim and a third-party bystander during the post-conflict period (de
Waal & van Roosmalen, 1979). Consolation, like reconciliation, may also have distress
alleviating properties for the victim (Aureli, 1997; de Waal & Aureli, 1997). Since the

14
original post-conflict study by de Waal and van Roosmalen, a distinction has been made
between the two forms of consolation: true consolation (also referred to as offered
consolation, see Wittig & Boesch, 2003a) and solicited consolation (reviewed in Watts et
al., 2000). True consolation (henceforth referred to as consolation) occurs when a third
party bystander initiates an affiliative interaction toward the recipient of aggression
(henceforth referred to as victim). Solicited consolation occurs when the victim initiates
an affiliative interaction toward a third party (Verbeek & de Waal, 1997).
Operationally, consolation is quite simple to identify, but functionally it is much
more complex. While many species have been shown to operationally demonstrate
solicited consolation (reviewed in Watts et al., 2000), [offered or true] consolation occurs
most convincingly in great apes (Pan troglodytes: de Waal & Aureli, 1996; de Waal &
van Roosmalen, 1979; Fraser & Aureli, 2008; Koski & Sterck, 2007; Kutsukake &
Castles, 2004; Palagi et al., 2006; Wittig & Boesch, 2003a; Pan paniscus: Clay & de
Waal, 2013a; Palagi et al., 2004; Palagi & Norscia, 2013; Gorilla gorilla: Cordoni et al.,
2006; Mallavarapu et al., 2006; Watts et al., 2000) and humans (Fujisawa et al., 2006).
However, there is some operational evidence of consolation in monkeys (Call et al.,
2002; Wittig et al. 2007), Asian elephants (Plotnik & de Waal, 2014), corvids (Fraser &
Bugnyar, 2010; Seeds et al., 2007), and canids (dogs: Cools et al., 2008; wolves: Palagi
& Cordoni, 2009), but other explanations are posed, and it is not yet clear if consolation
in these species is functionally equivalent to consolation in great apes. Of these studies,
Fraser and Bugnyar (2010), however, have come the closest to demonstrating ape-like
consolation in corvids, Palagi and Cordoni (2009) in Pistoia wolves, and recently Palagi

15
et al. (2014) in Tonkean macaques (egalitarian species) and Poltnik and de Waal (2014)
in Asian elephants. Interestingly, elephant bystanders not only offered physical and vocal
affiliation to distressed victims, but also adopted the victim’s emotional state (Poltnik &
de Waal, 2014). This study provides the first empirical evidence of empathy (Preston &
de Waal, 2002) through emotional contagion (de Waal, 2003; Hatfield et al., 1994) in
elephants (anecdotal evidence: Bates et al., 2008; Payne, 2003). Because so few species
exhibit true consolation in form and function, this behavior, until recently, has received
far less attention than reconciliation in studies and the literature (Watts et al., 2000).
While reconciliation and consolation are not necessarily mutually exclusive
behaviors, several chimpanzee and bonobo studies have found that consolation will often
occur in the absence of reconciliation (Palagi et al., 2006; Palagi et al., 2004; Wittig &
Boesch, 2003). This suggests that consolation and reconciliation may have some
overlapping functions. Four hypotheses address the functionality and benefits of
consolation (reviewed in Fraser et al., 2009):
1. The consolation (i.e., stress reduction or empathy) hypothesis suggests that the
bystander provides affiliation to the victim as an empathic response to the victim’s
distress (Aureli, 1997; de Waal & Aureli, 1996). Empathic consolation predicts that the
bystander and victim share a close, valuable relationship, in which the bond between the
bystander and aggressor is irrelevant (Wittig & Boesch, 2010). Romero et al. (2010)
found in two captive chimpanzee populations, offered consolation was best predicted by
the bystander’s close relationship with the victim, whereas there was no effect of the
bystander’s relationship with the aggressor. Since empathy is rooted in emotional

16
contagion (de Waal, 2008; Preston & de Waal, 2002) and prosocial behavior (de Waal &
Suchak, 2010; Eisenberg & Fabes, 1998), bystanders should be more likely to console
when the perceived distress of the victim is high (often measured by self-directed
behaviors). Under this premise, consolation should result in reduced anxiety in the victim
since valuable partners are more effective in alleviating distress due to a shared affiliation
and empathetic connection (Fraser et al., 2008; Wittig & Boesch, 2010). In semi-free
living bonobos, Clay and de Waal (2013a) found that bystanders not only offered
consolation to their close associate victims more often than to the aggressor’s close
associate victims, but also consolation reduced stress-related behaviors (i.e., self-
scratching). Furthermore, juvenile bonobos with better developed social and emotional
competence were more likely to offer consolation (Clay & de Waal, 2013b), showing that
a better understanding of empathy promotes consolatory behavior. Although empathic
consolation cannot serve a direct relationship repair function between former opponents,
as does reconciliation, recent studies have shown that consolation can serve an anxiety
reduction function similar to reconciliation when the latter fails to occur (Fraser et al.,
2008b; Palagi & Norscia, 2013; for negative evidence see Koski & Sterck, 2007).
Thus, consolation may be an important tactic for alleviating distress in belligerent
opponents, especially for victims, without having to risk renewed aggression between
opponents (Wittig & Boesch, 2003a). In the last decade, the consolation/stress alleviation
hypothesis has received the most support for explaining bystander-victim affiliation in
great apes (Clay & de Waal, 2013; Fraser et al., 2008b; Pagali & Norscia, 2013; Romero

17
& de Waal, 2010; for negative evidence see Koski & Sterck, 2007), however it has not
been convincingly demonstrated in wild chimpanzees (Wittig & Boesch, 2010).
2. The relationship repair (i.e., substitute for reconciliation) hypothesis predicts
that consolation serves as a proxy for reconciliation by restoring levels of tolerance
between opponents (Watts et al., 2000; Wittig et al., 2007). Bystanders that share a
valuable relationship with aggressors may be ‘reconciling’ with the victim on behalf of
their friend (Wittig et al., 2007). For this strategy to work, the bystander must not only be
a close associate of the aggressor, but the victim must recognize this relationship and
accept them as a reconciliation proxy. Wittig (2010) found support for this hypothesis in
the Tai chimpanzees. Bystanders were more likely to console victims when they had a
higher relationship benefit index with aggressors than with victims (Wittig & Boesch,
2010).
3. The self-protection hypothesis suggests that bystanders offer affiliation to the
victim to avoid redirected aggression (Call et al., 2002; Koski & Sterck, 2009). Studies
have shown that opponents use redirected aggression to deflect aggressive attention and
reduce stress (de Waal & van Hooff, 1981). Under this hypothesis, bystanders should
offer affiliation to higher-ranking victims with whom they share low affiliation since
valuable partner bystanders rarely receive redirected aggression (Call et al., 2002). Unlike
the consolation hypothesis, the benefit of the protection hypothesis favors the bystander.
Limited support for the self-protection hypothesis has been found in two populations of
captive chimpanzees (Koski & Sterck, 2009; Pagali et al., 2006), but no support has been

18
found in wild chimpanzees (Wittig & Boesch, 2010), bonobos (semi-wild: Clay & de
Waal, 2013a; captive: Palagi & Norscia, 2013), or captive gorillas (Cordoni et al., 2006).
4. The victim-protection hypothesis, introduced recently by Palagi and Norscia
(2013), suggests that bystander affiliation decreases the victim’s probability of receiving
further aggression from either the original aggressor or other bystanders. This hypothesis
predicts that recent victims should receive additional aggression from other group
members during the PC period, but the probability of aggression should reduce following
bystander affiliation. Given the recent proposal of this hypothesis, Palagi and Norscia
(2013) have provided the only confirmatory data. In captive bonobos, victims received
less aggression from group members following both offered and solicited consolation.
[However, they did not control for consolers delaying affiliation until the threat of
renewed aggression was low.] Functionally, consolation both protected the victim and
alleviated their distress, showing that bystander affiliation can serve multiple purposes. In
an earlier study, Pagali et al. (2006) found that consolation in captive chimpanzees did
reduce the likelihood of further conflicts between group members, but this trend toward
just the victim was not specifically tested. This new prosocial hypothesis should therefore
be tested in other studies to determine its applicability across species.
While consolation studies have been gaining popularity over the last decade,
results have been conflicting and inconclusive. Thus, the functionality of consolation is
still debated and may be multifaceted. Future studies, especially focusing on wild
populations, should strive to elucidate the role of consolation in conflict resolution and
further investigate the possible functions it may have.

19
Other Third-Party Interactions
While most post-conflict studies have focused on reconciliation and then
consolation, even fewer studies have focused on third-party bystander interactions with
aggressors. Similar to consolation, bystanders can either offer affiliation to aggressors or
aggressors can solicit affiliation from bystanders. While these interactions may seem
similar, the slight difference in initiation may have distinct functions. In the past, some
studies have lumped bystander affiliation with both victims and aggressors into the single
category of consolation (Wittig & Boesch, 2003a). However, this assumption is
presumptuous and each type of bystander affiliation should be investigated separately for
qualitative differences in form and function.
Appeasement. While consolation focuses on interactions with the victim,
appeasement behavior refers to third-party bystanders offering affiliation to the aggressor
during the post-conflict period (Das, 2000). In 1967, van Hooff was the first to
functionally define appeasement as subordinate bystanders offering affiliation to reduce
their risk of future attacks by potential aggressors. As more studies have emerged,
alternative hypotheses have been presented to test the functionality of appeasement.
Some of these hypotheses, while they differ operationally, functionally (and nominally)
overlap with conciliatory hypotheses:
1. The consolation (i.e., stress-alleviation) hypothesis suggests that aggressors,
like victims, may also need affiliation to reduce their distress after a fight. This
hypothesis predicts that the aggressor should show signs of distress following the conflict
and close associate bystanders should offer affiliation more than neutral associates.

20
Empathy is the motivating factor, which is often facilitated by similarity, familiarity, and
social closeness (de Waal, 2008; Hoffman, 2000).
2. The relationship repair hypothesis suggests that bystander affiliation with the
aggressor acts as a proxy for reconciliation between opponents. This hypothesis requires
triadic awareness (Tomasello & Call, 1997) and would predict that bystanders share a
strong or valuable association with the victim. While this form of substitute and triadic
reconciliation has been observed in monkeys (Cheney & Seyfarth, 1989; Judge, 1991;
Wittig et al., 2007), there is some initial evidence for apes (Wittig & Boesch, 2010).
3. The self-protection hypothesis suggests that appeasement functions as an
aggression avoidance strategy, whereby bystanders are attempting to interact affiliatively
with the aggressor to avoid redirected aggression. This hypothesis predicts likely targets
of aggression should offer affiliation to aggressors more than unlikely targets (Koski &
Sterck, 2009). Under this hypothesis, the action provides the bystander with a potential
benefit (i.e., aggression avoidance). In one population of captive chimpanzees, third
parties who offered affiliation to their potential aggressors were less likely to receive
aggression following the previous conflict (Koski & Sterck, 2009). Essentially, bystander
affiliation toward aggressors does not completely extinguish further aggression; rather it
redirects aggression away from that particular bystander.
4. The appeasement (i.e., prevention of aggression) hypothesis suggests that
bystander-aggressor affiliation functions to appease/calm the aggressor, restore levels of
tolerance in the group, and decrease the probability of further aggression. This differs
slightly from the self-protection hypothesis, as it does not focus solely on reducing the

21
risk of aggression for only the affiliative bystander, but instead on the general reduction
of aggression. This hypothesis predicts that close associate bystanders are more likely to
offer affiliation to the aggressor, and in doing so, will decrease the probability of further
aggression directed toward any group member. Aggressors may respond with group
tolerance after bystander affiliation because a valuable partner is showing the aggressor
support. While the bystander could have provided coalitionary support during the
conflict, that comes with its own set of risks. Appeasement may show just enough
support to satisfy the aggressor without having to fight alongside. This “prevention of
aggression” hypothesis has been demonstrated in captive chimpanzees (Koski & Sterck,
2009; Palagi et al., 2006) and lowland gorillas (Palagi et al., 2008), but has not been
demonstrated in wild apes.
Each hypothesis hinges on the quality of the relationship between the aggressor
and bystander. Studies must therefore test how the relationship quality influences the
occurrence and function of bystander-aggressor affiliation and how this may affect their
cost-benefit strategy. In fact, several possible benefits of the bystander-aggressor
interaction have been shown in previous primate studies, providing some suggestive
support for each of the hypotheses. If bystanders and aggressors are valuable partners,
interactions following conflicts may strengthen that bond (Call et al, 2002; Das, 2000). In
macaques and baboons, substitute (triadic) reconciliation, affiliative interaction between
aggressors and the kin of victims (Judge, 1991), may restore the relative’s (victim’s)
relationship with the aggressor (Wittig et al., 2007). Bystanders may reduce their
probability of redirected aggression by directly appeasing the aggressor (Call et al., 2002;

22
Palagi et al., 2008) or restoring levels of tolerance by not spreading aggression (Palagi et
al., 2006; Judge & Mullen, 2005).
However, appeasement is still a risky behavior. As a bystander approaches the
aggressor, they risk receiving redirected aggression (Pagali et al., 2008; Watts et al.,
2000). This risk is similar to the risk of renewed aggression in reconciliation. Thus like
reconciliation, the benefits of approaching and interacting with the aggressor must
outweigh the costs. More research is needed to better understand appeasement and test
whether the functionality of this behavior varies between and within species.
Recruitment and support. Very few studies have investigated aggressor-solicited
affiliation with third-party bystanders in the post-conflict period (Barbary macaques:
Aureli, 1997; olive baboons: Castles & Whiten, 1998; long-tailed macaques: Das et al.,
1998). Unlike appeasement, the function of this behavior likely benefits the aggressor as
opposed to the bystander. In more egalitarian species, like chimpanzees, aggressors may
experience some distress following a conflict especially if 1) the aggressor attacks a more
dominant individual, 2) there is no clear winner, and/or 3) the victim unexpectedly
retaliates. In two egalitarian species, Barbary macaques (Aureli, 1997) and olive baboons
(Castles & Whiten, 1996), and one despotic species, long-tailed macaque (Das et al.,
1998), aggressors exhibited an increase in stress-related behaviors following a conflict.
However in long-tailed macaques, aggressors’ anxiety levels did not decrease following
solicited affiliation with a bystander. Instead, this affiliation fueled the aggressor and
appeared to signal support and encouragement (Das et al., 1998). The limited data from
monkeys suggests that the possible functions of aggressor-initiated bystander affiliation

23
may be stress alleviation or third-party support/recruitment. Since this behavior is often
overlooked, hypotheses addressing the functionality of aggressor-solicited bystander
affiliation have not been well thought-out, developed, or investigated. Future studies
should focus on this largely untouched area of conflict resolution.
Measures of Stress and Distress in Primates
Self-Directed Behaviors
Displacement activities are behavior patterns that occur with no apparent
significance to the current context (Tinbergen, 1952). These behaviors, often in the form
of self-directed behaviors (SDB) in primates (i.e., self-grooming, self-scratching,
yawning, and body shaking), are typically observed in higher frequency in situations of
increased uncertainty or high tension (Aureli et al., 1989) and appear to be related to
autonomic arousal (Maestripieri et al., 1992; Troisi, 2002). There is both pharmacological
(Schino et al., 1996) and behavioral (Aureli et al., 1999; Castles et al., 1999) evidence to
suggest that scratching and some other self-directed behaviors may function as behavioral
signals of heightened arousal and increased physiological stress in chimpanzees.
Although SDBs can have a self-maintenance function and should not solely be
viewed as a ‘pure’ measure of anxiety, many studies have clearly illustrated their inflated
rate of occurrence in high-tension, high-stress contexts across many species, both captive
and wild, within the primate taxa (Aureli, 1997; Das et al., 1998; Kutsukake & Castles,
2001). In chimpanzees, rough scratching appears to be the most salient (Kutsukake,
2003) and reliable behavioral indicator of stress (Aureli & de Waal, 1997; Baker &
Aureli, 1997; Kutsukake, 2003; te Boekhorst et al., 1991). Several captive chimpanzee

24
studies have shown that chimpanzee SDBs increased when stranger vocalizations were
heard (Baker & Aureli, 1997), population density was high (Aureli & de Waal, 1997),
performing difficult cognitive task or committing an error (Itakura, 1993; Leavens et al.,
2001). In wild chimpanzees, Kutsukake (2003) found that low-ranking males produced
more rough scratches than high-ranking males when resting, and females produced more
rough scratches when other community members where present and also when those
individuals were non-affiliative. This demonstrates an apparent sex difference in the
context of anxiety. In general, females tended to exhibit higher levels of behavioral
anxiety compared to males.
Since aggressive contexts are also high-stress situations, it is hypothesized that the
victim’s SDBs will increase immediately after an aggressive conflict relative to baseline
conditions, but affiliative behavior between opponents or a bystander can reduce SDB
rates back to baseline (see integrated and consolation hypotheses). Several post-conflict
studies on various monkey species have shown inflated rough scratching rates to decline
as a function of time (Schino et al., 2007) and following reconciliation (Aureli & van
Schaik, 1991; Castles & Whiten, 1998; Cooper et al., 1997). However, until recently, the
relationship between aggression, conciliatory behaviors, and SDBs had not been
demonstrated in any ape species (for negative evidence see Koski & Sterck, 2007).
Arnold & Whiten (2001) were the first to investigate the relationship between
reconciliation and post-conflict stress in wild chimpanzees, but were unable to reach a
definitive conclusion due to conflicting results. Fraser (2008) was able to show that in her
captive chimpanzee population, SDBs were significantly lower following reconciliation,

25
and also consolation, as compared to other PC periods when reconciliation did not occur.
Bonobos have also shown a similar pattern of stress alleviation following consolation
(Clay & de Waal, 2013a; Palagi & Norscia, 2013).
Given that only a few ape studies have examined this relationship with SDBs and
the inconsistencies in their results, further research is necessary to understand the
functional relationship between SDBs and peaceful post-conflict interactions in
chimpanzees. However, researchers must be careful when using SDBs as distress markers
in conflict resolution studies because SDBs can also be high in baseline conditions due to
non-aggressive stress-invoking situations. Using a stress-related behavior or vocalization
that is confined to aggression (i.e., screaming) may be a more reliable indicator of
distress.
Distress and Recruitment Vocalizations.
In general, chimpanzee [bared-teeth] screams (van Hooff, 1973) are known to
provide detailed information about the social meaning/context (Slocombe et al., 2009)
and the caller’s emotional state (Siebert & Parr, 2006). Depending on an individual’s role
in aggression (i.e., aggressor or victim; Slocombe & Zuberbühler, 2005) and the intensity
of the conflict (Slocombe & Zuberbühler, 2007), chimpanzees produce acoustically
distinct agonistic screams. Listeners are then able to discern information based on the
scream’s acoustics (Slocombe et al., 2010) and can identify the vocalizer even when they
are not visible (Kojima, 2003). In chimpanzees, screaming in an aggressive context
appears to not only indicate distress, but may also be used to recruit help and/or seek
comfort (Goodall, 1986, p.129). If screams do have an advertisement and recruitment

26
function, distressed individuals (typically victims) should continue screaming until they
receive support (i.e., aggression intervention) or comfort (i.e., consolation). Whimpering,
a notably quieter vocalization, also indicates distress and may serve a similar comfort
recruiting function (Goodall, 1986, p. 129). Waa-barks, a loud sharp sound, also occur in
the agonistic context and are typically used by female victims of aggression to recruit
help and can indicate distress or “defiance” (Goodall, 1986, p. 130).
Given the fission-fusion structure of chimpanzee society, distress vocalizations
may be a more appropriate advertiser of distress than SDBs. When party members are
dispersed and are not in view of one another, individuals may not be aware of ensuing or
on-going conflicts. Screaming and waa-barks quickly advertise this information to all
party members and indicate important information about each caller’s emotional state.
SDBs, on the other hand, only provide distress information if individuals are in close
visual proximity of one another. Thus, SDB advertisement is likely limited to a small
audience reaching only a few meters. Furthermore, high-pitched screaming, whimpering,
and waa-barks in adults, unlike SDBs, occur almost exclusively in the context of
aggression (Goodall, 1986) and therefore function as direct measures of aggression-
related anxiety versus general anxiety. Since screaming and waa-barks reach a wider
audience (which increases the probability of recruitment) relative to SDBs and are more
confined to aggressive situations, they may therefore serve as a more appropriate and
effective measure of distress in wild chimpanzees following aggressive conflicts. Distress
vocalizations in chimpanzees should therefore encourage consolation, and subsequently
cease following affiliation.

27
Since SDBs have served as the standard in quantifying distress in primate conflict
resolution studies for the last 30 years, screaming has been largely ignored. To my
knowledge, only one study has investigated the relationship between screaming and
conflict resolution. Pagali et al. (2006) found that consolation was more frequent when
the victim screamed. However, they did not report whether the affiliative interaction
alleviated the victim’s distress (i.e., victim ceased screaming). More research should
incorporate screaming, whimpering, and waa-barks as both measures of distress and as
recruitment strategies in chimpanzee conflict resolution.
Conflict Resolution in Chimpanzees
Many studies of captive chimpanzees (Baker & Smuts, 1994; de Waal & van
Roosmalen, 1979; Fuentes et al., 2002; Koski et al., 2007a; Preuschoft et al., 2002) and
other primate species (Arnold & Barton, 2001; Cordoni et al., 2006; de Waal &
Yoshihara, 1983; Palagi et al., 2004; Silk et al., 1996) have demonstrated conflict
partners’ propensity to reconcile following acts of aggression. There is much ambiguity,
however, regarding chimpanzees’ reconciliation rates and explanations for doing so. The
wide variability in observed rates (14.4% to 47.5%, see Baker & Smuts, 1994 and Fraser
& Aureli, 2008, respectively), and even behaviors used during reconciliation (e.g.,
kissing and embracing behaviors, see de Waal & van Roosmalen, 1979; Fraser & Aureli,
2008), from one captive population to the next leaves unanswered questions about the
true function of the behavior and the underlying evolutionary mechanisms that mitigate
post-conflict interactions in group-living animals.

28
For decades, it was assumed that the results gleaned from captive studies could be
extrapolated to their wild counterparts. Only recently have post-conflict data on wild
chimpanzees been available. The four field studies – Tai North community (Wittig &
Boesch, 2003a, 2005), Mahale M-group (Kutsukake & Castles, 2004), Budongo Sonso
community (Arnold & Whiten, 2001), and Ngogo community (Watts, 2006) – have
provided an even more variegated response pattern of chimpanzee post-conflict behavior.
Only two field studies (Mahale & Budongo) used controlled procedures comparable to
the standard post-conflict methodology (de Waal & Yoshihara, 1983) used in captivity.
At Budongo, Arnold and Whiten (2001) examined the form, rates, and function of
reconciliation, consolation, and stress-related behavioral correlates, but their conclusions
were limited due to their small sample size (N = 120 Post conflict (PC) - Matched control
(MC) pairs). They were unable to demonstrate consolation or reliable behavioral
correlates of stress in their wild population.
Behavioral Patterns
Reconciliation. Comparing rates of reconciliation from one population to the next
is also challenging. Reported group conciliatory tendencies (CCT; Veenema et al., 1994)
for the wild populations, Mahale CCT: 15.5% (Kutsukake & Castles, 2004), Tai CCT:
15.85% (Wittig & Boesch, 2005) and Budongo CCT: 12.3% (Arnold & Whiten, 2001),
are similar to some captive groups (CCT 14.4%: Baker & Smuts, 1994 and CCT 17.25%:
Fuentes et al., 2002), but almost half the rate observed in others (CCT 27-35%: de Waal
& van Roosmalen, 1979, CCT 47.5%: Fraser & Aureli, 2008; CCT 41%: Preuschoft et
al., 2002). These CCT discrepancies demonstrate the wide variation between populations,

29
the likely dependence of the behavior on each group’s unique social and ecological
environment (Colmenares, 2006), and possibly cultural differences that may exist
between them. At present, there are too few studies to draw conclusions about the within
species source of variation in both captive and wild settings (Boesch et al., 2002;
Kutsukake & Castles, 2001).
Captive and wild chimpanzee studies have also shown that reconciliation rates are
often influenced by various intrinsic variables of the preceding conflict. In particular,
opponent sex combination has produced varying reconciliation patterns. Wild male
chimpanzees, relative to females, generally have higher rates of association due to shared
kinship, strong alliances, and strict linear dominance hierarchies (de Waal, 1982;
Goodall, 1986; Nishida, 1983; Watts, 1998). One would predict that this strong pattern of
male association would influence reconciliation rates between and among the sexes.
However, inconsistent patterns have emerged from one population to the next. In support
of males sharing strong social bonds, Koski et al. (2007b) found male-male opponents
reconciled more often than female-female dyads at the Burgers’ Zoo in Arnhem, the
Netherlands. However, many other studies have found contrasting patterns. Fraser and
Aureli (2008) noted that the sex-combination of opponents was one of the strongest
predictors of reconciliation in their captive chimpanzee group at the Chester Zoo, UK.
They found female-female opponent dyads reconciled significantly more than male-male
dyads. This pattern is in stark contrast to female-female relationships in wild populations
(Gilby & Wrangham, 2008; Wrangham, 2000; Wrangham et al., 1992), even compared to
the more gregarious West African females (Lehmann & Boesch, 2008). Alternatively in

30
Tai chimpanzees, Wittig & Boesch (2003a) observed male-female conflict partners to
reconcile at the highest frequency, while Mahale chimpanzees showed no evidence of sex
combination influencing reconciliation rates (Kutsukake & Castles, 2004).
At Mahale, reconciliation rates were instead influenced by conflict intensity
(Kutsukake & Castles, 2004). Individuals reconciled significantly more following high
intensity aggressive conflicts versus low intensity conflicts (for negative evidence see
Koski et al., 2007b) with victims being the more likely initiators. Similarly, Wittig &
Boesch (2005) found that the intensity of the conflict influenced the rate and complexity
of reconciliation at Tai. The more severe the conflict; the more behaviors involved in
reconciling. Additionally, rank difference between conflict partners at Tai also influenced
reconciliation rates (Wittig & Boesch, 2003a).
Other studies have shown the context in which the conflict occurred could also
influence opponent pairs’ tendencies to reconcile. Wittig and Boesch (2005) found that
reconciliation duration was shorter following conflicts over food and longer following
conflicts over sex. These studies suggest that reconciliation may not be as important or
needed in certain contexts. Chimpanzees may “expect” to have fights over food, making
reconciliation obsolete. The idea of context-specific reconciliation is a hypothesis that
should be explored further.
Consolation. Consolation has been demonstrated convincingly in several captive
and wild chimpanzee populations; however, some chimpanzee studies have failed to do
so (Arnold & Whiten, 2001; Fuentes et al., 2002). Similar to reconciliation, the degree of
variation between studies can be quite large. For example, Fraser and Aureli (2008)

31
documented a very high rate of consolation among their captive chimpanzee population.
Their consolation [triadic] conciliatory tendency (TCT) was 29.4%. Interestingly, this
TCT value is nearly ten times the rate documented in wild chimpanzees (Arnold &
Whiten, 2001; Wittig & Boesch, 2010). The variability among studies again shows the
need for future studies on the form and function of consolatory behaviors, especially in
wild populations.
Appeasement. Appeasement has received the least attention in post-conflict
studies. Very few studies have investigated the occurrence of appeasements, let alone its
functionality (Das, 2000). While several chimpanzee studies have documented the
occurrence of bystander-aggressor affiliation, a clear understanding of its function has yet
to emerge. Uniquely, Wittig and Boesch (2010) found support for the relationship repair
hypothesis in the Tai chimpanzees. No other chimpanzee studies have replicated this
functional explanation. Instead, Koski & Sterck (2009) found support for the self-
protection hypothesis in their captive chimpanzee group, and Palagi et al. (2006) found
support for the appeasement hypothesis. More studies are needed to understand the
functional underpinnings of this understudied behavior.
Behavioral specificity. In the first systematic study of reconciliation and
consolation, de Waal and van Roosmalen (1979) suggested that [captive] chimpanzees
preferentially use kisses during reconciliation and embraces during consolation. These
two specific behaviors rarely occurred in other social contexts outside of the post-conflict
period. Later, de Waal and Ren (1988) suggested that more socially tolerant species
might engage in explicit gestures during acts of reconciliation to restore the relative status

32
of the partners. Until recently (Fraser dissertation 2008), this finding of behavioral
specificity during post-conflict interactions has not been replicated (Arnold & Whiten,
2001; Fuentes et al., 2002; Kutsukake & Castles, 2004). Wittig and Boesch (2005) argue
that although the type of behavior used to elicit reconciliation may not be unique to the
context, the complexity of behavioral elements may be. They found that the complexity
of reconciliation (based on number of behavioral elements used) of the wild chimpanzees
of the Tai forest varied according to the intensity of the prior conflict. Kutsukake and
Castles (2004) found that intensity of the preceding conflict affected the reconciliation
rates of the chimpanzees, but unlike Wittig and Boesch, they did not investigate the
behavioral complexity of the reconciliatory actions. Since neither study’s prediction of
chimpanzee behavior during reconciliation has been consistently repeated in captive or
wild studies, further investigation of their occurrence and validity is necessary.
Following conflicts, there may also be ecological consequences to the conflict
partners if reconciliation or consolation does not occur. Aureli (1992) found that long-
tailed macaques, Macaca fascicularis, spent less time foraging following a conflict. This
skew in their typical activity budget may be the result of increased vigilance and/or
heightened anxiety, which are behaviors that typically decrease following reconciliation
(Fraser, 2008) and possibly consolation (Fraser et al., 2008b). This should be explored in
wild chimpanzee populations.
Vocalizations. Vocalizations have been largely ignored in chimpanzee post-
conflict studies. Silk (1996) found that female baboons used grunts to signal
reconciliation after conflicts and also to facilitate infant handling. In a later paper, Silk

33
(2002a) suggested that other studies have neglected to either record or report the role of
vocalizations in the post conflict behaviors of other primate species. Given the
importance they play in baboon society, subtle or specific vocalizations may also play a
similar role in chimpanzee post-conflict interactions. de Waal and Roosmalen (1979)
simply noted that their chimpanzee population sometimes paired submissive
vocalizations with their reconciliatory and consolatory behaviors, but did not analyze
their relationship to the conciliatory behavior. Future research should investigate the role
of vocalizations in this context and whether they can function as signals to elicit
conciliatory behavior.
Project Aims
Additional research on the post-conflict behaviors of other wild chimpanzee
populations is necessary to further understand 1) the broad spectrum of behaviors and
rates that have been observed, 2) what factors predict the occurrence of conciliatory
behavior, and 3) the functionality of these behaviors. However, for PC research to be
effective and suitable in a wild environment, it is important that traditional PC methods
be adapted to accommodate the fluidity in wild chimpanzee fission-fusion society. This
project aims to accomplish the above while accounting for the underlying functional
aspects of the post-conflict behaviors relative to wild chimpanzee socio-ecology.
A better understanding of the behaviors that appear to mitigate aggression in wild
chimpanzees can also provide insights into the social structure and the development of
peaceful post-conflict interactions and behaviors in early hominid populations. Although
the actual reconciliation behaviors of early hominids may always remain a mystery,

34
continued studies on the reconciliatory signals, behaviors, and strategies of chimpanzees
may further bring to light the evolution of human conciliatory behavior over the last five
million years.

35
Chapter 2

General Methods

36
Study Site
Kibale National Park (Figure 2.1) is the largest protected Ugandan reserve,
measuring 766 km
2
. It is located in southwestern Uganda (0°13’-0°41’N, 30°19’-30°32’
E) near the foothills of the Ruwenzori Mountains. Elevation in the national park
fluctuates between 1100 and 1600 m. The national park was officially established in
1993, prior to which it was a forest reserve that allowed sustainable logging (Chapman &
Chapman, 1997; Osmaston, 1959). Vegetation within the park includes a mosaic of moist
evergreen and semi-deciduous forest, grasslands, papyrus swamp, and exotic softwood
plantations (Chapman & Wrangham, 1993; Struhsaker, 1975). See Struhsaker (1997) for
a detailed description of Kibale National Park.

Figure 2.1. Map depicting Kibale National Park in Uganda, noting location of the
Kanyawara chimpanzee community (modified from Chapman et al., 2002).

37
Kibale is celebrated for its diverse populations of butterflies, birds, and primates.
The park offers one of the highest primate diversities and densities in Africa. Thirteen
species of nonhuman primates live within the park’s boundary: four species of prosimians
[dwarf bushbaby (Galago demidovi), eastern needle-clawed bushbaby (Galago inustus),
Thomas’s bushbaby (Galago thomasi), and potto (Perodicticus potto)], eight species of
monkeys [(red colobus (Procolobus rufomitratus), black & white colobus (Colobus
guereza), grey-cheeked mangabey (Cercocebus albigena), red-tailed monkey
(Cercopithecus ascanius), vervet monkey (Cercopithecus aethiops, L’Hoest’s monkey
(Cercopithecus lhoesti), blue monkey (Cercopithecus mitis), and olive baboon (Papio
anubis,)], and the endangered East African chimpanzee subspecies (Pan troglodytes
schweinfurthii). In 2003, more than 1200 East African chimpanzees were estimated to
live in the park at a density of 2.32 km
2
, which is the highest chimpanzee density of any
forest surveyed in Africa (Plumptre et al., 2003). Two of the park’s chimpanzee
communities, Kanyawara and Ngogo, are fully habituated to humans.
Kanyawara, the region inhabited by the study community, occupies
approximately 38 km
2
(Wilson, 2001) and is located on the park’s western edge (0°34’ N,
30°21’E) sitting at an average elevation of 1500 m (Chapman et al., 2002). The forest at
Kanyawara consists of pristine forest to the south and secondary (logged) forest in the
northern and central areas with papyrus swamps scattered throughout. The canopy height
averages 20-30 m (Chapman & Wrangham, 1993). Kanyawara has two distinct rainy
seasons typically occurring from March to May and August to November (Chapman et
al., 1999). The mean rainfall varies widely between months and across years. Chapman et

38
al. (2002) reported the mean annual rainfall from 1990-2001 was 1749 mm with mean
daily temperature averaging 14.9°C minimum and 20.2°C maximum.
Kanyawara offers a variety of food resources for chimpanzees in the form of
fruiting trees, terrestrial herbaceous vegetation (THV), honey, caterpillars, monkeys, and
crops (e.g., bananas, maize, and sugarcane). Availability of fruits, especially drupe
species (Mimusops, Pseudospondias, Uvariopsis), is often seasonal and therefore highly
variable. Fortunately, over 13 Ficus species (fig fruits) and abundant THV serve as
fallback foods during periods of scarcity (Wrangham et al., 1996). In drier months when
food is less abundant at Kanyawara, chimpanzees occasionally crop raid in village
gardens along the forest edge.
Study Community and Subjects
The Kanyawara chimpanzee community was the focus of this study. Struhsaker
(1975) was the first to systematically study primates in the Kibale forest. While his study
focused on red colobus, Struhsaker sporadically encountered chimpanzees. From 1976-
1978, Ghiglieri (1984, 1988) occasionally observed the Kanyawara chimpanzees, while
focusing mostly on the nearby Ngogo community. Isabirye-Basuta (1989) was the first to
systematically study the Kanyawara chimpanzees from 1983-1985, marking the
beginning of habituation without provisioning. Wrangham started the Kibale Chimpanzee
Project in 1987, and the chimpanzees have been continually studied ever since
(Wrangham et al., 1992). By 1990, all males and many females were fully habituated and
could be followed at a distance of 5 m. At the time of the present study, only recently
immigrated females (N = 3) were skittish and difficult to follow.

39
Data were collected from mid-November 2010 through mid-November 2011.
During that time, the community contained 53 individuals consisting of 12 adult males,
17 adult females, 6 subadults, 7 juveniles, and 11 infants (Appendix A). During the
course of the study, the community remained fairly constant. There was one known death
(infant), two known births (one in December 2010 and one in April 2011), no known
female emigration, and one known female immigration the last week of the study period
(not included in community total). One adult male, one adult female, one subadult, one
juvenile, and one infant live strictly in the far northern region of their territory and were
therefore never seen during the study period. Additionally, there were other northern and
southern territory females (and dependents) that were seldom seen. Observability was
therefore a key determinant of focal follow selection. Eleven adult males and six adult
females were selected for all-day focal follows (Table 2.1). These individuals could be
reliably followed alone or in parties at a distance of 5 m.
Behavioral Data Collection
Locating Chimpanzees
I typically followed chimpanzees all day from morning nest to night nest,
typically 7am to 7pm. On days where this was not possible, Kibale Chimpanzee Project
(KCP) field assistants normally nested at least one individual for focal follow the next
day. If no chimpanzees were nested the night before, I generally checked fruiting trees or
followed calls to locate chimpanzees the next morning. Chimpanzees were found on 97%
of days overall. However, during August 2011 chimpanzees were only observed on 17%
of days since most community members were frequently crop raiding in small, dispersed

40
parties near the park boundary. To promote a positive relationship with the local people, I
rarely followed chimpanzees when crop raiding.
Table 2.1. Kanyawara focal demographic information.
Name ID Sex Birthdate Mother Area Observed
Big Brown BB M 1966 C F
Bud PB M 1/20/1995 Stump C F
Eslom ES M 7/2/1994 Ekisigi C F
Johnny AJ M 1974 Lope C F
Kakama KK M 7/15/1985 Kabarole C F
Lanjo TJ M 8/13/1995 Tongo C F
Makoku LK M 1982 Lope C F
Stout ST M 1955 C F
Tofu TU M 1960 N I
Twig PG M 1988 Stump C F
Yogi YB M 1973 C F
Lia AL F 1982 C F
Outamba OU F 1979 C F
Quinto QT F 1992 C F
Rosa LR F 1989 Lope C F
Tongo TG F 1980 C F
Wilma WL F 1992 C F
Note. Birthdates were gleaned from Kibale Chimpanzee Project long-term data records.
Many dates are estimates. Area indicates core region of territory where majority of time
was spent (N=Northern; C=Central). Observed indicates how often individuals were seen
(F=Frequent, on a weekly basis; I=Infrequent, on a monthly basis). Names in italics
indicate natal adult females.

