University of Southern California Dissertations and Theses

Endocrine aspects of aggression and dominance in champanzees of the Kibale Forest

(USC Thesis Other)

Endocrine aspects of aggression and dominance in champanzees of the Kibale Forest

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content ENDOCRINE ASPECTS OF AGGRESSION AND DOMINANCE IN CHIMPANZEES OF IHE KIBALE FORES I by Martin N. Muller A Dissertation Presented to the F a c u l t y of th e G r a d u a t e S c h o o l U n iv e r s it y of S o u t h e r n C a l if o r n ia in Partial Fulfillment of the Requirements for the Degree D o c t o r of P h il o so p h y (A n t h r o p o l o g y ) May 2002 Copyright 2002 ' Martin N. Muller UNIVERSITY OF SOUTHERN CALIFORNIA The Graduate School University Park LOS ANGELES, CALIFORNIA 90089-1695 This d isserta tio n , w ritte n b y Martin N. Muller U nder th e d irectio n o f h.X&.. D isserta tio n C om m ittee, a n d a p p ro ved b y a ll its m em bers, has been p re se n te d to a n d a c c e p te d b y The G raduate School, in p a rtia l fu lfillm en t o f requ irem en ts fo r th e degree o f D O C TO R OF PH ILOSOPH Y 'an o f Graduate Studies DISSER TA TIO N COM M ITTEE Chairperson A c k n o w l e d g e m e n t s I am indebted to a large number of people for their assistance in planning and executing this study. At University of Southern California, I thank Craig Stanford, Christopher Boehm and Jane Goodall for inspiring my initial interest in chimpanzees, and facilitating my first trip to Africa to see them in the wild. I am particularly grateful to my advisor Craig Stanford for the extraordinary degree of autonomy that he granted me in pursuing this research. I am grateful to Richard Wrangham for inviting me to work in Kanyawara, and for first suggesting this study. He has been a patient and accomodating advisor, and I thank him, Elizabeth Ross, Ross, David, Ian and Barra for providing me with homes away from home in Uganda and Massachusetts. Fieldwork in Kibale is sponsored by Makerere University and the Uganda Wildlife Authority. I thank Gilbert Isabirye Basuta, John Kasenene, Kato Innocent, and the staff of the Makerere University Biological Field Station for assistance in the park. This research was generously funded by grants from the L.S.B. Leakey Foundation and the U.S. National Science Foundation (SBR-9729123 and SBR- 9807448). During the writing of this dissertation I was partially supported by a Haynes Foundation fellowship from University of Southern California. This research would not have been possible without long hours of hard work by the Kibale Chimpanzee Project’s Ugandan field assistants: John Barwogeza, Deo Kateeba, Christopher Katongole, Francis Mugurusi, Donor Muhangyi, Christopher 1 1 Muruuli and Peter Tuhairwe. I am grateful to my co-manager in the field, Mike Wilson, for shouldering the bulk of the administrative duties during 1998. I thank Sam Mugume for help managing the project, and Rachel Carmody and Geertrui Spaepen for assistance processing urine samples. Numerous researchers at Kibale offered logistical support, companionship, and hospitality in the field. For these I am indebted to Chrissy Apodaca, Keri Boomgarden, Colin and Lauren Chapman, Roger and Debbi Fonts, Chris Hoefer, Carole Hooven, John Mitani, William Olupot, John Paul, April Randle, Nat Seavy and David Watts. In Kampala, Oscar and Linda Rothen generously allowed me to stay in their home, and to store chimpanzee urine in their freezer. Debbie Cox also provided hospitality at the Wildlife Education Centre in Entebbe. At Harvard I thank Leah Domb for first suggesting that I go to Kibale. I am grateful to Peter Ellison for providing me with the resources of his Reproductive Ecology Laboratory and advice on data analysis. Throughout this project Susan Lipson offered invaluable assistance in the laboratory and exhibited exceptional patience when I used all of her reagents. I thank Grazyna Jasienska for introducing me to the radioimmunoassay procedures, and Cheryl Knott, Christina Perez and Dipti Kandlika for additional assistance in the laboratory. Mike Wilson provided pant-grunt observations from Kanyawara. Diana Sherry entered the Kanyawara feeding data into the computer. Randall Collura provided helpful computing advice. Ill At University of Southern California, the chair of the anthropology department, Zandy Moore, was extremely accommodating of my extended absences from campus. Rand Wilcox offered statistical guidance. William McClure provided valuable advice as the outside member of my dissertation committee. For helpful discussions and/or conunents on specific chapters in this dissertation, I thank Melissa Emery, Peter Gray, Sonya Kahlenberg, John Mitani, and Sherry Nelson. Ross Wrangham assisted on numerous aspects of this project, from the data collection in Kibale, to the laboratory work at Harvard. He also provided me with a welcome distraction from the writing of this dissertation by introducing me to the dolphins of Monkey Mia, Western Australia. I am grateful to him for all of his hard work, and for his friendship over the past five years. I am particularly indebted to Sherry Nelson for her patience and support during the writing of this dissertation. Finally, I am grateful to my parents, Martin and Marlene Muller, for years of unfailing encouragement and support. IV T a b l e o f C o n t e n t s PRRIJMTNARY PAGES 1. Acknowledgements 2. List of tables 3. List of figures C h a p t e r 1 : In t r o d u c t io n 1. Objectives 2. Steroid hormones 3. Testosterone a. Background b. T and aggression c. T and sexual behavior d. Other factors 4. Cortisol 5. Field studies of endocrine function in the great apes 6. Organization C h a p t e r 2: M e t h o d s 1. Study site 2. Study population 3. Behavioral data collection a. Field-assistant data b. Aggression data 4. Food availability and energy balance 5. Urine collection and preservation 6. Sample contamination 7. Hormone analysis 8. Statistical procedures C h a p t e r 3: P a t t e r n s of A g g r e ss io n 1. Introduction 2. Results a. Sex differences in aggression b. Dominance relationships c. Patterns of male aggression d. Intercommunity aggression 11 viii X 1 3 4 4 5 6 7 7 8 10 12 13 16 16 18 21 22 23 25 28 30 33 33 38 40 44 3. Discussion and conclusions 51 a. Female-female competition 51 b. Male-male competition 53 c. Intercommunity aggression 58 C h a p t e r 4: P a t t e r n s of S t e r o id P r o d u c t io n 1. Introduction 61 2. Methods 68 3. Results 70 4. Discussion and conclusions 87 a. Patterns of T and behavior 89 b. Diurnal patterns of sexual behavior 91 c. Diurnal patterns of aggression 94 C h a p t e r 5: T e s t o st e r o n e a n d E n e r g e t ic s 1. Introduction 96 2. Methods 103 a. T and chronic energetic status 103 b. T and acute changes in energetic status 104 3. Results 106 a. T and chronic energetic status 106 b. T and acute changes in energetic status 106 4. Discussion and conclusions 113 a. T and chronic energetic status 113 b. T and acute changes in energetic status 116 C h a p t e r 6: T h e C h a l l e n g e H y p o t h e sis 1. Introduction 121 2. Seasonally breeding primates and the challenge hypothesis 124 3. Male-male competition in the great apes 129 a. Mountain gorillas 130 b. Orangutans 132 c. Chimpanzees 136 d. Bonobos 139 e. Humans 140 4. Methods 144 a. Male aggression during mating competition 144 b. Male T during reproductive competition 146 5. Results 147 a. Male aggression during mating competition 147 b. Male T during reproductive competition 156 6. Discussion and conclusions 165 V I C h a p t e r 7: T e st o st e r o n e a n d D o m in a n c e 1. Introduction 172 2. Dominance and reproductive success in the great apes 175 a. Mountain gorillas 175 b. Orangutans 176 c. Chimpanzees 176 3. T, dominance and social stability 180 4. Methods 182 a. Determining dominance ranks 182 b. T measurements 184 5. Results 185 6. Discussion and conclusions 186 a. T and rank 186 b. T and the costs of social dominance 195 C h a p t e r 8: S t r e s s, D o m in a n c e a n d A g g r e ss io n 1. Introduction 197 2. Stress hormones and social dominance 199 3. Methods 204 a. Cortisol and dominance 204 b. Male cortisol and reproductive competition 204 c. Male sexual aggression and female stress 205 4. Results 206 a. Cortisol and dominance 206 b. Male cortisol and reproductive competition 204 c. Male sexual aggression and female stress 205 5. Discussion and conclusions 217 B ib l io g r a p h y 221 A p p e n d ix I: L a b o r a t o r y P r o c e d u r e s 246 1. Testosterone hydrolysis procedure 246 2. Testosterone extraction procedure 247 3. Testosterone assay set-up 248 4. Cortisol assay set-up 249 5. Charcoal procedure 250 6. Creatinine procedure 251 V ll L is t o f T a b l e s Table 2.1 Members of the Kanyawara community in 1998 15 Table 2.2 Foi*ty-minute aggression focals by adult male 19 Table 3.1 Frequency of attacks by males and females at two sites 35 Table 3.2 Contexts of aggression in three long-term study sites 36 Table 3.3 Victims of directed aggression in two long-term study sites 39 Table 3.4 Reconciled conflicts in wild and captive chimpanzees 45 Table 3.5 Intercommunity relationships in six long-term study sites 47 Table 4.1 Means of mean male T levels by time of day 75 Table 4.2 Means of mean male cortisol levels by time of day 79 Table 4.3 Copulation frequency and time of day at Kanyawara 85 Table 5.1 Energy availability and reproductive function in hominoids 99 Table 5.2 Mean urinary T levels for wild and captive chimpanzees 108 Table 5.3 Mean levels of T and creatinine by quarter 110 Table 5.4 Significance levels from pairwise comparisons of mean 110 quarterly T and creatinine Table 5.5 Monthly feeding time on major fruits at Kanyawara 111 Table 5.6 Percentage of daily samples from nine males collected on 119 days when fully swollen parous females were present Table 6.1 Mating competition and T variation in two seasonally 128 breeding primates Table 6.2 Individual rates of male aggression in reproductive and 148 non-reproductive contexts Table 6.3 Mean am T levels for 11 adult males in non-reproductive 157 and reproductive contexts Table 6.4 Mean am T values for 11 adult males in reproductive 160 and non-reproductive contexts Table 6.5 Mean male T in the presence of maximally tumescent 163 nulliparous females Table 7.1 Dominance levels for 11 Kanyawara males 189 Table 7.2 Dominance ranks for 11 Kanyawara males 189 Table 7.3 Mean morning and afternoon T values for 11 adult males 190 Vlll Table 8.1 Glucocorticoids and dominance rank in field studies of 202 social vertebrates Table 8.2 Mean morning and afternoon cortisol values for 11 207 adult males Table 8.3 Mean morning cortisol values for 11 adult males 212 in reproductive and non-reproductive contexts Table 8.4 Mean cortisol levels for two females in swelling and 215 non-swelling contexts IX L is t o f F ig u r e s Figure 3.1 Mean rates of aggression for male and female chimpanzees 34 Figure 3.2 Dominance rank and display rate in Kanyawara males 42 Figure 3.3 Frequency of aggression in relation to the number of adult 43 males in a party Figure 3.4 Ventral wounds on the extracommunity male killed by 49 male chimpanzees at Kanyawara Figure 3.5 Dorsal wounds on the extracommunity male killed by 50 male chimpanzees at Kanyawara Figure 4.1 Salivary T levels and age in four human populations 65 Figure 4.2 Steroid excretion and time of day 72 Figure 4.3 T and time of day 73 Figure 4.4 T and time of day in one male (LK) 74 Figure 4.5 Means of mean male T by time of day 76 Figure 4.6 Cortisol and time of day 77 Figure 4.7 Cortisol and time of day in one male (LK) 78 Figure 4.8 Means of mean male cortisol by time of day 80 Figure 4.9 Morning T and age in adult males 82 Figure 4.10 Male aggression and time of day 83 Figure 4.11 Male aggression and time of day including additional 84 observations Figure 4.12 Diurnal pattern of sexual behavior at Kanyawara 86 Figure 4.13 Testes and brain of an adult male chimpanzee 93 Figure 5.1 Hypothetical adaptive relationship between T and energetics 102 Figure 5.2 Mean urinary T levels in male chimpanzees 109 Figure 5.3 Chimpanzee feeding patterns in Kanyawara 112 Figure 6.1 The challenge hypothesis 122 Figure 6.2 Individual rates of total aggression in non-reproductive 149 and reproductive contexts Figure 6.3 Means rates of total aggression in non-reproductive versus 150 reproductive contexts Figure 6.4 Individual rates of high-level aggression in non-reproductive 152 and reproductive contexts Figure 6.5 Mean rates of high-level aggression in non-reproductive 153 versus reproductive contexts Figure 6.6 Mean rates of aggression during the swelling phase of 154 one female’s cycle Figure 6.7 Intensity of aggression in reproductive and non- 155 reproductive contexts Figure 6.8 Individual male T levels in reproductive and non- 158 reproductive contexts Figure 6.9 Means of mean male T levels in reproductive and non- 159 reproductive contexts Figure 6.10 Individual male T levels during and immediately following 161 reproductive competition Figure 6.11 Means of mean male T levels during and immediately 162 following reproductive competition Figure 6.12 Means of mean male T during mating periods with little 164 aggressive competition Figure 6.13 A theoretical version of the challenge hypothesis for 171 chimpanzees Figure 7.1 Pant-grunts among adult males at Kanyawara 187 Figure 7.2 Aggressive interactions among adult males at Kanyawara 188 Figure 7.3 Dominance level and mean morning T in Kanyawara males 191 Figure 7.4 Dominance level and mean afternoon T in Kanyawara males 192 Figure 7.5 Dominance rank and mean afternoon T in Kanyawara males 193 Figure 8.1 Dominance level and mean morning cortisol in Kanyawara 208 males Figure 8.2 Dominance rank and mean morning cortisol in Kanyawara 209 males Figure 8.3 Dominance level and mean afternoon cortisol in Kanyawara 210 males Figure 8.4 Dominance rank and mean afternoon cortisol in Kanyawara 211 males Figure 8.5 Individual male cortisol levels in reproductive and non- 213 reproductive contexts Figure 8.6 Means of mean male cortisol levels in reproductive and 214 non-reproductive contexts Figure 8.7 Female cortisol levels in swelling and non-swelling contexts 216 X I 1 In t r o d u c t io n Objectives In this study I examine the relationship between social behavior and the production of steroid hormones in wild chimpanzees (Pan troglodytes schweinfurthii). Specifically, I address the association between testosterone (T), cortisol, and aggression. The rationale is two-fold. First, elucidating the relationship between T, cortisol, and aggression will contribute to our understanding of chimpanzee social dynamics. Chimpanzees exhibit frequent aggression against a range of competitors in a variety of contexts. I examine both the costs and benefits of agonistic behavior, and attempt to quantify the former using measurements of steroid production. The goal is to understand both the proximate mechanisms involved in promoting aggression, and the ultimate evolutionary function of aggression. Second, characterizing the relationship between hormones and aggression in chimpanzees will contribute to the development of general principles that apply to other species with complex social behavior. Among male baboons {Papio anubis), correlations between endocrine physiology and social behavior vary according to the dominance style of the individual (e.g. Sapolsky and Ray 1989, Ray and Sapolsky 1992, Sapolsky 1992). Therefore, the style of dominance at the species level is also expected to influence relationships between hormones and behavior. Patterns of aggression and dominance in baboons differ significantly from those of chimpanzees. Accordingly, this study offers new insights into the endocrinology of primate aggression and dominance. This is the first study to examine the socioendocrinology of chimpanzees in the wild. One problem with field studies of behavioral endocrinology is that myriad environmental factors can affect hormone production, complicating hypothesis testing. Data from natural social groups in natural settings are preferable to those from captivity, however, because captivity’s artificial demographic, spatial, and psychological constraints make it difficult to interpret variation in both hormone levels and behavior. Fortunately the flexible nature of chimpanzee grouping patterns presents a diversity of contexts in which to investigate the role of T and cortisol. For example, chimpanzees alternate their time among traveling alone, with others of their own sex, and in mixed parties. Predictable variation of this kind can be useful in teasing apart potential effects of demography and social context on the production of hormones and behavior. This chapter briefly reviews some basic endocrine physiology, and introduces a number of environmental factors that could potentially affect steroid production in the wild. It may be necessary to control for these variables in later chapters, as I turn specifically to the relationship between aggression, dominance, T and cortisol. Steroid Hormones Steroid hormones are small, lipid-soluble molecules that are enzymatically synthesized from cholesterol. Steroids pass easily through cell membranes; thus, in order to circulate through the blood they are bound to specific plasma carrier proteins. In humans most cortisol (75-80%) is bound to transcortin, and most T (90%) to sex hormone binding globulin (SBHG). A proportion of both steroids binds weakly to albumin and other proteins, leaving 5-10% free cortisol and 1-2% free T. Only the free and weakly-bound forms of these molecules are biologically active (reviewed in Milgrom 1990, Genuth 1993). The classical mechanism of steroid action involves the regulation of gene transcription and protein synthesis. First, steroid hormones pass through the cell membrane and bind to specific receptors in the cell nucleus. Then these activated receptors bind to DNA and activate mRNA transcription and protein synthesis (Milgrom 1990, Genuth 1993). Since this process can take many hours (15-30), it was once believed that short-term fluctuations in hormone production (on the order of seconds to minutes) were biologically unimportant (Moore and Evans 1999). More recently, however, experimental evidence has shown that steroids can bind to cell membrane receptors, thereby regulating intracellular responses on a much shorter time scale (seconds to minutes) (Moore and Evans 1999, Falkenstein et al. 2000). These effects are independent of gene transcription and protein synthesis; thus, steroid specificity can differ between genomic and non-genomic responses (Moore and Evans 1999). The existence of steroid receptors in cell membranes suggests the possibility for extremely rapid effects of steroid release on behavior, but this field of research is still in its infancy. Testosterone Background T is a 19-carbon steroid synthesized by Ley dig cells in the testes under stimulation of luteinizing hormone (LH) from the pituitary. A small amount of T (and its androgenic precursors) is also produced in the zona reticularis of the adrenal glands. In males the latter is of little biological significance, because it represents such a small fraction of overall T production (reviewed in Genuth 1993). The effects of T on fetal development, pubertal growth, and adult behavior in vertebrate males are extensive and profound. Ellison (2001) suggests that T is best understood as a general modulator of male mating effort, because most of its physical and psychological effects on males relate to reproduction. For example, T and its metabolites drive male genital differentiation, reproductive maturation, and the development of secondary sexual characteristics (reviewed in Breedlove 1993, Genuth 1993, Nelson 1995). T regulates sperm production in the Sertoli cells of the testes (Genuth 1993). It promotes anabolism of sexually dimorphic muscle tissue that is critical in male-male reproductive competition. T also helps to modulate male mating interest (libido), and appears to facilitate male aggression, particularly in reproductive contexts (reviewed in Baum 1993, Carter 1993, Wallen 2000, Wingfield et al 1987, Wingfield et al 1990). The present study focuses on these behavioral aspects of T. T and aggression The central hypothesis for the relationship between T and aggression is that elevated T causes increased aggressivity. Three major lines of evidence support this view. First, in numerous studies of birds (Balthazart 1983), and mammals (e.g. red deer: Lincoln et al 1972), castration consistently reduces male aggressivity, and T treatment restores it. Second, in some species, individual T levels correlate with overall rates of aggression (e.g. rhesus monkeys: Higley et al 1996; Asian elephants: Lincoln and Ratnasooriya 1996). Finally, experiments show that T implants in normal individuals can cause increased aggression (e.g. song sparrows: Wingfield 1984), sometimes in a dose-dependent manner (e.g. red deer: Lincoln et al 1972). On the other hand, numerous studies have failed to find a strong correlation between T and aggression in normal individuals, and low concentrations of T are generally sufficient to facilitate aggression in castrated animals (Monaghan and Glickman 1992). This suggests a threshold effect for T on aggression rather than a dose-dependent response. In general, the relationship between T and aggression in mammals appears to be inconsistent because of social influences (e.g. ungulates: Bouissou 1983; rhesus monkeys: Rose et al 1971; Gordon et a l 1976, 1979; Michael and Zumpe 1978; Zumpe and Michael 1996; baboons: Sapolsky 1982, 1983; Alberts et al 1992). The relationship is further complicated by the fact that it can be reversed; that is, aggression can affect T secretion (e.g. Alberts et al 1992, Monaghan and Glickman 1992). The influence of T on aggressive behavior is generally considered to be both organizational and activational (Dixson 1980, Monaghan and Glickman 1992). These roles can be independent, however, because some organizational effects do not seem to require subsequent T exposure (Goy and Resko 1972, Dixson 1980), and some activational effects occur despite a lack of T exposure during development (Monaghan and Glickman 1992). To date it has proven difficult to specify the precise neural mechanisms linking T with aggressive behavior. Experimental studies on a range of species have shown that T enhances directed attention (reviewed in Archer 1988); sustained attention to reproductive competitors presumably increases the likelihood of attack. Additional studies, mostly on rodents, have shown that T activates vasopressin receptors in the hypothalamus (e.g. Delville et al 1996). This is the clearest evidence to date that T regulates a neurotransmitter known to be involved with offensive aggression. T and sexual behavior Most evidence indicates that T is required for normal male sexual behavior (e.g. birds: Balthazart 1983, rhesus monkeys: Trimble and Herbert 1968, humans: Bagatell et al 1994). Considerable evidence from primate studies suggests that T plays an important role in both male and female sexual motivation (Wallen 2000). T is also known to respond to sexual behavior, or even merely to sexual opportunity. For example, male T rises in the presence of strange females (domestic mice (even 6 without copulation): Macrides et al. 1975), in the presence of cycling females (talapoin monkeys: Eberhart et al. 1980), during seasonal breeding activity (rhesus monkeys: Gordon et al. 1976), and— in humans— after exposure to erotic material (Hellhammer et al. 1985). Other factors Two additional factors must be addressed as potential confounds in characterizing the relationship between T and behavior in chimpanzees. First, levels of free T in men decline steadily with age (Vermeulen et al. 1999, Ellison 2001). Second, there is a clear diurnal pattern of both T and cortisol production in men, with the highest levels in the early morning, and a steady decline throughout the day (Van Cauter 1990). A third factor that could potentially affect steroid production in free-ranging chimpanzees is energy availability (e.g. Bribiescas 1996). Numerous studies across a range of vertebrates indicate that low energy availability depresses female steroid production (reviewed in Schneider and Wade 2000, Ellison 2001). The evidence for a comparable effect in males is equivocal (e.g. Ellison and Panter-Brick 1996). However, I will investigate the link between food availability and male T as a possible confound. Cortisol Cortisol is a 21-carbon steroid produced by the adrenal cortex under stimulation of ACTH from the pituitary. Released in response to a wide range of physiological and psychological stressors, cortisol has numerous, far-reaching effects. Sapolsky (1993) suggests that these effects are best understood as a general effort to curtail long-term physiological processes (such as growth and reproduction) in the service of short-term energy mobilization (Chrousos and Gold 1992, Sapolsky 1993). The proximate effects of cortisol arc better understood than those of T; these arc described in chapter eight. Although the cortisol response is clearly adaptive in the short-term, chronic stressors can induce prolonged cortisol elevations that eventually result in pathology, including suppressed immune function (Sapolsky 1993). Because the adverse physiological consequences of chronic cortisol production are so striking, cortisol measurements have been used to quantify the costs and benefits of specific social strategies and interactions (e.g. Creel 2001). In this study I use cortisol measurements to examine the potential costs and benefits of social dominance and subordination in wild chimpanzees. Field Studies of Endocrine Function in the Great Apes Although laboratory studies have examined the relationship between social factors and levels of circulating hormones in a variety of primates, few such data were available from the wild until recently. The most detailed information about endocrine function in a free-ranging primate comes from Sapolsky’s studies of olive baboons in Kenya (Sapolsky 1983, 1991, 1992). Sapolsky took annual serum samples from darted baboons to examine individual variation in the stress response 8 and its effects on the testicular axis. Stress analysis was possible because the anaesthetic used in the darting procedure was itself a stressor, assumed to be equivalently powerful across different individuals. Sapolsky’s method was highly productive in showing, for example, how the stress response was related to social status and personality. It is not applicable, however, to species unsuitable for regular immobilization. For example, although gorillas, chimpanzees, and orangutans have all been successfully darted in the wild, the risks associated with the procedure are considerably greater than for terrestrial baboons or smaller-bodied arboreal monkeys. Ethical considerations therefore preclude the darting of free-ranging apes for the collection of serum. Fortunately, methods for assaying steroid metabolites in urine and feces have recently been developed, allowing for non-invasive physiological monitoring in the field (Whitten et al. 1998). Most of the field researchers studying endocrine function in great apes have chosen urine as an assay medium for both practical and theoretical reasons. First, apes urinate frequently and copiously. Second, they urinate predictably upon waking, so first morning samples are easy to collect. Third, since steroid metabolites accumulate in the bladder, urinary steroid levels represent an average of circulating levels between urinations. These averages are unaffected by the pulsatile release of hormones into the bloodstream that can confound basal measurements in serum. Finally, urinary assays are capable of detecting substantial acute increases in circulating steroid levels. In captive chimpanzees, for example, cortisol increases following anesthesia are apparent in both urinary and fecal assays (Whitten et al. 1998). Organization Chapter two describes the specific laboratory and field methods employed to quantify hormone levels and behavior in this study. Additional methodological details are provided as necessary in later chapters. Chapter three describes temporal patterns of steroid production and social behavior in chimpanzees. The association between steroid production and time of day is investigated, as well as the association between age and T. These basic patterns are established so that their effects can be controlled for when I examine the relationship between T and behavior in later chapters. Chapter four presents an overview of agonistic behavior in chimpanzees, and suggests ultimate explanations for patterns of both intragroup and intergroup aggression. It sets the stage for understanding the association between such aggression and its proximate hormonal mechanisms. Chapters five and six examine the determinants of short-term variation in male T levels. Chapter five tests the hypothesis that food availability has an effect on male T. Chapter six looks at the effects of social behavior on male T. Specifically, it examines the relationship between T, sexual behavior and aggressive mating competition. The goal is to test the hypothesis that variation in male T levels is associated primarily with male aggression in reproductive contexts, rather than 10 changes in reproductive physiology. The extent to which this idea, formally known as “the challenge hypothesis,” (Wingfield et al. 1990) applies to mammals is not yet clear; this is the first test using data from wild apes. Chapter seven follows up on the effects of social interactions on T by exploring the association between T and dominance in male chimpanzees. Chapter eight assesses the potential costs and benefits of social dominance in chimpanzee males by examining the relationship between dominance rank, aggression, and cortisol levels. It also presents some preliminary data on female cortisol levels in reproductive and non-reproductive contexts. This represents an attempt to document the physiological costs imposed on females by male sexual aggression. 11 2 M e t h o d s Study Site Kibale National Park is a 766-squarc-kilomctcr rcscrv^c located just north of the equator and east of the Ruwenzori Mountains in southwestern Uganda (0°13’ to 0°4U N and 30° 19’ to 30°32’ E). The study area, Kanyawara, sits at an elevation of 1500 meters, and is a mosaic of primary forest, logged forest, grassland, swamp, exotic softwood plantation, and agriculture. Average canopy height is 20-30 meters (Chapman and Wrangham 1993). Kibale soils appear to be more fertile than those found in many tropical forests. As a result, Kibale trees exhibit low levels of secondary compounds, and support one of the highest reported population and biomass densities of primates (Struhsaker 1997). Eleven species of non-human primate are represented in the park, including two colobines (Colobus badius, Colobus guerezd), five cercopithecines {Papio anubis, Cercocebus albigena, Cercopithecus ascanius, Cercopithecus Ihoesti, Cercopithecus mitis) three prosimians {Galago demidovi, Galago inustus, Perodicticuspotto), and one great ape {Pan troglodytes schweinfurthii). From 1977 to 1991, mean annual rainfall at Kanyawara was 1,622 mm (Struhsaker 1997). Rainfall is generally seasonal, with wetter periods from March to May and August to November. There is substantial variation in both monthly and annual rainfall, which precludes established breeding or fruiting seasons (Struhsaker 12 1997). Seasonal variation in fruit availability is marked, but the Kanyawara chimpanzees benefit from an abundance of fallback foods (in the form of terrestrial herbs) that appear to buffer them from extreme nutritional shortfalls (Wrangham et al. 1996). From 1977 to 1983, the average daily maximum and minimum temperatures (annual means of the 12 monthly mean temperatures) at Kanyawara were 23.3°C and 16.2°C (Struhsaker 1997). Struhsaker (1975, 1997) provides further details of the study site. Study Population The Kanyawara chimpanzees were studied briefly in the 1970’s by Ghiglieri (1984), and in the early 1980’s by Isabirye-Basuta (1989). They have been studied continuously since September 1987, when Richard Wrangham established the Kibale Chimpanzee Project. During this study all of the male chimpanzees and most of the females were well habituated to human observers, and could be observed at close range without disturbance. A small number of peripheral females were more timid, but could still be observed in groups. Researchers attempted to maintain a minimum distance of five meters between chimpanzees and themselves at all times. At the beginning of this study the community consisted of 50 chimpanzees, including 11 adult males, 15 parous females, one subadult male, two nulliparous 13 females, eight juveniles, and 13 infants (Table 2.1). Males were considered to have reached adulthood after successfully dominating all of the females in the community (e.g. Goodall 1986). During the course of the study, three infants were bom and five individuals died or disappeared. The Kanyawara community occupies a home range of at least 15 square kilometers (Chapman and Wrangham 1993). The chimpanzees primarily utilize mature forest and lightly-logged compartments. During periods of fruit scarcity, they occasionally move into agricultural areas, raiding crops such as sugarcane and banana stems (personal observation). They have never been provisioned. Other large mammals present in Kibale include buffalo (Syncerus caffer), duiker (Cephalophus callipygus and Cephalophus monticolo), bushbuck (Tragelaphus scriptus), bushpig (Potamochoerus porcus), giant forest hog {Hylochoerus meinertzhageni) and elephant (Loxodonta africand). In the past this list included lions {Panthera led) and leopards {Panthera pardus), but these are now extremely rare and only occasionally in the study area. Thus, natural predators do not appear to be a concern for the Kanyawara chimpanzees. Nor is deliberate human hunting of chimpanzees a serious problem in Kibale, because the local Batoro people do not eat primates. Poachers do set traps for antelopes and pigs, however, which sometimes snare chimpanzees (Muller 2000). Approximately one-quarter of the Kanyawara chimpanzees display conspicuous snare wounds, ranging from missing fingers or 14 Name Code Born Name Code Offspring Code Born Adult males Central females and dependents ^Badfoot BF 1966 Aunt Rose AR Mandela AM 1998 Big Brown BB 1966 Sanyu AS 1990 Imoso MS 1975 ^i*Finger FG ^i*PoIlen FP 1996 Johnny AJ 1974 ^i*Nectar FN 1989 Light Brown LB 1968 Kabarole KL Kaboyo KB 1998 Makoku LK 1982 Lia AL ^FKahila AK 1997 Slim SL 1971 Lope LP Ipasa LS 1997 Stocky SY 1964 Rosa LR 1989 Stout ST 1955 Outamba OU Tenkere OT 1998 Tofu TU 1960 Kilimi OK 1994 Yogi YB 1973 Tongo TG Lanjo TJ 1995 Subadult Males Edward ED 1988 Kakama KK 1985 Twig PG 1988 Southern females and dependents Bubbles BL Barbara BR 1989 Beetle BE 1995 Budongo BU 1998 Gombe GO Goodall GA 1991 Umbrella UM Uganda UG 1996 Nulliparous females Nile NL 1982 Nyenka NE 1983 Northern females and dependents Ekisigi EK Esiom ES 1994 Josta JO Kaana JK 1992 Mususu MU Max MX 1997 Pepsi PE Cola PC 1992 Stump PU Bud PB 1995 Table 2.1. Members of the Kanyawara community in 1998. Males are in bold. Individuals that died or disappeared during the study period are indicated with a Birth dates on the far right are for dependents. 15 hands to crippled feet. One individual (FN) died during the study period as a result of snare wounds. Behavioral Data Collection The data presented in this study were collected from November 1997 through December 1998. With the help of long-term field assistants, chimpanzees were followed, whenever possible, from the time that they woke in the morning until the time that they constructed their night nests. If observers lost track of a chimpanzee party, a new one was located by either listening for long-distance vocalizations (e.g. pant-hoots and screams), or by waiting near a firuiting tree. All-male and bisexual parties were followed preferentially, in order to facilitate data collection on male aggression. Field-assistant data The Kibale Chimpanzee Project (KCP) employs a staff of Ugandan field assistants who collect long-term data on the Kanyawara community. Assistants normally work in pairs; one fills in a party composition checksheet, while the other conducts ten-minute focal follows. Party composition is recorded every fifteen minutes. Because chimpanzee parties can be spread over large areas or through dense vegetation, determining which individuals to include in a party can be problematic (for discussion see Chapman et 16 al 1994). At Kanyawara a party is formally defined as all individuals traveling, resting, grooming, or socializing within 50 meters of each other. In practice such distances are sometimes difficult to estimate; thus, chimpanzees may be spread over a slightly larger area, and their inclusion in a party depends on the relative cohesion of the group (e.g. whether all individuals are moving together in the same direction). Observers use a simple scale to record the degree of tumescence of the sexual swelling for each adult female in a party (e.g. Wallis 1992). Females with sexual skins that are completely flat receive scores of (1) no swelling. Females with sexual skins that are partly inflated, but wrinkled and droopy, receive scores of (2) partial swelling. Females with sexual skins that are fully expanded (i.e. tense and shiny with no drooping) receive scores of (3) maximally tumescent. Party location is recorded on a map of the Kanyawara trail system at fifteen- minute intervals. If chimpanzees are feeding, both the species and portion of the plant being consumed are noted (i.e. bark, flower, mature leaf, young leaf, unripe fruit, ripe fruit, seed, pith, or wood). Ad libitum observations of conspicuous behaviors (such as charging displays, attacks, or pant-grunts) are also recorded on the party composition sheet. Separate checksheets are used to record predation and tool use. When observation conditions allow, ten-minute focal follows are conducted on all individuals save dependent offspring. Targets are selected randomly from visible 17 chimpanzees until all members of a party have been sampled, after which individuals may be resampled. The following information is recorded at two-minute intervals for the focal individual: activity, height (on the ground or in a tree), nearest neighbor, and individuals within five meters. Social interactions occurring during the ten minutes that involve the focal ID are also noted. Aggression data I employed forty-minute group locals to generate rates of aggression for individual chimpanzees. Such all-occurrence sampling (Altmann 1974) is the most efficient way to accumulate data on behaviors like aggression that occur regularly but at relatively low rates. One disadvantage of all-occurrence sampling is that it requires that no incident of a behavior be overlooked. The method was practicable in this study because the boisterous nature of chimpanzee agonism renders it highly conspicuous to observers. If a party could not be observed for the full forty minutes, then the focal follow was abandoned. If a party fissioned during the focal period, only data from individuals who were observed for the full 40 minutes were used in rate calculations. In practice such fissioning was rare, occurring in fewer than 8% of the forty-minute locals. Table 2.2 shows the distribution of focal observations by month across the adult males. I personally collected the data taken from January through November. Field 18 I H P - 4 P Q S on P Q P Q g GT) ^cncncn<N<N'^ ^ ^ c N ^ V - ) r - m oo C N C N m oo Tf \0 1 — 1 1 — ( c^ \o r - r - oo C N m ^ V - ) < N t J - 1 — 1 I o 1 — 1 1 — ( oo V O V O O s cn 1 — 1 1 — 1 o 1 — 1 cn so (N 1 — 1 lo m m C N < N CN ^ CN S O o o O O so oo CN ^ CN ^ CN T f m T f CN OO o so fHîîitlIIII iT i SO O O C O S SO ^ 1 —1 1 —1 CN c n 1 —1 CN rT S O O T T O O O O O O 1 — 1 1 — 1 oo 1 — I m rf 1 — I C N S o r - O c n s o O O ^ r r f ^ O C N m CN os O S Tf CN CN m ^ m o m o o c o s m CN ^ ^ m T f T f (N ( U ( U 'B § G - l > =3.1 T O < u 0 1 o o ê c d .S ' C O 73 o (B a 0 < Z ) CO < u 5) 00 c d < u 1 i C O O 0 1 o < u Q a C + H o 0 < u : < u 1 . I | 6 | 5 I ^ if e <2 OO COS COS 3 I I CO C O c d 3 < u la C O ^ . ê ^ 2l| H ^ S 19 assistants from the Kibale Chimpanzee Project collected most of the observations in December (these accounted for less than 4% of the total observations). Behavioral categories followed those of Bygott (1979) and Goodall (1986); these are summarized in Nishida et al. (1999). Charging displays involved exaggerated locomotion and piloerection, and were classified as either “vocal” or “nonvocal,” depending on the presence of the pant-hoot call. Non-vocal displays were more likely than vocal displays to be directed toward specific individuals (see chapter three). Chases were recorded when an individual pursued a fleeing conspecific, who was generally screaming. Attacks were recorded for all contact aggression. This included hits, kicks, or slaps delivered in passing (level one), more extended episodes of pounding, dragging, and biting (level two), and incidents lasting more than 30 seconds or leading to serious injury (level three). Responses to aggression were recorded in the following categories: no response, avoidance (moving away casually or fleeing rapidly), submission (presenting, crouching, mounting, pant-grunting, screaming, squeaking, or whimpering), tantrum (loud screaming resulting in glottal cramps, self-hugging, and ground or tree slapping), retaliation (directing aggression at the original aggressor), and redirected aggression (directing aggression at an innocent bystander). Whenever both members of a dyad could be observed for ten minutes following an aggressive interaction, affiliative 20 contact between the pair during that period was recorded as a reconciliation (as in de Waal 1993). Behavioral contexts of aggression were noted in the following categories: reunion (incidents occurring within five minutes of a reunion), social excitement (incidents occurring upon hearing distant chimpanzee calls, or arrival at a firuiting tree), sexual competition, meat competition, plant-food competition, protection (of offspring), and no obvious context. For further description of these contexts see Bygott (1974) or Goodall (1986). When group focals were not being conducted, ad libitum observations of aggression and submission were recorded. These were pooled with both focal aggression data, and KCP data (from the party composition and ten-minute focal sheets) to rank the adult males in a linear dominance hierarchy. The method for determining dominance ranks is described in detail in chapter seven. Ad libitum data were used only to assign male dominance ranks; they were not included in estimates of individual aggression rates. Food Availability and Energy Balance Fruit availability was estimated indirectly by calculating the total percentage of feeding observations for each month in which chimpanzees consumed fruit. This measure has previously been shown to correlate with direct estimates of fruit 21 abundance from phenology transects (Wrangham et al. 1996). Within the category “fruit,” drupes were distinguished from figs because drupes typically contain more sugar and are preferred by chimpanzees (Wrangham et al. 1996). For example, Wrangham et al. (1996) reported a significant negative correlation between drupe eating and fig eating by the Kanyawara chimpanzees, despite the fact that figs were relatively continuously available. As an independent measure of energy balance, urine samples from individual chimpanzees were screened for the presence of ketone bodies with Chemstrip urine analysis strips (Boehringer Mannheim, manufacturer). These strips are impregnated with chemical reagents that change color upon contact with urine, and are interpretable by comparison with color charts provided by the manufacturer. Ketone bodies appear in the urine When endogenous fat stores are rapidly broken down during periods of fasting. In a study of wild orangutans at Gunung Palung National Park in Indonesian Borneo, Knott (1998) found that the percentage of urine samples containing ketones increased dramatically during one fruit-poor season. Urine Collection and Preservation Chimpanzees regularly urinate upon waking in the morning, immediately before leaving their night nests. These first-moming urine samples were regularly collected by observers. When a chimpanzee urinated from a tree, an observer on the ground 22 trapped the urine in a disposable plastic bag attached to a two-meter pole. Urine samples were also collected opportunistically throughout the day. Whenever possible, samples were captured on plastic; if a bag could not be placed in time, urine was pipetted from leaves in the ground layer of vegetation. After collection, observers recorded the identity of the chimpanzee, the date, and the time of urination. One to twenty-four hours after collection, urine samples were processed and stored in a propane-powered freezer (mean: 6.5 hours). The freezer consistently maintained a temperature between -18° and -23° Celsius. Frozen samples were transported on both ice and dry ice to the Reproductive Ecology Laboratory at Harvard University, where I performed all hormone analyses. Sample Contamination The risk of sample cross-contamination was mitigated by collecting urine from vegetation only when it was clear that more than one individual had not urinated in the same area. This situation was easy to avoid with first-moming samples, because chimpanzees nest alone, and build fresh nests in new locations nightly. When collecting samples during the day, caution had to be exercised, particularly when more than one individual was feeding in the same tree. Fouled samples, however, were a rare nuisance. 