Cebus Capucinus) at Santa Rosa

UNIVERSITY OF CALGARY

Dominance Among Female White-Faced Capuchins (Cebus capucinus) at Santa Rosa

National Park, Costa Rica

Mackenzie Lee Bergstrom

A THESIS

SUBMITTED TO THE FACULTY OF GRADUATE STUDIES

IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE

DEGREE OF MASTER OF ARTS

DEPARTMENT OF ANTHROPOLOGY

CALGARY, ALBERTA

APRIL, 2009

ISBN: 978-0-494-51092-6

UNIVERSITY OF CALGARY

FACULTY OF GRADUATE STUDIES

The undersigned certify that they have read, and recommend to the Faculty of Graduate

Studies for acceptance, a thesis entitled "Dominance Among Female White-faced

Capuchins (Cebus capucinus) at Santa Rosa National Park, Costa Rica" submitted by

Mackenzie Lee Bergstrom in partial fulfilment of the requirements of the degree of

Master of Arts.

Supervisor, Dr. Linda Marie Fedigan, Department of Anthropology

Dr. Pascale Sicotte, Department of Anthropology

Dr. Kathreen E. Ruckstuhl, Department of Biological Sciences

Date

Abstract

Research on Old World primates has provided the foundation for our understanding of behavioral variation in competitive strategies resulting from social and ecological pressures. The aim of this project was to clarify the nature of competitive patterns among female white-faced capuchins, a New World primate. I examined five aspects of dominance: hierarchical linearity, strength, stability, nepotism, and dominance style. Females displayed linear and strong hierarchies in the dry season, but dominance expression was lower in the rainy season. Females formed stable matrilineal hierarchies and quickly acquired rank positions beneath their mother and older sisters upon reaching sexual maturity. Capuchin females also exhibited an intermediate dominance style based on unidirectional aggression, moderate levels of kin bias, and conciliatory behaviour.

These findings enhance our understanding of capuchin social systems and help to establish how the competitive strategies of white-faced capuchins compare to those of

Old World primates.

iii

Acknowledgements

First and foremost, I would like to thank my supervisor, Dr. Linda Fedigan, for providing the knowledge, support, patience, and constructive criticism necessary to guide me through this project. Thank you to Dr. Pascale Sicotte and Dr. Kathreen Ruckstuhl for serving as members of my defense committee and providing valuable feedback. This project would not have been possible without funding from the University of Calgary, and the following external organizations: NSERC (awarded to Dr. Linda Fedigan),

Alberta Ingenuity Fund, International Primatological Society, and Sigma Xi.

Thank you to Dr. Pascale Sicotte and Dr. Mary Pavelka for their instruction in primatological theory and methodology as well as their assistance with the development of my thesis proposal. For instruction in Access, helpful advice with analysis, and the creation of a project-specific data parser I thank Dr. John Addicott, who turned a potential data extraction nightmare into a manageable and organized database. I appreciate all of the help and advice received from the faculty and students within the

Department of Anthropology at the U of C, particularly Kira Delmore, Barb Kowalzik,

Amanda Melin, Eva Wikberg, Julie Teichroeb, and Greg Bridgett for advice and revisions on drafts of my work.

I am grateful to the ACG and staff members Sr. Roger Blanco and Sra. Maria

Luisa for granting me permission to conduct my research at Santa Rosa National Park. I am extremely grateful to Teresa Holmes, my dedicated and hard-working field assistant who stuck with me over the course of both field seasons through rainstorms, wasp and acacia stings, trips, falls, and army ant attacks. Without her positive attitude and

perseverance my field work would not have been as efficient or complete. Amanda

Melin, Nigel Parr, Brandon Klug, Fernando Campos, members of the Tulane capuchin team (Dr. Kathy Jack, Claire Sheller, Valerie Schoof, Zdanna King, and Andrew

Childers), the spider monkey team (Dr. Filippo Aureli, Dr. Colleen Schaffner, Dr. Norber

Asensio and Elvin Murillo Chacon), Lily and Matcha, and our trusty taxi driver and friend, Ronny also helped to make life in the field and at the Albergue enjoyable.

I would also like to acknowledge Dr. Jennifer Mathews, who introduced me to primatology at Trinity University and encouraged me to pursue my interests; Dr. Jeffrey

Rogers, who taught me valuable laboratory skills and extended the job flexibility that allowed me to explore my interest in social behavior at SFBR; Dr. Amanda Vinson, an amazing mentor and friend; and Dr. Susan Perry and Dr. Joseph Manson, who provided me with the opportunity and diligent training to conduct fieldwork on primates. I would not be where I am today without the support of these influential people.

I owe a great deal of thanks to my closest friends, especially, Brandon Klug,

Erin Baerwald, Kira Delmore, Eva Wikberg, Amanda Melin, Barb Kowalzik, Nigel Parr, and Amanda Bannister. Thank you for helping to make Calgary my second home, for providing an outlet through which I was able to gain perspective, spout stress-filled rants, and maintain as well-rounded a grad student lifestyle as possible.

Finally, I would like to thank my family. Mom, Dad, Ryan, Gram, Karen, Lin,

Jim, Kate, Bob, Kris, Bill, John and Erin, you have always shown the upmost support and interest in what I do. I love you all. Mom and Dad, I cannot thank you enough for the tremendous amount of time, effort, and endless support you have put forth to help me achieve my goals.

Table of Contents

Approval Page...... ii Abstract...... iii Acknowledgements...... iv List of Tables ...... ix List of Figures and Illustrations ...... xi

CHAPTER 1. A GENERAL INTRODUCTION TO DOMINANCE...... 1 Background...... 1 The evolution of sociality and group living ...... 1 Social interaction, conflict and contest competition...... 1 The expression and benefits of dominance in social mammals ...... 2 Previous research on primate socioecology ...... 3 Study species: White-faced capuchins (Cebus capucinus)...... 5 Research objectives and questions...... 7 Synopsis...... 8

CHAPTER 2. HIERARCHICAL STRUCTURE: LINEARITY, NEPOTISM, STRENGTH AND STABILITY ...... 10 Introduction...... 10 Rank variability and relatedness...... 11 Rank variability and resources ...... 12 Rank variability among groups ...... 13 Hierarchical linearity...... 13 Nepotism, rank acquisition and matrilineal rank inheritance...... 15 Hierarchical strength ...... 17 Hierarchical stability ...... 18 Research objective and study species...... 20 Research questions and predictions...... 21 Methods ...... 25 Study Site...... 25 Study Subjects ...... 25 Data Collection...... 26 Variables and Analyses ...... 28 Rank order and rate of dominance expression by season and group size...... 28 Hierarchical linearity ...... 28 Rank acquisition and stabilization ...... 29 Rank order patterning and matrilineal rank inheritance ...... 30 Strength...... 31 Stability...... 32 Results...... 33 Can all females be placed into a rank order based on the direction of submission and what is the rate of submission among individuals?...... 33

Is there variation in expression of dominance between seasons and among groups?...... 33 What is the degree of hierarchical linearity?...... 34 How long does it take for a female to acquire a stable rank position within the hierarchy upon reaching sexual maturity? And does a female’s rank early in life predict her position years later?...... 35 Do female ranks follow matrilineal rank inheritance patterning?...... 36 Are hierarchies strong?...... 36 Do a high proportion of agonistic interactions occur in the context of food? ...36 What is the hierarchical latency to detection? ...... 37 How consistent is the direction of fighting within dyads?...... 37 Are ranks stable over time? ...... 37 Discussion...... 38 Linearity and seasonal effects ...... 38 Rank acquisition and stabilization...... 40 Nepotism and matrilineal rank inheritance...... 42 Strength...... 44 Stability...... 45 Conclusions...... 46

CHAPTER 3. DOMINANCE STYLE AMONG FEMALE WHITE-FACED CAPUCHINS...... 61 Introduction...... 61 The socioecological basis for dominance style ...... 61 Behavioral co-variation ...... 62 Dominance style and Old World primates ...... 65 Research objective and study species...... 66 Methods ...... 70 Study Subjects ...... 70 Data Collection...... 71 Variables and Analyses ...... 72 Bidirectionality of Aggression...... 72 Kin Bias ...... 73 Intensity of kin bias...... 74 Maintenance of relationships and rank ...... 74 Results...... 75 Bidirectionality of aggression...... 75 Kin Bias...... 77 Discussion...... 78 Comparison with macaque species...... 81 Conclusion ...... 83

CHAPTER 4. GENERAL DISCUSSION ...... 96 Project summary and synthesis...... 97 Study limitations...... 101

vii

Areas for future research...... 104

REFERENCES ...... 107

APPENDIX 1...... 122

APPENDIX 2...... 126

viii

List of Tables

Table 2.1: Social categories among female primates designated according to competitive regime. Table reproduced from Sterck et al. (1997)...... 48

Table 2.2: The total number of focal hours collected per focal female for LV, CP and GC groups of white-faced capuchins across rainy and dry seasons...... 49

Table 2.3: Group compositions (number of individuals present per age/sex class) during the 2007 and 2008 study periods by group of white-faced capuchins (LV, CP, and GC) and season...... 50

Table 2.4: Identity, date of birth and genealogy information for female white-faced capuchins that were present in the adult female hierarchy for LV or CP group from 1997-2008 (Farah and Patch1, not present during this time period, were included to show a complete record of known matrilineal relationships) at Santa Rosa National Park, Costa Rica...... 51

Table 2.5: Dominance rank of adult female white-faced capuchins and Actor (row) / Recipient (column) matrices constructed using the direction of submissive interactions (avoid, cower, grimace, flee, supplantation (transposed)) for LV group in the rainy (i) vs. dry (ii) seasons...... 52

Table 2.6: Dominance rank of adult female white-faced capuchins and Actor (row) / Recipient (column) matrices constructed using the direction of submissive interactions (avoid, cower, grimace, flee, supplantation (transposed)) for CP group in the rainy (i) vs. dry (ii) seasons...... 53

Table 2.7: Dominance rank of adult female white-faced capuchins and Actor (row) / Recipient (column) matrices constructed using the direction of submissive interactions (avoid, cower, grimace, flee, supplantation (transposed)) for GC group in the rainy (i) vs. dry (ii) seasons...... 54

Table 2.8: Rate of submission, calculated as the number of interactions per focal hour, by group (LV, CP, and GC) and season among female white-faced capuchins...... 55

Table 2.9: The proportion of unknown dyads and Landau's index of linearity (h') with associated p-values by group of white-faced capuchins (LV, CP, and GC) and season...... 56

Table 2.10: Latency to detection (corrected per number of female white-faced capuchins) reported by group (LV, CP, and GC) and calculated as the number of observation hours needed to confidently place all females into the dominance matrix decided using all observation time for the dry season...... 57

Table 2.11: Directional inconsistency index (DII) reported by group of white-faced capuchins (LV, CP, and GC) and calculated as the percentage of submission that was directed in the less frequent direction within dyads...... 58

Table 3.1: Dominance style grading designed for macaque social organization adapted from Thierry (2000) to include additional species...... 85

Table 3.2: Female group composition, age in years (as of May 2007) and maternal relatedness for LV, CP, and GC groups of white-faced capuchins organized according to dry season rank within groups...... 86

Table 3.3: Bidirectionality of aggression among adult female white-faced capuchins across both seasons for LV, CP and GC groups...... 87

Table 3.4: Kin bias among adult female white-faced capuchins over both seasons for CP and LV groups: partial Kr correlation coefficients (p-value) between maternal relatedness coefficients and approach rate, grooming rate, percentage of time spent within 5 meters, and co-feeding rates, while controlling for rank difference...... 88

List of Figures and Illustrations

Figure 2.1: Submission rates (frequency per hour) reported for adult female white- faced capuchins by group (LV, CP, and GC) and season. There was a significant interaction effect between season and group (F(2,21) = 4.083, p = 0.032)...... 59

Figure 2.2: The mean ± SEM percentage of agonistic interactions that occurred in the context of resource versus social contexts for LV, CP and GC groups of white- faced capuchins during the dry season (χ2 (2, N = 3) = 6.44, p = 0.04)...... 60

Figure 3.1. Rates of dyadic aggression (frequency/hour) among female white-faced capuchins are shown for LV (n=5), CP (n=7) and GC (n=10) for rainy and dry seasons...... 89

Figure 3.2. Mean + SEM approach rates (frequency/hour) are shown for kin and non kin adult female white-faced capuchin partner combinations in LV (n = 5) and CP (n = 7) groups for rainy and dry seasons. * p ≤ 0.05, partial Kr test, controlling for rank distance...... 90

Figure 3.3. Mean + SEM groom rates (frequency/hour) are shown for kin and non kin adult female white-faced capuchin partner combinations in LV (n = 5) and CP (n = 7) groups for rainy and dry seasons. * p ≤ 0.05, partial Kr test, controlling for rank distance...... 91

Figure 3.4. Mean + SEM for the % of time spent within 5m (# points samples in proximity/total point samples per dyad) are shown for kin and non kin adult female white-faced capuchin partner combinations in LV (n = 5) and CP (n = 7) groups for rainy and dry seasons. Kin bias was not significant; partial Kr test, controlling for rank distance...... 92

Figure 3.5. Mean + SEM for co-feeding rate (frequency/hour) are shown for kin and non kin adult female white-faced capuchin partner combinations in LV (n = 5) and CP (n = 7) groups for the dry season only. Kin bias was not significant; partial Kr test, controlling for rank distance...... 93

Figure 3.6. Maintenance of proximity scores are shown for each adult female white- faced capuchin dyad (listed as dominant-subordinate) in LV group and compared between seasons (paired t(9) = -2.613, p = 0.028). Scores range from -1 (subordinate individual was responsible for maintaining proximity) to 1 (dominant individual was responsible for maintaining proximity)...... 94

(subordinate individual was responsible for maintaining proximity) to 1 (dominant individual was responsible for maintaining proximity)...... 95

xii 1

CHAPTER 1. A GENERAL INTRODUCTION TO DOMINANCE

Background

The evolution of sociality and group living

The evolution of sociality and group living has been discussed by evolutionary biologists and anthropologists for decades (Alexander, 1974). It is generally accepted that sociality developed and persists among many mammalian species because it lessens predation pressure (i.e., dilution effects, increased overall vigilance, and/or deterrence through mobbing behavior, van Schaik, 1989). The advantages of enhanced food acquisition or cooperative resource defense influence the formation of social groups as well (Alexander, 1974; Wrangham, 1980; van Schaik, 1983). These selective forces are widely believed to drive the formation and maintenance of group living and to enhance individual fitness despite the associated costs of increased intragroup feeding competition and exposure to parasites.

Social interaction, conflict and contest competition

Social interaction is inevitable among group living species. Various types of social organization have evolved as means to cope with the disadvantages of group-living such as increased disease transmission (including parasites) and competition for resources between group members (Alexander, 1974; Freeland, 1976; Dunbar, 1989).

Relationships between individuals form early in development and are based on learning from repeated social interactions. Over time, social relationships stabilize in such a way that individuals can often predict the outcome of an interaction based on previous experiences and proximate social and physical cues (Hinde, 1976; Mason, 1961; Rowell,

1974; Roney & Maestripieri, 2003). This predictive ability is suggested to be particularly important in contest situations. Maynard Smith and Price (1973), proposed that group living species should develop an evolutionary stable strategy (ESS) to avoid whenever possible the escalation of contests over desired resources and to only escalate if an opponent does so. Contests resulting in physical injury to one or both participants are very costly to individuals (Maynard Smith & Price, 1973). Therefore, it has been suggested that it is in the individual’s best interest to assess physical and behavioral cues to determine the resource holding power (i.e., likelihood of winning) of their opponent.

Following this rule, contests are expected to escalate only if 1) the pay-off of winning exceeds loss caused by injury, 2) asymmetry in fighting ability between opponents does not exist or cannot be accurately predicted through physical or behavioral cues or 3) the population is not part of an evolutionary stable strategy (Parker, 1974; Maynard Smith &

Parker, 1976).

The expression and benefits of dominance in social mammals

Kaufmann (1983) broadly defines dominance as “a relationship between two individuals in which one (the subordinate) defers to the other (the dominant) in contest situations.” The formation of dominance relationships through ritualized behavior and the formation of a dominance hierarchy allow the majority of contest situations to be settled without escalation and physical harm to one or more participating individuals.

Some species rely heavily upon physical differences as measures of fighting ability. That is, some animals rely on size (e.g., males vs. females in sexually dimorphic species) or weaponry [e.g., the interlocking of antlers and pushing matches by male red deer (Cervus elaphus), Clutton-Brock et al., 1986] to settle contests and reinforce previously

3 established dominance relationships. Other species, including primates, typically use aggressive and submissive behavioral displays such as supplantations and avoids [e.g., yellow baboons (Papio cynocephalus), Samuels et al., 1987; African elephants

(Loxodonta africana), Archie et al., 2006; nestling birds (Sula nebouxii), Valderrabano-

Ibarra et al., 2007; bighorn ewes (Ovis canadensis), Favre et al., 2008]. Regardless of the means through which dominance is expressed at the proximate level, increased opportunity for social interaction and increased access to resources can ultimately lead to the enhanced fitness of high ranking individuals. A lower amount of received aggression, resulting in lower levels of stress (Sapolsky, 2005) and increased access to higher quality food (reviewed by Koenig, 2002), are two factors thought to increase reproductive success of high ranking individuals. Among primates, high dominance rank has also been shown to be correlated with increased overall fitness, in particular, increased offspring survival, lower age at first birth for daughters, and longer life expectancy [e.g., chimpanzees (Pan troglodytes), Pusey et al., 1997]. Research regarding factors contributing to variation in dominance behavior has provided a greater understanding of the costs and benefits associated with dominance status and how these are related to other aspects of sociality across mammalian taxa (Bernstein, 1981).

Previous research on primate socioecology

As is the case with other social mammals, group-living primates are faced with the dilemma of needing to cooperate to defend resources and infants against other groups, while simultaneously competing over resources within their social group.

Socioecological models based on primate research (although applicable to many other taxa), attempt to explain the development of, and variation in, female social relationships

4 by examining the association between resources and social organization. These socioecological models focus on variation in resource distribution, nepotism vs. individualism, and competitive behavioral responses (Wrangham, 1980; van Schaik,

1983; van Schaik, 1989; Isbell, 1991; Sterck et al., 1997; Isbell & Young, 2002; Koenig,

2002; Koenig & Borries, 2006).

Wrangham’s (1979) research on ape social systems provided foundational research from which other socioecological models developed. He pointed out that there are differences between male and female strategies of achieving maximum inclusive fitness, due to differences in reproductive demands. Accordingly, he noted that females compete mainly over food resources and males compete mainly for access to females as mates. He classified females as bonded (FB) or non-bonded (non-FB) based on relatedness and dispersal patterns but also based on variation in resource distribution and intrasexual competition (Wrangham, 1980). Therefore, Wrangham argued that when desirable resources are clumped and defendable, increased access achieved via organization into bonded (kin) groups will outweigh the cost of competition for those individuals and female philopatry with nepotistic relationships and the formation of dominance hierarchies will result. On the other hand, when females do not face such competition they are expected to lack dominance relationships (i.e., they will be egalitarian) (Wrangham, 1980).

Later models (van Schaik, 1989; Sterck et al., 1997) added that intense within- group competition (WGC) should result in the formation of strong, linear dominance hierarchies with despotic (i.e., strict) social relationships whereas intense WGC in addition to intense between-group competition (BGC) should generate linear hierarchies

5 with increased social tolerance of subordinates by dominant individuals since cooperation is needed in resource defense at the group level. Female philopatric species that exhibit frequent coalitionary activity and grooming also tend to have linear hierarchies (Sterck et al., 1997; Isbell, 1991; van Schaik, 1989). Recent reviews have emphasized the need for additional behavioral measures such as hierarchical strength and stability (Isbell &

Young, 2002) as well as more detailed measures of resource size, distribution and usurpability, ranging patterns, and more direct measures of individual fitness (Isbell &

Young, 2002; Koenig, 2002; Koenig & Borries, 2006) to aid in species categorization and comparison. Studies have begun to address issues regarding dominance and behavioral variation in detail in Old World primate species, focusing on cercopithecine species.

However, studies of dominance hierarchies and dominance style are scarce among New

World primates. White-faced capuchins have similar social structure and behavioral patterns to cercopithecines although differences in factors such as female transfer and group size make capuchins an interesting species with which to compare our knowledge of competitive strategies and socioecology among Old World primates.

Study species: White-faced capuchins (Cebus capucinus)

Since socioecological models are fundamentally based on female relationships and competition (males compete for females and females compete for resources) and since previous studies related to dominance patterns focus on female relationships, I chose to focus my study on adult female white-faced capuchins, a New World primate.