Recording Tools and Equipment
I recorded all behavioral data in pen on a hand-held notebook (e.g., continuous
focal observations) and data sheets (e.g., party composition, aggression, post-conflict
observations). I used a hand-held digital voice recorder on days when I anticipated

41
intense aggression at a high rate. For example, days when parous females were in estrous
predicted an unusually high rate of aggression (Muller, 2002: no estrous female in party =
1.68 conflicts/hr, periovulatory female in party = 3.2 conflicts/hr). Then, I immediately
transcribed voice-recorded data in the evenings and transferred it to data sheets. A Timex
wristwatch kept real time throughout observations. Bushnell binoculars were useful to
continue recording data while chimpanzees were in the canopy. I used a Garmin 60csx
GPS to navigate through the forest, effectively use the trail system, and mark important
coordinates (e.g., night nests, aggression observations, intergroup encounters, etc.). I used
a Nikon digital camera to take opportunistic photographs and record HD video.
Focal (Baseline) Observations
I collected all-day continuous focal follows (Altmann, 1974) on 11 adult male and
six adult female chimpanzees. Daily focal selection and length of observation was largely
based on visibility, current representation in the data set, and previous focal
observation(s) characteristics (e.g., activity/context, group composition, etc.). While my
goal was to follow all focal candidates equally, this proved difficult to achieve.
Observations of females, in particular, were unreliable because they often joined the party
midday for short periods of time after a focal had already been selected. If a focal
traveled out of sight and could not be found after searching for 30 min, I chose another
focal for the remainder of the day. Occasionally, the focal fissioned from the main party.
If the focal remained alone for more than an hour and the main party could easily be
located, I abandoned the current focal and a different focal was selected.

42
I collected the following types of data during focal follows. (1) I continuously
recorded all activities, interactions, and vocalizations for the focal chimpanzee in real
time to the nearest second. (2) I recorded all affiliative (friendly contact behaviors),
aggressive (threats, non-contact aggression, contact aggression), and submissive (e.g.,
crouching, fleeing) interactions of the focal with other party members in detail. I recorded
the time, duration, behavior(s), the identity of the social partner(s) and, when possible,
the focal as the initiator or recipient of the behavior. (3) I documented the identity of all
females and the size of their genital swellings. Scores were assigned as follows: 1 –
completely deflated with maximum wrinkling, 2 – partially deflated or wrinkled with
partial swelling, or 3 – maximally tumescent (or swollen) with no wrinkling (Deschner et
al., 2003). Estrous females were defined as having maximal swelling tumescence.
Nonestrous females were defined as having deflated or partial swellings (Muller et al.,
2011). (4) Every 10 min, I performed an instantaneous scan sample (Altmann, 1974) of
the focal’s party size and composition. I recorded the presence of individuals in the party,
adult party members’ proximity to the focal (in meters using visual estimation), grooming
partners, play partners, and the party’s general activity using a majority rule. Party size
was defined as the number of individuals in view of the observer or field assistants in the
same place at the same time (Newton-Fisher et al., 2000). The measure of conspecific
proximity relative to the focal (using a visual estimation) was recorded at different
distances to assess baseline dyadic association at the level of the party, within 10 m, and
within 5 m (Gilby & Wrangham, 2008). See Appendix B for behavioral ethogram.

43
Aggression Observations
I collected directed aggression data ad libitum (Altmann, 1974). Directed
aggression (henceforth referred to as a conflict) was defined as aggressive behavior
targeted toward at least one individual. If a conflict occurred during focal observations, I
temporarily abandoned the focal to collect conflict (and subsequent post-conflict) data. I
assigned each conflict a unique identification number. I recorded the following
information about the conflict:
1) The date, time, duration, and location (forest floor vs. tree, GPS) of the
conflict. The conflict duration started with the first aggressive behavior
toward another individual and ended with submission, flight, third-party
intervention, reciprocal screaming, or a non-aggressive behavior.
2) The aggressor and victim identities and sex. I labeled the first individual to
direct an aggressive behavior toward another individual as the aggressor, and
the recipient of this aggressive action as the victim.
3) The type of aggression: threat (non-contact) or agonistic (contact).
4) The intensity of aggression was then given a score of 0-3.
a. A score of zero was assigned to non-contact threats (e.g., displays,
chases, throwing objects, aggressive vocalizations, etc.).
b. A score of one was assigned to contact aggression containing only one
physical element but never a bite (e.g., a single hit, kick, grab).

44
c. A score of two was assigned to contact aggression containing multiple
physical elements of moderate intensity but never a bite (e.g., multiple
hits, kicks, grabs).
d. A score of three was assigned to contact aggression of multiple
physical elements of high intensity possibly with bites. The aggression
was automatically scored as a three anytime biting, wounding, or
bleeding occurred.
5) The directionality of aggression: unidirectional (no retaliation by victim) or
bidirectional (victim retaliates).
6) The winner of the conflict, determined by resource monopolization and/or
submission. Conflicts with no clear winner were considered a draw.
7) The context in which the aggressive interaction occurred (subcategories from
Muller, 2002; general context categories from Wittig & Boesch, 2003a).
a. In food contexts, fights were over food accessibility. Subcategories
include meat competition or plant competition. Meat competition was
scored when an aggression occurred between the possessor of meat
and another individual. Plant competition was scored when one or both
opponents were feeding on a plant food resource.
b. In sex contexts, fights were either with (sexual coercion) or over
estrous females (sexual competition).
c. In social contexts, fights were over dominance, social partner access,
or other situations of social excitement. Subcategories include

45
reunions (aggression occurring within 5 min of party fusion),
protection (of offspring), dominance challenge, social partner
competition, or general social excitement (hear distance calls, arrive at
fruiting tree).
d. If the context was not clear, it was labeled unknown.
8) Identities of others involved and/or present (measured by relative visibility)
during the act of aggression. If involved, also the nature of their involvement
(e.g., giving support to either party member) was recorded.
Post-Conflict (PC) Observations
I investigated post-conflict behaviors following the observational procedures
described by de Waal and Yoshihara (1983) and Aureli et al. (1989) but modified them as
necessary for suitability in a wild setting (Whiten & Arnold, 2001; Wittig & Boesch,
2003a). Immediately following the end of the conflict, I began a 10-min PC observation
on either the aggressor or the victim or both depending on visibility. A 10-min PC
observation period was chosen because research has demonstrated that most post-conflict
partner interactions, even in the wild, typically occur within this time frame (Arnold &
Whiten, 2001). Beyond the 10-min observation period, behaviors do not significantly
differ from baseline data (Kutsukake & Castles, 2004). If aggression recurred within the
10 min, I restarted the 10-min PC observation following the last bout of aggression
(Fuentes et al., 2002; Preuschoft et al., 2002). I labeled the recurrence of aggression as
renewed (same opponents) or redirected (aggressor or victim attacked a third party).
Renewed data was then collapsed with the original bout of aggression by summarizing

46
the information. For example, the total duration of aggression combined the original
aggression duration with the renewed aggression duration, the overall intensity was the
highest intensity observed, the overall directionality was bidirectional if the victim
retaliated at any time or initiated the renewed aggression, and the winner was the
individual that monopolized the resource or did not show submission after the final
aggressive encounter. Redirected aggression was a novel conflict partner pairing and was
given a unique observation number.
During PC observations, I used the same behavioral recording and coding
methodology as previously described for baseline focal observations. I recorded all
interactions between opponents and third party bystanders in detail. In addition, I
continuously recorded the relative distance of the opponent to the focal throughout the
observation (Arnold & Whiten, 2001). If the focal moved out of sight for longer than one
minute and their opponent was relatively close (within 5 m) and was also not visible, I
abandoned the PC to ensure the occurrence of reconciliation was not underestimated and
excluded this data from analyses. See Appendix B for behavioral ethogram, and
Appendix C for operational definitions.
Self-Directed Behaviors (SDBs)
During each type of observation (baseline and PC), I recorded all self-directed
behaviors in frequency or duration using the behavioral definitions provided by Aureli &
de Waal (1997) with slight modification. The three possible SDBs are described below:
1) Self-grooming – Visual inspection of self using the mouth, hand(s), or tool to
pick through and/or bush aside one’s own hair. Duration of the self-grooming

47
bout was recorded. Fifteen seconds without grooming movements [or the
interruption of self-grooming behavior by a different behavior followed by a
return to self-grooming behavior] constituted a new bout of self-grooming
(Arnold & Whiten, 2001).
2) Self-scratching – Repeated movements of the hand(s) in which the
chimpanzee draws one’s own fingertips across own hair and/or skin. Gentle
scratching involves small movements of only the hand or fingers across the
hair or skin. Rough scratching is much more exaggerated including visible
movements of the arm across hair or skin. Kutsukake (2003) noted that rough
scratching was a more obvious movement, easily observable in wild
population where visibility is often restricted. In proximity, rough scratching
is also audible. Since rough scratching is a more reliable predictor of anxiety
in primates (te Boekhorst et al., 1991), I did not record gentle scratches.
Rough scratch frequency was recorded by counting the number of full arm
movements during the scratching bout. The majority of rough scratches
occurred in bouts, meaning multiple scratches occurred in a single bout. If
scratch frequency was not accurately counted, that bout was labeled with an
‘x’. Five-seconds without scratching movements [or the interruption of self-
scratching behavior by a different behavior followed by a return to self-
scratching] constituted a new bout of self-scratching (Arnold & Whiten,
2001).

48
3) Yawning – A gaping movement of the mouth generally exposing the teeth.
The number of yawns was recorded and a frequency was calculated.
Particular attention was paid to rough self-scratching behavior since previous studies
have shown this to be the most frequent and most reliable behavioral indicator of stress
(Aureli & de Waal, 1997; Baker & Aureli, 1997; Kutsukake, 2003; te Boekhorst et al.,
1991). When chimpanzees were in the trees, increased distance paired with leafy foliage
often obstructed clear views of the focal. SDBs were therefore never recorded when
chimpanzees were in the trees.
Data Analysis and Statistical Procedures
Dominance Hierarchy
In the wild, male chimpanzees outrank females, and adults outrank non-adults.
Males have a distinct linear hierarchy, which is classically determined by greeting
vocalizations (de Waal, 1978) commonly known as pant grunts. In the current study,
however, insufficient pant grunt data limited hierarchy distinction to alpha, high, mid,
and low rank categories. I therefore combined pant grunt exchange data with conflict
winner data to determine the linear hierarchy among adult males and females at
Kanyawara. I combined the two dominance matrices (Winner-Loser Matrix and Pant
Grunt Directionality Matrix) to test the linearity for males and females separately. I tested
the combined matrix using the I&SI method (de Vries & Appleby, 2000; Jameson et al.,
1999) in MatMan 1.1 (de Vries et al., 1993). This method was chosen because it 1)
accounts for the relative success of individuals in agonistic interactions, 2) can predict

49
dominance relationships for dyads that never interacted, 3) predicts the linearity of
relationships, and 4) assigns a rank order.
For analysis, tested the linear rank and also grouped rank. I grouped adult males
into three categories based on their individual rank: high, mid, and low (Table 3.3). I also
calculated the linear rank difference between the aggressor and the victim and then
categorized the dyad’s rank difference as small (1-3), medium (4-6), or large (7-10). For
example, as derived from Table 3.3, the linear rank difference for the alpha, KK, and the
lowest ranking individual, TU, was 10, and the rank difference was therefore categorized
as large. Females were lumped into a single category, which was ranked lower than all
male categories.
Linear rank was used to determine if the winner (determined by the opponent’s
action flee, crouch, or other subordinate behavior) of the dyadic conflicts (triadic
conflicts were excluded from this analysis) was dominant or subordinate to their
opponent. The winner was the individual that either 1) obtained the disputed resource or
2) caused their opponent to flee, crouch, or use another submission behavior. The winner
was labeled ‘dominant’ if their linear rank was higher than their opponent’s rank,
‘subordinate’ if their linear rank was lower than their opponent’s rank, or ‘unknown’ if
the conflict occurred between two females (insufficient data to determine female
linearity). For conflicts between a male and a female, the male was always considered
dominant to the female, and thus a male winner was labeled ‘dominant’ and a female
winner was labeled ‘subordinate.’

50
Proximity Association
Dyadic pairwise association indices (measured by affiliation time) were
calculated using party composition and focal proximity 10-min scan data. I calculated
one temporal association index (party-level association) and two spatial association
indices (within 5 m and nearest neighbor association). Calculations for these indices were
adopted from Cairns and Schwager (1987) and Gilby and Wrangham (2008).
Party-level association measures the dyad’s tendency to spend time in the same
party. I calculated party-level association using the Simple Ratio Index proposed by
Cairns and Schwager (1987). The party level association index (PLA) is calculated by
dividing the total number of 10-min scans that a dyad is in the same party together by the
total number of 10-min scans that the individuals are observed separately:
P
AB

P
A
+ P
B
- P
AB

where P
AB
= the total number of scans containing individuals A and B together, P
A
= the
total number of scans containing A, and P
B
= the total number of scans containing B.
Unlike Gilby and Wrangham (2008), I included all scans, not just one scan from each
unique party composition since they found no statistical difference between the two
methods. To make comparisons between dyads and sexes meaningful, the party-level
association score for each dyad (PLA
AB
) was then standardized against the PLA mean of
all adult dyads by dividing PLA
AB
by the PLA
MEAN
. Standardized values greater than one
indicated that dyad was found in the same party more often than the average dyad.
PLA
AB
=

51
Within 5 m association measures the dyad’s tendency to spend time within 5 m
while in the same party when one individual is the focal. This index is formulated to be
independent of party-level association using the following formula:
A
f
(B
5
) + B
f
(A
5
)
A
f
(B
p
) + B
f
(A
p
)
where A
f
(B
5
) = the number of scans A was the focal and B was within 5 m of A, B
f
(A
5
)
the number of scans B was the focal and A was within 5 m of B, A
f
(B
p
) = the number of
scans A was the focal and B was in the same party, and B
f
(A
p
) = the number of scans B
was the focal and A was in the same party. Again, to make comparisons between dyads
and sexes meaningful, the within 5 m association index was standardized against the
mean of all dyads by dividing 5M
AB
by the 5M
MEAN
. Standardized values greater than one
indicated that dyad was found within 5 m of one another more often than the average
dyad.
Nearest neighbor association measures the dyad’s tendency to spend time as the
focal’s nearest neighbor when within 5 m of one another while in the same party when
one individual is the focal. This index is formulated to be independent of within 5m
association using the following formula:
A
f
(B
nn
) + B
f
(A
nn
)
A
f
(B
5
) + B
f
(A
5
)
where A
f
(B
nn
) = the number of scans A was the focal and B was the nearest neighbor of
A, B
f
(A
nn
) = the number of scans B was the focal and A was the nearest neighbor of B,
A
f
(B
5
) = the number of scans A was the focal and B was within 5m when in the same
5M
AB
=

NN
AB
=

52
party, and B
f
(A
5
) = the number of scans B was the focal and A was within 5m when in
the same party. Again, to make comparisons between dyads and sexes meaningful, the
nearest neighbor association index was standardized against the mean of all dyads by
dividing NN
AB
by the NN
MEAN
. Standardized values greater than one indicated that dyad
was each other’s nearest neighbor when within 5 m more often than the average dyad.
Using the three independent measures of association described above, I then
calculated a composite association index (CAI) as outlined in Gilby and Wrangham
(2008). CAI combines both temporal (party-level association) and spatial (within 5 m and
nearest neighbor association) levels to evaluate each dyad’s overall association tendency.
The CAI is calculated by averaging the mean values for the three association indices for
each dyad:
(PLA
AB
/PLA
MEAN
) + (5M
AB
/5M
MEAN
) + (NN
AB
/NN
MEAN
)
3
where PLA
MEAN
= mean of PLA
AB
for all dyads, 5M
MEAN
= mean of 5M
AB
for all dyads,
NN
MEAN
= mean of NN
AB
for all dyads. For some dyads, it was impossible to calculate a
within 5 m and/or a nearest neighbor association index because that dyad was never
observed within 5 m. The CAI for those dyads was calculated by using only the party-
level association index or averaging only the party-level and within 5 m association
indices, respectively.
Contact Affiliation
During focal observations, I recorded the duration of all affiliative behavior with
others. Affiliative behavior included grooming, play, contact sit, mount, touch, sexual
CAI
AB
=

53
contact, etc. I then used all friendly behavioral exchanges to calculate composite contact
affiliation scores for each dyad. To calculate this, I used the total contact duration for
each dyad and divided by the total observation time for that dyad when one individual
was the focal and the other was in the same party (Preuschoft et al., 2002). I only used
affiliation data collected during focal (baseline) observations. Affiliation data collected
during post-conflict observations was not included in this calculation.
GRM
AB
+ PL
AB
+ CS
AB
+ MT
AB
+ TCH
AB
+ SC
AB

A
f
(B
p
) + B
f
(A
p
)
where GRM
AB
= total duration of A and B grooming, PL
AB
= total duration of A and B
playing, CS
AB
= total duration of A and B in body contact while sitting, MT
AB
= total
duration of A and B mounting, TCH
AB
= total duration of A and B touching, SC
AB
= total
duration of A and B in sexual contact, A
f
(B
p
) = total observation time when A was focal
and B was in the party, and B
f
(A
p
) = total observation time when B was focal and A was
in the party. To make comparisons between dyads and sexes meaningful, the contact
affiliation score for each dyad (CAS
AB
) was then standardized against the CAS mean of
all adult dyads by dividing CAS
AB
by the CAS
MEAN
. Standardized values greater than one
indicated that dyad affiliated more often than the average dyad.
Agonistic Support
During agonistic interactions, I recorded ad libitum coalitionary attacks and
favored interventions resulting in joint defense. Coalitionary attacks were defined as two
or more aggressors jointly attacking (measured by simultaneous or alternating aggressive
behavior) another individual(s). Favored interventions were defined as a third party
CAS
AB
=

54
joining a conflict to support/defend the victim thereby resulting in joint defense against
the aggressor. I calculated each dyad’s total agonistic support (AS) by summing the
frequency of coalitionary attacks and favored interventions and divided this value by
party observation time.
CA
AB
+ FI
AB

P
AB

where CA
AB
= total number of coalitionary attacks, FI
AB
= total number of favored
interventions, and P
AB
= total observation time when A and B were in the same party.
To make comparisons between dyads and sexes meaningful, the agonistic support score
for each dyad (AS
AB
) was then standardized against the AS mean of all adult dyads by
dividing AS
AB
by the AS
MEAN
. Standardized values greater than one indicated that dyad
provided agonistic support more often than the average dyad.
Aggression Level
During focal observations, I ad libitum recorded the aggressors and victims of all
aggression, which occurred between adults, even when post-conflict observations were
not possible. From this data, I calculated each dyad’s total aggression level (AL)
separately dependent upon the identities of the aggressor and victim. I calculated AL
AB

by summing the frequency of fights when A was the aggressor and B was the victim, and
AL
BA
by summing the frequency of fights when B was the aggressor and A was the
victim. I divided each of these values by the dyad’s total party observation time.
AGG
AB
AGG
BA

P
AB
P
AB

AS
AB
=

AL
AB
=
AL
BA
=

55
where AGG
AB
= total number of conflicts initiated by A against B, AGG
BA
= total
number of conflicts initiated by B against A, and P
AB
= total observation time when A
and B were in the same party. To make comparisons between dyads and sexes
meaningful, the aggression level score for each dyad (AL
AB
and AL
BA
) was then
standardized against the AL mean of all adult dyads by dividing AL
AB
and AL
BA
by the
AL
MEAN
. Standardized values greater than one for a dyad indicated that aggressor fought
against that victim more often than the average dyad.
Standardization of Indices
The above relationship parameters of association (PLA, 5M, and NN), affiliation,
agonistic support, and aggression level provided indices that were raw values for each
dyad. For each variable, I described how these measures were standardized against the
mean. For modeling analyses, I did not use the standardized scores. Instead, I grouped the
scores into appropriate categories for comparison (e.g., mutual, one-sided, neural
relationships). I chose two methods of comparison for the four relationship categories
based on the current literature because the first method distinguishes between mutual and
one-sided relationships whereas the second method does not.
The first method was the Preferred Social Partners (PSPs) method proposed by
Gilby and Wrangham (2008). This method distinguishes between mutual and one-sided
relationships. An example of a one-sided relationship would be chimpanzee A having a
strong affinity for chimpanzee B, but B not having a strong affinity for A. Mutual
relationships where both A and B have a strong affinity for one another are considered
stronger relationships than the former. To distinguish between the two relationship types,

56
mutual relationships were those that had a standardized value (e.g., PLA
AB
/PLA
MEAN
) a
half standard deviation greater than the mean of all dyads containing A and the mean of
all dyads containing B. Dyads that achieved this criterion associated/affiliated more often
than was typical for either member and was thusly labeled ‘mutual associates’ (associates
also referred to as PSPs). Dyads where only chimp A or B had a standardized value a half
standard deviation greater than their mean were labeled ‘one-sided associates.’
Remaining dyads were labeled ‘neutral associates.’ This comparison was made for each
dyad for each association (PLA, 5M, and NN) and relationship (CAS, AS, and AL)
parameter. For the combined association index (CAI), mutual PSPs were dyads that
exhibited mutual association for at least two of the three independent indices (PLA, 5M,
and NN). One-sided PSPs were dyads that exhibited association for either chimp A or B
for at least two of the three relationship parameters. Neutral partners were all remaining
dyads.
After the PSP status was calculated for each dyad for the three affiliative (CAI,
CAS, and AS) parameters, each dyad was then given an overall mutual PSP score ranging
from 0 to 3 with 0 representing the weakest relationship and 3 representing the strongest.
The overall mutual PSP calculation combined the above association and relationship
parameters into one score, Dyads with an overall mutual PSP score of 0 did not share
PSP status for any of the three parameters (association, contact affiliation, and agonistic
support). Dyads with an overall mutual PSP score of 1 shared PSP status on one
parameter. Dyads with an overall mutual PSP score of 2 shared PSP status on two

57
parameters. Dyads with an overall mutual PSP score of 3 shared PSP status on all three
parameters.
Aggression level PSPs were used as indicators of attack risk between A and B. A
dyad was categorized as ‘target of A’ if the standardized value (AL
AB
/AL
MEAN
) of
aggression directed by A against B was a half standard deviation greater than the mean of
all dyads containing A and the mean of all dyads containing B. If aggression directed by
B against A fit this rule, the dyad was categorized as ‘target of B.’ If A was a target of B
and B was a target of A, this dyad was categorized as ‘mutual targets.’ If neither A nor B
was a target of the other, this dyad was categorized as a ‘neutral nontarget’ dyad. This
categorization was also applied to bystanders that interacted with either A or B during the
post-conflict period. Bystanders were categorized as ‘aggressor’s target’ if they were a
target of the aggressor but not the victim, ‘victim’s target’ if they were a target of the
victim but not the aggressor, ‘target of both’ if they were a target of both the aggressor
and the victim, or ‘nontarget’ if they were not a target of the aggressor or the victim.
Method adapted from Romero and de Waal (2010).
The second method was the quartile method proposed by Cords and Aureli
(2000). This method has been used extensively in post-conflict research and is the
common method for assessing dyadic affiliation (Arnold & Whiten, 2001; Kutsukake &
Castles, 2004; Koski et al., 2007a). This method used quartile calculations to compare
raw indices between dyads and sexes. I used the raw indices (not the standardized) for
each relationship parameter to calculate the top quartile (above 75%), the mid quartile
(25%-75%), and the bottom quartile (below 25%) for that measure. Scores that fell within

58
the top quartile were labeled ‘strong/affiliative’, those in the middle were
‘moderate/neutral’, and those in the bottom were ‘weak/non-affiliative.’ Unlike the PSPs
method, the quartile method does not distinguish between mutual and one-sided
relationships.
Results from both methods were then tested in the multivariate analysis models to
determine which standardization method was a better predictor of the dependent variable.
This procedure is described in more detail in subsequent sections.
Reconciliation
Reconciliation was defined as affiliative contact between conflict partners within
10 min after the last aggressive behavior. In cases of renewed and redirected aggression,
the 10-min period was restarted after aggression again ceased. However, this definition
alone does not indicate whether reconciliation was a true product of the previous conflict
or merely a product of chance.
Traditionally, post-conflict (PC) studies have compared post-conflict observations
to a matched control (MC) to answer the question: Is the behavior a product of the
conflict or merely a product of chance (de Waal & Yoshihara, 1983)? The MC is an
observation period that is collected in the absence of aggression within several days after
the PC. In captivity, the MC usually seeks to match the PC on the following
characteristics: time of day, observation duration, proximity of conflict partners,
behavioral context, weather, etc. The basic premise is for the MC to mirror the PC in as
many characteristics as possible without aggression. Then the MC is compared to the PC.
If the conflict partners interact earlier or only in the PC when compared to the MC, the

59
interaction is a product of the conflict and is subsequently labeled reconciliation. Conflict
partners are then labeled as ‘attracted’ to one another after the conflict. If the conflict
partners interact earlier or only in the MC when compared to the PC, the interaction is
considered a product of chance. Conflict partners are then labeled as ‘dispersed.’ While
this method operates remarkably well under captive conditions (where party composition
is relatively constant), it is not suitable for wild chimpanzee populations (where parties
are constantly changing).
As it is never guaranteed that party composition or opponent proximity in the wild
will ever be perfectly matched to the original PC observation, wild chimpanzee
researchers can therefore waste valuable observation time waiting for an appropriate
matched-control (MC) observation to occur. Although strict criteria may easily be met in
captive conditions, the unpredictable nature of field studies provides an added challenge
(Arnold & Whiten, 2001; Kutsukake & Castles, 2004). Due to the above concerns,
instead of conducting an independent MC observation for each specific PC observation, I
used focal data to calculate a mean baseline affiliation (MBA) rate for each dyad. While
this method has been applied to monkey post-conflict studies (macaques: Judge, 1991;
baboons: Silk et al., 1996), this is its first application to chimpanzees.
I used the following method to calculate each dyad’s MBA rate. I used the
frequency of affiliative behavior (e.g., grooming, playing, touching, contact sitting,
mounting, etc.) during focal observations as the comparative measure. I concluded
duration would not be an appropriate measure of comparison for the following reasons:
1) the duration of affiliative behavior in the PC was not of importance, 2) only the

60
occurrence of affiliative behavior in the PC was needed to determine reconciliation, and
3) a corrected frequency could be calculated to control for the duration of affiliative
interactions during focal observations. In focal observations, I could not use the raw
affiliative interaction frequency because duration affected the raw score. For example, a
dyad that groomed for an hour would get a raw frequency of one, whereas a dyad that
groomed for 2 min, rested for 5 min, and then groomed again would get a raw frequency
of two. This method was biased toward ‘sporadic’ affiliation and did not accurately
reflect the dyad’s affiliative interaction frequency, which is why I adapted this method to
incorporate a duration component.
Instead of using the raw frequency of affiliation (Silk et al., 1996), I developed a
10-min sampling procedure to correct for ‘marathon’ and ‘sporadic’ affiliation. I selected
10-min because the PC observation period was also 10-min. When affiliative behavior
began, I started the 10-min sampling period. The dyad then was given a score of one for
every subsequent 10-min period containing affiliative contact. Using this method,
‘sporadic’ groomers could stop interacting for several minutes within the 10 min, but still
received a score of one if they did not interact beyond the 10-min period. ‘Marathon’
groomers, on the other hand, received a one for each 10-min period of continuing
affiliation (e.g., dyad’s interacting for 35 min would receive a score of four). The dyad’s
corrected affiliation frequency (CAF) was then normalized by dividing by the dyad’s
observation time. I calculated three different MBA rates based on different observation
time criteria: MBA
PARTY
, MBA
VIEW
, MBA
10M
.
MBA
PARTY
was calculated using the following formula:

61

ΣCAF
AB

A
f
B
p
+ B
f
A
p

where ΣCAF
AB
= the sum of corrected affiliation frequencies of A and B, A
f
B
p
= the
number of scans A was the focal and B was in the same party, and B
f
A
p
= the number of
scans B was the focal and A was in the same party.
Both MBA
VIEW
and MBA
10M
controlled for dyad distance in focal observations.
For example, dyads may be in the same party but may not affiliate because they are too
far apart or are not aware of each other’s presence. Presumably, dyads with visual contact
have an increased probability of affiliation. By controlling for partner proximity, it makes
the baseline focal affiliation data more applicable to post-conflict data because former
opponents are typically in close proximity (Kappeler & van Schaik 1992).
MBA
VIEW
was calculated using the following formula:
ΣCAF
AB

A
f
B
v
+ B
f
A
v

where ΣCAF
AB
= the sum of corrected affiliation frequencies of A and B, A
f
B
v
= the
number of scans A was the focal and B was in view of A, and B
f
A
v
= the number of
scans B was the focal and A was in view of B.
MBA
10M
was calculated using the following formula:
ΣCAF
AB

A
f
B
10
+ B
f
A
10

MBA
PARTY(AB)
=

MBA
VIEW(AB)
=

MBA
10M(AB)
=

62
where ΣCAF
AB
= the sum of corrected affiliation frequencies of A and B, A
f
B
10
= the
number of scans A was the focal and B was within 10m of A, and B
f
A
10
= the number of
scans B was the focal and A was within 10m of B.
Mean post-conflict affiliation (PCA) rates were calculated similarly to MBAs,
but used only the PC observation data. Since post-conflict observations were always 10
min in duration, a corrected affiliation frequency was not necessary. Instead, I used the
raw affiliation frequency (RAF) for each dyad. Opponents either received a one if they
affiliated during the PC, or a zero if they did not. I then divided the total RAF by the total
number of PC observations for each dyad. PCAs were calculated using the following
formula:
ΣRAF
AB

PC
AB

where ΣRAF
AB
= the sum of raw affiliation frequencies of A and B during post-conflict
observations, and PC
AB
= the total number of PC observations between A and B.
For dyads that affiliated at least once in a PC observation period, I then compared
that dyad’s PCA rates to their MBA rates to determine if conflict partners were
reconciling as a product of the conflict or chance. I used the following equation to answer
this question (adapted from Silk et al., 1996; Wittig & Boesch, 2003a):
PCA
AB

MBA
AB

where PCA
AB
= the mean post-conflict affiliation rate for A and B, MBA
AB
= the mean
baseline affiliation rate of A and B during focal observations. I calculated the PC
PC Relative
AB
=

PCA
AB
=

63
Relative
AB
three different times using the party, in view, and within 10m MBA value as
the denominator. All three MBA variants yielded the same pattern for each dyad.
Scores larger than one indicated that reconciliation occurred and were
subsequently labeled attracted. Scores smaller than one indicated the affiliation was not
true reconciliation and were subsequently labeled dispersed. I tested the statistical
significance of these values by comparing the PCA to the MBA using Wilcoxon
matched-pairs signed-ranks test (Silk et al., 1996).
In post-conflict literature, the standard measure of reconciliatory frequency is the
Veenema et al. (1994) Corrected Conciliatory Tendency (CCT):
A
AB
- D
AB

T
AB

where A is the number of attracted pairs of individuals A and B, D is the number of
dispersed pairs of individuals A and B, and T is the total number of PC-MC pairs (total
number of conflicts substituted for PC-MC pairs) of individuals A and B. CCTs were
calculated at the individual and group level. Nonparametric statistical tests, Wilcoxon
signed-rank and Mann-Whitney U tests, were used to test for significant differences in
the frequency of attracted verses dispersed pairs and in individual mean CCTs between
male-male, male-female, and female-female dyads, respectively.
Bystander Affiliation (Consolation and Appeasement)
Offered (or true) consolation was defined as affiliative contact offered by a third
party bystander to the victim. In cases where the conflict was bidirectional and the initial
aggressor became a victim, both conflict partners were eligible for offered consolation.
CCT
AB
=

x 100

64
Solicited consolation was defined as affiliative contact initiated by the victim towards a
third party bystander (Wittig & Boesch, 2003a). Appeasement was defined as affiliative
contact initiated by the third party bystander towards the aggressor (de Waal & Aureli,
1996). All bystander affiliations had to meet two requirements for inclusion in analyses.
Bystander affiliation must occur 1) within the 10-min post-conflict period, and 2) either
before or in the absence of reconciliation.
To determine whether either form of consolation (offered or solicited) was a
product of the conflict or chance, the same procedure as described for reconciliation was
used to investigate the consolation relative affiliation rate. Similar to CCT calculations,
mean individual triadic contact tendencies (TCTs) were calculated for each form of
consolation (Call et al., 2002). TCTs were calculated using the following formula:

A
AB
- D
AB

T
AB

where A is the number of attracted pairs of individuals A and B, D is the number of
dispersed pairs of individuals A and B, and T is the total number of PC-MC pairs (total
number of conflicts substituted for PC-MC pairs) of individuals A and B. Nonparametric
statistical tests, Wilcoxon signed-ranks and Mann-Whitney U tests, were used to test for
significant differences in the frequency of attracted verses dispersed pairs and in
individual mean TCT between male-male, male-female, and female-female dyads,
respectively. Appeasement analyses followed a similar model.

TCT
AB
=

x 100

65
Multivariate Analyses
Similar to Wittig and Boesch (2003a) and Fraser (2008), I employed a
multivariate approach using Generalized Linear Mixed Models and Linear Mixed Models
to investigate what conflict (e.g., duration, directionality, intensity, context, etc.) and
relationship (e.g., association, affiliation, agonistic support, etc.) variables predicted
conciliatory behavior (reconciliation and consolation) and aggression (renewed and
redirected) in the post-conflict period. Details regarding the specific variables tested are
explained in the appropriate chapter methods sections. All analyses were conducted at the
dyadic level.
Generalized Linear Mixed Models (GLMMs) and Linear Mixed Models (LMMs)
are extensions of the Generalized Linear Model, which is a generalization of the Logistic
Regression (McCullagh & Nelder, 1989). GLMMs were used to analyze binominal
dependent variables (e.g., reconciled or not), and LMMs were used to analyze continuous
dependent variables (e.g., reconciliation latency). GLMMs and LMMs both allow fixed
and random variables to be fitted to a predictive model. Independent variables are the
fixed variables that act as the predictors of the dependent variable(s). Identifying random
effect variables are a major advantage of mixed models because differences between
random effect levels are of no statistical interest (Tabachnick & Fidell, 2007). Random
variables are presumably values drawn from a larger population of values that
conceptually could have been selected. Basically, this means that the actual random
variables (e.g., subject identity) represent a random sample of all possible values (e.g.,
individuals) of that variable. Mixed models allow for distinction of the random effects in

66
analyses and therefore control for over- or under- representation of values. In studies that
deal with repeated measures, as in this study, this feature is incredibly important because
all individuals are often not equally represented in the various conditions. For example,
many individuals were involved in multiple post-conflict observations, sometimes
involving the same dyad. However, individual involvement varied greatly as some
individuals contributed more to the post-conflict data than others. Both GLMM and
LMM control for this variation and potential problem of pseudoreplication, and allow
dyadic variability to be detected (Goldstein, 2011). Random variables with variance close
to zero meant behavioral responses were relatively independent of the random variables
(e.g., subject identity).
For all mixed model analyses, the best model was built and chosen using the
Akaike’s information criterion (AIC), which identifies the most parsimonious model that
best explains dependent variable variance by comparing the adequacy of several models.
The AIC also penalizes for the number of independent variables in the model (Akaike,
1973). The lower the AIC score, the better the model (and the significant independent
variables) predicts the values of the dependent variable (Tabachnick & Fidell, 2007). I
used a step-up strategy whereby fixed (independent) variables were sequentially added to
build the best model. This means each fixed variable was first tested against the
dependent variable. I then sequentially added fixed variables with lower AIC scores to
the model in score order until the best model was achieved. The level of significance was
always set at p < 0.05.

67
The estimate coefficient β is an indication of effect strength that the independent
variable has on the dependent variable. Values of β farther away from 0 indicate a strong
effect of the independent variable on the dependent variable. Positive β indicates an
independent variable effect in favor of the first parameter of the dependent variable (e.g.,
reconciled), whereas a negative β indicates the opposite pattern (e.g., not reconciled). The
e
β
represents the odds ratio. For binomial independent variables, an e
β
of 1 indicates that
the dependent variable (e.g., reconciliation) is equally likely in both conditions (e.g.,
unidirectional or bidirectional), whereas greater than 1 indicates the dependent variable
(e.g., reconciliation) is e
β
more likely in the first parameter (e.g., unidirectional). For
continuous independent variables, an e
β
greater than 1 indicates the e
β
was more likely
with each unit increase of the independent variable (e.g., the longer the conflict lasted).
Unless otherwise stated, all multivariate analyses were performed in R version
3.0.1 using the lme4 package with a logit link function (Bates et al., 2013; R Core Team,
2012). Specifically, binominal dependent variables were tested via GLMMs and were
conducted using the ‘glmer’ function with a binominal family and logit link function in
the R package ‘lme4’. Continuous dependent variables were tested via LMMs and were
conducted using the ‘lmer’ function in the R package ‘lme4’ (Bates et al., 2013). In this
package, the lmer function does not generate p values. Instead, I used the ‘pvals.fnc’
function that uses a Markov chain Monte Carlo simulation with 10,000 iterations in the
‘languageR’ package to calculate the p values for these models (Baayen, 2011). Post-hoc
analyses, to test for the significant differences between the various levels of categorical
and ordinal variables, were conducted using the ‘glht’ function in the ‘multcomp’

68
package (Hothorn et al., 2008). All independent (predictor) variables in all models were
tested for multicollinearity between predictors using the variance inflation factor (VIF).
To test each predictor’s VIF, I used the ‘vif.mer’ function, which is tailored to ‘lmer’
models in the ‘lme4’ package created by Austin Frank (https://raw.github.com/aufrank/R-
hacks/master/mer-utils.R). To be conservative, predictors with VIF scores higher than
five were considered to be collinear with other variables and were thus not included in
that particular model (Craney & Surles, 2002; Rogerson, 2001).
Multivariate analyses in post-conflict research. Mixed models, like GLMM and
LMM, offer several advantages over other multivariate models especially in the field of
post-conflict research. Mixed models 1) control for participation variation in the data, 2)
permit unbalanced designs, 3) control for the number of independent variables, and 4)
control for random effects like focal and partner identity. While mixed models have been
used extensively in other areas of behavioral research (Deschner et al., 2004; Engh et al.,
2006), it was not until recently that they were implemented in post-conflict research.
Kutsukake and Clutton-Brock (2010) were the first to use mixed models in conflict
management research when studying aggression in meerkats. Fraser & Aureli (2008)
were the second to demonstrate the advantage of using mixed models to study post-
conflict behavior in a captive chimpanzee population. The present study is the first to
employ mixed model analyses to study post-conflict behavior in a wild chimpanzee
population. In the wild, this is particularly important because chimpanzee location and
behavior is often variable, making it much more difficult to meet statistical assumptions.