23 Care was taken to avoid collecting urine contaminated with feces. However, upon close inspection in the field laboratory, small amounts of fecal matter were visible in some samples. Heavily contaminated samples were discarded. When trivial amounts of fecal material were present, this was noted and the offending particles were removed from the sample prior to freezing. Because some urine samples have passed through leaves in the canopy before they reach the ground, and others are collected from the ground layer of vegetation, it was theoretically possible that contaminants from leaf surfaces could affect the results of the hormone assays. In July 1997,1 performed two tests at Kanyawara to address this issue. First, branches were collected from 22 of the tree species most frequently utilized by chimpanzees. Human volunteers were asked to make clean collections of urine. These were separated into an untreated control, and an experimental sample, which was poured over specified branches. After five to ten minutes, urine was collected from the leaves with a disposable pipet, following the procedures used for chimpanzee urine collection. The samples were frozen and returned to the Reproductive Ecology Laboratory, where I assayed them for T, cortisol, and creatinine, following the procedures detailed below. Paired comparisons of T, cortisol, and creatinine measurements from the control and experimental samples revealed no significant differences. T measurements between the two sets of samples 24 were highly correlated (r^=.97, p<.0001); the same was true for both cortisol (r^=.89, p<.0001) and creatinine (r^=.93, p<.0001). In a second test, 19 branches were collected from 18 of the most common species in the ground layer of vegetation. Approximately 250 ml of water was poured over each branch to simulate urination. Five to ten minutes later, water was collected from the leaves, following the procedures used for chimpanzee urine collection. The samples were frozen and returned to the Reproductive Ecology Laboratory, where I assayed them for T and creatinine. In all 19 samples, T levels were indistinguishable from zero, and creatinine was undetectable. Hormone Analysis Steroid levels were quantified by radioimmunoassay according to published protocols (Ellison 1988, Ellison et al. 1989) adapted for use with primate urine. Because most T appears in the urine as the metabolite testosterone glucuronide, samples were deconjugated before they were assayed. This was accomplished by hydrolysis; 100 pi of urine was combined in a test tube with 20 pi of the enzyme B- glucuronidase-arylsulfatase (Boehringer Mannheim, manufacturer) and 300 pi of pH 5 buffer (Fisher Scientific, manufacturer). This mixture was incubated overnight in a 37° C water bath. 25 The T assay is based on a four-position tritium competitor (Amersham-Searle, manufacturer) and an antiserum raised against testosterone-11-BSA provided by Gordon Niswender of Colorado State University (#250). This antiserum has reported cross-reactivities of 46% with DHT and 17% with androstenedione and dihydroepiandrosterone (Campbell 1990). Hydrolyzed urine samples were extracted twice in diethyl ether prior to assay, with recoveries individually monitored by the addition of trace amounts of tritiated T. Recoveries averaged 90%. Separation of bound and free steroid after 24 hour incubation at 4° C was accomplished by adsorption of free steroid to dextran-coated charcoal. Bound competitor was measured in a RackBeta liquid scintillation counter (LKB/Wallac, manufacturer). For a small number of samples (n=22), recoveries were less than 50%. These samples were hydrolyzed and extracted a second time, but recoveries were still extremely low. There was nothing obvious about these samples to suggest why recoveries were a problem, which highlights the importance of monitoring individual recoveries. These samples were excluded from all analyses. Quality control was maintained by monitoring values of urine pools at three different levels. Assay sensitivity, the least amount distinguishable from 0 with 95% confidence, averaged 11,000 pmol/L. Intra-assay variability (CV) at the 50% binding point of the standard curve was 6.6%. Inter-assay variability averaged 6.6%, 6.2%, and 6.4% for high, medium and low pools (n=17). Linearity of response was verified 26 by assaying serial dilutions of both T standard (predicted vs. observed values: r^=l, p<.0001) and chimpanzee urine (predicted vs. observed values: r^=.99, p<.0001). The cortisol assay is based on a four-position tritiated competitor (Amersham- Searle, manufacturer) and an antiserum raised against Cortisol-3-0- Carboxymethy lether-B SA (ICN Biomedicals #07-121016). This anti serum has reported cross-reactivities of 11.4% with 21 -desoxycorticosterone, 8.9% with 11- desoxycortisol, and 1.6% with corticosterone; cross-reactions with other naturally occurring steroids are non-significant. Details of the cortisol assay are similar to those of the T assay, except that the urine samples were not hydrolyzed, nor was the steroid extracted with ether. Cortisol was assayed directly from unpurified urine diluted 1:10 with distilled water. Appendix one contains complete laboratory protocols for both the T and the cortisol assays. For the female cortisol assay, sensitivity averaged 9600 pmol/L. Intra-assay variability (CV) at the 50% binding point of the standard curve was 4.9%. Inter assay variability averaged 6.5%, 3.7% and 12.8% for high, medium and low pools (n=12). For the male cortisol assay, sensitivity averaged 5250 pmol/L. Intra-assay variability (CV) at the 50% binding point of the standard curve was 6%. Inter-assay variability averaged 7.2%, 6.7% and 15.6% for high, medium and low pools (n=16). 27 To correct for variation in urine concentration, steroid levels were indexed to creatinine (Erb 1970, Cook and Beastall 1987). Creatinine is produced when creatine phosphate, a high-energy compound in skeletal muscle, is nonenzymatically dephosphorylated. This is assumed to occur at a relatively constant rate. Creatinine levels were quantified colorimetrically using the Jaffee reaction (Taussky 1954). Appendix one contains the complete laboratory protocol for the creatinine assay. Samples with creatinine measurements below .05 mg/ml (n=6; all from different males) were excluded from the analyses, because these measurements are less accurate and tend to inflate the final T values. Statistical Procedures Comparisons between independent groups were made with the Rust-Fligner rank based test (Rust and Fligner 1984). This nonparametric test is equivalent to the Mann-Whitney-Wilcoxon test when applied to two groups, and the Kruskal-Wallis test when applied to more than two groups. However, the Rust-Fligner procedure eliminates the assumption of equal variances that can result in poor power and wide confidence intervals in these other methods (reviewed in Wilcox 1996). Comparisons between dependent groups (i.e. paired comparisons) were made with the Agresti-Pendergrast rank-based procedure (Agresti and Pendergrast 1986). This nonparametric test is similar to the standard Wilcoxon Signed Rank test, but 28 appears to enjoy better power properties over a wide range of situations, particularly when sampling from heavy-tailed distributions (Kepner and Robinson 1988, Wilcox 1996). Comparisons between two independent binomials were made with the Storer- Kim procedure (Storer and Kim 1990). This procedure has better power than Fisher’s exact test, which is frequently used for such comparisons (Storer and Kim 1990, Agresti 1992). Unless noted, all correlations presented in this study employ Pearson’s product moment correlation coefficient (r). Regression lines were fitted with the standard least squares method. All statistical tests are two-tailed, and p values below .0001 are reported as “p<.0001.” Further details of the methodology are presented in the relevant chapters, below. 29 3 P a t t e r n s o f A g g r e s s io n Introduction Although wild chimpanzees may spend hours resting and grooming peacefully in mixed social groups, and affiliative interactions among them may be more common than agonistic ones, intraspecific aggression is nonetheless a frequent occurrence in chimpanzee society. Both males and females exhibit an array of aggressive behaviors (from mild threats to lethal attacks) in a variety of contexts (from infant protection to sexual competition) against a range of com petitors (from extracommunity males to newly immigrated females). Aggression, or merely the threat of aggression, can have a profound impact on individual patterns of ranging and association. For example, female immigrants at Gombe tend to settle in peripheral areas of the range— away from the dominant female— which may be a strategy to decrease the risk of infanticide (Williams 1999). The most dramatic examples of chimpanzee aggression come from observations of intercommunity encounters (e.g. Goodall et al. 1979). Male chimpanzees are philopatric, and aggressively defend their community range against incursions from neighboring males (Nishida 1979, Goodall 1986, Watts and Mitani 2001). In the course of such defense they sometimes cooperate to inflict lethal wounds on vulnerable strangers (reviewed in Wrangham 1999). Lethal coalitionary aggression is rare among mammals, having been documented as a major source of adult mortality only in wolves (Mech et al. 1998) and humans (van der Dennen 1995). In male chimpanzees it may be part of a larger strategy to reduce the coalitionary 30 strength of neighboring groups, and to expand territorially at their expense (Wrangham 1999). Within communities, male chimpanzees compete aggressively both for status (within a dominance hierarchy) and for access to estrous females (Bygott 1979, Goodall 1986). Coalitions can play an important role in both of these contexts, as males may cooperate either to challenge rivals (Nishida 1983, Nishida and Hosaka 1996) or to maintain exclusive access to an estrous female (Watts 1998). Intracommunity aggression is normally less brutal than that between communities, perhaps in part because relatedness among males is higher within a community (Morin et al. 1994). However, dominance struggles are sometimes marked by intense dyadic aggression and potentially lethal wounding (e.g. Goodall 1992, Nishida et al. 1995, Fawcett and Muhumuza 2000). Male chimpanzees appear to incur large costs as a result of their competition for dominance. These include not only the risk of severe injury in escalated fighting, but the energetic demands of agonistic display. The presumption that such costs must be offset by considerable reproductive benefits has been supported previously by behavioral data indicating that high rank is associated with increased mating success (e.g. Nishida 1983). More recently, genetic tests of paternity have allowed for direct measures of reproductive success, and corroboration of the behavioral evidence (Constable et al. 2001 ; for bonobos see Gerloff et al. 1999). The potential influence of dominance rank on female reproductive success is not as well understood. Dominance relationships among female chimpanzees are less 31, conspicuous than those of males, such that observers often find it difficult to rank them (Bygott 1974, 1979). On the one hand this is not surprising, as the limiting factor on female ape reproduction is normally considered to be food, and competition for food among female chimpanzees generally takes the form of scramble, rather than contest, competition (Wrangham 1980, Wrangham 2000). On the other hand, the lack of overt competition among female chimpanzees is puzzling, because recent evidence suggests that female dominance rank does influence factors such as infant survivorship and interbirth interval, probably through access to food (Pusey et al. 1997). This chapter describes patterns of intergroup and intragroup aggression among chimpanzees in Kibale National Park, Uganda. There are four main objectives. First, sex differences in rates, contexts, and targets of aggression are described, and accounted for in terms of ultimate reproductive strategies. Second, the relationship between aggression and dominance is explored, and the possible costs and benefits of high rank in this species are considered. Third, new incidents of intercommunity aggression, including a lethal coalitionary attack, are reported, and evaluated in light of the “imbalance-of-power” hypothesis (Manson and Wrangham 1991). Finally, these data are compared with observations from other long-term study sites in Tanzania (Bygott 1974, Goodall 1986, Nishida 1989, Nishida and Hosaka 1996) and Ivory Coast (Boesch and Boesch-Achermann 2000) in an attempt to identify the broader patterns underlying behavioral variation. An understanding of the larger 32 patterns of chimpanzee aggression will become important when considering the hormonal correlates of these behaviors in later chapters. Results Sex differences in aggression Male chimpanzees at Kanyawara exhibited much more frequent and intense aggression than females did. Figure 3.1 presents mean rates of aggression (charging displays, chases and attacks) for both sexes. Though variability exists within each sex, adult males were aggressive approximately 14 times more often than adult females (males: mean=.28 acts per observation hour, n=ll; females: mean=.02 acts per observation hour, n=10). This figure probably overestimates female aggression, because it excludes observations made when focal individuals were traveling alone (or solely with dependent offspring), and female chimpanzees are more likely to be solitary than males (Nishida 1979, Goodall 1986, Wrangham 2000). Table 3.1 shows attack rates for males and females at both Kanyawara and Gombe. The same pattern of male-biased aggression is evident. Aggression by female chimpanzees tended to take place in different contexts from aggression by males. Table 3.2 illustrates this sex-difference with data from three long-term study sites; it includes only episodes in which a context could be clearly determined. 52% of male aggression at Kanyawara and 14% at Gombe took place in indeterminate contexts. Most of these episodes appeared to be related to 33 I u I I Î < 0.4 0.3 - 0.2 - 0.1 - 0 M ales Fem ales Figure 3.1. Mean rates of aggression for male and female chimpanzees in Kanyawara, 1998. Aggression includes charging displays, chases, and attacks. Data are from 11 males and 10 females with at least 25 observation hours (male median=144.7, female median=69.7). Follows of lone targets have been excluded. Rates of male aggression are significantly higher than those of females (Rust-Fligner test, Q=42.8, p<.0001). Error bars represent the standard error of the mean. 34 § g I I c d o tS c d o tS g 00 I I gj I Î Q C O oo t - H od t - H C O V O V O 3 m V O m \o rt fN v — I V O ( L > t £ c d c d C d s S s s oo O O V O o C \ C \ a\ a\ a\ t - H t - H t - H T - H SÏÏ V O V O C\ o V O oo O v 55 3 & ^ oo ( U ( U ( U 7d 7d a a a < L > < L > < L > Ph Ph Ph O O O O V O a\ a\ a\ a\ r — i I g > > 1 1 1 1 > > 1 1 â a Ü o 1 1 6 V ] -S ' 3 1 0 Û c 0 s c / 3 Î 0 3 § C/3 s 1 i I £ ro < D I a 2 1 ' H - l o C d o Id O 03 1 c / 3 o o o 43 O C c / 3 O * H c d > o 43 < L > G c / 3 t O O O a O > O » - H T^ § o V O C N o a o c d a a g ? g s -3 1 ’c o < L > < L > 03 03 3 5 O 73 C c a V O c d 03 o\ c d t - H 03 § O O a a\ < L > 3^ < L > C o 'o p H T f - r - c\ o 0 Û ? = - . C Q .g I o c\ £ c d " S 03 ( U 'a o V £ > oo G\ 1 i I 4 3 Î I I <2 I (3 35 00 r - 1 0 0 1 g I I u I c / 3 a £ (D 7 d £ V O CN O O r- C O 1 - H oo oo CO o oo T f \ 0 VO CO CO 1 —I I o " a 3 X < D § I -- C O P l h VO I o T 3 O O I o % (D OO C O r - C O V O C O o o CN OO I o CN § • a I 3 O I 13 CO pL, T ± T ± VO CN CN C N § g C/3 - § 3 H £ £ 5 I X ^ B ^ rl o o I s f c d s u I (U I I s c/3 I U CN cd a g . ■ § < D (D a ^ c o o 1 < D (U 'Hr I cd 4 3 £ c / 3 I I I I s 03 < D g £ ( D 43 c/3 I - § % ( D a _ ( D I % C o 'g <D c/3 O (D 7d e e CN V O Ï I i C/3 c o 3 & (D c/3 O (D .(D O Ü a I I I c/3 g i j \ o od CN I c/3 g g C/) 03 13 C O > o 03 o B ' o o ( D c/3 £ < D * I D 1 -3 1Î (D .S ' ' ë (D 1 o \ o oo o oo (3 \ < D (D 43 g C/3 g 36 male dominance rivalry (Goodall 1986, personal observation). When contexts of male aggression could be determined, they were similar at Gombe and Kanyawara. A large proportion of male agonism took place either within five minutes of a reunion (34% at Kanyawara, 38% at Gombe) or during sexual competition (36% at Kanyawara, 17% at Gombe). Much aggression at Gombe also took place during competition for meat (17%); this figure was lower at Kanyawara (4%), where little hunting took place during 1998. Contexts of female aggression were somewhat less ambiguous than those of male aggression; 6.3% of observations at Gombe, 11% at Kanyawara, and 21% at Mahale could not be assigned to a distinct context. The most frequent contexts of female aggression were competition for plant food (80% at Kanyawara, 38% at Gombe, 35% at Mahale) and protection (12% at Kanyawara, 38% at Gombe, 35% at Mahale). If interspecific interactions are considered at Kanyawara, an even higher percentage of female aggression there took place during feeding competition; female chimpanzees occasionally chased blue monkeys or mangabeys from preferred feeding sites in fruiting trees. These episodes were excluded from the present data set, but the aggression appeared identical to that directed toward conspecific feeding competitors. Males and females tended to direct aggression at different targets. Table 3.3 shows the distribution of victims at two long-term study sites. At Kanyawara, most aggression by adult males (49.2%) was directed at other adult males; slightly less (24.6%) was directed toward parous females. Few parous females were cycling 37 during this study, but preliminary data from one popular female, Lia, suggest that parous females receive more aggression when they exhibit sexual swellings than at other times. During 33 hours of focal observation with no sexual swelling. Lia received aggression from adult males every 8.3 hours; during 46 hours of observation with a partial or maximal swelling. Lia was the victim of male aggression every 3.8 hours. This increased rate of received aggression was partially due to the fact that more adult males were present when Lia was swollen. However, males also became more aggressive when traveling in parties with estrous females (see below). Adult females were not observed to exhibit any aggression towards other adult females; more than 70% of their aggression was directed at juvenile and subadult females. At Gombe, adult males directed slightly more of their aggression at adult females (48%) than adult males (34%) (Goodall 1986). Adult females at Gombe also appeared to be involved in aggressive interactions with other adult females more frequently than at Kanyawara. 33% of observed aggression by adult females at Gombe was directed at other adult females (Goodall 1986). Dominance relationships Most pant-grunts by adult males (64%) were directed toward the alpha male (n=89). The two lowest-ranking males produced 39% of male pant-grunts. Consistent with data from Mahale (Nishida and Hosaka 1996), male dominance rank was positively and significantly associated with the total number of pant-grunts 38 0 0 OS 1 Ü 0 (/: 0 0 1 < 00 Os Os 1 I o > cn ' tJ - ^ cn cn cN m T j- m vn vn m T j - oo m T j- Os Tj- o vn w - i CN m (N VO O s Tf T j- (N (U I (U 1 Î m en o o oo Tt o o T j- T j- rn o O M m M e N 1 — 1 1 — 1 sg = lO T j- rn VO cn Os O O & o ■§ C T j -§■ O lO C N vn <N m o <o oo V O 1 — 1 1 — 1 oo I I CN CN < D ■3 3 00 00 (U a a (U (U . e n en - § H ^ 00 ^ (U < ^ I ^ S 2 c s (U 1 3 (U (U O ) 3 O B a 1 3 » I C T j O 3 I 13 ' > ^ 'B .a ^ I o < D C T j (U OD C T j S GO a _ G O C A C T j 13 I 3 O o I I 39 received (n=l 1, r=.66, p=.027) and the total number of agonistic initiations (n=ll, r=.69, p=.018) (both uncorrected for observation hours). During the study period, the alpha-male, Imoso, was just beginning his tenure. For three years starting in m id-1994, Big Brown had been the alpha-male at Kanyawara and his ally, Tofu, had held the beta position (Kibale Chimpanzee Project, long-term data). In mid-1997 Imoso, together with his ally Johnny, began to challenge these top-ranking males. By the end of that year they were clearly established as the new alpha and beta males. Tofu fell to the third place in the hierarchy, and Big Brown to the fourth. Although the observed dominance hierarchy was of relatively recent origin, it appeared to be quite stable during the study period; no reversals were observed in male-male pant-grunt interactions. And out of 107 decided agonistic bouts, only three reversals were recorded. It was not possible to rank the 15 adult females in a linear hierarchy, because dominance interactions between them were extremely rare. In 680 hours of focal observation, not one aggressive interaction was recorded between parous females. And although parous females frequently pant-grunted to adult males, they were rarely observed pant-grunting to other parous females. Patterns of male aggression The most frequent form of male aggression was the charging display. Display rates differed among males, but the alpha male, Imoso, was clearly the most aggressive member of the community. His display rate (.69 per observation hour) 40 was 4.5 times the male average (.15 per observation hour), and more than twice that of the next highest male (Tofu: third-ranking male, .29 per observation hour). Dominance rank and frequency of display were significantly correlated across all 11 adult males (r=.75, p=.008. Figure 3.2). Consistent with reports from Gombe (Bygott 1974, Goodall 1986), vocal displays were rarely directed toward specific individuals (only 6.8% of 148 vocal displays). Instead, they appeared to be directed either toward the group in general, or toward distant chimpanzees. Non-vocal displays, on the other hand, were frequently directed toward specific targets within a party (53.2% of 186 displays). This difference is statistically significant (Storer and Kim procedure, p<.0001). When only high-level agonistic acts (chases and attacks) are considered, the alpha-male was still the most frequent aggressor in the community. Furthermore, rank and the frequency of high-level aggression were positively and significantly correlated across the adult males (n=ll, r=.71, p=.014). Table 3.1 shows mean attack rates for males at both Kanyawara and Gombe. Ideally these would be presented as means of mean individual rates; however, to facilitate intersite comparisons I have used the same analyses as Goodall (1986). Male attack rates from Kanyawara in 1998 are comparable to those reported from Gombe in 1970; they are higher than those reported from Gombe in 1976 and 1978. The frequency of aggression in a given party at Kanyawara was positively and significantly correlated with the number of adult males in that party (r=.95, p<.0001. Figure 3.3). However, the rate of aggression per male did not increase with party 41 (/) & I I 0.7 0.6 Alpha male 0.5 - 0.4 - 0.3 ■ 0.2 - 0.1 ■ 0.5 1 1 -0.5 0 Dom inance Rank Figure 3.2. Dominance rank and display rate in Kanyawara males (1998). The alpha male displayed much more frequently than any other chimpanzee in the community. Across 11 adult males, there was a significant correlation between rank and display rate (r=.75, p=.008). Display rates are from 1429 hours of focal observation (male median= 144.7 hours). 42 I *0 i I I II @ 3 k A a, ♦ # Adult Males in Party Figure 3.3. Frequency of aggression in relation to the number of adult males in a party. The number of aggressive acts per hour was highly correlated with the number of adult males present (r=.95, p<.0001). However, males did not become more aggressive with increasing party size (i.e. there was no significant relationship between party size and the number of aggressive acts per male per hour). Parties containing maximally swollen females have been omitted. Data are from Kanyawara, 1998. 43 size. Party composition, on the other hand, had a substantial effect on rates of male aggression (Figure 3.3). When maximally swollen, parous females were present in a party, both the rate and intensity of male aggression increased. The average daily rate of aggression in parties containing eight or nine adult males and no estrous females was 1.68 incidents per hour. These incidents tended to be mostly low-level aggression (70% charging displays only). The average daily rate of aggression in parties containing eight or nine adult males and a parous female in her periovulatory period was significantly higher, at 3.2 incidents per hour (Rust-Fligner test: Q=4.01, p<.05) of mostly high-level aggression (59% chases and attacks). Table 3.4 shows rates of reconciliation at Kanyawara together with those reported from captivity. Adult males at Kanyawara rarely reconciled with each other following a conflict (less than 9% of interactions). Across all age and sex classes, only 10% of conflicts were reconciled within ten minutes. This is considerably lower than the rate of reconciliation reported from captivity (de Waal 1993). Similarly low rates of reconciliation reported from Taï (Wittig, in press) suggest that chimpanzees in the wild are less motivated to reconcile conflicts, perhaps because in the short-term they can always avoid particular conspecifics by fissioning from a party. Intercommunity aggression Intercommunity relations among male chimpanzees are predictably hostile. Border patrols, aggressive territorial defense, and border avoidance by lone 44 I I I 3 1 a I Ï t oo cn oo O n V O C O r ' V O V O o od c-i od vl' On vd T — H r — 4 T — H C O 04 04 oo oo oo Ù 0 D ? 0 \ V O O O cn a a 00 ov ov 1 — H 1 I I T 3 ( D ( U 5 I I £ Ô 6 I (N O V m V O V O o C " g yn V O o ov ^ cs cn V O S ; 0 1i V O V O O O O O O n O n O n I i II f l S ( D I Î O g § & o c 3 t3 I § 1 S I C D I I' g 1 O 1 3 "g c O 1 o T 3 3 0 § 1 g C • s I & ê ( 4 - 1 0 c /3 1 I I I 3 § B 1 § o g % I jc 3 ■ | ÎI < u § c d I c d Ü t3 o « cd f 'o ë § •o i I c I & U < D ■g I O I 3 I g < D 45 individuals have been reported from all the long-term study communities except Bossou, which has no neighbors (reviewed in Wrangham 1999; Table 3.5). Coalitionary intergroup attacks, sometimes lethal, appear to be a regular feature of chimpanzee society. I observed two such attacks at Kanyawara during this study, one of which resulted in a fatality. On August 12, 1998 five adult males from the Kanyawara community traveled to the southern boundary of their territory, where they encountered a nulliparous female and a subadult male from the neighboring community. The Kanyawara males attacked the pair, chasing them into a low tree. Both leapt approximately seven meters to the ground, and the subadult male fled to the south. Three Kanyawara males surrounded the female, striking her repeatedly. She was able to escape, however, and ran quickly to the southwest. Four adult males from the southern community arrived within two minutes of the female’s departure, presumably attracted by the screaming. They displayed and exchanged aggressive vocalizations with the Kanyawara males, who subsequently retreated further than one kilometer into their own territory. In a second incident on August 25, 1998, a group of ten adult males from the Kanyawara community patrolled the northern border of their territory. Observers lost the group at 17:05, but field assistant Christopher Katongole and I located them nearby the following morning with a dead male (aged 20-25) from the neighboring community. He had been killed the previous evening. 46 g c f l c / 3 O H 0 û û 1 P 9 3 X i 2 î I a ( N O n I I I + + + ^ + OO o - m m oo m o - o - O ’ + + + + + + 2 ^ ■ § ■ 'î 11 1 s (N VO + + + m i ex I I a o 2 I ' § a o I "O 'o I pq U U pq O) B 3 (D (D N g ex - 5 V T ) ro (D I H I g i & G o a û û g c / 3 Ç U g ■ g 2 o g 3 - 7 3 u o 3 7 3 I I c / 3 I ■ s a g c d o U O n O n O n i I i S c g c/3 f I 47 The site and the body exhibited signs of a protracted and vicious assault. The dead male lay at the base of a steep slope in a 7x12m patch of trampled vegetation. His arms and legs were extended, suggesting that he had been immobilized by some individuals while others attacked (as occurs in both captive chimpanzees and bonobos: Amy Parish, personal communication). The pattern of injuries was consistent with this scenario, as more than 30 puncture wounds and lacerations, ranging from 1/2 to 12 cm in diameter, were distributed across the male’s face and the ventral surface of his arms, legs, and abdomen (Figure 3.4). The dorsal surface was undamaged (Figure 3.5). Compound fractures in four of the left ribs attested to severe blows delivered by fists or feet. The testicles had been ripped from the body, and were recovered nearby. Five fingernails and one toenail had been torn from the digits with significant portions of flesh attached. The immediate cause of death appeared to be massive trauma to the throat, including a severed trachea. The Kanyawara males were found with the dead stranger on the morning of the 26*. Several of the males were pounding on the corpse when we arrived. As they had also been seen near the attack site on the evening of the 25*, when the male was killed, it seems likely that they were responsible. At approximately 10:45, at least three males from the neighboring community arrived near the attack site, and the two groups exchanged aggressive vocalizations. At 11:45 the Kanyawara males retreated to the south. 48 o Ç Q m o fi ^ £ i> f 3ii/i/ 3wyyx% (U 15 « S > 1 s • “ -s "3 g 1 2 "3 ( D Ü g S B ^ 1 a. (D I ' hJ Q a o o c d B I ^1 § ?5 Q - i 0 1 (D > C d C / 3 (D (D a «I E -S co (D 5 ■ 3 I . S ^ -S II -t ^ a ^ a i T T li 49 ( L > ^nvsvia} s a n f f io M V o > ï3U iby v y sw tÿ e w v 3 h /V iu s r 50 Discussion and Conclusions Female-female competition Consistent with reports from Gombe (Goodall 1986) and Mahale (Nishida 1989), female chimpanzees at Kanyawara were aggressive primarily in the context of feeding competition. Such aggression was rarely severe; it consisted primarily of mild threats or brief charging displays used to supplant feeding competitors. The prolonged attacks characteristic of male dominance rivalry were seen only on three occasions; these were all directed at one recently immigrated female. Parous females were rarely aggressive or submissive towards each other. Nishida (1989) suggests that parous females may direct aggression primarily at newly immigrated females, because such immigrants represent potential resource competitors. After dominance relationships are established between females, they remain stable, perhaps because the costs of escalated aggression (which include potential danger to offspring) outweigh the benefits of increased dominance standing. Because parous females, are already settled in their core areas, they “have no pressing reason to strive for higher rank” (Nishida 1989). Given that core areas may differ in both size and quality, however, and that core areas frequently overlap (e.g. Pusey et al. 1997), the rarity of aggressive interactions between adult females remains puzzling. Preliminary data from Gombe suggest that high dominance rank accords female chimpanzees significant reproductive advantage. For example, high-ranking females there appear to maintain access to 51 higher-quality core areas (Pusey et al. 1997, Williams 1999). These females live longer than low-ranking females, and enjoy shorter inter-birth intervals and higher offspring survival. They also produce daughters that reach sexual maturity earlier than those of low-ranking mothers, presumably because of improved nutrition (Pusey et al. 1997). With so much apparently at stake, one might expect female chimpanzees to show more overt competition over dominance than they appear to. One possible explanation for the low rates of aggressive competition observed among females at Kanyawara is that competition for space there is not as pronounced as it is at Gombe; thus, the benefits of high-rank to Kanyawara females are correspondingly less. This hypothesis has not yet been tested directly, but several observations suggest that female competition at Gombe is particularly intense. First, young females at Kasekela exhibit a relatively low rate of transfer (Pusey and Williams, in press). A female that stays in her natal community presumably bears increased costs associated with inbreeding, but may benefit from associating with a high-ranking mother, for example, by settling in her core area. Second, both infanticide and attempted infanticide by high-ranking females against low-ranking mothers appear to be more common at Gombe than at other sites (Pusey et al. 1997, Clark Arcadi and Wrangham 1999). Third, aggressive interactions between parous females appear to be more common at Gombe than at Kanyawara (Table 3.3). In the present study, no aggressive interactions were observed between adult females at Kanyawara in more than 680 hours of focal observation. At Gombe in 1978, adult females targeted other adult females approximately 33% of the time that they were 52 aggressive (n=159 aggressive acts; Goodall 1986). It is difficult to say whether this reflects a real difference, as Goodall does not provide overall rates of female aggression. Long-term data on female dominance rank and reproductive success at Kanyawara will eventually help to clarify this issue. Male-male competition Male-male competition in chimpanzees takes two general forms, both of which were evident in this study. In the long-term, male chimpanzees cooperate to defend a territory against neighboring males. This cooperation includes border patrols and coalitionary attacks on vulnerable rivals, and is discussed below. In the short-term, males within a community compete aggressively (and through sperm competition: Hasegawa and Hiraiwa-Hasegawa 1990) for access to estrous females (e.g. Watts 1998). Male chimpanzees are sometimes able to avoid short-term mating competition by escorting an estrous female to a peripheral part of the range in an exclusive consortship (Goodall 1986). Genetic data from Gombe indicate that this is sometimes a successful reproductive strategy, particularly for low-ranking males (Constable et al. 2001). O f thirteen infants genotyped in that community, five were sired in the consortship context, all by low and middle-ranking males. Behavioral data from three long-term study sites (Gombe, Mahale and Tai), however, indicate that most chimpanzee conceptions occur in the context of multi male parties (75-94%: Hasegawa and Hiraiwa-Hasegawa 1990, Wallis 1997, Boesch and Boesch-Achermann 2000). Although copulations sometimes occur in such 53 parties with little male aggression or coercion (Tutin 1979), this is generally limited to matings with nulliparous females, or parous females in the mid-follicular phase. Once parous females reach their periovulatory period, they become much more attractive to males, and coercion and aggression increase markedly (see chapter six). In the current study, males at Kanyawara showed both increased frequency and intensity of aggression in the presence of periovulatory females. Among male chimpanzees, success in short-term mating competition is closely predicted by dominance rank. Because low-ranking males may have access to parous females in the mid-follicular phase, and nulliparous females throughout the cycle, overall copulation rates may or may not correlate with dominance rank. However, alpha and high-ranking males consistently exhibit the highest copulation rates with parous females in the periovulatory period (see chapter six). Copulation frequency is not necessarily an accurate proxy for reproductive success. However, preliminary genetic data from Gombe support the idea that high- rank is reproductively advantageous. Of nine infants sired there in the group-mating context, five were sired by the alpha male, two by males who subsequently became alpha, one by a high-ranking male, and one by a middle-ranking male (Constable et al. 2001). The positive correlation between dominance rank and aggression reported here is consistent with observations from other long-term study sites. At Gombe in 1970, Bygott (1974) found significant correlations between male dominance rank and rates of both charging display and attack. At Tai in 1993, the two highest-ranking males 54 performed 80% of all agonistic displays (Boesch and Boesch-Achermann 2000). At Mahale in 1992, dominance rank was strongly correlated with total number of agonistic initiations (Nishida and Hosaka 1996). Finally, long-term data from Kanyawara indicate that two previous alpha males there also had the highest rates of display during their tenures (Wrangham, unpublished data). Among primates generally, however, high levels of aggression are not consistently associated with high dominance rank (Dixson 1980). Olive baboons {Papio anubis) are a well-studied example. Sapolsky (1982) found that among male baboons in Masai Mara, dominance rank in copulatory success was related to success in a number of competitive behaviors. During periods of dominance stability, however, frequency of aggression was not associated with high rank. In fact, the initiators of aggressive interactions lost those interactions 80% of the time when the hierarchy was stable. Only during periods of dominance instability do dominant male baboons consistently show the highest rates of aggression (reviewed in Sapolsky 1993). Sapolsky characterizes a baboon hierarchy as unstable when the overall rate of reversals in approach-avoidance interactions exceeds 10%. Such periods of instability are relatively rare in the wild, but may follow shifts in troop membership (e.g. Alberts et al 1992). The chimpanzee data, however, suggest that high rates of aggression are characteristic of high-ranking chimpanzee males, even in stable hierarchies. The associations between rank and aggression reviewed above all took place during periods of relative dominance stability. The rate of pant-grunt reversals among adult 55 males at Gombe during 1970 was only 0.5% (2 of 442 interactions: Table 5.7, Bygott 1974). At Mahale in 1992 the rate was 1.1% (3 of 268 interactions: Table 9.3, Nishida and Hosaka 1996). At Tai in 1993 the rate was 0.9% (1 of 109 interactions. Table 6.3, Boesch and Boesch-Achermann 2000). In the present study, no reversals were recorded in 89 pant-grunt interactions between adult males. During this study the overall rate of reversals for decided agonistic bouts between adult males was only 2.8%. A comparable figure is not available from Gombe during 1970, but the overall rate of reversals for attacks there was only 3.3% (calculated from Table 5.4, Bygott 1974). At Tai in 1993 the overall rate of reversals for agonistic interactions among adult males was slightly higher, at 12.9% (calculated from Table 6.3, Boesch and Boesch-Achermann 2000). Although the alpha male at Kanyawara had just started his tenure during the present study, at other sites high rates of aggression were exhibited by males who had been dominant for some time. During Bygott’s (1974) Gombe study, the alpha male, Mike, was in the sixth year of his tenure. At Tai in 1993, Fitz was in the third year of a four-year tenure (Boesch and Boesch-Achermann 2000). At Mahale, Ntologi reacquired his alpha status approximately five months prior to Nishida’s 1992 observations; he had been the alpha male in 1991, but was briefly ousted from the position by a coalition of two other males (Nishida and Hosaka 1996). Thus, male chimpanzees living in what are normally considered to be stable hierarchies tend to exhibit patterns of aggression similar to those of male baboons living in extremely unstable hierarchies. What accounts for this element of perpetual 56 instability in dominance relationships among chimpanzee males? One possibility is that the fission-fusion nature of chimpanzee social organization makes it relatively more difficult for high-ranking individuals to keep track of social relationships among other males in the group. Male baboons range in relatively stable groups, so their knowledge of the dominance hierarchy and their place in it should be fairly accurate. Chimpanzee males, on the other hand, frequently leave groups and rejoin them hours, days, or even weeks later. Because a high-ranking male can never be certain what political maneuvering has occurred in his absence, it may be necessary for him to continually reestablish his dominance when parties come together. This would help to explain the large proportion of aggression that takes place during the context of reunions. Coalitions are important to chimpanzee males in both maintaining and improving their dominance standing (Nishida 1983, Goodall 1986, Nishida and Hosaka 1996). These coalitions can be very fluid, with males showing high degrees of “allegiance fickleness” (i.e. a male may frequently turn against his former alliance partner if it is in his interest to do so) (Nishida 1983). The fluid nature of both male chimpanzee coalitions, and party composition, then, is likely to explain some, but not all of the apparent instability in male chimpanzee hierarchies. For despite their frequent use of coalitions, males can normally be ranked in a linear hierarchy (Boesch and Boesch- Achermann 2000). And although having a coalition partner may raise a male’s confidence to the point that he challenges a higher-ranking individual, actual 57 reversals in rank generally result from dyadic fights (Goodall 1986, Boesch and Boesch-Achermann 2000). The possibility that dominance hierarchies among male chimpanzees are inherently unstable has important implications for the presumed costs and benefits of social dominance in this species. When baboon hierarchies are unstable, high rank is often associated with elevated glucocorticoid secretion, probably reflecting severe psychological stress (Sapolsky 1992). Chronic exposure to high levels of circulating glucocorticoids is associated with a whole suite of adverse physiological effects, including depressed immune function (e.g. Sapolsky 1993). Furthermore, among baboons the most aggressive males also tend to exhibit the highest levels of circulating T (Sapolsky 1993). Numerous additional costs are associated with chronically high levels of circulating T, including increased metabolic rate and immunosuppression (reviewed in Wingfield et al. 1997). If the most aggressive, highest-ranking chimpanzee males are maintaining chronically elevated levels of cortisol and T, then these would represent previously unknown costs to social dominance. This possibility will be addressed with hormonal data in later chapters. Intercommunity aggression Coalitionary killing has previously been reported from two long-term chimpanzee study sites in Tanzania. In Gombe, males from the Kasekela community made systematic incursions into the neighboring Kahama group’s territory over three years, eliminating their rivals one by one in a series of vicious gang attacks (Goodall et al. 1979, Goodall 1986). Afterward the Kasekela males appropriated much of the 58 Kahama group’s range. A similar group extinction was later documented in Mahale National Park (Nishida et al. 1985). Because chimpanzees at both of these sites were provisioned, it has been argued that lethal intergroup attacks are an adverse effect of artificial feeding, not a natural aspect of chimpanzee behavior (Power 1991). The coalitionary attacks described above add to a growing body of evidence that lethal intergroup aggression is a common strategy employed by male chimpanzees to reduce the coalitionary strength of their neighbors, and expand their territories (Wrangham 1999). The evolutionary benefits of such expansion are clear. First, larger territories may include the ranges of more females. Second, females living in larger territories may have shorter inter-birth intervals and improved offspring survival (Williams 1999). After the group extinctions at Gombe and Mahale, the aggressors appropriated both territory and females from their defeated neighbors (Goodall et al 1979, Nishida et al. 1985, Goodall 1986). Animals generally avoid escalated dyadic aggression because of the inherent risk of severe injury from other, similarly armed adults (e.g. Maynard Smith and Pryce 1973). Coalitionary aggression may allow chimpanzees to mitigate these costs, however. At both Gombe and Kibale, cooperating males inflicted lethal injuries without sustaining appreciable wounds. Selection may therefore have favored a strategy by which male chimpanzees utilize lethal aggression against foreign males whenever there is an extreme imbalance of power (Manson and Wrangham 1991, Wrangham 1999). This idea is supported by the observations from Kanyawara, in which cooperating males attacked a subadult male and a nulliparous female, and 59 appear to have attacked and killed a lone male, but in both cases retreated upon the arrival of additional strangers. 