White-faced capuchins are gregarious New World monkeys (Platyrrhini) and belong to the family Cebidae. Their distribution ranges from northern Honduras to

6 northern Ecuador (Fragaszy et al., 2004). Their diet is primarily frugivorous, and they also eat insects and supplement their diet with small vertebrates during the dry season and at times of food scarcity (Rose, 1997). The mating system is polygynandrous and groups are composed of multiple males and females with an average adult sex ratio

(male:female) of 0.72 (Fedigan & Jack, 2001). Average group size ranges between 12 and 27 individuals (Fragaszy et al., 2004). Females are philopatric, and males typically disperse at a mean age of 4.5 years (Jack & Fedigan, 2004).

Clear aggressive and submissive behaviors which are commonly used to determine dominance hierarchies are exhibited in white-faced capuchins, even though formal submission (i.e., a ritualized fear behavior that is spontaneous rather than in response to received aggression) is not as prevalent as it is among baboons and macaques. Rank can easily be determined for the alpha male and female within a Cebus capucinus group, but determining rank order for more subordinate individuals has been difficult in previous studies. This difficulty suggests that dominance hierarchies may not be as linear or expressed as often as those in Old World species, or may differ in other ways than previously suspected (Fragaszy et al., 2004). Examination of additional dominance characteristics such as hierarchical structure and dominance assertion may help to clarify dominance relationships among female white-faced capuchins.

White-faced capuchins provide an ideal opportunity to explore aspects of dominance. As a highly gregarious species living in multi-male, multi-female social groups, they share a variety of behavioral patterns with Old World cercopithecine species, including female philopatry, male dispersal, and characteristics of female-bonded species (i.e., frequent female-female social interactions involving proximity, grooming,

7 alloparenting, aggression, and coalition formation, Jack, 2007). These similarities between capuchins and cercopithecines in female social structure and behavior may contribute to similar development and expression of dominance relationships. Thus, many of the traditional measures used to study dominance in Old World monkey species are applicable to white-faced capuchins and results should be comparable across species as well.

Research objectives and questions

The primary goal of my research was to clarify the type of dominance relationships present among adult female white-faced capuchins. Using non-invasive observational collection of social behavior data, I investigated the following research topics and questions regarding dominance characteristics:

1) Hierarchical linearity: Can white-faced capuchins be ranked in a linear order?

2) Rank acquisition and stabilization: When do rank acquisition and stabilization within

the adult hierarchy occur?

3) Rank order patterning and matrilineal rank inheritance: How do the presence of kin

and the occurrence of nepotistic behavior affect rank order patterning? Do females

follow the rules of matrilineal rank inheritance?

4) Hierarchical strength: Are hierarchies strong?

5) Hierarchical stability: Are ranks stable over time?

6) Dominance style: Where do female white-faced capuchins fall on the dominance style

continuum?

Synopsis

These research questions are addressed in detail in the following two data chapters. Chapter 2 focuses on hierarchical structure. I determined a rank order among females within each group based on dyadic interactions involving submission and assessed the degree of hierarchical linearity using Landau’s index. I also investigated variation in the rate of submission according to group size and season. Using long-term data, I examined rank acquisition among young adult females and assessed the impact of maternal relatedness on rank order to determine if female capuchins display the intense matrilineal rank inheritance patterns exhibited by cercopithecine females. I used three measures (context of agonistic interactions, latency to hierarchical detection, and the directional consistency of submission within dyads) to assess hierarchical strength. And finally, I used changes in female group membership, the rate of rank change among females, and the number of tied dominance relationships as indicators of hierarchical stability.

In Chapter 3, I used rank order results from Chapter 2 in conjunction with behavioral data to assess the type of dominance style exhibited among females. I used dyadic interactions to calculate three measures of bidirectionality of aggression

(directional inconsistency index, dyads up index, and counteraggression) and the degree of kin bias in social behavior (approach, grooming, proximity, and co-feeding). I also drew from previous research on conciliatory behavior in capuchins to estimate the level of tension-reducing behaviors exhibited by white-faced capuchins relative to macaque species. I attempted to compare my findings regarding behavioral patterns among white-

9 faced capuchins to a previously published comparative dominance style framework constructed for macaque species that used analogous behavioral measures.

Finally, in Chapter 4, I discussed the limitations of my study as well as areas of future research that may clarify dominance patterns seen among female white-faced capuchins. Ideas from both data chapters are integrated to allow for conceptualization of capuchin dominance in terms of socioecological theory and dominance concepts.

CHAPTER 2. HIERARCHICAL STRUCTURE: LINEARITY, NEPOTISM,

STRENGTH AND STABILITY

Introduction

Dominance is a fundamental aspect of social organization for gregarious mammalian species and is of particular importance to primates because they maintain structured social groups outside of the mating season. As in most mammalian species, competitive strategies for female primates are directly tied to food resources, and resource acquisition may be enhanced by the presence of related females who behave cooperatively (Trivers, 1972; Wrangham, 1980). As a consequence of group living, female primates are faced with the dilemma of needing to cooperatively defend resources against other groups, while simultaneously competing over resources within their social group (Wrangham, 1980). According to socioecological models (Wrangham, 1980; van

Schaik, 1989; Isbell, 1991; Sterck et al., 1997; Koenig, 2002; Isbell, 2004), female dominance relationships are shaped by this “compete versus cooperate” dilemma and can be broadly classified in terms of competitive regimes. These regimes refer to patterned social responses to intra and intergroup competitive pressures. For example, a female dominance system may be categorized as “resident nepotistic” if the females compete directly for food resources within their social group (leading to female residence and kin- based dominance hierarchies), but they lack high between-group competition for food.

The presence of between-group competition for food would lead dominants to become tolerant of subordinates in order to elicit their help in contests with other groups (Sterck

11 et al., 1997). Table 2.1, taken from Sterck et al. (1997), shows the four social categories that arise among female primates from different combinations of competitive pressure.

Although broad categorizations of competitive regimes into types may capture some of the variation seen in primate social systems, dominance relationships can also differ at the group, species, and population levels according to variation in localized (i.e., group-level) genetic, social and ecological factors, which makes modeling and “grand schemes” difficult. Therefore, expansion of the number and scope of relevant studies may reveal additional variation that will help to improve classification, cross-species comparisons and future modeling efforts. In this chapter, I will review key components of dominance (e.g., factors influencing rank variability, linearity, strength, and stability), and present my research project and predictions, and then discuss my findings concerning the structure of dominance relationships among female white-faced capuchins. My investigation of hierarchical rank patterns among female capuchins may provide one window into understanding the evolution of dominance behavior, its link to resource characteristics, and the costs and benefits of being dominant in terms of stress and fitness.

Information produced by my study of dominance behavior may also enhance our understanding of capuchin social systems and competitive patterns as well as help us to document how the competitive strategies of white-faced capuchins compare to those of better known Old World monkeys.

Rank variability and relatedness

Genetic relatedness among females within a social group has been shown to heavily influence dominance hierarchies, since dominance relationships often vary according to the degree of nepotistic behavior expressed among females (Kapsalis, 2004).

In nepotistic hierarchies, there is a relationship between rank order and relatedness such that offspring inherit their mothers’ rank. Individuals are able to increase their direct and inclusive fitness through alliances with their kin in agonistic interactions. Nepotistic behavior can also play a large role in other aspects of dominance including rank acquisition (Kawamura, 1958). Individualistic (i.e., non-nepotistic) hierarchies, on the other hand, more heavily reflect individual attributes like age, weight or weaponry, rather than affiliative relationships tied to kinship and alliances.

Rank variability and resources

Dominance relationships may also vary according to the distribution of resources

(Sterck et al 1997). Groups are expected to form clear dominance hierarchies when there is direct competition that results in a differential ability to capitalize on clumped, high- quality, and usurpable resources. On the other hand, groups are expected to lack decided dominance hierarchies or infrequently assert their dominance rank (and live in

“egalitarian societies”) if they compete only indirectly for ubiquitous resources.

Socioecological models suggest that intense within-group competition (WGC) results in the formation of strong, linear dominance hierarchies with despotic social relationships.

However, intense WGC that occurs along with intense between-group competition

(BGC) results in the formation of linear hierarchies that include increased social tolerance. This is because in the latter case, cooperation is needed for resource defense at the group level (Sterck et al., 1997). Resource characteristics (i.e., distribution, abundance, quality, and type) also vary annually, most drastically for species living in regions with distinct seasonal variation. Seasonal shifts in diet may impact competition and dominance interactions, leading to variation in dominance behaviors by season or an

13 overall behavioral response best suited to annual resource variation. Dominance behavior may also vary seasonally if resource characteristics change in such a way that access is limited at times to only a subset of the individuals in the group.

Rank variability among groups

Although rank in all groups is determined by the same set of selected species- specific behaviors (i.e., agonistic behavior), factors surrounding social relationships and the formation of rank (e.g., group composition and membership, relatedness and age distribution) may differ among groups. This intergroup variation creates unique dynamics influencing rank within each group. To understand the range of variation possible at the species level, it is best to include multiple study groups in rank analyses, whose adult female composition ranges in number of individuals, age, and number of related individuals.

Hierarchical linearity

Traditional methods used to construct dominance hierarchies have focused on building linear hierarchies. This approach generally assumes a transitive rank order based on the outcome of competitive interactions among individuals. However, perfect linearity within the hierarchy is uncommon and organization into a linear order ignores rank reversals, which may be important in understanding social dynamics at the group, population and even species level. Although hierarchies are often described as “linear”, quantification of this measure is needed for comparative purposes. Landau’s index of linearity (h’), corrected for “unknown” relationships (i.e., relationships in which individuals do not have any recorded submissive interactions) or tied relationships (i.e., relationships in which individuals have an equal number of directed submission) is a

14 common statistical measure used to assess the degree to which this order reflects the transitivity of dyadic relationships (de Vries, 1995). Unknown and tied relationships may also be of importance in the overall picture of dominance (Izar et al., 2006).

Often, the linearity of a hierarchy can be related to dispersal patterns and female social behavior since the formation of strong social bonds can help to reinforce and maintain dominance relationships (Chapais, 1992). Female philopatric species that exhibit frequent coalitionary activity and grooming tend to have strongly linear hierarchies (van Schaik, 1989; Isbell, 1991; Sterck et al., 1997). Macaques, baboons and vervets are examples of well-studied Old World monkey (OWM) species that follow these behavioral patterns and are consequently classified as “resident nepotistic”

(Kawamura, 1958; Samuels et al., 1987; Seyfarth, 1980). Exceptions exist, where only weak hierarchies arise despite female philopatry [e.g., patas monkeys (Erythrocebus patas), Isbell et al., 1998; Hanuman langurs (Semnopithecus entellus), Koenig et al. 1998; blue monkeys (Cercopithecus mitis), Cords, 2000], although this occurrence has been linked to the ability of single individuals to usurp food resources (Isbell et al., 1998).

These species are categorized as “resident egalitarian” (Sterck et al., 1997).

Many species of New World monkeys also exhibit behavioral patterns associated with linear hierarchies. Among the Atelines, individuals in groups of most species can be ranked in a linear order (Di Fiore & Campbell, 2007). Dominance relations among squirrel monkeys (Saimiri spp), sister taxa to capuchins, can also be predicted by looking at behavioral patterns that co-occur with or without the presence of dominance hierarchies such as dispersal and female social bonds. Saimiri boliviensis is a female philopatric, male dispersal species, in which females exhibit frequent coalitionary activity

15 and display strong, linear dominance hierarchies in response to competition for clumped resources (Mitchell, 1994; Boinski et al., 2002), whereas female S. oerstedii display weaker social bonds without coalitionary activity, disperse and lack dominance hierarchies since direct resource competition is minimal (Boinski, 1994). However, similar to Old World exceptions, female S. sciureus show strong, linear dominance hierarchies but lack coalitionary activity and strong female bonds, likely due to resource distribution and the ability of single individuals to usurp resources without needing stable coalitionary partners (Boinski et al., 2002).

Nepotism, rank acquisition and matrilineal rank inheritance

There are two different types of hierarchies proposed for rank order patterning among mammalian species. The first type of hierarchy is based on age or size, where individuals may form linear or non-linear hierarchies but receive few fitness benefits from association with kin [e.g., individualistic hierarchies: African elephant (Loxodonta africana), Archie et al. 2006]. The second type of hierarchy is nepotistic, where individuals directly or indirectly gain fitness benefits from associations with kin both in terms of acquiring rank and obtaining desired resources. Rank acquisition and dominance relationships are often greatly influenced by matrilineal descent in OWM species and thus, nepotistic. However, nepotism and individualism are not mutually exclusive characteristics and individual traits such as age or size can influence ranking as well. A primate’s rank is usually dependent upon the mother’s rank from birth until sexual maturity, at which point individual attributes or relatedness and social support help the individual to acquire a position within the adult hierarchy (Kawai, 1958).

Kawamura (1958) proposed the following principles governing matriarchal social ranks among a population of adult female Japanese macaques (Macaca fuscata): 1) daughters rank adjacent to maternal kin, 2) younger sisters rank above older sisters, and

3) mothers rank above daughters. Other OWM populations were thought to also generally follow these principles [e.g., rhesus macaques (Macaca mulatta), Sade, 1967; yellow baboons (Papio cynocephalus), Hausfater et al., 1982; and vervets (Chlorocebus pygerythrus), Horrocks and Hunte, 1983]. However, subsequent research on free-ranging and un-provisioned Japanese macaque populations revealed that contrary to Kawamura’s principles, mother-daughter agonistic support was less frequent and older daughters often ranked above younger siblings (Hill & Okayasu, 1995; Furuichi, 1983). Thus, captive environments and provisioning may create more intense resource competition that fosters a greater amount of support by mothers toward daughters in competitive situations. In groups of wild primates, unrelated or more distantly related individuals may play a larger role in sociality because life expectancy is shorter and the availability of closely related social partners is more erratic. Studies have also reported intragroup variation in matrilineal rank inheritance; higher ranking matrilines have been shown to adhere more closely to Kawamura’s principles than lower ranking matrilines, possibly as a consequence of variation in demographic factors such as group composition or size

(Kutsukake, 2000).

Although hierarchical transitivity, rank acquisition and nepotistic ranking patterns provide an important indication of the type of competitive regime influencing a species, decades of research on dominance in cercopithecines (e.g., macaques and baboons) have revealed a number of additional social patterns among female philopatric species. Social

17 patterns that may be essential for understanding variation in dominance relationships, social structure and fitness among social primates include hierarchical strength and stability.

Hierarchical strength

Strength of the hierarchy is another factor that may be important to understanding social structure and relationships, although it remains unexplored in white-faced capuchins. According to Isbell and Young (2002), there are several key quantifiable measures that indicate hierarchical strength. First, the context of agonism is important.

If a great proportion of interactions involving aggression and/or submission take place within the context of feeding (versus other situations), the hierarchy is probably well defined and strong (Isbell & Young, 2002). Second, the rate of submission and the amount of time it takes (i.e., group observation hours) to record the number of dominance interactions necessary to accurately place individuals into a hierarchy is termed ‘latency to detection’ and has been found to vary among species (Isbell & Young, 2002). A high rate of submission is associated with a short ‘latency to detection’ and indicates a strong hierarchy. The percentage of reversals within a matrix, also referred to as the “directional inconsistency index” (DII), is noted by Isbell and Young (2002) as a third measure of hierarchical strength. Interactions that occur below the diagonal within a constructed dominance matrix are termed dominance “reversals” since they represent aggression against the rank order. Strong hierarchies will exhibit low directional inconsistency index scores (<5%) (Isbell & Young, 2002). The directional inconsistency index has been used in many studies and is very useful for cross-species comparisons [e.g., yellow baboons (Papio cynocephalus), Hausfater et al. 1982; bonnet macaques (Macaca

18 radiata), Silk et al. 1981; vervets (Cercopithecus aethiops) and patas (Erythrocebus patas), Isbell et al., 1998; blue monkeys (Cercopithecus mitis), Cords, 2000].

Hierarchical stability

Finally, temporal stability is a comparative dominance measure (Isbell & Young,

2002). In species where the dominance hierarchy remains stable for relatively long periods of time, subordinate individuals withstand substantial amounts of social stress.

However, higher ranking individuals may also suffer stress during times of hierarchical instability when individuals are pressured to assert their rank (Sapolsky, 2005). The rate of rank change over time is a common measure of hierarchical stability. The proportion of tied relationships that are supported by a large number of submissive interactions may also be a good indicator of stability/instability within the hierarchy (Izar et al., 2006).

Many past studies have used temporal stability as a way to characterize dominance systems. Species classified as “resident nepotistic” tend to have stable hierarchies. Rhesus, pigtail, and captive stumptailed macaques are classified as having stable hierarchies (Bernstein & Williams, 1983; Bernstein, 1969; Rhine et al., 1989).

Chapais et al. (1991) classify Japanese macaque hierarchies as stable and argue that stability in this species can be partly attributed to a lack of destabilizing coalitionary activity by lower ranking individuals against more dominant ones. Hierarchies in savannah baboons have also been consistently classified as stable (Hausfater et al., 1982;

Samuels et al., 1987; Combes & Altmann, 2001).

On the other hand, individualistic hierarchies (i.e., “dominance based on individual ability” rather than nepotism) often occurs in species where resources are more ubiquitous and may only show moderate strength and stability, as shown in research on

19 mountain gorillas (Robbins et al., 2005) using the measures proposed by Isbell and

Young (2002). In species with individualistic hierarchies, rank is more likely to fluctuate over the course of an individual’s lifetime when it is dependant upon intrinsic characteristics that vary temporally, such as age (Robbins et al., 2005). Individual rank fluctuation in addition to lack of cooperative nepotistic support leads to an inconsistent long-term reinforcement of rank. Even species exhibiting more direct competition over resources may undergo periods of rank instability due to fluctuation in resource abundance that can in turn, impact reproductive success during that time (Gouzoules et al., 1982). Therefore, rank reinforcement and stability may not be adaptively important for species reliant upon foods with ubiquitous resource distribution and individualistic hierarchies since a high rank position would not provide the increase in resource acquisition that would lead to increased health and fitness benefits and higher lifetime reproductive success.

O’Brien and Robinson (1993) documented a consistent rank order in wedge- capped capuchin monkeys (Cebus olivaceus) over a 10-year study, which provides evidence supporting hierarchical stability and strength in capuchins. Although previous research has looked at hierarchical stability in white-faced capuchins in terms of rates of rank change, data from various sites show variation. For example, data collected on groups at Lomas Barbudal Biological Reserve indicate stability in the dominance hierarchies, whereas data indicated hierarchical instability at Santa Rosa National Park

(Manson et al., 1999). More research is needed to clarify whether this variation is due to methodological differences, limitations of the data, or true differences in behavior across groups and/or populations.

Research objective and study species

Despite the large body of information that has been acquired over the past 20 years for white-faced capuchins, there is still much to be learned about dominance interactions and hierarchies. My study was designed to clarify the nature of dominance relationships among adult female white-faced capuchins.

White-faced capuchins are female philopatric (Fedigan, 1993) and females are classified as female-bonded because of their high levels of association, including strong coalitionary behavior and nepotistic tendencies, as compared to other age-sex class dyads

(Perry, 1996). Past studies have ranked individuals in a linear order within groups of white-faced capuchins (Fedigan, 1993; Perry, 1996; Manson et al., 1999), and research on other well studied capuchin species (Cebus albifrons, apella and olivaceus) has documented linear rank order (reviewed in Fragaszy et al 2004). However, the degree of linearity is unknown. Past studies of female-bonded species show that higher ranking individuals tend to join fights more often, which continuously re-enforces rank (Veenema et al., 1997). If lower ranking individuals are involved in few agonistic interactions amongst each other, dominance relationships may remain unresolved.

White-faced capuchins often live in seasonal habitats that experience distinct rainy and dry seasons, causing fluctuations in resource characteristics. The capuchin diet is primarily frugivorous and seasonal resource variation may impact the level of direct competition and type of dominance relationships among females (Vogel et al., 2007).

However, resources for capuchins at Santa Rosa National Park, Costa Rica may not cause the intense, localized competition for resources that are created in a provisioned environment [e.g., Japanese macaques (Macaca fuscata), Kawamura, 1958], which seems

21 to foster more intense kin support by mothers toward youngest daughters. Regardless, fruit as a resource is generally considered to be a high-quality and clumped food (Isbell,

1991), and direct competition over resources is common in Cebus capucinus (Vogel,

2005). It is known that rates of triadic (vs. dyadic) aggression tend to be high in capuchins. Individuals often solicit help from social partners to form coalitions when aggressing another individual (Manson et al., 1999; Perry et al., 2004) and this pattern renders detection of dominance relationships among lower ranking individuals difficult

(Fragaszy et al., 2004).