69
Mixed models, however, can control for many of these random factors, such as unequal
representation in the data set.

70
Chapter 3

Agonistic and preferred social relations among Kanyawara
chimpanzees (Pan troglodytes): Revisited and expanded

71
Introduction
Aggression in chimpanzees has always been a topic of great interest because of
their close genetic relatedness to humans (King & Wilson, 1975; reviewed in Gagneux &
Varki, 2001) and because they sometimes display violent behavior similar to that seen in
humans (Manson et al., 1991; Newton-Fisher & Emery Thompson, 2012; Wrangham &
Wilson, 2004). Through the media, the general public has become captivated by our
“aggressive” next of kin, as chimpanzees are dramatically portrayed as violent,
unpredictable animals who sometimes fight each other to the death (intracommunity:
Watts, 2004; intercommunity: Watts et al., 2006; reviewed in Wrangham et al., 2006).
They may even sporadically attack humans (Laufer, 2011), but media reports neglect to
discuss the implications (e.g., adult chimpanzees kept inhumanely as pets, Ross et al.,
2008; high anthropogenic disturbance, Hockings et al., 2010) and the low rates of such
lethal behavior (Wrangham, 1999). Human aggression (e.g., murders, war, etc.) is
typically the leading story on any news program on any given day. However, lethal
aggression in wild chimpanzees, while very real, is often sensationalized and occurs at
rates similar to warfare among subsistence-society hunters and farmers (reviewed in
Wrangham et al., 2006). While some researchers have “blamed” higher rates of
chimpanzee lethal aggression on increased human impact in their natural habitat, a recent
study has confirmed it is an adaptive strategy employed to eliminate rivals and increase
their fitness through greater resource acquisition (Wilson et al, 2014). Lethal aggression,
while alluring, is the most extreme form of aggression, and like human warfare, it is not
the day-to-day norm in [chimpanzee] society.

72
Non-lethal aggression is, however, much more commonplace in chimpanzee
society. In fact, wild chimpanzees, on average, engage in non-lethal contact aggression
283 times higher than Australian aboriginal hunter-gathers (human data from Burbank,
1992). Including non-contact aggression would make this contrast even starker
(Wrangham et al., 2006). In the wild, a single day, let alone a single hour, rarely goes by
without some kind of fight breaking out among community members. Sometimes
squabbles are brief and may only involve a threatening wave of the arm, but other times
they are competitive, intense knock-down-drag-out fights over prized resources.
With intracommunity aggression rates this high, how do chimpanzees manage
conflicts and maintain social cohesion? While ample research has focused on why
chimpanzees within the same community initially fight (Muller, 2002; Muller &
Wrangham, 2004), less has focused on the variation that exists within the conflict and
what parameters predict this variation (Wittig & Boesch, 2003b). More data are therefore
needed to make within-species comparisons.
One aim of this study was to examine the Kanyawara chimpanzees’ conflict
patterns and variation to determine how the chimpanzees make decisions during conflicts
and what factors predict these decisions. The basis for these aims stems from the
Relational Model, which addresses the cost-benefit ratio of intragroup competition
through a decision-making process (de Waal, 1996). For more details on this model and
the theoretical basis for the following hypotheses (H) and predictions (P), please see
Chapter 1.

73
H
1
: Rates of aggression between dyads will be largely influenced by general time spent
in close association even after controlling for association time.
P
1
: Since males are generally more gregarious than females (even after
accounting for time spent in close association), I expected male dyads to fight
more often and more intensely than female dyads.
P
2
: Mixed sex dyads will affiliate, and therefore fight more, when females are in
estrous.
P
3
: In keeping with H
1
but opposite the prediction set forth by Wittig & Boesch
(2003b), I expected strongly associated (e.g., time in close proximity, time spent
in contact, time spent giving support) dyads to engage in more conflicts than
neutral associates regardless of sex. However, aggression should be less intense
between highly associated dyads (see H
7
).
H
2
: Aggression rates will increase when the potential benefit (i.e., resource acquisition)
to the aggressor is high.
P
1
: Male dyads will fight more over sex, dominance, meat, and social situations.
P
2
: Female dyads will fight more over food and when protecting their offspring.
H
3
: Estrous state and parity will influence aggression rates and variation between male-
female dyads.
P
1
: From males, parous females will receive higher rates of and more severe
aggression than nulliparous females.
P
2
: From males, fully swollen females will receive higher rates of and more
severe aggression than deflated and partially swollen females.

74
H
4
: Dominance rank will influence aggression rates and variation, especially in males.
P
1
: Opponents closer in rank will fight more often, but less intensely than more
distally ranked opponents.
P
2
: As rank difference decreases between opponents, victims will be more likely
to retaliate.
Cooperation and Valuable Relationships
As previously discussed, one costly byproduct of social living is that more time
spent together provides more opportunities to fight with conspecifics over resources.
However, there is an important counter to this. Social living also affords many benefits
(predator protection: Dunbar, 1998; van Schaik, 1983; food territory protection:
Wrangham, 1980), and in tolerant species like chimpanzees (Melis et al., 2006), it
provides opportunities to develop cooperative (Clements & Stephens, 1995; Mesterton-
Gibbons & Dugatkin, 1992; Muller & Mitani, 2005), collaborative (Boesch & Boesch,
1989; Melis et al., 2010), and valuable [long-lasting] social bonds (Gilby & Wrangham,
2008; male chimpanzees: Mitani, 2009; female chimpanzees: Langergraber et al., 2009)
between group members (reviewed in Boesch, 1996).
Just as chimpanzees are excellent models in which to study conflict variation,
they also are a suitable species in which to study association patterns. Chimpanzees’
fission-fusion society is structured so that party membership is constantly changing in
size and composition (Goodall, 1986; Nishida, 1968). This social fluidity affords
chimpanzees the liberty to be selective in choosing their preferred social partners (PSPs).
However, individual choices may not always align. For example, chimp A may prefer to

75
associate with chimp B when in the same party, whereas chimp B may prefer to associate
with chimp C. This demonstrates a one-sided relationship, which is not as strong as a
mutual relationship (e.g., chimp A prefers chimp B and vice versa). Individuals choosing
to avoid/not associate with one another indicates a neutral relationship (Gilby &
Wrangham, 2008). Building and maintaining preferred relationships requires a great deal
of time and energy, thus chimpanzees must be selective of their friends.
In chimpanzees, PSPs (also referred to as preferred associates and preferred
relationships) are most commonly measured by proximal association rates (Cairns &
Schwager, 1987; Gilby & Wrangham, 2008). However, for male chimpanzees who are
more gregarious (Pepper et al., 1999) and cooperative (reviewed in Muller & Mitani,
2005) than females (Wrangham & Smuts, 1980), time spent grooming (Watts, 2002) and
providing agonistic support (de Waal & Harcourt, 1992; Watts, 1998) are also
appropriate, and potentially more precise PSP measures. It should however be noted that
sex differences in association are more prominent in East African chimpanzees (Gilby &
Wrangham, 2008) compared to their West African relatives (Lehmann & Boesch, 2008).
The second aim of this study was to assess the Kanyawara chimpanzees’ PSPs by
determining which parameters influenced PSPs and how these varied by sex.
H
5
: Dyad sex combination will influence levels of PSP association.
P
1
: Male dyads will more strongly associate with one another on all PSP
parameters compared to mixed sex dyads and female dyads.
P
2
: Mixed sex dyads containing parous females will associate at higher PSP levels
compared to dyads containing nulliparous females.

76
P
3
: Female dyads will associate at much lower PSP levels compared to mixed sex
dyads.
Aggression Between PSPs
PSP status and strength between conspecifics must not only influence how
community members interact on a daily basis, but also when in direct competition for
resources. Since it is expected that PSPs spend more time together than neutral
associates, they must also be in direct competition for resources more frequently (van
Hooff and van Schaik, 1992; Walters & Seyfarth, 1987). Aggression therefore becomes
inevitable even between the best of friends. The predicament is that an out-competed PSP
(e.g., meat stolen by PSP and not shared) might deny support in a future situation where
cooperation is needed (e.g., withholding coalitionary support during mate guarding). To
minimize the relationship damage that occurs during aggression, PSP aggressors should
be strategic and use the minimum force (e.g., severity) necessary to obtain the resource.
Consequently, increased association between the same dyads has distinct costs and
benefits. Aggression between PSPs would therefore be the ultimate test of a dyad’s
relationship strength.
The final aim of this study was to determine how PSP status and strength
predicted the occurrence of and variation within directed aggression (also referred to as
conflicts).
H
6
: Mutual and one-sided PSPs would fight more often than neutral associates.
P
1
: As overall mutual PSP strength increases, fighting will also increase.
P
2
: Fighting between PSPs will be most prominent between male-male dyads.

77
P
3
: Overall mutual PSP opponents will remain in closer proximity immediately
following aggression (possibly to reconcile), whereas neutral partners would
disperse.
H
7
: Conflict intensity will be regulated by opponent PSP status and strength.
P
1
: Opponents sharing mutual or one-sided PSP status will engage in shorter
conflicts of lower severity than neutral associates.
P
2
: As the strength of mutual PSP increases, conflict duration and severity will
continue to decrease.
Methods
Study site and subjects. This study focused on the behaviors of the Kanyawara
chimpanzee community (0°34’ N, 30°21’E) located in Kibale National Park in
southwestern Uganda (0°13’-0°41’N, 30°19’-30°32’ E, Figure 2.1). See Chapter 2 for
study site details. The Kanyawara chimpanzees have been continuously studied since
1987 and are therefore habituated to human presence (Wrangham et al., 1992). Data were
collected from November 2010 through November 2011. At the time of study, the
community consisted of 52 individuals, including 12 adult males, 16 adult females, 6
subadults, 7 juveniles, and 11 infants. During this time frame, the community
composition remained fairly constant. One infant died, two infants were born, and no
females immigrated into or emigrated out of the community (Appendix A).
Focal follows. I collected all-day continuous focal follows (Altmann, 1974) on 11
adult male and six adult female chimpanzees (Table 2.1). I generally followed my focal
from morning nest to night nest; however, I elected to abandon my focal if they were

78
alone for more than 1 hr, and I knew the location of another party since I was particularly
interested in intragroup aggression. I would subsequently select a new focal in the multi-
individual party. I also employed instantaneous scan sampling (Altmann, 1974) on the
focal party every 10 min where I collected party composition, party size, party members’
proximity to focal, and general party activity (ruled by majority). Party size was defined
as the number of individuals in view of the observer or field assistants in the same place
at the same time (Newton-Fisher, Reynolds, & Plumptre, 2000).
Preferred social partners. Using the focal data, I calculated both one-sided and
mutual preferred social partners (PSPs) using the methods outlined by Gilby and
Wrangham (2008) on the following relationship value parameters: proximity association
(includes within party, within 5 m, and nearest neighbor proximity measurements),
contact affiliation (includes grooming, playing, mounting, embracing, contact sitting,
etc.), and agonistic support (includes coalitions and joint defense). I calculated an overall
PSP index by counting the number of PSP parameters each dyad mutually shared. Overall
mutual PSP scores ranged from 0 to 3. Neutral dyads were those that shared no mutual
PSP parameters (a score of 0). Weak PSPs were dyads that shared only one mutual PSP
parameter (a score of 1). Moderate PSPs were dyads that shared two PSP parameters (a
score of 2). Strong PSPs were dyads that shared all three PSP parameters (a score of 3).
See Chapter 2 for a detailed description and calculation of these parameters.
Dominance hierarchy. Traditionally, dominance rank in chimpanzees is
determined using pant-grunt vocalizations (Bygott, 1979) because pant-grunts are highly
directional and correlate with several measures of dominance. However, in a many

79
studies (Furuichi, 1997; Muller, 2002; Wroblewski et al., 2009), pant-grunt observations
alone are insufficient to determine linear rank and agonistic outcomes must often be
included. This study incorporated both pant-grunt and agonistic outcomes to determine
linear rank. To test this relationship among adults, I selected a probabilistic model
(Jameson et al., 1999) that 1) accounted for the relative success of individuals in agonistic
interactions, 2) predicted dominance relationships for dyads that never interacted, 3)
predicted the linearity of relationships, and 4) assigned a rank order. For more calculation
details, see Chapter 2.
Aggression data. I collected directed aggression data ad libitum (Altmann, 1974).
I only collected directed aggression data between adults. Directed aggression (also
referred to as a conflict) was defined as aggressive behavior targeted toward at least one
individual. If a conflict occurred during focal observations, I temporarily abandoned the
focal observation to collect conflict (and subsequent post-conflict) data. I assigned each
conflict a unique identification number. I recorded the following information about the
conflict: date, time, duration, location, aggressor and victim identities, type and
description of aggression, intensity of aggression, directionality of aggression, screams
during aggression, winner’s identity, context of aggression, bystander identities, and
distance between opponents. See Appendix B for behavioral ethogram and Appendix C
for operational definitions.
Statistical procedures. To calculate the linear dominance hierarchy, I tested the
combined (pant grunt recipient/performer + winner/loser) matrix using the I&SI method
(de Vries & Appleby, 2000; Jameson et al., 1999) in MatMan 1.1 (de Vries et al., 1993).

80
To test differences in dyadic preferred social partner means, I used an unpaired t-test. I
used Generalized Linear Mixed Models and Linear Mixed Models to determine what
factors best predicted variation in aggression between opponents. See Chapter 2 for a
detailed description of statistical analyses and the programs used.
Results
Preferred Social Partners (PSPs)
Data were sufficient to calculate preferred social partners (PSPs) for 279 unique
adult dyad combinations (male-male = 55 dyads, male-female = 154 dyads, female-
female = 70 dyads). When accounting for proximity association between dyads, all 279
dyads were within the same party for an average of 1601 (± 780 SD) party composition
10-min scans. Within same party scans, only 41 dyads were never within 5 m of each
other. Within 5 m scans, only 35 dyads were never each other’s nearest neighbor. To
some degree, most dyads were therefore proximal associates, even if they were not
proximity PSPs. Table 3.1 shows, on average, male-male dyads associated more than
expected (>1) and associated significantly more than both male-female and female-
female dyads. Male-female and female-female dyads however did not associate more
than expected (>1) and differences were not significant.
Of the 279 dyads, most dyads (79%, N = 221) did not provide each other with
agonistic support. The 58 dyads that did provide some form of agonistic support were
predominately same sex dyads. Table 3.1 shows male-male dyads provided the largest
amount of agonistic support, but female-female dyads also provide more agonistic

81
support than expected (>1). Male-female dyads provided very little agonistic support to
one another. All three dyads showed statistically significant differences.
Of the 279 dyads, 41% (N = 115) did not engage in affiliative contact during the
study period. The 59% of dyads that engaged in affiliative contact occurred
predominately between male-male dyads at a higher than expected rate (>1). Table 3.1
shows male-male dyads engaged in affiliative contact at a rate 4.7 times greater than
male-female dyads and 24 times greater than female-female dyads. All three dyads
showed statistically significant differences.
Table 3.1. Sex differences in proximity association, agonistic support, and contact
affiliation strength.

Dyad Mean (± SD) Comparison dyad t df p
Proximity Association
MM 1.44 (± 0.43) MF 7.60 207 <0.0001
MF 0.92 (± 0.43) FF 1.06 222 0.29
FF 0.86 (± 0.43) MM 7.49 123 <0.0001
Agonistic Support
MM 3.35 (± 3.67) MF 10.81 207 <0.0001
MF 0.10 (± 0.50) FF 4.31 222 <0.0001
FF 1.15 (± 2.94) MM 3.73 123 0.0003
Contact Affiliation
MM 3.13 (± 3.50) MF 7.26 207 0.0001
MF 0.67 (± 1.39) FF 3.09 222 0.0022
FF 0.13 (± 0.64) MM 7.01 123 0.0001
Note. MM indicates male-male dyads. MF indicates male-female dyads. FF indicates
female-female dyads. Mean indices were standardized by dividing by the mean of all
dyads (see Chapter 2 for calculation details). Mean values greater than one (in bold)
indicate dyads associated/interacted more than expected relative to the 279 dyad mean.

82
As depicted in Table 3.2, male-male dyads demonstrated a much higher tendency
toward mutual gregariousness for all three PSP parameters relative to dyads containing
females. On average, male-male dyads were at least twice as gregarious as other dyads.
For males, proximity played a key role in determining mutual PSPs as most males (78%)
were consistently in close proximity to one another. Males also shared strong PSPs for
agonistic support (36%) and contact affiliation (38%) relative to other dyads. Agonstic
support was not an important PSP measure for male-female dyads (1%), nor was contact
affiliation for female-female dyads (3%).
Table 3.2. Dyads classified as mutual PSPs for proximity association, agonistic
support, and contact affiliation.

Mutual PSP
Parameters
Male-Male Male-Female Female-Female
N % N % N %
Proximity Association 43 78.2 12 7.8 9 12.9
Agonistic Support 20 36.4 2 1.3 10 14.3
Contact Affiliation 21 38.2 11 7.1 2 2.9
Note. PSP indicates preferred social partner. N indicates number of mutual PSPs, and %
indicates percent of dyads classified as mutual PSPs.

Dominance Relationships
Linear hierarchy. In general, the male chimpanzees did not often pant-grunt to
one another during focal observations (N = 34). The alpha male received 82% of the
pant-grunts from the other males, and the second and third ranking male each received
9% of pant grunts. Males ranking fourth and below did not pant grunt to one another
during focal observations. The four lowest ranking males produced 65% of the male pant
grunts.

83
Males engaged in 172 dyadic conflicts with clear winners. The alpha male won
42% of those conflicts and only lost 3%. The second and third ranking males won 24%
and 9% of the conflicts, respectively. The four lowest ranking males lost 49% of the
conflicts.
From the focal pant-grunt and conflict data, I was able to calculate linear
dominance relationships among the 11 adult males. To do this, I combined pant grunt
exchange data (de Waal, 1978) with dyadic conflict winner data for males into a single
matrix. Coalitionary attacks were omitted. Based on the combined dominance matrix, the
corrected (for unknown relationships) Landau’s index of linearity was h’ = 0.75 (p <
0.0003) for males. This indicates that the male dominance hierarchy at Kanyawara was
significantly linear and stable during the study period. The reordered matrix indicated
that PB was involved in two relationship inconsistencies (as outlined below) with ES and
his brother, PG, but a linear rank order was unobstructed based on PB, ES, and PG
relationships with other males (Table 3.3).
No reversals were observed in male-male pant-grunt directionality, and only nine
winner-loser reversals were observed in the 172 decided dyadic conflicts. Of the nine, six
of the reversals were not labeled as inconsistencies in the linearity analysis because the
more dominant male won more conflicts than he lost to the subordinate male. For
example, while AJ won five conflicts against KK, AJ lost eight to KK. ST won one
conflict against ES, but lost three. The two inconsistencies in the linearity analysis
resulted from a subordinate male winning all conflicts against a dominant male. For

84
example, PB won the only decided conflict against his brother PG and won the only two
decided conflicts against ES.
Table 3.3. Male linear dominance rank.
Linear Rank Name ID Code Assigned Rank Category
1 Kakama KK Alpha
2 Makoku LK
High 3 Johnny AJ
4 Big Brown BB
5 Lanjo TJ
Mid 6 Eslom ES
7 Twig PG
8 Yogi YB
Low
9 Stout ST
10 Bud PB
11 Tofu TU
Note. Rank based on pant grunt directionality and conflict outcomes.
While females often pant-grunted to males, females rarely pant-grunted to each
other (N = 15). Females also rarely engaged in agonistic interactions with clear winners
(N = 47) with one another. I was, therefore, unable to rank the 14 adult females or
calculate female linear dominance due to insufficient data during focal observations.
Directed Aggression
Sex differences in aggression. Adult male chimpanzees at Kanyawara were more
likely to initiate directed aggression than adult females. Adult males acted as the
aggressor 86.4% of the time with adult females being the recipients of aggression
(henceforth referred to a victims) 54.2% of the time. The majority of conflicts were
between adult males and adult females (55.4%) followed by conflicts between only adult

85
males (33.4%) then by conflicts between only adult females (11.2%). Table 3.4 shows
that males most commonly directed aggression towards females, and females most
commonly directed aggression toward other females. However this does account for time
spent in the same party between dyads.
Table 3.4. Victims of directed aggression by sex in adult chimpanzees.
Victim Sex
Aggressor Sex Male Female Total
Male 211 (39%) 335 (61%) 546
Female 15 (17%) 71 (83%) 86
Total 226 406 632

Using the PSP calculation model, the aggression level was calculated for each
dyad accounting for time spent in the same party. PSP victims were determined for each
individual based on the number of observed conflicts and directionality of aggression
between dyads (see Chapter 2 for calculation details). After controlling for time spent
together, in general, males (M ± SD = 0.79 ± 1.92) initiated three times more aggression
than females (M ± SD = 0.26 ± 1.02, t(574) = 4.22, p < 0.0001), lending the same pattern
as presented in Table 3.6. In particular, males (M ± SD = 0.79 ± 1.92) initiated
significantly more aggression than parous (M ± SD = 0.33 ± 1.17, t(490) = 3.09, p =
0.002) and nulliparous females (M ± SD = 0.05 ± 0.20, t(346) = 3.48, p = 0.0006). Parous
females (M ± SD = 0.33 ± 1.17) initiated significantly more aggression than nulliparous
females (M ± SD = 0.05 ± 0.20, t(310) = 2.17, p = 0.03).

86
Table 3.5 shows the rate of aggression between aggressor and victim dyads in
more detail. Male dyads and female dyads engaged in the same amount of aggression
(t(266) = 1.13, p = 0.26, ns). However, parous females predominantly influenced female
aggression. Male dyads engaged in significantly more aggression than nulliparous
females (t(120) = 2.10, p = 0.038). Males aggressed against parous females more than
expected and significantly more than male dyads (t(218) = 3.27, p = 0.001), and parous
females aggressed against nulliparous females more than expected and significantly more
than parous female dyads (t(116) = 3.65, p = 0.0004).
Table 3.5. Rates of aggression, indicating number of PSP dyads, by dyad sex
combination.

Victim
Aggressor
Male Parous Female Nulliparous Female
Mean (± SD) PSP Mean (± SD) PSP Mean (± SD) PSP
Male 0.35 (± 0.58) 7 1.21 (± 2.70) 31 0.81 (±1.54) 9
Parous Female 0.06 (± 0.28) 2 0.32 (± 0.97) 9 1.51 (±2.57) 12
Nulliparous
Female
0.03 (± 0.13) 0 0.11 (±0.30) 2 0 ± 0 0
Note. PSP indicates the number of dyads classified by sex combination category as
aggression preferred social partners. Mean indices were standardized by dividing by the
mean of all dyads (see Chapter 2 for calculation details). Mean values greater than one (in
bold) indicate dyads aggressed toward each other more than expected relative to the
mean.

Out of the 110 possible male-parous female dyads, parous females were the males
preferred targets of aggression in 31 dyads. Of the 110 male-male dyads (accounting for
directionality of aggression), male victims were the preferred targets in only 7 dyads. In
general, females rarely initiated aggression with males and only two males were the

87
preferred targets of two females. Nulliparous rarely initiated aggression, but when they
did their preferred victim was a single parous female (LR, Table 3.5). Only three dyads
(one male-parous female dyad and two parous female-nulliparous female dyad) were
mutual preferred victims, meaning A was the preferred victim of B, and B was the
preferred victim of A. With the exception of five dyads, preferred victims were lower
ranking than their aggressors.
When accounting for estrous state in conflicts involving adult females (N = 421),
females without sexual swellings acted as aggressors 87.2% of the time, and were the
victims 54.9% of the time. Females with full swellings were the aggressors only 5.8% of
the time, but were the victims 31.7% of the time. Females with partial swellings were
aggressors 7.0% of the time, and were victims 13.5% of the time. When accounting for
parity in conflicts involving females, parous females were both the more likely
aggressors (94.2%) and victims (79.3%) as compared to nulliparous females (Table 3.5
and 3.6).
Table 3.6. Frequency of conflicts involving adult females when accounting for
parity, estrous state, and role in aggression.

Conflicts with Males
Conflicts with Other
Females
Sex Swelling Size Deflated Partial Full Deflated Partial Full
Parous
Females
Aggressor 11 1 1 63 2 3
Victim 177 39 73 32 1 0
Nulliparous
Females
Aggressor 0 1 1 1 2 0
Victim 7 12 28 7 3 27

88
Context also played a role in directed aggression, which differed by sex. Males
engaged in more directed aggression in the contexts of social excitement (43%), sexual
competition (23%), and reunions (19%), whereas females engaged in more directed
aggression in plant-food competition (47%), social excitement (31%), and protection
(15%). With the exception of the social excitement context, which also included high
arousal situations following prior aggression, the most common contexts in which males
and females aggressed did not overlap (Table 3.7).
Table 3.7. Context of directed aggression by sex at Kanyawara in 2011 compared to
data in 1998 from Muller (2002).

2011 (present study) 1998 (Muller, 2002)
Context Males Females Males Females
Reunion 19 4 43 0
Sexual Competition 23 1 36 8
Plant-food Competition 11 47 15 80
Meat Competition 3 2 4 0
Social excitement 43 31 8 0
Protection 2 15 1 12
Miscellaneous 0 0 2 0
Total Observations (N) 546 86 212 25
Note. Data represented as percentages.

Preferred social partner differences in aggression. The preferred social
partnerships of dyads influenced directed aggression, but patterns appeared different for

89
male-male dyads compared to dyads containing at least one female. Table 3.8 shows
male-male dyads who were mutual [proximity, agonistic support, and contact affiliation]
PSPs engaged in more conflicts than one-sided PSPs or neutral associates. Alternatively,
male-female and female-female dyads engaged in more conflicts when they were neutral
associates compared to one-sided and mutual PSPs. However, these raw frequencies of
aggression do not account for amount of time opponents spend in the same party.

Table 3.8. Frequency of conflicts between PSPs for measures of proximity
association, agonistic support, and contact affiliation by opponent sex combination.

PSP
Parameter
PSP
Direction
Male-Male Male-Female Female-Female Total
Proximity
Association
Mutual 122 33 4 159
One-sided 17 111 19 147
Neutral 72 206 48 326
Agonistic
Support
Mutual 107 14 4 125
One-sided 37 19 8 64
Neutral 67 317 59 443
Contact
Affiliation
Mutual 100 31 0 131
One-sided 30 68 3 101
Neutral 81 251 68 400
Note. PSP indicates preferred social partner. Mutual indicates both opponents were a PSP
of the other. One-sided indicates only one opponent was the PSP of the other. Neutral
indicates neither opponent was a PSP of the other.

When accounting for time spent together, the pattern of aggression between male
and female PSP opponents opposes the pattern presented in Table 3.4. Female opponents
also showed a different pattern, whereas for male opponents the pattern remains
unchanged. Figure 3.1 further illustrates this relationship between aggression and PSP

90
parameters after accounting for time spent together. In general, dyads involving males
resulted in higher rates of aggression when opponents were mutual PSPs compared to
one-sided PSPs and neutral partners. Male-female opponents sharing mutual contact
affiliation for one another (M ± SD = 3.45 ± 7.79) had significantly higher rates of
aggression compared to neutral partners (M ± SD = 0.91 ± 1.25; t(131) = 3.27, p =
0.001). However due to such large PSP variation within dyads (SD average = 1.42, SD
range = 0.41 – 7.79), the remaining PSP parameters across PSP status were trends and
not statistically significant from one relationship type to the next.

Figure 3.1. Mean rate of aggression relative to opponent PSP type/strength and sex.
MM indicates male-male opponents. MF indicates male-female opponents. FF indicates
female-female opponents. Mean indices were standardized by dividing by the mean of all
dyads (see Chapter 2 for calculation details). Mean values greater than one indicate dyads
aggressed toward each other more than expected relative to the mean.
Proximity
association

Contact
affiliation

Agonistic
support

PSP
Parameters

91
Female-female opponents presented a unique pattern in that one-sided PSPs had
higher rates of aggression than mutual PSPs or neutral partners, but again these patterns
were mostly trends due to high variation within groups (SD average = 1.41, SD range = 0
– 3.37). Female opponents sharing mutual agonistic support for one another (M ± SD =
0.28 ± 0.62) had significantly lower rates of aggression compared to one-sided PSPs (M
± SD = 1.55 ± 1.39; t(10) = 2.23, p = 0.05).
Table 3.9 further illustrates this difference between male-male dyads and dyads
involving females by quantifying mutual PSP strength. In general, male-male and male-
female dyads exhibited a similar pattern of aggression relative to mutual PSP strength.
The stronger a dyad’s PSP, the more likely they were to fight. Male-male dyads were the
only dyads with strong PSPs (score of 3) and engaged in conflicts at a rate of 23.6. [For
example, of the eight unique male-male dyads that shared all three PSP parameters, seven
of the dyads fought a total of 165 times. Only one male-male dyad (e.g., AJ-BB) with a
PSP strength of 3 was not observed in conflict during the study period.] Male-female
dyads with moderate PSPs (score of 2) were more than twice as likely to fight as neutral
and weak PSPs. Female-female dyads, on the other hand, were twice as likely to fight
with neural associates compared to weak and moderate PSPs.
Table 3.9 does not however account for the rate of aggression relative to time
spent together. After controlling for time together (Figure 3.2), the patterns of aggression
for male and female PSP dyads mirrors the data presented in Table 3.5. In general, as a
male dyads’ PSP strength declined, so did their risk of aggression. Specifically, male
dyads with a PSP strength of 3 (M ± SD = 1.08 ± 0.75) fought significantly more than

92
Table 3.9. Frequency of conflicts between mutual PSPs based on overall strength of
the opponents’ relationship, categorized by opponent sex combination.

Mutual
PSP
Strength
Male-Male Male-Female
Female-
Female
Total
Total
Conflicts
N = 632
0 65 (4.3) 290 (3.1) 65 (2.2) 420 (3.1)
1 28 (2.8) 42 (4.2) 4 (1.3) 74 (3.2)
2 53 (5.3) 18 (9) 2 (1.0) 80 (6.2)
3 165 (23.6) --- --- 165 (23.6)
Total
Unique
Conflict
Dyads
N = 181
0 15 94 29 138
1 10 10 3 23
2 10 2 1 13
3 7 --- --- 7
Total
Dyads
N = 279
0 20 137 55 212
1 14 14 13 41
2 13 3 2 18
3 8 0 0 8
Note. PSP indicates preferred social partner. PSP strength indicates the total number of
PSP parameters (see Table 3.1) shared mutually by opponents. Numbers in parentheses
represent the rate of aggression relative to mutual PSP strength when accounting for the
number of unique conflict dyads. Total unique conflict dyads indicates the number of
distinct dyads that engaged in at least one conflict. Total dyads refers to the number of
possible unique dyads within each PSP strength category.

male dyads with a PSP strength of only 1 (M ± SD = 0.47 ± 0.40; t(20) = 2.49, p = 0.02).
Due to high aggression variation between male dyads (SD average = 0.61, SD range =
0.40 – 0.75), there was no significant difference between the other PSP strengths for
males. Male dyads were the only dyads to achieve the highest mutual PSP strength of 3.
The pattern was opposite for female dyads. As a female dyads’ PSP declined, their risk of
aggression increased. Male-female dyads showed a nonlinear pattern. Male-female dyads
with a mutual PSP strength of 1 (M ± SD = 2.25 ± 4.15) had a significantly higher risk of

93
aggression compared to dyads sharing no PSP relationship (M ± SD = 0.99 ± 1.35; t(149)
= 2.52, p = 0.13).

Figure 3.2. Mean rate of aggression accounting for overall mutual PSP strength and
opponent dyadic sex combination. MM indicates male-male opponents. MF indicates
male-female opponents. FF indicates female-female opponents. PSP strength indicates
the number of PSP parameters shared mutually between opponents. Mean indices were
standardized by dividing by the mean of all dyads (see Chapter 2 for calculation details).
Mean values greater than one indicate dyads aggressed toward each other more than
expected relative to the mean.

Rank differences in aggression. Since linear rank could only be determined for
adult males, rank difference was calculated solely for male-male directed aggression.
Males with a rank difference of 1-3 were categorized as having a small difference in rank,
4-6 were medium, and 7-10 were large. Rank difference between opponents did not play

94
a major role in aggression. Males with small differences in rank engaged in directed
aggression (38%) slightly more than males with medium rank difference (35%), followed
by males with large rank difference (27%).
Components of directed aggression. The act of and response to directed
aggression between the Kanyawara chimpanzees varied considerably. However, most
aggression between adults was fairly mild, containing mostly threats (62%) and resulting
in no visible wounds (95%). When aggression escalated to include physical contact, most
contact was low in severity involving only one instance of contact (64%), followed by
mild severity involving multiple instances of contact (24%), and then instances of high
severity contact, which possibly included biting (12%). Unidirectional aggression (80%)
was more common than bidirectional (20%), and aggressors were more likely the winners
(85%). Victims (79%) more commonly screamed during conflicts than aggressors (6%),
with victims engaging in prolonged screams (continued screaming after the conflict
ended) only 16% of the time. Most opponents (48%) ended the conflict with 5-10 m
between one another, 28% of opponents were more than 10 m apart, and 23% of
opponents were within 5 m of one another. Most conflicts occurred on the ground (72%)
as compared to in the canopy (28%). Most aggression included only one conflict between
the aggressor and the victim and did not involve bystanders (58%). However, high
arousal from the initial conflicts sometimes spurred third-party conflicts (19%). Renewed
aggression between aggressors and victims occurred 12% of the time, and redirected
aggression oriented toward a third-party bystander occurred 14% of the time. While party
size varied widely, 61% of conflicts occurred in parties containing 11-20 adults, 23% of

95
conflicts occurred in parties of 6-10 adults, 8% of conflicts occurred in parties of 21 or
more adults, and 7% of conflicts occurred in parties of 2-5 adults.
Predictors of Conflict Variation
Given that conflicts varied considerably in form, I tested how the different
elements affected the opponents’ decisions throughout the conflict.
Aggression type. Aggression type was a binomial variable indicating if the
conflict was only threat based (no physical contact) or agonistic (included physical
contact). The best fit model (AIC = 803.5) for aggression type indicated that the best
linear predictor was the number of third-party bystanders (β ± SE = -0.065 ± 0.031, z =
-2.079, p = 0.038). As the number of bystanders increased, the probability of agonistic
aggression between conflict partners also increased. Overall mutual preferred social
partner was also a strong predictor of aggression type, but this relationship was quadratic,
not linear (β ± SE = 0.654 ± 0.252, z = 2.596, p = 0.009), resulting in no significant
difference between the different strengths of preferred social partnership.
Directionality. Directionality was a binomial variable indicating whether the
conflict was unidirectional or bidirectional. The best fit model (AIC = 491) indicated
three predictors of directionality, with the best predictor being severity, followed by the
victim’s estrous state, and aggressor’s linear rank. Severity showed a significant linear
relationship to directionality (β ± SE = -1.973 ± 0.327, z = -6.037, p < 0.001). Conflicts
with the lowest severity (only threats, scored as zero) were significantly more likely to be
unidirectional compared to conflicts with increasing agonistic severity (agonistic severity
scored from 1 to 3; 0 vs. 1: β ± SE = -2.409 ± 0.293, z = -8.232, p < 0.001, 0 vs. 2: β ± SE

96
= -2.496 ± 0.377, z = -6.588, p < 0.001, 0 vs. 3: β ± SE = -2.916 ± 0.471, z = -6.191, p <
0.001). However, there was no significant difference between agonistic severity scores 1
through 3. There was also a significant linear relationship between victim’s estrous state
and directionality with the probability of bidirectional aggression increasing for males
and non-swollen females (β ± SE = 1.052 ± 0.305, z = 3.449, p < 0.001). In post hoc
analyses, only male victims engaged in significantly more bidirectional conflicts than
fully swollen females (β ± SE = 1.594 ± 0.434, z = 3.667, p = 0.001). There was no
significant difference in directionality between male victims and deflated or partially
swollen female victims or between the different estrous states of female victims.
Aggressor’s rank showed a significant linear relationship to directionality with the
probability of bidirectional aggression increasing as the aggressor’s rank declined (β ± SE
= -1.912 ± 0.703, z = -2.721, p = 0.007). Female aggressors, who always rank below
every adult male, were much more likely to engage in bidirectional aggression with their
victim as compared to high-ranking males (p < 0.01).
Wounding. Wounding was a binomial variable indicating whether either
opponent had a visible open [bleeding] wound by the end of the conflict. The best fit
model (AIC = 246.1) indicated that the strongest predictor of wounding was the victim’s
estrous state. While there was a significant linear relationship showing that the
probability of wounding decreases from male to female victims and from fully deflated
females to fully swollen female victims (β ± SE = -1.532 ± 0.716, z = -2.139, p = 0.032),
there was no significant difference between each condition. The pattern was opposite for
the aggressor’s estrous state. The probability of wounding increased from male

97
aggressors to females, and from fully deflated females to fully swollen females, but this
pattern was only a trend approaching significance (β ± SE = 1.554 ± 0.858, z = 1.811, p =
0.070).
While wounds may be difficult to see under a chimpanzees’ blanket of black hair,
bleeding wounds were relatively easy to detect. Biting typically resulted in open wounds,
whereas other physical acts of aggression did not. Wounded individuals typically
advertised their wounds by self-grooming the injured area.
Severity. Severity was an ordinal variable indicating the extent of aggression
between opponents (0 = threat, 1 = one instance of physical contact, 2 = multiple
instances of physical contact, 3 = multiple and prolonged instances of physical contact
may include biting). The best fit model (AIC = 1550) indicated the best predictor to be
conflict type, followed by aggressor parity, then victim parity, and finally agonistic
support preferred social partnership. Conflict type delineated whether the conflict was a
single occurrence between opponents or contained renewed or redirected aggression.
Conflicts involving renewed aggression, meaning reoccurring aggression between the
original aggressor and victim within 10 min of the last bout, were significantly higher in
severity than conflicts without renewed aggression (β ± SE = 0.394 ± 0.116, t = 3.404, p
< 0.001). Parity of the aggressor and victim also predicted the severity of aggression.
Male aggressors were involved in less severe conflicts than parous female aggressors (β ±
SE = 0.371 ± 0.110, t = 3.369, p = 0.002). Nulliparous female victims, on the other hand,
were involved in less severe conflicts as compared to both male (β ± SE = -0.515 ± 0.156,
t = -3.308, p = 0.003) and parous female (β ± SE = 0.460 ± 0.144, t = 3.196, p = 0.004)

98
victims. Victims who were preferred social partners of aggressors for agonistic support
received less severe aggression than non-preferred partners, but this pattern was only
approaching significance (β ± SE = -0.188 ± 0.098, t = -1.914, p = 0.056).
Uncertain conflicts between non-friends resulted in higher severity than clear
definitive conflicts. Renewed aggression often occurred when there was no clear winner.
This possibly resulted in the escalation of severity until a clear winner emerged. Parous
females however engaged in more severe conflicts relative to other aggressors when they
were attacking other parous females with whom they did not share strong bonds. Severity
likely escalated in these contexts, as they were most often fighting to either protects their
offspring or over food resources.
Aggressor chase. Aggressor chase was a binomial variable indicating whether or
not the aggressor chased the victim during the conflict. By definition, chases required that
victims fled. The best fit model (AIC = 774.2) indicated three significant predictors for
aggressors chasing victims. The victim’s rank was the best predictor showing that as the
victim’s rank decreases, the probability of the aggressor chasing the victim (and the
victim fleeing) significantly increases (β ± SE = -2.258 ± 0.680, z = -3.319, p < 0.001).
Total party size was the second best predictor indicating that as party size increases the
likelihood of the aggressor chasing the victim significantly decreases (β ± SE = -0.057 ±
0.018, z = -3.139, p = 0.002). Overall mutual PSP was the final predictor showing that as
strength of preferred social partnership between opponents increased, the likelihood of
aggressor chase significantly decreased (β ± SE = -0.952 ± 0.360, z = -2.642, p = 0.008).
However, there was no significant difference between each of the PSP strengths.