60 4 P a t t e r n s o f St e r o id P r o d u c t io n Introduction Detailed human studies have shown that circulating cortisol levels exhibit a predictable diurnal pattern, peaking between the morning hours of 06:00 and 08:00, and steadily declining to a nadir around 01:00 (Van Cauter 1990). This circadian rhythm is endogenously driven by the central nervous system (Liotta and Krieger 1990). Its synchronization is influenced by both light-dark and sleep-wake cycles (Morin and Dark 1992). Although few data are available from primates in the wild, both Old World monkeys and non-human apes appear to conform to the human pattern of diurnal cortisol production (Whitten et al. 1998). Czekala et al. (1994) reported higher urinary cortisol values for morning versus afternoon samples in both free-ranging male mountain gorillas, and captive lowland gorillas. This finding was replicated with additional samples from male mountain gorillas (Robbins and Czekala 1997). Preliminary data from wild long-tailed macaques {Macaca fascicularis) reveal a distinct and statistically significant decline in urinary cortisol excretion throughout the day (van Schaik et al. 1991). Testicular T production in humans observes a similar diurnal pattern (Van Cauter 1990). The rhythm is, however, distinctly bimodal. T production reaches its zenith early in the morning (between 04:00 and 08:00), wanes throughout the day, effects a modest recovery in the late afternoon (between 16:00 and 18:00), and drops to a nadir around 24:00. This pattern is not independent, but seems to be driven by the 61 diurnal cycle of CRH-ACTH-cortisol production (James et al. 1978). Its adaptive significance is not currently understood. Temporal variation in T production appears to differ between hominoids and Old World monkeys. The former match the human pattern of high morning, and low evening, levels (e.g. Robbins and Czekala 1997). Many (though not all) of the latter— including the rhesus macaque— show peaks of T secretion between 21:00 and 24:00, and a nadir in the early morning (reviewed in Whitten et al. 1998, Dixson 1998; cf. talapoin monkeys: Martensz et al. 1987). The ultimate explanation for these differences is not clear. Numerous minor fluctuations in both T and cortisol secretion are superimposed on the diurnal rhythms described above. Between 12 and 24 times per 24 hours, the testes release a pulse of T in response to the pulsatile release of LH from the pituitary, and GnRH from the hypothalamus (Van Cauter 1990). Cortisol, on average, exhibits seven to nine such spikes in 24 hours (Liotta and Krieger 1990). The pulsatile nature o f steroid secretion complicates the use of plasma measurements to estimate basal hormone titers, since circulating levels of steroid can fluctuate significantly over a matter of minutes (e.g. Weick 1981). The use of urinary assays mitigates this problem, however, because excreted steroid represents an average of circulating levels between urinations (Beisel et al. 1964). As there is a time lag between steroid release and steroid excretion, urinary levels of hormone reflect endocrine status several hours prior to sampling (e.g. Bahr et al. 2000). Time lags to excretion are discussed in more detail below. 62 In addition to these established diurnal fluctuations, steroid production in both humans and chimpanzees manifests a predictable trajectory over the lifespan. T levels in males of both species are low in infancy, rising dramatically at adolescence. In captive chimpanzees, male T shows a significant increase approximately six months prior to the onset of puberty. It then continues to rise until adulthood, resulting in a 30-fold difference between prepubertal and postpubertal levels (McCormack 1971, Martin et al. 1977, Nadler et al. 1987, Mars on et al. 1991, Young et al. 1993). It is not clear from captive studies, though, whether there is a relationship between age and T among fully adult males. Following adolescence, T production in both men and women tends to diminish with age. However, the precise pattern of decay differs between plasma and salivary measurements. In men, plasma concentrations of T appear to be relatively stable until around age 60, after which they show a gradual decline (Vermeulen et al. 1972). Salivary T levels, on the other hand, show a clear linear decline with age (Vermeulen et al. 1999, Ellison 2001). This is probably because salivary T indexes nonbound (free) hormone— the biologically active form of the molecule. Most measures of plasma T index total steroid, which includes both free and protein-bound hormone. The latter is biologically inactive. Recent studies of non-western populations have further complicated the picture of androgens and aging. For although western men exhibit a steady decline in free T levels over the post-pubescent lifespan, this effect is either slight or non-existent in many non-western groups, including the Ache of eastern Paraguay and the Tamang 63 of central Nepal (reviewed in Ellison 2001). As illustrated in Figure 4.1, this difference appears to result from the fact that young males in non-western populations maintain relatively low levels of circulating T; older men in both settings maintain roughly equivalent levels. Why young males in non-western populations should maintain low basal T is not presently clear, though it has been suggested that chronic energy shortages in such populations might be responsible (e.g. Bribiescas 1996). This possibility is discussed in more detail in chapter five. Since both diurnal rhythms and age could hypothetically confound any relationship between circulating T and behavior, it is important to test for these effects. If confirmed, such effects could be controlled for in subsequent behavioral correlations. The expectation is that chimpanzee T and cortisol levels should show a steady decline from morning to evening, as occurs in humans and gorillas. Age, on the other hand, is not expected to show a significant correlation with T among adult males. This is because in humans the age effect is pronounced only in western populations, where food resources are abundant. Wild chimpanzees subsist in ecological settings with relatively limited energy availability; thus, the appropriate comparison is probably with humans living in more traditional settings (Bentley 1999). As noted above, the relationship between T and age in such populations is either weak or nonexistent. Because T is involved in the facilitation of both aggression and sexual behavior in numerous vertebrates (see chapter one), it is reasonable to ask whether the temporal fluctuations in steroid production described above correlate with temporal 64 Il H 600 500 - 400 - 300 - 200 - 100 - 0 o Boston A Lese + Tamang * Ache 10 20 — I— 30 — I— 40 50 60 70 80 A ge Figure 4.1. Salivary T levels and age in four human populations. The four lines represent average decline in T across the lifespan in males from (top to bottom) Boston, Zaire (Lese), Nepal (Tamang), and Paraguay (Ache). Despite a high degree of individual variability, T levels in Boston males show a steep and statistically significant decline with age. The three non-western populations also exhibit declining T across the lifespan, but not as extreme. For the Tamang and the Ache, the decline is not statistically significant. Unpublished data courtesy o f Peter Ellison. 65 changes in chimpanzee behavior. Adolescent aggression is perhaps the clearest example of such a correlation. Increased rates of agonistic behavior during adolescence— when circulating T levels also show a dramatic surge— have been clearly documented in a variety of primates, including chimpanzees (Kraemer et ai 1982, Goodall 1986, Nadler et al. 1987, Pusey 1990, Watts and Pusey 1993). Associations between diurnal patterns of behavior and circadian T variation have received much less attention in the literature. This is partly due to the difficulty of obtaining multiple daily hormone measurements from a study subject without affecting its behavior. Several bird studies, however, have reported correlations between daily patterns of sexual activity, aggression, and circulating T. Balthazart (1976, 1983), for example, collected plasma samples from two groups of domestic ducks every two hours. The first group was sampled in December, the second in March. Although diurnal rhythms of T production differed between the two groups, T levels correlated with the frequency of social display and sexual behavior across six two-hour observation periods. There was, however, no relationship between T and the frequency of several non-social behaviors. Thus, Balthazart concluded that sexual and aggressive behaviors in particular, and not simply activity levels in general, were correlated with circulating T. A similar relationship has been reported between diurnal patterns o f crowing (an androgen dependent vocalization) and circulating T in Japanese quail (Wada 1986). 66 Little is known about the daily pattern of aggressive behavior in wild chimpanzees. Goodall (1986) looked at the diurnal distribution of fights at Gombe during three of the hottest months in 1978 (July through September). She reported that the frequency of fighting peaked between the hours of 08:00 and 09:00, steadily decreased until 11:30, remained low through the afternoon, and then increased slightly at the end of the day (between 17:00 and 19:00). This pattern is very similar to that of diurnal T production in humans (Van Cauter 1990). Goodall, however, attributed the result to temperature, arguing that the Gombe chimpanzees are generally less active in the midday heat than in the morning and early evening. She noted that afternoon temperatures from July through September were generally in excess of 30° C. Diurnal variation in chimpanzee sexual behavior has previously been reported from several long-term study sites. At Mahale the daily copulation rate peaks between 07:00 and 08:00, decreases during the afternoon, and reaches a nadir in the evening (Hasegawa and Hiraiwa-Hasegawa 1990). Chimpanzees at Gombe and Budongo follow a similar pattern, with high morning copulation rates that decline steadily through the day (Tutin 1975, Wallis in press). At Wamba, Kano (1992) reported that 66% of observed bonobo copulations occurred before 09:00 (n=734). This chapter describes temporal variation in steroid production by male chimpanzees at Kanyawara. It begins by examining diurnal variation in both T and cortisol production, and goes on to look at the relationship between T and age in fully adult males. Finally, it investigates diurnal patterns of sexual and aggressive 67 behavior at Kanyawara, in order to see whether they correlate with daily steroid production. More detailed correlations between behavior and steroid production are left to later chapters. Methods I used standard regression analysis to test for an effect of time-of-day on steroid levels (both T and cortisol) across more than 500 urine samples from 11 male chimpanzees. Some males contributed more samples to the data set than others. Thus, I separately examined all of the samples from the most frequently represented male (LK), for comparison with the group results. I then used all of the adult male samples to generate mean male T and cortisol levels for fourteen one-hour intervals (05:00-18:00). (Group means of individual mean male T and cortisol levels were calculated for each interval.) Standard regression analysis was used to test for an effect of time-of-day on mean T across these intervals. I also compared mean morning and afternoon steroid levels for each of the adult males. All 1998 samples were included. Morning samples were those collected before 10:00; all others were considered afternoon samples. Paired values from each male were tested using the Agresti-Pendergrast procedure (Agresti and Pendergrast 1986). Approximately 70% of the samples examined here were collected in the morning. Nearly half of those were taken prior to 07:00. Early morning specimens 68 are prevalent in the data set because they are easy to collect: chimpanzees predictably urinate upon waking in the morning. I used Kendall’s Tau to test for a relationship between age and T in the Kanyawara males. This method tests the hypothesis that T has a monotonie decreasing relationship with age, without assuming that the relationship is linear. Male ages are estimates from observations by Isabirye-Basuta in the early 1980’s, and Wrangham from the late 1980’s to present. Young chimpanzees (15-20 years) exhibit a suite of morphological characteristics that include thick glossy black hair, unbroken teeth, and light facial creasing. Older chimpanzees (>35 years) display thinning brown or gray hair with less sheen, worn or broken teeth, and saggy, wrinkled faces. Additionally, older individuals may move more slowly and deliberately than young individuals. Because observations at Kibale have continued since 1983, the males in this study have been seen in a variety of conditions, allowing for reliable age estimates. In order to test for a correlation between diurnal patterns of agonistic behavior and circulating T, I calculated a mean rate of aggression for each of 12 intervals across the day (6:00 through 17:00). First, I assigned each 40-minute focal follow to the interval in which it began. For example, I placed follows that started between the hours of 06:00 and 07:00 in the 6:00 interval. I placed those that started between the hours of 07:00 and 08:00 in the 7:00 interval, and so on. Second, I summed the total number of aggressive acts (including charging displays, chases and attacks; see chapter four for details) observed in each interval, and divided this figure by the 69 number of male observation hours that fell into that interval. This yielded a rate expressed in aggressive acts per male observation hour. The two final intervals (16:00 and 17:00) each included fewer than 50 chimpanzee hours of observation (21 and 19, respectively), so I performed two separate analyses - one of which omitted these points. During this study, rates of male aggression at Kanyawara increased significantly in parties containing parous females with sexual swellings (see chapters four and six). Consequently, I also examined the relationship between aggression and time of day for parties containing no swollen females (stage 2 or 3). This was done to test whether diurnal patterns of aggression were independent from those of sexual behavior. To investigate diurnal patterns of sexual behavior at Kanyawara, I looked at five years of copulation data (collected between 1993 and 1997) from the Kibale Chimpanzee Project. I generated copulation rates for thirteen one-hour intervals (07:00-19:00) by dividing the total number of copulations observed in each interval by the total number of observation hours from that interval. Standard regression analysis was then used to test for an effect of time-of-day on copulation rate. Results Across the 11 adult males, mean morning T values were significantly higher than afternoon T values (Agresti-Pendergrast procedure, F=10.76, p<.01, Figure 4.2). The same was true for cortisol (Agresti-Pendergrast procedure, F=116.75, p<.0001, 70 n=l 1 males, Figure 4.2). When 522 T samples from 11 adult males were plotted against collection time, a statistically significant decline was apparent throughout the day (r^=.14, p<.0001. Figure 4.3). The same was true for all T measurements from the most frequently sampled male, LK (r^=.19, p<.0001, n=101. Figure 4.4). When means of mean male T levels were calculated for each hour, these showed a strong, statistically significant decline through the day (r^=.83, p<.0001, n=14. Table 4.1, Figure 4.5). The same was true when means were calculated across the samples from each interval without regard to which male they were collected from (r^=.86, p<.0001, n=14). When all cortisol samples from 11 adult males were plotted against collection time, there was also a clear, statistically significant decline throughout the day (r^=.22, p<.0001, n=505 samples from 11 males. Figure 4.6). The same pattern was evident in cortisol measurements from the most frequently sampled male, LK (r^=.29, p<.0001, n=93. Figure 4.7). When means of mean male cortisol levels were calculated for each hour, these showed a strong, statistically significant decline through the day (r^=.85, p<.0001, n=14. Table 4.2, Figure 4.8). The same was true when means were calculated across the samples from each interval without regard to which male they were collected from (r^=.83, p<.0001, n=14). Mean morning creatinine levels, on the other hand, did not differ significantly from mean afternoon values (Agresti-Pendergrast procedure, F=l.ll, p=.32, n=ll males). And when creatinine values from 575 male samples (collected from 11 adult and 4 subadult males) were plotted against collection time, the association between 71 700 600 - 500 - 400 - 300 - 200 - NN 100 - Testosterone Cortisol Figure 4.2. Steroid excretion and time of day. Mean morning T values for each adult male were significantly higher than afternoon T values (Agresti-Pendergrast procedure, F=10.76, p<.01, n=ll males). The same was true for cortisol (A gresti-Pendergrast procedure, F=116.75, p<.0001, n=ll males). Error bars represent the standard error of the mean. 72 I H 250 - T ---------r 5:00 7:00 9:00 11:00 13:00 15:00 17:00 19; Tim e o f Day Figure 4.3. T and time of day. When T values from 522 urine samples (collected from 11 adult males) are regressed against collection time, a statistically significant decrease is apparent (r^=.14, p<.0001, n=522). T values are generally higher and more variable in the morning than in the afternoon. 73 1250 I 1000 - a « I 750 - ♦ ♦ 500 - 250 - 5:00 7:30 10:00 12:30 15:00 17:30 Time of Day Figure 4.4. T and time of day in one male (LK). When T values from 101 urine samples (collected from 1 adult male) are regressed against collection time, a statistically significant decline is apparent (r^=.19, p<.0001, n=101). As in Figure 4.3, T values are generally higher and more variable in the morning than in the afternoon. 74 O u O s M ? /3 ' v - i p 9- 3 1 O c / 3 H T 3 c / 2 ( D ^ 'H j I I C/3 H ( D 13 .1 0 - 4 0 1 II ^\0(N<0T— (\C)T;)-Tr oo co < N ^ < N oo o\ o o ^ oo r~- oo oo m o o o o \ o o o v ^ ^ o o o r ' - ( N T j - \ C ) ^ ( N O \ o o o ^ o \ o \ ^ r 4 o o o r - - ^ O O O O O o o o o o v4 vo (X ) ô\ o o o o o o o o o o o o o o o o o o o ^ < N m '^ u ^ ^ r ~ 'O o i I cg I Q t 0 - 1 0 1 s I I 0 - 4 0 1 I 75 700 600 500 400 300 200 5:00 7:00 9:00 11:00 13:00 15:00 17:00 T im e o f D ay Figure 4.5. Means of mean male T by time of day. Means are from 540 urine samples collected from 11 adult males in 1998. A clear, statistically significant decline in mean male T is evident throughout the day (r^=.83, p<.0001, n=14). 76 o î U I u f 1750 1 1500 - 1250 - ♦ ♦ ♦ * 1000 - 750 - 500 - 250 - 5:00 7:00 9:00 11:00 13:00 15:00 17:00 19 Time of Day Figure 4.6. Cortisol and time of day. When T values from 505 urine samples (collected from 11 adult males) are regressed against collection time, a statistically significant decrease is apparent (r^=.22, p<.0001, n=505). Cortisol values are generally higher and more variable in the morning than in the afternoon. 77 1250 1! ô I J 1000 - 750 - 500 - 250 - T ---------------------r 5:00 7:30 10:00 12:30 15:00 17:30 Time of Day Figure 4.7. Cortisol and time of day in one adult male (LK). When cortisol values from 93 samples (collected from 1 adult male) are regressed against collection time, a statistically significant decline is apparent (r^=.29, p<.0001). As in Figure 4.5, cortisol values are generally higher and more variable in the morning than in the afternoon. 78 •S c / 3 ’ 9- 3 i 0 C / ! ) H "O V I ( D Î î oo c /3 Ü c3 (U 1 S 0 85%) in the urine (as opposed to feces). This is comparable to the time lag to peak urinary cortisol excretion reported for a single captive marmoset (approximately 2.5 hours) and a captive long-tailed macaque (approximately 5.5 hours) in the same study (Bahr et al. 2000). Whitten et al. (1998), reported that in captive chimpanzees cortisol increases following anesthesia were apparent in both fecal and urine samples, with a somewhat longer time lag to fecal cortisol excretion. In humans, urinary cortisol levels appear to lag approximately 2-4 hours behind plasma cortisol (reviewed in Pollard 1995). Diurnal fluctuations in steroid excretion introduce a significant amount of variance into the data set, which complicates comparisons both within and between individuals. Thus, throughout this thesis morning samples are compared and analyzed separately from afternoon samples. A second potential source of variance in this study, urine concentration, is affected by a number of factors, including fluid intake. As described in chapter two, I have standardized all steroid measurements to creatinine concentrations in order to correct for possible dilution effects (Cook and Beastall 1987). There was no 88 evidence for an effect of time of day on excreted creatinine. This was expected, as creatinine is assumed to be produced at a relatively constant rate. A third potential source of variance, age, did not show a significant correlation with T among the adult males sampled here. It is possible that this is the result of small sample size. Alternatively, male chimpanzees in the wild may not exhibit significant changes in T across the adult lifespan, as appears to be the case in some non-western human populations (Ellison 2001). As previously mentioned, such populations may be a better model for wild chimpanzees than Western males, because they subsist in more energetically restricted environments (Bentley 1999). Patterns of T and behavior Sexual behavior by male chimpanzees at Kanyawara peaked early in the morning, and showed a significant, linear decline through the day. The same was true for aggression, though preliminary data suggest a bimodal distribution, with a primary peak early in the morning followed by a slight, secondary peak in the late afternoon (Figure 4.9). These diurnal patterns of behavior are consistent with the circadian rhythm of urinary T excretion in chimpanzees. They are also consistent with the known pattern of circulating T in humans. Tutin (1975) reported the same circadian rhythm for sexual behavior among the Gombe chimpanzees. She also examined a number of non-sexual behaviors for circadian periodicity, and could detect no significant patterns for feeding, resting, or travelling. Grooming was more common in the late morning/early afternoon, but this pattern was independent from that of copulatory activity. Tutin therefore 89 suggested that the morning peak in chimpanzee copulations might result from the zenith of circulating T in each male. T is known to play a role in the activation and maintenance of sexual behavior in male primates (reviewed in Dixson 1998). However, surprisingly little is known about the acute effects of T on male libido (i.e. its effects within one to four hours of production). Experimental evidence from humans indicates that T injections can cause increased sexual arousal in women within three to four-and-a half hours (Tuiten et al. 2000). If the same is true for male chimpanzees, then a peak in T production between 04:00 and 05:00 might be expected to produce an increase in libido around 08:00, when the observed peak in sexual behavior occurs. This possibility remains speculative, however. Although correlations between T, aggression, and sexual behavior are suggestive, there is currently no direct evidence that the diurnal pattern of T production in chimpanzees plays a causal role in establishing diurnal patterns of behavior. Experimental data would be needed to address this point, and few such data exist for any species, let alone a great ape. Experiments with Japanese quail suggest that in birds diurnal patterns of T secretion are not always responsible for diurnal patterns of testosterone-dependent behavior. Wada (1986) showed that in Japanese quail crowing and increased locomotor activity are both androgen dependent behaviors exhibiting clear diurnal rhythms that match the diurnal pattern of circulating T. Castrated males fail to show normal crowing and locomotor activity. Males with T implants, however, exhibit these behaviors with their normal circadian periodicity. 90 Since subcutaneous implants restore circulating T levels without mimicking diurnal variation in T, Wada argues that circadian changes in T cannot be responsible for circadian rhythms of behavior. The relationship between circadian rhythms of T and behavior is further complicated by the observation that changes in social behavior can affect patterns of steroid secretion. In captive talapoin monkeys, for example, circadian T rhythms are altered by increased heterosexual interaction (Martensz et al. 1987). Males housed separately from females show the same general pattern of daily T secretion as males in heterosexual groups, but the latter exhibit a shift in peak T production, from 24:00 to 04:00. The pattern of circulating T during waking hours does not appear to be effected, however. Diurnal patterns of sexual behavior Although circadian rhythms of circulating T suggest a proximate explanation for diurnal rates of chimpanzee aggression and sexual behavior, the ultimate explanation for these patterns is not presently clear. Most chimpanzee copulations are initiated by males (Tutin and McGinnis 1981, Goodall 1986, Matsumoto-Oda 1999), who often seem eager to copulate in the early morning. At both Gombe (Goodall 1986) and Kibale (personal observation) males may even enter the nests of estrous females and court them before they have arisen. Thus, male interest appears to be responsible for the morning peak in copulations. Why, though, should morning copulations be more valuable to male chimpanzees than afternoon or evening copulations? Chimpanzees engage in intense sperm 91 competition (Hasegawa and Hiraiwa-Hasegawa 1990; Figure 4.13), so it is possible that copulating early in the day is advantageous in this regard. In support of this idea, baboons, who also engage in sperm competition, exhibit similarly high rates of copulation in the two to three hours following dawn {Papio anubis and Papio ursinus: Hall and DeVore 1965, Ransom 1981). Additionally, most consort changeovers between male baboons take place in the morning, suggesting that competition for mating at that time is particularly intense (Hausfater 1975, Smuts 1985, Bercovitch 1988, 1989). Very little is known about the mechanisms of sperm competition in primates, so it is not clear what the advantage to morning copulations would be. The general assumption regarding the timing of copulation in relation to paternity is that the male most likely to fertilize is the one who “copulates closest to the interval before ovulation that allows the sperm to become capacitated at the optimum time for fertilization” (Gomendio et al. 1998). From this perspective, only if there were a circadian rhythm in the timing of ovulation, would a circadian rhythm in mating interest make sense. Circadian rhythms in ovulation have been reported for a number of mammalian species (e.g. rodents: Goldman 1999). In humans there is clear evidence for a circadian rhythm in the onset of the preovulatory surge in luteinizing hormone (Edwards 1981, Cahill et al. 1998), suggesting the potential for circadian periodicity in ovulation. There are currently no systematic data on the diurnal patterning of ovulation in chimpanzees or baboons. 92 '■ -a ïj* m m Figure 4.13. Testes and brain of an adult male chimpanzee. In chimpanzees the testes are large in relation to the body w eight, suggesting selection via sperm com petition. Photograph by the author. 93 Male chimpanzees produce copulatory plugs which presumably function either to physically block competitors’ sperm from the fem ale’s reproductive tract (the “chastity belt” hypothesis), to minimize the loss of spermatozoa from the female’s reproductive tract, or both (reviewed in Dixson 1998). Thus, an alternative possibility is that male chimpanzees mating early in the morning obtain an advantage in sperm competition by (1) placing their copulatory plugs before other males can mate and (2) copulating before other males have placed their plugs. However, no studies have yet been conducted that could help to evaluate this hypothesis. Against the idea that high rates of morning copulation are associated with sperm competition, species that clearly do not engage in such competition, like the hamadryas baboon (Biquand 1994, Birkhead 2000), also largely restrict their mating behavior to the morning hours (Kummer 1968). This may represent an exceptional case, since the majority of social interactions among hamadryas take place on their sleeping cliffs, before the day’s march across the savanna has begun (Kummer 1968). A systematic investigation of diurnal mating patterns across the primate order is needed to fully address this issue. Diurnal patterns of aggression Why chimpanzees should be more aggressive in the morning than at other times is similarly unclear. The pattern cannot be explained as a byproduct of the morning peak in sexual behavior (i.e. as mating competition), because it persists on days when tumescent females are absent. Goodall (1986) suggested that high afternoon temperatures were responsible for the diurnal pattern o f attacks during three 94 particularly hot months at Gombe. She argued that chimpanzees were less active during the hottest part of the day, and thus, less likely to be involved in contact aggression. This does not explain the similar pattern found in Kibale, however, where average maximum temperatures are considerably lower than those reported from Gombe (e.g. Struhsaker 1997). Another possibility is that chimpanzees show peaks in aggression in the morning and afternoon, because those are the times when party composition is most likely to change (e.g. Wrangham 1975). As described in the previous chapter, aggression is common in the context of reunions, when males reestablish dominance relations with individuals that they have not seen for hours, days, or weeks. The frequent aggression that occurs when chim panzees unnest is consistent with this interpretation, since males are in effect coming together after being separated by sleep. 95 5 T e s t o s t e r o n e a n d E n e r g e t ic s Introduction Ape mothers, like most mammalian females, invest more in parental effort than ape fathers do. This investment, in the form of internal gestation, lactation, and infant transport, reduces the mother’s capacity to invest in subsequent reproduction. Theory therefore suggests that access to environmental resources such as food should be the primary constraint on female reproductive success, and an important determinant of female reproductive timing (Trivers 1972, Emlen and Oring 1977, Wrangham 1980). For example, females should conceive only when the resources necessary for successful pregnancy and parturition are available (Wasser and Barash 1983, Ellison 1990, Ellison 2001). In line with this expectation, in humans (reviewed in Ellison et al. 1993) and other great apes (Bentley 1999; Knott 1999, 2001) ovarian function is extremely sensitive to fluctuations in both energy balance and activity. For example, acute increases in either workload or nutritional stress predictably reduce circulating levels of ovarian hormones (Ellison et al. 1993). This reversible, short-term suppression can adversely affect fecundity. Subfecund or anovulatory cycles and amenorrhea are common in Western women who are exercising heavily (e.g. Ellison and Lager 1985, 1986, Jasienska and Ellison 1993) or losing weight (e.g. Schweiger et al. 1992). They are also prevalent in non-Western populations, where workloads tend to be higher, and seasonal nutritional stress more severe (Ellison et al. 1989, Panter-Brick 96 et al 1993). This means that women have a lower probability of conception during periods when a successful reproductive outcome is less likely. Adaptive mechanisms of this sort, sensitive to environmental energy availability, appear to be prominent in female primates. By contrast, little is known about the factors affecting male gonadal function. One possibility is that, as in females, energy availability plays a critical role. Preliminary evidence from Western human populations, indeed, indicates that under some circumstances weight loss, dietary composition, and exercise can affect testicular function (reviewed in Campbell and Leslie 1995). For example, in some clinical studies men exhibit decreased levels of circulating T in response to extreme nutritional stress (Klibanski et a l 1981). Consequently, Campbell and Leslie (1995) have argued that environmental stress in non-Western populations may adversely affect spermatogenesis, leading to decreased male fecundity (Campbell and Leslie 1995). While such an effect is theoretically possible, empirical demonstrations have proven problematic for several reasons. First, the relationship between acute nutritional stress and decreased T is ambiguous (Bentley et a l 1993), and consistently appears only under pathological conditions (e.g. in anorexics). Second, rates of spermatogenesis are largely independent of circulating T concentrations, and testicular T concentrations do not correlate with number of spermatozoa (Rea et al 1986, Rommerts 1988, Weinbauer and Neischlag 1990). Finally, modest reductions in sperm count do not appear to have a significant impact on male fecundity (Polanski and Lamb 1988). 97 Nevertheless, a number of studies have been conducted on non-Western human populations to test the idea that circulating levels of T are sensitive to acute changes in energy availability. These are summarized in Table 5.1. Ellison and Panter-Brick (1996) monitored salivary T levels in two populations of Nepalese men, and found little response to seasonal nutritional stress. Nepalese women in the same populations, by contrast, demonstrated pronounced variation in ovarian function between periods of high and low energy availability (Panter-Brick et al. 1993). A similar pattern was observed among Lese horticulturalists in Zaire, with women showing conspicuous seasonal variation in gonadal steroid production, and men showing no such variation (Bailey et al. 1992, Bentley et al. 1993). Bribiescas (1997) monitored salivary T levels in two populations of Ache men in eastern Paraguay. He found no significant differences between a comparatively wealthy population, which enjoyed high net energy availability, and a poorer, more energetically stressed population. The only comparable hormonal data from a non-human ape come from Knott’s (1998, 1999) study of free-ranging orangutans in Gunung Palung National Park, Indonesia. Knott compared gonadal steroid levels in both males and females with data on seasonality in food availability and diet. She documented dramatic fluctuations in food availability between a mast-fruiting period, when both male and female orangutans consumed more than 7000 calories a day, and a non-mast period, when caloric intake fell by more than half (Knott 1998). During the food-poor season, orangutans were energetically stressed and lost weight. Preliminary 98 I en I .8 I a 0 1 I < I 8 I <D cn I IS ■ c i I e S V O I 2 X ) I ■ g a I (U I " O o {I a ÿ 6 j 8 ' g I C/5 I i 8 II il f I C /5 cd I 1 c d N i I ÎI M O V C \ C3\ I ' S X I T 3 I g (D è X (D I y C/5 8 C /3 g 0 3 T 3 O . O II a I b £ I C I h I O §- 8 g 1 C o X o cd :3 W ) C /3 I 1 8 X I (D > X 0 X 1 1 § s o . § 8 (D I C d X I I •s s X o o '+ H g 'cd s b B I in I I d C o s c/3 § o IX a 8 i I § I cd 8 % I o a j i ■ s 8 a I I o n X I d 99 horm onal data from this study appear consistent with results from human populations. Female orangutans showed significant seasonal differences in levels of urinary estrogen metabolites, suggesting reduced ovarian function during periods of food scarcity (Knott 1999). A comparison of urinary T levels between the same seasons in both adolescent and adult male orangutans, however, revealed no differences (Knott 1999). To date, then, there is little empirical support for the idea that testicular function is sensitive to local ecology in the way that ovarian function is. Nor is this surprising. Since the costs of gamete production for human males are small, energy balance is not expected to be the limiting factor on male fecundity that it is for females. Accordingly, the benefits of regulating fecundity in response to acute changes in energetic status are expected to be lower for males than for females. Does testicular function respond to chronic changes in energetic status, however? This possibility is suggested by reports from a number of investigators that men living in non-Western populations maintain relatively low levels of circulating T. For example, both Efe foragers and Lese horticulturalists in the Ituri Forest, Zaire, exhibit significantly lower levels of salivary T than Western controls in Boston (Ellison et al. 1989, Bentley et al. 1993). This is also true for the !Kung San of northern Botswana (Christiansen 1991), the Ache of eastern Paraguay (Bribiescas 1996), and the Tamang and Kami of central Nepal (Ellison and Panter-Brick 1996). It is possible that this pattern of decreased steroid production in non-W estern populations is a result of chronic energy shortages. The evidence for such shortages 100 in these populations includes both low levels of stored fat and short stature (reviewed in Ellison 2001). Bribiescas (1996) has argued that under conditions of chronic energy shortage, men might benefit by reducing levels of circulating T, because T promotes anabolism of m etabolically expensive muscle tissue. He suggests that the suppression of testicular function in response to chronic nutritional stress might be an adaptive mechanism to optimize somatic energy allocation and therefore, increase survival (Figure 5.1). Although this interesting idea is consistent with observed population differences in basal T levels, it has not yet been tested with sensitive measures of muscle mass nor muscle anabolism. In this chapter I evaluate the evidence that T is suppressed in wild chimpanzees during periods of low energy availability. The effects of both acute and chronic shortages are considered. The former can be assessed with data from Kanyawara, by comparing mean male T levels during periods of high versus low food abundance. The latter are more difficult to test, since the Kanyawara chimpanzees are currently the only wild population with well-characterized T profiles. In order to evaluate the effects of chronic energetic status on basal T levels, measurements from free-ranging individuals are compared with those from captive males. Because captive chimpanzees are more sedentary and better fed than their wild cousins, the wild/captive comparison is a useful analogue to the W e stern/non-We stern contrast prevalent in the literature on T and energetics in human males (Bentley 1999). 101 C h r o n ic e n e r g e t ic s t r e s s T e s t o s t e r o n e s u p p r e s s io n I D e c r e a s e d m u s c l e a n a b o l is m D e c r e a s e d s o m a t ic m a in t e n a n c e c o s t s Figure 5.1. Hypothetical adaptive relationship between T and energetics. (Bribiescas 1996) For example, life history variables of female chimpanzees typically accelerate in response to the increased energetic status of captivity. Changes include earlier age at reproductive maturity, earlier age at first birth, and shorter interbirth intervals (reviewed in Bentley 1999). If males are regulating T levels in order to modulate somatic maintenance costs, as suggested by Bribiescas (1996), then captive individuals should exhibit higher T levels than free-ranging individuals. 102 Methods T and chronic energetic status Chronic effects of energetic status on male gonadal function were assessed by comparing mean urinary T levels between chimpanzees in the wild and in captivity. Three-hundred and twelve samples were considered from 11 adult males at Kanyawara. These were all collected prior to 10:00 in order to control for diurnal variation in hormone production (see chapter four). Details of the collection procedures and the T assay are presented in chapter two. Comparisons were made using the Rust-Fligner test (see chapter two). Twenty-eight samples were considered from 11 adult males in captivity: one individual from the Yerkes Regional Primate Center, and ten from the Laboratory for Experimental Medicine and Surgery in Primates (LEMSIP). These samples were collected in the fall and winter of 1995-1996 by a Harvard undergraduate, as part of a thesis project investigating urinary androgen levels in captive chimpanzees and bonobos (Gibb 1996). Keepers at the facilities were instructed to pipet samples off of clean floors, and to “avoid contamination with feces or urine from other individuals” (Gibb 1996:27). Samples were collected in the morning whenever possible, in order to control for diurnal variation in T production. After collection, samples were frozen until they could be shipped on dry ice to the Reproductive Ecology Laboratory at Harvard University. Gibb assayed the captive samples for T and creatinine in 1996, using the same protocol employed in this study, in the same laboratory. The samples were then 103 stored frozen in the Reproductive Ecology Lab at Harvard University. I reanalyzed them in May of 2000. T and acute changes in energetic status In order to assess whether acute changes in energy balance had an effect on steroid excretion by male chimpanzees, I calculated quarterly (January-March, April- June, July-September, October-December) mean T levels for nine of the 11 adult males at Kanyawara (there were insufficient samples to do so for one male, and another disappeared halfway through the study). I then performed all pairwise comparisons to test the null hypothesis that mean male T did not vary by quarter. Rom’s procedure (Rom 1990) was used to control the experimentwise Type I error probability (a=.05). T measurements in this study have been indexed to creatinine in order to correct for variation in urine concentration (see chapter two). Thus, I also calculated quarterly mean creatinine levels for each of the adult males, and performed the same pairwise comparisons to test whether creatinine varied across the year. This was necessary to address the possibility that any observed differences in T levels could be a byproduct of changes in creatinine excretion. Because creatinine is a product of skeletal muscle, creatinine production could theoretically vary with food availability. In humans, for example, deficits in skeletal muscle mass are accompanied by diminished creatinine production and excretion (Weinsier and Morgan 1993). As a direct measure of energy balance, I used Chemstrip urine analysis strips (Boehringer Mannheim, manufacturer) to screen chimpanzee urine samples for the 104 presence of ketone bodies. Ketone bodies appear in the urine when endogenous fat stores are rapidly broken down during periods of fasting. As an indirect measure of energy balance, I examined feeding records from Kanyawara in order to quantify the amount of time that chimpanzees spent eating fruit during each month. When ripe fruit is available, chimpanzees eat it almost exclusively (Wrangham et al. 1998). Because fruit abundance varies temporally, however, chimpanzees are occasionally forced to fall back on lower quality piths, which are distributed throughout the study site (Wrangham et al. 1991, Conklin- Brittain et al. 1998). Thus, when fruit is scarce, the Kanyawara chimpanzees subsist on a diet that is significantly lower in simple sugars, non-structural carbohydrates, and fat, than when fruit is abundant (Conklin-Brittain et al. 1998). These periods of low fruit availability should represent the times of greatest energetic stress. Field assistants in Kibale regularly note at fifteen-minute intervals whether chimpanzees are feeding. If they are, both the species and portion of the plant being consumed are recorded. The percentage of monthly feeding observations in which chimpanzees consumed fruit was used as a proxy for fruit availability. This measure has previously been shown to correlate with direct estimates of fruit abundance from phenology transects (Wrangham et al. 1996). Within the category “fruit,” drupes were distinguished from figs. Drupes typically contain more sugar than figs, and are preferred by chimpanzees (Wrangham et al. 1996). For example, Wrangham et al. (1996) reported a significant negative correlation between drupe eating and fig eating 105 by the Kanyawara chimpanzees, despite the fact that figs were relatively continuously available. Results T and chronic energetic status M ean T levels for the 11 captive males in G ibb’s (1996) study were approximately 50% higher than those of the 11 Kanyawara males (Table 5.2, Figure 5.2). This difference was statistically significant (Rust-Fligner test: Q=5.6, p<.05, n=l 1). In order to confirm the 1996 values, I assayed the same frozen samples four years later (May 2000). Mean T levels for each captive male in 2000 were highly correlated with those from the original study (r^=.91, p<.0001, n = ll). The 2000 values were, however, uniformly higher by approximately 12%. Regardless of which values are used in the analysis, captive males showed significantly higher levels of urinary T compared to those in the wild. T and acute changes in energetic status Mean quarterly T levels were calculated for nine adult males, based on 331 morning urine samples. Multiple morning samples were occasionally collected from a single individual on the same day. In such cases, the mean of the samples was taken in order to preserve independence. This left a total of 290 morning T values from the nine males. No quarter had fewer than 50 such values (Quarter 1=68, Quarter 2=79, Quarter 3=51, Quarter 4=92). 