Prior data on stability for white-faced capuchins is inconclusive, although longitudinal analysis of hierarchies show long periods of stability interspersed with short periods of instability (Manson et al., 1999). Fluctuation of female group membership has been shown to directly influence rank position (Chapais, 1985; Marsden, 1968). Changes in the adult female membership of groups can occur either through immigration, emigration, death/disappearance, or with the maturation of young adult females (5 years old) to full adult status (6 years old).

Research questions and predictions

I collected and analyzed behavioral data to investigate the following aspects of dominance: linearity of hierarchies, nepotistic behavior in relation to rank acquisition and rank order, hierarchical strength and rank stability. I formed the following research questions and subsequent predictions based on previous research and on patterns predicted by the socioecological model:

1) Hierarchical linearity: Can white-faced capuchins be ranked in a linear order?

a) Can all females be placed into a rank order based on submissive interactions

and what is the rate of dominance expression among individuals?

I predicted that I would be able to place females in a rank order based on the

direction of submission, although I expected submission rates would be low.

b) Is there variation in the expression of dominance between seasons and among

groups?

I predicted that submissive interactions would vary seasonally, with more linear

and strict enforcement of rank during the dry season, when the diversity and

abundance of fruit is lower. I also expected seasonal variation to have a greater

impact on the dominance hierarchies of larger groups. I predicted the latter

because smaller groups are able to forage simultaneously in smaller areas and

may not need to make as large a change in foraging strategy to reduce competition

or enforce rank in adjusting for seasonal differences in food availability.

c) What is the degree of hierarchical linearity?

I predicted that rank order among adult individuals would be linear but the degree

of hierarchical linearity would be moderate since submission may occur

infrequently among subordinate individuals, making them more difficult to rank.

2) Rank acquisition and stabilization: When do rank acquisition and stabilization

within the adult hierarchy occur?

a) How long does it take for a female to acquire a stable rank position within the

hierarchy upon reaching sexual maturity?

Since female kin are readily available as social partners, coalitionary support is

common and WGC is high (Vogel, 2005; Vogel et al., 2007), I predicted that

female white-faced capuchins would quickly acquire a rank position similar to

that of their kin.

b) Does a female’s rank early in life predict her position years later?

I predicted that females would remain in a stable rank position throughout their

lifetimes, similar to many species of cercopithecines.

3) Rank order patterning and matrilineal rank inheritance: How do the presence of

kin and the occurrence of nepotistic behavior affect rank order patterning? Do

females follow the rules of matrilineal rank inheritance?

a) Do females rank adjacent to kin?

I predicted that females would rank adjacent to close maternal kin.

b) Do daughters rank according to youngest ascendency?

I predicted that rank order would follow the youngest ascendency rule.

c) Do mother’s rank above daughters?

I predicted that mothers would rank above daughters, similarly to species of

cercopithecines.

d) Do daughters occupy similar ranks in their adult lives to those that are or

were occupied by their mothers?

Once a stable rank position was established, I expected daughters to maintain

similar rank positions in their adult lives to those occupied by their mothers.

e) If not, is there another pattern to the observed rank order?

I did not expect other factors to influence rank order.

4) Hierarchical strength: Are hierarchies strong?

a) Do a high proportion of agonistic interactions occur in the context of food?

I predicted that a high proportion of aggression and submission would occur in

the context of feeding (Vogel, 2005), a pattern that is characteristic of a strong

hierarchy.

b) What is the hierarchical latency to detection?

I predicted that latency to detection would be long due to the low rates of dyadic

(vs. triadic aggression), characteristic of white-faced capuchins (Manson et al.,

1999), a pattern that is characteristic of a strong hierarchy.

c) How consistent is the direction of fighting within dyads?

I predicted that the presence of coalitionary activity in female relationships would

increase directional inconsistency in dyadic agonistic situations, a characteristic of

a weak hierarchy.

5) Hierarchical stability: Are ranks stable over time?

a) How often does female group membership fluctuate?

I predicted that female group membership would remain stable for long periods of

time, since female immigration, emigration, and death/disappearance are

relatively uncommon.

b) What is the rate of rank change over time?

I predicted that dyadic relationships would remain constant and stable unless large

changes in female group membership occurred.

c) Are there a large proportion of tied dyadic relationships?

I predicted that long-term rank consistency is likely and would result in a low

proportion of tied relationships.

Methods

Study Site

I conducted my research at Santa Rosa National Park (SRNP) in the Area de

Conservación Guanacaste (ACG), Guanacaste Province, Costa Rica. This area consists of 108 square kilometres of seasonal tropical dry deciduous forest (Fedigan & Jack,

2001). Average annual rainfall is 1472 mm (Fedigan & Jack, 2001), with a dry season extending from mid-December until mid-May. Three primate species coexist at SRNP: white-faced capuchin monkeys (Cebus capucinus), mantled howler monkeys (Alouatta palliata) and black-handed spider monkeys (Ateles geoffroyi).

Study Subjects

Data were collected on adult females (≥ 5 years) from three study groups: Los

Valles (LV), Cerco de Piedra (CP), and Guanacaste (GC). Individuals were distinguished based on age, sex, body size, and distinguishing features such as facial structure, skin pigmentation, hair coloration and scars. The number of adult female study subjects for each group was 5, 7, and 10, respectively, totalling 22 individuals over both field seasons

(Table 2.2). Males were not included in this study since socioecological theory suggests that males and females compete for different resources and therefore may have different dominance structures. Juveniles and infants (< 5 years old) were also excluded, as they are too young to have established dominance within the group and their ranks were likely dependent on their mothers’ (Kawai, 1958). All study groups are well-habituated; CP has

26 been studied continuously since 1986, LV since 1990 and GC since 2007. Group size, including all age/sex classes, fluctuated over the two study periods; the total number of individuals present in each group ranged from 16-19 in LV, 18-22 in CP, and 23-33 in

GC (see Table 2.3 for group compositions). The extent of available long-term data on

LV and CP and the variation in number of females among the LV, CP and GC groups allows for comparative study of long-term stability and the effects of group size on social behavior. Dominance hierarchies have been constructed on an annual basis for LV and

CP over the past 19 and 22 years respectively and these are used in my analyses. One caveat for the use of these long-term data is that although only dyadic agonistic behavioral data was included, the exact set of behaviors used to construct each hierarchy in past years may have fluctuated annually, and may not have included all the behaviors used in my construction of dominance hierarchies (e.g., supplantation, cower, avoid, grimace, flee).

Data Collection

A field assistant and I collected data over two field seasons to control for behavioral differences that might result from seasonal variation. The first field season occurred during 3.5 months of the early rainy season between May and August 2007 (22 focal individuals, 127.85 focal hours, 345.57 h observational contact). The second field season occurred during 4.5 months of the dry season between January and May 2008 (22 focal individuals, 331 focal hours; 714.27 h observational contact). Groups were followed on a rotational basis approximately one week per group (varying slightly with group size; i.e., larger groups were followed longer than smaller groups to achieve similar focal data hours per individual) for three weeks per month. Behavioral data were

27 collected approximately five days per week, 10-12 hours per day, using a PSION

Workabout hand-held computer. In addition, a digital voice recorder was used and later transcribed when interactions occurred too rapidly to be collected on PSIONs, such as in the case of intense aggression. I primarily watched the focal individual and dictated behaviors while my field assistant recorded them into the PSION, although roles occasionally reversed. The individual recording also watched the focal subject and disagreements regarding dictation of behaviors were discussed to ensure agreement and data consistency between observers.

I collected 10-minute continuous-time focal animal samples (Altmann, 1974) to record all behavioral states and events as well as interactants involving the focal female.

Instantaneous scan samples were recorded at 2.5 minute intervals during focal follows

(Altmann, 1974). Scans included the activity of the focal animal as well as all individuals within close proximity (i.e., in contact, <1 body length, 1-5 body lengths, and

5-10 body lengths; (Appendix 2) (Perry, 1996). Ad libitum data were collected opportunistically for vocalizations (e.g., signalling fights or predators), intragroup agonism and intergroup encounters.

Focal individuals were selected randomly and sampled as equally as possible each day. Following Perry (1996), the first available focal was selected according to the following rotational rules: a) each focal but one must be sampled before moving to the next round of sampling; b) one focal can be skipped within a round until located if they cannot be found after sufficient search effort or are suspected to be absent from the group; c) a focal cannot be sampled if they interacted with the previous focal during the last 5 minutes of the previous follow to ensure sampling independence; and d) a focal

28 cannot be sampled if they were within 5m of the previous focal in the last scan sample of the previous follow. All social behaviors that involved the focal (see Appendix 1 for ethogram), approaches and leaves, and IDs of interactants were recorded during focal sessions.

Variables and Analyses

Rank order and rate of dominance expression by season and group size

I determined a rank order based on the direction of agonistic dominance

interactions among dyads. A dominance interaction entailed submissive behavior (i.e.,

supplantation, avoid, cower, flee and grimace behaviors) or a submissive behavior

immediately following aggression from another individual. These behaviors were selected because they are unidirectional within dyadic agonistic behavioral sequences. I constructed dominance matrices, including all focal females, for each group based on the outcome of these interactions for the rainy season and dry season. I also calculated rates

of submission by season as the total number of interactions per focal hour for each group

and compiled them monthly to determine if the rate varied by season and/or group size. I

used a one-way ANOVA using Tukey’s post-hoc test for these analyses.

Hierarchical linearity

I performed a dominance matrix analysis on constructed dominance matrices for

the wet and dry seasons and determined the degree of hierarchical linearity using

Landau’s modified index of linearity (h’) and Noldus MatMan 1.0 (Noldus Information

Technology, 1998). This measure is a modified version of Kendall’s coefficient of

linearity (Appleby, 1983), but has been corrected for unknown relationships and is based

on a randomization test using 10,000 randomizations (de Vries, 1995). Landau’s

29 linearity index (h) was calculated using a square matrix of dyadic submissive interactions according to the following formula where N is the number of females in the group, S is the number of females dominated by individual i and unknown and tied dyads are assigned a value of ½ (Appleby, 1983):

2 12 N ⎛ 1 ⎞ h = NS −− 1 3 ∑⎜ ()⎟ − NN i=1 ⎝ 2 ⎠

The index (h) was then corrected for unknown and tied relationships using the following formula to yield Landau’s modified linearity index (h’) where μ is the number of unknown dyads and N is the number of females (de Vries, 1995):

6u 'hh += ()3 − NN

Landau’s index varies from 1 (complete linearity) to 0 (absence of linearity). I interpreted values along this continuum according to the following categorization for degree of linearity: weak (0-0.5), moderate (0.5-0.8) and strong (0.8-1.0). I calculated the percentage of unknown relationships as the number of F-F dyads for which there were no interactions divided by the total number of F-F dyads for that group.

Rank acquisition and stabilization

I determined the amount of time it takes for a female to acquire a stable rank position within the hierarchy using long-term data on LV (1990-2008) and CP groups

(1986-2008). It is difficult to precisely define a stable rank due to continuous fluctuation in female group membership. For example, a female may move from second ranking to top ranking within a year simply due to the disappearance of the alpha female rather than a true challenge of rank. To help account for this, I considered the latency to adult rank

30 stabilization as the amount of time between a female’s entry into the hierarchy at 5 years of age until the first year of a time period where rank position is maintained

(unchallenged or challenged without resulting fluctuation) for ≥ 3 years. I then determined the strength of correlation between this rank (calculated as the percentage of individuals dominated) and the rank in the final year a female was studied, using a

Spearman’s rho correlation analysis to see if the female’s earliest stable rank in adulthood predicted her rank later in life.

Rank order patterning and matrilineal rank inheritance

To analyze matrilineal rank inheritance, I used long-term data collected on LV and CP groups from 1997-2008, when maternal relatedness for all females within the study groups was known, excluding immigrants (see Table 2.4 for genealogies). To analyze rank order patterning, I determined the proportion of sister-sister dyads for which the older sisters outranked younger sisters and mother-daughter dyads for which the mother outranked her daughter. I determined if a female occupied a similar rank during adulthood to that occupied by her mother by determining the strength of correlation between a daughter’s rank (percentage of individuals dominated) and the mother’s rank using a Spearman’s rho correlation analysis. Both the mother’s and daughter’s rank at the age of 9 years were used in this calculation since: 1) a female’s rank is probably well established at this age since the female has likely given birth to at least one infant, 2) it is unlikely for a female to have fallen in rank as a result of aging at this age and 3) this age provided the largest dataset of mother-daughter dyads with known maternal relatedness during the observation period.

Strength

A strong hierarchy is one that is regularly reflected through submissive interactions. Isbell and Young (2002) concluded that the occurrence of agonism during feeding is common among species with strong dominance hierarchies. Therefore, as an indicator of hierarchical strength, I measured the percentage of dyadic submission that took place in the context of feeding out of all occurrences of agonism interactions for which the context could be determined. Due to increased visibility during the dry season,

I recorded the context of agonistic interactions during this season as either “resource related” (i.e., aggression or submission occurring at a water source or in a tree while the group was feeding on a food item such as fruit, insects, or vertebrates) or as “social” (i.e., aggression or submission occurring while the group was resting or socializing in the absence of a contestable food resource). I also measured the hierarchical ‘latency to detection’ using dry season data only (due to limited hours during the rainy season) by determining the total hours of observation needed to place females confidently into a dominance hierarchy (Isbell et al., 1998). Koenig and Borries (2006) demonstrated that group size affects the detection times for hierarchies and must also be considered in determining strength of hierarchy. Following Isbell and Young (2002), I accounted for group size by dividing the observation hours by the number of females in each group.

This produces an average latency time per female for each study group (Isbell et al.,

1998). Short latencies to detection are characteristic of groups with strongly enforced dominance hierarchies. Finally, I used the directional inconsistency index to test the asymmetry of agonism; that is, the percentage of all interactions that were directed in the less frequent direction within dyads (de Waal, 1977). This measure was proposed as a

32 good indicator of hierarchical strength that also indicates stability (Isbell & Young,

2002).

Stability

I assessed temporal stability via analysis of annual dominance hierarchies collected on LV and CP groups over 19 and 24 consecutive years, respectively. From these longitudinal data, I calculated rates of change in female membership as 1) Female

Entry = E/Y, where E is the number of times females entered the group and Y is the number of total years the group was studied and 2) Female Departure = D/Y, where D is the number of times females left the group and Y is the number of total years the group was studied. To assess the stability of female ranks, I calculated the average annual rank change per female (Cheney et al., 1988), for females that were present in the group for more than one year according to the following formula: Rank Change = N/Y, where N is the total number of rank changes for an individual female and Y is the total number of years the female was present in the group. Then I calculated an overall mean including all females. I considered a rank change as an “active” movement up or down the hierarchy (i.e., resulting from challenge by an adjacent or new individual rather than the outcome of a disappearance/death or a shift up due to an individual entering the hierarchy below without known dispute). Finally, I assessed the proportion of dyadic relationships that are tied but supported by a high number of submissive interactions. This is an indicator of hierarchical instability since higher proportions might indicate contest for rank positions (i.e., undecided relationships) (Izar et al., 2006). I considered all results of statistical tests significant at p ≤ 0.05.

Results

Can all females be placed into a rank order based on the direction of submission and what is the rate of submission among individuals?

All females were placed into a rank order either above or below another female without discrepancies. Dominance interactions consisted of submissive behaviors or aggression followed by clear submissive behaviors. Avoids accounted for the majority of recorded dominance interactions (40%). Cowers, supplantations, and flees were less common (16.4%, 17.6% and 22.4%, respectively) and grimaces occurred infrequently

(3%). Actor/recipient matrices (Tables 2.5, 2.6, and 2.7) show the frequency of submission for each study group by season as well as the resulting rank order among adult females. I recorded 39 interactions involving submission during 345.57 hours of observation in the rainy season (LV: 7 interactions/99.23 hours; CP: 11/109.57; GC:

21/136.77) and 296 interactions during 714.27 hours of observation in the dry season

(LV: 24 interactions/191.6 hours; CP: 117/228.13; GC: 155/294.53). The average overall rate of submission across groups was 0.21 interactions per focal hour in the rainy season

(LV: 0.16 (5/31); CP: 0.19 (8/42); GC: 0.27 (15/54.83)) and 0.53 in the dry season (LV:

0.21 (17/79.2); CP: 0.74 (83/112); GC: 0.63 (88/140)) (Table 2.8).

Is there variation in expression of dominance between seasons and among groups?

I used a one-way ANOVA to determine if submission rate varied by season and/or group. I compared monthy submission rates per group during the rainy season with monthly rates during the dry season. Rates of submission differed significantly between seasons, with lower submission rates in the rainy season (F(1,21) = 33.253, p <

0.000). Rates of submission also differed significantly among groups (F(2,21) = 10.268,

34 p = 0.001). Tukey’s post-hoc comparison of the three groups indicated that the smallest group, LV, (M = 0.166) had significantly lower rates of submission than either CP (M =

0.499, p = 0.001) or GC (M = 0.454, p = 0.002) and that rates of submission were not significantly different between CP and GC groups. The interaction effect between season and group was significant (F(2,21) = 4.083, p = 0.032) indicating that the influence of season on submission rate was greater in the two larger groups (CP and GC) than in LV, as shown in Figure 2.1.

What is the degree of hierarchical linearity?

During the rainy season, Landau’s linearity index (h’) scores were relatively low for LV, CP and GC groups and hierarchies were not significantly linear (Table 2.9).

During this period, the percentage of unknown relationships was high for all three groups as well.

During the dry season, Landau’s linearity index (h’) scores were higher and significant values suggest linear hierarchies for CP and GC groups (Table 2.9). Although

Landau’s linearity index was high for LV group in dry season, h’ was insignificant (p =

0.113). This is probably because the small sample size of five individuals for LV limits the power of linearity analysis since six individuals are required to achieve statistical significance and to eliminate the mathematical possibility that the determined hierarchy was obtained by chance (Appleby, 1983). However, unlike the rainy season, the percentage of unknown relationships was very low for all three groups in the dry season.

How long does it take for a female to acquire a stable rank position within the hierarchy upon reaching sexual maturity? And does a female’s rank early in life predict her position years later?

Over the 19 and 23 year study periods covered by our database, a total of 33 adult females have been observed in LV and CP groups. Seven adult females entered the hierarchies of these groups through immigration and 15 females entered the hierarchies via maturation to adult status. Only two of the seven immigrant females remained in the study groups for more than three years (Feisty: six years and Tuft: four years) making it difficult to determine dominance patterns among this category of females. However, both immigrant females dropped to the lowest rank within one year of entering into their respective groups and held a stable lowest-rank position for the remainder of their time in the group, suggesting that lack of related allies may be important in ascending the hierarchy. It is possible that the immigrant females who stayed for less than three years did not have the necessary support to successfully compete amongst the natal females and acquire a stable rank position within the hierarchy. Of the 15 individuals who matured into adults in their natal groups, two (CA and MA) were only present for one year, three

(SA, ED, and KI) were included in the hierarchy for less than three years, and three (SL,

TI, and ZA) have not yet maintained a stable rank order for a period of more than two years. For the seven remaining females (KL, DL, CH, SE, NY, PU, SI), the mean latency to adult rank stabilization equalled 1.14 ± 0.90 years (range = immediate acquisition to three years). Following Pusey et al.’s (1997) approach, I investigated whether a female’s early rank predicted her rank later in adulthood, using females that had been studied at least three years. I found a strong correlation between the stabilized rank a female

36 acquired early in her adult life and her position in the final year she was studied

(Spearman’s rho = 0.887, n=7, p = 0.008, mean # of years observed = 8.7).

Do female ranks follow matrilineal rank inheritance patterning?

In 8 of 8 (100%) sister-sister dyads, older sisters outranked younger sisters for all years where they were present together in the hierarchy with the exception of one dyad during one year (ZA ranked above TI her first year in the hierarchy but TI’s rank was not yet considered stable). Of the five mother-daughter dyads present during this time period, mothers outranked daughters in one dyad (BB-CH), ranks of both mother and daughter were considered unstable for two dyads (BB-SL and SE-ZA), and a rank reversal occurred where the daughter overtook the mother’s position in two dyads (LI-

NY and SE-TI). In both cases of rank reversal, mothers were at least 14 years old. LI

(~20 years old) disappeared the year after NY overtook the alpha female position. There was a strong correlation between the daughter’s rank and her mother’s rank at the same age (nine years old) (Spearman’s rho = 0.662, n=10, p = 0.037). Only 5 of 15 (33.3%) mothers were still present in the hierarchy when their daughters acquired a rank position.