99
Since victims must flee in order to be chased by aggressors, the above predictors
suggests that friends and high ranking individuals may be less likely to engage in chase-
flee behaviors because they anticipate low risk and high potential for reconciliation due
to their shared bond and/or high status. Large parties may also constrain this behavior due
to space limitations.
Victim flee. Victim flee was a binomial variable indicating whether or not the
victim fled from the aggressor during the conflict. By definition, victims could flee
without being chased. The best fit model (AIC = 743.6) indicated three predictors. The
best predictor was the overall conflict score, which was a composite measure of
aggression type (threat = 0, agonistic = 1), directionality (unidirectional = 0, bidirectional
= 1), and wounding (no wound = 0, yes wound = 1) ranging from 0-3. As the conflict
score increased, the victim was less likely to flee (β ± SE = -2.906 ± 0.732, z = -3.968, p
< 0.001). Specifically, victims involved in conflicts with scores of zero were significantly
more likely to flee than victims involved in conflicts with scores of one (β ± SE = -1.926
± 0.525, z = -3.668, p = 0.001), two (β ± SE = -3.326 ± 0.646, z = -5.153, p < 0.001), or
three (β ± SE = -3.865 ± 1.041, z = -3.712, p < 0.001). Victims involved in conflicts with
a score of one were significantly less likely to flee than victims involved in conflicts with
scores of two (β ± SE = -0.400 ± 0.363, z = -3.853, p < 0.001). The second best predictor
was severity showing a linear relationship between increasing severity and the victim’s
likelihood to flee (β ± SE = 1.771 ± 0.587, z = 3.018, p = 0.003). Conflicts of the lowest
severity (i.e., threats) were more likely to result in the victim fleeing than conflicts of the
highest severity (β ± SE = 2.750 ± 0.866, z = 3.175, p = 0.007). The final predictor was

100
third-party bystanders. As the number of bystanders increased the likelihood that the
victim fled decreased (β ± SE = -0.068 ± 0.337, z = -2.016, p = 0.043).
Victims appeared to flee to avoid aggression escalation and the impeding
associated risks. A quick escape immediately reduced the cost of attack. However, if
aggression quickly escalates before the victim has the opportunity to flee, the victim is
essentially forced to stand their ground and defend himself/herself. Fleeing however
likely increases a victim’s risk of third-party attacks if there are many bystanders. The
more bystanders nearby, the more individuals a victim must flee passed/avoid.
Aggressor scream. Aggressor scream was a binomial variable indicating whether
or not the aggressor screamed during the conflict. The best fit model (AIC = 166.4)
indicated three predictors. The best predictor was the aggressor’s estrous state showing a
linear relationship (β ± SE = 5.003 ± 1.289, z = 3.882, p < 0.001). Specifically, male
aggressors were less likely to scream than fully deflated (β ± SE = 2.524 ± 0.565, z =
4.465, p < 0.001), partially swollen (β ± SE = 4.743 ± 1.587, z = 2.989, p = 0.013), and
fully swollen female (β ± SE = 6.718 ± 1.764, z = 3.808, p < 0.001) aggressors. Conflicts
of the lowest severity and with the lowest conflict scores were less likely to elicit
aggressor screams compared to conflicts with higher severity and scores (β ± SE = -3.641
± 1.139, z = -3.196, p = 0.005, β ± SE = 4.927 ± 1.210, z = 4.071, p < 0.001,
respectively).
In general, aggressors rarely screamed (6%). The above pattern indicates that
aggressors were more likely to scream when they were also experiencing distress from
severe conflicts or were potentially fearful. Female aggressors likely screamed more than

101
males because they were fearful of losing the conflict or being attacked by a male
bystander (i.e., policing behavior).
Victim scream. Victim scream was a binomial variable indicating whether or not
the victim screamed during the conflict. The best fit model (AIC = 560.7) indicated seven
predictors. The best predictor was the type of aggression with agonistic aggression
eliciting more victim screams than threats (β ± SE = -1.432 ± 0.287, z = -4.991, p <
0.001). The second best predictor was conflict type in which redirected aggression
produced fewer victim screams than renewed (β ± SE = 2.921 ± 0.859, z = 3.402, p =
0.003) or single occurrence (β ± SE = 1.022 ± 0.292, z = 3.499, p = 0.002) aggression.
Sex and parity also predicted victim screaming. Male aggressors elicited victim screams
more often than nulliparous (β ± SE = -4.313 ± 1.404, z = -3.072, p = 0.005) and parous
(β ± SE = -1.124 ± 0.493, z = -2.280, p = 0.052) female aggressors, and male victims
screamed less than both nulliparous (β ± SE = 1.379 ± 0.523, z = 2.641, p = 0.021) and
parous (β ± SE = 0.820 ± 0.299, z = 2.741, p = 0.016) female victims. Victims produced
screams more when chased by the aggressor (β ± SE = 0.604 ± 0.283, z = 2.136, p =
0.033) and when they fled from the aggressor (β ± SE = 0.700 ± 0.262, z = 6.672, p =
0.008). Finally, victims were more likely to scream when they lost to a coalition than
when they simply lost to a dominant aggressor (β ± SE = -1.264 ± 0.605, z = -2.089, p =
0.037). A similar pattern was present for no winner draws in which victims of
coalitionary wins were more likely to scream than victims involved in conflicts with no
clear winner, but this was only a trend (β ± SE = -1.294 ± 0.692, z = -1.869, p = 0.062).
These predictors demonstrate a clear pattern in which victim’s screams are calibrated to

102
the overall seriousness of the conflict and likely advertise this information to other party
members.
Intervention. Intervention was a binomial variable indicating whether or not a
third party bystander intervened in the conflict. The best fit model (AIC = 288.7)
indicated three predictors. The best predictor was severity showing a significant linear
relationship in which the likelihood of an intervention increased as conflict severity
increased (β ± SE = 1.797 ± 0.383, z = 4.695, p < 0.001). The second best predictor was a
preferred social partners measure showing that if the victim was a contact affiliation
preferred social partner of the aggressor then a third party bystander was less likely to
intervene (β ± SE = -1.026 ± 0.445, z = -2.303, p = 0.021). Total party size was the last
predictor showing that as party size increased so did the probability of a third party
intervention (β ± SE = 0.074 ± 0.034, z = 2.180, p = 0.029), possibly because a PSP was
more likely to be in the party.
Opponent end distance. Opponent end distance was an ordinal measure
indicating the distance between opponents immediately after the conflict ended (i.e.,
victim stopped fleeing, aggressor stopped chasing, no more contact, etc.). End distance
was ranked as 1 (within 5 m), 2 (between 5 and 10 m), or 3 (beyond 10 m). The best fit
model (AIC = 1374) indicated six predictors. Partly by definition, victim flee was the
best predictor showing that opponents were further apart following the conflict if the
victim fled (β ± SE = 0.292 ± 0.062, t = 4.708, p < 0.001). However, aggressors could
chase the victim to potentially reduce opponent distance, but this pattern was not found.
The next best predictor was number of bystanders present, showing that as the number of

103
bystanders increased the end distance between opponents decreased (β ± SE = -0.045 ±
0.010, t = -4.597, p < 0.001). Aggression type was the third best predictor indicating that
threats resulted in less distance between opponents than agonistic aggression (β ± SE =
-0.204 ± 0.062, t = -3.279, p = 0.001). Total party size predicted an opposite pattern than
number of bystanders. As party size increased, end distance between opponents also
increased (β ± SE = 0.018 ± 0.005, t = 3.213, p = 0.001). As overall mutual preferred
social partners strength increased between opponents, end distance between opponents
decreased (β ± SE = -0.262 ± 0.087, t = -3.000, p = 0.003). Specifically, opponents that
shared no mutual preferred social partner parameters had a greater end distance than
opponents that shared all three preferred social partner parameters (β ± SE = -0.332 ±
0.118,t = -2.815, p = 0.024). Victim parity also predicted opponent end distance with
parous female victims staying closer to their opponents immediately after the conflict
than both male (β ± SE = -0.282 ± 0.096, t = -2.939, p = 0.009) and nulliparous female (β
± SE = -0.396 ± 0.120, t = -3.287, p = 0.003) victims.
Conflict duration. Conflict duration was a continuous variable indicating the time
elapsed in seconds between the first aggressive action and the last between opponents.
Conflict duration was reclassified as an ordinal variable for analysis (very short = < 6 s,
short = 6 to 9 s, long = 10 to 20 s, very long = 20+ s). The best fit model (AIC = 1738)
contained seven predictors of conflict duration. The best predictor was if the aggressor
chased the victim. Conflicts involving aggressor chases were on average longer than
conflicts without chases (β ± SE = 0.593 ± 0.081, t = 7.304, p < 0.001). The second best
predictor was directionality with bidirectional conflicts on average lasting significantly

104
longer than unidirectional conflicts (β ± SE = -0.635 ± 0.121, t = -5.256, p < 0.001).
Aggression type was the third best predictor with agonistic conflicts on average lasting
significantly longer than conflicts only containing threats (β ± SE = -0.457 ± 0.091, t =
-5.029, p < 0.001). Conflict type was the next best predictor. Conflicts containing
renewed aggression lasted significantly longer than other conflicts (Figure 3.3).
Figure 3.3. Influence of conflict type on overall conflict duration. Redirected indicates
a second bout of aggression occurring between the aggressor and a bystander or the
victim and a bystander during the post-conflict period. Renewed indicates a new bout of
aggression reoccurring between the original opponents during the post-conflict period.
Single indicates one independent bout of aggression between opponents.

Specifically, single occurrence conflicts were significantly shorter than renewed
aggression conflicts (β ± SE = -0.880 ± 0.130, t = -6.780, p < 0.001) and redirected
aggression conflicts that also contained renewed aggression (β ± SE = -0.883 ± 0.285, t =
-3.098, p = 0.009). Renewed aggression conflicts were significantly longer than single
redirected aggression conflicts (β ± SE = 1.007 ± 0.165, t = 6.088, p < 0.001).
Average
Conflict
Duration

105
Furthermore, redirected aggression conflicts containing renewed aggression were
significantly longer than single redirected aggression conflicts (β ± SE = 1.011 ± 0.301, t
= 3.354, p = 0.004). There was no significant difference between single occurrence
conflicts and redirected aggression conflicts or renewed aggression conflicts and
redirected aggression conflicts that contained renewed aggression.
Conflict winners also influenced conflict duration as conflicts involving
coalitionary wins lasted significantly longer than conflicts between female dyads (β ± SE
= -0.746 ± 0.199, t = -3.739, p = 0.002), and conflicts where the dominant opponent won
(β ± SE = -0.751 ± 0.145, t = -5.191, p < 0.001), the subordinate opponent won (β ± SE =
-0.792 ± 0.241, t = -3.289, p = 0.008), and no winner draws (β ± SE = -0.591 ± 0.192, t =
-3.087, p = 0.016). If the victim screamed during the conflict, this also predicted longer
duration (β ± SE = 0.229 ± 0.096, t = 2.375, p = 0.038).
Discussion
Preferred Social Partners and Aggression
Male-male dyads. Gilby and Wrangham (2008) reported that from 1999-2004 that
male dyads had an average proximity association index (using CAI calculation, see
Chapter 2) of 1.44. Interestingly, male dyads in this study, during a different alpha tenure
with new adult male dyads, replicated this finding. The relative value of male social
bonds therefore appears to be consistent across time. During the study period, male dyads
demonstrated a more frequent, consistent, and mutual pattern of association relative to
dyads containing females (supports H
5
P
1
). Male dyad preferred social partnership was
also steady across measures of proximal association, agonistic support, and contact

106
affiliation. In fact, 12.7% of male dyads exhibited the strongest level of mutual PSP,
sharing PSP on all three measures. This pattern of male-male association supports what is
known about male chimpanzee gregariousness (Pepper et al., 1999), cooperation
(grooming: Mitani, 2005; Watts, 2002; meat sharing: Mitani & Watts, 2001; Stanford,
1998), and long-term (Goodall, 1986; Mitani, 2009) and short-term (Fedurek et al., 2013)
social bonds (reviewed in Muller & Mitani, 2005). During the study period, there were
two maternal brother dyads (LK-AJ and PG-PB), suggesting that kinship is not the only
factor promoting PSPs among males (Langergraber et al., 2007).
For male dyads, mutual PSPs fought more than one-sided PSPs or neutral partners
(supports H
6
for males), however male PSPs fought less than mixed sex PSPs (H
6
P
2
unsupported) but more than female PSPs. For mutual PSPs, the chance of fighting
increased as time spent together [either in close proximity, agonistic support, or contact]
increased. The chance of fighting continued to increase as male PSP strength increased
(supports H
6
P
3
). This presents an obvious tradeoff for males. While having PSPs has
obvious benefits (e.g., coalitionary support, grooming partners, dominance status, meat
sharing, etc.), it seems to come at the price of increased aggression toward your PSP.
However, the stronger the relationship, the more likely it is to withstand the frequent
bouts of aggression through relationship repair (Chapter 4).
Male-female dyads. Mixed sex dyads, as expected, were much less gregarious
than males. However, PSP status of mixed sex dyads was skewed toward dyads
containing parous estrous females (supports H
5
P
2
; Pepper et al, 1999). All but one (i.e.,
PB-LN) mixed PSP dyad included a parous female. Of the 12 male-female mutual

107
proximity PSP dyads, 11 of the dyads consisted of each adult male with the same parous
female (TG). Early into the study period, TG had lost her baby and started cycling again.
Since TG was the only cycling parous female during the entire study period, all adult
males were very interested in maintaining close proximity to TG whenever she was
partially and fully swollen. This pattern with TG, however, did not extend to the other
PSP parameters. Other females appeared to avoid prolonged proximity with males,
possibly to avoid aggression towards themselves or their offspring (supports infant safety
hypothesis: Otali & Gilchrist, 2006). Male-female dyads also rarely provided agonistic
support to each other (mutual agonistic support PSP = 1% of mixed sex dyads). The two
mutual agonistic support PSPs were KK-OU and ST-LR. While female linear rank could
not be calculated, OU and LR (resident female) are arguably the two highest-ranking
females. Male PSPs provided the majority of their agonistic support when a low-ranking
male repeatedly attacked the high-ranking female, and these females supported their male
PSPs when attacking low-ranking females.
Like males, females are equally likely to be victims of male aggression (Muller,
2002). This study found that male-female mutual PSP dyads engaged in more aggression
than one-sided or neutral partners (H
6
supported for mixed sex dyads). As overall mutual
PSP strength increased, the probability of aggression also increased (supports H
6
P
1
).
Again, there is a tradeoff. While there may be benefits for mixed sex dyads to have some
PSP affinity, too much increases the dyad’s aggression rates.
Female-female dyads. In general, wild female chimpanzees are much less
gregarious than males (Wrangham & Smuts, 1980; Wrangham, 2000), rarely grooming

108
(Wrangham et al., 1992) or associating (Gilby & Wrangham, 2008) with one another.
While this study supports these generalizations about the East African subspecies,
Kanyawara female dyads’ association patterns nearly doubled (0.86) relative to rates
reported in 1999-2004 (mean = 0.49; Gilby & Wrangham, 2008). However, the
Kanyawara females, on average, still associated much less often than males, having fewer
PSPs, and rarely grooming one another (mutual contact affiliation PSP = 3% of female
dyads).
Interestingly, female dyads provided agonistic support more than expected and
higher than mixed sex dyads (H
5
P
3
unsupported). While agonistic support between non-
kin females is reportedly rare (Townsend et al., 2007; Kahlenberg et al., 2008), this study
found agonistic support for non-kin female dyads to be higher than expected (14%) and
was an important PSP parameter. Non-kin females gave each other agonistic support in
two predominant conditions, parous females jointly: 1) attacked nulliparous (immigrant)
females, and 2) protected infants or parous females when being brutally attacked by adult
males. This result is particularly meaningful. In protection contexts, the infant or parous
female being attacked, typically by a male, is likely incurring serious physical costs
before the protector can intervene. Furthermore, in the case of an infant attack, the costs
can be fatal. Thus, it is important for females to quickly respond and support each other
agonistically under this circumstance. Females have a much better chance of winning
(i.e., successful protection) against males, if multiple individuals are working together to
retaliate. There were several instances of mobbing where three or more females jointly
intervened to rescue a female under brutal attack. If protection were the driving force

109
behind female association, contact affiliation and proximal association among females
would not be as important. The Kanyawara females’ association patterns support this
idea. One might expect that females would need to be in close proximity to provide
protection to their agonistic support PSPs. However, this was not the case. Screams from
female PSP victims often recruited help from far away (> 10 m) females that were out of
sight but in auditory range.
Unlike male and mixed sex dyads, female dyads were less likely to engage in
aggression if they were mutual PSPs compared to one-sided PSPs and neutral partners
(H
6
unsupported for female dyads). Uniquely for female dyads, the higher their PSP
strength, the lower the risk of aggression (H
6
P
1
unsupported only for female dyads). This
result supports the idea that female social (Sapolsky, 1993; reviewed in Muller & Mitani,
2005) and rank (Pusey et al., 1997) relationships are more stable than those of males.
Most aggression is directed from resident females to immigrant females (neutral
partners), and not towards other resident females (possible PSPs; Goodall, 1986; Muller,
2002; Nishida, 1989). Further supporting this notion, parous females preferred
nulliparous females as victims more than fellow parous females or males.
Maximizing the Benefits of Aggression
In 2002, Muller reported on the agonistic relations among the Kanyawara
chimpanzees. Like Muller, the current study found strong linear rank among males with
the alpha receiving the majority of pant grunts, but failed to determine a linear hierarchy
for females due to insufficient data. In chimpanzees, rank is largely determined by the
outcome of repeated dyadic aggressive contests between individuals, with coalitions

110
potentially playing an important role in achieving and maintaining status (Nishida, 1983;
Nishida & Hosaka, 1996). While there are obvious costs to maintaining high rank (i.e.,
metabolic costs: Muller & Wrangham, 2004b; health: Sapolsky, 2005; lethal wounding:
Fawcett & Muhumuza, 2000, Goodall, 1992, Nishida et al., 1995), it can afford many
benefits to both males and females through resource monopolization (i.e., food and mate
exploitation).
Adult males were the more likely initiators of aggression (supports H
1
P
1
) but this
difference was not as pronounced as that reported by Muller (present study: males three
times more aggressive than females, Muller study: males 14 times more aggressive than
females). However, during Muller’s study there were more estrous females, which might
have resulted in higher rates of male-male and male-female aggression due to increased
sexual competition between males and increased sexual coercion between males and
females. As a “rule”, male chimpanzees are socially dominant to females (Goodall, 1986;
Takahata, 1990). Consistent with theory, males exhibited higher rates of aggression
relative to females (Muller & Wrangham, 2004a) since the general pattern in mammals is
dominant individuals are more aggressive than subordinates (Clutton-Brock et al., 1976;
Cote, 2000; Creel et al., 1997; Hrdy & Hrdy, 1976). The alpha male was responsible for
initiating the majority of conflicts (19%), and the four highest-ranking males were in total
responsible for 52% of conflict initiations. This pattern is consistent with data from other
long-term chimpanzee study sites (i.e., Gombe: Bygott, 1979, Kanyawara: Muller, 2002;
Mahale: Nishida & Hosaka, 1996; Tai: Boesch & Boesch-Achermann, 2000). With
dominant individuals initiating more conflicts than subordinates, conflicts likely afforded

111
a clear social benefit to the aggressor by reaffirming their dominant position in the
hierarchy (Wittig & Boesch, 2003b) and gaining preferred access to the fought over
resource.
Interestingly in Muller’s (2002) study, adult female-female conflicts were never
observed, whereas this study analyzed 71 adult female-female conflicts and observed
many more. The rate of adult female aggression appears to have changed over the last
decade at Kanyawara and now is more similar to rates reported at Gombe, Mahale, and
Tai (Goodall, 1986, Kutsukake & Castles, 2004, Wittig & Boesch, 2005) and may
suggest an increase in food competition, core area overlap, and/or immigrant females.
During the study period, there were four relatively new nulliparous female immigrants
who did not have established core areas [compared to only one in Muller’s study].
Female-female aggression, especially between parous and nulliparous females, likely
afforded a high resource benefit, in which winners had greater access to preferred food
items or feeding spots. In support of this idea, during Muller’s study female core areas
were more evenly distributed throughout the chimpanzees home range: 7 central females,
5 northern females, 3 southern, and 2 immigrant females. Females in the present study
appeared to have much greater overlap. While there were still 7 consistent central
females, 3 (of the 4) northern females and the one southern female spent much of their
time in the central areas. Additionally, the four immigrant females were often in central
areas. While females were not always in the same party, spending significant periods of
time in the central areas meant females often “bumped” into one another. The more even
distribution of female core areas throughout the home range during Muller’s study may

112
therefore have influenced lower levels of aggression relative to the present study in which
more females were centrally located and frequently encountering one another.
In large part, the context of aggression for males and females mirrors what Muller
and others (Goodall, 1986; Nishida, 1989, Wittig & Boesch, 2003b) have reported with
males exhibiting more aggression during reunion and sexual competition (supports
H
2
P
1
), and females in plant-food competition and protection (supports H
2
P
2
).
Alternatively, the context social excitement played a more dominant role in aggression
for both sexes in the present study. The sex differences in context demonstrate that adult
chimpanzees fought more in situations of high personal benefit. Females focused on food
and protecting their offspring have higher overall fitness (Emery Thompson et al., 2007;
Pusey et al., 1997). Overall fitness benefits outweighed the fighting costs. Males,
however, focused their fights more in social (i.e., reunions) and sex contexts, which are
known to increase their potential fitness through social status maintenance (Baker &
Smuts, 1994; de Waal & Hoekstra, 1980; Goodall, 1986; Nishida, 1989) and mating
opportunities (de Waal, 1982; Hasegawa & Hiraiwa-Hasegawa, 1990). In the Pan genus,
high-ranking males sire more offspring than low-ranking males (Boesch et al., 2006;
Gerloff et al., 1999; Constable et al., 2001) either through female choice (Matsumoto-
Oda, 1999; Stumpf & Boesch, 2005, 2006) or [more likely] sexual coercion (Muller et
al., 2007; Muller et al., 2009).
Estrous state and parity also played roles in directed aggression towards females.
Males attacked parous females more than nulliparous females (supports H
3
P
1
), but
estrous state influenced these attacks. Males attacked more parous females when deflated

113
as compared to swollen, but attacked more nulliparous females when swollen as
compared to deflated (partially supports H
3
P
2
). At first glance, these finding may seem
contradictory to the sexual coercion hypothesis reviewed by Muller & Wrangham (2001,
2009) in which coercion and aggression increase when males are in the company of
periovulatory parous females. However, the present study did not collect hormonal data
to show that the parous females were actually ovulating when maximally swollen
(Deschner et al., 2004). Only four parous females (UM, TG, OU, and WA, Table 2.1)
exhibited full swellings during study period. By the start of the study period, UM was
already confirmed pregnant but continued to swell for several months (captive: Wallis &
Lemmon, 1986; wild: Wallis & Goodall, 1993). UM was ignored by the males during
these pregnancy swellings. OU and WA had full swellings several times during the study
period but both still had young suckling infants ages 2 and 3, respectively. Typically, the
interbirth interval for female chimpanzees is around 5 years (Wallis, 1997). This ‘false
start’ swelling when females are still lactating, exhibited by OU and WA, can be sporadic
and occur several months or a year before females are actually ovulating (Wallis, 1992).
While males showed some sexual interest in these females during their ‘false start’
swellings (Wallis & Bettinger, 1993), they did not compete with other males for mating
opportunities or exhibit increased aggression towards the females.
TG was the only parous female that generated high interest and aggression when
fully swollen. [TG’s infant died and she started cycling within a few months.] In general,
TG received high rates of aggression as her swelling increased but once fully swollen,
she quickly submitted to soliciting males, seemingly to avoid further aggression. Sexual

114
coercion, in this one case I would argue, occurred just prior to the female’s periovulatory
period, in which the males established an aggressive history with her that resulted in her
immediate submission when maximally swollen. Given that no other parous females were
cycling at the time, this particular female was a limited commodity and likely highly
anticipated by the males. So while male aggression did increase in the presence of the
periovulatory parous female when she was maximally swollen, it was not necessary
directed toward her. Instead, their aggressive energy was often directed toward other
males in the context of mate guarding (Watts, 1998). Copulatory submissions reduced the
parous female’s probability of receiving aggression. Furthermore, after experiencing such
high aggression and excitement surrounding her return to cycling, during TG’s second
periovulatory period she managed to sneak away on a two-month-long consortship
(Goodall, 1986) with a young, low ranking male (PB). While this study presents some
evidence of increased aggression when parous females are in estrous (supports H
3
P
2
),
long-term data from Kanyawara presents a much clearer pattern of male aggression and
sexual coercion during parous periovulatory periods (Muller et al., 2007).
The differences in aggression toward parous and nulliparous females is likely due
to the males’ differential interest in mating with old (parous) versus young (nulliparous)
females. Male chimpanzees prefer mating with older, not younger, females (Muller &
Wrangham, 2004; Muller et al., 2006) likely because these females have experience
raising young and have increased fecundity (Muller et al., 2007). While males generally
prefer mating with parous females, during this study period the lack of periovulatory

115
parous females may have by default made periovulatory nulliparous females more
attractive.
Minimizing the Costs of Aggression
While many costs of aggression are visibly obvious (e.g., wounding, energy
expenditure, etc.), social costs may be more cryptic. In general, the more valuable a
relationship, the more potential damage a conflict can cause to future levels of tolerance
and cooperation (Aureli & Smucny, 2000). Physical and social costs to the aggressor can
however be managed and possibly minimized during, and potentially after (e.g.,
reconciliation: Chapter 4, consolation: Chapter 5), the conflict. The Kanyawara
chimpanzees used several strategies to minimize both the physical and social costs of
aggression.
Sex, rank, and preferred social partnerships. Several intrinsic and relationship
factors of the opponents affected their decisions during the conflict. Sex, estrous state,
and parity were the three most common predictors of conflict variation in severity,
wounding, and screaming (supports H
3
). Parous and fully swollen female aggressors
were involved in conflicts of higher severity with wounding more than nulliparous and
non-swollen female aggressors and male aggressors. Victims of female aggression were
less likely to retaliate and were typically nulliparous females. Nulliparous females are
known to receive intense aggression from parous females due to resource competition
(Nishida, 1989). Thus resident female aggressors expected to win these conflicts (89%
won) and perceived their risk to be low. Additionally, female aggressors and victims
screamed more than their male counterparts (Slocombe & Zuberbühler, 2005), possibly

116
to recruit aid from a high-ranking male (de Waal, 1982; Slocombe & Zuberbühler, 2007)
especially if aggression costs were unexpectedly high.
For male aggressors, the potential physical and social costs were increased
compared to females as a whole. Male aggression was typically more severe with greater
wounding potential compared to female aggression (against other females), possibly
because male victims more commonly retaliated during the conflict compared to female
victims (supports H
1
P
1
). This increased the wounding potential and energy expenditure
for male aggressors. Muller (2002) found higher levels of aggression intensity when
periovulatory parous females were present in the party, and in this study conflict severity
was highest for male and parous female victims (supports H
3
P
1
, but H
3
P
2
is
unsupported).
Despite an undetermined female dominance hierarchy, it is well known that all
adult male chimpanzees are socially dominant to females (Goodall, 1986; Takahata,
1990). Therefore, all females were ranked as a group below the lowest ranking adult
male. As the aggressors’ rank decreased and rank difference between opponents
narrowed, conflicts were more often bidirectional (supports H
4
P
2
). As the victims’ rank
decreased and the rank difference between opponents widened, aggressors chased fleeing
victims more often. Victims recognized the potential high costs of attacks and chases
from the highest-ranking individuals and fled instead of fighting back because their
chance of winning was low. However, when unexpectedly attacked by a lower ranking
individual, victims retaliated because their likelihood of winning was high (Silk, 2002b).

117
While PSP status influenced the occurrence of aggression, it minimally affected
actual conflict variation. Conflicts between aggressors who were preferred agonistic
supporters of victims were lower in severity than conflicts between neutral opponents
(partially supports H
7
P
1
). Opponents sharing an overall mutual preferred social
partnership were less likely to chase one another during conflicts (partially supports
H
7
P
2
) and maintained closer proximity immediately after conflicts (supports H
6
P
3
).
Interestingly, if the aggressor was a contact affiliation PSP of the victim, third party
bystanders were less likely to intervene. This seemingly simple result provides some
support for the social intelligence hypothesis (Cheney & Seyfarth, 1990; Tomasello &
Call, 1997), showing that chimpanzees not only have a good understanding of their own
personal relationships, but they are also aware of other individuals’ preferred
relationships and adjust their behavior according to these external relationships.
Overall conflict intensity. An unanticipated result was that conflict behaviors had
a much larger impact on opponent response than PSP status. Simply put, opponent
response was based more on the actions of the other individual during the conflict than
whether or not opponents were friends. As the overall intensity (accounting for agonistic
aggression, wounding, bidirectional attacks, chasing, screams, and duration) of a conflict
climbed, the costs grew considerably, and it triggered a chain reaction whereby each
opponent responded with increasingly aggressive behavior either in attack or defense.
Intensity amplified the overall physical costs of aggression and prolonged the conflicts,
theoretically causing further damage to PSP relationships. As conflict intensity increased,
the likelihood of opponent screams also increased. Screams advertised to other party

118
members the degree of conflict intensity and distress (Slocombe & Zuberbühler, 2005)
and were used to recruit help (Slocombe & Zuberbühler, 2007). In fact, third-party
interventions were more common during conflicts of increased severity. Following these
intense costly conflicts, victims fled more often and opponents kept their distance, likely
to reduce the risk of renewed aggression; however, this tactic was not always successful.
Intense conflicts were more likely to result in renewed aggression between opponents,
possibly because intense conflicts 1) cause more damage to the relationship between
opponents or 2) result in unclear winners (Wittig & Boesch, 2003a).
Bystander and party size also influenced conflict intensity. As the number of
bystanders (i.e., party members in view of the conflict) increased, so did the conflict
intensity. Since social status is often decided by conflict outcomes (Riss & Goodall,
1977; Silk, 2002b), opponents may fight harder when others are watching to assert their
dominance over not only their opponent but also, indirectly, over onlookers. It may be a
way of communicating [with the entire party] a message of “look at how tough I am, no
one better mess with me.” When fewer individuals are nearby, each opponent may not
need to overtly defend their reputation, especially if surrounded by PSPs.
However, there was an interesting difference in opponent distance relative to
bystander and party size. Immediately after the fight ended, opponents remained in closer
proximity if there were many bystanders, but if party size was high, opponents then
dispersed. There are several possible explanations for this. As the number of bystanders
increased, opponents remained close because 1) it may be spatially more difficult for
opponents to disperse when crowded by bystanders, 2) walking past fewer bystanders

119
reduces the risk of redirected aggression from/toward bystanders, or 3) it could increase
the likelihood of nearby consolation, especially if a PSP bystander is nearby (Chapter 5).
Under these conditions, opponents should remain in close proximity to maximize the
benefits of possible conciliatory behavior between opponents (i.e. reconciliation) or with
bystanders (i.e., consolation), and minimize the cost of potential redirected third-party
aggression. However, as party size increased, opponents dispersed because they may 1)
selectively seek out interactions with more preferred, distally located party members, or
2) be avoiding renewed aggression.
The rate of third-party interventions also increased with total party size, perhaps
because PSPs were more likely to be nearby to intervene. If PSPs did not directly witness
the conflict, but were in the party, they would likely be in auditory range of recruitment
screams and would be able to assess the situation (Slocombe et al., 2010; Slocombe et al.,
2009). Then based on [vocal] triadic awareness (de Waal, 1982, p. 182), party members
could then quickly adjust their behavior to provide the necessary assistance or
intervention.
Conclusion
While the Kanyawara chimpanzees exhibited wide variability in directed
aggression, there were distinct patterns to their choices during the conflict. The
chimpanzees appeared to adjust their responses to their opponent based on preceding
behaviors within the conflict bout and intrinsic attributes of their opponent. Chimpanzee
aggressors quickly weighed the various costs and benefits of each situation to determine
if the conflict was “worth it.” However, their decision-making process did not stop there.

120
Once engaged in conflict, opponents continued to evaluate the situation and manage the
conflict. Certain mechanisms were more effective at managing the ongoing cost-benefit
battle (de Waal & Aureli, 2000). While the chimpanzees did not always perfectly
evaluate the net costs and benefits (e.g., initiators sometimes lost the conflict), the overall
patterns indicate they were skilled “decision-makers” who typically judged the conflict
accurately. Furthermore, the potential long-term costs to opponent relationships after
aggression could be further diminished if conciliatory behaviors followed (see Chapter 4
on reconciliation).