106 Mean urinary T levels for the nine adult males did not differ significantly between the first, second, and fourth quarters o f 1998 (Tables 5.3 and 5.4). However, mean T levels for these males were significantly higher in quarter three (July through September) than in all other quarters (Table 5.4). Mean creatinine excretion, on the other hand, showed no significant variation between any of the quarters (Tables 5.3 and 5.4). Thus, the increased levels of T excretion observed in quarter three are unlikely to be a byproduct of the method for standardizing hormone measurements to creatinine. I screened 926 urine samples for ketone bodies. These samples were collected throughout the study period. Chimpanzees from all age and sex classes were represented, including pregnant females, lactating females, and juveniles. None tested positive for ketones. At least 250 feeding observations were available for each month, from November 1997 to December 1998 (mean=619). Periods of relatively high and low food abundance were clearly discernible from these data (Table 5.5, Figure 5.3). The period of lowest fruit-availability fell between July and October. During that time the chimpanzees fed primarily on piths; no feeding points included observations of chim panzees eating drupes, and fewer than 20% included observations o f chimpanzees eating figs. Fruit consumption was slightly higher in the months of June and November, but still accounted for less than 50% of feeding observations. 107 I I I i 1 I 0 5 i I & & & c / 3 & & & & & & c / 3 c / 3 O O c / 3 s s 'fe s s s s s s W w w p ) H M w w w w W M X X X X h J X X h J h J X Tf Cl Cl C O Cl Cl T f Cl X X I T ) oo m o C N C O T f m X c~ c\ CO T f m iC3 CO ^ O m ^ m CO T t O c^ m X t H § ( Ü O CO m X o o m m ^ Cl Tf rn ^ X m o CO ^ C l C l CO T f ^ 0 0 '^ 0 0 - ^ O X < ^ C 'C 1 1 C 3 l o r ^ c ^ x c s o o x w i x o C ^ C l i O i O - r j - X x T f C ^ X T f X I (N 0\ 0 \ I O N O N V I X § X . o c cb e I c I H I J CO V I ( Ü I ■ g ( D (Ü ? > g I 108 a i l % o 1200 S 1000 - £ H £ î 800 600 400 200 0 Captive Maies Kanyawara Maies Figure 5.2 M ean urinary T levels in male chimpanzees. Captive samples were collected by Gibb (1996) from 11 adult and subadult males at LEMSIP and Yerkes (mean=992, SE=143). Wild samples were collected from 11 adult males at Kanyawara in 1998 (m ean=599, SE=49). Captive m ales exhibited significantly higher levels of urinary T than wild males (Rust-Fligner test: Q=5.6, p<.05). Error bars represent the standard error of the mean. 109 Quarter Mean S Mean SE Q1 (Jan-Mar) 525 6 .533 .02 Q2 (Apr-Jun) 591 7 .520 .03 Q3 (Jul-Sep) 760 6 .539 .06 Q4 (Oct-Dec) 576 4 .556 .02 Table 5.3. Mean levels of T and creatinine by quarter (1998). Values are means of mean am T and creatinine levels from 9 adult males. Testosterone Creatinine Comparison P Reject P Reject Ho jLiQl— j l iQ2 .220 NS .393 NS Ho j l iQ1=j l iQ3 .003 YES .187 NS Ho pQl=p.Q4 .320 NS .152 NS Ho jLiQ2=jLiQ3 .012 YES .962 NS Ho jLiQ2=jLiQ4 .647 NS .400 NS Ho jLiQ3=jLiQ4 .004 YES .561 NS Table 5.4. Significance levels from pairwise comparisons of mean quarterly T and creatinine (n=9 adult males) using the Agresti-Pendergrast procedure. Rom’s (1990) procedure was used to control the experimentwise probability of Type I error (a=.05). 110 5 & I I & Q I I o o r - H o ^ r < i m ^ ' ? t » o O f ^ c ^ o o v - ) v ~ ) o o x t ^ c ^ x X ' ^ ’-Hr r ' U u O i n O o o ' ^ ° ° c n O O X 0 0 0 ^ 0 X Tf T — I C 4 o o (N ra O O O o o X m vn X r - H r < l T i- ï i o o I - S S K S S ^ ' = > ' = > f ' ^ g ^ ' = ' ' = ° o O C n | T — * o o o o o r ' ^ ' ^ o o o o ^ p : ! oooooooooooooo [ ^ ^ Ç : o o o o o o o o o o o a 1 § oo o o o cn ra O N Tf Tf rn rn c> i/N S V ") X o o o o o o V ") ^ m X M m X f-4 m m C " * -4 m cs lo V ") i I I 'S O O O N O N X 1 < D X X g X o S S % z O 'c? G c O < D 0 0 I I 1 o m in ■ g H I in o X g < D X I 00 c X < D £ X G tjO ^ < D I C L , X o ■ g X c o (J G O C O .g I < D ■g I I O G I I I ,s X < D < D § % a & I I I I o 0 1 < D c /5 a 00 C o 00 I ' + H ^ t S , C o < D K 00 J U I I s I I 13 < D X G O a ^ I ■ C O O < D O O C O C < D O o O h 1 3 < D X X C O c/5 2 & X & Q I s Co % U I 2 g C O pH X W ) C O K I O V J Co I I 1 Q 2 2 V J I Î C o S V J I Q I s I C o S u Ê I I C o § Co & X g c/5 1 X 2 < D I I 1 3 X I 111 Il .V ji II -A---Drupes --0—Figs All fruit 0.9 - | 0.8 - 0.7 - 0.6 - 0.5 - 0.4 - 0.3 - 0.2 - 0.1 - A—A N D J F M A M J J A S O N D M onth Figure 5.3. Chim panzee feeding patterns in Kanyawara, November 1997 to December 1998. The solid line shows the total percentage of feeding scans during each month in which chimpanzees were feeding on fruit. The dashed lines show the relative contribution of drupes and figs to the overall total. There is one clear period of fruit scarcity, from July to October. 112 Over the rest of the year, fruit consumption was relatively high, accounting for 60%-80% of feeding observations. The latter figure masks some variation within the high-fruit period, since chimpanzees relied almost entirely upon figs in some months, and ate mostly drupes in others. Despite the small number of months in the sample, the negative correlation between drupe-eating and fig-eating reported by Wrangham et a l (1996) was evident (KendalTs Tau: x=-.45, p<.05, n=14 months). High-quality food was noticeably scarce in the period between July and October, however. Surprisingly, T excretion by male chimpanzees peaked during the period of lowest food availability. The overall relationship between quarterly mean T levels and fruit consumption is illustrated in Figure 5.4. Unfortunately a more fine-grained (monthly) analysis could not be performed, because not all males were sampled in each month. Discussion and Conclusions T and chronic energetic status Very few urine samples were available from captive chimpanzees. From this limited data set, however, it appears that chimpanzees maintain higher T levels in captivity than in the wild. This finding is generally consistent with Bribiescas’s (1996) suggestion that chronic energetic stress should lead to lower levels of circulating T, and decreased muscle mass. Sensitive tests of this hypothesis could 113 -O Fruit Testosterone o « li IÎ II Qtr 1 Qtr2 Qtr 3 Qtr 4 1998 Quarters Figure 5.4 Quarterly fruit availability and mean T levels in male chimpanzees at Kanyawara. Male T levels were highest during the period of lowest fruit availability. (T levels are means of male means.) 114 not be performed, however, since measures of muscle mass were not obtained from these chimpanzees. Caution is necessary in interpreting this result, since there were several potential sources of error in comparing the Kanyawara chimpanzees with captive individuals. First, although weights were not available from G ibb’s subjects, captive chimpanzees tend to be considerably larger than wild chimpanzees. For example, two prime males who died at Kanyawara during the study period weighed 48 kg and 44 kg. This is comparable to one 49 kg male at Yerkes who was bom in the wild. It is, however, 40-50% less than six captive-born males at Yerkes, who average 70 kg (Katherine Paul, personal communication). Comparing T levels between males of such different sizes is problematic. This is particularly true when using urinary measurements that have been corrected for creatinine, because the amount of creatinine produced is dependent upon the amount of skeletal muscle present. This source of error is not likely to have biased the overall result presented here, however, since higher creatinine levels in captive males should make them appear to have lower T levels than they actually do. The small number of samples collected in Gibb’s study is a second potential source of error. Several males were represented by single urine samples, which may not provide an accurate estimate of basal hormone levels. Additionally, despite the fact that I reassayed all of Gibb’s samples, it is still possible that differences in urine collection procedures and assay technique could have influenced the results. 115 Additional information about captive chimpanzees is needed before a definitive comparison of T production in captive and wild individuals can be made. Finally, it should be noted that even if the results reported here are replicated with additional samples, the differences between captive and wild chimpanzees are not necessarily driven by energetics. Instead, they might be a byproduct of some other aspect of captivity. The artificial demographic and spatial constraints in captivity can affect vertebrates in ways which are not easily predictable. For example, Wingfield (1990) notes that captive birds generally maintain significantly lower levels of circulating T than birds in the wild, despite the fact that they are provisioned. In some cases captive T levels are an order of magnitude lower. This pattern does not appear to be related to social stress, since captive birds also exhibit significantly lower levels of corticosterone. Wingfield (1990) suggests that male birds produce less T in captivity because they are not exposed to the full range of social input that they would encounter in the wild. T and acute changes in energetic status T levels in the Kanyawara males were generally stable throughout the year, except during the third quarter, when they rose significantly. This quarter represented the period of lowest food availability. This result is consistent with findings from studies of non-Western human populations, in which T levels are not suppressed in response to short-term changes in energetic status. In a number of human studies, women have been found to suppress ovarian steroid production during periods of low energy availability in which men failed to 116 show a response (Table 5.1). It is not yet known whether female chimpanzees at Kanyawara exhibited decreased steroid production during any of the months examined here. Thus, it is possible that the periods of low food availability in this study were relatively, but not absolutely, bad. Abundant fallback foods, in the form of stems and piths, may have buffered the chimpanzees from the effects of low fruit availability, such that they would not be expected to show a dramatic endocrine response. It is clear, for example, that food availability was never so bad that chimpanzees were rapidly mobilizing endogenous fat stores over a sustained period. Ketone bodies were not detected in any of the urine samples assayed in this study. This contrasts with Knott’s (1998) study of wild orangutans, in which 25-100% of urine samples tested positive for ketones during a period of low-fruit availability, depending on the month. Wrangham notes that in long-term observations from Kanyawara, the chimpanzees never appear to lose weight seasonally, nor do they lose their hair sheen (Wrangham et a l 1996). This contrasts with other sites, such as Gombe National Park, where chimpanzees can appear thin and dull-haired when fruit is in short supply (Wrangham et al 1996). What can account for the fact that chimpanzees in this study not only failed to show decreased T levels during periods of low-food availability, but actually showed increases in T excretion? One possibility is that male T levels in quarter three were responding to the presence of cycling females. T production in a number of vertebrate species increases in response to short-term mating competition (Wingfield 117 et al. 1990). If maie chimpanzees exhibit such an increase, then it might account for these findings, since a large percentage of the samples in quarter three were collected on days when fully swollen, parous females were present (Table 5.6). The possibility that male chimpanzees exhibit T increases in response to mating competition is the subject of chapter six. In order to control for the possibility that reproductive competition was responsible for the pattern of increased T excretion in quarter three, I recalculated the values in tables 5.3 and 5.4 using only samples collected on days when no swollen females were present. This procedure did not significantly affect the overall result. However, such a small number of samples remain (29 for 9 males) when those collected on days with swollen females are excluded, that the results may no longer be meaningful (i.e. one or two samples may not accurately reflect an individual’s basal T levels over a three-month period). Thus, from the current data set it is impossible to distinguish whether the high levels of T observed in quarter three could have resulted from reproductive competition, or whether this pattern would persist in periods of low food availability even in the absence of swollen females. What is clear is that during the period when the least amount of food was available, T levels were highest. Thus, T levels in the Kanyawara males either (1) did not respond to short-term decreases in energy availability or (2) did respond to decreases in energy availability, but the effect 118 I ■ § g I I t I . > 3 3 • c 3 Q I m O N I g = 3 oo 9 V O C O (N r~ CO CO ON 04 ON ^ 0 4 CO a a a cy V 3 t § T O i o G O < D '3 I c s 'H. " O I N O V I I g I I I 5 I 6 C l. i I 119 was overridden by increases in T resulting from social factors. In either case, it seems likely that variation in male T will be better explained with reference to social factors than to energetic status. In light of the fact that males and females face very different reproductive constraints, this conclusion is not surprising. 120 6 T h e C h a l l e n g e H y p o t h e sis Introduction Theory suggests that while the reproductive success of female mammals is constrained primarily by access to resources, male reproduction is limited principally by access to mates (Trivers 1972, Emlen and Oring 1977, Wrangham 1980). Thus, as suggested by the results from chapter five, male gonadal function might be expected to be more responsive to social influences than to energetics. In accordance with this idea, a substantial body of research has shown that in vertebrates the steroid hormone T plays a critical role in facilitating male aggression, specifically in reproductive contexts (e.g Higley et al 1996, Moore 1986, Wingfield 1984, Lincoln et al. 1972).. The evidence for this effect is particularly clear in birds, which show dramatic interspecific and individual differences in temporal patterns of T secretion, explicable by variation in the intensity of male mating competition (Beletsky et al. 1995, Wingfield et al. 1990, Wingfield et al. 1987). Basal T levels in free-ranging birds are slightly higher in the breeding season than in the nonbreeding season. This modest increase is sufficient to support basic reproductive functions, such as spermatogenesis and courtship behavior, but does not interfere with parental behavior. Following this initial rise in basal T, circulating levels of T increase further during periods of heightened male aggression, up to a maximum physiological level. According to the “Challenge Hypothesis,” 121 Maximum Breeding baseline Non-breeding baseline B A MATURATION BREEDING REGRESSION Figure 6.1. The challenge hypothesis. Basal T levels in free-ranging birds are slightly higher in the breeding season (B) than in the nonbreeding season (A). This modest increase is sufficient to support basic reproductive functions. Follov^ing this initial rise, circulating levels of T increase further during periods of heightened male aggression, up to a maximum physiological level (C). Based on Wingfield et al (1990). 122 (Wingfield et al 1990, Figure 6.1), T levels increase when males must respond to threats from conspecifics, particularly during territory formation and mate guarding. Experiments in the laboratory have shown that visual and auditory stimuli from male conspecifics are sufficient to induce elevations in circulating T (Wingfield 1994). T levels decrease during periods when males need to provide parental care to offspring. Experimental manipulations of male birds have shown that high levels of T suppress parental behavior in favor of aggression (e.g. Hegner and Wingfield 1987). Consistent with the predictions of the challenge hypothesis, polygynous birds maintain higher levels of circulating T during the breeding season than monogamous birds do (Wingfield et al 1990). Furthermore, experiments in the wild have shown that T implants can induce polygyny in normally monogamous birds (Wingfield 1984, De Ridder et a l 2000). Bird species that exhibit high levels of paternal care, however, generally maintain low basal T levels. These species also show a greater T response to social challenges than do species with low paternal care, presumably because the latter are already maintaining T levels close to the physiological maximum (Wingfield gr a/. 1987, 1990). The challenge hypothesis thus appears to have wide application for birds and some other vertebrates (e.g. spiny lizards: Moore 1986). It has rarely been applied to mammals. Preliminary data on ring-tailed lemurs (Cavigelli and Pereira 2000) are consistent with the predictions, while data on dwarf mongooses present a puzzling 123 challenge (Creel et al. 1993). The first part of this chapter considers the reproductive ecology of male hominoids in the framework of the challenge hypothesis. Since none of the great apes is a seasonal breeder, and most male apes do not engage in parental care, the specific formulation of the hypothesis must be modified for these species. The emphasis will be on data collected from wild populations, because captivity imposes dietary and demographic constraints that make it difficult to interpret variation in both behavior and hormone levels. The second part of this chapter uses data fi” om Kanyawara to test the predictions of the challenge hypothesis for chimpanzees. Seasonally breeding primates and the challenge hypothesis Although the evidence for an acute effect of energetic status on gonadal function in male primates is equivocal (chapter five), the influence of social interactions is relatively clear. In seasonally breeding primates, basal T levels tend to increase substantially during the breeding season (Dixson 1998). The timing of this increase appears to be coordinated by social cues rather than simply changes in the physical environment (Herndon 1983). For example, troops of monkeys living in the same forest may show intergroup differences in the timing of the onset of mating, but intragroup synchrony is maintained (e.g. talapoin monkeys: Rowell and Dixson 1975). Furthermore, introducing estrogen-treated females to a free-ranging group 124 before the breeding season begins can advance the onset of mating (e.g. rhesus monkeys: Vandenbergh and Drickamer 1974). As is the case with birds, some of the additional T produced by primate males during the breeding season may be necessary for spermatogenesis and the normal expression of sexual behavior (e.g. Zamboni et al. 1974). This requirement seems insufficient to explain the dramatic increases exhibited by some species, however, because relatively low concentrations of circulating T are generally adequate to maintain male reproductive function (reviewed in Dixson 1998). A threshold effect of this kind is apparent in the rhesus macaque (Macaca mulatta). Non-breeding- season levels of T are sufficient to maintain sexual behavior in male rhesus monkeys. Researchers consistently report that intact males exposed to extraspecific females or receptive conspecifics outside of the breeding season immediately exhibit normal sexual behavior; sustained increases in circulating T occur subsequently (Herndon et al. 1981, Ruiz de Elvira et al. 1982, Bernstein et al. 1983, Michael and Zumpe 1993). Castrated rhesus, on the other hand, show a gradual decline in sexual behavior over time (reviewed in Dixson 1998). The challenge hypothesis posits that increases in male T during the breeding season facilitate male-male aggression in reproductive contexts. Although a number of studies have investigated the relationship between T and aggression in rhesus monkeys, most of these have not tested the challenge hypothesis directly. Michael 125 and Zumpe (1981), for example, looked at T and aggression in a captive group of rhesus males over three years, but their subjects were housed in breeding pairs, so the only possible recipients of aggression were female cagemates. The relevance of data collected under such artificial social conditions for understanding male rhesus behavior in the wild is uncertain. Even under these circumstances, however, high levels of circulating T were associated with increased aggression (Michael and Zumpe 1978, 1981). A more appropriate test of the challenge hypothesis requires data from free- ranging primate groups where male-male competition can occur during the breeding season. The best data of this kind come from a study by Higley et al. (1996), who measured CSF free T in free-ranging rhesus monkeys. They reported that individual levels of CSF free T were positively correlated with overall rates of aggression in their subjects. During the breeding season, both T levels and rates of aggression increased dramatically, as males competed for access to females (Mehlman et al. 1997). In captive studies where males have been allowed to interact in outdoor enclosures, both T levels and rates of aggression have also been found to increase during the breeding season (Gordon et al. 1976, Ruiz de Elvira et al. 1982). Thus, in the best-controlled studies, T tends to be positively correlated with aggression. A more specific prediction of the challenge hypothesis suggested by work on birds is that the species which show the most dramatic increases in aggression during 126 the breeding season should also exhibit the largest increases in basal T during that time. The present paucity of hormonal data on free-ranging primates precludes a thorough test of this hypothesis. Data on fecal T in free-ranging muriquis (Brachyteles arachnoïdes), however, are suggestive. Like rhesus monkeys, muriquis are seasonal breeders that engage in intense sperm competition (Strier 1992). In contrast to rhesus, however, muriquis do not compete aggressively for access to females during the breeding season (Milton 1985, Strier 1992). Rather, males take turns copulating with receptive females in a relaxed atmosphere. Milton (1985) describes males lined up on a branch, peacefully awaiting an opportunity to copulate. As predicted by the challenge hypothesis, fecal T levels in male muriquis do not differ between breeding and non-breeding seasons (Strier et al. 1999). This finding is consistent with the idea that acute changes in T have more to do with facilitating aggression in reproductive contexts than they do with sexual behavior, even in species that exhibit intense sperm competition (Strier et al. 1999, Table 6.1). Finally, a review of published serum T levels across species by Whitten (2000) suggests that, as with birds, interspecific patterns of T secretion in primates may be related to social systems. In cercopithecoids, species living in multimale groups show significantly higher T levels than those living in unimale groups. Male-male challenges are expected to be more persistent in mutimale groups. The comparison 127 s I o > W) c/5 ^ to û c 3 u X ) g P Q CD ÙO C C3 § a C D X ) B ' g g e 1 3 CO ■ § 1 c/5 C D C/5 C D I J g I J I 1 ( T j C D C/5 I ■g % g CD C D 1 3 I I I I s 1 % • Is O K -s; C j 5 S C Q Q I c 3 C O S c/5 I 1 W ) 1 13 I CD O § 'i ' i ON ON ON c 3 g H c y 5 I § I a t O Q o O N W) ON I Z il 128 with birds suggests that in unimale groups, T responses to acute challenges should be more pronounced, but this has not been tested. Male-male competition in the great apes Among the extant great apes, operational sex ratios (Mitani et al. 1996), patterns of parental investment (Trivers 1972), and potential reproductive rates (Clutton- Brock and Vincent 1991) are all consistent with relatively high levels of male-male competition. This competition includes both sperm competition and aggressive competition, which takes two general forms. In the long-term, males can compete to maintain permanent access to females; in the short-term (i.e. a single reproductive cycle), males can compete to mate with estrous females. The relative intensity of these forms varies between species. For gorillas, male-male competition within any particular reproductive cycle is rninimal, but competition to attract females to a group and retain them is pronounced (Watts 1996). For chimpanzees, who live in multi-male communities, competition for access to estrous females is much more pronounced. The predictions of the challenge hypothesis should therefore vary between these species, according to the relative importance of short and long-term mating competition. These predictions are reviewed below for each of the great apes, with relevant evidence from field studies of endocrine function. As very few field studies of endocrine function in male apes have been pursued, many of these 129 predictions have yet to be tested. The predictions for chimpanzees are tested below, with data from Kanyawara. Mountain gorillas Mountain gorillas (Gorilla gorilla beringei) live in relatively stable groups consisting of at least one silverback male, one or more unrelated females, and a variable number of offspring (Watts 1996). Both male and female mountain gorillas tend to emigrate from their natal groups (Harcourt 1978). Emigrating males either join an all-male band or range by themselves; females either join a new breeding group or take up with a solitary male (Watts 1996). Females do not range by themselves, and normally leave a group only when a new group or a solitary male are nearby. Silverbacks acquire females by attracting them from other males during intergroup encounters (Harcourt 1978, Watts 1996). Solitary males sometimes pursue established groups in order to challenge silverbacks (e.g. Yamagiwa 1986). Aggressive displays between solitary males and resident silverbacks may serve to advertise the fighting abilities of males for the benefit of females (Harcourt 1978, Sicotte 1993). During these encounters males may also coerce females by committing infanticide; most mothers of infanticide victims transfer to the infanticidal male (Fossey 1984, Watts 1989). To prevent their females from associating with other males, silverbacks employ two strategies. First, they sometimes herd females away fi*om other males (Sicotte 130 1993). This is most likely to happen in newly formed, rather than well-established, groups. Second, they threaten or attack strange males. More than 70% of encounters between strange male mountain gorillas in the wild involve aggression (Harcourt 1978, Sicotte 1993). Silverbacks perform threat displays against intruding males, which involve chest-beating, charging, and branch- breaking. Around 17% o f these encounters escalate to full contact aggression (Sicotte 1993). Fossey (1983) reported that 74% of silverback remains showed signs of healed head wounds, and 80% had broken or missing canines (n=64), which she attributes primarily to fights between males. The intensity of male aggression in such encounters increases with the number of potential female emigrants (Sicotte 1993). Short-term mating competition in mountain gorillas, on the other hand, is relatively subdued, partly because many groups have only one adult male (Harcourt et al. 1981). In groups with more than one male, a clear dominance hierarchy is evident; dominant males frequently interrupt copulations by subordinates, and direct aggression towards them (Watts 1996). The challenge hypothesis makes two predictions about levels of circulating T in male gorillas. First, dominant silverbacks should have higher T levels than subordinate males, because they direct more aggression at subordinate males, and are more involved in aggressive intergroup encounters. Preliminary data on urinary T values from three gorilla groups at Karisoke indicate that this is the case (Robbins 131 and Czekala 1997). Second, silverbacks should show acute rises in circulating T levels during intergroup encounters, when their females are at risk. Tests of this hypothesis have not yet been attempted, but should be possible because such encounters typically last between one and three days (Sicotte 1993). Since encounters between newly formed groups tend to be longer than those between well- established ones (Sicotte 1993), they are expected to lead to greater increases in circulating T. Orangutans Adult male orangutans {Pongo pygmaeus) occupy large and mutually overlapping home ranges that incorporate the overlapping ranges of multiple females. Resident males attempt to maintain exclusive access to sexually attractive females, reacting aggressively to rivals (Rodman and Mitani 1986). Not all adult males attempt to establish residency in a core area. Some range widely, seeking opportunistic matings with estrous females (Galdikas 1981, Rodman and Mitani 1986, te Boekhorst et al 1990). Thus, van Schaik and van Hooff (1996), have characterized the orangutan social system as “roving male promiscuity.” Two relatively distinct morphs of adult male orangutan have been identified in captivity (Kingsley 1988) and the wild (te Boekhorst et al 1990). “Developed” and “undeveloped” males are both sexually mature, but developed males exhibit a suite of secondary sexual characteristics that includes large body mass, fatty cheek pads. 132 longer and thicker hair, and musty body odor. Developed males also emit a long-call vocalization which appears to mediate spacing between males. Low ranking individuals avoid moving toward the calls of high-ranking individuals, and high- ranking individuals move toward the calls of low-ranking individuals (Mitani 1985a). Relationships between developed male orangutans are consistently antagonistic (Galdikas 1981, Mitani 1985a). Developed males show a high rate of wounding and disfigurement from aggressive interactions with other males (Galdikas 1981, van Schaik and van Hooff 1996). Knott (1999) observed visible wounds fi*om male-male competition in 6 of the 12 developed males in her study site at Gunung Palung National Park. Two of these individuals died during the study, both apparently from infected wounds inflicted by other males. Relationships between developed males and undeveloped or subadult males are less consistently antagonistic. Galdikas (1979, 1981) reported that the presence or absence of females had a predictable affect on the level of aggression in these relationships. In the absence of females, developed males were more tolerant of undeveloped males and subadults. In the presence of females, especially consort partners, adult males were more aggressive, chasing other males, shaking branches at them, and making long-calls. Most mating associations between a single male and female orangutan are short term consortships (Galdikas 1981, van Schaik and van Hooff 1996). Dominant males 133 often attempt to terminate associations between females and subordinate males. Undeveloped males will sometimes follow a consorting pair, and attempt to mate when the male is distracted (van Schaik and van Hooff 1996). A large percentage of matings involving undeveloped males occur as a result of male coercion (Galdikas 1981, Mitani 1985b). The challenge hypothesis suggests that the occurrence of mate-guarding, high rates of long-calling, and high levels of aggression by developed male orangutans should correlate with high levels of circulating T. Unfortunately this hypothesis is difficult to test with data from the wild, because in the field it is often impossible to distinguish between subadult males, undeveloped adult males, and adult males who are in the process of developing secondary sexual characteristics (Knott, personal communication). Detailed endocrine studies of captive orangutans, however, provide preliminary support. Kingsley (1982, 1988) and Maggioncalda et al (1999, 2000) examined T levels in juvenile orangutans, undeveloped adults, developing adults, and developed adults. In both studies, undeveloped adults exhibited significantly lower T levels than developed males. Males in the process of developing exhibited peaks of growth hormone (Maggioncalda et al 2000), and in one study were found to exhibit a peak in T slightly higher than the developed males (Maggioncalda et al 1999). Kingsley (1982, 1988), however, reported T levels for developing males intermediate between developed and undeveloped males. Whether or not developing males exhibit 134 a transitory increase in T, fully developed males appear to exhibit significantly higher levels than individuals in developmental arrest. Maggioncalda et al (1999) suggest that the risk of intense aggression from fully developed males makes it necessary for subadult males to carefully monitor their social environment before their final development. They suggest that the frequency of long-call vocalizations and rates of encounter with adult males give growing males a means of evaluating the density of their reproductive competitors, and adjusting their rates of development accordingly. For example, long-call vocalizations could modulate hormone production via neuronal connections from auditory receptors to the hypothalamus (Maggioncalda et a l 1999). Evidence for some mechanism of this sort in the wild comes from Ketambe, where four males were seen to exhibit developmental arrest for more than 10 years (te Boekhorst et al 1990). Anecdotal evidence from captivity (e.g. Maple 1980) suggests that subadults housed with developed males tend to experience delayed development. Caution is necessary in interpreting such reports, however, since in some cases subadults have been reported to develop in the presence of dominant males. Given that styles of dominance relationships between dyads can vary considerably, controlled experiments are needed to better understand the interaction between social context and male development. 135 Two more specific hypotheses can be suggested for the acute effects of mating competition on T in developed male orangutans. First, dominant males should exhibit higher T levels when they are being challenged by subordinates. Such challenges can take place over a series of months, and are accompanied by increased rates of aggression, including aggressive vocalizations, and actual fighting (e.g. Utami and Setia 1995). Second, developed males consorting with females should show higher levels of T when they are mate-guarding. Chimpanzees Chimpanzees {Pan troglodytes) live in fission-fusion societies (Nishida 1979). Every individual belongs to a particular community (or “unit group,” Nishida 1968) containing from 20 to more than 140 individuals. Within communities individuals form temporary associations, traveling sometimes alone and sometimes in parties that may include all of the community’s adult males, and many of the females. The range of a community is defined as the area used by its males, who are more gregarious and travel further than females (Nishida 1968, Goodall 1986, Wrangham 2000). Males are philopatric; that is, they live in the community where they were born, and they cooperate with each other to defend the community range by expelling male intruders. In addition, males occasionally attack and kill individuals on borders or in neighboring community ranges, apparently as part of a strategy to reduce the 136 coalitionary power of neighbors and expand their own territory (Wrangham 1999a). Since the territory encompasses the ranges of numerous females, males appear to benefit in two ways from its expansion. First, larger territories may lead to more females breeding with the resident males. For example, aggressors at Gombe and Mahale appropriated both territory and females from their defeated neighbors (Goodall et al 1979, Nishida 1985, Goodall 1986). In addition, there is evidence that in larger territories mothers have relatively high reproductive rates, through both shorter inter-birth intervals and improved offspring survival (Williams 1999). Improved food density per mother is presumably partly responsible for this effect. Within communities, reproductive competition among males occurs through sperm competition (Hasegawa and Hiraiwa-Hasegawa 1990) and aggression over mating opportunities (Watts 1998). Most copulations occur in multi-male parties without male herding, aggression, or coercion. These opportunistic matings tend to be with nulliparous females, or early in the mid-follicular phase of parous females. Towards the end of the follicular phase (in the periovulatory period) parous females become more attractive to males, as evidenced by a marked increase in male coercion and male-male aggression (Watts 1998). Male-male aggressive competition is thus relatively intense around the time when parous females ovulate. Females can escape the aggressive attentions of males by accompanying a single adult male to a peripheral part of the range in an exclusive “consortship”. The male then maintains 137 exclusive access to her from a period of several days to more than a month (Goodall 1986). Although consortships vary in frequency among populations, they are never the predominant male strategy. For example, data from Gombe, Mahale and Tai indicate that conceptions normally occur in multi-male parties, rather than in consortships (75-94% in parties, Hasegawa and Hiraiwa-Hasegawa 1990, Wallis 1997, Boesch and Boesch 2000). Male-male contests over mating are therefore an important determinant of reproductive success. Victory in these contests is closely predicted by the dominance hierarchy, a set of dominance relationships which invariably includes a clear alpha- male, and often a linear set of relationships among all males (Bygott 1979, Nishida 1979, Goodall 1986, Hayaki et al. 1989). In larger groups, dominance relationships among some middle or low-ranking dyads may be unresolved (Bygott 1979). Chimpanzee males employ frequent charging displays to maintain or to challenge the existing dominance hierarchy, including exaggerated locomotion, piloerection, slapping, stamping, branch swaying, and throwing (Bygott 1979). Dominance reversals are regularly preceded by a period of heightened aggression and increased rates of display by one or both members of the dyad. Reversals are normally the result of fights, which sometimes result in severe wounds (Goodall 1986). Although challenges are frequent, skilled use of coalitions, social grooming, and well-timed aggression normally allow an alpha-male to maintain his top status for several years 138 at a time. During this period, he is often able to dominate mating access to parous females in the periovulatory period. The nature of dominance relations among male chimpanzees therefore suggests two specific predictions from the challenge hypothesis. First, high-ranking males are expected to maintain higher levels of circulating T than low-ranking males. Second, males should show increased T during periods when parous females are in estrous in the late-follicular phase, which is a period of heightened aggression. The former hypothesis is tested in chapter seven. The latter is addressed in the second half of this chapter, below. Bonobos The social lives of bonobos {Pan paniscus) are in many ways similar to those of chimpanzees. They also live in communities of up to 100 or more, within which individuals travel in parties of varying size and composition. Males are philopatric, and like chimpanzees they defend the community range as a territory. Within communities, males form dominance hierarchies with a recognizable alpha-male. Conceptions normally occur in multi-male parties (Kano 1992). Male bonobos generally exhibit less aggression in reproductive contexts than male chimpanzees do, however. They rarely attempt to herd or coerce females, and do not show increased aggression during the periovulatory period of parous females. Finally, status competition appears to be less frequent and dramatic among male 139 bonobos than among chimpanzees. On the other hand, there is evidence that male copulation rate in small multi-male parties is correlated with dominance rank (Kano 1996). Female choice might be partly responsible for this pattern. Accordingly, the challenge hypothesis suggests that, as in chimpanzees, high- ranking male bonobos will have higher levels of circulating T than low-ranking males. However, unlike chimpanzees, male bonobos are not expected to show increased T during periods when parous females are in estrous, since there is no evidence for increased aggression. Finally, the reduced levels of aggression overall among male bonobos suggest either that circulating levels of T may be lower, or that peripheral sensitivity to T is reduced. Humans Humans are the only great ape in which males provide considerable paternal care. Thus, in contrast to the other apes, human males face a clear trade-off between parenting effort and mating effort (Trivers 1972). For men, then, the general predictions of the challenge hypothesis should closely parallel those for monogamous birds. First, men competing for an opportunity to mate should have higher levels of T than those who are not engaged in mating competition. Second, when men choose to invest in offspring rather than mating effort, they should exhibit decreases in circulating T. 140 The difficulty of obtaining long-term behavioral data on humans makes the testing of these predictions problematic. One approach to the first hypothesis is to compare married and unmarried men. In two studies of military personnel, married men were found to have slightly lower T levels than unmarried men, though the difference was modest (Booth and Dabbs 1993, Mazur and Michalek 1998). Although these results are suggestive, this approach is not ideal because it fails to take into account individual variation in mating effort. Some married men are presumably more committed to monogamous relationships than others, so significant variation in reproductive competition can be expected among married men. The hypothesis that paternal care in men should correlate with decreased levels of T is somewhat easier to test. In birds periods of paternal care are predictably accompanied by decreases in male T (Beletsky et al. 1995). In mammals that exhibit paternal care, T levels also decrease from the period of gestation to lactation (e.g. prairie voles: Roberts et al. 1996 Mongolian gerbils: Brown et al. 1995 dwarf hamsters: Rebum and Wynne-Edwards 1999). In line with these findings, the challenge hypothesis predicts that T levels in men should decrease when they are caring for dependent offspring. Storey et al. (2000) tested this idea in 34 couples taking natal classes. Males in this sample, who presumably intended to invest in offspring, exhibited chronic declines in T production over the course of the pregnancy. They also exhibited acute decreases in circulating T in response to visual. 