Are hierarchies strong?

Do a high proportion of agonistic interactions occur in the context of food?

Figure 2.2 shows the percentage of interactions involving aggression and/or submission that were either resource-related or social of the 230 interactions with known context that were recorded for LV, CP and GC groups during the dry season. A chi- squared test showed a significant relationship between the percentage of agonism and its context – indicating strong hierarchies; more agonistic interactions occurred in resource- related contexts versus a social contexts, χ2 (2, N = 3) = 6.44, p = 0.04.

What is the hierarchical latency to detection?

The number of observational hours needed to place females confidently into a dominance hierarchy increased with group size (Table 2.10). When latency to detection times were corrected for the number of females, I found a consistent value across groups of 21 hours/female (M = 21.02, SD = 1.08).

How consistent is the direction of fighting within dyads?

For LV, of the dyads for which interactions occurred (60% rainy, 100% dry), all were unidirectional relationships and the higher ranked individual always directed aggression (i.e., supplantation) or received submission from the lower ranked individual

(DII = 0%). For CP, a greater proportion of relationships were known in the dry versus rainy season (42.86% rainy, 95.24% dry). Of the known relationships, all were unidirectional, and there were no reversals against the hierarchy (DII = 0%) during the rainy season. However, in the dry season, 20% were bidirectional (4 of 21 dyads: lower ranking individuals) resulting in an increased DII of 4.27%. For GC, of the 35.56% known dyads in the rainy season, all were one-way and lacked reversals (DII = 0%). In the dry season 80% of the dyads were known, and 5.56% of relationships were bidirectional (2 of 45 dyads: mid-ranking individuals including LL, who was common to both dyads) resulting in a DII score of 1.94% (Table 2.11).

Are ranks stable over time?

The rate that females enter the adult hierarchy/group each year through emigration, death/disappearance, or maturation to adult status based on long-term data from LV and CP (19 and 23 years, respectively) is 0.9 females/year and the rate they leave is 1.1 females/year. I determined that rank change per female for LV was 0.287

(ranging from 0 to 0.632, N=10) and for CP it was 0.510 (ranging from 0 to 1.333,

N=17). Thus, the overall annual rank change per female was 0.427. At this rate, rank change per female as a result of active challenge is expected every 2.34 years. None of the dyadic relationships across all three study groups were tied, providing evidence for stable relationships among females.

Discussion

Linearity and seasonal effects

My results indicate that during the dry season, female capuchin hierarchies are linear for all three study groups. The relatively low percentage of unknown relationships during the dry season likely increased the reliability of constructed rank orders in this season since a greater proportion of the relationships are known and supported by multiple interactions. Preliminary data indicate that during the dry season a greater proportion of time was spent foraging on/competing for clumped resources than during the rainy season (Bergstrom, unpubl.), which may explain the increase in submissive behavior exhibited by subordinate females towards dominant females during this season.

The linearity results I found for this population are consistent with those found for other female philopatric species that exhibit similar patterns of social behavior such as high levels of association, strong coalitionary behavior and nepotistic tendencies. Macaques, baboons and vervets are examples of highly studied Old World monkey species that exhibit similar behavioral patterns as well as linear hierarchies (Kawamura, 1958;

Samuels et al., 1987; Seyfarth, 1980). My Landau’s h’ scores were slightly higher than those reported for white-faced capuchin groups with the same or similar numbers of

39 females during comparable seasonal conditions at Lomas Barbudal, a neighboring field site (h’ = 0.5, p = 0.13; Perry et al. 2008). Data from my study groups (CP and GC) suggest that slight hierarchical instability for dyads with low rank disparity between individuals may explain the slight drop in the degree of linearity from perfect transitivity

(h’ = 1.0). In CP, 4 of 5 (80%) triangular submissive interactions (i.e., interactions that included submission of a higher ranking female towards a lower ranking female) included

KI, who has not yet established a set rank position, and all triangular submissive interactions (five in CP and three in GC) were between females separated by two or fewer rank positions. Between-site variability may simply reflect differences in group composition and relationship dynamics within “low rank disparity” dyads, or from a socioecological perspective, may indicate slight differences in social response to competition and food availability between the two field sites.

I also found a seasonal effect on submission rates. It is likely that seasonal variation in food availability affects the types of resources on which capuchins forage.

The scan samples that I collected during my focal animal follows indicate that a greater proportion of time is spent during the rainy season foraging on ubiquitous resources (e.g., insects and pith) versus more clumped resources (e.g., fruit, flowers, vertebrates or localized water sources). Dispersed and abundant resources could increase group spread and reduce direct competition and the occurrence of submission during the rainy season

(Koenig & Borries, 2006). It is also important to note that the high proportion of dyads in all three groups that failed to exhibit submission during the rainy season likely contributed to difficulty in the construction of reliable rank orders for this time period.

That being said, season affected the smallest group (LV) the least. Rank order was

40 consistent across both seasons for LV and only 1 of 21 dyads (4.76%) changed rank positions in CP indicating that rank order may be reliable for these smaller groups even with limited interactions and known relationships. These groups may be small enough that 1) there is not the pressure to compete as often since the abundance of localized resources may satisfy energetic demands for all individuals and/or 2) the limited number of available social partners makes relationships more predictable, thus reducing the occurrence of direct competition and the need to assert dominance. Indeed, the interaction effect found between season and group on submission rate supports this idea since the increase in submission rate was much greater for the two larger groups (CP and

GC) from the rainy to dry season. The low linearity scores and lack of significance during the rainy season could also have been a result of insufficient data. It is possible that low submission rates and reduced focal hours due to environmental conditions restricted the number of interactions below that which is needed to produce linear hierarchies. On the other hand, factors such as group spread or great rank disparity might cause a true lack of interaction among particular dyads under conditions during this season. More research regarding food availability and distribution, ranging, foraging patterns, and group spread is needed to support these ideas.

Rank acquisition and stabilization

Although very little is known about the transition from adolescence to adulthood in white-faced capuchins, an individual’s rank seems to be dependent upon her mother’s rank until she is sexually mature. As females reach sexual maturity around 5 years of age

(Carnegie et al., 2005; Fedigan & Rose, 1995; Fragaszy et al., 2004), they begin to acquire a position in the dominance hierarchy among other adult females. This process

41 takes approximately one year and the resulting position is likely to remain constant throughout the female’s lifetime/residency within the group, suggesting that initial rank acquisition is an important process and reflects more than immediate social relationships.

This pattern of rank acquisition and stability has important implications for health and fitness. If higher ranking females gain priority of access to usurpable resources, consistency in dominance rank over time will result in a greater impact of dominance on lifetime fitness (reviewed by Fedigan 1983). Certainly, rank order in many species of

OWM [e.g., baboons (Papio spp), Hausfater et al., 1982; Altmann and Alberts, 2003] is consistent throughout an individual’s lifetime and high rank has been correlated with increased fitness (including increased infant survival, shorter inter-birth intervals and lower age of sexual maturity for offspring; Pusey et al., 1997) and reduced stress

(Sapolsky, 2005). However, a previous study of this population of white-faced capuchins did not find an effect of female rank on fitness, probably due to the interference or confounding influence of male reproductive strategies such as infanticide on the overall reproductive success of females (Fedigan et al., 2008). While male strategies can lower overall reproductive success in females during times of instability and after male takeovers, high dominance rank may still have a positive impact on fitness during periods of stability as well as on individual health. A comparison of reproductive output during periods of stability and instability as well as measures of health such as nutritional intake, body weight and condition, and stress (i.e., glucocorticoid levels) in white-faced capuchins would further test this suggestion. Also, research regarding factors influencing the acquisition of rank such as coalitionary support from others, most importantly

42 mothers and full-siblings, would help us to better understand rank order patterns in relation to dominance and inclusive fitness benefits in this species.

Nepotism and matrilineal rank inheritance

Dominance relationships among females in LV and CP groups were nepotistic and females ranked adjacent to kin but rank order patterning did not follow Kawamura’s principle that younger daughters would rank over older ones. Instead, once rank was established by a female within the hierarchy, she tended to rank according to her age, with older sisters ranking higher than younger ones. This capuchin pattern of rank order suggests two things. First, age and individual attributes such as size may play an equally important role in the organization of hierarchies; measures of body size and weight would help to clarify this finding. Second, mothers may only infrequently support daughters in agonistic interactions (e,g,, against their sisters) [only 5 of 15 (33.3%) mothers were still present in the hierarchy when their daughters acquired a rank position] and/or that an individual’s dependent rank with her mother might diminish before rank establishment in the adult hierarchy commences. Accordingly, when a young capuchin female reaches an age at which she needs to compete for resources, she must work her way up the hierarchy by challenging those with previously established ranks to develop dominance relationships and solidify her new status. Effectively, such rank challenges and changes may occur until a female obtains a rank position below but adjacent to her sibling(s). A possible reason for hindrance of further rank ascendency within the hierarchy is that once positioned adjacent to kin, a female would be unlikely to overtake her maternal kin and ascend the hierarchy any further due to strong social bonds and possible lack of social

43 support. More detailed studies of social behavior, including coalitionary behavior, before and during rank acquisition are needed to come to a conclusion.

Nonetheless, my results suggest that nepotism plays a key role in the structure of female hierarchies in our study capuchins. Despite the deviation of the capuchin pattern from Kawamura’s principles, the importance of nepotistic support among white-faced capuchins is still apparent since maternal kin acquire rank adjacent to one another.

Examination of hierarchies for LV and CP from 1997 until 2007, (excluding temporary immigrants), provides support for this argument. Almost all individuals with matrilineal kin simultaneously present in the hierarchy (N = 14, two matrilines in LV and three matrilines in CP) ranked adjacent to matrilineal kin in both groups. The one exception was KI, who was the youngest adult female in CP group and may not have fully established or stabilized her rank. Her rank increased between 2007 and 2008 and if this trend continues she will settle into rank position below her older sister (ED) once her rank is fully established, thus following the proposed rank pattern for this population.

There were not enough instances where either 1) a mother was present during the same time period as her daughters or 2) ranks were stable enough to make a valid assessment of patterns of mother-daughter rank order. In both cases of mother-daughter rank reversal, mothers were at least 14 years old. Past studies of macaques have found an effect of age on social behavior that varies by species and degree of social withdrawal

[e.g., Japanese macaques (Macaca fuscata), Pavelka, 1990; toque macaque (Macaca sinica), Ratnayeke, 1994; long-tailed macaque (Macaca fascicularis), van Noorwijk and van Schaik, 1987; Veenema et al. 1997]. Veenema et al. (1997) found that older individuals (≥ 14 years) spent less time socializing and that lower ranking individuals

44 were affected more by age than higher ranking individuals. Little is known about the effect of age on sociality in white-faced capuchins. Although age does not seem to affect social behavior in captive individuals (Fragaszy, Leighty and Branch cited in Fragaszy et al., 2004), it is possible that age might cause social withdrawal and an ensuing drop in rank position for old females in wild populations where ecological pressures impacting competition and social behavior are greater. A larger dataset is needed to investigate this hypothesis.

Strength

Measures of hierarchical strength indicate that white-faced capuchins exhibit strong dominance hierarchies. A significant percentage of agonism occurred in a resource-related context, indicative of a strong hierarchy. The latency to detection during the dry season was consistent for all three groups at approximately 21 hours/female (N =

22). This is a very short latency when compared to other species such as vervets (46 hr/female, N = 9) or patas (51 hr/female, N = 15) (Isbell & Young, 2002), suggesting that white-faced capuchins have stronger or more highly expressed dominance hierarchies compared to these species. However, the latency to detection measure tends to be negatively related to the rate of agonism (i.e., greater levels of agonism yield more submissive interactions and thus, faster detection of hierarchies). And I found rates of agonism to be seasonally dependent for this population; therefore, the strength of dominance hierarchies may be seasonally variable as well. The minimum number of observational hours needed to confidently rank all individuals was only met for LV group during the rainy season (latency x number of females = 99.32, 148.55, and 219.92 for

LV, CP and GC respectively), using the latency to detection measures as determined for

45 the dry season. Therefore, the latency to detection values calculated for LV, CP, and GC groups provide further evidence that there was an insufficient amount of data collected during the rainy season for CP and GC groups to place all individuals into a dominance matrix. Additionally, rates of submission for all three groups were lower during the rainy season, indicating that latency may be longer and dependent upon seasonal distribution of resources and resulting levels of competitive interaction. More data are needed to assess the variability of latency to detection and hierarchical strength according to season for white-faced capuchins. Finally, directional inconsistency scores for all three groups were

0% in the rainy season and less than 5% in the dry season. This provides another indicator of strong dominance hierarchies. The directional inconsistency index measure is said to correlate with hierarchical stability (Isbell & Young, 2002) because hierarchies with fewer dominance reversals against the hierarchy tend to be more stable over time.

Stability

Hierarchical stability has been reported among many species of cercopithecines

(Nakamichi et al., 1995; Hausfater et al., 1982; Cheney et al., 1981). My limited analysis of stability shows that female white-faced capuchin hierarchies are stable. Dominance relationships are either decided or avoided altogether; no relationships were tied or heavily disputed during this study period, which would have been indicated by a high level of bidirectional submission within a dyad. However, a closer look at factors influencing linearity reveals that even a few bidirectional instances of submission occurring among dyads of low rank distance indicate low-grade instability and may decrease the degree of linearity from perfect transitivity. Adult female group membership in this population changes by approximately two females per year and

46 females are likely to change rank every 2.34 years. Changes in group membership likely provide a constant low-grade disruption of hierarchical stability, contributing to rank change, since each year females present in the hierarchy fill the new void and a new female will enter and challenge others until her rank position has been established.

Despite small scale fluctuations in rank over the course of long-term study for this population, there were no changes in the ordering of matrlines over the past eight years in

CP and ten years in LV (despite one male takeover in CP and three in LV), indicating considerable stability in dominance relationships among female white-faced capuchins in the face of low-grade fluctuation in females or more drastic changes in male composition and membership. Monitoring for more rapid changes in rank order or larger changes such as reversals of matrilines may be crucial for identifying disruptive pressures contributing to dominance instability and an important step in evaluating the adaptive significance of female dominance rank for capuchins.

Conclusions

1. All females could be placed into a rank order based on the direction of

submission.

2. The degree of hierarchical linearity was high during the dry season.

3. Submission rates and thus dominance interactions varied significantly by group

size and by season.

4. Females acquired a stable rank position within approximately one year of

reaching sexual maturity, which predicted their rank years later.

5. Hierarchies are nepotistic but age may also influence hierarchical structure.

Females rank adjacent to maternal kin and occupy ranks similar to those held by

their mothers. However, they do not follow Kawamura’s principle of youngest

daughter ascendency.

6. Hierarchies are strong. A high proportion of agonistic interaction occurred within

the context of food, latency to detection is relatively short (21 hrs/female), and

directional inconsistency scores for interactions were low (<5% across seasons

and groups).

7. Ranks are stable. Despite the rate of change in female group membership of

approximately two females/year (one exit, one entry), females are only expected

to change rank an average of once every 2.34. There were no tied dyadic

relationships, indicating that relationships were either strongly enforced and

known, or dyads never interacted.

Table 2.1: Social categories among female primates designated according to competitive regime. Table reproduced from Sterck et al. (1997).

Social category Competitive regime Social response Within- Between- Female Female group group philopatry ranking contest contest DE Dispersal- Low Low No Egalitarian Egalitarian RE Resident-Egalitarian Low High Yes Egalitarian RN Resident-Nepotistic High Low Yes Nepotistic and despotic RNT Resident-Nepotistic- High High Yes Nepotistic but Tolerant (Potentially) tolerant

Table 2.2: The total number of focal hours collected per focal female for LV, CP and GC groups of white-faced capuchins across rainy and dry seasons. Rainy season Dry season Group Focal Focal ID Focal samples Total focal hours Focal samples Total focal hrs LV Kathy Lee KL 38 6.33 94 15.67 Dos Leches DL 37 6.17 95 15.83 Salsa Lizano SL 37 6.17 95 15.83 Blanquita BB 37 6.17 95 15.83 Chutney CH 37 6.17 95 15.83 186 31.01 474 79 CP Simba SI 36 6 96 16 Sarabi SA 36 6 96 16 Ed ED 36 6 96 16 Timone TI 37 6.17 96 16 Kiara KI 34 5.67 96 16 Seria SE 36 6 96 16 Zazu ZA 37 6.17 96 16 252 42.01 672 112 GC Lily LY 33 5.5 84 14 Maxine MX 33 5.5 84 14 Minerva MV 33 5.5 84 14 Luna Lovegood LL 33 5.5 84 14 Petunia PT 33 5.5 84 14 Fleur FL 33 5.5 84 14 Rosmerta RM 33 5.5 84 14 Mrs Weasley MW 33 5.5 84 14 Rita Skeeter RS 32 5.33 84 14 Lavender LV 33 5.5 84 14 329 54.83 840 140 Overall 767 127.85 1986 331 49

Table 2.3: Group compositions (number of individuals present per age/sex class) during the 2007 and 2008 study periods by group of white-faced capuchins (LV, CP, and GC) and season.

LV CP GC Rainy Dry Rainy Dry Rainy Dry Age/Sex Class Adult Female 5 5 7 7 10 10 Adult Male 2 2 3 2 3 4 Subadult Male 1 1 0 0 0 4 Juvenile Female 4 4 4 4 2 2 Juvenile Male 2 2 2 0 1 4 Infant Female 2 2 4 5 2 2 Infant Male 2 3 2 1 6 7 Total 18 19 22 19 24 33

Note: Table includes all individuals present at any time each season and not minimum values.

Table 2.4: Identity, date of birth and genealogy information for female white-faced capuchins that were present in the adult female hierarchy for LV or CP group from 1997- 2008 (Farah and Patch1, not present during this time period, were included to show a complete record of known matrilineal relationships) at Santa Rosa National Park, Costa Rica.

* indicates estimated birth year i indicates immigrant female with unknown matriline

Rank (i) KL DL BB SL CH 1 KL * 0000 2 DL 0 * 0 0 0 3 BB 1 0 * 0 0 4 SL 0 2 1 * 0 5 CH 1 0 1 1 *

Rank (ii) KL DL BB SL CH 1 KL * 0000 2 DL 7 * 0 0 0 3 BB 3 1 * 0 0 4 SL 1 1 1 * 0 5 CH 4 2 1 3 *

Rank (i) SI SA ED TI SE KI ZA 1 SI * 000000 2 SA 0 * 0 0 0 0 0 3 ED 0 0* 0000 4 TI 1 0 0 * 0 0 0 5 SE 1 0 0 0 * 0 0 6 KI 2 2110* 0 7 ZA 1 00011*

Rank (ii) SI SA ED TI KI SE ZA 1 SI * 000000 2 SA 3 * 0 0 0 0 0 3 ED 9 4* 0000 4 TI 6 6 7 * 2 0 0 5 KI 14 8 1 4 * 1 1 6 SE 9 4813* 1 7 ZA 9 30274*

Rank (i) LY MX LL PT RM FL MV MW RS LV 1 LY * 0 0 0 0 0 0 0 0 0 2 MX 1 * 0 0 0 0 0 0 0 0 3 LL 1 0 * 0 0 0 0 0 0 0 4 PT 0 1 2 * 0 0 0 0 0 0 5 RM 0 0 0 2 * 0 0 0 0 0 6 FL 0 0 0 1 0 * 0 0 0 0 7 MV 1 0 0 0 1 0 * 0 0 0 8 MW 0 0 1 1 0 0 0 * 0 0 9 RS 0 1 0 0 4 0 1 0 * 0 10 LV 0 0 0 1 0 1 0 0 1 *

Rank (ii) LY MX MV PT LL RS FL MW RM LV 1 LY * 0 0 0 0 0 0 0 0 0 2 MX 0 * 0 0 0 0 0 0 0 0 3 MV 0 1 * 0 0 0 0 0 0 0 4 PT 5 12 4 * 0 0 0 0 0 0 5 LL 1 2 0 3 * 2 1 0 0 0 6 RS 2 4 7 9 5 * 0 0 0 0 7 FL 2 2 3 7 3 2 * 0 0 0 8 MW 3 0 1 6 0 1 1 * 0 0 9 RM 4 5 0 6 6 6 0 6 * 0 10 LV 7 4 4 10 5 0 1 2 0 *

Table 2.8: Rate of submission, calculated as the number of interactions per focal hour, by group (LV, CP, and GC) and season among female white-faced capuchins.