121
Chapter 4

Conflict and Relationship Determinants of the Occurrence and
Variability in Reconciliation in Wild Chimpanzees (Pan troglodytes)

122
Introduction
Aggression is an inevitable and universal component of social living. In
mammals, intergroup aggression is often driven by territorial displays and defense, and
typically results in spatial distribution and avoidance of opponents following the fight
(King, 1973; Scott, 1958). Alternatively, intragroup aggression often results in proximal
attraction and affiliative association of opponents afterwards (de Waal & van Roosmalen,
1979), especially in primates (de Waal, 2000). It is important for group-living animals to
maintain group cohesion and mitigate the costs of aggression within the group, despite
occasional conflict (Aureli & de Waal, 2000). Many nonhuman primate studies have
shown that intragroup aggression can be quite costly, beyond the obvious risk of injury
and high-energy expenditure, and can damage social relationships in immediate and long-
term ways (Aureli & van Schaik, 1991; Cords, 1992; Koski et al., 2007; Kutsukake &
Castles, 2004; Silk, Cheney, & Seyfarth, 1996; Watts, 1995; Wittig & Boesch, 2003a).
While conflict avoidance strategies (i.e., employing submissive and/or appeasing
behaviors) can minimize the occurrence and escalation of aggression (de Waal, 1996;
Preuschoft & van Schaik, 2000), these behaviors do not act to resolve conflict after it has
already occurred. Thus, individuals must find ways to minimize the negative effects of
intragroup aggression and restore levels of tolerance or affinity in the post-conflict
period.
Reconciliation
Peacemaking strategies are necessary to mitigate aggression and maintain group
stability and cohesion in social animals. Specifically, reconciliation, defined as affiliative

123
contact between opponents after fighting, is one mechanism through which opponents
can mitigate the negative effects of aggression. Primates use reconciliation to 1) repair
social relationships, particularly between valuable (high social or reproductive
importance) partners (Aureli & de Waal, 1997; Cords, 1992; de Waal & van Roosmalen,
1979; Wittig & Boesch, 2003a), 2) reduce elevated behavioral stress levels (Cooper et al.,
2007; te Boekhorst et al., 1991), and 3) reduce the probability of renewed aggression
(Aureli & van Schaik, 1991; Cords, 1992; Silk et al., 1996). Some non-primate species
also exhibit reconciliatory behaviors for similar reasons (domestic goats: Schino, 1998;
bottlenose dolphins: Samuels & Flaherty, 2000; Weaver, 2003; spotted hyena, Hofer &
East, 2000; Wahaj et al., 2001; wolves: Cordoni & Palagi, 2007).
While reconciliation has been studied extensively in captive primate populations
since the late 1970s (de Waal & van Roosmalen, 1979), little is known about the post-
conflict behaviors of their wild counterparts, especially chimpanzees. To date, only four
field studies have investigated the post-conflict behaviors of wild chimpanzees (Arnold &
Whiten, 2001; Kutsukake & Castles, 2004; Watts, 2006; Wittig & Boesch, 2003a). In
particular, wild chimpanzees have become well-known for their high rates of intragroup
aggression relative to other animal species including humans (Muller, 2002; Muller &
Wrangham 2004; Wrangham et al., 2006). During conflicts, chimpanzees make decisions
based on cost and benefit strategies (Wittig & Boesch, 2003b). These attributes make
wild chimpanzees good models in which to study reconciliatory strategies, variation, and
function. For example, high benefit partners (i.e., mutual grooming associates,
coalitionary partners, proximity associates, etc.) should show a higher tendency to

124
reconcile relative to low benefit partners (Aureli et al., 2002; Wittig & Boesch, 2005).
Concurrently, conflicts resulting in high costs (e.g., high intensity or severity) should
show a lower tendency to reconciliation unless the high benefit of reconciling outweighs
the high cost incurred during the conflict (e.g., reconcile to repair a valuable social
relationship despite incurring injuries during the conflict). Some studies have detected a
positive relationship between conflict intensity and relationship damage (de Waal &
Yoshihara, 1983; Koyama, 2001; Schino et al., 1998), whereas others have not (Aureli,
1997, Kutsukake & Castles, 2001; Castles & Whiten, 1998). With this variability in
response, it is therefore important to investigate this cost/benefit relationship further.
Additionally, there is a wide degree of variation in the rates and patterns of
reconciliation observed in both captivity and the wild (Arnold & Whiten, 2001; Baker &
Smuts, 1994; de Waal, 1986; de Waal & van Roosmalen, 1979; Fraser & Aureli, 2008;
Fuentes et al., 2002; Koski et al., 2007b; Wittig & Boesch, 2005). For example,
reconciliatory rates in captivity typically range from 22-48% (Casperd, 1997; de Waal &
van Roosmalen, 1979; Fraser & Aureli, 2008; Fuentes et al., 2002; Preuschoft et al.,
2002), whereas wild rates are lower (15-19%) with little deviation from the mean (15-
19%, Arnold & Whiten, 2001; Kutsukake & Castles, 2004; Wittig & Boesch, 2003a).
This variation is likely the result of differences in environment (i.e., captive vs. wild),
socio-ecology, and methodologies, further suggesting that while the act of reconciliation
is not an artifact of captivity (Colmenares, 2006), the frequency of reconciliation is.
These discrepancies make drawing broad conclusions about intraspecific variation quite
challenging.

125
General hypotheses. Based on patterns from past studies, general hypotheses (H)
and predictions (P) were as follows:
H
1
: The Kanyawara chimpanzees would have a similar group reconciliatory rate
(measures using the Corrected Conciliatory Tendency, CCT, Veenema et al., 1994) to
other wild chimpanzee populations, but significantly lower than most captive
populations.
P
1
: Male-male opponents will have the highest dyad CCT, followed by male-
female opponents, and then by female-female opponents.
P
2
: Male-female reconciliation rates will be influenced by parity.
H
2
: The occurrence of reconciliation will be influenced by conflict behaviors and the
associated costs.
P
1
: Victims will initiate reconciliation more than aggressors (to avoid renewed
aggression).
P
2
: High intensity conflicts (measured by aggression type, severity, and
directionality) will be reconciled more than low intensity conflicts.
H
3
: The occurrence of reconciliation will be influenced by rank.
P
1
: Opponents with a small rank difference will be more likely to reconcile than
opponents with a larger rank difference.
H
4
: The occurrence of reconciliation will be influenced by the context of the conflict.
P
1
: Opponents will be less likely to reconcile following conflicts over limiting
resources like food and sex.

126
P
2
: Opponents will be more likely to reconcile following conflicts involving
social factors like reunions, dominance assertions, displacements, etc.
Proximity and Reconciliation
While several studies have suggested that close-proximity was equivalent to using
affiliative contact behaviors in the post-conflict (PC) period, none had actually tested this
hypothesis (Call, 1999; Cords, 1993; Patzelt et al., 2009). Only recently have studies
started focusing on the importance of distancing strategies in PC behavior. Patzelt et al.
(2009) and McFarland and Majolo (2011) were the first studies to use close-proximity
(<1.5 m) as an affiliative behavior in the PC period. McFarland and Majolo (2013)
published an article emphasizing this huge gap in the PC literature stating that
interopponent distance regulation is often overlooked, ignored, and not tested, and were
the first to give evidence that close-proximity functions similarly to affiliative contact in
the PC period. They found that close-proximity approaches (<1.5 m) reduced the victim’s
anxiety in the PC period and served to reconcile the opponents. In another study
investigating the role of proximity in the PC period, Clay and de Waal (2013a) reported
that the best predictor of contact consolation in bonobos was the distance from the
bystander to the conflict. Independent of kinship, bystanders in close proximity (< 5 m)
to the victim were significantly more likely to console than those further away. These
recent studies on interopponent distance regulation in the PC period have provided new
insights in the study of conflict resolution, reviving interest in this field. Specifically, I
argue that interopponet distance regulation, defined as distance between opponents in the
PC period, may not be as relevant in the study of captive populations because dispersal

127
strategies are limited or absent in captivity. However, wild chimpanzee fission-fusion
social structure in a free-ranging habitat complicates applications of captive data to these
naturalistic populations.
In particular, I hypothesized that:
H
5
: Distancing strategies following aggression would play a key role in wild chimpanzee
PC behavior.
P
1
: Opponents in close proximity immediately following aggression will be more
likely to reconcile than those more distally located. Victims in close proximity of
their aggressor are at a greater risk of renewed aggression than victims further
away.
P
2
: Opponents within 5 m immediately following aggression will be the most
likely to reconcile, followed by those within 5 to 10 m, and then by those beyond
10 m.
Valuable Relationships and Reconciliation
Relationships play a key role in the lives of social animals, making many
decisions based on their personal relationships with other group members. The Valuable
Relationship Hypothesis (VRH) states that opponents with high relationship value [in
baseline conditions] should preferentially reconcile over those with neutral relationships
(de Waal & Aureli, 1997) to repair their relationship (Cords & Aureli , 1996; van Schaik
& Aureli, 2000), restore tolerance (Cords, 1992), and promote future affiliative
interactions between “friends” (Cords, 1997; Cords & Aureli, 2000). Many studies have
found support for the VRH (baboons: Castles et al, 1996; bonobos: Palagi et al., 2004;

128
capuchins: Leca et al., 2002; Manson et al., 2005; Verbeek & de Waal, 1997;
chimpanzees: Preuschoft et al., 2002; Wittig & Boesch, 2005; gorilla: Watts, 1995;
macaques: Aureli et al., 1989; de Waal & Yoshihara, 1993), while others have not
(baboons: Silk et al., 1996; chimpanzees: Casperd, 1997; Kutsukake & Castles, 2004;
macaques: Veenema et al, 1994; tamarins: Schaffner & Caine, 2005) or found mixed
results (langurs: Sommer et al., 2002; chimpanzees: Arnold & Whiten, 2001; Castles et
al., 1996) even within the same species.
Reconciliatory studies also vary on what factors determine relationship value (i.e.,
time in proximity, time spent grooming, agonistic/coalitionary support), and how
relationship value should be calculated (i.e., association indices: Cairns & Schwager,
1987; value, compatibility, and security: Cords & Aureli, 2000; preferred social
partnerships: Gilby & Wrangham, 2008), making comparisons difficult. Given the wide
range of behaviors/interactions used to determine relationship value in previous studies,
Watts (2006) called for a more detailed investigation on “what we mean by value and
how we measure it” with regard to conflict resolution. The above variation in response
and measurement led to the following hypothesis and predictions:
H
6
: Wild chimpanzee opponents who share valuable relationships will preferentially
reconcile because of the associated benefits compared to opponents who have neutral
relationships.
P
1
: Relationship value measured using Preferred Social Partnerships calculations
(PSPs, Gilby & Wrangham, 2008) will better predict reconciliation than other
methods calculating value (Cairns & Schwager, 1987; Cords & Aureli, 2000).

129
P
2
: Reconciliation will be more likely between PSPs that share multiple levels of
value (i.e., time in close proximity, time spent grooming, time providing agonistic
support) compared to PSPs that only share one.
P
3
: PSPs of mutual value will reconcile more than PSPs with one-sided
relationships.
Variation in the Components of Reconciliation
Previous research, including the present study, has repeatedly explored the five
“wh” questions (who, what, when, where, why) of reconciliation, but few studies have
tackled the “how” question in detail (Wittig & Boesch, 2005). To simply measure the
occurrence of reconciliation and the context in which it occurs is overlooking how
variable reconciliation is once the act begins. Reconciliation can vary considerably in
terms of its latency, approacher/initiator, pre-reconciliatory signal, action, opponent
response, directionality, complexity, and duration. Furthermore, this variation in the
components of reconciliation can be influenced by not only conflict and relationship
variables, but also by the preceding reconciliatory variables that occur during the act
itself. These reconciliatory variables can be viewed linearly as a chain reaction with one
preceding the next in time. For example, the latency is dependent on the first opponent to
approach the other. The initiator then follows based on the approach. The reconciliatory
action is based on the initiation, and so on. Using this model, it is possible to test how
each component of reconciliation influences the next over time. It is then possible to start
answering questions about 1) why there is so much variation in the components of

130
reconciliation, and 2) what variables best predict how an opponent will respond once they
have chosen to reconcile. It becomes an intricate web of behavioral and social factors.
This study aimed to determine how wild chimpanzees reconciled after conflicts
and what factors influenced/predicted their variation in reconciliatory components.
H
7
: Reconciliatory variation will be dependent on conflict behaviors and the associated
costs.
P
1
: Reconciliation between opponents that were involved in more intense
conflicts (measured by aggression type, severity, and directionality) will be
initiated more often by the victim, occur almost immediately after the conflict,
and be one-sided and brief compared to those reconciling after less intense
conflicts.
H
8
: Reconciliatory variation will be dependent on sex and relationship parameters and
the associated benefits.
P
1
: Female estrous state will affect male-female reconciliatory variation. Males
who reconcile with fully swollen females would be more likely to initiate and
prolong reconciliation compared to males who reconcile with deflated or partially
swollen females.
P
2
: Reconciliation between PSPs will occur shortly after the conflict with the
action being bidirectional and long in duration, whereas reconciling neutral
partners would have a longer latency to reconcile and the act would be
unidirectional and quick.

131
In summation, aims of this study were to investigate the rates, occurrence, and
variation in reconciliation relative to conflict, relationship, and preceeding reconciliatory
variables in the wild chimpanzee community of Kanyawara, Kibale National Park,
Uganda. This study was developed because 1) very little is known about reconciliation in
wild chimpanzees, 2) there are inconsistencies in the literature, and 3) methodologies
have not been strict and comprehesive. My project investigated the Kanyawara
chimpanzees’ post-conflict behaviors as complex interactive behaviors that affect the
occurrence and rate of one another. To better address methodological issues in past wild
studies (Arnold & Whiten, 2001; Kutsukake & Castles, 2001; Wittig & Boesch, 2003), I
adapted the PC methodology employed by Silk et al. (1996) on baboons to accommodate
the fluidity in wild chimpanzee party composition and size so that results are more
comparable to traditional post-conflict studies. I also incorporated more conflict,
relationship, and reconciliatory component variables than prior studies to better
determine which factors have the highest degree of predictability on the occurrence and
variation in reconciliation. This degree of detail allowed for better conclusions regarding
of the proximate and ultimate functions of reconciliation.
Methods
Study Site and Subjects
Data were collected from November 2010 through November 2011 in the
Kanyawara, Kibale Chimpanzee Project area in Kibale National Park, Uganda (0°13’-
0°41’N, 30°19’-30°32’ E; Figure 2.1). See Chapter 2 for study site and chimpanzee
community (Appendix A) details. Eleven adult male and six adult female chimpanzees

132
were selected for all-day focal follows (Table 2.1). Daily chimpanzee focals were
selected based on observability, party composition, and general location.
Data Collection
Baseline (focal) data. I conducted all-day continuous focal sampling (Altmann,
1974), following the focal chimpanzee from nest to nest. I recorded in detail each focal
chimpanzee’s general activity, interactions, and vocalizations in real time (Appendix B).
For interactions, I recorded the occurrence, directionality, duration, and partner of all
affiliative, aggressive, and submissive behaviors of the focal. I documented the estrous
state of each visible female daily (Deschner et al., 2003). Every 10 min, I recorded the
focal chimpanzee’s party size, composition, and proximity of other individuals to the
focal via instantaneous scan sampling (Altmann, 1974). I collected directed aggression
data between adults for all party members ad libitum (Altmann, 1974), recording the
conflict’s opponents (i.e., aggressor(s) and victim(s) identities), time, duration, location,
aggressive actions, severity, directionality, winner, context, and third-party bystanders.
Post-conflict (PC) data. Following a directed bout of aggression between two (or
more) adult chimpanzees within my focal’s party, I immediately began focal following
either the aggressor or victim (or both, depending on their accessibility and visibility) of
the conflict for a 10-min period. I collected all behaviors of the temporary focal
(aggressor and/or victim), including any conflict partner or third-party interactions and
proximities. I generally followed the observational procedures described by de Waal and
Yoshihara (1983) and Aureli et al. (1989), but modified them as necessary for suitability
in a wild setting (Whiten & Arnold, 2001; Wittig & Boesch, 2003a). If aggression

133
reoccurred within the 10-min period, I restarted the 10-min PC observation following the
last bout of aggression (Fuentes et al., 2002; Preuschoft et al., 2002).
Operational Definitions
Reconciliation was defined as affiliative contact between conflict partners
occurring within the 10-min PC period. However, this definition alone does not
sufficiently determine whether reconciliation was a true product of the previous conflict
or merely a product of chance. Traditionally, this is determined by comparing a PC
observation to a matched control (MC) observation (de Waal & Yoshihara, 1983).
However, this method is best suited for captivity, and given the fission-fusion nature of
wild chimpanzee society, is not suitable for determining reconciliation in wild
populations (Arnold & Whiten, 2001; Kutsukake & Castles, 2004). Instead of conducting
an independent MC observation for each specific PC observation, I used focal data to
calculate a mean baseline affiliation (MBA) rate for each dyad to compare to the post-
conflict period affiliation rate. While this method has been applied to monkey post-
conflict studies (macaques: Judge, 1991; baboons: Silk et al., 1996), this is its first
application to chimpanzees. See Chapter 2 for equation and calculation details. See
Appendix C for operational definitions of all dependent and independent variables.
Statistical Analyses
General statistics. To test for significant differences in each dyad’s baseline and
PC affiliation rates, I used Wilcoxon signed-ranks tests (Silk et al., 1996). From this, I
determined the number of attracted and dispersed pairs, and then calculated the overall
Corrected Conciliatory Tendency (CCT; Veenema et al., 1994) as well as the individual

134
CCTs for male-male, male-female, and female-female dyads. Wilcoxon signed-ranks
were used to test for significant differences in the frequency of attracted versus dispersed
pairs, and Mann-Whitney U tests were used to test individual mean CCTs between male-
male, male-female, and female-female dyads. See Chapter 2 for general analysis details.
Multivariate analyses. I used generalized linear mixed models (GLMMs) with a
binominal error structure and a logit link function to determine what variables predicted
the occurrence of reconciliation. The unit of analysis was each conflict, and each conflict
partner’s identity was designated as a random effect. I tested both conflict and
relationship variables in the models and used a step-up strategy to build the best fitting
model with the lowest Akaike’s information criterion (AIC; Tabachnick & Fidell, 2007).
All GLMM analyses were conducted in R version 3.0.1 statistical software (R Core
Development Team, 2012) using the ‘lme4’ package and the ‘glmer’ function (Bates et
al., 2013).
To determine the best predictors of reconciliation variation, I used GLMMs (as
described above) for the binomial dependent variables and linear mixed models (LMMs)
for the continuous dependent variables (see Appendix C). When building the models, I
again employed a step-up strategy whereby predictor variables (based on hypotheses)
were added to the model sequentially. Several models were built and tested, and the best
model was selected based on the lowest AIC. All LMMs were analyzed in R version
3.0.1 statistical software (R Core Development Team, 2012) using the ‘lme4’ package
and the ‘lmer’ function (Bates et al., 2013). To calculate p values of the LMMs, I used

135
the ‘pvals.fnc’ function that uses a Markov chain Monte Carlo stimulation with 10,000
iterations in the ‘languageR’ package (Baayen, 2011).
Post-hoc analyses for all models (both GLMMs and LMMs) were conducted
using the ‘glht’ function in the ‘multcomp’ package in R (Hothorn et al., 2008). All
predictor variables in all models were also tested for multicollinearity between predictors
using the variance inflation factor (VIF). Predictors with VIF scores higher than five were
considered to be collinear with other variables and were excluded from the model
(Craney & Surles, 2002; Rogerson, 2001). See Chapter 2 for a detailed description of
statistical analyses.
A large number of descriptive, conflict, and relationship dependent variable
categories (Appendix C) were tested for their predictor candidacy of reconciliation, both
individually and in applicable interactions, in separate GLMMs. Variables with low AIC
scores showing or approaching statistical significance were added sequentially to an
overall GLMM to build a best fit model for reconciliation.
To investigate the variation within reconciliation, I tested the following binomial
dependent variables in separate GLMMs to build the best fit predictor model for each:
prereconciliation signal, approacher, initiator, reconciliatory action, recipient response,
and reconciliation directionality. I tested the following continuous dependent variables in
separate LMMs to build the best fit predictor model for each: reconciliation latency,
complexity, and duration. Given that reconciliatory behaviors occurred as a sequence,
linearly in time, I could only include reconciliatory variables as predictors of another
reconciliatory variable if that variable occurred earlier in time than the reconciliatory

136
dependent variable for that model. The established time order of reconciliatory factors
was as follows (from occurring first to last in the reconciliation response sequence):
prereconciliation signal è reconciliation latency è approacher è initiator è
reconciliatory action è recipient response è reconciliation directionality è
reconciliation complexity è reconciliation duration. For example, when building the
models, prereconciliation signal, latency, approacher, initiator, action, and recipient
response could all be included as predictors of reconciliation directionality, along with
other conflict and relationship variables, but complexity and duration were excluded from
the model since they would have occurred afterwards and therefore could not be
considered a predictor.
Results
A total of 1054 conflicts were recorded between two or more chimpanzees over a
one-year period. Since I was only following some adults as focal chimpanzees (11 males,
6 females), I excluded conflicts that included 1) a non-adult conflict partner (N = 290), 2)
two conflict partners that were never focal subjects (N = 5), and 3) a conflict partner that
was generally observed less than 30 hr throughout the entire study period (N = 3). I
excluded any conflicts where the conciliatory outcome was unknown (N = 24), which
was usually the result of poor visibility. I also had to correct for multiple conflicts
resulting from renewed aggression, which is aggression reoccurring between conflict
partners within the 10-min PC period. I collapsed these conflicts into a single conflict
since these aggression events were not independent of one another and shared the same
conflict partners (resulted in 174 renewed conflicts collapsed into 74 analyzable

137
conflicts). For analyses, this resulted in 632 residual conflicts occurring between 181
adult conflict partner dyads out of the 279 possible adult dyads containing at least one
focal individual.
Reconciliation
Reconciliation occurred 123 times in the 632 conflicts with 57 of the 181 conflict
partner dyads using reconciliation at least once. The majority of reconciliation (52%)
occurred within the first minute of the PC period (Figure 4.1). Applying the Veneema et
al. (1994) method for standardizing the occurrence of reconciliation in a population, the
mean (±SD) group Corrected Conciliatory Tendency (CCT) was 15.8% ± 11.0%, and the
mean
Figure 4.1. Percent of reconciled dyads across 10-min post-conflict period.

(±SD) dyadic CCT was 14.4% ± 26.1%. As depicted in Figure 4.2, both male-male (CCT
23.2% ± 31.8%, N
1
= 42, N
2
= 33, U = 959.5, p = 0.004) and male-female dyads (CCT
14.4% ± 25.6%, N
1
= 106, N
2
= 33, U = 2223.0, p = 0.018) reconciled significantly more
than female-female dyads (CCT 3.0% ± 12.1%; supports H
1
P
1
), but there was no

138
significant difference between dyads involving males (male-male vs. male-female, N
1
=
42, N
2
= 106, U = 2552.5, p = 0.164). Looking closer at male-female dyads, males
reconciled significantly more with parous (CCT 18.9% ± 28.0%) versus nulliparous
females (CCT 0.6% ± 3.3%, N
1
= 80, N
2
= 26, U = 1467.5, p = 0.001; supports H
1
P
2
).
Parity did not influence reconciliatory tendencies in female-female dyads (parous female
– parous female CCT 5.6% ± 16.2% vs. parous female – nulliparous female CCT 0%, N
1

= 18, N
2
= 15, U = 150, p = 0.605).
When comparing each reconciled dyad’s rates of affiliation in the PC period to
affiliation rates in focal (baseline) observation, the data show that all baseline affiliation
rates (controlling for various partner proximity) were significantly lower than affiliation
Figure 4.2. Mean conciliatory tendency (CCT) across sex composition. Sex is
indicated by M for male and F for female. Parity is indicated by F
p
for parous female and
F
n
for nulliparous female. Asterisks indicate significant differences between specific
dyads.

*
*
*

139

Figure 4.3. Overall affiliation rates in post-conflict (PC) period compared to three
overall baseline observation rates, varying by distance of opponents, for reconciled
dyads.

rates in the PC period (Figure 4.3). Specifically, dyads (N = 57) were significantly more
likely to engage in contact affiliation immediately following a fight (within the 10-min
PC observation) than during baseline conditions when 1) in the same party (T = 1653, df
= 56, z = 6.56, p < 0.0001), 2) in view of one another (T = 1653, df = 56, z = 6.56, p <
0.0001), and 3) within 10 m of one another (T = -1496, df = 56, z = -6.10, p < 0.0001).
This finding conclusively shows that contact affiliation between former opponents within
the 10-min post-conflict period was indeed true reconciliation and was a product of the
previous conflict, not a product of chance affiliation.
Predictors of Reconciliation
The overall model that best predicted the occurrence of reconciliation, as
measured by the lowest AIC score (AIC = 510.3, N = 632), included only five predictor
variables, three of which were significant, with no collinearity [mean = 1.201] between

140
any independent variables (VIF of the independent variables were considerably below the
recommended cutoff value of 10, Craney & Surles, 2002; Kutner et al., 2004).
Opponent distance. The best predictor of reconciliation was opponent end
distance, which refers to the distance (in meters) between opponents immediately
following the last aggressive behavior. The opponent end distance was categorized as
being within 5m, between 5m and 10m, or beyond 10m. There was a significant negative
linear relationship between opponent end distance and opponent contact reconciliation (β
± SE = -1.875 ± 0.296, z = -6.346, p < 0.00001, supports H
5
), meaning the closer the
opponents were after a conflict, the more likely they were to reconcile. Opponents within
5 m immediately following the conflict were more likely to reconcile than opponents that
were 5 to 10 m apart (β ± SE = -1.162 ± 0.264, z = -4.399, p < 0.0001) and opponents
that were beyond 10 m (β ± SE = -2.652 ± 0.418, z = -6.346, p < 0.0001). Opponents that
were 5 to 10 m apart immediately following the conflict were more likely to reconcile
than opponents that were beyond 10 m (β ± SE = -1.490 ± 0.395, z = -3.770, p = 0.0004,
supports H
5
P
1
and H
5
P
2
).
Preferred social partners. The second best predictor of reconciliation was a
relationship variable that evaluated the overall Preferred Social Partner (PSP) status of
the dyad (supports H
6
P
1
). An overall PSP was someone who shared a high affinity for
their partner on at least two of the three relationship measures: proximal association (at
the level of party, within five m, and nearest neighbor), contact affiliation, and agonistic
support (see Chapter 2 for calculations of relationship variables). Overall PSP status was
determined by combining the PSP status of the three relationship variables into one

141
“super” relationship variable. Opponent dyads were then given a score ranging from 0-3
(0 = neutral, 1 = less preferred, 2 = moderately preferred, 3 = highly preferred). Each
opponent dyad’s overall score reflected the actual number of PSP relationship variables
(out of the three possible) shared between opponents, either mutually or one-sided.
Each relationship variable (i.e., proximal association, contact affiliation, and
agonistic support) was first tested individually in the contact reconciliation model and
each relationship variable was found to be a significant predictor of contact reconciliation
individually for the aggressor (p < 0.001) and the victim (p < 0.001), and also mutually
for both conflict partners (p < 0.005). Each of the relationship variables’ AIC scores were
then compared, which resulted in selecting the overall “super” PSP variable for the victim
(AIC = 576.6) as the best opponent relationship predictor for contact reconciliation.
There was a significant positive linear relationship between the overall PSP status of the
aggressor relative to the victim and contact reconciliation (β ± SE = 1.162 ± 0.317, z =
3.663, p = 0.0002), meaning as the aggressor’s PSP status increased for the victim, the
conflict partners were increasingly likely to reconcile (H
6
P
3
unsupported). When the
aggressor was a highly or moderately preferred overall PSP of the victim (sharing three
or two PSP variables), the conflict partners were significantly more likely to reconcile
than when the aggressor was a neutral (sharing no PSP variables) social partner of the
victim (highly PSP vs. neutral: β ± SE = 1.506 ± 0.448, z = 3.363, p = 0.004; moderately
PSP vs. neutral: β ± SE = 1.112 ± 0.365, z = 3.048, p = 0.012; supports H
6
P
2
). While
there was a trend for the victim’s highest scoring overall PSP (sharing three PSP
variables) to reconcile more frequently than less preferred PSPs (sharing only one PSP

142
variable), this trend only approached significance (β ± SE = 1.071 ± 0.430, z = 2.491, p =
0.060). There was no significant difference between highly preferred and moderately
preferred PSPs of the victim (β ± SE = 0.394 ± 0.419, z = 0.939, p = 0.781), nor
moderately preferred and less preferred PSPs (β ± SE = 0.677 ± 0.361, z = 1.874, p =
0.237), or less preferred PSPs and neutral social partners (β ± SE = 0.435 ± 0.342, z =
1.273, p = 0.577).
To determine which of the three relationship predictors (proximal association,
agonistic support, and contact affiliation) contributed the most to the overall “super” PSP
variable for the victim, the three predictors were tested against one another in a separate
model with only these three relationship variables (AIC = 572.5). In this post-hoc model,
the best predictor of contact reconciliation was proximal association (p = 0.001) followed
by agonistic support (p = 0.027) then contact affiliation (p = ns). This means
reconciliation was more likely to occur when the aggressor was a PSP of the victim for
the relationship variables of proximal association (β ± SE = 0.951 ± 0.293, z = 3.240, p =
0.001) and agonistic support (β ± SE = 0.638 ± 0.289, z = 2.211, p = 0.027), and in turn,
these two variables contributed significantly to the general predictability of the overall
“super” PSP variable (supports H
6
P
2
). Contact affiliation PSP status of the victim’s
aggressor did not significantly contribute to the predictability of the overall “super” PSP
variable (β ± SE = 0.117 ± 0.291, z = 0.403, p = 0.687).
Sex and parity. The third best predictor of reconciliation was the aggressor’s sex.
Male aggressors were significantly more likely to reconcile than female aggressors (β ±
SE = 1.817 ± 0.724, z = 2.511, p = 0.012). Victim parity was also included in the best fit

143
model but was not a significant predictor of contact reconciliation; instead there was only
a trend for nulliparous female victims to reconcile less than parous female (β ± SE =
2.374 ± 1.120, z = 2.139, p = 0.073) and male (β ± SE = -2.093 ± 1.135, z = -1.844, p =
0.065) victims.
When opponent dyad sex composition (i.e., male-male, male-female, female-
female) was substituted into the model for both aggressor’s sex and victim’s parity,
female-female conflict partners reconciled significantly less as compared to both male-
female (β ± SE = 2.239 ± 0.864, z = 2.591, p = 0.010) and male-male (β ± SE = 2.155 ±
0.917, z = 2.350, p = 0.019) conflict partners. There was no difference in reconciliation
frequency between male-male and male-female conflict partners (β ± SE = -0.085 ±
0.402, z = -0.211, p = 0.974). This model was, however, not ultimately selected because
it had a slightly higher AIC score (alternative model AIC = 516.4). All other variable
patterns in this alternative model remained the same.
Directionality. Directionality was the final variable included in the best fit model,
and while unidirectional conflicts were reconciled slightly more than bidirectional
conflicts, this predictor did not achieve statistical significance (β ± SE = 0.558 ± 0.332, z
= 1.680, p = 0.093).
Reconciliation Variation
There was a wide degree of variation in reconciliatory responses and behaviors. I
therefore was interested in determining what conflict, relationship, and applicable
reconciliatory precursors best predicted the timing, initiation, and behaviors involved
during reconciliation. Specifically, did reconciliatory behaviors vary according to

144
intrinsic opponent variables (i.e., sex, relationship, rank), conflict variables (i.e.,
directionality, severity, aggression type, etc.), or both? Wittig and Boesch (2005)
investigated some reconciliatory variation in the Taï chimpanzees, but did not report
beyond reconciliation latency, initiator, complexity, and duration. I expanded upon these
categories to include other factors/behaviors involved in reconciliation to understand how
the chimpanzees adapt their reconciliatory behavior during the act.
Predictors of Reconciliation Variation
Pre-reconciliatory signal. Pre-reconciliatory signal was a binomial variable (yes
or no) indicating whether or not a conflict partner used an affiliative vocalization (e.g.,
breathy pant, pant grunt, whimper/scream, etc.) or non-contact behavior (e.g.,
approach/crouch, extend hand, open arms, present, etc.) when in close proximity (< 1 m)
to their opponent immediately before (< 5 s) offering a reconciliatory gesture. I wanted to
know if certain factors determined whether or not chimpanzees used an affiliative signal
to “test the waters” prior to reconciliation. The best fit model (AIC = 143.6) indicated
that mutual PSP proximal association was the strongest predictor of a pre-reconciliatory
signal, followed by whether the aggressor chased the victim during the conflict, and
conflict type (single occurrence, renewed, or redirected conflict). Specifically, mutual
PSP proximal associates used pre-reconciliatory signals less than neutral associates (β ±
SE = -1.116 ± 0.498, z = -2.239, p = 0.025). Pre-reconciliatory signals occurred more
often following a conflict where the aggressor chased the victim as opposed to a conflict
that did not involve a chase (β ± SE = 1.059 ± 0.481, z = 2.204, p = 0.028). Pre-
reconciliatory signals were more commonly used following a renewed conflict as

145
opposed to a redirected conflict (β ± SE = 2.142 ± 1.058, z = 2.025, p = 0.043). There
was no difference between single occurrence conflicts and renewed or redirected
conflicts.
Reconciliation latency. Reconciliation latency was a continuous variable
indicating elapsed time (in seconds) between the end of the conflict and reconciliation.
More than 50% of reconciliations were initiated within the first minute (median 40 s) of
the post-conflict period (average 122.86 s, range 0 s to 8.33 min; Fig. 4.1). The best fit
model (AIC = 1549) resulted in only one significant predictor. Reconciliation latency
significantly decreased if aggressors chased their victims during the conflict as compared
to conflicts with no chase (β ± SE = -73.818 ± 34.795, t = -2.122, p = 0.036).
Approacher. Approacher was a binomial variable (aggressor or victim) measuring
which conflict partner first approached their former opponent and stopping within 1 m.
Aggressors approached their opponent after aggression less than victims (41.9% vs.
58.1%). The best fit model (AIC = 158) resulted in only one reliable predictor: victim’s
estrous state. Males were given a score of 0, indicating no estrous, and females’ estrous
was scored as a 1 – deflated swelling, 2 – partially swollen, or 3 – fully swollen (see
Chapter 2). The aggressor was significantly more likely to approach a fully swollen
female victim than a male (β ± SE = -1.938 ± 0.605, z = -3.203, p = 0.007) or a female
victim with a deflated swelling (β ± SE = -2.847 ± 0.730, z = -3.902, p < 0.001). There
was no significant pattern for partially swollen female victims (Figure 4.4).

146
f.V_Estrous
Approacher
0 1 2 3
A V
0.0 0.2 0.4 0.6 0.8 1.0
Initiator. Initiator was a binomial variable (aggressor or victim) measuring which
conflict partner first initiated reconciliatory contact with their former opponent.
Aggressors initiated slightly more reconciliation than victims (51.4% vs. 48.6%; H
2
P
1

Figure 4.4. Effect of victim’s sex and estrous state on likelihood of aggressor or
victim approach to reconcile. E1 indicates deflated swelling. E2 indicates partial
swelling. E3 indicates full swelling. Bar width visually represents victim’s estrous state
totals (N) for each condition.

unsupported). Not surprisingly, the best fit model (AIC = 129) indicated that the
approacher was the best predictor of the initiator followed by the victim’s estrous state
and then the conflict duration. Whomever chose to approach immediately before
reconciliation was also the individual that most likely initiated reconciliation (β ± SE =

1.0
0.8
0.6
0.4
0.2
0.0
Victim’s
Estrous
State
Male

E1

E2

E3

Female

Approacher
Aggressor

Victim
Rate of approach following conflict

147
Approacher
Initiator
A V
A V
0.0 0.2 0.4 0.6 0.8 1.0
-2.606 ± 0.600, z = 4.346, p < 0.0001; (Figure 4.5). As was true for the approacher, the
aggressor was more likely to initiate reconciliation with a fully swollen estrous female
victim than with a male victim (β ± SE = -3.077 ± 1.301, z = -2.364, p = 0.018). There
was no significant difference in initiator response between the other estrous states.
Aggressors were also more likely to initiate reconciliation after a short (5-9 s) conflict as
compared to a very long (> 20 s) conflict (β ± SE = 2.402 ± 0.916, z = 2.621, p = 0.043).
There was no significant difference in initiator response between the other conflict
duration categories.

Figure 4.5. Proportional relationship between the approacher and initiator of
reconciliation. Bar width visually represents approacher totals (N) for each condition.