141 tactile, olfactory, and auditory stimuli associated with paternal care (i.e. when listening to a tape of a crying baby, and holding a doll that had been wrapped in a blanket worn by a real baby). Are high levels of T in men generally associated with higher rates of aggression, as suggested for the other great apes? The literature on T and aggression in humans is extensive, but on this point it is inconclusive. In general, studies that compare relatively aggressive groups with less aggressive groups have reported higher T levels in the aggressive groups (reviewed in Archer 1991). However, two widespread methodological difficulties make this test of the challenge hypothesis for human males problematic. First, most studies of human aggression utilize self-reports, peer reports, or psychological tests of aggressiveness, rather than behavioral data on rates of aggression (reviewed in Mazur and Booth 1998). Second, studies that do attem pt to measure behavior directly rarely investigate aggression in the context of male-male reproductive competition. Instead, they commonly include aggression in non- reproductive contexts, such as stealing, or prison violence. One of the few studies which avoids these pitfalls is Flinn’s (1988) elegant work on mate-guarding in rural Dominica, an island in the Caribbean. Because houses in the village were quite open, Flinn was able to make direct behavioral observations of male-male competition, among other behaviors. Two related findings from this study are relevant here. First, males were found to exhibit higher rates of agonistic 142 interactions with other males when their mates were fecund, than when they were infecund. Flinn characterizes this as aggression in the context of mate-guarding. Second, males were involved in fewer agonistic interactions when their offspring were in early infancy. Unfortunately, no hormonal data were collected in this study. However, these represent the kind of rich behavioral data that are needed before a sophisticated account of T and aggression in men can be made. Finally, one may expect the relationship between T, dominance and aggression to be somewhat different in humans compared to the other apes, because male-male competition less routinely takes the form of contact aggression. Among human foragers, for example, competitive aggressive displays are uncommon, and access to females and resources is not generally decided by threats or physical attacks (Boehm 1999). Mazur and Booth (1998) have suggested that T in humans should show stronger correlation with dominance behavior than it does with aggression, because humans strive for dominance in a variety of ways that do not involve aggression. They review a growing body of evidence that T in men responds in predictable ways to competition for status, that not only includes physical challenges, such as wrestling, but contests such as chess or video games. Chimpanzees in Kanvawara In the remainder of this chapter I test the predictions of the challenge hypothesis outlined above using data from the Kanyawara community. A number of questions 143 will be addressed. First, do rates of male aggression increase in parties with maximally tumescent females? Second, are rates of male aggression higher, and is aggression more intense, during a female’s periovulatory period than during earlier stages of maximal tumescence? Third, are increased rates of aggression in reproductive contexts accompanied by increases in T? And if so, are the increases associated primarily with aggression, or with sexual behavior? Methods Male aggression during mating competition I calculated individual rates of aggression by dividing the number of agonistic acts perpetrated by each adult male by his total number of observation hours. Observations from parties containing fewer than two adult males were excluded. Rates of high-level aggression (chases and attacks) and overall aggression (displays, chases, and attacks) were calculated separately. To determine whether rates of aggression differed between reproductive and non- reproductive contexts, I performed two tests, one general and one specific. The former utilized all observations from 1998. A paired, two-tailed test was used to compare individual rates of male aggression on days when maximally tumescent parous females were present with those on days when such females were absent. Only males for whom at least 20 observation hours were available in each condition 144 were considered (nine of eleven). Three parous females were observed cycling at Kanyawara in 1998: Lia (AL), Ekisigi (EK), and Gombe (GO). Cycles from all three were included in this data set A second, more specific test was based upon detailed observations from one of Lia’s cycles. Lia was maximally tumescent from September 25 through October 8, 1998. Observers stayed with her throughout this 14-day period, so it was possible to clearly discern the periovulatory period (the last five days of maximal swelling). I was thus able to test the hypotheses that male aggression is (1) more frequent and (2) more intense during the periovulatory period than at other times. To ascertain whether the frequency of aggression increased during Lia’s periovulatory period, I compared party rates of aggression from seven of the first nine days of maximal tumescence with those from the final five days. Mean daily rates of male aggression were calculated by summing the number of aggressive acts observed during one day’s 40 minute locals, dividing by the number of locals performed, and adjusting the resulting figure to an hourly rate. The Rust-Fligner procedure was used to compare rates of party aggression under both conditions. To ascertain whether the intensity of male aggression increased during Lia’s periovulatory period, I compared the proportion of aggressive acts observed during that period that were chases and attacks (high-level aggression) with the same proportion from parties containing no maximally tumescent parous females. All 145 parties observed during Lia’s periovulatory period contained eight or nine adult males. To control for the fact that male aggression is more common in parties with more males (see chapter three), I utilized observations only from parties that contained eight or nine adult males and no maximally tumescent parous females. Male T during reproductive competition I tested the hypothesis that male T increases during periods of reproductive competition in two ways. In the first, general test, I compared mean male T levels on days when maximally swollen parous females were present with those from days when such females were absent. Samples from both periods were available for all 11 adult males. In a second, more localized test, I compared mean male T levels across the fourteen days when Lia was maximally tumescent with mean levels from the 14 days following her detumescence. Samples from both of these periods were available for eight of the adult males. Paired comparisons were made with the Rust-Fligner test. In a separate analysis, I examined whether males showed increases in urinary T levels in the presence of maximally tumescent nulliparous females. Nulliparous females were considered separately from parous females because the former are less attractive to adult males, and do not incite aggressive mating competition (see discussion below). By considering nulliparous females it is possible to examine the effect that sexual behavior has on male T levels in the absence of intense agonistic 146 behavior. Samples were available from eight of the 11 adult males on days when nulliparous females were present. Two nulliparous females were observed cycling at Kanyawara in 1998: Nile (NL) and Nyenka (NE). Cycles from both are included in this data set. Mean T values were calculated for each male from daily T values for the period in question. To control for diurnal fluctuations in T excretion, only morning samples (i.e. those collected prior to 10:00) were used to calculate daily values. When multiple morning samples were collected from an individual on a single day, the average of these was taken as the daily value. Results Male aggression during mating competition Individual rates of aggression are summarized in Table 6.2. The frequency of male aggression, particularly high-level aggression, increased during periods of reproductive competition. Males were aggressive on average 0.4 times per observation hour in parties containing maximally tumescent parous females, versus 0.28 times per hour in other parties (Agresti-Pendergrast procedure, F=4.8, p=.06, n=9. Figures 6.2 and 6.3). For high-level aggression the difference was more striking. Males engaged in chases or attacks on average 0.17 times per hour in parties 147 C O (D 1 ù O I O. c 1 I ■ g CO (U I CO I CO I p . c ' o I I 1 1 <D 1 I s < 4 1 CO I u § a I o I 0 3 (D % u § 1 o 13 I ii I I Ii I o D O D O o o (N O V O V T j- m m o m T j- V O m m m < N p p ^ — 4 o o o o O o o o o O o o o O V D O T j- V O o O V cn oo O V m < N oo m cn oo m o < N m o < N o o o o O o o o d d d d o m r~ r-" m cn 5 m " (N 5 vd T j- c4 T j- in m m m o 0 0 ^ 0 0 0 0 ^ O O O O O O C D O O V o oo T f T f O V oo oo < N < N T j - m O V T j - O V T j - m o C N O O < N o < N O O < N o d d d d d d d d d d d m v4 o oo On O cn cn o \ ^ p q p H W P 3 > ^ M ^ J c / 3 h -1 V O cn V O s p o o lO ^ G \ V O C S w 0 T 3 O j 0 3 td (D - I C Ih CO i c ! 2 I § ■ p o o Ü # > Î I T 3 § <D " O I I § c S I « 0 < o ^ • S " S g (D 1 S S s ; o 4 3 8 > o 0 3 8 (D B < N VO (D i C % I CO I i " d I I I 43 0 3 O S (U B o 43 . O 0 3 w i Î (D I oj I O â § - a c3 . ( U (D DA 11 148 I I Parties without maximally tumescent parous females H Parties containing maximally tumescent parous females Makoku Stocky Stout Light Brown Big Brown Johnny Imoso Total Aggression (Mean acts per observation hour) Figure 6.2. Individual rates of aggression in non- reproductive and reproductive contexts. Includes data from nine adult males for whom at least 20 observation hours were available in each condition. The four individuals at the bottom of the figure are the highest-ranking males. Parties containing fewer than two adult males have been excluded. Total aggression includes non-vocal displays, vocal displays, chases, and attacks. For additional information, see Table 6.2. 149 Il % I I s . 0.5 0.4 - 0.3 - 0.2 0.1 0 Parties without maximally swollen parous females Parties with maximally swollen parous females Figure 6.3. Mean rates of total aggression in non- reproductive and reproductive contexts. Includes data from nine males for whom at least 20 hours of observation were available in each condition (see Figure 6.2 and Table 6.2). Parties containing fewer than two adult males have been excluded. “Total aggression” includes non-vocal displays, vocal displays, chases, and attacks. The difference is almost significant using a paired, two-tailed test (Agresti-Pendergrast procedure, F=4.8, p=.06, n=9). Error bars represent the standard error of the mean. 150 containing maximally tumescent parous females, versus 0.07 times per hour in other parties (Agresti-Pendergrast procedure, F=22.23, p=.002, n=9, Figures 6.4 and 6.5). Detailed observations from a single cycle support the idea that party rates of aggression increase dramatically in the late follicular phase. During Lia’s periovulatory period, agonistic acts were observed, on average, 3.2 times per hour; in the preceeding days of maximal tumescence this figure was 0.89 times per hour (Rust-Fligner test: Q=7.2, p=.007, ni=7, n2=5. Figure 6.6). When only high-level aggression is considered, a similar pattern is evident. Chases and attacks were observed, on average, two times per hour during Lia’s periovulatory period; in the nine preceeding days this figure was 0.66 times per hour (Rust-Fligner test: Q=8, p=.005, n%=7, n2=5, Figure 6.6). Observations from the same cycle indicate that the intensity of male aggression also increases in the late follicular phase. Eight or nine adult males were present throughout Lia’s periovulatory period, during which 60% of recorded aggression consisted of chases and attacks (33 of 55 incidents). In parties containing eight or nine adult males, but no maximally tumescent females, only 30% of recorded aggression consisted of chases and attacks (19 of 63 incidents). This difference is statistically significant (Storer-Kim procedure, p=.001. Figure 6.7). 151 I I Parties without maximally tumescent parous females ^ 0 Parties containing maximally tumescent parous females Makoku Stocky Stout Light Brown Big Brown Johnny Imoso High Level Aggression (Mean acts per observation hour) Figure 6.4. Individual rates of high-level aggression in non-reproductive and reproductive contexts. Includes data from nine adult males for whom at least 20 observation hours were available in each condition. The four individuals at the bottom of the figure are the highest-ranking males. Parties containing fewer than two adult males have been excluded. High-level aggression includes chases and attacks. For additional information, see Table 6.2. Error bars represent the standard error of the mean. 152 11 ii a 0.25 0.2 - 0.15 - 0.1 - 0.05 - 0 Parties without maximally swollen parous females Parties with maximally swollen parous females Figure 6.5. Mean rates of high-level aggression in non- reproductive and reproductive contexts. Includes data from nine males for whom at least 20 hours of observation were available in each condition (see Figure 6.4 and Table 6.2). Parties containing fewer than two adult males have been excluded. “High-level aggresssion” includes chases and attacks. The difference is statistically significant using a paired, two-tailed test (Agresti- Pendergrast procedure, F=22.23, p=.002, n=9). Error bars represent the standard error of the mean. 153 I I Days of maximal swelling prior to Lia’s POP H Days of Lia’s Periovulatory period II ! g ? A 3.5 ■ 3 ■ 2.5 ■ 2 - 1.5 ■ 1 . 0.5 0 X All A ggression H igh-Level A ggression Figure 6.6. Mean rates of aggression during the swelling phase of one female’s cycle (Lia). Rates of both aggression (displays, chases, and attacks) and high-level aggression (chases and attacks) were significantly higher during the periovulatory period (the last five days of maximal swelling prior to detumescence) than in the preceeding days of maximal swelling (Rust-Fligner test: All aggression Q=7.2, p=.007, ni=7, n?=5; High-level aggression Q=8, p=.005, ni=7, n2=5). Error bars represent the standard error of the mean. 154 I I Low level aggression: charging displays High level aggression: chases and attacks 8-9 Adult Males No Tumescent Females 8-9 Adult Males Lia’s Periovulatory Period Figure 6.7. Intensity of aggression in reproductive and non- reproductive contexts. The figure on the left shows the proportion of charging displays (low-level aggression) versus chases and attacks (high-level aggression) for parties containing 8-9 adult males and no maximally tumescent females. The figure on the right shows the same data for parties observed during Lia’s periovulatory period. The difference is statistically significant (19 of 63 incidents vs. 33 of 55 incidents; Storer-Kim Procedure: p=.001). 155 Male T during reproductive competition Table 6.3 and Figure 6.8 show mean morning T values for 11 adult males in reproductive and non-reproductive contexts. In parties containing maximally tumescent parous females, urinary T levels averaged 784 pmoFmg creatinine. Mean morning T levels in parties without such females were significantly lower, at 570 pmoFmg creatinine (Agresti-Pendergrast procedure, F=8.34, p=.016, n=ll. Figure 6.9). Male T levels during one of Lia’s cycles were consistent with this general pattern (Table 6.4 and Figure 6.10). In the 14-day period when Lia was maximally tumescent, mean morning T levels for eight adult males averaged 885 pmoFmg creatinine. In the 14-day period following Lia’s detumescence, T levels in the same males averaged only 597 pmoFmg creatinine (Agresti-Pendergrast procedure, F=8.6, p=.022, n=8. Figure 6.11). Male T levels did not appear to increase in the presence of maximally tumescent nulliparous females. In parties containing such females, urinary T levels for eight males averaged 537 pmoFmg creatinine. In parties containing no maximally tumescent females, mean morning T levels for the same males were slightly higher, at 652 pmoFmg creatinine. This difference is not statistically significant (Agresti- Pendergrast procedure, F=4.01, p>.05, n=8. Table 6.5, Figure 6.12). 156 K T k 3 § P 4 1 Ü C L , H g o Tf- o o ID ^ es s ID (N (N 4> = I 3 tin ( S ^ _ _ _ r- g (N G\ 0\ ID V £ ) V O V O 0 \ OO V O ID T f- V O r-4 cn V O C D oo C D r- r- T j- g g g 6 II ■§i 1 1 g - CD OS CD ^ 04 Tf" O /—s OS CD I D ID CD 0 4 OS 0 4 r - O ^ 0 4 0 4 T f ID O O V O C4 0 4 I CD t| - C D ' ^ 0 ^ 0 4 V O C D O V O ^ V O C D I D V O C D O ^ O O I D O O I D O O I D C D O O r - I D T f C D i D r - T j - i D V O o O ' ID > i I T 3 g g § C • i c 2 r 3 c3 g (U C D vd I H c (U 0 c/5 1 I I I x : o L. I c3 r g N a 0 (/) t . S ' 1 c /3 1 I I o o CD c« g i o CD c 3 CO g I g C/3 ’g > • o I I I <D c/5 o ■s g 8 O "O CD cr P (/) t C/3 CD I Î c CD a .S' ■g I c/3 I > G H c/3 I O/ -S s I 0 I I 0 3 CD I I P4 N CD I pH C/3 I > 157 I I 1998 days with no maximally tumescent parous females 1998 days with maximally tumescent parous females Imoso Stout Johnny Badfoot Stocky Light Brown Makoku Big Brown 1----------- 1 ----------- r 750 1000 1250 1500 Mean Testosterone Pmol/mg Creatinine Figure 6.8. Individual male T levels in reproductive and non-reproductive contexts. The white bars indicate mean T levels for males on observation days when no maximally swollen, parous females were present. The shaded bars represent mean male T levels on days when maximally tumescent, parous females were present. Only morning samples are included. For additional information, see Table 6.3. 158 h| II If I 1998 days with maximally tumescent parous females 1998 days with no tumescent females Figure 6.9. Means of mean male T levels in reproductive and non-reproductive contexts. The difference is statistically significant using a paired, two-tailed test (Agresti-Pendergrast procedure, F=8.34, p=.016, n=ll). Only morning samples are included. Error bars represent the standard error of the mean. For additional information, see Figure 6.8 and Table 6.3. 159 I i I g I I i C D g CD C 4 e C 4 C 4 C D O C D C 4 O S C D ID S O C D 1 — ( O S I O S ID C 4 C 5 0 ( X ) ID -rf C D C 5 0 ID r-4 H t * H Ï — H s s o r-4 ID C 4 ID CD CD P 3 ^ ^ o o ^ C X ) o o ID ID i s ID s; ID O O ID OO o o O S i s e d cd C D M o o 160 14 days of Lia’s maximal tumescence I I 14 days immediately following detumescence Imoso Stout Johnny Light Brown Makoku Stocky Big Brown 500 750 1000 1250 1500 Mean Testosterone Pmol/mg Creatinine Figure 6.10. Individual male T levels during and immediately following reproductive competition. The white bars indicate mean T levels for males during 14 days when one popular female (Lia) was maximally tumescent. The shaded bars represent mean male T levels for the 14 days immediately following detumescence. Only morning samples are included. For additional information, see Table 6.4. 161 H I II I f I 1000 900 800 700 600 500 400 300 200 100 0 14 days of Lia’s maximal swelling 14 days following detumescence Figure 6.11. Means of mean male T levels during and immediately following a period of reproductive competition. The difference is statistically significant using a paired, two-tailed test (Agresti-Pendergrast procedure, F=8.6, p=.022, n=8). Only morning samples are included. Error bars represent the standard error of the mean. For additional information, see Figure 6.10 and Table 6.3. 162 g 0 1 % I t g O f X i I C N C\ o o O o o C N < N r- O < N < N ^ 04 (N m Tf 04 ^ Tj- v o o o o m m 0 4 en v o 0 4 0 4 m o o r- o o Ci e m 0 4 0 4 co m m r- m r- S Tf en m o o T — I V O m o o o v o o o * ^ 0 4 o \ m 0 4 e s v o e n e n O s r - o o o o T f o o e n I eu g 0 K n 1 c (L> c « K g C Æ I g 0 f X i 1 S = 3 'I K T ï t § ( D I ( D ( D m sd • s H g E I I 3 O O T 3 ( D « 3 3 « 3 § > B H I g I ? c / ) t I g a g C A - g I - § o î - ë § « 3 ( D N c « ( D l C / ) c/3 (D O > ' O (D c/3 0 ■ s 1 1 r j O T 3 ë c / 3 t C/3 I I ■ g I I g : ja ! s c 3 s g (D t C/3 ë 1 0 c 1 f . o (D I § I M c/3 I C/3 < C/3 (U c/3 1 eu g c/3 K g (D I ; & 1 I O c/3 (D I I 163 nj II If 800 - 700 - 600 - 500 - 400 - 300 - 200 - 100 - N.S. 1998 days with maximally tumescent nulliparous females 1998 days with no tumescent females Figure 6.12. Means of mean male T levels during mating periods with little aggressive competition. The difference is not statistically significant (Agresti- Pendergrast procedure, F=4.01, p>.05, n=8). For more information see Table 6.5. Error bars represent the standard error of the mean. 164 Discussion and Conclusions Male chimpanzees at Kanyawara showed clear increases in aggression when traveling in parties with maximally tumescent parous females. Because I was not always able to observe cycling females tliroughout the entire period of maximal swelling, it was sometimes impossible to distinguish the periovulatory period. As a result, all days of maximal tumescence have been weighted equally in the general analysis (Figures 6.2 through 6.5). A more specific analysis of one complete cycle indicates that rates of aggression differed between the initial days of maximal swelling and those of the periovulatory period (Figure 6.7). Thus, it seems likely that had only days in the periovulatory period been considered, the individual increases in male aggression depicted in figures 6.2 through 6.5 would have been even more dramatic. Captive studies indicate that ovulation in female chimpanzees is most likely to occur either on the final day of maximal tumescence, or on the day of detumescence (Graham 1981). It sometimes occurs a day or two earlier. Thus, it is not surprising that both male mating interest and sexual aggression peak during the final days of maximal swelling. Urinary T levels in the adult males also increased significantly in the presence of estrous females. This finding is consistent with the prediction of the challenge hypothesis that changes in T will be associated with reproductive aggression. This 165 observation alone, however, does not rule out an alternative hypothesis that the observed change in T was associated primarily with reproductive function, rather than aggression. This possibility is suggested by several studies showing that male T rises in the presence of strange females (in domestic mice, even without copulation: Macrides et al. 1975), in the presence of cycling females (talapoin monkeys: Eberhart et al. 1980), and— in humans— after exposure to erotic material (Hellhammer et al. 1985). With chimpanzees the difficulty lies in separating the effects of male sexual behavior from those of male reproductive aggression. Fortunately, we are provided with a natural control in the form of nulliparous females. For, as was the case with muriquis (above), nulliparous chimpanzee females experience male sexual behavior without the intense sexual aggression. Specifically, copulation rates for nulliparous females and parous females do not differ at Kibale. Wrangham (in press) estimated copulation rates for maximally tumescent females per 100 hours of time spent with individual adult males. Across 12 adult males, the median copulation rate was 4.7 copulations per 100 hours for 6 parous females versus 4.3 copulations per 100 hours for 7 nulliparous females. This difference is not significant. There are, however, important rank-related differences in male copulation patterns. Low ranking males tend to copulate with nulliparous females at higher rates, and with parous females at lower rates than high-ranking males. High-ranking 166 males are apparently less interested in nulliparous females, perhaps because the latter copulate at least twice as many times for a given birth as the former, suggesting that each copulation with a nulliparous female is less valuable (Wrangham in press). Whatever its explanation, the consequence of this lack of high-ranking male interest in nulliparous females, is that males do not generally show possessive mating behavior when nulliparous females are cycling. Thus, nulliparous females permit us to examine whether male T levels in reproductive contexts are influenced primarily by aggression, or whether sexual behavior in and of itself is enough to induce significant increases. The data presented here seem to uphold the predictions of the challenge hypothesis. Male T rose only in response to parous females, when reproductive aggression was also intense. Sexual behavior alone did not induce significant changes in urinary T levels. How, then, do chimpanzees generally fit into the challenge hypothesis, as illustrated in Figure 6.1? As previously mentioned, seasonally breeding birds maintain a non-breeding baseline (A) level of T which increases slightly to a breeding baseline (B) in order to facilitate reproductive function during the breeding season. A similar effect in chimpanzees seems unlikely, since they are not seasonal breeders, and are therefore unlikely to predictably encounter cycling females at a certain time of year. (Despite equivocal evidence for a seasonal distribution of births at Gombe 167 (Wallis 1992), it is clear that both conceptions and births at all sites occur throughout the year.) Similarly, for birds living in highly unpredictable environments, where signals such as photoperiod and temperature are not reliable indicators of food abundance, the breeding season may not occur at the same time every year. Birds living in such unpredictable environments may permanently maintain breeding baseline levels (B) of T (reviewed in Wingfield et al 2000). When should chimpanzee T levels rise from the breeding baseline (B) to the maximum physiological level (C), however? In birds, the nature of this transition can be predicted by the breeding system. Birds that exhibit high levels of paternal care generally maintain low basal T levels. The same species show a greater T response to social challenges than do species with low paternal care, presumably because the latter are already maintaining T levels close to the physiological maximum (Wingfield et al 1987, 1990). Since chimpanzees do not exhibit paternal care, the comparison with birds suggest that they should show weak or nonexistent T responses to aggressive males or receptive females, because they are always maintaining T levels near the physiological maximum (C). The data from Kanyawara, however, suggest that male chimpanzees show a relatively pronounced T response to male-male competition over estrous females. What can account for this discrepancy? 168 One possibility is the rarity with which males encounter cycling parous females. In the wild, a female chimpanzee may not have her first infant until she is 13-15 years old. Depending on the study site, interbirth intervals average 5-7 years (reviewed in Knott in press). Finally, at Kanyawara parous females average only about 5 cycles per birth (Wrangham in press). As a result, cycling parous females are expected to be temporally rare. In the present study, for example, maximally swollen parous females were present on only about 13% of observation days. Their actual frequency was likely lower, since it was easier to find and follow chimpanzees when they were assembled in large parties with estrous females. Because there are significant physiological costs associated with increased androgen secretion (see chapter seven), it may not pay for male chimpanzees to always maintain high T levels when opportunities for conception are rare. The relative rarity of such opportunities for male chimpanzees results in a different strategy, whereby T increases are tied to brief periods of intense competition. The suggested theoretical model is presented in Figure 6.13. Because the presence of estrous females is unpredictable, male chimpanzees consistently maintain breeding baseline levels (B) of T. When females are maximally tumescent, male T levels rise to level C, facilitating competition. The magnitude of these increases appears to be unrelated to dominance rank. Rates of aggression in the absence of estrous females, however, are clearly related to dominance rank, and dominance rank 169 is predictive of success in mating competition. This raises the possibility that males differ by rank in their breeding baseline (B) levels of T, as illustrated in Figure 6.13. This possibility is examined in the next chapter. 170 Maximum Breeding baseline Non-breeding baseline MATURATION BREEDING iŒGRESSION No E strous Female POP No E strous Females Females Figure 6.13. A theoretical version of the challenge hypothesis for chimpanzees. In contrast with most birds, males never fall from their breeding baseline (B) to a non-breeding baseline (A). However, males may exhibit rank-related differences in their breeding baseline (B). During periods of heightened male aggression (especially the periovulatory period), T levels rise to a maximum physiological level (C). Compare with Figure 6.1. 171 7 T e s t o s t e r o n e a n d D o m i n a n c e Introduction Male chimpanzees compete persistently and aggressively for status within their communities (Goodall 1986). Such competition involves the performance of elaborate agonistic displays, and the maintenance of complex social alliances (e.g. Nishida and Hosaka 1996), activities that demand significant investments of time, energy, and resources (e.g. meat: Nishida et al. 1992, Mitani and Watts 2001). Attendant risks include severe injury, or even death, in agonistic encounters (e.g. Goodall 1992). The substantial costs associated with dominance-striving in chimpanzees imply the existence of compensatory benefits. Although high-rank could theoretically confer either a survival advantage (through enhanced access to resources), or an indirect reproductive advantage (via kin selection), it is the direct reproductive benefits of male dominance that have received the most attention in the primate literature (e.g. Cowlishaw and Dunbar 1991). Altmann’s (1962) “priority of access” model proposed that, across primates, dominance rank and reproductive success should generally be correlated, because high-ranking males can be expected to monopolize matings with estrous females. Tests of the priority-of-access model have been conducted with field data from a wide range o f primates, producing mixed results. Both methodological and conceptual difficulties contribute to the disparate findings (Fedigan 1983). The primary methodological difficulty lies in assessing male reproductive success. Most 172 studies employ indirect measures, such as copulation frequency, that may or may not reflect actual paternities (Fedigan 1983). However, recent advances in extracting, amplifying, and sequencing DNA from hair and fecal samples have facilitated the direct assessment of male reproductive success in the wild. The result has been a series of molecular genetic studies that strongly support the priority of access model (e.g. mandrills: Dixson et al. 1993, baboons: Altmann et al. 1996, rhesus macaques: Bercovitch and Nürnberg 1997, mangabeys: Gust et al. 1998, Hanuman langurs: Launhardt et al. 2001), as well as a number that do not (e.g. ring-tailed lemurs: Pereira and Weiss 1991, rhesus macaques: Berard et al. 1993). The primary conceptual difficulty lies in operationalizing the notion of dominance. Most researchers have attempted to construct linear dominance hierarchies based upon the outcome of agonistic or approach/avoidance interactions. However, this approach can be problematic for several reasons. First, patterns of aggression and submission between dyads may not be consistent. Second, dominance hierarchies may not be linear. Third, agonistic dominance may not predict dominance in other contexts (e.g. resource competition) (Fedigan 1983, 1992). The confusion surrounding both the concept of dominance, and tests of the priority-of-access model is mitigated by the realization that primate societies differ in the degree to which aggressive competition is an effective social strategy. Thus, the degree to which groups are socially stratified varies along a continuum from relatively despotic (where the priority of access model generally holds, dominance 173 hierarchies are linear, and the benefits of high rank are considerable) to relatively egalitarian, (where dominance relations are unstable and poorly defined) (e.g. Vehrencamp 1983). Differences in the degree of despotism appear to be driven by predictable ecological factors, such as the costs of dispersal, and the precise nature of intergroup and intragroup competition (e.g. Wrangham 1980, Vehrencamp 1983, van Schaik 1989). The effects of ecology on social system and dominance style are complex, and the details are beyond the scope of this review. It is sufficient to note that African apes clearly lie on the despotic end of the political continuum (Boehm 1999), where conceptual difficulties associated with the notion of dominance are less problematic. Chimpanzees, for example, have a distinct vocalization (the pant-grunt) that functions as a formal signal of subordinance (e.g. Bygott 1974, Goodall 1986). Pant- grunt orientation has repeatedly been shown to correlate with a range of aggressive and submissive interactions (Bygott 1974, Hayaki et al. 1989, Nishida and Hosaka 1996, Boesch and Boesch-Achermann 2000). It is often possible to rank male chimpanzees in a clear linear hierarchy. And in cases when insufficient dyadic interactions have been observed to produce a linear hierarchy, males can normally be assigned to a dominance level, i.e. alpha, high, middle, or low (e.g. Bygott 1974, Watts 1998). As noted in chapter three, dominance rank and frequency of aggression are generally correlated in male chimpanzees. High-ranking males also enjoy greater access to preferred resources (Goodall 1986). 174 Does social dominance lead to increased reproductive success in chimpanzees and other great apes, however? I begin this chapter by reviewing both behavioral and genetic data that relate to this question. Next I look at the relationship between dominance and T production across a range of primates, and review the evidence that social stability is an important intervening variable. Finally, 1 use data from Kanyawara to test the prediction from chapter six that dominance rank and T in chimpanzees are correlated. This leads to a consideration of the potential physiological costs associated with dominance-striving in this species. Dominance and Reproductive Success in the Great Apes Mountain gorillas For male mountain gorillas, the reproductive benefits of high rank appear to be relatively straightforward. Females mate primarily in the 2-3 days around ovulation, copulating only about 20 times per conception. Male aggression is more frequent and intense when more estrous or potentially emigrant females are available (Watts 1996). Between groups, males who win fights tend to attract females and to defend them successfully against other males. Because most mountain gorilla groups (60%) contain only one adult male (Robbins 1995), intragroup competition over females is rare. Within multi-male groups, alpha males dominate mating access by guarding estrous females and interrupting the mating attempts of lower-ranking males. Low-ranking males do obtain some copulations, however (Robbins 1999). Tolerance towards rival males 175 within groups may be shown in two circumstances. First, when their daughters reach reproductive age, silverbacks allow young males to mate them. Second, silverbacks sometimes develop tolerant relationships with “follower” males, who may assist them in intergroup encounters. Follower males are probably normally sons or brothers. Thus, it is expected that an alpha-male s reproductive success can be closely predicted by his duration of tenure and the number of females in his group. Orangutans For male orangutans, on the other hand, the reproductive significance of high rank remains mysterious. The fact that males can undergo delayed maturation (see chapter six) implies that survival falls when males become dominant, and that high rank must therefore confer reproductive benefits. In support, developed males are intolerant of each other’s presence, and they successfully supplant subordinate males from consortships with females (van Schaik and van Hooff 1996). Nevertheless, females appear to mate with several males per conception, and copulations by low- ranking males (often involving coercion) are common (Mitani 1985). Genetic data are needed to assess the reproductive significance of such copulations. Chimpanzees The mating patterns of chimpanzees are complex, and it took twenty years of field study to outline their basic structure. In every population that has been examined in detail, group mating is the predominant pattern. When estrous females are present, chimpanzees gather in large parties containing members of both sexes. 176 Females mate repeatedly with multiple male— sometimes every adult male in the communit— resulting in several hundred copulations per conception. Much mating occurs without any attempts by males to mate-guard. Thus, male dominance rank does not always correlate with total copulation rate (e.g. Gombe: Tiitin 1979; Kanyawara: Wrangham in press; Mahale: Hasegawa and Hiraiwa- Hasegawa 1990, Nishida 1997; Taï: Boesch and Boesch-Achermann 2000). This led early observers to conclude that male rank was unrelated to reproductive success. Long-term observations, however, have proven to be consistent with the priority- of-access model. First, even in the absence of overt mate-guarding, total copulation rate and dominance rank are sometimes correlated in the expected way, with higher- ranking males mating more often (e.g. Gombe 1973-1974: Goodall 1986; Mahale 1992: Nishida 1997). For chimpanzees, this is more likely when there are few adult males in the community, and a stable hierarchy (Goodall 1986). Second, high-ranking males attempt to maintain exclusive access to estrous females during the periovulatory period. Intense aggression during the final days of a fem ale’s maximal tumescence leads to a consistent relationship between dominance rank and mating success when conceptions are most likely to occur. Alpha males in Gombe regularly had the highest copulation rates during periovulatory periods (Tutin 1979, Goodall 1986). And in Kanyawara, male rank was positively correlated with copulation rate with parous, but not nulliparous females (Wrangham in press). In Mahale, copulation rate with parous females in the periovulatory period was three times higher for the alpha-male than for any other 177 individual (Hasegawa and Hiraiwa-Hasegawa 1990), and was correlated with dominance rank overall (Nishida 1997). At Ngogo, periovulatory females were guarded by top-ranking males, either alone or in coalitions of two to three males (Watts 1998). Female mating cycles can also take place wholly or partially in consortships of isolated male-female pairs (e.g. 8% of cycles, Boesch and Boesch-Achermann 2000). Consortships have been estimated to be responsible for a small but significant proportion of conceptions (Gombe 25%, Wallis 1997; Mahale: < 20%, Hasegawa and Hiraiwa-Hasegawa 1990; Taï: 6%, Boesch and Boesch-Achermann 2000). Consorting males in Taï were always high-ranking (Boesch and Boesch-Achermann 2000). No clear effect of rank is known in Gombe. Consortships may, however, offer opportunities for males to escape the constraints of competing alliances of lower-ranking males (Boesch and Boesch-Achermann 2000), or of domination by higher-ranking males (Goodall 1986). The benefits to females are unclear, but probably include reduced social stress and increased feeding time. A recent study by Constable et al. (2001) suggests an additional benefit to females, and may provide an explanation for the relatively high frequency of consortships reported from Gombe compared to other sites. Constable and colleagues note that at Gombe females tended to consort with low-ranking males when they had high-ranking male relatives in the community. For such females, consortships may represent a tactic to avoid inbreeding since male chimpanzees sometimes attempt to force copulations with their unwilling mothers or maternal 178 sisters. Two predictions follow. First, between communities, consortships should be more common in cases where females disperse less often. Second, within communities, non-dispersing females should show higher rates of consortship than immigrant females. Additional study will be needed to evaluate these predictions. For both chimpanzees and bonobos, behavioral observations have now been corroborated by preliminary genetic data from three communities (Taï, Gombe, and Wamba). These data indicate that some low-ranking males do indeed achieve paternity, but that, as expected, high-ranking males tend to do better. Reports from Gagneux et al (1997, 1999) that 7 of 13 infants born into one community at Taï were sired by extra-community males initially confused this picture. For although it was known that extra-community copulations sometimes occur (i.e. when females make temporary visits to neighboring communities (Goodall 1986), or when consortships are formed with individuals from neighboring communities) it had generally been assumed that these were relatively rare. Gagneux et a l (1997, 1999), however, reported that several mothers had disappeared for only a few days around the inferred time of conception, suggesting that they were purposefully seeking copulations with strange males during their most fertile periods. More recently, however, Constable et a l (2001) have reanalyzed the genetic data from Taï, identifying probable fathers within the community for six of the seven infants originally thought to have been sired outside the community. In addition, a new study by Vigilant and Boesch (2001) suggests that extracommunity conceptions at 179 Taï occur rarely, if ever (31 of 31 infants genotyped were fathered by community males). Genetic data from Gombe are remarkably consistent with behavioral indications that high-rank is reproductively advantageous (Constable et al. 2001). Five of thirteen infants bom there were likely to have been sired in the consortship context, while nine were sired in the group-mating context. Low and middle-ranking males were responsible for all five paternities in the consortship context. O f the nine infants sired in a group setting, five were sired by the alpha male, two by males who subsequently became alpha, one by a high-ranking male, and one by a middle- ranking male. A similar result was reported from a genetic study of bonobos at Wamba (Gerloff et al. 1999). There between five and seven of ten infants were fathered by the two highest-ranking males in the community. The two lowest ranking males fathered none of the offspring who were genotyped. In neither Gombe nor Wamba were any extra-community conceptions detected. T, Dominance and Social Stability The relationship between male rank and T has been examined in a range of primates, producing a variety of results. Dominant males had higher T in wild mountain gorillas (Robbins and Czekala 1997) and several captive groups (squirrel monkeys: Coe et al. 1979, Mendoza et al. 1979; talapoin monkeys: Keverne et al. 1982; mandrills: Wickings and Dixson 1992; marmosets: Abbott 1993; rhesus monkeys (in an all-male group): Rose et al. 1971). However, rank and T were 180 unrelated in wild long-tailed macaques (van Schaik et al 1991), and in other captive groups (squirrel monkeys: Steklis et al 1986; Japanese macaques: Eaton and Resko 1974; vervet monkeys: Steklis et al 1985; rhesus monkeys (in heterosexual groups): Gordon et al 1976, Bercovitch 1993). One explanation for these conflicting results is that when dominance relationships are unresolved (i.e. when hierarchies are unstable), rank is more likely to be correlated with aggression, and therefore with T (Sapolsky 1993). The best evidence for this effect comes from studies of free-ranging baboons. As discussed in chapter three, during periods of rank instability, high-ranking baboon males exhibit extremely high rates of aggression. During these periods they also maintain higher rates of circulating T than low-ranking males. During periods of dominance stability, however, the opposite occurs: low-ranking males exhibit both the highest rates of aggression, and the highest levels of circulating T (reviewed in Sapolsky 1993). As discussed in chapter three, high-ranking males at Kanyawara maintained the highest rates of aggression throughout this study, despite the fact that the dominance hierarchy appeared to be stable. Thus, the prediction is that high-ranking males at Kanyawara should exhibit the highest rates of urinary T. This prediction is tested below. An additional prediction is that changes in individual rank should be accompanied by changes in circulating T, an effect that has been found in numerous primate studies (e.g. Rose et al 1975, Bernstein et a l 1983). Unfortunately this 181 prediction could not be tested at Kanyawara, because no changes in the dominance hierarchy were observed during 1998 (see chapter three). The idea that social defeat or success causes a corresponding change in T has been tested extensively in humans. Following various forms of competition, winners have been shown to have higher increases in T than losers (chess: Mazur et at. 1992, wrestling: Elias 1981, tennis: Mazur and Lamb 1980, running: Dessypris et al. 1976 Experimental evidence suggests that androgen administration alone has little effect on social rank in primates (Gordon et al. 1979, Green et al. 1972, Bernstein et al. 1983), but rises in rank can occur (Joslyn 1973, Clark and Birch 1946). Methods Determining dominance ranks Dominance ranks are commonly assigned to male chimpanzees based on the distribution of pant-grunt vocalizations (e.g. Bygott 1979). As previously mentioned, pant-grunt orientation is highly directional, and reliably correlates with several other measures of dominance (Bygott 1974, Hayaki et al. 1989, Nishida and Hosaka 1996, Boesch and Boesch-Achermann 2000). Thus, it is possible to assign ordinal dominance ranks to chimpanzees by simply arranging males in a dominance order such that no male pant-grunts to an individual higher in rank. However, it is not uncommon that the observed distribution of pant-grunts is insufficient to distinguish male rank beyond the basic categories of alpha, high, middle, and low (e.g. Bygott 1974, Watts 1998). Within these categories there may 182 not be enough data to determine dominance between individuals. Such was the case in the present study, so in order to compare dominance rank and T, I assigned males to one of these four dominance levels based on pant-grunt orientation. In order to acquire enhanced resolution on dominance relationships among the adult males, I also assigned dominance ranks based on the outcome of decided agonistic bouts. To do this I employed a probabilistic model (Jameson et al. 1999) that takes into account the number of opponents that an individual has successfully defeated, and the relative success of those opponents in their own agonistic encounters. This method has two major advantages over standard techniques for determining ordinal dominance rankings. First, it can be used to predict dominance relationships for dyads that have not yet been observed to encounter each other. Second, it provides information about how dominant individuals are over others (i.e. the scaled values indicate probabilities associated with pairs of ranks). This probabilistic method is similar to those used for ranking chess players. Further discussion can be found in Jameson et al. (1999) and de Vries and Appleby (2000). To begin, each male is assigned a scale score based on the equation: <^,)=[«=(2WrN,)/2N,] (1) W, =the number of observed encounters that a, won Ny =the number of encounters in which at was involved o c =2.50663 (a constant taken from the Taylor expansion of the normal distribution) 183 After each individual receives an initial score from equation one, equation two (below) is used repeatedly to rescale individuals until their scores converge, and further iterations of the procedure no longer induce changes. 