Rate of submission Group Season # Interactions # Focal hours (freq/focal hrs) LV Rainy 5 31.00 0.16 Dry 17 79.17 0.21 CP Rainy 8 42.00 0.19 Dry 83 112.00 0.74 GC Rainy 15 54.83 0.27 Dry 88 140.00 0.63

Table 2.9: The proportion of unknown dyads and Landau's index of linearity (h') with associated p-values by group of white-faced capuchins (LV, CP, and GC) and season.

Group Season # females (N) # unknown dyads (%) Landau's (h') p-value

LV Rainy 5 4/10 (40.00) 0.600 0.506 Dry 5 0/10 (0.00) 1.000 0.113 CP Rainy 7 12/21 (57.14) 0.536 0.319 Dry 7 1/21 (4.76) 0.893 0.016* GC Rainy 10 29/45 (64.44) 0.279 0.508 Dry 10 9/45 (20.00) 0.776 0.001** *p≤0.05 and **p≤0.01

Latency to detection Group N females Total hrs observed N placement hrs (hrs/female) LV 5 191.60 99.32 19.86 CP 7 228.13 148.55 21.22 GC 10 294.53 219.92 21.99

Group Season % known dyads DII LV Rainy 60.00 0% Dry 100.00 0% CP Rainy 42.86 0% Dry 95.24 4.27% GC Rainy 35.56 0% Dry 80.00 1.94%

0.8

0.7

0.6

0.5

0.4

0.3

0.2 Rate of submission (frequency/hour) LV 0.1 CP GC 0 Rainy Dry Season

Figure 2.1: Submission rates (frequency per hour) reported for adult female white-faced capuchins by group (LV, CP, and GC) and season. There was a significant interaction effect between season and group (F(2,21) = 4.083, p = 0.032).

80 Resource Social 70

30 Mean + SEM (% interactions) (% SEM + Mean 20

0 Context of aggression

Figure 2.2: The mean ± SEM percentage of agonistic interactions that occurred in the context of resource versus social contexts for LV, CP and GC groups of white-faced capuchins during the dry season (χ2 (2, N = 3) = 6.44, p = 0.04).

CHAPTER 3. DOMINANCE STYLE AMONG FEMALE WHITE-FACED

CAPUCHINS

Introduction

The socioecological basis for dominance style

Intra-specific competition for cost-effective food resources is the most important factor shaping social relationships among female primates (Wrangham, 1979;

Wrangham, 1980). When resources are ubiquitous, direct competition for acquisition is not necessary and egalitarian social systems without dominance hierarchies exist. When desirable resources are clumped and defendable, increased access through organization into kin groups (i.e., female-bonded, “FB” groups) outweighs the cost of competition

(Wrangham, 1980; van Schaik, 1989). Furthermore, the level of competitive pressure is expected to affect the “dominance style” that females exhibit in FB groups. Dominance style is the level of tolerance displayed by dominant individuals toward more subordinate ones, and varies on a continuum from despotic to relaxed relationships (de Waal, 1989).

It is a larger, more encompassing social construct than the dominance hierarchy that is formed within a group and depicts the “nature of the entire relationship between dominant and subordinate” in terms of dominance, competitiveness, and tension regulation (de Waal, 1986; de Waal, 1989). Whether or not a species exhibits a despotic or a more relaxed dominance style seems to be correlated with environmental pressures, including intergroup competition and predation (de Waal, 1989). Behavioral mechanisms of tension regulation are not expected in species where females rarely compete over food

62 resources (de Waal, 1989). When individuals rely on other group members for cooperative defense in response to high levels of intergroup competition or predator defense, a greater level of tolerance is expected between individuals with disparity in rank (de Waal, 1989; van Schaik, 1989; Sterck et al., 1997).

Behavioral co-variation

Three measurable traits co-vary in macaque species according to the type of dominance style exhibited (de Waal, 1989; Thierry, 1985). These traits include the bidirectionality of aggression (i.e., the level of symmetry of aggressive interactions that occurs within a dyadic relationship), kin bias in social behaviors such as approaching, grooming, proximity, and co-feeding (i.e., feeding within close proximity), as well as the presence or absence of post-conflict conciliatory behavior. The extent to which these three behavioral traits are exhibited and influence social relationships provides a quantifiable way to classify the dominance style for each species. Interspecies comparisons provide a means by which dominance style can be assessed across primate genera in relation to factors such as social organization and ecological pressures.

The bidirectionality of aggression represents the type of relationships dominants and subordinates exhibit in conflict situations. Aggression is unidirectional when group- level cooperation is not necessary and dominant individuals do not tolerate subordinates as aggressors during within-group resource competition. On the other hand, aggression tends to be bidirectional when high ranking females tolerate some level of competition and aggression and when ecological pressures may promote the maintenance of a higher level of cohesion and cooperation among group members (de Waal, 1989).

The amount of variation that occurs within female relationships, specifically with respect to dominance and philopatry, is also a major influence on social structure (Sterck et al., 1997). In many species, the development of close social bonds relies heavily upon a high degree of relatedness. According to Sade (1972), dominance and kinship are often linked. Species that have strong, linear dominance hierarchies exhibit a high degree of nepotism and kin bias. A high degree of kin bias in relation to rank is likely in species with matrilineal dominance hierarchies, as matrilineally related females hold ranks adjacent to one another under these circumstances (Thierry, 1990). Kin bias is usually apparent in patterns of affiliation between closely related (r = 0.25) individuals (Chapais et al., 1997). Therefore, it is important to determine the degree of kin bias associated with affiliation (approaching and grooming), proximity, and co-feeding in order to determine the level of social bonding and tolerance exhibited towards distant or non-kin.

As an additional measure of tolerance, maintenance of proximity can be used to determine if dominant or subordinate individuals are responsible for maintaining behavioral association. Maintenance of proximity by subordinate females with dominant females may indicate tolerance of subordinates on behalf of dominants, representative of a relaxed dominance style.

Reconciliation is not common in primates and has been proposed to represent an effort to pacify strained relationships in order to maintain strong bonds after agonistic encounters. de Waal (1986) suggested that aggression and subsequent reconciliation may act to strengthen bonds in dyadic relationships. A species that exhibits frequent reconciliation may face socioecological pressures to maintain cooperative relationships.

Reconciliation may function directly after a dispute as a means of resolving conflict or more regularly though species-specific social behaviors that are thought to reduce tension among more relaxed macaque species such as soft biting [e.g., stumptail macaques

(Macaca arctoides), Thierry, 2000].

According to de Waal (1989), patterns of bidirectional aggression, kin bias and reconciliation all represent a species’ response to socioecological pressure. The analysis of covariation of these social patterns allows us to assess dominance style along a continuum from despotic to relaxed. Maintenance of social relationships by higher ranking individuals, greater tolerance of lower ranking individuals in competitive situations, and higher rates of reconciliation among dyads with greater rank disparity result in cooperative relationships that work to counteract pressures such as between- group competitions and thereby yield a more relaxed dominance style. On the other hand, asymmetric aggression between individuals, infrequent reconciliation, and high levels of kin bias in social behaviors characterize a despotic dominance style. In despotic primate species, there is very little tolerance of low ranking individuals by high ranking individuals in competitive and social situations.

Since males and females respond differently to socioecological influences and compete for different resources (males for females and females for food), it is important to classify dominance style across all partner combinations to gain a better overall species profile (Hemelrijk & Gygax, 2004). Research among mammals to date is most complete regarding female-female primate relationships.

Dominance style and Old World primates

Studies have begun to address the topic of dominance style in detail in Old World primate species, most thoroughly within the genus Macaca. Drawing from his own and previous research, Thierry (2000) investigated behavioral covariation for nineteen species of macaques (confidently for eleven species and tentatively for eight species due to limited data) and divided the continuum of characteristics described by de Waal (1989) into a four-grade scale. The first grade consists of macaque species that exhibit the most despotic behavior whereas the fourth grade consists of the most relaxed species (Table

3.1). While Thierry’s graded scale provides categorization of macaque species, it does not provide behavioral cutoffs for each grade, leaving precise categorization (particularly between two adjacent grades) unclear. Berman et al. (2004) clarified research on this topic for Tibetan macaques (Macaca thibetana) by categorizing their dominance style as despotic and placing them within the second grade of Thierry’s scale. Research has recently extended beyond Macaca; captive studies classify gelada baboons

(Theropithecus gelada) as despotic, but more relaxed than all macaque species (Reichler et al., 1998), and the dominance style of guereza colobus (Colobus guereza) as very relaxed (Grunau & Kuester, 2001). According to reported behaviors, guereza colobus would likely fall into either grade 4 of Thierry’s scale or into an even more relaxed category. Data are still scarce regarding dominance style in Old World primate species outside of the subfamily Cercopithecinae and this topic has not been explored in New

World primates. Therefore, I conducted this study of dominance patterns in a New

World monkey to gain a more general understanding of dominance style across primate

66 species, as well as to better understand how cercopithecines and cebines fit into the greater primate pattern.

Research objective and study species

This study investigates the dominance style of female white-faced capuchins

(Cebus capucinus), a gregarious New World primate species. White-faced capuchins live in multi-male, multi-female social groups and share a variety of behavioral patterns with

Old World primate species including female philopatry, male dispersal, and characteristics of female-bonded species (Perry, 1996; Jack, 2007). Many of the traditional measures used to quantify dominance in Old World species are applicable to white-faced capuchins and results can be compared across species. Similarities in female social structure (i.e., patterns of social interactions; Kappeler, 2002) are likely to reflect similarities in many behaviors related to dominance such as rates and direction of aggression, association and affiliation patterns, kin bias and tolerance levels. Here, I examine the bidirectionality of aggression and patterns of kin bias in approach, grooming, proximity, and co-feeding behaviors to make an initial assessment of the type of dominance style exhibited by female white-faced capuchins. As studies of dominance style have not yet been conducted on a New World primate, I will attempt to compare the results obtained in my study to the large body of work published on macaque species.

To explore the topic of dominance style in white-faced capuchins, I investigate two key research questions regarding behavioral co-variants and propose the following predictions based on species-specific behavior already reported for white-faced capuchins:

1) Is aggression bidirectional among dyads of female white-faced capuchins?

Previous research on wild white-faced capuchins has shown that females direct aggression down the hierarchy (Perry, 1996). Bidirectionality of aggression (defined by counteraggression in dyadic conflicts involving physical contact among females, males, and juveniles) was high in captive white-faced capuchins (Leca et al., 2002). However, overall rates of aggression tended to be lower in female-female dyads than male-male and male-female dyads (Rose, 1994). Accordingly, overall bidirectionality of aggression among females should be low, characteristic of a despotic dominance style. I made the following predictions regarding three measures of bidirectionality of aggression: 1) There will be very few instances of aggressive events where lower ranking individuals aggress towards dominant ones; 2) At the group level, I predict that the predominant direction of aggression will be down the hierarchy for all dyads; 3) Levels of counteraggression (i.e., instances where an individual responds and directs any type of aggression back at the aggressor within the same aggressive bout) will be moderate (e.g., 10-30%) in dyadic encounters between females.

2) Do female white-faced capuchins exhibit kin bias in social behaviors?

Female white-faced capuchins are philopatric and form matrilineal dominance hierarchies (see Chapter 2). Association patterns (i.e., proximity) and affiliation among females are very high (Fragaszy et al., 2004), particularly among females of similar rank, who share a higher degree of kinship (Perry, 1996). Grooming is also more common among females than males in dyads with small rank differences (Manson et al., 1999), where a higher level of relatedness is expected, and is mostly directed towards dominant

68 individuals. However, because females minimize the number of individuals in association while feeding (Phillips, 1995), bias in dominance and kinship may not hold under foraging conditions. According to these findings from earlier studies, overall levels of kin bias in social behaviors should be high (i.e., characterized by a positive and significant correlation between behavior and level of relatedness) and thus characteristic of a despotic dominance style. I made the following predictions regarding kin bias in social behavior: 1) Kin bias in rates of approach, groom, and proximity behaviors will be high; 2) Kin bias in rates of co-feeding will be low, representing low tolerance in resource defense; and 3) Dominant individuals will control association patterns and maintenance of proximity.

Additional factors regarding female social behavior are indicative of a despotic dominance style. The capuchin diet is composed of high quality and monopolizable fruit resources, which according to socioecological theory, should lead to high levels of direct competition (van Schaik, 1989; Sterck et al., 1997). Like macaques, female white-faced capuchins form linear, strong and stable dominance hierarchies that can be constructed from clear interactions involving submission (see Chapter 2). While females invariably compete within groups for these resources, they do not typically participate along with the males in between-group competitions for resources and may not need to exhibit the tolerance in dominant-subordinate relationships required to induce cooperation at the group level.

On the other hand, white-faced capuchins exhibit some cooperative behavioral patterns such as alloparenting and social traditions that may function to strengthen social

69 bonds or regulate tension and encourage group-level cooperation typical of a relaxed dominance style. Females show high rates of alloparenting and allonursing (Perry,

1996). Allonursing is a cooperative behavior not seen in many species of primates. Also,

Manson (2005) provided evidence for conciliatory tendencies among adult females in 1 of 2 studies. The “wheeze dance”, described as a sequence of behaviors normally observed in an affiliative context or during consort that co-occur with tense situations

(e.g., fighting) (Perry, 1998). This behavior occurs most often among males but has been observed among females and suggests the existence of appeasement during or after tense situations (Perry, 1998). Finally, “handsniffing,” a behavioral tradition exhibited between two females whereby one or both participants inserts at least one digit into the nostril of their partner and holds the position in a calm and seemingly “meditative” state, facilitates social contact between partners (Perry, 1996). This behavior may function to strengthen social bonds and reduce tension among participants, as it occurs across kinship and dominance lines (Bergstrom, unpubl.).

Nonetheless, given that a large proportion of behavioral patterns in white-faced capuchins indicate despotic relationships and there are few behavioral indicators of relaxed dominance style, I predict that the overall dominance style among females will be despotic.

Methods

Study Subjects

This study is based on observational data collected on white-faced capuchins at

Santa Rosa National Park (SNRP) in the Area de Conservación Guanacaste (ACG),

Guanacaste Province, Costa Rica. SNRP is composed of 108 square kilometers of seasonal tropical dry deciduous forest (Fedigan & Jack, 2001) and experiences two distinct seasons: the rainy season from mid-May and November and the dry season from

December until mid-May. Data were collected from May-August 2007 (127.85 focal hours, 345.57 h observational contact) and January-May 2008 (331 focal hours; 714.27 h observational contact) to account for possible seasonal differences in behavior.

I excluded juveniles and infants (< 5 years old) from data collection since their ranks are dependent upon their mother’s preceding sexual maturity. Socioecological theory suggests that female fitness is more directly tied to resource competition than that of males, whose fitness relies upon access to females (Wrangham, 1980). Consequently, dominance relationships and hierarchical structures may differ between sexes and I limited analysis of dominance style to adult females for this study. Behavior can also vary with group size since demographic factors such as the proportion of related social partners may differ (Berman et al., 1997). I studied three habituated study groups of varying sizes, Los Valles (LV), Cerco de Piedra (CP), and Guanacaste (GC) consisted of

5, 7, and 10, adult females (≥ 5 years old) respectively, totalling 22 individuals over both field seasons. Table 3.2 shows demographic information for focal individuals including rank (dry season), birth date, age at the beginning of the study, and mother. I used age,

71 sex, body size and distinguishing features to identify individuals and to estimate the age of individuals with unknown birth dates.

I constructed dominance hierarchies using dyadic interactions involving supplantation, cower, avoid, grimace and flee behaviors (see Chapter 2). I limited analyses involving maternal relatedness to LV and CP groups for which long-term data have been collected since 1990 and 1986 respectively; two matrilines were present in LV and three in CP during the study period. GC has been studied since 2007 and kinship is unknown.

Data Collection

I collected behavioral data for 10-12 hours per day, five days per week using a

PSION Workabout hand-held computer. I noted rapidly occurring interactions such as intense aggression using a digital voice recorder. Monthly data collection varied with group size (i.e., larger groups were followed longer than smaller groups) to attain similar focal data hours per individual.

I conducted 10-minute continuous-time focal animal follows to record all social behaviors that involved the focal and IDs of interactants (Altmann, 1974). I recorded instantaneous scan samples at 2.5 minute intervals during follows and included the activity of the focal animal as well as all individuals within close proximity (i.e., in contact, >1 body length, 1-5 body lengths, and 5-10 body lengths; (Appendix 2) (Perry,

1996). I collected ad libitum data opportunistically for vocalizations, intragroup agonism and intergroup encounters. To sample individuals as equally and independently as possible I randomly selected available focal subjects who had not interacted or been in

72 proximity of the previously sampled individual within the last 2.5 minute scan interval of the follow. Data were collected by two individuals and always dictated by one and monitored by the other to avoid inter-observer reliability issues.

Variables and Analyses

Bidirectionality of Aggression

I measured the rate of aggression as the number of aggressive events per hour collected during focal samples (Perry, 1996). I also calculated the proportion of mild

(i.e., behaviors lacking physical contact: swipe at, tooth grind, bounce, nip, snap at, lunge, glare, or open-mouth threat) versus intense (i.e., behaviors involving physical contact: bite, chase, hit, pull, push, pounce on, or wrestle) aggression from focal samples as a measure of escalation in aggressive encounters (see Appendix 1 for ethogram). I included ad libitum data in all measures of aggression other than rate calculations.

Similar to methodology used by Berman et al. (2004), I used three measures to test the bidirectionality of aggression among females. The degree of bidirectionality was determined using the directional inconsistency index (de Waal, 1977). This index calculates the asymmetry of aggression as the percentage of all interactions that were directed in the less frequent direction within dyads. Second is a measure of the direction of aggression at the group level: I used the dyads-up index to calculate the percentage of dyads for which the main direction of aggression was up the dominance hierarchy. Third is counteraggression, which I measured as the percentage of bouts of dyadic aggression of any kind that included one or more aggressive responses from the target. A bout was defined as an aggressive sequence between two individuals that included one or more

73 aggressive events. All aggressive events within the same dyad were considered the same bout until 10 minutes without aggression elapsed. Although these measures are not mutually exclusive, they provide different perspectives from which to analyze the bidirectionality of aggression (i.e., at the level of the dyad and group).

Kin Bias

Relatedness, measured via the coefficient of relatedness (r), represents the percentage of genes that two individuals share by common descent. Values used in this study reflect known relatedness through matrilineal lines in LV and CP groups.

However, paternity is unknown. I assessed the degree of kin bias for approach, grooming, proximity, and co-feeding behaviors to determine if tolerance was exhibited more towards close maternal kin than distant/non-kin. I categorized individuals as close kin at r ≥ 0.25 and distant kin or non-kin at r < 0.25, a standard cutoff for kin bias in primate studies (Chapais et al., 1997).

I measured approach rate as the frequency per hour that the focal animal approached another individual within 2.5 meters. I measured rate of grooming as the frequency per hour that a focal subject initiated a bout of grooming with another individual. To determine the amount of time spent within five meters of another individual, I measured the percentage of scan samples in which the focal subject was within five meters of another specific individual out of the total scan samples taken for that dyad. To determine if individuals preferentially co-feed with close kin, I recorded per dyad the frequency per hour that a focal subject and another individual fed within 5m of one another.

I used matrix correlations from partial Kendall’s (Kr) rowwise tests with 2000 permutations (Hemelrijk, 1990) to determine kin bias in rates of approach, grooming, and co-feeding as well as the percentage of time spent within 5m. This nonparametric method of correlation analysis tests for a relationship between each social factor and the relatedness of individuals while controlling for possible effects of rank distance, since females rank adjacent to matrilineal kin (Noldus Information Technology, 1998). Kr coefficients range between -1 and 1, with positive values categorized as weak (0-0.5), moderate (0.5-0.8) and strong (0.8-1.0) positive correlations between datasets. I only used groups with known maternal relatedness (LV and CP) for these analyses.

Intensity of kin bias

I also used the intensity of kin bias as a measure of kin preference among group members. To determine the intensity of kin bias, I used a method described by Berman et al. (2004). I compared the null hypothesis that there was an equal number of grooming bouts among all females in the group to the amount of observed grooming [(the total number of grooming bouts given and received by a female during her focal follows) x

(the proportion of females in the group that were related by ≥ 0.25)].

Maintenance of relationships and rank

I also calculated maintenance of proximity scores as a measure of tolerance to determine if dominant or subordinate individuals were responsible for maintaining proximity and association with each other. I used approaches and leaves during focal follows to calculate maintenance of proximity for each dyad using the formula: A’s responsibility for proximity = UA/(UA + UB) – SA/(SA+ SB), where U = the # of occasions

75 where the pair united to a distance of less than or equal to 2.5 meters and S = the # of occasions where the pair separated (Martin & Bateson, 1993). I used a paired t test to determine if there was a significant seasonal difference in maintenance of proximity. I considered all results of statistical tests significant at p ≤ 0.05.