1.0
0.8
0.6
0.4
0.2
0.0
Approacher
Aggressor

Victim
Initiator
Aggressor

Victim
Rate of initiation following conflict

148
Reconciliatory action. Reconciliatory action was the first affiliative contact
behavior used by the initiator to reconcile with their opponent. This variable was
classified as a binomial variable by grouping actions into two categories: prolonged-bout
behaviors (> 15 s duration, e.g., groom and contact sit) or short-momentary behaviors (<
15 s duration, e.g., embrace, mount, copulate, touch, etc.). The best fit model (AIC =
148.9) indicated that the best predictor of reconciliatory action type was conflict duration
(β ± SE = 1.465 ± 0.516, z = 2.839, p = 0.005). Short conflicts (5-9 s) were more likely to
result in prolonged-bout reconciliatory actions whereas very long conflicts (> 20 s) were
more likely to result in short-momentary reconciliatory actions (β ± SE = 2.148 ± 0.804, z
= 2.670, p = 0.037). There was no significant difference in reconciliatory action outcomes
between the other conflict duration categories. Other significant predictors included
whether the aggressor chased the victim during the conflict, the presence of prolonged
screaming by the victim after the conflict ended, and who initiated reconciliation. If
aggressors chased their victims during the conflict (β ± SE = 1.279 ± 0.533, z = 2.399, p
= 0.016) or victims gave prolonged screams after the conflict ended (β ± SE = 1.818 ±
0.767, z = 2.370, p = 0.019), opponents were more likely to engage in short-momentary
reconciliatory actions as compared to prolonged-bouts. . Aggressors that initiated the
reconciliation were more likely to engage in short-momentary reconciliatory actions as
opposed to victim initiators who chose to use prolonged-bout reconciliatory actions more
regularly (β ± SE = -1.101 ± 0.519, z = -2.122, p = 0.034).
Recipient response. Recipient response was a binomial variable (allow or
reciprocate) measuring how the recipient responded to the initiator’s reconciliatory action

149
immediately (within 3 s) after the reconciliatory action was offered. The recipient’s
response was classified as ‘reciprocate’ if the recipient countered with a reciprocal action
within 3 s of the initiator’s action, and as ‘allow’ if the recipient tolerated the initiator’s
action but did not respond with a reciprocal action within the 3-second time frame. The
best fit model (AIC = 79.8) indicated that the only predictor of recipient response was the
specific reconciliatory action. Touches, which included general body touches and
extended hand gestures, more often resulted in recipient reciprocation than did contact
sitting (β ± SE = 2.587 ± 1.258, z = 2.057, p = 0.04) or mounts/embraces (β ± SE = 2.085
± 0.801, z = 2.603, p = 0.033).
Reconciliation directionality. Reconciliation directionality was a binominal
variable measuring whether both opponents (bidirectional) actively exchanged affiliative
behavior during the reconciliatory bout or if only one opponent acted affiliatively
(unidirectional) at any point during the reconciliatory bout. Most reconciliatory bouts
were unidirectional (79.1%) versus bidirectional (20.9%). The best fit model (AIC =
106.9) indicated that the best predictor of directionality was bystander count, followed by
victim’s estrous state, reconciliatory action, and reconciliation latency. As the number of
bystanders increased, the probability of unidirectional reconciliation also increased (β ±
SE = 0.586 ± 0.16, z = 3.668, p = 0.0002). Male (β ± SE = 4.770 ± 1.561, z = 3.056, p =
0.011), and deflated swelling female victims (β ± SE = 4.698 ± 1.554, z = 3.023, p =
0.012) were more likely to engage in bidirectional reconciliation as compared to fully
swollen female victims. Reconciliatory actions involving touches more likely resulted in
bidirectional reconciliation as compared to groom (β ± SE = -2.597 ± 0.793, z = -3.274, p

150
= 0.005), contact sit (β ± SE = -3.893 ± 1.374, z = -2.834, p = 0.022), or embrace/mount
(β ± SE = -2.978 ± 1.056, z = -2.821, p = 0.023). Reconciliation occurring after a very
short latency (< 5 s) following the original conflict more commonly resulted in
bidirectional reconciliatory actions as compared to short (5-9 s) latencies (β ± SE = 2.648
± 0.839, z = 3.158, p = 0.008).
Reconciliation complexity. Reconciliation complexity was an ordinal variable
(ranging from 1 to 4) measuring the sum of different affiliative behaviors used by both
partners during the reconciliatory bout (Wittig & Boesch, 2005). The best fit model (AIC
= 190) indicated that recipient response best predicted reconciliation complexity followed
by conflict duration, and victim’s decision to flee at the end of the conflict. Recipients of
reconciliatory actions that responded with grooming and soft bites more commonly
resulted in more complex reconciliatory bouts as compared to recipient’s that responded
with touches (groom vs. touch: β ± SE = -0.723 ± 0.281, z = -2.578, p = 0.043; soft bite
vs. touch: β ± SE = -0.877 ± 0.336, z = -2.612, p = 0.039) or allowance (no active
reciprocation) (groom vs. allowance: β ± SE = 0.658 ± 0.253, z = 2.602, p = 0.04; soft
bite vs. allowance: β ± SE = 0.812 ± 0.323, z = 2.518, p = 0.049). There was a significant
linear relationship between reconciliation complexity and conflict duration (β ± SE =
-0.194 ± 0.080, t = -2.425, p = 0.017) with short (5-9 s) conflicts averaging more
complex reconciliatory bouts as compared to very long (> 20 s) conflicts (β ± SE =
-0.388 ± 0.132, z = -2.931, p = 0.018). Victims that fled from their aggressor after a
conflict averaged more complex reconciliatory bouts than victims that did not flee (β ±
SE = 0.183 ± 0.078, t = 2.348, p = 0.021).

151
Reconciliation duration. Reconciliation duration was a continuous variable
measuring the amount of time (in seconds) opponents engaged in an uninterrupted
reconciliatory bout. Half of the reconciliation bouts last less than 15 s in duration, and
89% of reconciliatory bouts lasted less than 4 min in duration (median 87 s, average 137
s, range 1 s to 18 min; Fig. 4.6). Only 1.6% of reconciliation bouts lasted beyond 10 min.
Figure 4.6. Percent of reconciled dyads across duration of reconciliation.

The best fit model (AIC = 1486) indicated that the best predictor of reconciliation
duration was reconciliation directionality followed by recipient response, reconciliatory
action type, aggression type, and mutual contact PSP. Bidirectional reconciliatory actions
resulted in longer reconciliatory bouts than did unidirectional reconciliatory actions (β ±
SE = -231.09 ± 41.12, t = -5.620, p < 0.0001). Immediately reciprocated reconciliation
lasted for shorter periods of time than non-reciprocated (β ± SE = -199.12 ± 56.89, t =
-3.500, p < 0.001). Opponents using embraces/mounts and touches reconciled for shorter
periods of time than did partners using grooming (embrace/mount vs. groom: β ± SE =
123.219 ± 29.175, z = 4.223, p < 0.001; touch vs. groom: β ± SE = -128.824 ± 37.797, z

152
= -3.408, p = 0.004) and contact sitting (embrace/mount vs. contact sit: β ± SE =
-133.225 ± 45.771, z = -2.911, p = 0.018; touch vs. contact sit: β ± SE = -138.831 ±
50.644, z = -2.741, p = 0.03). Conflicts involving contact aggression resulted in shorter
reconciliatory bouts than did conflicts involving only threats (β ± SE = 59.53 ± 24.03, t =
2.477, p = 0.015). Contact affiliation PSP that shared a mutual preferred social
partnership reconciled for significantly longer than did neutral partners (β ± SE = 51.11 ±
24.27, t = 2.106, p = 0.037).
Discussion
The Kanyawara chimpanzees showed clear evidence of reconciliatory behavior
with notable variation throughout the study period. Given that this is the first time
reconciliation was quantified at this long-term field site, this study provides new
information on this community’s reconciliatory tendencies, conflict resolution decision-
making patterns, and associated predictors. These wild conciliatory (peacemaking) data
are valuable for a number of reasons. First, chimpanzee post-conflict behavior has been
extensively studied in captive populations (Cords & Aureli, 2000; de Waal & van
Roosmalen, 1979, Fraser & Aureli, 2008; Koski et al., 2007; Preuschoft et al., 2002), but
field data are limited to only five sites (Budongo: Arnold & Whiten, 2001; Mahale:
Kutsukake & Castles, 2004; Ngogo: Watts, 2006; Tai: Wittig & Boesch, 2003a, 2005),
including the present study. Second, previous studies have argued that captive post-
conflict data can easily be extrapolated to explain wild populations and their patterns of
behavior (Colmenares, 2006); however, the results of the present study suggest otherwise.

153
The Kanyawara chimpanzees had a reconciliatory CCT (15.8%) similar to other
wild populations [Budongo CCT 19%, Mahale CCT 15%, Tai CCT 16% (Arnold &
Whiten, 2001; Kutsukake & Castles, 2004; Wittig & Boesch, 2003a)], but notably lower
than their captive counterparts (ranging 22-48%, de Waal & van Roosmalen, 1979;
Casperd, 1997; Fraser & Aureli, 2008; Fuentes et al., 2002; Preuschoft et al., 2002),
lending support to H
1
. This suggests that while the sheer occurrence of reconciliation is
likely not a by-product of captivity, the rate at which it occurs might be. Captive residents
may, by enclosure design, already be at an increased risk of renewed aggression because
they cannot physically distance themselves from their opponent, and instead elect to
reconcile to reduce the probability of renewed aggression. Alternatively, the fission-
fusion social structure of wild chimpanzees allows for greater distancing/dispersion
options following a conflict, i.e., opponents can distance themselves while remaining in
the same party, completely fission from the party, or anything in between. The present
study shows that these distancing options may have affected the Kanyawara
chimpanzees’ tendencies to reconcile in the 10-min period immediately following a
conflict.
Predictors and Function of Reconciliation
The present study found three definitive predictors of reconciliation: opponent
distance, relationship value, and aggressor sex. This study, however, did not find support
for several conflict parameters reported in other PC studies. For example, conflict
intensity (Kutsukake & Castles, 2004; Wittig & Boesch, 2003a), conflict context (Wittig
& Boesch, 2003a) and dominance rank [difference] (Arnold & Aureli, 2007) had no

154
bearing on the Kanyawara chimpanzees’ reconciliation, lending no support for H
2
P
2
, H
3
,
or H
4
.
Opponent distance. The Kanyawara chimpanzees post-conflict decisions to
reconcile or not were best predicted by opponent distance immediately following the
conflict after the victim had stopped fleeing or the aggressor had stopped chasing. In
support of H
5
, opponents within 5 m were the most likely to reconcile, followed by
opponents within 5 to 10 m, and lastly opponents beyond 10 m. A simple explanation
might be when opponents are in close proximity 1) energy expenditure to reconcile is low
and 2) visibility of opponents is high. In many parts of the forest, visibility of opponents,
and other party members, does not extend beyond 10 m unless on a cut trail. Simply put,
opponents in close-proximity immediately following the fight were more likely to
reconcile as a result of proximal convenience.
While these results may not seem surprising, to date, no other studies have shown
the importance of interopponent distance regulation in relation to chimpanzee
reconciliatory decisions. Arnold and Whiten (2001) found a positive relationship between
conflict intensity and opponent distance (i.e., distance between opponents was shorter
following low intensity conflict), but this resulted in only a trend for low-intensity
conflicts to be reconciled more often than high-intensity. Kutsukake and Castles (2004)
failed to find a significant relationship between opponent distance and reconciliation, but
they only measured opponents to be within 5 m or beyond 5 m. This proximity category
may have been too general, or reconciliatory patterns may truly not be influenced by
opponent proximity at Mahale. [In fact, compared to other wild chimpanzee PC studies

155
(Arnold & Whiten, 2001; Wittig & Boesch, 2005), reconciliatory patterns at Mahale were
uncharacteristic, i.e., sex combination and affiliative association did not influence
reconciliation.] However, this study suggests that proximity not only plays an important
and pivotal role in the conflict resolution strategies in wild chimpanzees, but also, as
indicated by the low CCT values (15-19%), the preferred choice among wild
chimpanzees was to avoid their opponent by increasing interopponent distance and
ultimately not reconcile. Since approaching your opponent after a fight is risky and
increases your probability of renewed aggression (Aureli & van Schaik, 1991; Cords,
1992), the “safer” option in the wild was to distance themselves (using the many
available distancing options) from their opponent to avoid receiving more aggression in
the immediate future. Data from the present study supported this pattern. While
opponents within 10 m had a higher risk of renewed aggression, this risk greatly
diminished if/after opponents reconciled.
This finding serves as a more proximate explanation of reconciliatory behavior,
whereby the opponents are reconciling to gain short-term benefits (i.e., access to food
resources or preferred partners) or reduce the uncertainty of the situation (i.e., reduce
probability of renewed aggression) and signal that the conflict is over (e.g., probability of
renewed aggression decreases following reconciliation). This finding also provides some
support for the Benign Intent Hypothesis (Silk 1996, 2000, see Chapter 1). While this
hypothesis was proposed as an alternative to the Valuable Relationships Hypothesis (see
Chapter 1), these hypotheses are not necessarily mutually exclusive. Instead, they offer
different perspectives, which address the proximate and ultimate mechanisms that may

156
both motivate reconciliatory behavior (Cords & Aureli, 1996). Clearly, there are
important immediate (e.g., physical health) and long-term (e.g., social relationships)
benefits to successful reconciliation between opponents.
Only recently have studies even started focusing on the importance of distancing
strategies in post-conflict (PC) behavior by qualifying close proximity as an affiliative
reconciliatory behavior (Patzelt et al., 2009; McFarland & Majolo, 2011) or using
proximity to predict conciliatiory behavior (Clay & de Waal, 2013a). These recent studies
on interopponent distance regulation have provided new insights in the study of conflict
resolution and further support the present study’s results. Future captive and wild PC
studies should therefore incorporate opponent distance into their study design and
analyses.
Preferred social partners. Not all opponents in close proximity reconciled and
not all distal opponents avoided reconciliation; the best fit model revealed several other
factors that predicted the occurrence of reconciliation. Opponents were significantly more
likely to reconcile if the aggressor was a moderate or a high overall PSP of the victim
(one-sided relationship). The data supports both H
6
P
1
and H
6
P
2
. This overall PSP
relationship parameter was chosen for the best fit model because it most accurately
(lowest AIC) predicted opponent reconciliation. However, all relationship param
(proximal association, contact affiliation, and agonistic support) showed a significant and
identical pattern of reconciliatory behavior, with both one-sided and mutual prefered
social partnerships between opponents often resulting in reconciliation. This indicates

157
that support for H
6
P
3
is mixed. While one-sided relationships in favor of the victim best
predicted reconciliation, mutual relationships were still a significant predictor.
This study’s effects of preferred social partnership on reconciliation are consistent
with other PC studies on chimpanzees (Arnold & Whiten, 2001; de Waal & Van
Roosmalen, 1979; Fraser et al., 2008a; Watts, 2006; Wittig & Boesch 2003a, 2003b,
2005) and other great apes (Clay & de Waal, 2013a; Palagi et al., 2004; Watts, 1995), and
are congruent with the Valuable Relationships Hypothesis (VRH; Cords & Aureli, 1996;
de Waal & Aureli, 1997). Continued associations with PSPs may confer considerable
benefits to both parties (i.e., agonistic support, food sharing, resource/mate defense),
hence repairing these valued relationships following a conflict may be in the best interest
of one or both opponents depending on the nature and strength of their [one-sided or
mutual] relationship (Aureli & de Waal, 2000; Watts, 2006). The consistent results of the
PSP param indicate the importance of social relationships in chimpanzee PC decision-
making, and also suggest that the VRH maybe be a more appropriate ultimate explanation
for the evolution of reconciliatory behavior in great apes (Cords & Aureli, 2000). In the
context of social living animals, individuals should focus on maintaining friendships and
not making enemies. While reconciliation may provide the obvious benefit of relationship
repair amongst valuable partners, it may also prevent low benefit partners from becoming
enemies (Wittig, 2004).
Aggressor sex. Male aggressors reconciled with their opponents (i.e. actively
initiated and received more reconciliatory behaviors) significantly more than female
aggressors. When incorporating dyadic composition of opponents into the model, it

158
lended the same results as the best fit model, with male-male and male-female dyads both
reconciling significantly more than female-female dyads. While male-male dyads did
reconcile more than male-female, this result was not significant (partially supports H
1
P
1
).
Considering the patriarchal nature of wild chimpanzee society, the well-known
gregariousness of males (Mitani et al., 2000; Pepper et al., 1999; Watts, 2002, 2006;
Wrangham, 2000), and the closer dyadic party/proximal association among males (Gilby
& Wrangham, 2008), this result is not surprising and coincides with other captive and
wild chimpanzee PC studies (Arnold & Whiten, 2001; de Waal, 1986; Koski et al.,
2007b). While Wittig & Boesch (2005) found that Tai chimpanzees’ male-female dyads
reconciled the most, this result is still consistent with the present study’s findings that
male aggressors predict higher reconciliatory tendencies. Moreover, during the study
period at Tai, it should be noted that there were only three adult males (one of which
died/disappeared during study period) and 11 adult females. This disproportionate sex
ratio might have influenced their dyadic sex composition results (Wittig & Boesch,
2005).
Other chimpanzee PC studies have found opposing results to the present study
with female-female dyads in captivity reconciling more than male-male dyads (Baker &
Smuts, 1994; Fraser & Aureli, 2008; Fuentes et al., 2002; Preuschoft et al., 2002), or no
effect of sex or dyad composition on reconciliatory behavior in the wild (Kutsukake &
Castles, 2001). With the high variability in dyadic sex combination results, this suggests
sex combination may not be a defining factor in determining reconciliatory variation, but
instead might vary according to group/community history, overall adult sex composition,

159
environment (i.e., captive or wild), and/or socio-ecology. In fact, female chimpanzees are
reportedly more gregarious in West Africa as compared to their East African counterparts
(Boesch & Boesch-Achermann, 2000; Lehmann & Boesch, 2005), which also might
explain some of the variation in reconciliatory sex composition patterns in wild
chimpanzees. It is still unclear whether this may be a cultural or subspecies difference or
related to community composition.
Predictors and Function of Reconciliatory Variation
Not only is there a high degree of interspecific variation in reconciliatory
behavior, there is also a high degree of intraspecific variation in both captive and wild
chimpanzee reconciliatory behavior. Even within a single population, the Kanyawara
chimpanzees exhibited a wide degree of variation in behaviors and responses during the
reconciliatory bout. However, wild chimpanzees are a good model to study intraspecific
variation because they 1) naturally use reconciliation (Arnold & Whiten, 2001;
Kutsukake & Castles, 2003; Wittig & Boesch, 2003a) and 2) optimize their conflict
response based on cost and benefit strategies (Wittig & Boesch, 2003b). When the
chimpanzees chose to reconcile, the way in which they engaged in reconciliation was
dependent on a variety of factors ranging from conflict variables, relationship with the
opponent, and variation in reconciliation precursors that seemed to maximize their cost-
benefit ratio. Since there are many variables that constitute a reconciliatory bout, some of
these behaviors acted as precursors to subsequent behaviors, thereby influencing the
reconciliatory bout as it progressed.

160
The Kanyawara chimpanzees seemed to evaluate the costs and benefits of
reconciliation, which influenced the way in which they reconciled. The cost of
reconciling was the risk of renewed aggression. The benefit of reconciling seemed to be
indicated by the relationship variables, which measured the value of each opponent’s
relationship to one another in terms of general proximity, affiliation, and agonistic
support. When the cost of reconciling was high and the benefit was low, the chimpanzees
tended to make reconciliation brief, one-sided, and less personal. Whereas when the cost
of reconciling was low and the benefit was high, the chimpanzees tended to prolong
reconciliation by engaging in more personal, mutual interactions.
Pre-reconciliatory signal. Arnold and Whiten (2001) found that Budongo
chimpanzees sometimes used pre-reconciliatory signals, and that these signals resulted in
successful reconciliation 18% of the time. While the present study did not find pre-
reconciliatory signals to be an overall predictor of reconciliation, the chimpanzees
decision to use them was dependent on several relationship and conflict parameters.
Opponents were more likely to use pre-reconciliatory signals when they were neutral
proximal associates, chased during the conflict, and victims of renewed aggression.
Opponents that were mutual PSP proximal associates, not chased, and victims of
redirected aggression were less likely to use pre-reconciliatory signals. This result
indicates that high risk situations and uncertain, less valuable relationships warrant
opponents to first “test the waters” and signal their intentions before committing to
reconcile with their opponent.

161
Reconciliation latency. Even though tension is known to decrease over time
(Aureli & van Schaik, 1991), like most other post-conflict studies (Aureli & de Waal,
2000), the majority of reconciliation occurred relatively quickly within the post-conflict
period. Unlike Wittig and Boesch (2005), this study found that reconciliation latency was
only predicted by whether or not the aggressor chased the victim during the conflict.
Wittig and Boesch (2005) found that reconciliation latency was predicted by the victim’s
sex, opponent rank difference, and conflict duration. However, it should be noted that
Wittig and Boesch did not analyze the act of chasing as a predictor. The act of chasing
may be a sensitive behavior that encourages opponents to take increased pause before
deciding to reconcile, possibly out of fear that they may be chased again and actually
caught if the opponent chooses to renew aggression instead of reconcile (partially
supports for H
7
P
1
).
Approacher and initiator. At Taï, Wittig and Boesch (2005) found victims
initiated more reconciliation than aggressors. While I found a similar pattern in
approachers (i.e., victims were more likely to approach their opponent than aggressors),
both aggressors and victims initiated reconciliation equally (H
2
P
1
unsupported).
Specifically, aggressors were more likely to initiate reconciliation when the victims were
fully swollen estrous females or males as compared to fully deflated females. Victims
were more likely to initiate reconciliation when the aggressors were fully deflated
females as compared to fully swollen estrous females or males. Since 86% of aggressors
were male, this explains why aggressors preferentially initiated reconciliation with both
male and fully swollen estrous female victims. Males form long-term equitable social

162
bonds with one another and these male-male bonds are stronger and more valuable than
all male-female bonds (Nishida, 1983; Gilby & Wrangham, 2008; Mitani, 2009).
However, males also form short-term valuable social bonds with females when they are
fully swollen and in estrous (Matsumoto-Oda, 2002; Pieta, 2008; Stumpf & Boesch,
2005). Both partners, under this circumstance, present high reconciliatory benefit. For
example, copulatory reconciliation benefits males by providing mating opportunities and
females by reducing their probability of renewed/continued aggression. Alternatively,
since 64% of victims were female (and 55% of female victims had fully deflated
swellings), it is not surprising that these victims preferentially initiated reconciliation
with fully deflated female aggressors. Given that their social bonds with males and fully
deflated estrous females are weaker, these fully deflated female victims might have
initiated reconciliation to strengthen their bonds with the aggressors and avoid further
aggression.
Aggressors were also more likely to initiate reconciliation following short (5-9 s)
versus very long (> 20 s) conflicts. Since the risk of further aggression is known to
increase with conflict duration (Wittig & Boesch, 2003b), conflicts of relatively short
duration present a low probability that aggression will reignite between former
opponents, thus reducing the overall cost of reconciling.
Reconciliatory action. While the Kanyawara chimpanzees used a variety of
affiliative behaviors (i.e., groom, contact sit, embrace, mount, copulate, touch, etc.)
during reconciliatory bouts, their initial reconciliatory behaviors fell into two distinct
categories based on their average baseline duration: prolonged-bout behaviors (i.e.,

163
groom and contact sit) and short-momentary behaviors (i.e., embrace, mount, touch, etc.).
While prolonged-bout behaviors are potentially riskier and costlier than short-momentary
behaviors because they solicit extended contact between opponents, they also can offer
great benefits by better repairing the relationship with continuous affiliative contact.
Short conflicts that did not involve chases or prolonged screaming afterwards were more
likely to result in prolonged-bout behaviors that were initiated by the victim. These types
of conflicts showed patterns of low distress to the victim and as a result were perceived as
being less costly or lower in risk for the victim to then initiate reconciliation. When the
risk/cost of engaging in these behaviors is low, there is no reason not to engage in
affiliative contact for an extended period of time. Alternatively, very long conflicts that
involved chases and/or prolonged screaming afterwards were more likely to result in
short-momentary behaviors that were initiated by the aggressor. These types of conflicts
showed patterns of high distress to the victim and as a result were perceived as being
more costly or higher in risk (partial support for H
7
). While short-momentary
reconciliatory behaviors likely provide an immediate signal that the conflict is over, they
do not require a huge time investment, which allows both opponents to quickly separate
and return to their previous or subsequent activity.
Recipient response. Most recipients of reconciliation did not immediately (< 3 s)
reciprocate with an affiliative behavior (15%), but instead allowed their opponent to
contact them (85%). However, when recipients chose to immediately reciprocate, the
type of reconciliatory action that their opponent used first generally predicted this.
Touches, which included extended hand gestures, were reciprocated much more often

164
than mounts/embraces or contact sitting. Touching and extended hand gestures seemed to
mirror the human handshake in terms of form and function. Most often when one
chimpanzee extended their hand out towards their opponent, that opponent immediately
responded by briefly touching fingertips with the initiator. The same was true for general
touches. Most often when one chimpanzee briefly touched their opponent on the shoulder
or head, that opponent would immediately respond in a similar fashion. Touching and
extended hand gestures seemed to inherently elicit the immediate response of their
opponent for the affiliative action to be completed. Like the human handshake, these
affiliative interactions were brief and mutual. Contact sitting and mount/embrace
behaviors, however, do not implicitly require a mutual action from the recipient in order
to be completed. While they can incorporate a bidirectional action, they only require one
individual to act and the other individual to tolerate.
Reconciliation directionality. Most opponents only engaged in unidirectional
reconciliation (75%) likely because bidirectional reconciliation requires mutual effort
from both opponents, which can be viewed as time consuming and energetically
expensive. However, when opponents did exchange mutual affiliation (25%), this was
predicted by having fewer bystanders nearby, male victims or fully deflated females, and
touching reconciliatory actions following a very short (< 5 s) reconciliation latency. If
there are fewer bystanders nearby during a fight, this implies that there are fewer
alternative social partners available. Since social partner choices were limited, opponents
may have “settled” by mutually interacting with one another as opposed to finding a new
social partner. The decision to engage in bidirectional or mutual reconciliation may have

165
also been influenced by the sex, social status, and social relationship of the bystanders
relative to the opponents. However this was not tested and could be investigated in future
studies.
The victim’s sex and estrous state also influenced reconciliatory directionality for
several reasons (partially supports H
8
P
1
). First, males are more gregarious than females
and develop long-term social bonds (Gilby & Wrangham, 2008), thus it was not
surprising that male victims were more likely to engage in mutual reconciliation with
aggressors (who were predominantly male). Fully deflated females were also more likely
to engage in mutual reconciliation, possibly to appease their aggressors and avoid future
aggression by reciprocating. Fully swollen estrous female victims, however, more
commonly engaged in unidirectional reconciliation. Given that fully swollen estrous
females are highly sought after by males (Watts, 1998; Wrangham, 2002), they can be
viewed as a limited commodity that is of high potential benefit. Thus, the estrous females
did not need to engage in mutual reconciliation with their aggressors. Instead, they were
“exempt” from typical female etiquette because male aggressors were more interested in
mating than in how the females responded to the conflict. Bidirectionality in these
situations was a clear way for both opponents to acknowledge that the conflict was over
and their relationship had been repaired.
Very short latencies (< 5 s) also predicted bidirectionality in reconciliation. This
result is interesting because tension is known to decrease over time (Aureli & van Schaik,
1991), so why not wait to reduce risk? This short time window does not allow for much
other than immediately approaching to reconcile. This short reconciliatory latency

166
suggests that opponents mutually “planned” on reconciling immediately after the conflict
ended. In short, one opponent was not avoiding or indifferent to the other. If both
opponents did have intentions of reconciling, then it is logical to expect that the latency to
reconcile would be short and both opponents would mutually interact during the
reconciliatory bout.
Reconciliation complexity. The Kanyawara chimpanzees mostly used simple
(80%) forms of reconciling as compared to advanced (18%) or complex (2%) forms.
While Wittig and Boesch (2005) found that aggressors initiated more simple
reconciliation and victims initiated more advanced/complex reconciliation, the
Kanyawara chimpanzees did not show this pattern. Instead, increased complexity
occurred more commonly when 1) recipients responded with grooming and soft bites as
opposed to touches or allowances, 2) conflict duration was short as opposed to very long,
and 3) victims fled from their aggressors during the conflict. In these situations,
complexity appeared to increase when recipients responded with a more “intimate”
behavior of high social benefit and the uncertainty or risk was low.
Reconciliation duration. Similar to Wittig and Boesch (2005), half of
reconciliatory bouts were shorter than 15 s in duration. Reconciliatory duration increased
when reconciliation was 1) bidirectional but slightly delayed, 2) the first reconciliatory
action was a prolonged-bout behavior versus a short-momentary behavior, 3) opponents
were mutual contact affiliation PSP versus neutral partners (partially supports H
8
P
2
), and
4) aggression only involved threats with no agonistic contact. Under these circumstances,
increased duration was dependent on prolonged mutual contact familiarity (high social

167
benefit) after low intensity aggression (low cost/risk). Alternatively, the Taï
chimpanzees’ reconciliatory duration was influenced by conflict context, opponent rank
difference, and victim’s sex (Wittig & Boesch, 2005).
Overall Patterns of Reconciliatory Behavior in Wild Chimpanzees
It is difficult to address overall patterns in wild chimpanzee reconciliatory
behavior because of the wide degree of variation between study methodologies, predictor
variables tested, and chimpanzee community socio-ecological differences. Despite this
variation, however, there are some common reconciliatory themes. Most studies,
including both wild and captive, have found that preferred social partners play a vital role
in conflict resolution and individuals view these partners as more valuable both inside
and outside of conflicts. In particular, Kanyawara chimpanzees put an even bigger
emphasis on mutual versus one-sided preferred social partners, with one-sided
relationships better predicting the occurrence of reconciliation but mutual relationships
better predicting the variation within reconciliation, a detail that other studies have not
investigated. Males in general are also more actively involved in reconciliatory strategies,
which is consistent with chimpanzee social behavior. However, this study’s major finding
that immediate opponent proximity following a conflict predicts reconciliation was either
not explored or found (Kutsukake & Castles, 2004) in any prior wild chimpanzee post-
conflict study, and thus we cannot know if other chimpanzee communities’ reconciliatory
strategies are also influenced by opponent proximity. Similarly, this study tested several
other conflict and relationship parameters that other studies overlooked, making it
impossible to make direct comparisons between sites and determine what consistencies

168
there are between and across wild chimpanzee reconciliatory behavior. Future studies
should address this issue and develop a multi-faceted approach with more comprehensive
and consistent methodology for testing the vast array of parameters that might predict or
influence reconciliation.
Conclusion
Following a conflict, the Kanyawara chimpanzees seem to have clear-cut
strategies for evaluating the post-conflict period and whether or not they should
reconcile. They quickly assessed the costs and benefits of the situation and made
informed decisions based on that analysis. Since each opponent had the option to vary
their response, certain conflict and relationship variables were pivotal and may have
largely influenced each opponent’s decision in similar or even opposite ways. In many
ways, each partner’s decision generated a formal ballroom dance. While there are general
“rules,” their movements and decisions were full of negotiations and must have been
quick responses to their partner’s next step.

169
Bibliography

Akaike, H. (1973). Maximum likelihood identification of Gaussian autoregressive
moving average models. Biometrika, 60(2), 255-265.

Altmann, J. (1974). Observational study of behavior sampling methods. Behaviour, 49,
227-265.

Amici, F., Aureli, F., & Call, J. (2008). Fission-fusion dynamics, behavioral flexibility,
and inhibitory control in primates. Current Biology, 18(18), 1415-1419.

Anderson, D. P., Nordheim, E. V., Boesch, C., & Moermod, T. C. (2002). Factors
influencing fission-fusion grouping in chimpanzees in the Tai National Park, Côte
d’Ivoire. In C. Boesch, G. Hohmann, & L. F. Marchant (Eds.), Behavioural
diversity in chimpanzees and bonobos (pp. 90-101). Cambridge: Cambridge
University Press.

Arnold, K., & Aureli, F. (2007). Postconflict reconciliation. In C. J. Campbell, A.
Fuentes, K. C. Mackinnon, M. Panger, & S. K. Bearder (Eds.), Primates in
perspective (pp. 592-608). Oxford: Oxford University Press.

Arnold, K., & Barton, R. (2001). Postconflict behavior of spectacled leaf monkeys
(Trachypithecus obscurus). I. Reconciliation. International Journal of
Primatology, 22, 243-266.

Arnold, K., & Whiten, A. (2001). Post-conflict behaviour of wild chimpanzees (Pan
troglodytes schweinfurthii) in the Budongo forest, Uganda. Behaviour, 138, 649-
690.

Aureli, F. (1992). Post-conflict behavior among wild long-tailed macaques (Macaca
fascicularis). Behavioral Ecology and Sociobiology, 31, 329-337.

Aureli, F. (1997). Post-conflict anxiety in nonhuman primates: the mediating role of
emotion in conflict resolution. Aggressive Behavior, 23, 315-328.

Aureli, F., Cords, M., & van Schaik, C. P. (2002). Conflict resolution following
aggression in gregarious animals: a predictive framework. Animal Behaviour, 64,
325-343.

Aureli, F., & de Waal, F. B. M. (1997). Inhibition of social behavior in chimpanzees
under high-density conditions. American Journal of Primatology, 41, 213-228.

Aureli, F., & de Waal, F. B. M. (2000). Natural conflict resolution. Berkley, CA:
University of California Press.

170

Aureli, F., Preston, S. D., & de Waal, F. B. M. (1999). Heart rate responses to social
interactions in free-moving rhesus macaques (Macaca mulatta): A pilot study.
Journal of Comparative Psychology, 113, 59-65.

Aureli, F., & Schaffner, C. M. (2007). Aggression and conflict management at fusion in
spider monkeys. Biology Letters, 3, 147-149.

Aureli, F., & Smucny, D. (2000). The role of emotion in conflict and conflict resolution.
In F. Aureli & F. B. M. de Waal (Eds.), Natural conflict resolution (pp. 199-224).
Berkeley, CA: California University Press.

Aureli, F., & van Schaik, C. P. (1991). Post-conflict behaviour in long-tailed macaques
(Macaca fascicularis): II. Coping with uncertainty. Ethology, 89, 101-114.

Aureli, F., van Schaik, C. P., & van Hooff, J. A. R. A. M. (1989). Functional aspects of
reconciliation among captive long-tailed macaques (Macaca fascicularis).
American Journal of Primatology, 19, 39-51.

Aureli, F., Veenema, H. C., van Eck, C. J. V., & van Hooff, J. A. R. A. M. (1993).
Reconciliation, consolation, and redirection in Japanese macaques (Macaca
fuscata). Behaviour, 124, 1-21.

Baayen, R. H. (2011). Language R: Data sets and functions with "Analyzing linguistic
data: A practical introduction to statistics". R package version 1.4.
http://CRAN.R-project.org/package=languageR

Baker, K. C., & Aureli, F. (1997). Behavioural indicators of anxiety: An empirical test in
chimpanzees. Behaviour, 134, 1031-1050.

Baker, K. C., & Smuts, B. B. (1994). Social relationships of female chimpanzees:
diversity between captive social groups. In R. W. Wrangham, W. C. McGrew, F.
B. M. de Waal, & P. G. Heltne (Eds.), Chimpanzee cultures (pp. 227-242).
Cambridge, MA: Harvard University Press.

Bates, L. A., Lee, P. C., Njiraini, N., Poole, J. H., Sayialel, K., Sayialel, S., Moss, C. J., &
Byrne, R. W. (2008). Do elephants show empathy? Journal of Consciousness
Studies, 15, 204–225.

Bates, D., Maechler, M., & Bolker, B. (2013). lme4: Linear mixed-effects models using
S4 classes. R package version 0.999999-2. http://CRAN.R-
project.org/package=lme4.

171
Boehm, C. (1999). Hierarchy in the forest: The evolution of egalitarian behavior.
Cambridge, MA: Harvard University Press.

Boesch, C. (1996). Social grouping in Tai chimpanzees. In W.C. McGrew, L.F.
Marchant, & T. Nishida (Eds.), Great ape societies (pp. 101-113). Cambridge:
Cambridge University Press.

Boesch, C., & Boesch, H. (1989). Hunting behavior of wild chimpanzees in the Tai
National Park. American Journal of Physical Anthropology, 78, 547-573.

Boesch, C., & Boesch-Achermann, H. (2000). The chimpanzees of the Tai Forest:
Behavioural ecology and evolution. Oxford: Oxford University Press.

Boesch, C., Hohmann, G., & Merchant, L. F. (2002). Behavioural diversity in
chimpanzees and bonobos. Cambridge, UK: Cambridge University Press.

Boesch, C., Kohou, G., Néné, H., & Vigilant, L. (2006). Male competition and paternity
in wild chimpanzees of the Taï forest. American Journal of Physical
Anthropology, 130(1), 103-115.

Burbank, V. K. (1992). Sex, gender, and difference: Dimensions of aggression in an
Australian Aboriginal community. Human Nature, 238, 251-277.

Butovskaya, M., Verbeek, P., Ljungberg, T., & Lunardini, A. (2000). A multicultural
view of peacemaking among young children. In F. Aureli & F. B. M. de Waal
(Eds.), Natural conflict resolution (pp. 243-260). Los Angeles, CA: University of
California Press.

Bygott, J. D. (1979). Agonistic behaviour, dominance and social structure in wild
chimpanzees of the Gombe National Park. In D. Hamburg & E. R. McCown
(Eds.), The great apes (pp. 404-427), Menlo Park, CA: Benjamin Cummings.

Cairnes, S. J., & Schwager, S. J. (1987). A comparison of association indices. Animal
Behaviour, 35, 1454-1469.

Call, J., Aureli, F., & de Waal, F. B. M. (1999). Reconciliation patterns among
stumptailed macaques: A multivariate approach. Animal Behaviour, 58, 165-172.

Call, J., Aureli, F., & de Waal, F. B. M. (2002). Postconflict third party affiliation in
stumptailed macaques. Animal Behaviour, 63, 209-216.

Casperd, J. M. (1997). The evolution of reconciliation within the primate order.
Unpublished doctoral dissertation, University of Liverpool, UK.

172
Castles, D. L., Aureli, F., & de Waal, F. B. M. (1996). Variation in conciliatory tendency
and relationship quality across groups of pigtail macaques. Animal Behaviour, 52,
389-403.

Castles, D. L., & Whiten, A. (1998). Post-conflict behaviour of wild olive baboons II.
Stress and self-directed behaviour. Ethology, 104, 148-160.

Castles, D. L., Whiten, A., & Aureli, F. (1999). Social anxiety, relationships and self-
directed behaviour among wild female olive baboons. Animal Behaviour, 58,
1207-1215.

Chapman, C. A., & Chapman, L. J. (1997). Forest regeneration in logged and unlogged
forests of Kibale National Park, Uganda. Biotropica, 29, 396-412.

Chapman, C. A., & Chapman, L. J. (2000). Determinants of group size in primates: The
importance of travel costs. In S. Boinski & P. A. Garber (Eds.), On the move:
How and why animals travel in groups (pp. 24-41). Chicago: University of
Chicago Press.

Chapman, L. J ., Chapman, C. A., Brazeau, D., McGlaughlin, B., & Jordan, M. (1999).
Papyrus swamps and faunal diversification: Geographical variation among
populations of the African cyprinid Barbusn eumayeri. Journal of Fish Biology,
54, 310-327.

Chapman, C. A., Chapman, L. J., & Gillespie, T. R. (2002). Scale issues in the study of
primate foraging: Red colobus of Kibale National Park. American Journal of
Physical Anthropology, 117, 349-363.

Chapman, C. A., Chapman, L. J., Struhsaker, T. T., Zanne, A. E., Clark, C. J., & Poulsen,
J. R. (2005). A long-term evaluation of fruiting phenology: Importance of climate
change. Journal of Tropical Ecology, 21, 31-45.

Chapman, C. A., & Wrangham, R. W. (1993). Range use of the forest chimpanzees of
Kibale: Implications for the understanding of chimpanzee social organization.
American Journal of Primatology, 31(4), 263-273.