5(ay)=[2(WrLy/NJ+Qy (2) Ly =the number of encounters that ay lost Qy=the mean scale score of the individuals that ay met in agonistic encounters The final term (Qy) takes into account the number of times that an individual has met each of his opponents (i.e. Q y =(scale value of opponent from encounter 1 + scale value of opponent from encounter 2 + scale value of opponent from encounter 4...)/total number of encounters). Males in the current study received scale values falling between -1 and 1. Testosterone measurements Mean T values were calculated for each male from daily T values for all observation days in 1998. To control for diurnal fluctuations in T excretion, means from morning samples (i.e. those collected prior to 10:00) were calculated separately from afternoon samples. When multiple morning or afternoon samples were collected from an individual on a single day, the average of these was taken as the daily morning or afternoon value. Two tests were performed to examine the relationship between T and dominance. In the first test, mean T values for individual males were correlated with dominance level (i.e. (1) alpha, (2) High, (3) Middle, (4) Low). In a second test, mean T values 184 for individual males were correlated with scaled dominance ranks. Morning and afternoon samples were analyzed separately. Results Only 89 pant-grunts were observed among the 11 adult males (Figure 7.1). Sixty-four percent of these (57) were directed toward the alpha male, Imoso. Thus, it was impossible to determine precise dominance relationships among several of the males solely from the pant-grunt data. There were sufficient observations, however, to assign males to a dominance level (Table 7.1). No reversals were recorded in any of the pant-grunt interactions (i.e. within all possible dyads, pant-grunts were given by only one of the males). Figure 7.2 shows the distribution of decided agonistic encounters among dyads. These data were used to generate a probabilistic dominance rank for each male. Males received scaled values between -1 and 1, with 1 representing high rank. These are presented in Table 7.2. The distance between the scaled values indicates a difference in the magnitude of dominance (i.e. males with similar values are more difficult to distinguish in rank than those further apart.) The correlation between dominance level and mean morning T across the 11 adult males was statistically insignificant (Table 7.3, Figure 7.3, r=.28, p=.4). This was due largely to low T levels in two high-ranking males, BB and TU (shown in gray in figure 7.3). When these males are excluded (see discussion below), the correlation is much higher (r=.66, p=.05). 185 Across the adult males, the correlation between mean afternoon T and dominance level was statistically significant (Table 7.3, Figure 7.4, n=ll, r=.62, p=.04). The correlation between mean afternoon T and dominance rank was also statistically significant (Figure 7.5, n=l 1, r=.62, p=.04). Discussion and Conclusions T and rank The relationship between male rank and T in the Kanyawara males differed by time of day. For morning samples, the correlation was relatively weak, and statistically insignificant. For afternoon samples, the correlation was considerably stronger, and statistically significant. Given that individual variance in hormone production is greater in the morning than in the afternoon (e.g. Van Cauter 1990, Dabbs 1990, Czekala et al. 1994), it is perhaps not surprising that morning samples show a weaker correlation with behavioral measures. Interestingly, however, in the present case the discrepancy between the morning and afternoon correlations is due almost entirely to low morning T values in two high-ranking males. Big Brown and Tofu. When these males are omitted from the morning sample, the correlation is comparable to that found in afternoon measurements (Figures 7.3 and 7.4). Is there something exceptional about these individuals that could help to explain this effect? It is perhaps significant that Big Brown and Tofu were the alpha and beta males, respectively, from mid-1994 to late 1997. They were ousted from their top-ranking 186 MS A .T TU BB LB SL ST BE YB SY LK MS * 6 4 4 13 12 1 1 7 9 AJ * 1 4 4 2 TU * 2 1 1 2 BB * 1 1 1 1 LB * 2 1 SL * 1 1 ST * 1 1 BE * YB * 2 2 SY * LK * Figure 7.1. Pant-grunts among adult males at Kanyawara. Entries are the number of times that the column male pant- grunted to the row male. Data are from both focal follows and ad libitum observations in 1998. 187 MS AJ BB TU LB SL ST BF YB SY LK MS * 5 3 2 11 3 3 2 1 4 4 AJ 1 * 2 1 4 4 1 1 6 3 BB * 1 2 4 1 2 2 TU 1 * 1 2 2 4 1 LB * 2 2 2 2 2 SL 1 * 3 2 1 ST ■ k 3 BF * 1 1 YB ■ k SY k 1 LK k Figure 7.2. Aggressive interactions among adult males at Kanyavv^ara. Entries are the number of times that the column male lost a dyadic agonistic bout ^vith the ro^v male. Data are from both focal fbllo^vs and ad libitum observations in 1998. 188 Rank Male ID Alpha MS High-ranking AJ, BB, TU Middle-ranking BF, LB, SL, ST Low-ranking LK, SY, YB Table 7.1. Dominance levels for 11 Kanyawara males, 1998. Male ID Scaled Rank Imoso MS 0.739 Johnny AJ 0.465 Tofu TU 0.320 Big Brown BB 0.258 Slim SL -0.078 Light Brown LB -0.138 Badfoot BF -0.467 Stout ST -0.505 Stocky SY -0.641 Yogi YB -0.800 Makoku LK -0.853 Table 7.2. Dominance ranks for 11 Kanyawara males, 1998. Ranks have been assigned based on the outcome of decided agonistic bouts using a probabilistic model (Jameson et al. 1999). 189 s a en I < î en b J c H c < N so OS r j - 1 — H os U ~ ) OO os c d m so < N 1 —4 V ) CS| ( N r j - r - 1 m r j - r j - r j - r j - m r j - en r j - r j - < N O i î ( D I I I V ) O O V ) \o T j- V ) < N < N o\ < N rj- o < N C < l rj- o o <N 1 - H co f o o o < N r ^ v o m v ~ ) m < N v ~ ) m m (N r o^ mmr j - r j - 0^ co oo vo oo rJ -O Nr ^v~ )rJ - o v - i v o r - ' o v o o ^ v o m o o ^ O N r ^ r j - m r j - v ~ ) V ~ > r ^ v o ^ r j - a î a -C -C Æ Û X ) t) f l Û X ) a a I ’ -a CD I CD I î I s s s n s s s s s s s s î e s r f s s f e s s a a I en o o o\ os t c d C / 3 I II 11 - ^ .o o 1 3 > g I I b O C I § S r n I H t3 O C / 3 I > o î a 2 § N C / 3 < D C / 3 ( D I p. c I > a 1 o 0 2 1 o 190 I I H a 0 ^ 1000 1 900- 1 800 - I 700 - u U 600 - a 500 a 400 - P L h 300 - 200 - Alpha High Medium Low M ale R ank Figure 7.3. Dominance level and mean morning T in Kanyawara males. Across the 11 adult males the relationship is statistically insignificant (solid line: r=.28, p=.4). This is due largely to low T levels in two high- ranking males, BB and TU (shown in gray). The dashed line shows the correlation when these males are excluded (r=.66, p=.05). 191 600 5 0 0 - H ^ 400 g f S 2 300 - 200 Medium Alpha High Low M ale R ank Figure 7.4. Dominance level and mean afternoon T in Kanyawara males. The relationship is statistically significant (r=.62, p=.04) 192 600 i l I .S 500 1 O C Q « a i f II 400 - 300 - 200 -1.0 -0.8 -0.5 -0.3 0.0 0.3 0.5 M aie A ggression R ank 0.8 1.0 Figure 7.5. Dominance rank and mean afternoon T in Kanyawara males. The relationship is statistically significant (r=.62, p=.04). 193 positions by Imoso and Johnny in late 1997, when this study began (see chapter three). Since the loss of high-rank in numerous primates is associated with declines in circulating T (e.g. Bernstein et al 1983), this recent deposition might explain the conspicuously low morning levels observed in these males. It should be noted, however, that the declines in T associated with the loss of rank in captivity are relatively transient. This was not the case for Big Brovm and Tofu, whose urinary T levels were no higher at the end of 1998 than they were at the beginning. Furthermore, if the low morning T levels observed in Big Brown and Tofu can be attributed to their loss of rank, then why are their afternoon T levels so high in relation to those of the other males? One possibility is that whereas morning T levels reflect baseline status, afternoon T levels in some way reflect the outcome of social interactions through the day (e.g. wins and losses in aggressive interactions), and such outcomes are closely predicted by rank. Thus, despite the fact that Big Brown and Tofu lost their top-ranking positions, they remained high-ranking throughout the study period. Thus, they still won a large proportion of their agonistic encounters with other males, resulting in higher afternoon T. Additional observations of males rising and falling in rank at Kanyawara will clearly be necessary to test this hypothesis. In the meantime, however, it appears that among the males at Kanyawara rank and T are generally correlated, even when the dominance hierarchy is stable. Particularly striking is the “alpha-male effect” (Whitten 2000). Throughout the study top-ranking Imoso persistently exhibited the highest levels of urinary T, in both the morning and the afternoon. 194 As suggested in chapter three, the association between high rank and T during a period of rank stability presented here is consistent with the idea that chimpanzee dominance hierarchies are inherently less stable than those o f baboons. Alternatively, it is possible that the 14-month period of data collection at Kibale represented an unusually prolonged period of rank instability, and that the behavioral m easures used to evaluate rank stability in baboons are inappropriate for chimpanzees. If the latter is true, then the correlation between rank and T would not be expected in other, more stable chimpanzee hierarchies. I am currently collecting hormonal data from two communities in Gombe National Park, that may allow for a test of this hypothesis in the future. T and the costs of social dominance If high T facilitates aggression and dominance in male chimpanzees, and these are indeed associated with enhanced reproductive success, then why not maintain perpetually elevated T? In birds such elevation interferes with paternal care, but since most male apes do not invest in offspring, this is not a relevant constraint. The main problem with perpetually elevated T in chimpanzees would seem to be the costly nature of frequent, escalated aggression. Combat between adult male chimpanzees is always risky, with high potential for severe or fatal wounds (e.g. Goodall 1992). Such a cost is seen in birds; for example, male brown-headed cow birds with T implants develop serious injuries from escalated fighting with other males (Beletsky et al. 1995). Thus, it is obvious why some individuals might, at 195 times, accept low T and subordinate status to avoid the costs of male-male competition. Do high-ranking males, though, suffer additional costs from maintaining increased levels of circulating T? Wingfield et al. (1997) have identified several costs associated with high T, three of which may apply to chimpanzees: direct energy costs, indirect energy costs, and immunosuppression. Direct energy costs refer to the fact that high T levels may increase metabolic rate and decrease fat reserves. Indirect energy costs of high T reflect the increased energetic expenditure associated with aggressive displays, territorial aggression, and fighting. Finally, there is some evidence that high levels of T may suppress the development of immune responses. This effect is not well-documented, however. One problem is that many studies compare T levels with some proxy of immune function, such as parasite load (e.g. Saino and Moller 1994). Experimental investigations using more sophisticated measures of immune function have failed to reveal a deleterious effect of high T on immune function in birds (Hasselquist et al. 1999, Evans et al. 2000). There is some evidence, however, that this effect may be more pronounced in the great apes. Women, for example, generally have stronger immune responses than men, and exhibit a higher incidence of some autoimmune diseases (Grossman 1984, Talal et al. 1987); these patterns may be partly related to decreased levels of circulating androgens in women. 196 8 St r e s s, D o m in a n c e a n d A g g r e s s io n Introduction The basic physiology of the vertebrate stress response has been well described in both proximate and functional terms (e.g. Selye 1956, Chrousos and Gold 1992, Sapolsky 1993, Genuth 1993). The initial response, mediated by the sympathetic nervous system, involves the secretion of catecholamines (epinephrine and norepinephrine) from the adrenal medulla. These molecules facilitate a general state of arousal and vigilance (the classic “fight-or-flight response”). Their wide-ranging physiological effects include an increase in heart rate, and changes in arteriolar constriction that function to shunt blood toward exercising muscles (Genuth 1993). The secondary stress response involves an increase in the concentration of circulating glucocorticoids, 21 -carbon steroids produced by the adrenal cortex under stimulation of ACTH from the pituitary. This increase occurs within a few minutes of the onset of a stressor. In many birds and rodents the primary glucocorticoid is corticosterone; however, in primates, including chimpanzees and humans, it is cortisol (Nelson 1995). The most conspicuous of cortisoTs far-reaching physiological effects involve the regulation of metabolism. Cortisol increases the availability of glucose in the bloodstream by promoting the conversion of protein to glucose via hepatic gluconeogenesis (Genuth 1993). At the same time it inhibits the ability of insulin to promote glucose uptake and glycogen synthesis. These effects are understandable 197 from an adaptive perspective: long-term energy storage is curtailed, as energy reserves are mobilized in response to crisis (Sapolsky 1993). Cortisol also suppresses immune function, through a complex set of interactions with the cells of the immune system and their chemical messengers (reviewed in Sapolsky 1993, Genuth 1993, Rabin 1999). Most dramatically, it decreases the number of circulating T-cell lymphocytes, interrupts their transport, and interferes with their function. For this reason glucocorticoids are used clinically to suppress immune responses, for example, when transplanted organs or tissues are threatened with rejection (Genuth 1993). The adaptive significance of glucocorticoid-induced immunosuppression is not presently clear. It seems unlikely that such suppression represents a physiological constraint, because it involves active removal of lymphocytes from the bloodstream (Buchanan 2000). A popular adaptive hypothesis is that cortisol restrains the immune system during periods of stress in order to prevent a surge in immune function that could lead to autoimmune disorders (reviewed in Sapolsky 1993, Genuth 1993, Buchanan 2000). This idea remains largely untested. Although acute rises in cortisol represent an essential, adaptive response to short term stressors, it is clear that chronically high levels of glucocorticoid secretion can lead to pathology, including gastric ulcers and atherosclerosis (Sapolsky 1993). Other adverse effects o f sustained glucocorticoid exposure include protein breakdown and muscle wasting (Genuth 1993). Finally, prolonged suppression of 198 the immune system by glucocorticoids can lead to increased risk of infection. This effect is well-documented in human studies (reviewed in Rabin 1999). For example, chronic, severe stressors such as divorce or the death of a family member are frequently associated with illness (O ’Leary 1990). And in some studies basal cortisol levels correlate with observed incidence of illness (e.g. Flinn and England 1997). Because the adverse physiological consequences of chronic exposure to high levels of glucocorticoids are so striking, measurements of these hormones have been used to quantify the costs of specific social strategies and interactions. Particular interest has focused on the relative costs and benefits of social dominance (Creel 2001). Stress Hormones and Social Dominance Numerous studies have demonstrated that the magnitude of the cortisol response marshaled by an individual depends upon both the physiological and the psychological aspects of the stressor that induces it (reviewed in Sapolsky 1993). In general, the less predictable that a stressor is, and the less control that an individual has over it, the greater the cortisol response. Studies of captive rodents provide an illustration. If two rats are simultaneously given a series of electric shocks, then the rat that is able to decrease the shock rate by pressing a lever will show a less dramatic glucocorticoid response. The same is true for the rat that receives a 199 warning bell before each shock is administered. This difference occurs despite the fact that in both cases the rats end up receiving an identical program of shocks (Weiss 1970). The observation that both unpredictability and loss of control are associated with increased levels of stress led to an early expectation among researchers that in social animals subordinate individuals should generally m aintain higher levels of circulating glucocorticoids than dominants (Creel and Sands in press). Tests of this hypothesis in primates have been performed primarily on captive populations, producing mixed results. For some groups this relationship generally appears to hold true (reviewed in Sapolsky 1992). In others, however, researchers find no correlation between cortisol and dominance (squirrel monkeys: Steklis et al. 1986; talapoin monkeys: Martensz et al. 1987; rhesus macaques: Bercovitch and Clarke 1995). To complicate matters, in still other groups a previous correlation between cortisol and rank can disappear when the dominance hierarchy is unstable (squirrel monkeys: Coe et al. 1979; talapoin monkeys: Keverne et al. 1982; macaques: Shively and Kaplan 1984). Sapolsky (1992) explained the latter observation in terms of a shift in the rank distribution of stress. Specifically, in stable hierarchies, high-rank incurs psychological benefits associated with predictability and control. The result is relatively low basal levels of stress hormones. In unstable hierarchies, however, the 200 social position of high-ranking individuals is threatened, leading to a heightened stress response. This interpretation is supported by long-term observations of olive baboons in Kenya. Sapolsky (1992) reported that during a period in which the dominance hierarchy was stable, high-ranking baboon males were less aggressive than low- ranking males, and exhibited lower circulating levels of cortisol and T. During a period of extreme dominance instability, however, high-ranking males were more aggressive and showed higher levels of cortisol and T than low-ranking males. Although Sapolsky’s baboon data are largely consistent with findings from captivity, more recent data from the field suggest that low levels of circulating glucocorticoids are rarely associated with high rank in free-ranging populations (Creel 2001). For example, of nine field studies conducted on social vertebrates, only Sapolsky’s reported higher levels of glucocorticoids in subordinate individuals (Table 8.1). In two populations no significant difference was reported between dominants and subordinates (mountain gorillas: Robbins and Czekala 1997, long tailed macaques: van Schaik et al. 1991). In the final six populations, dominants maintained higher basal levels of circulating glucocorticoids than subordinates (Florida scrub jay: Schoech et al. 1991, alpine marmot: Arnold and Dittami 1997, ring-tailed lemur: Cavigelli 1999, dwarf mongoose: Creel et al. 1996, African wild dog: Creel et al. 1997, wolf: Creel and Sands in press). 201 <D J Ü g 00 I g o 00 C Z 3 (D I 00 0 \ 0 \ a a S o o u 00 O N O N ' r - H II g c o Q Q 1 X ) 1 § o Q 8 Î C / 3 •g I * I I ■ § 00 « g - C I •I N O N O O N O N O N O N O N O N T— c T -M r — 4 c d c d c d CD O CD CD CD O O O O »-( U U U g c o Q 1 P h P h n X ) X ) c/3 eu O o O c o o O ' d p q P Q £ g a o c /) 8 P h < 3 < I a i f ! -§ gj S' g c / 3 C/3 li O N O N o 3 X ) i I I I I 1 I § c a C e • Ü S C i. S ' K O .% , 1 J b O X ) § o § c Q Q Q o o Ph P Q P Q c / 3 c / 3 c / 3 (U (D (D o o o CD CD CD Ph Ph Ü H 2 L . % I X ) 'c 3 i I n J . 2 k . I I X ) I a i s g g 2 a s Æ I c § I I ■ § 00 i 'T 3 < ü is - C 5 I § I & k . O c d I 1 ■ gJ i C /3 g g c o Q i § t I g P4 U I a •§ c / 3 0 1 1 CD g C - § g o O I O 3 00 1 202 Thus, it appears that contrary to the predictions derived from captive observations, high levels of stress hormones may represent a general cost of social dominance in group-living species (e.g. Creel 2001). However, the underlying behaviors that link social dominance with increased stress are not well understood (Creel and Sands in press). Aggression is a prime candidate for such a behavior, because in the wild high-ranking individuals are often involved in agonistic interactions at higher rates than subordinates. This is certainly the case for high- ranking chimpanzees (see chapter three). Thus, characterizing the relationship between cortisol and dominance in chimpanzees should help to clarify the role of frequent aggressive interactions on the production of social stress. The remainder of this chapter examines the relationship between stress, dominance and aggression in Kanyawara chimpanzees. I begin by testing the hypothesis that dominance rank positively correlates with basal cortisol level. Next, I test the hypothesis that individual males show increased cortisol secretion during periods in which parous females are maximally tumescent. As shown in chapter six, such periods are marked by heightened levels of male aggression. Finally, the idea that increased involvement in aggressive interactions can lead to increased cortisol secretion suggests that parous females should exhibit higher levels of circulating cortisol when they are maximally tumescent compared to other times. As previously demonstrated, parous females receive aggression at significantly higher rates during such periods of maximal swelling. Nulliparous females, on the 203 other hand, are not expected to show cortisol differences between swelling and non swelling periods, because males do not possessively mate-guard nulliparous females. Data from two females are examined in detail here, offering a preliminary test of this hypothesis. Methods Cortisol and dominance Two tests were performed to examine the relationship between cortisol and dominance. In the first test, mean cortisol values for individual males were correlated with dominance level (i.e. (1) alpha, (2) High, (3) Middle, (4) Low). In a second test, mean cortisol values for individual males were correlated with scaled dominance ranks (see chapter seven). Morning and afternoon samples were analyzed separately. Mean cortisol values were calculated for individual males using daily cortisol values from the entire twelve-month study period. To control for diurnal fluctuations in cortisol excretion, only morning samples (i.e. those collected prior to 10:00) were used to calculate daily values. When multiple morning samples were collected from an individual on a single day, the average of these was taken as the daily value. Male cortisol during reproductive competition I tested the hypothesis that male cortisol increases during periods of reproductive competition by comparing mean levels from individual males on days when 204 maximally swollen parous females were present with those from days when such females were absent. Samples from both periods were available for all 11 adult males. Paired comparisons were made using the Agresti-Pendergrast procedure. Three parous females were observed cycling at Kanyawara in 1998: Lia (AL), Ekisigi (EK), and Gombe (GO). Cycles from all three were included in this data set. Male sexual aggression and female stress I tested the hypothesis that female cortisol levels increase in response to male mating aggression by comparing mean levels within individual females between swelling and non-swelling periods. Comparisons were made with the Rust-Fligner test. Only a small number of females were cycling during 1998. Thus, there were sufficient samples from only two females to make this comparison: the parous female Lia (AL) and the nulliparous female Nile (NL). Few samples were available from each female; thus, I included both morning and afternoon samples in the analysis. Although diurnal fluctuations in cortisol production represent a potential confound with this method, the pooling was justified in this case by the even distribution of samples. Morning and afternoon samples were available from both females in both periods; therefore, neither high morning nor low evening samples were expected to bias the results by unduly influencing any particular category. 205 Results Cortisol and dominance Table 8.2 shows mean morning and afternoon cortisol levels for 11 adult males. High-ranking males maintained higher cortisol levels than low-ranking males. The correlation between male dominance level and mean morning cortisol was statistically significant (n = ll, r=.69, p=.02, Figure 8.1). The correlation between agonistic dominance rank and mean morning cortisol was similar, just failing to reach significance (n=l 1, r=.54, p=.087. Figure 8.2). The correlation between male dominance level and mean afternoon cortisol was also statistically significant (n = ll, r=.65, p=.03. Figure 8.3). Agonistic dominance rank and afternoon cortisol levels followed a similar pattern (n = ll, r=.68, p=.02. Figure 8.4). Male cortisol during reproductive competition Table 8.3 and Figure 8.5 show mean morning cortisol values for 11 adult males in reproductive and non-reproductive contexts. In parties containing maximally tumescent parous females, urinary cortisol levels averaged 562 pmol/mg creatinine. Mean morning cortisol levels in parties without such females were significantly lower, at 363 pmol/mg creatinine (Agresti-Pendergrast procedure, F=16.9, p=.002, n=l 1, Figure 8.6). 206 (D î C O § O I V 3 (D C 3 C O i % X— \ X— \ o o X— \ T—4 ooooT-HO\comwnr 1 ■ S o s G 4 — 1 U - i V 3 a f O O o 0 3 > > O o V 3 V 3 c 1 o O O > ~ » C O 0 3 o T 3 c 4 - 4 C D o o " O G g 1 bû C 0 3 4 - ' " G O ^ 4 O ( D G c : G C D H C D 03 S V 3 C D N C N V 3 OO C D C D 04 X i G 03 03 H t/3 I eu c V 3 I > (D V 3 I i 73 o 'S V 3 C D 03 C 207 600 500 - 9 L) 4 0 0 - < a § % « 2 300 - 200 High Medium Low Alpha M ale R ank Figure 8.1. Dominance level and mean morning cortisol in Kanyawara males. The relationship is statistically significant (r=.69, p=.02). For additional information, see Table 8.2. 208 600 *3 500- s | § 300 - 200 1.0 -0.8 -0.5 -0.3 0.0 0.3 0.5 0.8 1.0 M ale A ggression R ank Figure 8.2. Dominance rank and mean morning cortisol in Kanyawara males (r=.54, p=.087). For additional information, see Table 8.2. 209 400 300 - 200 - 100 - Alpha High Medium Low M ale R ank Figure 8.3. Dominance level and mean afternoon cortisol in Kanyawara males. The relationship is statistically significant (n=ll, r=.65, p=.03). For additional information, see Table 8.2. 210 400 300 - 2 0 0 - O iD 100 - -1.0 -0.8 -0.5 -0.3 0.0 0.3 0.5 0.8 1.0 M ale A ggression R ank Figure 8.4. Dominance rank and mean afternoon cortisol in Kanyawara males. The relationship is statistically significant (n=ll, r=.68, p=.02). For additional information, see Table 8.2. 211 I C L h g o c /3 H s O h G c /3 O c 3 = i 3 Ph 1 , G < U C O < u c /3 c /3 < u I ex H c /3 < D 13 g 1 G PX '% c /3 c3 13 § § % fX 0 c /3 1 U i ? u ,(D 1 — H m o o o \ V ) (N Tf vn < N C M o \ O s vn ^ v n m m o\ T - 4 ^ o m 'O o V T 3 T j- Tf 04 Tj- O O O m v n m v o m o m m o\ Tf \o \o en 1 — I m 04 m 1 — I ^ ^ 0 4 0 4 r o v n Tj- 0 4 '-H 0 4 ^ o o '-H Tj- 1 —I 0 4 0 4 0 4 0 4 T —t /—\ m m O O 0 \ o 0 4 m o o m m < ^ o o o v o o o v n v n m r ^ o 4 ' ^ o 4 vn T f , -4 o oo ov n o\ r ^ o\ \0 ' ^ ' ^ ' ^ ' ^ o 4 r o o 4 o 4 r o o 4 r o e n r n \0 _4 m çsi | s I '.G I I G O G c /3 B -g c c j V h c /3 < L > > g o O û G § I 1 1 1 i g a s g C /3 I i C u i I c c j c /3 (U G • § 0 î i o 'S 1 2 O g » " O i c /3 I C/3 î ex Ô « 1 1 I O s 212 1998 days with no maximally tumescent parous females 1998 days with maximally tumescent parous females Imoso Stout Johnny Badfoot Stocky Light Brown Makokn Big Brown 1000 M ean C ortisol P m ol/m g C reatinine Figure 8.5. Individual male cortisol levels in reproductive and non-reproductive contexts. The white bars indicate mean cortisol levels for males on observation days when no maximally swollen, parous females were present. The shaded bars represent mean male cortisol levels on days when maximally tumescent, parous females were present. Only morning samples are included. For additional information, see Table 8.3. 213 l i U 2 « U l i 700 600 500 400 300 200 100 0 1998 days with maximally tumescent parous females 1998 days with no tumescent females Figure 8.6. Means of mean male cortisol levels in reproductive and non-reproductive contexts. The difference is statistically significant using a paired, two- tailed test (Agresti-Pendergrast procedure, F=16.9, p=.002, n=ll). Only morning samples are included. Error bars represent the standard error of the mean. For additional information, see Figure 8.5 and Table 8.3. 214 Male aggression and female stress When the parous female Lia was not maintaining a sexual swelling, her urinary cortisol levels averaged 279 pmol/mg creatinine. Mean cortisol levels during periods of partial or maximal tumescence were more than twice as high, at 695 pmol/mg creatinine. This difference is statistically significant (Rust-Fligner test: Q=7.34, p=.007, ni=ll, n2=7. Table 8.4, Figure 8.7). The nulliparous female, Nile, on the other hand, showed no significant difference between a mean cortisol level of 23 8 pmol/mg creatinine during non-swelling periods versus 177 pmol/mg creatinine in swelling periods (Rust-Fligner test: Q=0.39, p=0.53, ni=6, n2=20. Table 8.4, Figure 8.7). Female Swelling Mean Cortisol SE Samples AL None 279 102 11(13) AL Partial or Full 695 - 90 7(8) NL None 238 50 6(6) NL Partial or Full 177 26 20(22) Table 8.4. Mean cortisol values for two females in swelling and non swelling contexts. Lia (AL) was a parous female, and Nile (NL) was nulliparous. Both morning and afternoon samples are included. Sample sizes are the total number of daily cortisol values, followed by the number of urine samples used to calculate those values (in parentheses). For additional information see Figure 8.7. 215 No sexual swelling Partial or full sexual swelling II I f P L h Lia (Parous female) Nile (Nulliparous female) Figure 8.7. Female cortisol levels in swelling and non swelling contexts. For Lia, cortisol levels were significantly higher during partial and m axim al tumescence (Rust-Fligner test Q=7.34, p=.007). For Nile there was no difference between the two periods (Rust-Fligner test Q=.39, p=.53). 216 Discussion and Conclusions In contrast to T, the association between male rank and cortisol did not appear to differ by time of day. High-ranking males exhibited higher cortisol levels in both morning and afternoon samples. This finding is inconsistent with Sapolsky’s (1983) studies of wild baboons, because the dominance hierarchy at Kanyawara was relatively stable during the sampling period (see chapter three). Sapolsky’s model predicts that high rank should generally be associated with low cortisol in stable hierarchies, because rank is presumed to confer psychological benefits associated with predictability and control. Such is the case in numerous captive studies, where social subordination is associated with high levels of circulating glucocorticoids (reviewed in Creel and Sands in press). As indicated in Table 8.1, however, most field studies of social vertebrates have reported a correlation between rank and glucocorticoid secretion similar to the one described here. What can account for this apparent difference? One possibility, suggested by Creel et al. (1996), is that patterns of aggression differ between wild and captive populations. Specifically, subordinates in the wild are often able to avoid dominant individuals, and thus may be infrequently involved in agonistic interactions. Such avoidance is not possible in captivity, which might explain the tendency for higher levels of stress hormones in captive subordinates. In addition, dominant individuals might be expected to exhibit higher levels of circulating 217 cortisol in the wild because they are frequently involved in aggressive interactions as they assert their status (Creel et al. 1996). By this logic, dominant individuals might be expected to exhibit higher levels of circulating glucocorticoids in any species in which they are engaged in aggressive interactions more often than subordinates (Creel 2ÜÜ1). f his idea fits well with both the chimpanzee data presented here and Sapolsky’s baboon data. As shown in chapter three, high-ranking chimpanzee males exhibit significantly higher rates of aggression than low-ranking males, even in stable dominance hierarchies. Similarly, the relationship between cortisol and rank in baboons parallels the relationship between aggression and rank. In unstable hierarchies dominant males are more aggressive and have higher cortisol levels; in stable hierarchies subordinate males are more aggressive and have higher cortisol levels. Thus, aggression appears to be a prime candidate for the behavior that mediates the relationship between stress and dominance. Unfortunately, compelling data from social carnivores do not fit with this idea. Among wild dogs, for example, dominant individuals are not involved in aggressive interactions more frequently than subordinates are except during the mating season. Dominants maintain higher basal levels of circulating glucocorticoids throughout the year, however (Creel et al. 1997). And in free-ranging wolves dominant individuals are no more aggressive than subordinates, yet they also maintain higher levels of glucocorticoids throughout the year (Creel and Sands in press). 218 Consequently, despite the fact that dominance rank appears to be a reasonable predictor of glucocorticoid levels across a variety of species, the underlying behaviors that are responsible for this relationship remain a mystery (Creel and Sands in press). One possibility is that these will differ by species. Aggression appears to be a critical variable for primates, but not for carnivores. Additional ûeld studies may help to clarify this question. The observation that increases in male cortisol were associated with periods of reproductive competition in male chimpanzees is unsurprising, given cortisol’s well- known role in short-term energy mobilization. More interesting is the apparent difference between nulliparous and parous female chimpanzees in their cortisol response to reproductive competition. This study only examined two females, so caution is warranted in interpreting the results. Nevertheless, the differences reported here are suggestive. Parous, estrous females are frequent targets of male aggression, particularly when high-ranking males are attempting to maintain exclusive mating access to them. Being mate-guarded is thus likely to be extremely stressful for females. Consistent with this idea, the parous female Lia showed dramatic increases in urinary cortisol levels during maximal tumescence. Lia was intensively mate-guarded on many of her maximal swelling days, and the aggression that she received from males on these days increased substantially. The nulliparous female Nile, on the other hand, did not exhibit increases in urinary cortisol when she was maximally swollen. 219 Nor did she receive high rates of aggression or male mate guarding. These data represent an important first step in documenting the physiological costs imposed on females by male sexual aggression. Additional sample collection is currently underway at both Kanyawara and in Gombe, that will allow for a more detailed picture of these costs. 220 B ib l io g r a p h y Abbott DH (1993) Social conflict and reproductive suppression in marmoset and tamarin monkeys. In WA Mason and SP Mendoza (eds.): Primate Social Conflict. Albany: SUN Y Press. Agresti A (1992) A survey of exact inference in multidimensional tables. Statistical Science 7:131-152. Agresti A, and Pendergrast J (1986) Comparing mean ranks for repeated measures data. Communications in Statistics - Theory and Methods 15:1417-1433. Alberts SC, Sapolsky RM, and Altmann J (1992) Behavioral, endocrine, and immunological correlates of immigration by an aggressive male into a natural primate group. Hormones and Behavior 26:167-178. Altmann J (1974) Observational study of behavior: sampling methods. Behaviour 49:227-267. Altmann J, Alberts SC, Haines SA, Dubach J, Muruthi P, Coote T, Geffen E, Cheesman DJ, Mututua RS, Saiyalel SN, Wayne RK, Lacy RC, and Bruford MW (1996) Behavior predicts genetic structure in a wild primate group. Proceedings o f the National Academy o f Sciences, U.S.A. 93:5797-5801. Altmann SA (1962) A field study of the sociobiology of rhesus monkeys, Macaca mulatto. Annals o f the New York Academy o f Sciences 102:338-435. Archer J (1988) The Behavioural Biology o f Aggression. Cambridge: Cambridge University Press. Archer J (1991) The influence of testosterone on human aggression. British Journal o f Psychology 82:1-28. Arnold W, and Dittami J (1997) Reproductive suppression in male alpine marmots. Animal Behaviour 53:53-66. Bagatell C, Heiman J, Rivier J, and Bremner W (1994) Effects of endogenous testosterone and estradiol on sexual behavior in normal young men. Journal o f Clinical Endocrinology and Metabolism 75:1503-1507. 221 Bahr NI, Palme R, Mohle U, Hodges JK, and Heistermann M (2000) Comparative aspects of the metabolism and excretion of cortisol in three individual nonhuman primates. General and Comparative Endocrinology 117:427-438. Bailey RC, Jenike MR, Bentley GR, Harrigan AM, and Ellison PT (1992) The ecology of birth seasonality among agriculturalists in central Africa. Journal o f Biosocial Science 24:393-412. Balthazart J (1976) Daily variations of behavioural activities and plasma testosterone levels in the domestic duck. Journal o f Zoology 180:155-173. Balthazart J (1983) Hormonal correlates of behavior. In DS Earner, JR King and KC Parkes (eds.): Avian Biology. New York: Academic Press. Baum MJ (1993) Neuroendocrinology of sexual behavior in the male. In JB Becker, SM Breedlove and D Crews (eds.): Behavioral Endocrinology. Cambridge: MIT Press. Beisel WR, Cos JJ, Horton R, Chao PY, and Forsham PH (1964) Physiology of urinary cortisol excretion. Journal o f Clinical Endocrinology and Metabolism 24:887-893. Beletsky LD, Gori DF, Freeman S, and Wingfield JC (1995) Testosterone and polygyny in birds. In DM Power (ed.): Current Ornithology. New York: Plenum Press. Bentley GR (1999) Aping our ancestors: comparative aspects of reproductive ecology. Evolutionary Anthropology 7:175-185. Bentley GR, Harrigan AM, Campbell B, and Ellison PT (1993) Seasonal effects on salivary testosterone levels among Lese males of the Ituri Forest, Zaire. American Journal o f Human Biology 5:711-717. Berard JD, Numberg P, Epplen JT, and Schmidtke J (1993) Male rank, reproductive behavior, and reproductive success in free ranging rhesus macaques. Primates 34:481-489. Bercovitch FB (1988) Coalitions, cooperation, and reproductive tactics among adult savanna baboons. Animal Behaviour 34:1198-1209. Bercovitch FB (1989) Body size, sperm competition, and determinants of reproductive success in male savanna baboons. Evolution 43:1507-1521. 222 Bercovitch FB (1993) Dominance rank and reproductive maturation in male rhesus macaques {Macaca mulatto). Journal o f Reproduction and Fertility 99:113-120. Bercovitch FB, and Clarke SA (1995) Dominance rank, cortisol concentrations, and reproductive maturation in male rhesus macaques. Physiology and Behavior 58:215- 221. Bercovitch FB, and Numberg P (1997) Genetic determination of paternity and variation in male reproductive success in two populations of rhesus macaques. Electrophoresis 18:1701-1705. Bernstein IS, Gordon TP, and Rose RM (1983) The interaction of hormones, behavior, and social context in nonhuman primates. In BB Svare (ed.): Hormones and Aggressive Behavior. New York: Plenum Press. Biquand S, Boug A, Biquand-Guyot V, and Gautier J-P (1994) Management of commensal baboons in Saudi Arabia. Revue dEcologie 49:213-222. Birkhead TR (2000) Promiscuity: An Evolutionary History o f Sperm Competition. Cambridge, MA: Harvard University Press. Boehm C (1999) Hierarchy in the Forest: The Evolution o f Egalitarian Behavior. Cambridge, MA: Harvard University Press. te Boekhorst IJA, Schurmann CL, and Sugardjito J (1990) Residential status and seasonal movements of wild orang-utans in the Gunung Leuser Reserve (Sumatra, Indonesia). Animal Behaviour 39:1098-1109. Boesch C, and Boesch-Achermann H (2000) The Chimpanzees o f the Tai Forest: Behavioral Ecology and Evolution. Oxford: Oxford University Press. Booth A, and Dabbs J (1993) Testosterone and men's marriages. Social Forces 72:463-477. Bouissou M (1983) Androgens, aggressive behaviour and social relationships in higher mammals. Hormone Research 18:43-61. Bradley BJ, Robbins MM, Williamson EA, Boesch C, and Vigilant L (2001) Assessment of reproductive strategies in wild mountain gorillas {Gorilla gorilla beringei) using DNA from feces. American Journal o f Physical Anthropology, Supplement 32:43. 223 Bribiescas RG (1996) Testosterone levels among Ache hunter/gatherer men: a functional interpretation of population variation among adult males. Human Nature 7:163-188. Bribiescas RG (1997) Testosterone as a proximate determinant of somatic energy allocation in human males: evidence from Ache men of eastern Paraguay. PhD, Harvard University, Cambridge. Brown RE, Murdoch I, Murphy PR, and Moger WH (1995) Hormonal responses of male gerbils to stimuli from their mate and pups. Hormones and Behavior 29:474- 491. Buchanan KL (2000) Stress and the evolution of condition-dependent signals. Trends in Ecology and Evolution 15:156-160. Bygott JD (1974) Agonistic behaviour and dominance in wild chimpanzees. Ph.D., University of Cambridge, Cambridge. Bygott JD (1979) Agonistic behavior, dominance and social structure in wild chimpanzees of the Gombe National Park. In D Hamburg and ER McCown (eds.): The Great Apes. Menlo Park, CA: Benjamin Cummings. Cahill DJ, Wardle PG, Harlow CR, and Hull MGR (1998) Onset of the preovulatory luteinizing hormone surge: diurnal timing and critical follicular prerequisites. Fertility and Sterility 70:56-59. Campbell BC (1990) Menstrual profile of salivary testosterone and its implications for changes in coital frequency across the menstrual cycle. Ph.D., Harvard University, Cambridge, MA. Campbell BC, and Leslie PW (1995) Reproductive ecology of human males. Yearbook o f Physical Anthropology 38:1-26. Carter CS (1993) Hormonal influences on human sexual behavior. In JB Becker, SM Breedlove and D Crews (eds.): Behavioral Endocrinology. Cambridge: MIT Press. Cavigelli SA (1999) Behavioural patterns associated with faecal cortisol levels in free-ranging female ring-tailed lemurs, Lemur catta. Animal Behaviour 57:935-944. Cavigelli SA, and Pereira ME (2000) Mating season aggression and fecal testosterone levels in male ring-tailed lemurs {Lemur catta). Hormones and Behavior 37:246-255. 