Results

Bidirectionality of aggression

I recorded a total of 331 aggressive interactions during focal animal sampling over both seasons (130 rainy and 201 dry) to calculate the rate of aggression for each group. Rates of aggression increased with group size (Figure 3.1) and were an average of

1.74 times greater in the rainy season (LV: 0.58 interactions/hour; CP: 1.00; GC: 1.28) than the dry season (LV: 0.30; CP: 0.58; GC: 0.80), which may indicate seasonal variation in ecological factors (i.e., fruiting species, size, and distribution) affecting aggressive interactions. When aggression does occur, the majority is mild (M = 73.6%,

SD = 1.2%) versus intense aggression (M = 25.9%, SD = 1.4%).

Table 3.3 shows results for the three measures of bidirectionality of aggression

(directional inconsistency index, dyads-up index and percentage counteraggression) among females for both seasons across LV, CP and GC groups. I recorded a total of 516 aggressive interactions (focal and ad libitum sampling) across all three groups (178 rainy,

338 dry). The directional inconsistency index reports the highest bidirectionality in aggressive interactions of the three measures; however, scores were still at a moderate level. Scores fell between 16.9–26.7% for all three groups during the rainy season and

76 between 17–18.8% for two of three groups during the dry season, indicating that females direct aggression up the hierarchy in approximately one of five behavioral interactions.

LV, the smallest group, was an exception at only 6.9% for the dry season.

Low dyads-up scores indicate that although some aggressive interactions are directed up the hierarchy, the majority of aggression at the dyad level among females is directed downward from dominant toward subordinate individuals. For the rainy season, dyads-up scores were 14.3% and 7.1% for LV and CP respectively (a score was not calculated for GC since a reliable hierarchy could not be constructed due to a low number of submissive interactions). Scores were lower (0–3.2%) for all three groups for the dry season. Changes in the direction of aggression in two of the 31 dyadic relationships (one in LV and one in CP) accounted for the seasonal variation in dyads-up scores. These changes were supported by a very low number of behavioral interactions per dyad (one in

LV and two in CP) and more data are needed to confirm whether they reflect a true change in the direction of aggression for those relationships.

I recorded a total of 244 dyadic bouts of aggression (i.e., a sequence of at least one aggressive behaviors) during both seasons across all three study groups. The overall percentage of counteraggression was low (4.5%) indicating that contests between white- faced capuchin females are asymmetric; only 11 bouts included an aggressive response from the target individual. Results suggest that a relationship may exist among these variables: the percentage of counteraggression, female group size, and season. There was an eight-fold difference (7.7% rainy, 0% dry) in percentage of counteraggression between seasons for the smallest group (LV), a four-fold difference (16.0% rainy, 3.6% dry) for

77 the intermediate group (CP), and little variation across seasons (2.6% rainy, 3.2% dry) for the largest group (GC). Seasonal variation in resources may influence the occurrence of counteraggression, having less of an influence on larger groups. Within the two smaller groups (LV and CP), dominant individuals directed aggression towards subordinates in six of the seven (86%) counteraggressive bouts.

Kin Bias

Females in CP group showed significant but weak kin bias (τ < 0.5) in approaching and grooming behaviors during both seasons. Females in LV group showed a significant moderate preference (0.5 ≤ τ ≤ 0.8) for kin as grooming partners during the dry season but lacked kin bias in approach rate during both seasons and grooming rate during the rainy season (Table 3.4).

Mean approach and groom rates for maternal kin (r ≥ 0.25) and distant/non-kin (r

< 0.25) dyads are shown in Figures 3.2 and 3.3, respectively. There was no significant kin bias in either group for the amount of time spent within 5m of another individual.

Although not significant, kin were likely to be in proximity more often during the dry season for LV group and in the rainy season for CP group, suggesting that the amount of time spent near others is not random (see Figure 3.4 for mean percentages). There was no significant kin bias in co-feeding behavior for either group (Figure 3.5 displays mean co-feeding rates for kin and distant/non-kin dyads). Overall, females showed significant bias in approaching and grooming close maternal kin but did not spend significantly more time in proximity or co-feeding with kin, suggesting that individuals actively seek out kin to engage in these affiliative behaviors.

I also investigated the intensity of the kin bias expressed in grooming behavior among females. To measure intensity, I compared the frequency of observed grooming among kin to the amount of expected grooming, based on the number of available kin within the groups following (Berman et al., 2004). Females in LV group lacked kin bias in grooming during the rainy season (mean intensity = 0.95 ± 0.31, n = 5) but groomed female kin at an intensity of 1.51 ± 0.08 (n = 5) in the dry season. As expected, intensity increased with group size and females in CP group showed preference for maternal kin as grooming partners across both seasons at a stronger average intensity of 2.00 ± 0.20

(rainy: 2.08 ± 0.25; dry: 1.92 ± 0.32, n = 7) times the expected amount given matrilineal relatedness within the group.

As an additional measure of tolerance, I calculated maintenance of proximity scores for female dyads. In LV group, there was a significant shift in responsibility for maintaining proximity from subordinate individuals to dominant individuals between rainy and dry seasons (paired t(9) = -2.613, p = 0.028) (Figure 3.6). In CP group there was also a significant difference in the maintenance of proximity among dyads (paired t(20) = -2.604, p = 0.017), although dominant individuals were responsible for maintaining proximity with subordinates during both seasons.

Discussion

To date, there have been no studies of dominance style in New World monkeys.

The patterns of social behavior documented in this study suggest that dominance style among female white-faced capuchins is very similar to Old World cercopithecine species

79 such as macaques. I investigated the bidirectionality of aggression and kin bias in social behavior among female white-faced capuchins to provide an initial assessment of dominance style and to make comparisons with Old World species in which dominance behavior is better understood. Based on behavioral patterns of white-faced capuchins published in previous studies, I predicted that females in my study population would exhibit behavioral patterns indicative of a despotic dominance style. My results suggest that an intermediate classification of dominance style may be more appropriate.

Levels of aggression across all three study groups were low (approximately one or fewer interactions per hour) and similar to rates reported previously among female white- faced capuchins in this population (Fedigan, 1993; Rose, 1994). Aggressive encounters mostly consisted of mild, non-contact interactions. All three measures of bidirectionality of aggression suggest that the majority of aggression is unidirectional among females, which is characteristic of a despotic dominance style. The unidirectionality of overall aggression (directional inconsistency index) and the direction of aggression within dyads

(dyads-up index) indicate few attempts by low ranking individuals to challenge higher ranking ones. In addition, 86% of the few instances of counteraggression among females in LV and CP groups occurred from dominant individuals toward subordinates in response to mild threats, indicating possible enforcement/reinforcement of dominance status by higher ranking individuals when challenged. Bidirectionality of aggression was higher in the rainy season for all three measures, suggesting that during this time of year, dominant-subordinate relationships may be slightly more relaxed. Correlation of aggressive behavior with ecological measures may help to clarify this pattern.

Females showed significant kin bias in rates of approaching (CP dry season) and grooming (LV dry and CP rainy and dry seasons) as well as an effect of dominance on association patterns, which is suggestive of a despotic dominance style. The intensity of kin bias in grooming behavior was 1.5 times greater than expected in LV group during the dry season and 2.0 times greater in CP across both seasons, providing further support for the existence of a despotic dominance style. Additionally, dominant individuals were responsible for maintenance of proximity with subordinate individuals during the dry season in LV, which contrasted with rainy season patterns. Dominant individuals maintained proximity with subordinate individuals during both seasons in CP, but to a much greater extent during the dry season. The difference in maintenance of proximity between the two groups across seasons suggests that group size may seasonally affect proximity patterns. The pattern of dominant individuals maintaining proximity with subordinate individuals corresponds to significant kin bias in approach and grooming for both groups and may be indicative of despotic behavioral patterns if initiation of social interactions by subordinates are not tolerated by dominants and occur less often.

However, another possibility is that dominant individuals are approaching to initiate association and form stronger bonds with subordinates, indicative of relaxed dominance style. Lack of kin bias in approach and grooming during the rainy season corresponded to a reversal in maintenance of proximity by females in LV group. LV is the smallest group and individual differences in behavior may have a greater effect on patterns seen at the group level. The two highest ranking females had dorsal infants during the rainy season which may have contributed to the lack of kin bias and more approaches by

81 individuals from the lower ranking matriline. In accordance with this finding, research on macaques has shown kin bias in social behaviors to increase with group size (Berman et al., 1997).

Low levels of kin bias in the percentage of time spent in proximity and co-feeding in this population of white-faced capuchins are indicative of a more relaxed dominance style among females. These behaviors may be considered “passive” since they allow a greater distance between individuals and involve less direct interaction than either approaching or grooming. Therefore, female capuchins show kin bias in “active” social behaviors that is not extended to the “passive” behaviors of time spent in proximity and co-feeding, indicative of an overall intermediate dominance style classification.

Comparison with macaque species

Although measures of dominance style for macaque species are often calculated across all male and female partner combinations, broad comparisons can be drawn with behavioral patterns of female white-faced capuchins. Published results for despotic species of macaques (Macaca fascicularis, M. fuscata, M. mulatta, and M. thibetana; reviewed by Berman et al. 2004) classified within grades 1 and 2 of Thierry’s (2000) scale, display low percentages of bidirectional aggression across the directional inconsistency index, dyads-up (0-5%), and counteraggressive measures (0-30%). Female white-faced capuchins display percentages characteristic of grades 1 and 2 for the dyads- up index and counteraggression measures during the dry season, although directional inconsistency index scores are somewhat higher (mean = 17.7%) than scores reported for all macaque species. Considering all three measures of aggression, a grade 2

82 categorization seems fitting during the dry season. However, scores for capuchins across all three measures of aggression are higher during the rainy season implying that dominance style may become more relaxed towards grade 3 during the rainy season.

Female white-faced capuchins showed a lower level of kin bias (in only two of four measured behaviors) relative to grade 2 M. thibetana (for which the most extensive kin bias results have been published) across all behaviors. Results were not significant in proximity and co-feeding situations. While this may reflect a difference in social preference, it may also be the outcome of smaller group sizes (capuchin mean = 18;

Fragaszy et al. 2004) and reduced matrilineal kin availability, in turn leading to greater amounts of social behavior with more distantly related individuals (Chapais, 2001).

Despite lack of kin preference in proximity and co-feeding, the presence of kin bias in affiliative interactions among female capuchins is still characteristic of grade 1 and 2 macaques and grades 3 and 4 typically do not show strong kin preference in affiliative behaviors (Thierry, 2000). The intensity of grooming among female capuchins (LV: 1.5;

CP: 2.0) fell between grade 1 and 2 macaques (Macaca fuscata: 2.3, M. mulatta: 2.6, and

M. thibetana: 2.7) (reviewed by Berman et al. 2004) and published data on grade 3 macaques (M. arctoides: 1.3). This suggests that capuchins could fall in either grade 2 or grade 3 based on kin bias and grooming behavior.

Conciliatory behavior was not measured in this study due to difficulties in implementing post-conflict – matched control (PC-MC) methodology under field conditions where the number of variables influencing social behavior is greater. PC-MC methodology measures the occurrence of reconciliation by comparing rates of behavior

83 within a dyad directly after a conflict (PC) to rates of behavior within the same dyad during a matched situation (i.e., similar environment) at a later time. Using PC-MC analysis with wild male and female white-faced capuchins, Manson et al. (2005) showed that reconciliation occurred in 1 of 2 studies at levels (9-44% among all male and female dyad types) similar to two species of macaques (grade 2 Macaca fascicularis and grade 4

Macaca maurus) and one species of baboon (Papio anubis, not rated on Thierry’s scale).

Reported conciliatory tendency for female-female dyads was lower (9%) but still supports presence of reconciliation behavior. Evidence for low levels of reconciliation among females in addition to cooperative behavior such as alloparenting and handsniffing behavior, suggest an intermediate dominance style classification for this category but additional measures are needed to determine a more exact position within the graded scale.

Conclusion

Female white-faced capuchins displayed intermediate levels of bidirectionality of aggression when considered across both rainy and dry seasons, akin to an intermediate dominance style. Kin bias was high in “active” social behaviors such as approach and grooming but not present in “passive” social situations such as proximity and co-feeding, which are characteristic of an intermediate dominance style. Evidence for low-grade conciliatory tendencies has been published for another white-faced capuchin population

(Manson et al., 2005) and the presence of alloparenting and social traditions suggest that cooperation exists among adult females, which is also suggestive of an intermediate

84 classification. These behavioral patterns are comparable with macaques in grades 2 and 3 of Thierry’s (2000) scale.

Studies of macaques provide an excellent base with which to compare behavioral data from species with similar behavioral patterns. Detailed studies of other genera will help us to expand our understanding of dominance style, clarify behavioral differences between grades of the continuum, and assess whether the despotic to relaxed continuum of dominance style, designed for macaque species, represents the full spectrum of variation possible across the primate order. I found that behavior among capuchin females varied between seasons, and therefore, species may exhibit a small range of dominance styles dependent on seasonal fluctuation of resources. More comparable data on female-female dyads is needed across seasons, populations, and species to confirm the style of dominance exhibited by female white-faced capuchins. Additional studies within the genus Cebus, or sister taxa Saimiri, will create a comparative framework from which we can begin to expand our knowledge of dominance style and assess behavioral variation among platyrrhines.

Table 3.1: Dominance style grading designed for macaque social organization adapted from Thierry (2000) to include additional species.

1st Grade 2nd Grade 3rd Grade 4th Grade

Behavioral gradient Low conciliatory High conciliatory tendency and tendency and social social tolerance levels. tolerance levels.

High asymmetry of Low asymmetry of contests, contests, dominance dominance gradient and kin bias. gradient and kin bias.

Species Rhesus macaque Longtail macaque Stumptail macaque Tonkean macaque Japanese macaque Pigtail macaque Barbary macaque Moor macaque (Taiwan macaque) Tibetan macaquea Liontail macaque Crested macaque Bonnet macaque (Gorontalo macaque) (Toque macaque) (Heck’s macaque) (Assamese macaque) (Booted macaque) (Muna-Butung macaque)

Parentheses indicate least-known species a Berman et al. (2004) 85

Table 3.2: Female group composition, age in years (as of May 2007) and maternal relatedness for LV, CP, and GC groups of white-faced capuchins organized according to dry season rank within groups.

Group ID Rank (dry) Birth date (mm/dd/yyyy) Age Mother LV KL 1 4/1/1989 18 Gringa LV DL 2 5/29/1991 16 Gringa LV BB 3 1/1/1983* 24* unknown LV SL 4 3/1/1996 11 BB LV CH 5 8/2/1999 7 BB CP SI 1 8/5/1998 8 Limp CP SA 2 1/1/2001 6 Limp CP ED 3 5/14/2000 7 Pumba CP TI 4 5/16/1996 11 SE CP KI 5 4/29/2002 5 Pumba CP SE 6 6/10/1989 17 Patch1 CP ZA 7 2/2/1999 8 SE GC LY 1 1/1/1997* 10* unknown GC MX 2 1/1/1990* 17* unknown GC MV 3 1/1/1990* 17* unknown GC PT 4 1/1/2001* 6* unknown GC LL 5 1/1/1990* 17* unknown GC FL 6 1/1/2001* 6* unknown GC RS 7 1/1/2000* 7* unknown GC MW 8 1/1/1994* 13* unknown GC RM 9 1/1/1987* 20* unknown GC LV 10 5/1/2002* 5* unknown *estimated based on physical characteristics (e.g., brow length, scars, comparison to other females of known age)

Table 3.3: Bidirectionality of aggression among adult female white-faced capuchins across both seasons for LV, CP and GC groups.

Rainy Dry A Directional Inconsistency Index LV 20.0% 6.9% CP 26.7% 18.8% GC 16.9% 17.0% B Dyads Up Index LV 14.3% 0.0% CP 7.1% 0.0% GC N/A 3.2% C Counteraggression (%) LV 7.7% 0.0% CP 16.0% 3.6% GC 2.6% 3.2%

Group Season Approach Groom % time in 5m Co-feed LV Rainy 0.3562 (0.1185) -0.1077 (0.6705) 0.3859 (0.1015) N/A Dry 0.7354 (0.0825) 0.6538 (0.0445) 0.6743 (0.0725) 0.0580 (0.4315) CP Rainy 0.3490 (0.0250) 0.4022 (0.0340) 0.2428 (0.0890) N/A Dry 0.2741 (0.0520) 0.3917 (0.0335) 0.1361 (0.1690) 0.2013 (0.1940)

1.40

1.20

1.00

0.80

0.60

Rate of aggression (freq/hr) 0.40

0.20 Rainy Dry 0.00 LV (n=5) CP (n=7) GC (n=10) Group

Figure 3.1. Rates of dyadic aggression (frequency/hour) among female white-faced capuchins are shown for LV (n=5), CP (n=7) and GC (n=10) for rainy and dry seasons.

1 Kin 0.9 Non-Kin

0.8

0.7

0.6 * 0.5

0.4 Mean + SEM (freq/hr)Mean + 0.3

0.2

0.1

0 LV Rainy LV Dry CP Rainy CP Dry Group/Season

1.6 Kin NonKin 1.4

1.2

* 0.8

0.6

Mean + SEM (freq/hr)Mean + *

0.4 *

0.2

0 LV Rainy LV Dry CP Rainy CP Dry Group/Season

8.00% Kin NonKin 7.00%

6.00%

5.00%

4.00%

Mean + SEM (%) SEM + Mean 3.00%

2.00%

1.00%

0.00% LV Rainy LV Dry CP Rainy CP Dry Group/Season

0.4 Kin NonKin 0.35

0.3

0.25

0.2

0.15 Mean + SEM (freq/hr)

0.1

0.05

0 LV CP Group

DL-CH

DL-SL

DL-BB

KL-CH

KL-SL

Dyad KL-BB

SL-CH

BB-CH

BB-SL Dry KL-DL Rainy

-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 Maintenance Score

KI-ZA

TI-KI

ED-SE

SA-ZA

SA-SE

SA-ED Dyad

SI-KI

SI-TI

TI-ZA

TI-SE Dry Rainy SI-SA

-1.00 -0.80 -0.60 -0.40 -0.20 0.00 0.20 0.40 0.60 0.80 1.00 Maintenance Score

Figure 3.7. Maintenance of proximity scores are shown for each female white-faced capuchin dyad (listed as dominant-subordinate) in CP group and compared between seasons (paired t(20) = -2.604, p = 0.017). Scores range from -1 (subordinate individual was responsible for maintaining proximity) to 1 (dominant individual was responsible for maintaining proximity).

CHAPTER 4. GENERAL DISCUSSION

Investigation of social dynamics in gregarious primate species has been pivotal for understanding the costs and benefits of group-living. Research on Old World monkey species has provided the foundation on which we have built an understanding of the ecological pressures and social responses that contribute to the expression of social organization in primates (Wrangham, 1979; Wrangham, 1980; van Schaik, 1989; Isbell,

1991; Sterck et al., 1997; Isbell & Young, 2002; Koenig, 2002). Patterns of competitive interaction and the development of dominance relationships are essential components of social organization (Sterck et al., 1997). Previous research on dominance patterns has documented the occurrence of inter-group and inter-population variation as well as differences across species. Understanding this variation helps us develop models that link behavioral adaptations to environmental pressures such as predation and resource characteristics (i.e., distribution, abundance, quality, seasonality). White-faced capuchins

(Cebus capucinus) have many behavioral patterns in common with cercopithecines (a subfamily of Old World monkeys), including female philopatry, male dispersal and kin- biased interactions (Perry, 1996). Accordingly, this species offers an ideal opportunity to expand research on dominance patterns to include neotropical primate species. Research on dominance can provide a more nuanced understanding of New World primate sociality by elucidating competitive patterns and social responses to competition.

Project summary and synthesis

I investigated five key aspects of dominance behavior (hierarchical linearity, nepotism, strength, stability, and dominance style) among white-faced capuchins. My objective was to clarify the dominance patterns and range of behavioral variation exhibited by females. In Chapter 2, I examined structural aspects of female hierarchies. I was able to place all females into a highly linear rank order during the dry season.

However, rates of submission used to construct dominance hierarchies during the rainy season were lower and varied according to group size. As a result, it was difficult to reliably rank females and determine the degree of hierarchical linearity during the rainy season. Seasonal variation in resource characteristics likely contributed to this change in dominance expression.

The rank order patterns in my study population suggest that both nepotistic behavior and individualistic qualities such as age contribute to hierarchical organization.

White-faced capuchin females follow two of Kawamura’s three principles regarding matrilineal rank inheritance: mothers tend to rank above daughters and daughters rank adjacent to maternal kin. Hierarchies were nepotistic and females quickly acquired a stable rank position adjacent to maternal kin upon reaching sexual maturity. However, age also influenced rank within matrilines, as older females ranked above younger ones; therefore, Kawamura’s youngest ascendency rule was not exhibited in my study animals.

It is possible that mother capuchins, as is the case in some wild populations of macaques, do not play as important a role in their daughter’s rank acquisition as in provisioned macaque societies that experience more intense and localized competition for resources.

In my study, mothers in each matriline tended to rank above their oldest daughters, who in turn ranked above their younger sisters. The stable rank position initially acquired by a female was strongly predictive of her rank later in life. The social relationships formed by young adult female white-faced capuchins as they enter the adult hierarchy contribute to long-term hierarchical stability, as suggested by O’Brien and Robinson (1993) of relationships among juvenile/young adult female wedge-capped capuchins (C. olivaceus).

Female white-faced capuchin hierarchies were also strong. The majority of submission was resource related. The latency to detection was 21 hours/female, a relatively short detection time relative to Old World monkeys (e.g., vervets and patas,

Isbell and Young, 2002). The directional inconsistency of submission was low and at least 95% of agonistic interactions were directed down the hierarchy. These findings suggest that dominance is clearly and regularly asserted by dominants toward subordinates in competitive situations among female white-faced capuchins.

Hierarchies among females at Santa Rosa National Park show a high level of stability. Low-grade changes in group membership such as immigration, emigration, death/disappearance, or maturation of young females to adult status occurred at a fairly steady rate (average: one exit and one entry per year). The overall rank order of matrilines was not disrupted by either changes in female group membership or changes in alpha male membership. Over the past eight years in CP and ten years in LV there were no changes in the ordering of matrilines. Therefore, alpha male takeovers and low-level but on-going fluctuation in female membership do not seem to greatly affect hierarchical stability and larger scale changes. A combination of changes (e.g., male takeover and the

99 simultaneous loss of more than one female) may be required to destabilize female dominance relationships in this population.

Over a six year period, females in a neighboring population of C. capucinus at

Lomas Barbudal Biological Reserve also demonstrated hierarchical linearity (during dry season conditions) and hierarchical stability (Perry et al., 2008; Manson et al., 1999).

Lomas Barbudal females exhibit matrilineal rank inheritance and do not follow

Kawamura’s youngest ascendency rule; kin bias is also apparent in social behaviors, as in my study population (Perry et al., 2008). Although females of other capuchin species can also be placed into a rank order, there is limited literature regarding the specific aspects of hierarchical structure that I investigated for C. capucinus (e.g., linearity, nepotism, strength, and stability). Results from my study of hierarchical structure are congruent with the linear and stable hierarchies reported for wedge-capped capuchins (C. olivaceus:

O’Brien and Robinson, 1993). However, in brown capuchins (C. apella), rank reversals of matrilines following alpha male takeover have been observed despite the formation of linear hierarchies among females in this species (Janson, pers. comm.). Alpha male– female social bonds do not seem to be as strong in white-faced as in brown capuchins, which may be one reason why alpha male takeovers have a greater impact on female hierarchical stability in the latter species.

In Chapter 3, I made a preliminary assessment of dominance style by evaluating how patterns of aggression and kin-biased social behavior were expressed in relation to the hierarchies I had generated for Chapter 2. Following Thierry’s (2000) classification of dominance style into four grades from despotic to relaxed, I found that behavioral

100 patterns exhibited by female white-faced capuchins were consistent with an intermediate dominance style classification (grade 2 or 3). Seasonal changes in linearity seemed to fluctuate in accord with variation in dominance style in this study population. In the rainy season, when submission rates were low and hierarchies were difficult to determine, aggression was moderately bidirectional. Conversely, in the dry season, bidirectional aggression was almost non-existent. The higher rates of submission displayed in the dry season occurred in concert with an increase in the uni-directionality of aggressive interactions. In other words, the more “clear-cut” dominance hierarchies observed in the dry season corresponded to more aggression directed towards subordinate individuals and fewer attempts by subordinates to behave agonistically toward dominants. These findings suggest that during the dry season resources may be distributed in a way that promotes more strict and clear dominance interactions than during the rainy season.

Females showed significant kin bias in active social interactions such as approaching and grooming but these indicators of kin preference did not extend to passive social behaviors such as proximity and co-feeding. Evidence from a previous study of capuchins at the neighboring site of Lomas Barbudal suggests that behaviors promoting group cohesion, such as reconciliation, do exist for this species (e.g., one of two studies confirmed conciliatory behavior, Manson et al., 2005). A comparison of my findings with results from a review of 19 macaque species suggests that the overall expression of tension-reducing behaviors is moderate in white faced capuchins in comparison to despotic species of macaques, which lack conciliatory tendencies (de Waal

& Luttrell, 1989; Aureli et al., 1993; Chaffin et al., 1995).

101

I attempted to place my findings on capuchin dominance in the context of a four- grade scale that classified macaque species in terms of behavioral traits associated with dominance style (Thierry, 2000). Based on their aggression patterns, kin-biased behavior and conciliatory tendencies, female white-faced capuchins can be placed within grades 2 or 3 of Thierry’s scale. A more exact placement was not possible since quantitative values (and their ranges) for each grade have not been assigned. Female white-faced capuchins lack formal submissive signals, which are seen in grade 1 (despotic) macaque species, but my study animals exhibit moderate kin bias and higher levels of aggression than is characteristic of grade 4 (relaxed) macaques. Capuchins have linear, nepotistic, strong and stable hierarchies like many macaque species. However, direct comparisons between capuchins and macaques have to be made and interpreted with caution since the relevant behaviors may not co-vary in the same way in the two taxa. Perhaps in the future, comparative behavioral measures from Cebus and Saimiri could be compiled to create a framework from which to explore dominance style among New World monkeys.

Study limitations

A number of interesting issues emerged from my research that I was unable to decisively explore due to the time constraints of fieldwork and the scope of the study.

These issues include: 1) seasonal variation in dominance behavior; 2) the effect of presence of infants on submission and tolerance levels; 3) the role of triadic behavior in rank acquisition and patterning; 4) the contribution of paternal relatedness and the full extent of maternal relatedness; and 5) the effect of group size on social relationships and

102 dominance. Longer-term research on dominance should help to clarify these issues and their relevance to the larger study of social dynamics among capuchins.

My research was conducted over two field seasons. I collected a total of 3.5 months of rainy season data and 4.5 months of dry season data and found a significant difference in the expression of submission between the rainy and dry seasons. There was also a significant difference in the degree to which season affected groups of different sizes, with larger groups (CP and GC) exhibiting more of an increase in rates of submission during the dry season than the smaller group (LV). Low rates of submission during the rainy season in both CP and GC groups made testing the linearity of hierarchies difficult since they led to a larger percentage of unknown relationships due to lack of submissive interactions for certain dyads. There was also a difference in the bidirectionality of aggression between seasons, made apparent by the directional inconsistency index, dyads-up index and counteraggression measures. I now know that due to low submission rates, a greater amount of data must be collected during the rainy season to construct a reliable rank order and to test measures such as hierarchical linearity and the dyads-up index (which is reliant upon rank order). White-faced capuchins have a wide geographic distribution and not all populations live in highly seasonal habitats. To better understand the influence of seasonal ecology on intra-specific variation in dominance behavior it would be helpful to compare results from my study population to the behavior of white-faced capuchins living in regions with less seasonal variation.

Due to the limited number of females per group, and thus small sample size, I was unable to control for presence of dorsal infants during my study. Additional research

103 conducted during both seasons would allow me to determine if females behave differently with and without an infant during each season. Additional research would also allow me to clarify the style of dominance that characterizes capuchins by observing whether the level of tolerance towards lower ranking females increases when they are carrying a dorsal infant.

My study focused solely on dyadic behavioral interactions since dominance can be difficult to determine when based on conflicts involving multiple individuals.

However, the role of coalitionary activity in rank acquisition and matrilineal rank inheritance should be investigated in this population. Female capuchins often engage in triadic interactions and provide support to individuals with whom they have the closest affiliative relationships (Perry et al., 2004). Coalitionary support via triadic interactions almost certainly influences access to resources in this population, as it does in other populations (Vogel et al., 2007).

Although long-term data have been collected on LV and CP groups for 19 and 23 years respectively, relatedness among females present at the beginning of each group study was unknown. Therefore, analyses using relatedness information could only be used from 1997 on, when maternal relatedness among all females was known. Paternity analysis is currently being conducted on both of these groups and includes individuals present prior to my study. This genetic information will increase the size of the dataset and strengthen analyses of rank acquisition and matrilineal rank inheritance.

GC group is the newest addition (studied continuously since 2007) to the long- term study of white-faced capuchins at Santa Rosa National Park and it is the largest

104 group (N = 10 adult females) in this sample. Although genetic relatedness is in the process of being determined for individuals in this group, it is currently unknown and thus GC females were not included in my analyses of kin-biased behavior. Kin-biased behavior can be sensitive to group size since demographic factors (such as the proportion of related social partners) may differ across groups (Berman et al., 1997). In larger groups, individuals face time constraints in terms of the number of individuals with which they are able to interact. Among macaques, individuals in larger groups have been shown to develop social networks that include a greater proportion of close kin [e.g., rhesus macaques (Macaca mulatta), Berman et al. 1997; Japanese macaques (Macaca fuscata), Chapais et al. 1997]. The addition of GC to the Santa Rosa dataset will strengthen dominance analyses by helping us to determine whether variation in group size affects dominance patterns in capuchins to the same extent as in primate species with larger average group sizes (e.g, macaques).

Areas for future research

Although my Master’s project clarified behavioral aspects of hierarchical structure (e.g., linearity, strength and stability), it also revealed a number of directions for further research. Socioecological theory assumes that behavioral variation in social relationships results from ecological pressures. Therefore, the next step to understanding dominance relationships in white-faced capuchins is to investigate possible ecological influences. Boinski et al. (2002) have complied data to categorize the socioecology of three species of squirrel monkey (Saimiri oerstedii, S. boliviensis, and S. scuireus), the

105 taxa of New World monkey most closely related to Cebus. Classification of capuchin species using behavioral and ecological data, and comparison to Saimiri socioecology, will help us to further test predictions regarding social and ecological pressures as well as the influence of phylogenetic inertia that may contribute to the development and maintenance of social organization and competitive regimes.

Within-group competition occurs in this species and dominance rank has been shown to affect individual feeding rates (Vogel, 2005). But it is currently unknown if the dietary resources of capuchins at Santa Rosa are actually distributed in the manner predicted by the socioecological model (i.e., clumped), and if seasonal changes in food distribution, abundance and quality account for the variation in dominance I have observed in this population. It would be helpful to do further research to establish how variation in the pattern of resource distribution corresponds to changes in group spread and competitive interactions (Koenig & Borries, 2006). Determining the nutritional value of dietary resources would also help to clarify whether there are more submissive interactions over resources with higher nutritional values – or, if dominant individuals adjust their dominance style to be more tolerant of subordinates when foraging on less nutritious resources, even if the resources are present in a clumped distribution.

Once the behavioral and ecological factors surrounding dominance are better clarified, we can begin to investigate the influence of rank on stress, health, and overall fitness among female white-faced capuchins. At the proximate level, the costs of low dominance rank include restricted access to limited resources (Wrangham, 1980;

Whitten, 1983) and it is hypothesized that physiological costs associated with chronic

106 stress may result from low dominance rank and reduced access to resources. When a

‘difficult’ and unpredictable situation occurs, such as a dominance interaction involving aggressive behavior, the body produces cortisol as part of the ensuing stress response

(Sapolsky, 1993). Although an increase in cortisol can be adaptive by mobilizing energy in response to short-term stressors, sustained levels can be maladaptive by decreasing health and fitness (Sapolsky, 2005). There are two competing hypotheses regarding dominance and stress: 1) Dominant individuals have elevated cortisol due to the need to aggressively maintain and reassert their dominance (Creel, 2001; Muller & Wrangham,

2004). 2) Subordinate individuals have higher cortisol because a) they receive greater levels of aggression and b) competition reduces their access to limited resources

(Sapolsky, 1991). Research regarding levels of stress as inferred from cortisol measures will help to complete our picture of dominance by linking proximate environmental factors and social cues influencing dominance relationships to physiological consequences and the ultimate fitness of individuals.

107

REFERENCES

Alexander, R. D. 1974. The evolution of social behavior. Annual Review of Ecological

Systems, 5, 324-382.

Altmann, J. 1974. Observational study of behavior: sampling methods. Behaviour, 49,

227-265.

Altmann, J. & Alberts, S. C. 2003. Variability in reproductive success viewed from a life-

history perspective in baboons. American Journal of Human Biology, 15, 401-

409.

Appleby, M. C. 1983. The probability of linearity in hierarchies. Animal Behavior, 31,

600-608.

Archie, E. A., Morrison, T. A., Foley, C. A. H., Moss, C. J. & Alberts, S. C. 2006.

Dominance rank relationships among wild female African elephants, Loxodonta

africana. Animal Behavior, 71, 117-127.

Aureli, F., Veenema, H. C., van Panthaleon van Eck, C. J. & van Hooff, J. A. R. A. M.

1993. Reconciliation, consolation, and redirection in Japanese macaques (Macaca

fuscata). Behaviour, 124, 1-21.

Berman, C. M., Ionica, C. S. & Li, J. 2004. Dominance style among Macaca thibetana on

Mt. Huangshan, China. International Journal of Primatology, 25, 1283-1312.

Berman, C. M., Rasmussen, K. L. R. & Suomi, S. J. 1997. Group size, infant

development and social networks in free-ranging rhesus monkeys. Animal

Behavior, 53, 405-421.

108

Bernstein, I. S. 1969. Stability of the status hierarchy in a pigtail monkey group (Macaca

nemestrina). Animal Behavior, 17, 452-458.

Bernstein, I. S. 1981. Dominance: the baby and the bathwater. The Behavioral and Brain

Sciences, 4, 419-457.

Bernstein, I. S. & Williams, L. E. 1983. Ontogenetic changes in the stability of rhesus

monkey dominance relationships. Behavioural Processes, 8, 379-392.

Boinski, S. 1994. Affiliation patterns among Costa Rican squirrel monkeys. Behaviour,

130, 191-209.

Boinski, S., Sughrue, K., Selvaggi, L., Quatrone, R., Henry, M. & Cropp, S. 2002. An

expanded test of the ecological model of primate social evolution: competitive

regimes and female bonding in three species of squirrel monkeys (Saimiri

oerstedii, S. boliviensis and S. sciureus). Behaviour, 139, 227-261.

Carnegie, S. D., Fedigan, L. M. & Ziegler, T. E. 2005. Behavioral Indicators of Ovarian

Phase in White-Faced Capuchins (Cebus capucinus). American Journal of

Primatology, 67, 51-68.

Chaffin, C. L., Friedlen, K. & de Waal, F. B. M. 1995. Dominance style of Japanese

macaques compared with rhesus and stumptail macaques. American Journal of

Primatology, 35, 103-116.

Chapais, B. 1985. An experimental analysis of a mother-daughter rank reversal in

Japanese macaques (Macaca fuscata). Primates, 26, 407-423.

Chapais, B. 1992. The role of alliances in social inheritance of rank among female

primates. In: Coalitions and alliances in humans and other animals (Ed. by

109

Harcourt, A. H. & de Waal, F. B. M.), pp. 29-59. New York: Oxford University

Press.

Chapais, B. 2001. Primate nepotism:what is the explanatory value of kin selection?

International Journal of Primatology, 22, 203-229.

Chapais, B., Gaulthier, C., Prud'homme, J. & Vasey, P. 1997. Relatedness threshold for

nepotism in Japanese macaques. Animal Behavior, 53, 1089-1101.

Chapais, B., Girard, M. & Primi, G. 1991. Non-kin alliances, and the stability of

matrilineal dominance relations in Japanese macaques. Animal Behavior, 41, 481-

491.

Cheney, D. L., Andelman, S. J., Seyfarth, R. M. & Lee, P. C. 1988. Reproductive success

in vervet monkeys. In: Reproductive Success: Studies of Individual Variation in

Contrasting Breeding Systems (Ed. by Clutton-Brock, T. H.), pp. 384-402.

Chicago: University of Chicago Press.

Cheney, D. L., Lee, P. C. & Seyfarth, R. M. 1981. Behavioral correlates of non-random

mortality among free-ranging female vervet monkeys. Behavioral Ecology and

Sociobiology, 9, 153-161.

Clutton-Brock, T. H., Albon, S. D. & Guinness, F. E. 1986. Great expectations:

dominance, breeding success and offspring sex ratios in red deer. Animal

Behavior, 34, 460-471.

Combes, S. L. & Altmann, J. 2001. Status change during adulthood: life-history by-

product or kin selection based on reproductive value? Proceedings of the Royal

Society of London, B268, 1367-1373.

110

Cords, M. 2000. The agonistic and affiliative relationships of adult females in a blue

monkey group. In: Old World Monkeys (Ed. by Jolly, C. J. & Whitehead, P. F.),

pp. 453-479. Cambridge: Cambridge University Press.

Creel, S. 2001. Social dominance and stress hormones. Trends in Ecology and Evolution,

16, 491-497. de Vries, H. 1995. An improved test of linearity in dominance hierarchies containing

unknown or tied relationships. Animal Behavior, 50, 1375-1389. de Waal, F. B. M. 1977. The organization of agonistic relations within two captive

groups of Java monkeys (Macaca fascicularis). Zeitschrift Fuer Tierpsychologie,

44, 225-282. 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. 1989. Dominance 'style' and primate social organization. In:

Comparative Socioecology (Ed. by Standen, V. & Foley, R. A.), pp. 243-264.

Oxford: Blackwells. de Waal, F. B. M. & Luttrell, L. M. 1989. Toward a comparative socioecology of the

genus Macaca: different dominance styles in rhesus and stumptailed macaques.

American Journal of Primatology, 19, 83-109.

Di Fiore, A. & Campbell, C. J. 2007. The atelines: variation in ecology, behavior, and

social organization. In: Primates In Perspective (Ed. by Campbell, C. J., Fuentes,

A., MacKinnon, K. C., Panger, M. & Bearder, S. K.), pp. 155-185. Oxford:

Oxford University Press.

111

Dunbar, R. I. M. 1989. Social systems as optimal strategy sets: the costs and benefits of

sociality. In: Comparative Socioecology: The Behavioural Ecology of Humans

and Other Mammals (Ed. by Standen, V. & Foley, R. A.). Oxford: Blackwell

Scientific Publications.

Favre, M., Martin, J. G. A. & Festa-Bianchet, M. 2008. Determinants and life-history

consequences of social dominance in bighorn ewes. Animal Behavior, 76, 1373-

1380.

Fedigan, L. M. 1983. Dominance and reproductive success in primates. Yearbook of

Physical Anthropology, 26, 91-129.

Fedigan, L. M. 1993. Sex differences and intersexual relations in adult white-faced

capuchins (Cebus capucinus). International Journal of Primatology, 14, 853-877.

Fedigan, L. M., Carnegie, S. D. & Jack, K. M. 2008. Predictors of reproductive success in

female white-faced capuchins (Cebus capucinus). American Journal of Physical

Anthropology, 137, 82-90.

Fedigan, L. M. & Jack, K. M. 2001. Neotropical primates in a regenerating Costa Rican

dry forest: A comparison of howler and capuchin population patterns.

International Journal of Primatology, 22, 689-713.

Fedigan, L. M. & Rose, L. M. 1995. Interbirth Interval Variation In Three Sympatric

Species of Neotropical Monkey. American Journal of Primatology, 37, 9-24.

Fragaszy, D., Visalberghi, E. & Fedigan, L. M. 2004. The Complete Capuchin: The

Biology of the Genus Cebus. Cambridge: Cambridge University Press.

112

Freeland, W. J. 1976. Pathogens and the evolution of primate sociality. Biotropica, 8, 12-

24.

Furuichi, T. 1983. Interindividual distance and influence of dominance on feeding in a

natural Japanese macaque troop. Primates, 24, 445-455.

Gouzoules, H., Gouzoules, S. & Fedigan, L. 1982. Behavioural dominance and

reproductive success in female Japanese monkeys (Macaca fuscata). Animal

Behavior, 30, 1138-1150.

Grunau, T. & Kuester, J. 2001. Dominance style in female guerezas (Colobus guereza

RUPPELL 1985). Primates, 42, 301-307.

Hausfater, G., Altmann, J. & Altmann, S. 1982. Long-term consistency of dominance

relations among female baboons (Papio cynocephalus). Science, 217, 752-754.

Hemelrijk, C. K. 1990. A matrix partial correlation test used in investigations of

reciprocity and other social interaction patterns at group level. Journal of

Theoretical Biology, 143, 405-420.

Hemelrijk, C. K. & Gygax, L. 2004. Dominance style, differences between the sexes and

species. Interaction Studies, 5, 131-146.

Hill, D. A. & Okayasu, N. 1995. Absence of 'youngest ascendancy' in the dominance

relations of sisters in wild Japanese macaques (Macaca fuscata yakui). Behaviour,

132, 367-379.

Hinde, R. A. 1976. Interactions, relationships and social structure. Man, 11, 1-17.

Horrocks, J. & Hunte, W. 1983. Maternal rank and offspring rank in vervet monkeys: An

appraisal of the mechanisms of rank acquisition. Animal Behavior, 31, 772-782.

113

Isbell, L. A. 1991. Contest and scramble competition: patterns of aggression and ranging

behavior among primates. Behavioral Ecology, 2, 143-155.

Isbell, L. A. 2004. Is there no place like home? Ecological bases of female dispersal and

philopatry and their consequences for the formation of kin groups. In: Kinship

and Behavior in Primates (Ed. by Chapais, B. & Berman, C. M.), pp. 71-108.

Oxford: Oxford University Press.

Isbell, L. A., Pruetz, J. D. & Young, T. P. 1998. Movements of vervets (Cercopithecus

aethiops) and patas monkey (Erythrocebus patas) as estimators of food resource

size, density, and distribution. Behavioral Ecology and Sociobiology, 42, 123-

133.

Isbell, L. A. & Young, T. P. 2002. Ecological models of female social relationships in

primates: similarities, disparities, and some directions for future clarity.

Behaviour, 139, 177-202.

Izar, P., Ferreira, R. G. & Sato, T. 2006. Describing the organization of dominance

relationships by dominance-directed tree method. American Journal of

Primatology, 68, 189-207.

Jack, K. M. 2007. The cebines: toward an explanation of variable social structure. In:

Primates In Perspective (Ed. by Campbell, C. J., Fuentes, A., MacKinnon, K. C.,

Panger, M. & Bearder, S. K.), pp. 107-123. New York, Oxford: Oxford

University Press.

114

Jack, K. M. & Fedigan, L. 2004. Male dispersal patterns in white-faced capuchins, Cebus

capucinus part 1: Patterns and causes of natal emigration. Animal Behaviour, 67,

761-769.

Kappeler, P. M. & van Schaik, C. P. 2002. Evolution of primate social systems.

International Journal of Primatology, 23, 707-740.

Kapsalis, E. 2004. Matrilineal kinship and primate behavior. In: Kinship and Behavior in

Primates (Ed. by Chapais, B. & Berman, C. M.), pp. 153-176. Oxford: Oxford

University Press.

Kaufmann, J. H. 1983. On the definitions and functions of dominance and territoriality.

Biological Reviews, 58, 1-20.

Kawai, M. 1958. On the rank system in a natural group of Japanese monkey. Primates, 1,

111-133.

Kawamura, S. 1958. Matriarchal social ranks in the Minoo-B troop: A study of the rank

system in Japanese monkeys. Primates, 1, 149-156.

Koenig, A. 2002. Competition for resources and its behavioral consequences among

female primates. International Journal of Primatology, 23, 759-783.

Koenig, A., Beise, J., Chalise, M. K. & Ganzhorn, J. U. 1998. When females should

contest for food - testing hypotheses about resource density, distribution, size, and

quality with Hanuman langurs (Presbytis entellus). Behavioral Ecology and

Sociobiology, 42, 225-237.

Koenig, A. & Borries, C. 2006. The predictive power of socioecological models: a

reconsideration of resource characteristics, agonism, and dominance hierarchies.

115

In: Feeding Ecology in Apes and Other Primates (Ed. by Hohmann, G., Robbins,

M. M. & Boesch, C.), pp. 261-282. Cambridge: Cambridge University Press.

Kutsukake, N. 2000. Matrilineal rank inheritance varies with absolute rank in Japanese

macaques. Primates, 41, 321-335.

Leca, J. B., Fornasieri, I. & Petit, O. 2002. Aggression and reconciliation in Cebus

capucinus. International Journal of Primatology, 23, 979-998.

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., Rose, L. M., Perry, S. & Gros-Louis, J. 1999. Dynamics of female-female

relationships in wild Cebus capucinus: data from two Costa Rican Sites.

International Journal of Primatology, 20, 679-701.

Marsden, H. M. 1968. Agonistic behaviour of young rhesus monkeys after changes

induced in social rank of their mothers. Animal Behavior, 16, 38-44.

Martin, P. C. & Bateson, P. 1993. Measuring Behaviour. Cambridge: Cambridge

University Press.

Mason, W. A. 1961. The effects of social restriction on the behavior of Rhesus monkeys.

III. Dominance tests. Journal of Comparative Physiology and Psychology, 54,

694-699.

Maynard Smith, J. & Parker, G. A. 1976. The logic of asymmetric contests. Animal

Behavior, 24, 159-175.

Maynard Smith, J. & Price, G. R. 1973. The logic of animal conflict. Nature, 246, 15-18.

116

Mitchell, C. L. 1994. Migration alliances and coalitions among adult male South

American squirrel monkeys (Saimiri sciureus). Behaviour, 130, 169-190.

Muller, M. N. & Wrangham, R. W. 2004. Dominance, cortisol and stress in wild

chimpanzees (Pan troglodytes schweinfurthii). Behavioral Ecology and

Sociobiology, 55, 332-340.

Nakamichi, M., Itoigawa, N., Imakawa, S. & Machida, S. 1995. Dominance relations

among adult females in a free-ranging group of Japanese monkeys at Katsuyama.

American Journal of Primatology, 37, 241-251.

Noldus Information Technology. 1998. Reference manual. MatMan 1.0.

O'Brien, T. G. & Robinson, J. G. 1993. Stability of social relationships in female wedge-

capped capuchin monkeys. In: Juvenile Primates: Life History, Development, and

Behavior (Ed. by Periera, M. E. & Fairbanks, L. A.), pp. 197-210. New York:

Oxford University Press.

Parker, G. A. 1974. Assessment strategy and the evolution of fighting behavior. Journal

of Theoretical Biology, 47, 223-243.

Pavelka, M. S. M. 1990. Do old female monkeys have a specific social role? Primates,

31, 363-373.

Perry, S. 1996. Female-female social relationships in wild white-faced capuchin

monkeys, Cebus capucinus. American Journal of Primatology, 40, 167-182.

Perry, S. 1998. Male-male social relationships in wild white-faced capuchins, Cebus

capucinus. Behaviour, 135, 139-172.

117

Perry, S., Barrett, C. & Manson, J. H. 2004. White-faced capuchin monkeys show triadic

awareness in their choice of allies. Animal Behaviour, 67, 165-170.

Perry, S., Manson, J. H., Muniz, L., Gros-Louis, J. & Vigilant, L. 2008. Kin-biased social

behaviour in wild adult female white-faced capuchins (Cebus capucinus). Animal

Behaviour, 76, 187-199.

Phillips, K. A. 1995. Resource patch size and flexible foraging in white-faced capuchins

(Cebus capucinus). International Journal of Primatology, 16, 509-519.

Pusey, A., Williams, J. & Goodall, J. 1997. The influence of dominance rank on the

reproductive success of female chimpanzees. Science, 277, 828-831.

Ratnayeke, S. 1994. The behavior of postreproductive females in a wild population of

toque macaques (Macaca sinica) in Sri Lanka. International Journal of

Primatology, 15, 445-469.

Reichler, S., Ducker, S. & Preuschoft, S. 1998. Agonistic behavior and dominance style

in gelada baboons (Theropithecus gelada) [Abstract]. Folia Primatologica, 69,

200.

Rhine, R. J., Cox, R. L. & Costello, M. B. 1989. A twenty-year study of long-term and

temporary dominance relations among stumptailed macaques (Macaca arctoides).

American Journal of Primatology, 19, 69-82.

Robbins, M. M., Robbins, A. M., Gerald-Steklis, N. & Steklis, H. D. 2005. Long-term

dominance relationships in female mountain gorillas: strength, stability, and

determinants of rank. Behaviour, 142, 779-809.

118

Roney, J. R. & Maestripieri, D. 2003. Social development and affiliation. In: Primate

Psychology (Ed. by Maestripieri, D.), pp. 171-204. Cambridge, MA: Harvard

University Press.

Rose, L. M. 1994. Benefits and costs of resident males to females in white-faced

capuchins, Cebus capucinus. American Journal of Primatology, 32, 235-248.

Rose, L. M. 1997. Vertebrate predation and food sharing in Cebus and Pan. International

Journal of Primatology, 18, 727-765.

Rowell, T. E. 1974. The concept of social dominance. Behavioral Biology, 11, 131-154.

Sade, D. S. 1967. Determinants of dominance in a group of free-ranging rhesus monkeys.

In: Social Communication Among Primates (Ed. by Altmann, S.), pp. 99-114.

Chicago: The University of Chicago Press.

Sade, D. S. 1972. A longitudinal study of social behavior of rhesus monkeys. In:

Functional and Evolutionary Biology of Primates (Ed. by Tuttle, R.), pp. 378-

398. Chicago: Aldine.

Samuels, A., Silk, J. B. & Altmann, J. 1987. Continuity and change in dominance

relationships among female baboons. Animal Behaviour, 35, 785-793.

Sapolsky, R. M. 1991. Testicular function, social rank and personality among wild

baboons. Psychoneuroendocrinology, 16, 281-293.

Sapolsky, R. M. 1993. The physiology of dominance in stable versus unstable

hierarchies. In: Primate social conflict (Ed. by Mason, W. A. & Mendoza, S. P.),

pp. 171-204. Albany: SUNY Press.

119

Sapolsky, R. M. 2005. The influence of social hierarchy on primate health. Science, 308,

648-652.

Seyfarth, R. M. 1980. The distribution of grooming and related behaviors among adult

female verve monkeys. Animal Behaviour, 28, 798-813.

Silk, J. B., Samuels, A. & Rodman, P. S. 1981. Hierarchical organization of female

Macaca radiata. Primates, 22, 84-95.

Sterck, E. H. M., Watts, D. P. & van Schaik, C. P. 1997. The evolution of female social

relationships in nonhuman primates. Behavioral Ecology and Sociobiology, 41,

291-309.

Thierry, B. 1985. A comparative study of aggression and response to aggression in three

species of macaque. In: Primate ontogeny, cognition, and social behaviour (Ed.

by Else, I. & Lee, P. C.). Cambridge: Cambridge University Press.

Thierry, B. 1990. Feedback loop between kinship and dominance: the macaque model.

Journal of Theoretical Biology, 145, 511-521.

Thierry, B. 2000. Covariation of conflict management patterns across macaque

species. In: Natural Conflict Resolution (Ed. by Aureli, F. & de Waal, F. B. M.),

pp. 106-128. Berkley: University of California Press.

Trivers, R. L. 1972. Parental investment and sexual selection. In: Sexual Selection and

the Descent of Man (Ed. by Campbell, B.), pp. 136-179. Chicago: Aldine-

Atherton.

Valderrabano-Ibarra, C., Brumon, I. & Drummond, H. 2007. Development of a linear

dominance hierarchy in nestling birds. Animal Behavior, 74, 1705-1714.

120 van Noordwijk, M. A. & van Schaik, C. P. 1987. Competition among female long-tailed

macaques, Macaca fascicularis. Animal Behavior, 35, 577-589. van Schaik, C. P. 1983. Why are diurnal primates living in groups? Behaviour, 87, 120-

144. van Schaik, C. P. 1989. The ecology of social relationships among female primates. In:

Comparative Socioecology: the Behavioral Ecology of Humans and Other

Mammals (Ed. by Standen, V. & Foley, R. A.), pp. 195-218. Oxford: Blackwell

Press.

Veenema, H. C., Spruijt, B. M., Gispen, W. H. & van Hooff, J. A. R. A. M. 1997. Aging,

dominance history, and social behavior in Java-monkeys (Macaca fascicularis).

Neurobiology of Aging, 18, 509-515.

Vogel, E. R. 2005. Rank differences in energy intake rates in white-faced capuchin

monkeys, Cebus capucinus: the effects of contest competition. Behavioral

Ecology and Sociobiology, 58, 333-344.

Vogel, E. R., Munch, C. H. & Janson, C. H. 2007. Understanding escalated aggression

over food resources in white-faced capuchin monkeys. Animal Behaviour, 74, 71-

80.

Whitten, P. L. 1983. Diet and dominance among female vervet monkeys (Cercopithecus

aethiops). American Journal of Primatology, 5, 139-159.

Wrangham, R. W. 1979. On the evolution of ape social systems. Social Science

Infomation, 18, 335-368.

121

Wrangham, R. W. 1980. An ecological model of female-bonded primate groups.

Behaviour, 75, 262-300.

APPENDIX 1: Ethogram of social behaviors recorded during continuous focal follows and ad libitum for white-faced capuchins.

Affiliative Approach Movement of one individual into close proximity of another individual Groom Picking through the fur of another individual in attempt to remove parasites and insects Groom request Gesture to receive grooming by laying down or crouching near another individual Touch Purposely making contact with another individual, usually with the hand Inspect Visually or manually searching part of another’s body (i.e., putting one's face near another individual) Climb Crawling on top of another with benign/friendly intent Suck Placing one's mouth on another individual's body part and moving mouth to create suction Hug Standing or sitting in close contact with one or both arms around another individual Grope Continuously rubbing against another individual, usually to absorb a substance from an object Palpate nipple Gently touching/manipulating another individual’s nipple Aggressive (Intense) Bite Forcefully closing mouth on an individual; contact with teeth Chase Pursuing another individual at a rapid rate Hit Using one's hand to make forceful, physical contact with another individual Pull Grabbing another individual and bringing them toward oneself in a forceful manner Push Forceful physical contact with motion to move an individual away from oneself Lunge An intense motion toward another individual without physical contact Pounce Forcefully jumping directly onto another individual Wrestle Intense aggression and physical contact from one individual to another involving tumbling/rolling and entwined limbs; often defensively or aggressively reciprocated Intense vocal threat Loud and intense pulses, roars, and shrieks; varies by individual Cough threat An abrupt and forced exhale to create a short one-syllable vocal burst Aggressive (Mild) Swipe Motion to grab or hit at another individual 122

Bounce Jumping up and down while making eye contact with another individual Glare Directing one's gaze at another individual; often involves furrowing of the brow Open-mouth threat Staring gaze towards another individual while simultaneously opening one's mouth to expose the canines Nip Un-forcefully closing of one's mouth on another individual; contact with teeth Snap at An attempt/warning gesture to bite or nip another individual Tooth grind Low friction noise caused by forceful clenching and movement of upper and lower teeth against each other Mild vocal threat Less intense pulses, roars, and shrieks compared to intense vocal threat; varies by individual Submissive Avoid Actively moves away from another individual Cower Motion to shudder or withdraw from another individual Flee Moving away/running from another individual rapidly Grimace Horizontally widening of mouth and tensing of lips; teeth are exposed and clenched Head shake Quick and sometimes repetitive movement of head back and forth Wrinkle face Scrunching or tensing facial muscles in response to another individual Yelp Short cry; usually in response to being startled Scream Cry of distress; depending on intensity can range in pitch, tone, and duration Supplantation Taking the place of an individual after they move, without coming into contact Sexual Duck face Wrinkling one's face, purse lips together and protruding them from the face while directing one's gaze towards another individual Pace Movement/walking back and forth while directing gaze towards another individual Pirouette Spinning or turning one's body while attempting to maintain gaze towards another individual Mount Obtaining a superior dorsal/ventral position over another individual Copulate Intercourse between two individuals involving mounting, thrusting and intromission Ejaculate Excretion of semen from the male's penis, usually during copulation Locomote Movement/walking by the individual in the inferior dorsal/ventral mounted position 123 Bottom thrust Movement of hips back and forth by the individual in the inferior dorsal/ventral mounted position

Dismount Removing oneself from a superior dorsal/ventral position over another individual Squeak vocalization High-pitched vocal bursts/squeals uttered while engaged in sexual behaviors Grunt vocalization Low-pitched "throaty" vocal bursts uttered while engaged in sexual behaviors Social Food Co-feeding Consuming food items within close proximity to another individual Co-feeding rejection Attempting to consume food items within close proximity to another individual but deterred by that individual usually through aggressive behavior Food guard Individual with food moves away or into a position to protect food item from another individual Food handle Touching another individual’s food item Food interest Visually inspecting or attempting to touch another's food item Food push Forceful physical contact with motion to move an individual away from one's food item Food share Giving part or all of one's food item to another individual Tolerated theft Allowing another individual to take one's food item without contesting Untolerated theft Having one's food item forcefully taken by another individual Coalitionary Head flag Soliciting help from another individual by quickly looking at them and then toward an opponent Embrace Standing next to/putting one's arm around another individual and both directing gazes toward an opponent Cheek to cheek Soliciting help from an individual by pressing one's cheek to theirs and directing gazes toward an opponent Overlord Climbing on top of another individual and directing gazes toward an opponent Hip grasp request Grabbing another individual by the hips and pulling them toward oneself to initiate an overlord Back into request Backing up toward another individual to initiate an overlord against an opponent Play Bite Closing one's mouth on an individual; contact with teeth but in the context of play Chase Pursuing another individual at a rapid rate in the context of play Flee Movement away/runs from another individual rapidly in the context of play Hit Using one's hand to make forceful, physical contact with another individual in the context of play 124 Invite/Face

Bounce Jumping up and down while making eye contact with another individual in the context of play Chicken fight Two individuals hang upside-down by their tales and engage in one (or many) play behavior(s) Pull Grabbing another individual and bringing them toward oneself in a forceful manner in the context of play Overlord Climbing on top of another individual and directing gazes toward an opponent in the context of play Push Forceful physical contact with motion to move an individual away from oneself in the context of play Lunge An intense motion toward another individual without physical contact in the context of play Staring gaze towards another individual plus open mouth expression which exposes the canines in the context of Threat play Pounce Forcefully jumping directly onto another individual in the context of play Wrestle Intense physical contact from one individual to another involving tumbling/rolling and entwined limbs; usually reciprocated in the context of play Infant related Bridge Creating a self-infant pathway over broken terrain/canopy using one's body part, usually an arm or leg Mount Allowing an infant to assume a dorsal position Fetch Retrieving an infant from another individual Invite Crouching to allow an infant to approach or mount Nurse Suckling/feeding on milk from the nipple of another individual Reject nursing Attempting to suckle/feed on milk from an individual who is not lactating or responds to deter behavior Tantrum Flailing of infant's limbs/tail, often accompanied by screaming vocalizations, in response to social interaction Restrain Using one's arm to keep an infant in a dorsal/mounted position Wrestle Attempting to remove an infant from a dorsal/mounted position 125

APPENDIX 2: State behaviors and proximity classifications recorded for scan samples within focal samples of white-faced capuchins.

Behavioral State Affiliative Engaging in friendly social behavior Aggressive Engaging in aggressive social behavior Coalitionary Individual is social and interacting triadically or with >3 individuals Drink Drinking from a water source; source can be specified as river, tree hole, or unnatural (i.e., provisioned) source Forage Chewing or manipulating a food item; item is specified as insect, leaf, fruit, vertebrate prey, or other Forage Travel Chewing or manipulating food items while locomoting Forage Visually Visually scanning food items and surrounding environment Groom Engaging in grooming-related social behavior Infant Interacting with another individual's infant or nursing/carrying one's own infant Play Engaging in social play Rest Lying or sitting Self-direct Focusing behavior toward one’s own body, such as autogrooming or scratching Sexual Engaging in sexual social behavior Sleep Sleeping Social Food Engaging in a social food-related behavior within 5m of another individual Stand Standing still Submissive Engaging in fearful social behavior Travel Locomoting

126

Travel Vigilant Cautiously locomoting while actively scanning the surrounding environment Proximity Category Contact All individuals that are within physical contact with focal Within 1 All individuals that are within 0.5m of focal (~1 body length) Within 5 All individuals that are within 2.5m of focal (~5 body lengths) Within 10 All individuals that are within 5.0m of focal (~10 body lengths) Dorsal Carry Infant that focal may be carrying Nurse Receive Infant that may be nursing from focal

127