Cheney, D. L., & Seyfarth, R. M. (1989). Redirected aggression and reconciliation
among vervet monkeys, Cercopithecus aethiops. Behaviour, 110, 258-275.

Cheney, D. L., & Seyfarth, R. M. (1990). How monkeys see the world. Chicago: Chicago
University Press.

Clay, Z., & de Waal, F. B. M. (2013a). Bonobos respond to distress in others:
Consolation across the age spectrum. PLoS ONE, 8(1), e55206.

173

Clay, Z., & de Waal, F. B. M. (2013b). Development of socio-emotional competence in
bonobos. Proceedings of the National Academy of Sciences, 110(45), 18121-
18126.

Clements, K. C., & Stephens, D. W. (1995). Testing non-kin cooperation: Mutualism and
the Prisoner’s Delimma. Animal Behaviour, 50, 527-535.

Clutton‐Brock, T. H., Greenwood, P. J., & Powell, R. P. (1976). Ranks and relationships
in Highland ponies and Highland cows. Zeitschrift für Tierpsychologie, 41(2),
202-216.

Colmenares, F. (2006). Is postconflict affiliation in captive nonhuman primates an artifact
of captivity? International Journal of Primatology, 27, 1311-1335.

Constable, J. L., Ashley, M. V., Goodall, J., & Pusey, A. E. (2001). Noninvasive
paternity assignment in Gombe chimpanzees. Molecular ecology, 10(5), 1279-
1300.

Cools, A. K. A., van Hout, A. J. M., & Nelissen, M. H. J. (2008). Canine reconciliation
and third-party-initiated postconflict affiliation: Do peacemaking social
mechanisms in dogs rival those of higher primates? Ethology, 114, 53-63.

Cooper, M., Aureli, F., & Singh, M. (2007). Sex differences in reconciliation and post-
conflict anxiety in bonnet macaques. Ethology, 113, 26-38.

Cordoni, G., & Palagi, E. (2007). Reconciliation in wolves (Canis lupus): New evidence
for a comparative perspective. Ethology, 114, 298-308.

Cordoni, G., Palagi,, E., & Tarli, S. (2006). Reconciliation and consolation in captive
western gorillas. International Journal of Primatology, 27, 1365-1382.

Cords, M. (1988). Resolution of aggressive conflicts by immature long-tailed macaques,
Macaca fascicularis. Animal Behaviour, 36, 1124-1135.

Cords, M. (1992). Post-conflict reunions and reconciliation in long-tailed macaques.
Animal Behaviour, 44, 57-61.

Cords, M. (1993). On operationally defining reconciliation. American Journal of
Primatology, 29(4), 255-267.

Cords, M., & Aureli, F. (2000). Reconciliation and relationship qualities. In F. Aureli &
F. B. M de Waal (Eds.), Natural conflict resolution (pp. 177-198). Berkeley:
University of California Press.

174

Cords, M., & Killen, M. (1998). Conflict resolution in human and nonhuman primates. In
J. Langer & M. Killen (Eds.), Piaget, evolution, and development (pp. 193-218).
Mahway, NJ: Lawrence Erlbaum Associates, Inc.

Cote, S. D. (2000). Dominance hierarchies in female mountain goats: Stability,
aggressiveness and determination of rank. Behaviour, 137, 1541-1566.

Craney, T. A., & Surles, J. G. (2002). Model-dependent variance inflation factor cutoff
values. Quality Engineering, 14(3), 391-403.

Creel, S., Creel, N. M., Mills, M. G., & Monfort, S. L. (1997). Rank and reproduction in
cooperatively breeding African wild dogs: Behavioral and endocrine correlates.
Behavioral Ecology, 8(3), 298-306.

Crockford, C., Wittig, R. M., Seyfarth, R. M., & Cheney, D. L. (2007). Baboons
eavesdrop to deduce mating opportunities. Animal Behaviour, 73, 885-890.

Crockford, C., Wittig, R. M., Whitten, P. L., Seyfarth, R. M., & Cheney, D. L. (2008).
Social stressors and coping mechanisms in wild female baboons (Papio
hamadryas ursinus). Hormones and Behavior, 53(1), 254-265.

Das, M. (2000). Conflict management via third parties: Post-conflict affiliation of the
aggressor. In Aureli, F. & de Waal, F. B. M. (Eds.), Natural conflict resolution
(pp. 263-280). Berkeley, CA: University of California Press.

Das, M., Penke, Z., & van Hooff, J. A. R. A. M. (1998). Postconflict affiliation and
stress-related behavior of long-tailed macaque aggressors. International Journal
of Primatology, 19, 53-71.

Deschner, T., Heistermann, M., Hodges, K., & Boesch, C. (2003). Timing and probability
of ovulation in relation to sex skin swelling in wild West African chimpanzees,
Pan troglodytes verus. Animal Behaviour, 66, 551–560.

Deschner, T., Heistermann, K., Hodges, K, & Boesch, C. (2004). Female sexual swelling
size, timing of ovulation, and male behavior in wild West African chimpanzees.
Hormones and Behavior, 46, 204-215.

de Vries, H., & Appleby, M. C. (2000). Finding an appropriate order for a hierarchy: A
comparison of the I&SI and the BBS methods. Animal Behaviour, 59(1), 239-245.

de Vries, H., Netto, W. J., & Hanegraaf, P. L. (1993). Matman: A program for the
analysis of sociometric matrices and behavioural transition matrices. Behaviour,
125(3), 157-175.

175

de Waal, F. B. M. (1982). Chimpanzee politics: Power and sex among apes. New York:
Harper and Row.

de Waal, F. B. M. (1986). The integration of dominance and social bonding in primates.
The Quarterly Review of Biology, 61, 459-479.

de Waal, F. B. M. (1996). Consolation, reconciliation and a possible difference between
macaques and chimpanzees. In A. E. Russon, K. A. Bard, & S. T. Parker (Eds.),
Reaching into thought: The minds of great apes (pp. 80-110). Cambridge:
Cambridge University Press.

de Waal, F. B. M. (2000). Primates: A natural heritage of conflict resolution. Science,
289(5479), 586-590.

de Waal, F. B. (2003). On the possibility of animal empathy. In A. S. R. Manstead, N.
Frijda, & A. Fischer, Feelings and emotions: The Amsterdam symposium (pp.
379-99). Cambridge: Cambridge University Press.

de Waal, F. B. M. (2008). Putting the altruism back into altruism: The evolution of
empathy. Annual Review of Psychology, 59, 1-22.

de Waal, F. B. M., & Aureli, F. (1996). Consolation, reconciliation and a possible
cognitive difference between macaques and chimpanzees. In A. E. Russon, K. A.
Bard, & S. Taylor Parker (Eds.), Reaching into thought: The minds of great apes
(pp. 80-110). Cambridge: Cambridge University Press.

de Waal, F., & Aureli, F. (1997). Conflict resolution and distress alleviation in monkeys
and apes. Annals of the New York Academy of Sciences, 807(1), 317-328.

de Waal, F. B. M., & Ren, R. (1988). Comparison of the reconciliation behavior of
stumptail and rhesus macaques. Ethology, 78, 129-142.

de Waal, F. B. M., & Suchak, M. (2010). Prosocial primates: Selfish and unselfish
motivations. Philosophical Transactions of the Royal Society B: Biological
Sciences, 365(1553), 2711-2722.

de Waal, F. B. M, & van Hooff, J. A. R. A. M. (1981). Side-directed communication and
agonistic interactions in chimpanzees. Behaviour, 77, 164–198.

de Waal, F. B. M., & van Roosmalen, A. (1979). Reconciliation and consolation among
chimpanzees. Behavioral Ecology and Sociobiology, 5, 55-66.

176
de Waal, F. B. M., Yoshihara, D. (1983). Reconciliation and redirected affection in
rhesus monkeys. Behaviour, 85, 224-241.

Dobson, A. J. (1990). An introduction to generalized linear models. London: Chapman &
Hall.

Dunbar, R. I. (1988). Primate social systems. London: Croom Helm.

Durham, D. L. (2006). Primates in print: Using text analysis of journal articles to
examine publication trends. International Journal of Primatology, 27(Supplement
1), 437 Abstract.

Engh, A. L., Beehner, J. C., Bergman, T. J., Whitten, P. L., Hoffmeier, R. R., Seyfarth, R.
M., & Cheney, D. L. (2006). Female hierarchy instability, male immigration and
infanticide increase glucocorticoid levels in female chacma baboons. Animal
Behaviour, 71(5), 1227-1237.

Emery Thompson, M., Kahlenberg, S. M., Gilby, I. C., & Wrangham, R. W. (2007). Core
area quality is associated with variance in reproductive success among female
chimpanzees at Kibale National Park. Animal Behaviour, 73(3), 501-512.

Eisenberg, N., & Fabes, R. A. (1998). Prosocial development. In W. Darmon & N.
Eisenberg (Eds.), Handbook of child psychology: Vol. 3. Social, emotional, and
personality development (pp.701–778). New York: Wiley.

Fawcett, K., & Muhumuza, G. (2000). Death of a wild chimpanzee community member:
Possible outcome of intense sexual competition. American Journal of
Primatology, 51, 243-247.

Fedurek, P., Machanda, Z. P., Schel, A. M., & Slocombe, K. E. (2013). Pant hoot
chorusing and social bonds in male chimpanzees. Animal Behaviour, 86(1), 189-
196.

Fraser, O. N. (2008). Reconciliation, consolation and relationship quality in chimpanzees.
Doctoral dissertation, Liverpool John Moores University, UK.

Fraser, O. N., & Aureli, F. (2008). Reconciliation, consolation and postconflict
behavioral specificity in chimpanzees. American Journal of Primatology, 70, 1-
10.

Fraser, O. N., & Bugnyar, T. (2010). Do ravens show consolation? Responses to
distressed others. PLoS One, 5(5), e10605.

177
Fraser, O. N., Koski, S. E., Wittig, R. M., & Aureli, F. (2009). Why are bystanders
friendly to recipients of aggression? Communicative & Integrative Biology, 2(3),
285-291.

Fraser, O. N., Schino, G., & Aureli, F. (2008a). Components of relationship quality in
chimpanzees. Ethology, 114, 834-843.

Fraser, O. N., Stahl, D., & Aureli, F. (2008b). Stress reduction through consolation in
chimpanzees. PNAS, 105(25), 8557-8562.

Fuentes, A., Malone, N., Sanz, C., Matheson, M., & Vaughan, L. (2002). Conflict and
post-conflict behaviour in a small group of chimpanzees. Primates, 43, 223-235.

Fujisawa, K. K., Kutsukake, N., & Hasegawa, T. (2006). Peacemaking and consolation in
Japanese preschoolers witnessing peer aggression. Journal of comparative
psychology, 120(1), 48-57.

Furuichi, T. (1997). Agonistic interactions and matrifocal dominance rank of wild
bonobos (Pan paniscus) at Wamba. International Journal of Primatology, 18(6),
855-875.

Gagneux, P., & Varki, A. (2001). Genetic differences between humans and great apes.
Molecular phylogenetics and evolution, 18(1), 2-13.

Gerloff, U., Hartung, B., Fruth, B., Hohmann, G., & Tautz, D. (1999). Intracommunity
relationships, dispersal pattern and paternity success in a wild living community
of Bonobos (Pan paniscus) determined from DNA analysis of faecal samples.
Proceedings of the Royal Society of London. Series B: Biological Sciences,
266(1424), 1189-1195.

Ghiglieri, M. P. (1984). The chimpanzees of Kibale Forest. New York: Columbia
University Press.

Ghiglieri, M. P. (1989). Hominoid sociobiology and hominid social evolution. In P.G.
Heltne & L.A. Marquardt (Eds.), Understanding chimpanzees (pp. 370-379).
Cambridge, MA: Harvard University Press.

Gilby, I. C., & Wrangham, R. W. (2008). Association patterns among wild chimpanzees
(Pan troglodytes schweinfurthii) reflect sex differences in cooperation.
Behavioral Ecology and Sociobiology, 62(2), 1831-1842.

Gilman, A., & Hill, J. (2006). Data Analysis Using Regression and Multilevel/
Hierarchical Models. Cambridge University Press.

178
Goldstein, H. (2011). Multilevel Statistical Models. UK: John Wiley & Sons, Ltd.

Goodall, J. (1986). The chimpanzees of Gombe: Patterns of behaviour. Cambridge, MA:
Belknap Press.

Goodall, J. (1992). Unusual violence in the overthrow of an alpha male chimpanzee at
Gombe. In T. Nishida, W.C. McGrew, P. Marler, M. Pickford, & F.B.M. de Waal
(Eds.), Topics in primatology. Tokyo: University of Tokyo Press.

Goodall, J. (2007). Bixby symposium on population and conservation: Keynote address.
Population Environment, 28, 274-282.

Hare, B., & Tomasello, M. (2005). Human-like social skills in dogs? Trends in cognitive
sciences, 9(9), 439-444.

Hasegawa, T., & Hiraiwa-Hasegawa, M. (1990). Sperm competition and mating
behavior. In T. Nishida (Ed.), The chimpanzees of the Mahale Mountain: Sexual
and life history strategies (pp. 115-132). Tokoyo: University of Tokoyo Press.

Hatfield, E., Cacioppo, J. T., & Rapson, R. L. (1994). Emotional contagion. Cambridge,
UK: Cambridge University Press.

Hockings, K. J., Yamakoshi, G., Kabasawa, A., & Matsuzawa, T. (2010). Attacks on
local persons by chimpanzees in Bossou, Republic of Guinea: Long‐term
perspectives. American journal of primatology, 72(10), 887-896.

Hofer, H., & East, M. L. (2000). Conflict management in female-dominated spotted
hyenas. In F. Aureli & F. B. M. de Waal (Eds.), Natural conflict resolution (pp.
232-234). Berkeley: University of California Press.

Hoffman, M. L. (2000). Empathy’s limitations: Over-arousal and bias. In M. L. Hoffman
(Ed.), Empathy and moral development: Implications for caring and justice (pp.
197-220). Cambridge: Cambridge University Press.

Holobinko, A., & Waring, G. H. (2010). Conflict and reconciliation behavior trends of
the bottlenose dolphin (Tursiops truncatus). Zoo biology, 29(5), 567-585.

Hothorn, T., Bretz, F., & Westfall, P. (2008). Simultaneous inference in general
parametric models. Biometrical Journal 50(3), 346-363.

Hrdy, S. B., & Hrdy, D. B. (1976). Hierarchical relations among female Hanuman
langurs (Primates: Colobinae, Presbytis entellus). Science, 193(4256), 913-915.

179
Isabirye-Basuta, G. (1989). Feeding ecology of chimpanzees in the Kibale Forest,
Uganda. In P.G. Heltne & L.A. Marquardt (Eds.), Understanding chimpanzees
(pp. 116-127). Cambridge, MA: Harvard University Press.

Itakura, S. (1993). Emotional behavior during the learning of a contingency task in a
chimpanzee. Perceptual and motor skills, 76(2), 563-566.

Jameson, K. A., Appleby, M. C., & Freeman, L. C. (1999). Finding an appropriate order
for a hierarchy based on probabilistic dominance. Animal Behaviour, 57(5), 991-
998.

Janson, C. H., & Goldsmith, M. L. (1995). Predicting group size in primates – Forging
costs and predation risks. Behavioral Ecology, 6, 326-336.

Judge, P. G. (1991). Dyadic and triadic reconciliation in pigtail macaques (Macaca
nemestrina). American Journal of Primatology, 23, 225–237.

Judge, P. G., & Mullen, S. H. 2005. Quadratic postconflict affiliation among bystanders
in a hamadryas baboon group Animal Behaviour, 69, 1345–1355.

Kahlenberg, S. M., Emery Thompson, M., & Wrangham, R. W. (2008). Female
competition over core areas in Pan troglodytes schweinfurthii, Kibale National
Park, Uganda. International Journal of Primatology, 29(4), 931-947.

Kappeler, P. M., & Schaik, C. P. (1992). Methodological and evolutionary aspects of
reconciliation among primates. Ethology, 92(1), 51-69.

King, J. A. (1973). The ecology of aggressive behavior. Annual Review of Ecology and
Systematics, 117-138.

King, M. C., & Wilson, A. C. (1975). Evolution at two levels in humans and
chimpanzees. Science, 188(4184), 107-116.

Kinzey, W. G. (1987). Evolution of human behavior: Primate models. New York: SUNY
Press.

Kojima, S., Izumi, A., & Ceugniet, M. (2003). Identification of vocalizers by pant hoots,
pant grunts and screams in a chimpanzee. Primates, 44(3), 225-230.

Koski, S. E., de Vries, H., van den Tweel, S. W., & Sterck, E. H. M. (2007a). What to do
after a fight? The determinants and inter-dependency of post-conflict interactions
in chimpanzees. Behaviour, 144, 529-555.

180
Koski, S. E., Koops, K., & Sterck, E. H. M. (2007b). Reconciliation, relationship quality,
and postconflict anxiety: Testing the integrated hypothesis in captive
chimpanzees. American Journal of Primatology, 69, 158-172.

Koski, S. E., & Sterck, E. H. M. (2007). Triadic postconflict affiliation in captive
chimpanzees: Does consolation console? Animal Behaviour, 73, 133-142.

Koski, S. E., & Sterck, E. H. M. (2009). Post-conflict third-party affiliation in
chimpanzees: What’s in it for the third party? American Journal of Primatology,
71, 409-418.

Koyama, N. F. (2001). The long-term effects of reconciliation in Japanese macaques
Macaca fuscata. Ethology, 107, 975-987.

Koyama, N. F., & Palagi, E. (2006). Managing conflict: Evidence from wild and captive
primates. International Journal of Primatology, 27, 1235-1240.

Kutner, M. H.; Nachtsheim, C. J.; Neter, J. (2004). Applied Linear Regression Models
(4th ed.). McGraw-Hill Irwin.

Kutsukake, N. (2003). Assessing relationship quality and social anxiety among wild
chimpanzees using self-directed behaviours. Behaviour, 140, 1153-1171.

Kutsukake, N., & Castles, D. L. (2001). Reconciliation and variation in post-conflict
stress in Japanese macaques (Macaca fuscata fuscata): Testing the integrated
hypothesis. Animal Cognition, 4, 259-268.

Kutsukake, N., & Castles, D. L. (2004). Reconciliation and post-conflict third-party
affiliation among wild chimpanzees in the Mahale Mountains, Tanzania.
Primates, 45, 157-165.

Kutsukake, N., & Clutton-Brock, T. H. (2010). Grooming and the value of social
relationships in cooperatively breeding meerkats. Animal Behaviour, 79, 271-279.

Langergraber, K. E., Mitani, J. C., & Vigilant, L. (2007). The limited impact of kinship
on cooperation in wild chimpanzees. Proceedings of the National Academy of
Sciences, 104, 7786-7790.

Langergraber, K., Mitani, J., & Vigilant, L. (2009). Kinship and social bonds in female
chimpanzees (Pan troglodytes). American Journal of Primatology, 71(10), 840-
851.

Laufer, P. (2011). Forbidden creatures: Inside the world of animal smuggling and exotic
pets. Rowman & Littlefield.

181
Leavens, D. A., Aureli, F., Hopkins, W. D., & Hyatt, C. W. (2001). Effects of cognitive
challenge on self‐directed behaviors by chimpanzees (Pan troglodytes). American
journal of primatology, 55(1), 1-14.

Leca, J. B., Fornasieri, I., & Petit, O. (2002). Aggression and reconciliation in Cebus
capucinus. International Journal of Primatology, 23(5), 979-998.

Lehmann, J., & Boesch, C. (2005). Bisexually bonded ranging in chimpanzees (Pan
troglodytes verus). Behavioral Ecology and Sociobiology, 57(6), 525-535.

Lehmann, J., & Boesch, C. (2008). Sexual differences in chimpanzee sociality.
International Journal of Primatology, 29(1), 65-81.

Ljungberg, T., Horowitz, L., Jansson, L., Westlund, K., & Clarke, C. (2005).
Communicative factors, conflict progression, and use of reconciliatory strategies
in pre-school boys: A series of random events or a sequential process? Aggressive
Behavior, 31, 303-323.

Maestripieri, D., Schino, G., Aureli, F., & Troisi, A. (1992). A modest proposal -
displacement activities as an indicator of emotions in primates. Animal Behaviour,
44, 967-979.

Mallavarapu, S., Stoinski, T. S., Bloomsmith, M. A., & Maple, T. L. (2006). Postconflict
behavior in captive western lowland gorillas (Gorilla gorilla gorilla). American
Journal of Primatology, 68, 789-801.

Majolo, B., Ventura, R., & Koyoma, N. F. (2005). Postconflict behaviour among male
Japanese macaques. International Journal of Primatology, 26, 321-336.

Manson, J. H., Perry, S., & Stahl, D. (2005). Reconciliation in wild white-faced
capuchins (Cebus capucinus). American Journal of Primatology, 65, 205–219.

Manson, J. H., Wrangham, R. W., Boone, J. L., Chapais, B., Dunbar, R. I. M., Ember, C.
R., et al. (1991). Intergroup Aggression in Chimpanzees and Humans [and
Comments and Replies]. Current Anthropology, 32(4), 369-390.

Matsumoto-Oda, A. (1999). Female choice in the opportunistic mating of wild
chimpanzees (Pan troglodytes schweinfurthii) at Mahale. Behavioral Ecology and
Sociobiology, 46, 258-266.

Matsumoto-Oda, A. (2002). Behavioral seasonality in Mahale chimpanzees. Primates,
43(2), 103-117.

182
McCullagh, P., & Nelder, J. A. 1989. Generalized linear models. New York: Chapman &
Hall.

McFarland, R., & Majolo, B. (2011). Reconciliation and the costs of aggression in wild
Barbary macaques (Macaca sylvanus): A test of the integrated hypothesis.
Ethology, 117(10), 928-937.

McFarland, R., & Majolo, B. (2013). The importance of considering the behavioral form
of reconciliation in studies of conflict resolution. International Journal of
Primatology, 34(1), 15-29.

Melis, A. P., Hare, B., & Tomasello, M. (2006). Engineering cooperation in
chimpanzees: Tolerance constraints on cooperation. Animal Behaviour, 72, 275-
286.

Melis, A. P., Warneken, F., & Hare, B. (2010). Collaboration and helping in
chimpanzees. In. E.V. Lonsdorf, S.R. Ross, & T. Metsuzawa (Eds.), The mind of
the chimpanzee: Ecological and experimental perspectives (pp. 278-393).
Chicago: University of Chicago Press.

Mesterton-Gibbons, M, & Dugatkin, L. A. (1992). Cooperation among unrelated
individuals: Evolutionary factors. Quarterly Review of Biology, 67, 267-281.

Mitani, J. C. (2009). Male chimpanzees form enduring and equitable social bonds.
Animal Behaviour, 77, 633-640.

Mitani, J. C., Merriwether, D. A., & Zhang, C. (2000). Male affiliation, cooperation, and
kinship in wild chimpanzees. Animal Behaviour, 59(4), 885-893.

Mitani, J. C., & Watts, D.P. (2001). Why do chimpanzees hunt and share meat? Animal
Behaviour, 61, 915-924.

Muller, M. N. (2002). Agonistic relations among Kanyawara chimpanzee. In C. Boesch,
G. Hohmann, & L. Marchant (Eds.), Behavioural diversity in chimpanzees and
bonobos (pp. 112-124). Cambridge: Cambridge University Press.

Muller, M. N., Emery Thompson, M, & Wrangham, R. W. (2006). Male chimpanzees
prefer mating with old females. Current Biology, 16(22), 2234-2238.

Muller, M. N., Kahlenberg, S. M., Emery Thompson, M., Wrangham, R. W. (2007).
Male coercion and the costs of promiscuous mating for female chimpanzees.
Proceedings of the Royal Society B: Biological Sciences, 274(1612), 1009-1014.

183
Muller, M. N., Kahlenberg, S. M., & Wrangham, R. W. (2009). Male aggression against
females and sexual coercion in chimpanzees. In M. Muller & R.W. Wrangham
(Eds.), Sexual coercion in primates and humans: An evolutionary perspective on
male aggression against females (pp. 184-217). Cambridge, MA: Harvard
University Press.

Muller, M. N., & Mitani, J. C. (2005). Conflict and cooperation in wild chimpanzees.
Advances in the Study of Behavior, 35, 275-331

Muller, M. N., Thompson, M. E., Kahlenberg, S. M., & Wrangham, R. W. (2011). Sexual
coercion by male chimpanzees shows that female choice may be more apparent
than real. Behavioral Ecology and Sociobiology, 65(5), 921-933.

Muller, M. N., & Wrangham, R. W. (2001). The reproductive ecology of male
hominoids. In P. T. Ellison (Ed.), Reproductive ecology and human evolution (pp.
397-427), New York: Aldine de Gruyter.
Muller, M. N., & Wrangham, R. W. (2004a). Dominance, aggression and testosterone in
wild chimpanzees: A test of the ‘challenge hypothesis.’ Animal Behaviour, 67,
113-123.

Muller, M. N., & Wrangham, R. W. (2004b). Dominance, cortisol and stress in wild
chimpanzees (Pan troglodytes schweinfurthii). Behavioral Ecology and
Sociobiology, 55(4), 332-340.

Neat, F. C., Taylor, A. C., & Huntingford, F. A. (1998). Proximate costs of fighting in
male cichlid fish: The role of injuries and energy metabolism. Animal Behaviour,
55, 875-882.

Newton-Fisher, N. E., & Emery Thompson, M. (2012). Comparative evolutionary
perspectives on violence. In T. K. Shackelford & V. A. Veekes-Shackelford
(Eds.), Oxford handbook of evolutionary perspectives on violence, homicide, and
war (pp. 41-60). Oxford: Oxford University Press.

Newton-Fisher, N. E., Reynolds, V., Plumptre, A. J. (2000). Food supply and chimpanzee
(Pan troglodytes schweinfurthii) party size in the Budongo Forest Reserve,
Uganda. International Journal of Primatology, 21(4), 613-628.

Nishida, T. (1968). The social group of wild chimpanzees in the Mahale mountains.
Primates, 9, 167-224.

Nishida, T. (1970). Social behavior and relationship among wild chimpanzees of the
Mahale mountains. Primates, 11, 1-47.

184
Nishida, T. (1979). The social structure of chimpanzees of the Mahale mountains. In
D.A. Hamburg & E. R. McCown (Eds.), The great apes (pp. 73-121). Menlo
Park, CA: Benjamin & Cummings.

Nishida, T. (1983). Alpha status and agonistic alliance in wild chimpanzees (Pan
troglodytes schweinfurthii). Primates, 24(3), 318.

Nishida, T. (1989). Social interactions between resident and immigrant female
chimpanzees. In P. Heltne & L. Marquardt (Eds.), Understanding chimpanzees
(pp. 68-89). Cambridge, MA: Harvard University Press.

Nishida, T., & Hosaka, K. (1996). Coalition strategies among adult male chimpanzees of
the Mahale Mountains, Tanzania. In W.C. McGrew, L.F. Marchant, & T. Nishida
(Eds.), Great ape societies (pp. 114-134). Cambridge: Cambridge University
Press.

Nishida, T., Hosaka, K., Nakamura, M, & Hamia, M. (1995). A within-group gang attack
on a young adult male chimpanzee: Ostracism of an ill-mannered member?
Primates, 36, 207-211.

Nishida, T., Kano, T., Goodall, J., McGrew, W. C., & Nakamura, M. (1999). Ethogram
and ethnography of Mahale chimpanzee. Anthropological Science, 107, 141-188.

Okamoto, K., Agetsuma, N., & Kojima, S. (2001). Greeting behavior during party
encounters in captive chimpanzees. Primates, 42, 161-165.

Osmaston, H. A. (1959). Working plan for the Kibale and Itwara Forests. Uganda Forest
Department, Entebbe.

Otali, E., & Gilchrist, J. S. (2006). Why chimpanzee (Pan troglodytes schweinfurthii)
mothers are less gregarious than nonmothers and males: The infant safety
hypothesis. Behavioral Ecology and Sociobiology, 59(4), 561-570.

Palagi, E. Chiarugi, E., & Cordoni, G. (2008). Peaceful post-conflict interactions between
aggressors and bystanders in captive lowland gorillas (Gorilla gorilla gorilla).
International Journal of Primatology, 70, 949-955.

Palagi, E., Cordoni, G., & Borgognini Tarli, S. (2006). Possible roles of consolation in
captive chimpanzees (Pan troglodytes). American Journal of Physical
Anthropology, 129, 105-111.

Palagi, E., & Cordoni, G. (2009). Postconflict third-party affiliation in Canis lupus: Do
wolves share similarities with the great apes? Animal Behaviour, 78(4), 979-986.

185
Palagi, E., Dall’Olio, S., Demuru, E., & Stanyon, R. (2014). Exploring the evolutionary
foundations of empathy: Consolation in monkeys. Evolution and Human
Behavior, 35(4), 341-349.

Palagi, E., & Norscia, I. (2013). Bonobos protect and console friends and kin. PloS One,
8(11), e79290.

Palagi, E., Paoli, T., & Borgognini Tarli, S. M. (2004). Reconciliation and consolation in
captive bonobos (Pan paniscus). American Journal of Primatology, 62, 15-30.

Parr, L. A., Cohen, M., & de Waal, F. B. M. (2005). The influence of social context on
the use of blended and graded facial displays in chimpanzees (Pan troglodytes).
International Journal of Primatology, 26, 73-103.

Parr, L. A., Waller, B. M., Vick, S. J., & Bard, K. A. (2007). Classifying chimpanzee
facial expressions using muscle action. Emotion, 7(1), 172.

Patzelt, A., Pirow, R., & Fischer, J. (2009). Post-conflict affiliation in barbary macaques
is influenced by conflict characteristics and relationship quality, but does not
diminish short-term renewed aggression. Ethology, 115(7), 658-670.
Payne, K. (2003). Sources of social complexity in the three elephant species. In F. B. M.
de Waal & P. L. Tyack (Eds.), Animal social complexity: Intelligence, culture,
and individualized societies (pp. 57-85). Cambridge, MA: Harvard University
Press.

Pellegrini, A.D. (2008). The roles of aggressive and affiliative behaviours in resource
control: A behavioural ecology perspective. Developmental Review, 28, 461-487.

Pepper, J. W., Mitani, J. C., & Watts, D. P. (1999). General gregariousness and specific
social preferences among wild chimpanzees. International Journal of
Primatology, 20(5), 613-632.

Pieta, K. (2008). Female mate preferences among Pan troglodytes schweinfurthii of
Kanyawara, Kibale National Park, Uganda. International Journal of Primatology,
29(4), 845-864.

Plotnik, J. M., De Waal, F. B., & Reiss, D. (2006). Self-recognition in an Asian elephant.
Proceedings of the National Academy of Sciences, 103(45), 17053-17057.

Plumptre, A. J., Cox, D., & Mugume, S. (2003). The Status of Chimpanzees in Uganda.
Albertine Rift Technical Report Series No. 2. Wildlife Conservation Society.

Preston, S. D., & de Waal, F. B. M. (2002). Empathy: Its ultimate and proximate bases.
Behavioral and Brain Sciences, 25, 1-72.

186

Preuschoft, S., & van Schaik, C. P. (2000). Dominance and communication. Conflict
management in various social settings, 77-105.

Preuschoft, S., Wang, X. F. A., & de Waal, F. B. M. (2002). Reconciliation in captive
chimpanzees: A reevaluation with controlled methods. International Journal of
Primatology, 23, 29-50.

Pusey, A., Williams, J., & Goodall, J. (1997). The influence of dominance rank on the
reproductive success of female chimpanzees. Science, 277, 828-831.

R Core Team. (2012). R: A language and environment for statistical computing. R
Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0,
URL http://www.R-project.org/.

Riss, D., & Goodall, J. (1977). The recent rise to the alpha-rank in a population of free-
living chimpanzees. Folia primatologica, 27(2), 134-151.

Roeder, J. J., Fornasieri, I., & Gosset, D. (2002). Conflict and post-conflict behaviour in
two lemur species with different social organizations (Eulemur fulvus and
Eulemur macaco): A study on captive groups. Aggressive Behavior, 28, 62-74.

Rogerson, P. A. (2001). Statistical methods for geography. London: Sage.

Romero, T., Castellanos, M. A., & de Waal, F. B. M. (2010). Consolation as possible
expression of sympathetic concern among chimpanzees. Proceedings of the
National Academy of Sciences, 107(27), 12110-12115.

Romero, T., Castellanos, M. A., & de Waal, F. B. M. (2011). Post-conflict affiliation by
chimpanzees with aggressors: Other-oriented versus selfish political strategy.
PloS One, 6(7), e22173.

Romero, T., & de Waal, F. B. M. (2010). Chimpanzee (Pan troglodytes) consolation:
Third-party identity as a window on possible function. Journal of Comparative
Psychology, 124(3), 278.

Rooney, N. J., & Bradshaw, J. W. S. (2006). Social cognition in the domestic dog:
Behaviour of spectators towards participants in interspecific games. Animal
Behaviour, 72, 343-352.

Ross, S. R., Lukas, K. E., Lonsdorf, E. V., Stoinski, T. S., Hare, B., Shumaker, R., &
Goodall, J. (2008). Inappropriate use and portrayal of chimpanzees. Science, 319,
1487.

187
Sackin, S., & Thelen, E. (1984). An ethological study of peaceful associative outcomes to
conflict in preschool children. Child Development, 55, 1098-1102.

Samuels, A., & Flaherty, C. (2000). Peaceful conflict resolution in the sea? In F. Aureli &
F. B. M. de Waal (Eds.), Natural conflict resolution (pp.229-231). Berkeley, CA:
University of California Press.

Sapolsky, R. M. (1993). The physiology of dominance in stable versus unstable social
hierarchies. In W. Mason & S. Mendoza (Eds.), Primate social conflict (pp. 171-
204). Albany, NY: SUNY Press.

Sapolsky, R. M. (1995). Social subordinace as a marker of hypercortisolism: Some
unexpected subtleties. Annals of the New York Academy of Sciences, 711, 626-
639.

Sapolsky, R. M. (2005). The influence of social hierarchy on primate health. Science,
308(5722), 648-652.

Schaffner, C. M., Aureli, F., & Caine, N. G. (2005). Following the rules: Why small
groups of tamarins do not reconcile conflicts. Folia Primatologica, 76, 67-76.

Schino, G. (1998). Reconciliation in domestic goats. Behaviour, 135, 343-356.

Schino, G. (2000). Beyond the Primates. In F. Aureli & F. B. M. de Waal (Eds.), Natural
conflict resolution (pp.225-242). Berkeley, CA: University of California Press.

Schino, G., Perretta, G., Taglioni, A., Monaco, V., & Troisi, A. (1996). Primate
displacement activities as an ethnopharmacological model of anxiety. Anxiety, 2,
186-191.

Schino, G., Rosati, L., Geminiani, S., & Aureli, F. (2007). Post-conflict anxiety in
Japanese macaques (Macaca fuscata): Aggressor’s and victim’s perspectives.
Ethology, 113, 1081-1088.

Scott, J. P. (1958). Aggression. Chicago: University of Chicago Press.

Seeds, A. M., Clayton, N. S., & Emery, N. J. (2007). Postconflict third-party affiliation in
rooks, Corvus frugilegus. Current Biology, 17, 152-158.

Siebert, E. R., & Parr, L. A. (2003). A structural and contextual analysis of chimpanzee
screams. Annals of the New York Academy of Sciences, 1000(1), 104-109.

Silk, J. B. (1996). Why do primates reconcile? Evolutionary Anthropology, 5, 39-42.

188
Silk, J. B. (1997). The function of peaceful post-conflict contact among primates.
Primates,38, 265-279.

Silk, J. B. (1998). Making amends: Adaptive perspectives on conflict remediation in
monkeys, apes, and humans. Human Nature, 9(4), 341.

Silk, J. B. (2000). The function of peaceful post-conflict interactions: An alternate view.
In F. Aureli & F. B. M. de Waal (Eds.), Natural conflict resolution (pp.179-181).
Berkeley, CA: University of California Press.

Silk, J.B. (2002a). The form and function of reconciliation in primates. Annual Review in
Anthropology, 31, 21-44.

Silk, J. B. (2002b). Practice random acts of aggression and senseless acts of intimidation:
The logic of status contests in social groups. Evolutionary Anthropology: Issues,
News, and Reviews, 11(6), 221-225.

Silk, J. B., Cheney, D., & Seyfarth, R. (1996). The form and function of post-conflict
interactions between female baboons. Animal Behaviour, 52, 259-268.

Slocombe, K. E., Kaller, T., Call, J., & Zuberbühler, K. (2010). Chimpanzees extract
social information from agonistic screams. PLoS One, 5(7), e11473.

Slocombe, K. E., Townsend, S. W., & Zuberbühler, K. (2009). Wild chimpanzees (Pan
troglodytes schweinfurthii) distinguish between different scream types: Evidence
from a playback study. Animal cognition, 12(3), 441-449.

Slocombe, K. E., & Zuberbühler, K. (2005). Agonistic Screams in Wild Chimpanzees
(Pan troglodytes schweinfurthii) vary as a Function of Social Role. Journal of
Comparative Psychology, 119(1), 67.
Slocombe, K. E., & Zuberbühler, K. (2007). Chimpanzees modify recruitment screams as
a function of audience composition. Proceedings of the National Academy of
Sciences, 104(43), 17228-17233.

Sommer, V., Denham, A., & Little, K. (2002). Postconflict behaviour of wild Indian
langur monkeys: Avoidance of opponents but rarely affinity. Animal Behaviour,
63, 637.

Stanford, C. B. (1998). Chimpanzee and red colobus. Cambridge: Harvard University
Press.

Stanford, C. B. (2002). Avoiding predators: Expectations and evidence in primate
antipredator behavior. International Journal of Primatology, 23, 741-757.

189
Struhsaker, T. T. (1975). The red colobus monkey. University of Chicago Press, Chicago.

Struhsaker, T. T. (1997). Ecology of an African rain forest: Logging in Kibale and the
conflict between conservation and exploitation. University Press of Florida.

Stumpf, R. M., & Boesch, C. (2005). Does promiscuous mating preclude female choice?
Female sexual strategies in chimpanzees (Pan troglodytes verus) of the Tai
National Park, Cote d’Ivoire. Behavioral Ecology and Sociobiology, 57(5), 511-
524.

Stumpf, R.M., & Boesch, C. (2006). The efficacy of female choice in chimpanzees of the
Taï Forest, Côte d’Ivoire. Behavioral Ecology and Sociobiology, 60(6), 749-765.

Tabachnick, B. G., & Fidell, L. S. (2007). Multilevel linear modeling. Using multivariate
statistics, 781-857.

Takahata, Y. (1990). Social relationships among adult males. The Chimpanzees of the
Mahale Mountains, University of Tokyo Press, Tokyo, 149-170.

te Boekhorst, I. J. A., de Weerth, C., & van Hooff, J. A. R. A. M. (1991). Does scratching
signal stress in chimpanzees? In I. J. A. te Boekhorst (Ed.), Social structure of
three great ape species (pp.157-175). Utrecht: University of Utrecht Press.

Tinbergen, N. (1952). “Derived” activities; their causation, biological significance,
origin, and emancipation during evolution. The Quarterly Review of Biology,
27(1), 1-32.

Tomasello, M. J., & Call, J. (1997). Primate cognition. Oxford: Oxford University Press.

Townsend, S. W., Slocombe, K. E., Emery Thompson, M., & Zuberbühler, K. (2007).
Female-led infanticide in wild chimpanzees. Current Biology, 17, R355-R356.

Troisi, A. (2002). Displacement activities as a behavioural measure of stress in
nonhuman primates and human subjects. Stress, 5, 47-54.

van der Dennen, J., & Falger, V. (1990). Sociobiology and conflict. London: Chapman &
Hall.

van Hooff, J. A. R. A. M. (1967). The facial displays of the Catarrhine monkeys and
apes. In D. Morris (Ed.), Primate ethology (pp. 7-68). Chicago: Aldine.

190
Van Hooff, J. A. R. A. M. (1973). A structural analysis of the social behaviour of a semi-
captive group of chimpanzees. In M. von Cranach & I. Vine (Eds.), Social
communication and movement: Studies of interaction and expression in man and
chimpanzee (pp.75-162). London: Academic Press.

van Hooff, J. A. R. A. M., & van Schaik, C. P. (1992). Cooperation in competition: The
ecology of primate bonds. In A. H. Harcourt & F. B. M. de Waal (Eds.),
Coalitions and alliances in humans and other animals (pp. 357-389). Oxford:
Oxford University Press.

van Schaik, C. P. (1983). Why are diurnal primates living in groups? Behaviour, 87, 120-
144.

Veenema, H. C., Das, M., & Aureli, F. (1994). Methodological improvements for the
study of reconciliation. Behavioural Processes, 31, 29-38.

Verbeek, P., & de Waal, F. B. M. (2001). Peacemaking among pre-school children.
Journal of Peace Psychology, 7, 5-28.

Vigilant, L., Hofreiter, M., Siedel, H., & Boesch, C. (2001). Paternity and relatedness in
wild chimpanzee communities. Proceedings of the National Academy of Sciences,
98, 12890-12895.

Wahaj, S. A., Guse, K. R., & Holekamp, K. E. (2001). Reconciliation in the spotted
hyena (Crocuta crouta). Ethology, 107, 1057-1074.

Wakefield, M. L. (2008). Grouping patterns and competition among female Pan
troglodytes schweinfurthii at Ngogo, Kibale National Park, Uganda. International
Journal of Primatology, 29(4), 907-929.

Wallis, J. (1992). Socioenvironmental effects on timing of first postpartum cycles in
chimpanzees. In T. Nishida, W.C. McGrew, P. Marler, M. Pickford, and F. de
Waal (Eds.), Topics in primatology, vol. 1: Human origins (pp. 119-130). Tokyo:
University of Tokyo Press.

Wallis, J. (1997). A survey of reproductive parameters in the free-ranging chimpanzees
of Gombe National Park. Journal of Reproduction and Fertility, 109, 297-307.

Wallis, J., & Bettinger, T. (1993). Sexual attraction in wild chimpanzees: A comparison
of cyclic, pregnant, and lactating females with full anogenital swellings. American
Journal of Primatology, 30, 353-354 (abstract).

Wallis, J., & Goodall, J. (1993). Genital swelling patterns of pregnant females in Gombe
National Park. American Journal of Primatology, 31, 89-98.

191
Wallis, J., & Lemmon, W. B. 1986. Social behavior and genital swelling in pregnant
chimpanzees (Pan troglodytes). American Journal of Primatology, 10, 171-183.

Wallner, B., Moestle, E., Dittami, J., & Prossinger, H. (1999). Fecal gluccocorticoids
documents stress in female Barbary macaques (Macacca sylvanus). General and
Comparative Endocrinology, 113, 80-86.

Walters, J. R., & Seyfarth, R. M. (1987). Conflict and cooperation. In B. B. Smuts, D. L.
Cheney, R. M. Seyfarth, R. W. Wrangham, & T. T. Struhsaker (Eds.), Primate
societies (pp. 306-317). Chicago: University of Chicago Press.

Watts, D. P. (1995). Post-conflict social events in wild mountain gorillas. II. Redirection,
side-direction and consolation. Ethology, 100, 158-174.

Watts, D. P. (1998). Coalitionary mate guarding by male chimpanzees at Ngogo, Kibale
National Park, Uganda. Behavioral Ecology and Sociobiology, 44, 43-55.

Watts, D. P. (2002). Reciprocity and interchange in the social relationships of wild male
chimpanzees. Behaviour, 139(2-3), 343-370.

Watts, D. P. (2004). Intracommunity coalitionary killing of an adult male chimpanzee at
Ngogo, Kibale National Park, Uganda. International Journal of Primatology,
25(3), 507-521.

Watts, D. P. (2006). Conflict resolution in chimpanzees and the valuable-relationships
hypothesis. International Journal of Primatology, 27, 1337-1364.

Watts, D. P., Colmenares, F., & Arnold, K. (2000). Redirection, consolation, and male
policing: How targets of aggression interact with bystanders. In F. Aureli & F. B.
M. de Waal (Eds.), Natural conflict resolution (pp. 281-301). Berkeley:
University of California Press.

Watts, D. P., & Mitani, J. C. (2002). Hunting behavior of chimpanzees at Ngogo, Kibale
National Park, Uganda. International Journal of Primatology, 23(1), 1-28.

Watts, D. P., & Mitani, J. C. (2001). Boundary patrols and intergroup encounters in wild
chimpanzees. Behaviour, 138, 299-327.

Watts, D. P., Muller, M. N., Amsler, S. J., Mbabazi, G., & Mitani, J. C. (2006). Lethal
intergroup aggression by chimpanzees in Kibale National Park, Uganda.
American Journal of Primatology, 68(2), 161-180.

Weaver, A. (2003). Conflict and reconciliation in captive bottlenose dolphins, Tursiops
truncatus. Marine Mammal Science, 19, 836-846.

192
Wilson, M. L. (2001). Imbalances of power: How chimpanzees respond to the threat of
intergroup aggression. Doctoral dissertation, Harvard University, Cambridge,
MA.

Wilson, M. L., Boesch, C., Fruth, B., Furuichi, T., Gilby, I. C., Hashimoto, C., Hobaiter,
C. L., Hohmann, G., Itoh, N., Koops, K., Lloyd, J. N., Matsuzawa, T., Mitani, J.
C., Mjungu, D. C., Morgan, D., Muller, M. N., Mundry, R., Nakamura, M.,
Pruetz, J., Pusey, A. E., Riedel, J., Sanz, C., Schel, A. M., Simmons, N., Waller,
M., Watts, D. P., White, F., Wittig, R. M., Zuberbühler, K., & Wrangham, R. W.
(2014). Lethal aggression in Pan is better explained by adaptive strategies than
human impacts. Nature, 513(7518), 414-417.

Wilson, A. C., & Sarich, V. M. (1969). A molecular time scale for human evolution.
Proceedings of the National Academy of Sciences, 63(4), 1088-1093.

Wittig, R. M. (2004). Conflict management in wild chimpanzees (Pan troglodytes).
Doctoral dissertation, University of Leipzig, Germany.

Wittig, R. M. (2010). Function and cognitive underpinnings of post-conflict affiliation in
wild chimpanzees. In E. V. Lonsdorf, S. R. Ross, & T. Matsuzawa (Eds.), The
mind of the chimpanzee (pp. 208-219). Chicago: University of Chicago Press.

Wittig, R. M., & Boesch, C. (2003a). The choice of post-conflict interactions in wild
chimpanzees (Pan troglodytes). Behaviour, 140, 1527-1559.

Wittig, R. M., & Boesch, C. (2003b). “Decision-making” in conflicts of wild
chimpanzees (Pan troglodytes): An extension of the Relational Model. Behavioral
Ecology and Sociobiology, 54(5), 491-504.

Wittig, R. M., & Boesch, C. (2003c). Food competition and linear dominance hierarchy
among female chimpanzees of the Tai National Park. International Journal of
Primatology, 24, 847-867.

Wittig, R. M., & Boesch, C. (2005). How to repair relationships - Reconciliation in wild
chimpanzees (Pan troglodytes). Ethology, 111, 736-763.

Wittig, R. M., & Boesch, C. (2010). Receiving post-conflict affiliation from the enemy's
friend reconciles former opponents. PLoS One, 5(11), e13995.

Wittig, R. M., Crockford, C., Wikberg, E., Seyfarth, R. M., & Cheney, D. L. (2007). Kin-
mediated reconciliation substitutes for direct reconciliation in female baboons.
Proceedings of the Royal Society B: Biological Sciences, 274(1613), 1109-1115.

193
Wroblewski, E. E., Murray, C. M., Keele, B. F., Schumacher-Stankey, J. C., Hahn, B. H.,
& Pusey, A. E. (2009). Male dominance rank and reproductive success in
chimpanzees, Pan troglodytes schweinfurthii. Animal Behaviour, 77(4), 873-885.

Wrangham, R. W. (1980). An ecological model of female-bonded primates. Behaviour,
75, 262-300.

Wrangham, R. W. (1999). Evolution of coalitionary killing. Yearbook of Physical
Anthropology, 42, 1-30.

Wrangham, R. W. (2000). Why are male chimpanzees more gregarious than mothers? A
scramble competition hypothesis. In P. M. Kappeler (Ed.), Primate males (pp.
248-258). Cambridge: Cambridge University Press.

Wrangham, R. W. (2002). The cost of sexual attraction: Is there a tradeoff in female Pan
between sex appeal and received coercion? In C. Boesch, G. Hohmann, & L.
Merchant (Eds.), Behavioural diversity in chimpanzees and bonobos (pp. 204-
215). Cambridge: Cambridge University Press.

Wrangham, R. W. (2008). The International Primatological Society as a coalition:
Primatologists and the future of primates. International Journal of Primatology,
29, 3-11.

Wrangham, R. W., Chapman, C. A., Clark-Arcadi, A. P., & Isabirye-Basuta, G. (1996).
In W. C. McGrew, L. F. Marchant, T. Nishida (Eds.), Great ape societies (pp. 45-
57). Cambridge University Press.

Wrangham, R. W., Clark, A. P., & Isabirye-Basuta, G. (1992). Female social
relationships and social organization of Kibale Forest chimpanzees. In T. Nishida,
W. C. McGrew, P. Marler, M. Pickford, & F. B. M. de Waal (Eds.), Topics in
primatology vol. 1, Human origins (pp. 81-98). Tokyo: University of Tokyo
Press.

Wrangham, R. W., & Smuts, B. B. (1979). Sex differences in the behavioural ecology of
chimpanzees in the Gombe National Park, Tanzania. Journal of Reproduction and
Fertility, S28, 13-31.

Wrangham, R. W., & Wilson, M. L. (2004). Collective violence: Comparisons between
youths and chimpanzees. Annuals of New York Academic Science, 1036, 233-256.

Wrangham, R. W., Wilson, M. L., & Muller, M. N. (2006). Comparative rates of violence
in chimpanzees and humans. Primates, 47, 14-26.

194
Appendix A

Kanyawara Community Demographic Information

Name ID Sex Age-Class Birthdate Mother Area Observed
Big Brown BB M A 1966 C F
Bud PB M A 1/20/1995 Stump C F
Eslom ES M A 7/2/1994 Ekisigi C F
Imoso MS M A 1979 Finger N N
Johnny AJ M A 1974 Lope C F
Kakama KK M A 7/15/1985 Kabarole C F
Lanjo TJ M A 8/13/1995 Tongo C F
Makoku LK M A 1982 Lope C F
Stout ST M A 1955 C F
Tofu TU M A 1960 N I
Twig PG M A 1988 Stump C F
Yogi YB M A 1973 C F
Bubbles BL F A 1960 S I
Hillary HL F A 1978 N N
Harare HH F A 1993 Hillary N R
Leona LN F A I S
Lia AL F A 1982 C F
Michelle ML F A 1997 I S
Mususu MU F A 1970 N I
Outamba OU F A 1979 C F
Quinto QT F A 1992 C F
Rosa LR F A 1989 Lope C F
Rwanda RD F A 1996 I S
Tenkere OT F A 2/13/1998 C F
Tongo TG F A 1980 C F
Tripole
§
PO F A I R
Umbrella UM F A 1981 N R
Wangari WA F A 1991 N I
Wilma WL F A 1992 C F

195
Max MX M SA 1/15/1998 Mususu N I
Tacugama OG M SA 4/10/2001 Outamba C F
Tuber TT M SA 11/1/2000 Tongo C F
Tuke AT M SA 11/2/1999 Lia C F
Gaga GG F SA 7/2/2002 Gladys N N
Special NP F SA 2/9/2000 Nile C I
Bono BO M J 10/22/2003 Bubbles S I
Hexly HX M J 2005 Hillary N N
Likizo AZ M J 12/28/2004 Lia C F
Unasema UN M J 11/1/2004 Umbrella N I
Euro EU F J 5/25/2004 Ekisigi C F
Omusisa OM F J 6/23/2005 Outamba C F
Tsunami TS F J 1/23/2005 Tongo C F
Basuta BT M I 6/23/10 Bubbles S I
Hamilton HT M I 7/2/2008 Harare N R
Moon MN M I 12/26/2008 Mususu N I
Mpaka HM M I 1998 Hillary S N
Quiver QV M I 8/3/2008 Quinto C F
Teddy
x
M I 6/12/2009 Tongo C F
Wallace WC M I 8/5/2008 Wilma C F
Azania AN F I 5/1/2009 Lia C F
Betty LB F I 1/2/2011 Rosa C F
Buke UK F I 4/5/2011 Umbrella N I
Gola OL F I 3/3/2009 Outamba C F
Note. Age-class was divided into four categories (A=Adult; SA=Subadult; J=Juvenile;
I=Infant). Birthdates were gleaned from Kibale Chimpanzee Project long-term data
records. Many dates are estimates. Area indicates core region of territory where majority
of time was spent (N=Northern; C=Central; S=Southern; I=Immigrant with no
established core area). Observed indicates how often individuals were seen (F=Frequent,
on a weekly basis; I=Infrequent, on a monthly basis; R=Rare, only a few times during the
year; N=Never; S=Skittish but frequent). Names in bold indicate subjects of regular focal
follows. Names in italics indicate natal adult and subadult females. Teddy
x
died in
January 2011. Tripoli
§
was not included in the community size estimation because she
immigrated the last week of the study period.

196
Appendix B

Behavioral Ethogram

Behavior Code Definition
Movement, Proximity, and Directional Behaviors
Approach APP One individual actively approaches another
Avoid AV One individual actively avoid another by changing
their direction and moving away
Bipedal BI Moving upright on legs
Carry CAR Transporting something or someone from one location
to another
Climb CLB Climbing up, down, or laterally in tree
Cling CL Hanging while supporting the weigh of self in a tree
Crouch CR Hunched down in a crouching position
Dart DT Rapidly moving away from an individual. Females
often dart away from males after copulation
Displacement DISP One individual occupies the same space previous
occupied by another
Follow FO Walking/traveling after an individual while
maintaining close proximity
Hold HO Holding an item or individual
Ignore IGN Not responsive to stimuli, which could be a noise,
vocalization, action, individual, etc.
Interrupt INT One individual interrupting the behavior of another
either actively or indirectly
Lean away LA Position body to increase distance between self and
another individual

197
Leap LEA Jump propelled by both feet, typically occurring in the
tree canopy
Leave LEV Move away from an individual or party increasing
proximity beyond 10 m
Lie down LD Lie horizontally on back, side, or belly
Move away MA Distancing self from an individual while remaining in
fairly close proximity (within 10 m)
Move toward TO Decreasing distance between self and an individual
while moving in a line toward that individual
Nearest
neighbor
NN Individual closest in proximity
Nest NST Construction of a day or night nest
Rest RT Lying or sitting while not engaged in another activity,
includes sleeping
Reject REJ Individual offers affiliative interaction but is rejected
by partner (e.g., extend hand offered but not returned)
Reposition RP Changing posture while remaining in the same location
Retreat RTR Walk or run away when in pursuit, typically occurs
during intercommunity encounters or hunting
Reunite RUN Two parties or individuals coming together after an
absence
Ride RID Offspring cling to mother’s front or back while she
travels on group or in the tree
Sit SIT Upright posture
Stand STD Quadrupedal stance
Travel TR

Moving on the ground or in the trees when not feeding
or foraging
Wait WAIT Pausing for another individual. Mothers often wait for
their dependents when traveling

198
Feeding Behaviors
Beg BEG One individual begging (e.g., open hand) for the
resource possess by another. Often accompanied by
whimpering, occurring between a mother and offspring
Drink DR Ingesting water
Feed FD Ingesting food in the tree or on the ground, includes
species of food item
Forage FG Traveling short distances while feeding periodically
along the way, typically when feeding on various THV
(terrestrial herbaceous vegetation) species, includes
species of forage items
Geophagy GEO Consuming dirt
Hunt HT Stalking, pursing, capturing, and killing a monkey
species (i.e., red colobus, black and white colobus, blue
monkey)
Leaf sponge LSP Using a wadged or crumpled leaf to absorb water,
typically from a tree hole, and drink
Leaf swallow LSW Swallowing the leaf of certain plant species whole,
possibly for medicinal purposes
Food share FDS One individual offering food to another individual
(e.g., from mother to offspring, meat between males)
Vomit VOM Regurgitation of digested food
Wadge WAD Juice extract of food items (i.e., fruit, leaves, piths,
meat mixed with leaves, etc.) by chewing and
compressing item between the lower lip and teeth
Affiliative Behaviors
Branch shake BS Males grabbing and shaking a small tree or branch to
solicit copulation
Contact Sit CS Two or more individuals passive sitting in physical
contact of one another

199
Copulation COP Intromission and pelvic thrusting between males and
estrous females
Embrace EMB Open arms wrapped around an individual either
mutually or one-sided
Extended Hand EXHD Offering an open hand (palm up) with arm extended to
another individual, reciprocal gesture often used in
reconciliation and consolation
Groom GRM Inspection of another individuals body by combing
through the hair with the fingers or lips, either mutual
or one-sided
Handhold HH Two individuals grasping hands
Handclasp HC During grooming, holding hands above the head to
mutually groom the underarm area
Head bob HB Moving the head up and down, typically in a play or
greeting context
Inspect INSP Visual or physical inspect of a females swelling, often
accompanied by finger insertion and smelling
Kiss KISS Pressing of lips onto another individual
Leaf-groom LGRM Using the lips, placing of a parasite or debris removed
from the body onto a leaf for visual inspection
Lick LCK Using the tongue to lick
Mount MT One individual bipedally stands behind another while
embracing them from behind or partially climbing onto
their lower back
Offer OFF One individual gives an item or initiates an affiliative
behavior with another
Open arms OAR Arms expanded horizontally to signal an embrace
Patrol PAT Individuals (typically males) moving as a unit along
their territory boundary in search of neighboring
community intruders

200
Play PL Individuals engaged in playful behaviors, often
accompanied by a play face and play pant. Distinction
made between social, self, and object play
Present PRE Quadrupedal stance with rump pointed toward
dominant individual. Females often present to males
for swelling inspection
Protect PRO One individual providing aid to another in a socially
dangerous or stressful situation. Mothers protect their
dependents
Rump-rump
rub
RRR Two individuals each present their rumps and rapidly
rub them together, typically used in the context of
reassurance
Self-groom SG Inspection of self using hands or mouth to remove
debris and parasites from body
Soft bite SB One individual provides a gentle open mouth bite on a
body part of another
Touch TCH One individual makes intentional contact with another
individual using their hand or foot
Toy play TOY Playful behaviors used when interacting with an object
(i.e., stick). Most common in infants and juveniles
Aggressive Behaviors
Arm threat ARM Waving extended arm in an exaggerated motion toward
another individual
Attack ATT One individual making agonistic contact with another
individual
Branch drag BG Grabbing and dragging a branch behind during a
display
Bipedal swagger BIS While in a bipedal stance, swaying from side to side.
Typically marks the start of a display
Bite BIT Using teeth to penetrate the skin of another individual

201
Charge CHG One individual quickly and aggressively moving
toward another
Chase CH One individual in pursuit of another individual as their
flee
Display DSY Rapid, exaggerated movement (often bipedal)
involving charges and branch dragging, breaking, and
throwing
Flee FL One individual retreats from another
Foot stomp FSP Pounding of foot against the ground to produce noise,
often used in displays
Grab GR One individual tightly holding another in an aggressive
manner
Ground slap GSL Hitting the ground with hand, typically used in displays
Intervene INV Third party bystander involving themselves in an
ongoing conflict either to stop the conflict or in support
of an opponent
Joint display JTDSY Multiple individuals cooperatively displaying at the
same time
Jump JMP One individual jumping on top of another individual in
a conflict
Kick KCK One individual contacting another with an extended
foot during aggression
Kill KILL Using physical force to end the life of an individual
Piloerect hair PILO Hair standing on end, typically during high arousal and
displays
Retaliate RET Reacting aggressively to an attack instead of fleeing
Slap SL One individual hitting another with an open hand,
extended arm
Threat TH Non-contact aggression involving charges, chases, arm
threats, etc.

202
Throw THR Launching an object or individual into the air
Wound WND Skin puncture or split, and often bleeding, resulting
from a fight
Facial Expressions
Compressed lip CL Upper lip puffed and pressed hard against lower lip,
often expressed during displays
Droop lip DL Open mouth with lower lip hanging down
Fear grin FG Retraction of lips maximally exposing the upper and
lower teeth, often accompanied by screaming
Play face PF Relaxed face with an open mouth exposed only the
lower teeth, often accompanied by a play pant
Pout face PT Pursed lips slightly open while drooping down, often
accompanied by whimpering
Relaxed face RF Resting, natural face with relaxed brow and lips
Vocalizations and Gesture-Producing Sounds
Alarm call
Wraaa
AC High pitched, long drawn out ‘wraaa’ occurring when
new or dangerous objects are encountered (e.g., snake,
human, etc.)
Breathy pant BP Low rapid breaths taken in succession often used in
greeting
Buttress bang BB Stomping large tree buttresses to make a loud thumping
sound, often used in displays or when pant hooting by
males
Copulation
scream
CSCR Short high pitched scream made by females after
copulation
Food bark FDBK Short barks (can be high or low pitched) made in rapid
succession when entering a feeding tree and when
[initially] feeding
Hoo HOO Soft short successive ‘hoo,’ often used by infants

203
Lipsmack LS Moving lips together to make a smacking or popping
sound when grooming
Pant hoot PH Series of long hooting calls, involving pushing air in
and out of the diaphragm, building until climaxing in a
scream
Pant-grunt PGRT Deep throaty grunt made by subordinates toward
dominant individuals
Play pant PP Breathy pant used during play
Scream SCR High pitched, loud, often prolonged distress
vocalization
Soft grunt SFT GRT Quiet grunt
Waa bark WAA Loud, sharp bark typically made during aggression
Whimper WH Soft, low pitched sounds mimicking crying
Note. For interactive behaviors, identity of focal’s affiliative interaction partner(s) was
always recorded in conjunction with appropriate code. Behavioral definitions described
in Goodall (1986) and Nishida et al. (1999). Facial expression described in Parr et al.
(2005, 2007).

204
Appendix C

Operational Definitions

Variable Type Definition
Opponent Variables
Aggressor sex Binomial Sex of initiator of directed aggression: m =
male; f = female
Victim sex Binomial Sex of recipient of directed aggression: m =
male; f = female
Dyadic sex
composition
Multinomial Sex combination of aggressor and victim:
MM = male-male; MF = male-female; FF =
female-female
Aggressor estrous Ordinal Estrous state of aggressor: 0 = male; 1 =
deflated female; 2 = partially swollen
female; 3 = fully swollen female
Victim estrous Ordinal Estrous state of victim: 0 = male; 1 =
deflated female; 2 = partially swollen
female; 3 = fully swollen female
Aggressor parity Multinomial Parity state of aggressor: M = male, no
parity; N = nulliparous female; P = parous
female
Victim parity Multinomial Parity state of victim: M = male, no parity; N
= nulliparous female; P = parous female
Aggressors rank Ordinal Aggressor’s rank based on linear rank order
of males (1 thru 11): high = rank 1-4; mid =
rank 5-7; low = rank 8-11
Victims rank Ordinal Victim’s rank based on linear rank order of
males (1 thru 11): high = rank 1-4; mid =
rank 5-7; low = rank 8-11
Rank difference Ordinal Degree of rank difference between aggressor
and victim: small = 1-3; medium = 4-7; large
= 8-11

205
Winner rank Multinomial Rank of winner relative to opponent: D =
dominant; S = subordinate; N = no winner; C
= coalitionary win
Kinship Binomial Relatedness of aggressor and victim: Y =
yes; N = no
Relationship Variables
Wittig relative party
association
Binomial F = frequent; R = rare (Wittig & Boesch,
2005)
Cords association Multinomial Calculated for within party, within view,
within 10 m, and within 5 m: A = affiliative;
N = neutral; NA = non-affiliative (Cords &
Aureli, 2000)
Combined association
index (CAI)
Binomial Overall dyadic association of opponents
during baseline conditions by combining
proximity indices of in same party, within 5
meters, and nearest neighbor: Y = yes; N =
no (Gilby & Wrangham, 2008)
Preferred social
partner (PSP)
------ Measures both one-sided (PSP of aggressor
to victim and from victim to aggressor) and
mutual relationships between opponents in
baseline conditions for various valuable
relationship parameters (Gilby &
Wrangham, 2008)
Proximity association
PSP
Binomial Uses the CAI to determine proximity
association PSPs. Calculated one-sided and
mutual: P = PSP; N = neutral partner
Wittig agonistic
support
Binomial Y = yes; N = no (Wittig & Boesch, 2005)
Agonistic support PSP Binomial Determined PSPs for coalitionary support
during aggression in baseline conditions.
Calculated one-sided and mutual PSPs: P =
PSP; N = neutral partner
Arnold baseline
affiliation
Ordinal 1 = absent; 2 = moderate; 3 = strong (Arnold
& Whiten, 2001)

206
Contact affiliation
PSP
Binomial Determined PSPs based on time spent in
affiliative contact (i.e., contact sitting,
grooming, mounting, etc.) between dyads.
Calculated one-sided and mutual PSPs: P =
PSP; N = neutral partner
Overall PSP Binomial Dyad was a PSP on 2 or more PSP
relationship parameters. Calculated one-
sided and mutual Overall PSPs: M =
majority PSP; N = no
Overall PSP count Ordinal Ranked strength of overall PSP by counting
dyad’s PSP status for all PSP relationship
parameters. Calculated one-sided and mutual
Overall “super” PSPs: 0 = dyad shared no
PSPs; 1 = dyad shared one PSP; 2 = dyad
shared two PSPs; 3 = dyad shared three PSPs
Conflict Variables
Aggression bout ------ All aggressive, agonistic, and submissive
behaviors that continuously occur between
opponents
Directed aggression Binomial Aggression directed toward one or more
individuals, may or may not involve contact.
Excludes bluff displays with no intended
target
Renewed aggression Binomial Reinitiated directed aggression between
former opponents within the PC period
Redirected aggression Binomial Redirected aggression towards a third-party
bystander that occurs within the PC period
initiated by either the original aggressor or
victim
Third-party
aggression
Binomial Directed aggression occurring between third-
party bystanders in tandem with the original
bout of aggression
Conflict (fight) ------ Directed aggression between two or more
individuals, may or may not involve contact

207
Aggression type Binomial Directed aggression containing agonistic
contact (AG) or only threats (TH)
Severity Ordinal Degree of severity during conflict: 0 =
threats only; 1 = one instance of agonistic
contact; 2 = multiple instances of agonistic
contact; 3 = multiple instances of prolonged
agonistic contact with possible biting
Wounding Binomial Open wound or blood visible on at least one
opponent: Y = yes; N = no
Directionality Binomial Whether or not victim reciprocates
aggression. Unidirectional (UNI) aggression
= victim does not reciprocate. Bidirectional
(BI) aggression = victim reciprocates.
Location Binomial Initial location of aggression: G = ground; C
= canopy
Wittig context Multinomial Context in which aggression occurs: FOOD
= feeding context; SEX = sexual context;
SOCIAL = social context (Wittig & Boesch,
2005)
Muller context Multinomial Context in which aggression occurs: FD =
plant feeding; MT = eating meat; SEX =
sexual coercion/competition; SOCIAL =
general social; REUNION = party fusion;
PRO = one individual protecting another
(Muller, 2002)
Parous estrous female
count
Continuous Number of parous estrous females present in
party during conflict
Bystander count Continuous Number of third-party bystanders in view
during conflict
Total party size Continuous Number of individuals present in party
during conflict
Leave party Binomial Whether or no either opponent left the party
in the 10-minute period following conflict: Y
= yes; N = no

208
Opponent end
distance
Ordinal Distance between opponents immediately
after the last aggressive action or response
(i.e., flee): 1 = less than 5 meters; 2 =
between 5 and 10 meters; 3 = beyond 10
meters
Aggressor chase Binomial Whether or not the aggressor chased the
victim during the conflict: Y = yes; N = no
Victim flee Binomial Whether or not the victim fled during the
conflict: Y = yes; N = no
Aggressor scream Binomial Whether or not the aggressor screamed
during the conflict: Y = yes; N = no
Victim scream Binomial Whether or not the victim screamed during
the conflict: Y = yes; N = no
Prolonged scream Binomial Whether or not the victim continued
screaming after the conflict had ended: Y =
yes; N = no
Intervention Binomial Whether or not a third-party bystander
intervened in the conflict on behalf of the
victim: Y = yes; N = no
Conflict type Multinomial Indicated whether the conflict was a single
occurrence between opponents, contained
renewed aggression and/or redirected
aggression
Aggression level Determined each individual’s preferred
victims of aggression while controlling for
time spent in the same party. P = preferred
target of aggressor, N = neutral target of
aggressor
Post-Conflict Interactions
Post-conflict (PC)
period
------ 10-minute period immediately following that
last aggressive interaction or response
between opponents

209
Conciliatory behavior ------ Affiliative interactions between opponents or
third-party bystanders during the PC period
Reconciliation bout ------ All affiliative behaviors and interactions that
continuously occur between opponents
during the PC period
Contact reconciliation Binomial First affiliative contact between opponents
during the PC period: Y = yes; N = no
Third party bystander Continuous Identity of party members in view of a
conflict, but not involved in the conflict
Offered (or true)
consolation
Binominal Third party bystander offers affiliative
contact to the victim within the PC period: Y
= yes; N = no
Solicited consolation Binomial Victim solicits affiliative contact with a third
party bystander within the PC period: Y =
yes; N = no
Appeasement Binomial Third party bystander offers affiliative
contact to the aggressor within the PC
period: Y = yes; N = no
Recruitment Binomial Aggressor offers affiliative contact to a third
party bystander within the PC period: Y =
yes; N = no
Conciliatory Variation Variables
Latency Continuous/
Ordinal
Time (in seconds) from the end of the
conflict to the first affiliative interaction
between opponents or third party bystanders
during the PC period: 1 = very short < 5 s; 2
= short between 5-10 s; 3 = long between 11-
20 s; 4 = very long > 20 s
Preconciliatory signal Binomial Whether or not either opponent or bystander
offers an affiliative vocalization or gesture
towards the other when in close proximity (<
1 m) prior to conciliatory behavior: Y = yes;
N = no

210
Approacher Binomial Identity of opponent or bystander that first
approaches (within 1 meter) the other during
the PC period: A = aggressor; V = victim
Initiator Binomial Identity of opponent or bystander that first
initiates reconciliation during the PC period:
A = aggressor; V = victim
Action Binomial Categorization of initial behavior used
during conciliation: Short-momentary
behavior = embrace, mount, touch;
Prolonged-bout behavior = grooming,
contact sitting
Recipient response Binomial Immediate response (within 5 s) of the
recipient of the initial conciliatory action:
ALLOW = allows conciliatory action but
does not reciprocate: RECIPROCATE =
allows conciliatory action and reciprocates
with an affiliative behavior
Conciliatory
directionality
Binomial Whether or no both opponents actively
engage in affiliative behaviors with one
another during the conciliatory bout:
Unidirectional (UNI) = only one individual
gives an affiliative behavior; Bidirectional
(BI) = both individuals give an affiliative
behavior to one another
Complexity Ordinal Number of different behaviors used during a
conciliatory bout: SIMPLE = one affiliative
behavior; COMPLEX = two affiliative
behaviors; ADVANCED = three or more
affiliative behaviors
Conciliatory duration Continuous/
Ordinal
Time elapsed (in seconds) during the
conciliatory bout, from the start of the first
affiliative interaction between opponents or
with bystanders until contact ceases: VS =
very short < 5 s; S = short between 5-10 s; L
= long between 11-20 s; VL = very long >
20 s

Abstract (if available)

Abstract Peacemaking strategies are often necessary to mitigate aggression and maintain group stability in social animals. While this topic of behavior has been extensively studies in captive primate populations, little is known about the post-conflict behaviors of their wild counterparts. In particular, wild chimpanzees have become well-known for their high rates of intragroup aggression relative to other animal species and even humans making them a model candidate in which to study aggression mitigation strategies. Captive chimpanzees, like humans, use reconciliation and consolation as mechanisms to mitigate the negative effects of aggression. However, the fission-fusion social structure of wild chimpanzees makes it difficult to extrapolate captive results to wild populations. In particular, distancing strategies following aggression may play a key role in wild post-conflict behavior. This research investigated the post-conflict behaviors of wild chimpanzees at Kanyawara. In a one-year period, 632 aggression post-conflict observations between 181 dyads were collected and analyzed. ❧ While the Kanyawara chimpanzees exhibited wide variability in directed aggression, there were distinct patterns to their choices during the conflict. In general, male chimpanzees directed aggression most often towards parous females. Parous females directed their aggression most often towards nulliparous females. Conflicts were also influenced by social relationships. Male dyad preferred social partners engaged in more aggression than one-sided and neutral partners, whereas female dyad preferred social partners engaged in less aggression than neutral partners. Mixed-sex dyad partners engaged in more aggression only if they were mutual contact preferred social partners. In general, the chimpanzees appeared to adjust their responses to their opponent based on preceding behaviors within the conflict bout and intrinsic attributes of their opponent, quickly weighing the various costs and benefits of each situation to determine which mechanisms were more effective at managing the ongoing cost-benefit battle. Furthermore, the potential long-term costs to opponent relationships after aggression could be further diminished if conciliatory behaviors followed. ❧ After a fight, opponents can chose whether to reconcile, renew aggression, redirect aggression, or avoid one another. In the course of this study, reconciliation occurred 123 times between 57 dyads. The Kanyawara chimpanzees had a dyadic corrected conciliatory tendency of 14.4% and a group corrected conciliatory tendency of 15.7%. Overall, the Kanyawara chimpanzees had a corrected conciliatory tendency similar to other wild populations (15-19%), but notably lower than their captive counterparts (22-48%). Male-male dyads had the highest conciliatory tendency (23%), followed by male-female (14%) and female-female (3%) dyads. Generalized linear mixed models determined opponent distance was the best predictor of reconciliation. Opponents within 10 meters after the conflict were more likely to reconcile than those beyond 10 meters (p<0.001), indicating distancing strategies influence reconciliation. The next best predictor of reconciliation was the social relationship between opponents. Mutual preferred social partners were significantly more likely to reconcile than neutral partners (p<0.001). If the aggressor was an overall preferred social partner of the victim, this was an even better predictor of reconciliation (p<0.001), providing support for the valuable relationship hypothesis.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

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

PDF

You always say that: physical aggression perpetration, linguistic behavior, and conflict intensity in young adult dating couples

PDF

The influence of gonadal steroid hormones and social environment on sexually differentiated behavior patterns in wild and captive immature chimpanzees

Asset Metadata

Creator Hartel, Jessica Andrea (author)

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

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Integrative and Evolutionary Biology

Publication Date 07/07/2015

Defense Date 01/20/2015

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

Tag aggression,Chimpanzees,conflict resolution,OAI-PMH Harvest,peacemaking,Reconciliation,valuable relationships

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Stanford, Craig B. (committee chair), McNitt-Gray, Jill L. (committee member), Muller, Martin (committee member), Wood, Justin N. (committee member)

Creator Email hartel@usc.edu,hartelj@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-587049

Unique identifier UC11302090

Identifier etd-HartelJess-3553.pdf (filename),usctheses-c3-587049 (legacy record id)

Legacy Identifier etd-HartelJess-3553.pdf

Dmrecord 587049

Document Type Dissertation

Format application/pdf (imt)

Rights Hartel, Jessica Andrea

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