224 Chapman CA, White FJ, and Wrangham RW (1994) Party size in chimpanzees and bonobos: a réévaluation of theory based on two similarly forested sites. In RW Wrangham, WC McGrew, FBM de Waal and PG Heltne (eds.): Chimpanzee Cultures. Cambridge, MA: Harvard University Press. Chapman CA, Wrangham R, and Chapman LJ (1995) Ecological constraints on group size: an analysis of spider monkey and chimpanzee subgroups. Behavioral Ecology and Sociobiology 36:59-70. Chapman CA, and Wrangham RW (1993) Range use of the forest chimpanzees of Kibale: implications for the understanding of chimpanzee social organization. American Journal o f Primatology 31:263-273. Christiansen KH (1991) Serum and saliva sex hormone levels in !Kung San men. American Journal o f Physical Anthropology 86:37-44. Chrousos GP, and Gold PW (1992) The concepts of stress and stress system disorders: Overview of physical and behavioral homeostasis. Journal o f the American Medical Association 267:1244-1251. Clark Arcadi A, and Wrangham RW (1999) Infanticide in chimpanzees: review of cases and a new within-group observation from the Kanyawara study group in Kibale National Park. Primates 40:337-351. Clark G, and Birch HG (1945) Hormonal modifications of social behavior. I. The effect of sex-hormone administration on the social status of a male-castrate chimpanzee. Psychosomatic Medicine 7:321-329. Clutton-Brock TH, and Vincent ACJ (1991) Sexual selection and potential reproductive rates of males and females. Nature 351:58-60. Coe CL, Mendoza SP, and Levine S (1979) Social status constrains the stress response in the squirrel monkey. Physiology and Behavior 23:633-638. Conklin-Brittain NL, Wrangham RW, and Hunt KD (1998) Dietary response of chimpanzees and cercopithecines to seasonal variation in fruit abundance. II. Macronutrients. International Journal o f Primatology 19:971-998. Constable JL, Ashley MV, Goodall J, and Pusey AE (2001) Noninvasive paternity assignment in Gombe chimpanzees. Molecular Ecology 10:1279-1300. Cook B, and Beastall GH (1987) Measurernent of steroid hormone concentrations in blood, urine and tissues. In B Green and RE Leake (eds.): Steroid Hormones. Oxford: IRL Press. 225 Cowlishaw G, and Dunbar RIM (1991) Dominance rank and mating success in male primates. Animal Behaviour 41:1045-1056. Creel S (2001) Social dominance and stress hormones. Trends in Ecology and Evolution 16:491-497. Creel S, Creel NM, Mills MGL, and Monfort SL (1997) Rank and reproduction in cooperatively breeding African wild dogs: behavioral and endocrine correlates. Behavioral Ecology. Creel S, Creel NM, and Monfort SL (1996) Social stress and dominance. Nature 379:212. Creel S, and Sands JL (in press) Is social stress a consequence of subordination or a cost of dominance? In P Tyack and F de Waal (eds.): Animal Social Complexity. Cambridge, MA: Harvard University Press. Creel SR, Wildt DE, and Monfort SL (1993) Androgens, aggression and reproduction in wild dwarf mongooses: a test of the challenge hypothesis. American Naturalist 141:816-825. Czekala NM, Lance VA, and Sutherland-Smith M (1994) Diurnal urinary corticoid excretion in the human and gorilla. American Journal o f Primatology 34:29-34. Dabbs JM (1990) Salivary testosterone measurements: reliability across hours, days, and weeks. Physiology and Behavior 48:83-86. De Ridder E, Pinxten R, and Eens M (2000) Experimental evidence of a testosterone-induced shift from paternal to mating behaviour in a facultatively polygynous songbird. Behavioral Ecology and Sociobiology 49:24-30. Delville Y, Mansour KM, and Ferris CF (1994) Testosterone facilitates aggression by modulating vasopressin receptors in the hypothalamus. Physiology and Behavior 60:25-29. Dennen Jvd (1995) The Origin o f War: The Evolution o f a Male-Coalitional Reproductive Strategy. Groningen, Netherlands: Origin Press. Dessypris A, Kuoppasalmi K, and Adlecreutz H (1976) Plasma cortisol, testosterone, androstenedione and luteinizing hormone (LH) in a non-competitive marathon run. Journal o f Steroid Biochemistry 7:33-37. 226 Dixson AF (1980) Androgens and aggressive behavior in primates: a review. Aggressive Behavior 6:37-67. Dixson AF (1998) Primate Sexuality: Comparative Studies o f the Prosimians, Monkeys, Apes, and Human Beings. New York: Oxford University Press. Dixson AF, Bossi T, and Wickings EJ (1993) Male dominance and genetically determined reproductive success in the mandrill. Primates 34:525-532. Eaton GG, and Resko JA (1974) Plasma testosterone and dominance in a Japanese macaque {Macaca fuscata) troop compared with repeated measures of testosterone in laboratory males. Hormones and Behavior 5:251-259. Eberhart JA, Keverne EB, and Meller RE (1980) Social influences on plasma testosterone leves in male talopoin monkeys. Hormones and Behavior 14:247-266. Edwards RG (1981) Test-tube babies, 1981. Nature 293:253-256. Elias M (1981) Serum cortisol, testosterone, and testosterone-binding globulin responses to competitive fighting in human males. Aggressive Behavior 7:215-224. Ellison PT (1988) Human salivary steroids: methodological considerations and applications in physical anthropology. Yearbook o f physical anthropology 31:115- 142. Ellison PT (1990) Human ovarian function and reproductive ecology: new hypotheses. American Anthropologist 92:933-952. Ellison PT (2001) On Fertile Ground: A Natural History o f Human Reproduction. Cambridge, MA: Harvard University Press. Ellison PT, and Lager C (1985) Exercise-induced menstrual disorders. New England Journal o f Medicine 313:825-826. Ellison PT, and Lager C (1986) Moderate recreational running is associated with lowered salivary progesterone profiles in women. American Journal o f Obstetrics and Gynecology 154:1000-1003. Ellison PT, and Panter-Brick C (1996) Salivary testosterone levels among Tamang and Kami males of central Nepal. Human Biology 68:955-965. Ellison PT, Panter-Brick C, Lipson SF, and O'Rourke MT (1993) The ecological context of human ovarian function. Human Reproduction 8:2248-2258. 227 Ellison PT, Peacock NR, and Lager C (1989) Ecology and ovarian function among Lese women of the Ituri Forest, Zaire. American Journal o f Physical Anthropology 78:519-526. Emlen ST, and Oring LW (1977) Ecology, sexual selection, and the evolution of mating systems. Science 197:215-223. Erb RE (1970) Urinary creatinine as an index compound for estimating rate of excretion of steroids in the domestic sow. Journal o f Animal Science 30:79-85. Evans MR, Goldsmith AR, and Norris SRA (2000) The effects of testosterone on antibody production and plumage coloration in male house sparrows {Passer domesticus). Behavioral Ecology and Sociobiology 47:156-163. Falkenstein E, Tillmann H-C, Christ M, Feuring M, and Wehling M (2000) Multiple actions of steroid hormones - a focus on rapid, nongenomic effects. Pharmacological Reviews 52:513-555. Fawcett K, and Muhumuza G (2000) Death of a wild chimpanzee community member: Possible outcome of intense sexual competition. American Journal o f Primatology 51:243-247. Fedigan LM (1983) Dominance and reproductive success in primates. Yearbook o f Physical Anthropology 26:91-129. Fedigan LM (1992) Primate Paradigms: Sex Roles and Social Bonds. Chicago: University of Chicago Press. Flinn MV (1988) Mate guarding in a Caribbean village. Ethology and Sociobiology 9:1-28. Flinn MV, and England BG (1997) Social economics of childhood glucocorticoid stress response and health. American Journal o f Physical Anthropology 102:33-53. Fossey D (1984) Infanticide in mountain gorillas {Gorilla gorilla beringei) with comparative notes on chimpanzees. In G Hausfater and SB Hrdy (eds.): Infanticide: Comparative and Evolutionary Perspectives. Hawthorne, NY: Aldine. Gagneux P, Boesch C, and Woodruff DS (1999) Female reproductive strategies, paternity and community structure in wild West African chimpanzees. Animal Behaviour 57:19-32. Gagneux P, Woodruff DS, and Boesch C (1997) Furtive mating in female chimpanzees. Nature 387:358-359. 228 Galdikas BMP (1979) Orangutan adaptation at Tanjung Puting Reserve: mating and ecology. In DA Hamburg and ER McCown (eds.): The Great Apes. Menlo Park, CA: Benj amin/ Cummings. Galdikas BMP (1981) Orangutan reproduction in the wild. In CE Graham (ed.): Reproductive Biology o f the Great Apes. New York: Academic Press. Genuth SM (1993) The endocrine system. In RM Berne and MN Levy (eds.): Physiology. St. Louis: Mosby Year Book. Gerloff U, Hartung B, Truth B, Hohmann G, and Tautz D (1999) Intracommunity relationships, dispersal pattern and paternity success in a wild living community of bonobos {Pan paniscus) determined from DNA analysis of faecal samples. Proceedings o f the Royal Society o f London, Series B 266:1189-1195. Ghiglieri MP (1984) The Chimpanzees o f Kibale Forest: A Field Study o f Ecology and Social Structure. New York: Columbia University Press. Gibb AL (1996) Dimorphism, aggression, and sex: urinary androgens in common chimpanzees {Pan troglodytes) and bonobos {Pan paniscus). BA, Harvard University, Cambridge. Goldman BD (1999) The circadian timing system and reproduction in mammals. Steroids 64:679-685. Gomendio M, Harcourt AH, and Roldan ERS (1998) Sperm competition in mammals. In TR Birkhead and AP Moller (eds.): Sperm Competition and Sexual Selection. New York: Academic Press. Goodall J (1986) The Chimpanzees o f Gombe: Patterns o f Behavior. Cambridge, MA: Harvard University Press. Goodall J (1992) Unusual violence in the overthrow of an alpha male chimpanzee at Gombe. In T Nishida, WC McGrew, P Marier, M Pickford and FBM de Waal (eds.): Topics in Primatology. Tokyo: University of Tokyo Press. Goodall J, Bandora A, Bergmann E, Busse C, Matama H, Mpongo E, Pierce A, and Riss D (1979) Inter-community interactions in the chimpanzee population of the Gombe National Park. In DA Hamburg and ER McCown (eds.): The Great Apes. Menlo Park, CA: Benjamin/Cummings. 229 Gordon TP, Rose RM, and Bernstein IS (1976) Seasonal rhythm in plasma testosterone levels in the rhesus monkey: A three year study. Hormones and Behavior 7:229-243. Gordon TP, Rose RM, Grade CL, and Bernstein IS (1979) Effects of increased testosterone secretion on the behavior of adult male rhesus living in a social group. Folia Primatologica 32:149-160. Goy RW, and Resko JA (1972) Gonadal hormones and behavior of normal and pseudohermaphroditic female primates. Recent Progress in Hormone Research 28:707-733. Graham CE (1981) Menstrual cycle of the great apes. In CE Graham (ed.): Reproductive Biology o f the Great Apes. New York: Academic Press. Green R, Whalen RE, Rutley B, and Battle C (1972) Dominance hierarchy in squirrel monkeys (Saimiri sciureus): Role of the gonads and androgen on genital display and feeding order. Folia Primatologica 18:185-195. Grossman CJ (1984) Regulation of the immune system by sex steroid. Endocrine Reviews 5:435-454. Gust DA, McCaster T, Gordon TP, Gergits W, Casna N, and McClure HM (1998) Paternity in sooty mangabeys. International Journal o f Primatology 19:83-94. Hall KRL, and DeVore I (1965) Baboon social behavior. In I DeVore (ed.): Primate Behavior: Field Studies ofMonleys and Apes. New York: Holt, Rinehart and Winston. Harcourt AH (1978) Strategies of emigration and transfer by primates, with particular reference to gorillas. Zeitschrift fu r Tierpsychology 408:401-420. Harcourt AH, Fossey D, and Sabater-Pi J (1981) Demography of Gorilla gorilla. Journal o f Zoology 195:215-233. Hasegawa T, and Hiraiwa-Hasegawa M (1990) Sperm competition and mating behavior. In T Nishida (ed.): The Chimpanzees o f the Mahale Mountains: Sexual and Life History Strategies. Tokyo: University of Tokyo Press. Hasselquist D, Marsh JA, Sherman PW, and Wingfield JC (1999) Is avian humoral immunocompetence suppressed by testosterone? Behavioral Ecology and Sociobiology 45. 230 Hausfater G (1975) Dominance and reproduction in baboons {Papio cynocephalus): a quantitative analysis. Contributions to Primatology 7:1-150. Hayaki H, Huffman MA, and Nishida T (1989) Dominance among male chimpanzees in the Mahale Mountains National Park, Tanzania: a preliminary study. Primates 30:187-197. Hegner RE, and Wingfield JC (1987) Effects of experimental manipulation of testosterone levels on parental investment and breeding success in male house sparrows. Auk 104:462-469. Hellhammer D, Hubert W, and Schurmeyer T (1985) Changes in saliva testosterone after psychological stimulation in men. Psychoneuroendocrinology 10:77-81. Herndon JG (1983) Seasonal breeding in rhesus monkeys: influence of the behavioral environment. American Journal o f Primatology 5:197-204. Higley JD, Mehlman PT, Poland RE, Taub DM, Vickers J, Suomi SJ, and Linnoila M (1996) CSF testosterone and 5-HIAA correlate with different types of aggressive behaviors. Biological Psychiatry 40:1067-1082. Isabirye-Basuta G (1989) Feeding ecology of chimpanzees in the Kibale Forest, Uganda. In PG Heltne and LA Marquardt (eds.): Understanding Chimpanzees. Cambridge, MA: Harvard University Press. James VHT, Tunbridge D, Wilson GA, Hutton JD, Jacobs HS, Goodall AB, Murray MAF, and Rippon AE (1978) Central control of steroid hormone secretion. Journal o f Steroid Biochemistry 9:429-436. Jameson KA, Appleby MC, and Freeman EC (1999) Finding an appropriate order for a hierarchy based on probabilistic dominance. Animal Behaviour 57:991-998. Jasienska G, and Ellison PT (1993) Heavy workload impairs ovarian function in Polish peasant women. American Journal o f Physical Anthropology, Supplement 16:17-18. Joslyn WD (1973) Androgen-induced social dominance in infant female rhesus monkeys. Journal o f Child Psychology and Psychiatry 14:137-145. Kano T (1992) The Last Ape: Pygmy Chimpanzee Behavior and Ecology. Stanford, CA: Stanford University Press. 231 Kano T (1996) Male rank order and copulation rate in a unit-group of bonobos at Wamba, Zaire. In WC McGrew, LF Marchant and T Nishida (eds.): Great Ape Societies. Cambridge: Cambridge University Press. Kepner JL, and Robinson DH (1988) Nonparametric methods for detecting treatment effects in repeated-measures designs. Journal o f Statistical Association 83:456-461. Keveme EB, Meller RE, and Eberhart A (1982) Dominance and subordination: concepts or physiological states? In AB Chiarelli and RS Corruccini (eds.): Advanced Views in Primate Biology. New York: Springer-Verlag. Kingsley S (1982) Causes of non-breeding and the development of secondary sexual characteristics in the male orangutan: a hormonal study. In L de Boer (ed.): The Orangutan: Its Biology and Conservation. The Hague: Junk Publishers. Kingsley S (1988) Physiological development of male orangutans and gorillas. In J Schwartz (ed.): Orangutan Biology. Oxford: Oxford University Press. Klibanski A, Beitins IZ, Badger T, Little R, and McArthur JW (1981) Reproductive function during fasting in men. Journal o f Clinical Endocrinology and Metabolism 53:258-263. Knott CD (1998) Changes in orangutan caloric intake, energy balance, and ketones in response to fluctuating food availability. International Journal o f Primatology 19:1061-1079. Knott CD (1999) Reproductive, Physiological and Behavioral Responses of Orangutans in Borneo to Fluctuations in Food Availability. PhD, Harvard University, Cambridge. Kraemer HC, Horvat JR, Doering C, and McGinnis PR (1982) Male chimpanzee development focusing on adolescence: Integration of behavioral with physiological changes. Primates 23:393-405. Kummer H (1968) Social Organization o f Hamadryas Baboons. Chicago: University of Chicago Press. Lincoln GA, Guinness F, and Short RV (1972) The way in which testosterone controls the social and sexual behavior of the red deer stag {Cervus elaphus). Hormones and Behavior 3:375-396. Lincoln GA, and Ratnasooriya WD (1996) Testosterone secretion, musth behaviour and social dominance in captive male Asian elephants living near the equator. Journal o f Reproduction and Fertility 108:107-113. 232 Liotta AS, and Krieger DT (1990) ACTH and related peptides. In E-E Baulieu and PA Kelly (eds.): Hormones: From Molecules to Disease. New York: Chapman and Hall. Macrides F, Bartke A, and Dalterio S (1975) Strange females increase plasma testosterone levels in male mice. Science 189. Maggioncalda AN, Czekala NM, and Sapolsky RM (2000) Growth hormone and thyroid stimulating hormone concentrations in captive male orangutans: Implications for understanding developmental arrest. American Journal o f Primatology 50:67-76. Maggioncalda AN, Sapolsky RM, and Czekala NM (1999) Reproductive hormone profiles in captive male orangutans: Implications for understanding developmental arrest. American Journal o f Physical Anthropology 109:19-32. Manson JH, and Wrangham RW (1991) Intergroup aggression in chimpanzees and humans. Current Anthropology 32:369-390. Maple T (1980) Orangutan Behavior. New York: van Nostrand Reinhold. Marson J, Meuris S, Cooper RW, and Jouannet P (1991) Puberty in the male chimpanzee: Time-related variations in luteinizing hormone, follicle-stimulating hormone, and testosterone. Biology o f Reproduction 44:456-460. Martensz ND, Vellucci SV, Fuller LM, Everitt BJ, Keveme EB, and Herbert J (1987) Relation between aggressive behaviour and circadian rhythms in cortisol and testosterone in social groups of talapoin monkeys. Journal o f Endocrinology 115:107-120. Martin DE, Swenson RB, and Collins DC (1977) Correlation of serum testosterone levels with age in male chimpanzees. Steroids 29:471-481. Matsumoto-Oda A (1999) Female choice in the opportunistic mating of wild chimpanzees {Pan troglodytes schweinfurthii) at Mahale. Behavioral Ecology and Sociobiology 46:258-266. Maynard Smith J, and Price GR (1973) The logic of animal conflict. Nature 246:15- 18. Mazur A, and Booth A (1998) Testosterone and dominance in men. Behavioral and Brain Sciences 21:353-397. 233 Mazur A, Booth A, and Dabbs JM (1992) Testosterone and chess competition. Social Psychology Quarterly 55:70-77. Mazur A, and Lamb TA (1980) Testosterone, status, and mood in human males. Hormones and Behavior 14:236-246. Mazur A, and Michalek J (1998) Marriage, divorce, and male testosterone. Social Forces 77:315-330. McCormack SA (1971) Plasma testosterone concentration and binding in the chimpanzee: Effect of age. Endocrinology 89:1171-1177. Mech DL, Adams LG, Meier TJ, Burch JW, and Dale BW (1998) The Wolves o f Denali. Minneapolis: University of Minnesota Press. Mehlman PT, Higley JD, Fernald BJ, Sallee FR, Suomi SJ, and Linnoila M (1997) CSF 5-HIAA, testosterone, and sociosexual behaviors in free-ranging male rhesus macaques in the mating season. Psychiatry Research 72:89-102. Mendoza SP, Coe CL, Lowe EL, and Levine S (1979) The physiological response to group formation in adult male squirrel monkeys. Psychoneuroendocrinology 3:221- 229. Michael RP, and Zumpe D (1978) Annual cycles of aggression and plasma testosterone in captive male rhesus monkeys. Psychoneuroendocrinology 3:217-220. Michael RP, and Zumpe D (1981) Relation between the seasonal changes in aggression, plasma testosterone and the photoperiod in male rhesus monkeys. Psychoneuroendocrinology 6:145-158. Michael RP, and Zumpe D (1993) A review of hormonal factors influencing the sexual and aggressive behavior of macaques. American Journal o f Primatology 30:213-241. Milgrom E (1990) Steroid hormones. In E-E Baulieu and PA Kelly (eds.): Hormones: From Molecules to Disease. New York: Chapman and Hall. Miller MW, Hobbs NT, and Sousa MC (1991) Detecting stress responses in Rocky Mountain bighorn sheep {Ovis canadensis canadensis): Reliability of cortisol concentrations in urine and feces. Canadian Journal o f Zoology 69:15-24. Milton K (1985) Mating patterns of woolly spider monkeys, Brachyteles arachnoïdes: Implications for female choice. Behavioral Ecology and Sociobiology 17:53-59. 234 Mitani JC (1985) Mating behaviour of male orangutans in the Kutai Game Reserve, Indonesia. Animal Behaviour 33:392-402. Mitani JC (1985) Sexual selection and adult male orangutan long calls. Animal Behaviour 33:272-283. Mitani JC, Gros-Louis J, and Richards AF (1996) Sexual dimorphism, the operational sex ratio, and the intensity of male competition in polygynous primates. The American Naturalist 147:966-980. Mitani JC, and Watts DP (2001) Why do chimpanzeees hunt and share meat? Animal Behaviour 61:915-924. Monaghan EP, and Glickman SE (1992) Hormones and aggressive behavior. In JB Becker, SM Breedlove and D Crews (eds.): Behavioral Endocrinology. Cambridge: MIT Press. Moore FL, and Evans SJ (1999) Steroid hormones use non-genomic mechanisms to control brain functions and behaviors: A review of evidence. Brain, Behavior and Evolution 54:41-50. Moore MC (1986) Circulating steroid hormones during rapid aggressive responses of territorial male mountain spiny lizards, Sceloporus jarrovi. Hormones and Behavior 21:511-521. Morin LP, and Dark J (1992) Hormones and biological rhythms. In JB Becker, SM Breedlove and D Crews (eds.): Behavioral Endocrinology. Cambridge: MIT Press. Morin PA, Moore JJ, Chakraborty R, Jin L, Goodall J, and Woodruff DS (1994) Kin selection, social structure, gene flow, and the evolution of chimpanzees. Science 265:1193-1201. Muller MN (2000) The knuckle-walking wounded. Natural History 109(8):44-48. Nadler RD, Wallis J, Roth-Meyer C, Cooper RW, and Baulieu EE (1987) Hormones and behavior of prepubertal and peripubertal chimpanzees. Hormones and Behavior 21:118-131. Nelson RJ (1995) An Introduction to Behavioral Endocrinology. Sunderland, Massachusetts: Sinauer. Nishida T (1968) The social group of wild chimpanzees in the Mahali Mountains. Primates 9:167-224. 235 Nishida T (1979) The social structure of chimpanzees of the Mahale Mountains. In D Hamburg and ER McCown (eds.): The Great Apes. Menlo Park, CA: Benjamin Cummings. Nishida T (1983) Alpha status and agonistic alliance in wild chimpanzees {Pan troglodytes schweinfurthii). Primates 24:318-336. Nishida T (1989) Social interactions between resident and immigrant female chimpanzees. In PG Heltne and LA Marquardt (eds.): Understanding Chimpanzees. Cambridge, MA: Harvard University Press. Nishida T (1997) Sexual behavior of adult male chimpanzees of the Mahale Mountains National Park, Tanzania. Primates 38:379-398. Nishida T, Hasegawa T, Hayaki H, Takahata Y, and Uehara S (1992) Meat sharing as a coalition strategy by an alpha male chimpanzee? In T Nishida, W McGrew, P Marier, M Pickford and F de Waal (eds.): Topics in Primatology. Volume 1. Human Origins. Tokyo: Tokyo University Press. Nishida T, Hiraiwa-Hasegawa M, Hasegawa T, and Takahata Y (1985) Group extinction and female transfer in wild chimpanzees in the Mahale National Park, Tanzania. Zeitschrift fur Tierpsychologie 67:284-301. Nishida T, and Hosaka K (1996) Coalition strategies among adult male chimpanzees of the Mahale Mountains, Tanzania. In WC McGrew, LF Marchant and T Nishida (eds.): Great Ape Societies. Cambridge: Cambridge University Press. Nishida T, Hosaka K, Nakamura M, and Hamia M (1995) A within-group gang attack on a young adult male chimpanzee: Ostracism of an ill-mannered member? Primates 36:207-211. Nishida T, Kano T, Goodall J, McGrew WC, and Nakamura M (1999) Ethogram and ethnography of Mahale chimpanzees. Anthropological Science 107:141-188. O'Leary A (1990) Stress, emotion, and human immune function. Psychological Bulletin 108:363-382. Panter-Brick C, Lotstein DS, and Ellison PT (1994) Seasonality of reproductive function and weight loss in rural Nepali women. Human Reproduction 8:684-690. Pereira ME, and Weiss ML (1991) Female mate choice, male migration, and the threat of infanticide in ringtailed lemurs. Behavioral Ecology and Sociobiology 28:141-152. 236 Polanski FF, and Lamb EJ (1988) Do the results of semen analysis predict future fertility? Fertility and Sterility 49:1059-1065. Pollard TM (1995) Use of cortisol as a stress marker: practical and theoretical problems. American Journal o f Human Biology 7:265-274. Power M (1991) The Egalitarians - Human and Chimpanzee: An Anthropological View o f Social Organization. Cambridge: Cambridge University Press. Pusey AE, Williams J, and Goodall J (1997) The influence of dominance rank on the reproductive success of female chimpanzees. Science 277:828-830. Pusey AE (1990) Behavioural changes at adolescence in chimpanzees. Behaviour 115:203-246. Rabin BS (1999) Stress, Immune Function, and Health. New York: Wiley-Liss. Ransom TW (1981) Beach Troop o f the Gombe. Lewisburg: Bucknell University Press. Rea MA, Marshall GR, Weinbauer GF, and Nieschlag E (1986) Testosterone maintains pituitary and serum FSH and spermatogenesis in gonadotropin-releasing hormone antagonist-suppressed rats. Journal o f Endocrinology 108:101-107. Rebum CJ, and Wynne-Edwards KE (1999) Hormonal changes in males of a naturally biparental and a uniparental mammal. Hormones and Behavior 35:163-176. Robbins MM (1995) A demographic analysis of male life history and social structure of mountain gorillas. Behaviour 132:21-47. Robbins MM (1999) Male mating patterns in wild multimale mountain gorilla groups. Animal Behaviour 57:1013-1020. Robbins MM, and Czekala NM (1997) A preliminary investigation of urinary testosterone and cortisol levels in wild male mountain gorillas. American Journal o f Primatology 43:51 -64. Roberts RL, Zullo A, Gustafson EA, and Carter CS (1996) Perinatal steroid treatments alter alloparental and affiliative behavior in prairie voles. Hormones and Behavior 30:576-582. 237 Rodman PS, and Mitani JC (1986) Orangutans: sexual dimorphism in a solitary species. In BB Smuts, DL Cheney, RM Seyfarth, RW Wrangham and TT Struhsaker (eds.): Primate Societies. Chicago: University of Chicago Press. Rom DM (1990) A sequentially rejective test procedure based on a modified Bonferroni inequality. Biometrika 77:663-666. Rommerts FFG (1988) How much androgen is required for maintenance of spermatogenesis? Journal o f Endocrinology 116:7-9. Rose RM, Bernstein IS, and Gordon TP (1975) Consequences of social conflict on plasma testosterone levels in rhesus monkeys. Psychosomatic Medicine 37:50-61. Rose RM, Holaday JW, and Bernstein IS (1971) Plasma testosterone, dominance rank, and aggressive behavior in male rhesus monkeys. Nature 31:366-368. Rowell TE, and Dixson AF (1975) Changes in social organization during the breeding season of wild talapoin monkeys. Journal o f Reproduction and Fertility 43:419-434. Ruiz de Elvira M, Herndon J, and Collins D (1983) Effect of estradiol-treated females on all female groups of rhesus monkeys during the transition between the nonbreeding and breeding season. Folia Primatologica 41:191-203. Rust SW, and Fligner MA (1984) A modification of the Kruskal-Wallis statistic for the generalized Behrens-Fisher problem. Communications in Statistics - Theory and Methods 13:2013-2027. Saino N, and Moller (1994) Secondary sexual characters, parasites and testosterone in the barn swallow, Hirundo rustica. Animal Behaviour 48:1325-1333. Sapolsky RM (1982) The endocrine stress response and social status in the wild baboon. Hormones and Behavior 16:279-292. Sapolsky RM (1983) Endocrine aspects of social instability in the olive baboon {Papio anubis). American Journal o f Primatology 5:365-379. Sapolsky RM (1991) Testicular function, social rank and personality among wild baboons. Psychoneuroendocrinology 16:281-293. Sapolsky RM (1992) Cortisol concentrations and the social significance of rank instability among wild baboons. Psychoneuroendocrinology 17:701-709. 238 Sapolsky RM (1993) Neuroendocrinology of the stress-response. In JB Becker, SM Breedlove and D Crews (eds.): Behavioral Endocrinology. Cambridge: MIT Press. Sapolsky RM (1993) The physiology of dominance in stable versus unstable social hierarchies. In WA Mason and SP Mendoza (eds.): Primate Social Conflict. Albany: SUNY Press. van Schaik CP (1989) The ecology of social relationships amongst female primates. In V Standen and RA Foley (eds.): Comparative socioecology: the behavioral ecology o f humans and other mammals. Oxford: Blackwell. van Schaik CP, and van Hooff JARAM (1996) Toward an understanding of the orangutan's social system. In WC McGrew, LF Marchant and T Nishida (eds.): Great Ape Societies. Cambridge: Cambridge University Press. van Schaik CP, van Noordwijk MA, van Bragt T, and Blankenstein MA (1991) A pilot study of the social correlates of levels of urinary cortisol, prolactin, and testosterone in wild long-tailed macaques (Macaca fascicular is). Primates 32:345- 356. Schneider JE, and Wade GN (2000) Inhibition of reproduction in service of energy balance. In K Wallen and JE Schneider (eds.): Reproduction in Context. Cambridge, MA: MIT Press. Schoech S, Mumme RL, and Moore MC (1991) Reproductive endocrinology and mechanisms of breeding inhibition in cooperatively breeding Florida scrub jays (Aphelocoma c. coerulescens). The CowJor 93:354-364. Schweiger U, Tuschl RJ, Platte P, Broocks A, Laessle RG, and Pirke K-M (1992) Everyday eating behavior and menstrual function in young women. Fertility and Sterility 57:771-775. Selye H (1956) The Stress o f Life. New York: McGraw Hill. Shively C and Kaplan J (1984) Effects of social factors on adrenal weight and related physiology of Macaca fascicularis. Physiology and Behavior 33:777-782. Sicotte P (1993) Inter-group encounters and female transfer in mountain gorillas: Influence of group composition on male behavior. American Journal o f Primatology 30:21-36. Smuts B (1985) Sex and Friendship in Baboons. New York: Aldine de Gruyter. 239 Steklis HD, Brammer GL, Raleigh MJ, and McGuire MT (1985) Serum testosterone, male dominance, and aggression in captive groups of vervet monkeys (Cercopithecus aethiops sabaeus). Hormones and Behavior 19:154-163. Steklis HD, Raleigh MJ, Kling AS, and Tachiki K (1986) Biochemical and hormonal correlates of dominance and social behavior in all-male groups of squirrel monkeys. American Journal o f Primatology 11:133-145. Storer BE, and Kim C (1990) Exact properties of some exact test statistics for comparing two binomial proportions. Journal o f the American Statistical Association 85:146-155. Storey AE, Walsh CJ, Quinton RL, and Wynne-Edwards KE (2000) Hormonal correlates of paternal responsiveness in new and expectant fathers. Evolution and Human Behavior 21:79-95. Strier KB (1992) Faces in the Forest: The Endangered Muriqui Monkeys o f Brazil. New York: Oxford University Press. Strier KB, Ziegler TE, and Wittwer DJ (1999) Seasonal and social correlates of fecal testosterone and cortisol levels in wild male muriquis (Brachyteles arachnoïdes). Hormones and Behavior 35:125-134. Struhsaker TT (1975) The Red Colobus Monkey. Chicago: University of Chicago Press. Struhsaker TT (1997) Ecology o f an African Rain Forest. Gainesville: University Press of Florida. Talal N, and Ansar AS (1987) Sex hormones and autoimmune diseases: A short review. International Journal o f Immunotherapy 3:65-70. Taussky HH (1954) A microcolorimetric determination of creatine in urine by the Jaffe reaction. Journal o f Biological Chemistry 208:853-861. Trimble MR, and Herbert J (1968) The effect of testosterone or oestradiol upon the sexual and associated behavior of the adult female rhesus monkey. Journal o f Endocrinology 42:171-185. Trivers RL (1972) Parental investment and sexual selection. In B Campbell (ed.): Sexual Selection and the Descent o f Man, 1871-1971. Chicago: Aldine. 240 Tutin C (1975) Sexual behaviour and mating patterns in a community of wild chimpanzees {Pan troglodytes schweinfurthii). Ph.D., University of Edinburgh, Edinburgh. Tutin CEG (1979) Mating patterns and reproductive strategies in a community of wild chimpanzees. Behavioral Ecology and Sociobiology 6:39-48. Tutin CEG, and McGinnis PR (1981) Chimpanzee reproduction in the wild. In CE Graham (ed.): Reproductive Biology o f the Great Apes. New York: Academic Press. Utami S, and Setia TM (1995) Behavioral changes in wild male and female Sumatran orangutans {Pongo pygmaeus abelii) during and following a resident male take-over. In RD Nadler, BFM Galdikas, LK Sheeran and N Rosen (eds.): The Neglected Ape. New York: Plenum Press. Van Cauter E (1990) Diurnal and ultradian rhythms in human endocrine funtion: a minireview. Hormone Research 34:45-53. Vandenbergh JG, and Drickamer L (1974) Reproductive coordination among free- ranging rhesus monkeys. Physiology and Behavior 13:373-376. Vehrencamp SL (1983) A model for the evolution of despotic versus egalitarian societies. Animal Behaviour 31 .*667-682. Vermeulen A, Goemaere S, and Kaufman JM (1999) Testosterone, body composition and aging. Journal o f Endocrinological Investigations 22:110-116. Vermeulen A, Rubens R, and Verdonck L (1972) Testosterone secretion and metabolism in male senescence. Journal o f Clinical Endocrinology and Metabolism 34:730-735. Vigilant L, and Boesch C (2001) Reproductive strategies of West African chimpanzees {P. t. verus) of the Tai National Park: A reduced role for extra-group paternity. American Journal o f Physical Anthropology, Supplement 32:156. de Vries H, and Appleby MC (2000) Finding an appropriate order for a hierarchy: A comparison of the I&SI and the BBS methods. Animal Behaviour 59:239-245. de Waal FBM (1993) Reconciliation among primates: A review of empirical evidence and theoretical issues. In WA Mason and S Mendoza (eds.): Primate Social Conflict. Albany, NY: SUNY Press. 241 Wada M (1986) Circadian rhythms of testosterone-dependent behaviors, crowing and locomotor activity, in male Japanese quail. Journal o f Comparative Physiology A 158:17-25. Wallen K (2000) Risky business: Social context and hormonal modulation of primate sexual desire. In K Wallen and JE Schneider (eds.): Reproduction in Context. Cambridge: MIT Press. Wallis J (1992) Chimpanzee genital swelling and its role in the pattern of sociosexual behavior. American Journal o f Primatology 28:101-113. Wallis J (1997) A survey of reproductive parameters in the free-ranging chimpanzees of Gombe National Park. Journal o f Reproduction and Fertility 109:297-307. Wallis J (in press) Seasonality of reproductive behaviour in female chimpanzees at Budongo and Gombe. In C Boesch, G Hohmann and L Marchant (eds.): Behavioural Diversity in Chimpanzees andBonobos. Cambridge: Cambridge University Press. Wasser SK, and Barash DP (1983) Reproductive suppression among female mammals: implications for biomedicine and sexual selection theory. The Quarterly Review o f Biology 58:513-538. Watts DP (1989) Infanticide in mountain gorillas: new cases and a reconsideration of the evidence. Ethology 81:1-18. Watts DP (1996) Comparative socio-ecology of gorillas. In WC McGrew, LF Marchant and T Nishida (eds.): Great Ape Societies. Cambridge: Cambridge University Press. Watts DP (1998) Coalitionary mate guarding by male chimpanzees at Ngogo, Kibale National Park, Uganda. Behavioral Ecology and Sociobiology 44:43-55. Watts DP, and Mitani JC (2001) Boundary patrols and intergroup encounters in wild chimpanzees. Behaviour 138:299-327. Watts DP, and Pusey AE (1993) Behavior of juvenile and adolescent great apes. In ME Pereira and LA Fairbanks (eds.): Juvenile Primates: Life History, Development, and Behavior. Oxford: Oxford University Press. Weick RF (1981) The pulsatile nature of luteinizing hormone secretion. Canadian Journal o f Physiology and Pharmacology 59:779-785. 242 Weinbauer GF, and Nieschlag E (1990) The role of testosterone in spermatogenesis. In E Nieschlag and HM Behre (eds.): Testosterone: Action, Deficiency, Substitution. Berlin: Springer-Verlag. Weinsier RL, and Morgan SL (1993) Fundamentals o f Clinical Nutrition. St. Louis: Mosby. Weiss J (1970) Somatic effects of predictable and unpredictable shock. Psychosomatic Medicine 32:397-414. Whitten PL (2000) Evolutionary endocrinology of the cercopithecoids. In P Whitehead and C Jolly (eds.): Old World Monkeys. Cambridge: Cambridge University Press. Whitten PL, Brockman DK, and Stavisky RC (1998) Recent advances in noninvasive techniques to monitor hormone-behavior interactions. Yearbook o f Physical Anthropology 41:1 -23. Whitten PL, Stavisky R, Aureli F, and Russell E (1998) Response of fecal cortisol to stress in captive chimpanzees {Pan troglodytes). American Journal o f Primatology 44:57-69. Wickings EJ, and Dixson AF (1992) Testicular function, secondary sexual development and social status in male mandrills {Mandrillus sphinx). Physiology and Behavior 52:909-916. Wilcox RR (1996) Statistics for the Social Sciences. New York: Academic Press. Wilcox RR (1997) Introduction to Robust Estimation and Hypothesis Testing. New York: Academic Press. Williams J (1999) Female strategies and the reason for territoriality in chimpanzees: lessons from three decades of research at Gombe. Ph.D., University of Minnesota. Wingfield JC (1984) Androgens and mating systems: testosterone-induced polygyny in normally monogamous birds. The Auk \ 0\ :665-671. Wingfield JC, Ball GF, Duffy AM, Hegner RE, and Ramenofsky M (1987) Testosterone and aggression in birds. American Scientist 75:602-608. Wingfield JC, Hegner RE, Duffy AM, and Ball GF (1990) The "challenge hypothesis": Theoretical implications for patterns of testosterone secretion, mating systems, and breeding strategies. The American Naturalist 136:829-846. 243 Wingfield JC, Jacobs J, and Hillgarth N (1997) Ecological constraints and the evolution of hormone-behavior interrelationships. Annals o f the New York Academy o f Sciences 807:22-41. Wittig R (in press) Conflict management in Tai chimpanzees. In C Boesch, G Hohmann and L Marchant (eds.): Behavioural Diversity in Chimpanzees and Bonobos. Cambridge: Cambridge University Press. Wrangham RW (1975) The behavioural ecology of chimpanzees in Gombe National Park, Tanzania. Ph.D., University of Cambridge, Cambridge. Wrangham RW (1980) An ecological model of female-bonded primate groups. Behaviour 75:262-300. Wrangham RW (1999) The evolution of coalitionary killing: The imbalance-of- power hypothesis. Yearbook o f Physical Anthropology 42:1 -30. Wrangham RW (2000) Why are male chimpanzees more gregarious than mothers? A scramble competition hypothesis. In PM Rappeler (éd.): Primate Males. Cambridge: Cambridge University Press. Wrangham RW (in press) The cost of sexual attraction: Is there a tradeoff in female Pan between sex appeal and received coercion? In C Boesch, G Hohmann and M L. (eds.): Behavioural diversity in chimpanzees and bonobos. Cambridge: Cambridge University Press. Wrangham RW, Chapman CA, Clark-Arcadi AP, and Isabirye-Basuta G (1996) Socioecology of Kanyawara chimpanzees: Implications for understanding the costs of great ape groups. In WC McGrew, LF Marchant and T Nishida (eds.): Great Ape Societies. Cambridge: Cambridge University Press. Wrangham RW, Conklin NL, Chapman CA, and Hunt KD (1991) The significance of fibrous foods for Kibale Forest chimpanzees. Philosophical Transactions o f the Royal Society o f London, Series B 334:171-178. Wrangham RW, Conklin-Brittain NL, and Hunt KD (1998) Dietary response of chimpanzees and cercopithecines to seasonal variation in fruit abundance. I. Antifeedants. InternationalJournal o f Primatology 19:949-970. Yamagiwa J (1986) Activity rhythm and the ranging of a solitary male mountain gorilla (Gorilla gorilla beringei). Primates 27:273-282. Young LG, Gould KG, and Smithwick EB (1993) Selected endocrine parameters of the adult male chimpanzee. American Journal o f Primatology 31:287-297. 244 Zamboni L, Conaway CH, and van Pelt L (1974) Seasonal changes in production of semen in free-ranging rhesus monkeys. Biology o f Reproduction 11:251-267. Zumpe D, and Michael RP (1996) Social factors modulate the effects of hormones on the sexual and aggressive behavior of macaques. American Journal o f Primatology 38:233-261. 245 A p p e n d i x 1: L a b o r a t o r y P r o c e d u r e s Urinary Androgen Hydrolysis Procedure 1. Defrost the urine samples. 2. Aliquot 100 pi of each sample into a labeled 16 x 150 mm test tube. 3. Aliquot 20 pi of p-Glucoronidase/Arylsulfatase (Roche) into each tube. 4. Aliquot 300 pi of pH 5 buffer (Fisher Scientific) into each tube. 5. Vortex the tubes briefly. 6. Stopper the tubes and incubate overnight in 37° C water bath. 246 Urinary Androgen Extraction Procedure 1. Add 100 pi of testosterone recovery counts to each 16 x 150 mm tube containing a hydrolyzed sample. 2. Pipet 100 pi of testosterone recovery counts into each of three scintillation vials. Add scintillation fluid to the vials, cap, and shake. Label the vials TR (total recoveries). 3. Take three empty scintillation vials and add scintillation fluid to them. Cap vials and label BKG (background). 4. Briefly vortex the sample tubes. 5. Under ventilation hood, add 5 ml ether to each sample tube. A new can of ether is weighed and used for each extraction. 6. Vortex the tubes for two minutes. 7. Set the tubes in dry ice until aqueous phase has frozen (approximately 2 minutes). 8. Decant ether into labeled 16 x 100 mm test tubes. Set tubes in warm drier under nitrogen gas. 9. When aqueous phase has defrosted, add 5 ml of ether and vortex for one minute. 10. Set tubes in dry ice until aqueous phase has frozen. 11. Decant ether into the same 16 x 100 mm tubes and leave in warm drier under nitrogen until ether has completely evaporated. 12. Add 1 ml pH 7 buffer to each tube. 13. Vortex tubes for one minute. Stopper and refrigerate if leaving overnight. 247 Urinary Androgen Assay Set-Up Procedure 1. Label 12 x 75 mm tubes for samples in duplicate and two standard curves in triplicate. 2. Add 10 pi aliquots of extracted sample to each 12 x 75 tube. 3. Add 390 pi of pH 7 buffer to each sample tube and vortex briefly. 4. Aliquot 150 pi of extracted samples into scintillation vials and add scintillation fluid. Cap, shake, and label with sample number for recovery counts. 5. Set up each standard curve as follows: Label T Standard Buffer 00 (NSB) none 500 0(TB) none 400 5 2.5 400 10 5 395 20 10 390 50 25 375 100 50 350 200 100 300 400 200 200 800 400 none 6. Add 100 pi of testosterone antibody to each tube except 00. 7. Add 100 pi of hot testosterone to each tube (including 00). 8. Add 100 pi of hot testosterone to each of three scintillation vials. Add scintillation fluid, cap, and shake. Label the vials TC (total counts). 9. Vortex each of the 12 x 75 mm tubes briefly. 10. Cover the tubes with parafilm and aluminum foil, and refrigerate overnight. 248 Urinary Androgen Charcoal Procedure 1. Turn on the centrifuge and cool to 4° C. 2. Put charcoal on stirring plate in a bowl of ice. Put racks and carriers in iced tubs. (Charcoal solution is only good for one week. After initial preparation, stir charcoal solution for 30 minutes; thereafter, stir for at least 5 minutes before reusing.) 3. Rapidly add 200 pi charcoal to each tube. Vortex briefly. Put tubes in order in centrifuge carriers. 4. Incubate at 4° C for 15 minutes. 5. Spin for 10 minutes. 6. Decant into scintillation vials. 7. Add scintillation fluid to each vial, cap, and label. Shake vials. 8. Rinse out charcoal tubes in sink and throw away in plastic bag. Record amount of radioactivity washed down sink on sink disposal record. 9. Let the assay sit overnight. Be sure that the vials are in the correct order. The rack order is as follows: BKG TR sample recoveries TC standard curve in order (triplicates) samples (duplicates) standard curve in order (triplicates) 10. Count the assay and analyze the results. Aliquot sizes for the analysis program are as follows: Recovery . 15 Sample .001 249 Urinary Cortisol Assay Set-Up Procedure 1. Dilute each urine sample 1:10 with distilled water in a 16 x 100 mm tube. 2. Label 12 x 75 mm tubes for samples in duplicate and two standard curves in triplicate. 3. Add 30 pi aliquots of diluted sample to each 12 x 75 tube. 4. Add 370 pi of pH 7 buffer to each sample tube and vortex briefly. 5. Set up each standard curve as follows: Label C Standard Buffer 00 (NSB) none 500 0 (TB) none 400 5 2.5 400 10 5 395 20 10 390 50 25 375 100 50 350 200 100 300 400 200 200 800 400 none 6. Add 100 pi of cortisol antibody to each tube except 00. 7. Add 100 pi of hot cortisol to each tube (including 00). 8. Add 100 pi of hot cortisol to each of three scintillation vials. Add scintillation fluid, cap, and shake. Label the vials TC (total counts). 9. Vortex each of the 12 x 75 mm tubes briefly. 10. Cover the tubes with parafilm and aluminum foil, and refrigerate overnight. 250 Urinary Creatinine Procedure 1. Prepare a creatinine assay sheet. Each assay can accomodate up to 20 samples. 2. Pipet 100 pi of each urine sample into a labeled 16 x 100 mm test tube. 3. Add 900 pi of distilled water to each tube, so that the original urine sample has been diluted 1:10. 4. Vortex the tubes briefly. 5. Pipet 100 pi of distilled water into the first four wells (A1-A4) of a Dynatech styrene microtiter plate. 6. Pipet 100 pi of 1 mg/dl creatinine standard into the next 4 wells (A5-A8) of the plate. 7. Pipet 100 pi of 3 mg/dl creatinine standard into the next 4 wells (A9-A12) of the plate. 8. Pipet 100 pi of 10 mg/dl creatinine standard into the next 4 wells (B1-B4) of the plate. 9. Pipet 100 pi of the first diluted urine sample into the next 4 wells of the plate. Repeat for each sample. Be careful to place each sample in the proper wells. Refer to the creatinine assay sheet. 10. Pipet 50 pi of NaOH (0.75 M) into each well. 11. Pipet 50 pi of picric acid (0.02 M) into each well. 12. Cover the plate, and wait for one hour. 13. Measure the optical density of each well, using a 490 nm test filter and a 630 nm reference filter. Use wells A1-A4 as blanks. 14. Compute the slope of optical density between 0 mg/ml of creatinine, 1 mg/dl creatinine, 3 mg/dl creatinine, and 10 mg/dl creatinine. 15. Extrapolate creatinine concentrations of samples. (Use the median value from each set of 4 samples.) 251 UMI Number: DP23579 All rights reserved INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. Disssftaîi.oin PiiWisMng UMI DP23579 Published by ProQuest LLC (2014). Copyright in the Dissertation held by the Author. Microform Edition © ProQuest LLC. All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code ProQuest LLC. 789 East Eisenhower Parkway P.O Box 1346 Ann Arbor, Ml 48106- 1346

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Lab coats in the dream factory: science and scientists in Hollywood

PDF

Morphosyntactic feature chains and phonological domains

PDF

A comparison between aboriginal archery in western North America and eastern Asia

PDF

Out of many...: A social history of the homosexual rights movement as originated and continued in Los Angeles, California

PDF

The culture of the Taironas

PDF

Studies of two naturally occurring compounds which effect release of acetylcholine from synaptosomes

PDF

Numerical analysis of ecological survey data

PDF

The role of microbes in the formation of the tubeworm fouling community in Fish Harbor, Los Angeles Harbor, California

PDF

Resocialization of older adults to work roles through worklife autobiography

PDF

The innervation of the teeth and periodontium of the rat

PDF

Studies on the cytology of bacteria and yeasts by means of ultraviolet photolysis

PDF

Studies on the bacterial oxidation of saturated fatty acids and their derivatives

PDF

The 'Widmar' (1981) decision and the status of religious liberty on the public college and university campus: Antecedents, effects, and prospects

PDF

The sphere of the Keris: power and people in a Balinese princedom

PDF

Detection of marine neurotoxins and characterization of the presynaptic activity of iotrochotin from the sponge 'Iotrochota birotulata'

PDF

Studies on the aerobic oxidation of fatty acids by bacteria

PDF

The Allied Health Professions Admission Test : its role in selection for physical therapy programs

PDF

Cation localization in growth hormone and prolactin cells of the rat pars distalis

PDF

Self-perceived job satisfaction and dissatisfaction of urban elementary classroom teachers

PDF

A comparison of root resorption between Invisalign treatment and contemporary orthodontic treatment

Asset Metadata

Creator Muller, Martin N. (author)

Core Title Endocrine aspects of aggression and dominance in champanzees of the Kibale Forest

School Graduate School

Degree Doctor of Philosophy

Degree Program Anthropology

Degree Conferral Date 2002-05

Defense Date 05/01/2002

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

Tag biological sciences,OAI-PMH Harvest,social sciences

Language English

Contributor Digitized by ProQuest (provenance)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c30-203302

Unique identifier UC11225104

Identifier 397563 (dmrecord),https://doi.org/10.25549/usctheses-c3-397563 (doi),DP23579.pdf (filename),etd-muller-425749.pdf (filename),usctheses-c30-203302 (legacy record id),usctheses-c3-397563 (legacy record id)

Legacy Identifier DP23579.pdf

Dmrecord 203302

Document Type Dissertation

Rights Muller, Martin N.

Type texts

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

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

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA