Seasonal Effects on the Nutrition and Energetic Condition of Female White-Faced Capuchin Monkeys

University of Calgary PRISM: University of Calgary's Digital Repository

Graduate Studies The Vault: Electronic Theses and Dissertations

2015-05-05 Seasonal Effects on the Nutrition and Energetic Condition of Female White-Faced Capuchin Monkeys

Bergstrom, Mackenzie Lee

Bergstrom, M. L. (2015). Seasonal Effects on the Nutrition and Energetic Condition of Female White-Faced Capuchin Monkeys (Unpublished doctoral thesis). University of Calgary, Calgary, AB. doi:10.11575/PRISM/27720 http://hdl.handle.net/11023/2245 doctoral thesis

University of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission. Downloaded from PRISM: https://prism.ucalgary.ca UNIVERSITY OF CALGARY

Seasonal Effects on the Nutrition and Energetic Condition of

Female White-Faced Capuchin Monkeys

Mackenzie Lee Bergstrom

A THESIS

SUBMITTED TO THE FACULTY OF GRADUATE STUDIES

IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE

DEGREE OF DOCTOR OF PHILOSOPHY

GRADUATE PROGRAM IN ANTHROPOLOGY

CALGARY, ALBERTA

APRIL, 2015

Seasonal variation in food availability can ultimately affect female reproductive success. I investigated the effect of seasonality on the behavior, nutritional intake and physical condition of female white-faced capuchins in Sector Santa Rosa, Costa Rica. Capuchins are omnivorous neotropical primates that focus foraging effort on fruit and invertebrate resources. My project objectives were to 1) document the dietary profile of females; 2) measure the extent of seasonal variation in diet; 3) determine if seasonal variation in fruit abundance affects the physical condition of females; and 4) identify which ecological and social variables most strongly predict variation in energy balance. I collected 12 months of behavioral observations on 25 adult females living in three habituated groups between September 2009 and April 2011. I conducted phenological surveys and nutritional analyses of foods to document energy availability and nutritional intake. I measured the creatinine concentration and specific gravity of urine samples to estimate the relative muscle mass of females, and I ran radioimmunoassays to quantify urinary

C-peptide, a biomarker for energy balance. Females exhibited seasonal changes in foraging behavior, diet and physical condition. During low fruit abundance months, invertebrates comprised a larger proportion of total energy and protein intake than did fruit. Lepidopteran larvae (i.e., caterpillars) were a particularly important invertebrate resource and comprised a large proportion of energy and protein intake during the early rainy season (May – August).

Despite seasonal dietary shifts, females did not meet estimated energy requirements in some months. The relative muscle mass of females was higher during months with high versus low fruit abundance, and varied by group according to the energy available from ripe fruit. In addition, females produced urinary ketones during months with low fruit and energy intake. C- peptide analysis accurately reflected energy balance and indicated that females experienced low

ii energy balance during low fruit months. Ripe fruit energy density (kJ/ha) was the single most important predictor of female energy balance. This extensive examination of dietary flexibility and the physiological costs associated with living in a seasonal habitat will help us to elucidate the relationship among environmental conditions, behavior and reproductive success in female primates.

iii Acknowledgements

This dissertation is the culmination of years of hard work and would not have been possible without help from mentors, colleagues, friends and family. I cannot express how grateful I am to all of you for your support and guidance throughout this process – I truly could not have done it without you. There are a few people and organizations I would like to thank more extensively.

Academic Mentors

First and foremost, I would like to thank my supervisor, Dr. Linda Fedigan, for providing me with the opportunity, support and resources to pursue this degree. Linda’s guidance and feedback helped to shape my project from the early stages of development and grant writing to finalization of the dissertation chapters. She also contributed in many ways to my professional development, including teaching me to work independently, guiding me to think critically and broadly about results, and helping me to develop polished presentations of my work. Without her support, this project would not have been possible. Also, thank you to Drs. Steig Johnson,

Pascale Sicotte, Mary Pavelka, Charles Mather, and many other faculty members and staff within the Department of Anthropology & Archaeology at the University of Calgary, for providing help and guidance during various stages of my masters and doctoral programs.

Thank you to Dr. Melissa Emery Thompson, who hosted me in her laboratory and provided invaluable support and feedback throughout many stages of this project. Melissa provided essential logistical advice from the setup of my project through shipping my samples from Costa Rica. She also trained and supervised me to analyze capuchin urine and fecal samples in the Hominoid Reproductive Ecology Laboratory at the University of New Mexico.

She and Dr. Michelle Brown dedicated their time (and patience) to long Skype and email conversations while I was dealing with urine standardization issues and trying to understand and

iv streamline the associated statistical analyses. These conversations, as well as those with Drs.

Tak Fung, Fernando Campos, and Eva Wikberg, were instrumental in helping me to assess and integrate the results of my study.

Lastly, I would like to thank again the various mentors that have helped to guide me along the path that lead me to pursue masters and doctorate degrees in the first place. At Trinity

University (San Antonio, TX), Drs. James Shinkle and Thomas Koppenheffer shared their passion for biology with me and inspired me to conduct research. Also at Trinity, Dr. Jennifer

Mathews introduced me to the field of primatology and encouraged me to pursue my interests.

Dr. Jeffrey Rogers taught me valuable laboratory skills and extended the job flexibility that allowed me to explore my interest in social behavior at Southwest Foundation for Biomedical

Research (SFBR). Also at SFBR, Dr. Amanda Vinson was a role model, mentor and friend who encouraged me to pursue a graduate degree and to try fieldwork. Finally, Drs. Susan Perry and

Joseph Manson provided me with the opportunity and diligent training to carry out a field-based research project. The work I did while employed as part of The Lomas Barbudal Capuchin

Monkey Project team provided much of the groundwork that has contributed to my success in this field, and most importantly, it sparked my love of capuchin monkeys. I would not be where

I am today without the support of these influential people.

Funding Agencies

This project would not have been possible without funding from a number of sources – I would like to express my gratitude to all of them. Specifically, I would like to thank Alberta Innovates

Technology Futures (Alberta Innovates Student Scholarship), the University of Calgary (Dean’s

Entrance Scholarship, Faculty of Graduate Studies Doctoral Scholarship, and the Dean’s

Doctoral Scholarship), and the Department of Anthropology & Archaeology (Teaching

v Assistantships and Graduate Research Scholarships) for funding that allowed me to focus full- time on my dissertation. The following research grants supported my dissertation project: The

Leakey Foundation Research Grant, Alberta Innovates Student Research Grant, International

Primatological Society Research Grant, Animal Behaviour Society Student Research Grant,

Sigma Xi Grants-in-Aid of Research, and the University of Calgary Dissertation Research Grant.

Dr. Linda M. Fedigan’s Natural Sciences and Engineering Research Council of Canada

(NSERCC) and Canada Research Chairs Program grants also provided project funding.

Assistance at Sector Santa Rosa, Costa Rica

I would like to thank the Costa Rican park staff and the Ministry of the Environment, Energy and

Technology (MINAET), for permission to carry out my research in Sector Santa Rosa, Costa

Rica. In particular, thank you to Sr. Roger Blanco Segura and María Marta Chavarria Diaz for your extended help throughout this project.

Thank you to my field assistants: Chelsea Lees, Barb Kowalzik, Bhavisha Thankey,

Heidi Clouse, and Monica Myers. I realize that working alongside a Type A perfectionist is not always the easiest of tasks, and I thank you for your hard work and dedication, patience and understanding, and friendship during what turned out to be an insanely demanding project.

There were many crazy and memorable field moments, including The Magical House of Green

Bees, The Peanut (aka “pee net”), and Rick the Kung Fu Tamandua. I cannot emphasize enough how incredibly grateful I am that you persevered through the many unfortunate circumstances we endured while in pursuit of urine samples. These include, but are not limited to, being peed on, pooed on, vomited on, stung by bees or wasps, and attacked by acacia or bullet ants. Your willingness (even when reluctant) to trek around in soggy boots and field clothes after wading through mucky snake-filled quebradas, or hang yourselves off of cliffs in desperate attempt to

vi catch a sample after following a female for countless hours, was appreciated and will not be forgotten. In light of the above circumstances, I would like to give a special shout out to Salsa

(“donor” of my first high-volume urine sample) and Padma (the monkey equivalent of Old

Faithful), as well as a bit of a cold shoulder to Dos Leches, “Merve” and Lily for seemingly purposefully withholding their valuable biological resources more often than not. Generally speaking however, I owe a great deal of thanks to all of the females in LV, CP and GN groups for providing data, lots of laughs, and for withstanding our presence.

I am deeply grateful to the other capuchin researchers with whom I shared my field seasons, especially Valerie Schoof and Claire Sheller. Having also conducted PhD projects, I know you know that words cannot express the importance of your friendship and support. I am also appreciative of Teresa Holmes, who was with me at the very beginning of my graduate career and assisted during my masters project, and then returned to Costa Rica during the final stages of my PhD project to help with invertebrate sample collection (alongside Dr. Caroline

Turner Hogan, Monica Myers and Saúl Chevez). Additional thanks to the gusaneros, Ruth

Franco and Johan Vargas, who provided logistical support while I was dehydrating and storing invertebrate samples. Roberto Espinoza and Adrián Guadamuz Chavarria also helped with the identification of a number of plants species. Ronny Senteno Garnier started out by providing taxi services for our research group but quickly became my friend and Costa Rican dad. Thank you to Ronny for the huge amount of help with logistical aspects of my project (especially in ensuring the urine samples I collected were shipped safely), and for his welcomed company, kindness, and hilarious stories (particularly the golden sweater and the Bagaces wave). Finally, thank you to all of the friends I spent time with in Costa Rica during this project, who helped to make my field experiences memorable including, Rafa Sandoval, Julia Tarquino, Yvette Foulds-

vii Davis, Elvin Murillo Chacόn, Norberto Asensio, Lilly Morales, Francisco Pizarro, and Edith

Lόpez Lara.

Assistance with project logistics, data and writing

I would like to give a special thank you to Dr. John Addicott for being a data guru, and for kindly offering his support. As I stated in my master’s thesis, I have no idea how I could have dealt with and analyzed my data without the countless hours of help he provided in developing my data parser and database. I am sure that having my project off his plate is cause for immense celebration. Additional thanks to my cousin, Jim McQuade, for helping to automate some of the initial data checking. Also on the topic of data, thank you to Dr. Erin Vogel for sharing a portion of her unpublished nutritional dataset, which allowed me to conduct a more thorough analysis of the Santa Rosa capuchin diet. I would also like to thank Dr. Amanda Melin and Dr. Melissa

Emery Thompson for their extensive guidance and feedback as collaborators on papers that we plan to publish from Chapters 2-5. Sarah Carnegie, Josie Vayro, Juli Finlay, Monica Myers,

Jeremy Hogan, Kayla Hartwell, my partner, Javier Borau Garcia, and my mom, Terry

Bergstrom, also provided helpful feedback on drafts of various dissertation chapters. Thank you to Tracy Wyman for technical support with my ranging data and for lending a listening ear during stressful dissertation moments. Finally, thank you to Greg Bridgett for all of the logistical help he provided related to supply orders, financial claims, sample shipment and nutritional processing.

Friends and Family

Last but certainly not least, I would like to thank my friends and family. Amanda Melin,

Fernando Campos, Nigel Parr, Monica Myers and Sarah Carnegie were part of my Santa Rosa family at the University of Calgary. Eva Wikberg, Julie Teichroeb, Sheila Holmes, Teresa

viii Holmes, Josie Vayro and Juli Finlay were among my closest friends within the wonderfully supportive Department of Anthropology & Archaeology at the U of C. Thank you also to

Amanda Bannister-Lefurgey, Erin Baerwald, Brandon Klug, and Clint Westgard for your friendship and sharing in much needed non-academic fun in Calgary. I am also grateful to the many friends I made in New Mexico while I was there carrying out my lab work, especially

Tone Jackson, Lizzy Eadie, Alden Rice Gilligan (and family), and Lori Berman.

Thank you to my amazingly fabulous, wonderful and caring family: Mom, Dad, Ryan,

Laurel, Gram, Karen, Lin, Jim, Kate, Jimmy, Bob, Kayla, Kris, Bill, John, Erin and Javier. I’m sure it hasn’t been easy explaining to people all of these years why in the world I study monkeys in the jungles of Central America and what I’m going to do career-wise with this degree! Thank you for always being there for me and believing in me – I could not have done this without you.

Over the years, you have taught me to work hard, encouraged me to set high goals, and motivated me to continue to strive to pursue them. I am who I am, and I am where I am today, largely because of your unconditional love and support. Finally, thank you to Javier for your love, patience and understanding as I’ve dealt with many project related ups and downs. I am so lucky to have had you by my side as this project has come to an end. Thank you for the endless encouragement – you mean the world to me!

ix To My Family

Thank you for your love and support.

x Table of Contents

Abstract ...... ii Acknowledgements ...... iv Table of Contents ...... xi List of Tables ...... xv List of Figures and Illustrations ...... xvi

CHAPTER ONE: GENERAL INTRODUCTION ...... 1 1.1 Study site: Sector Santa Rosa, Área de Conservaciόn Guanacaste, Costa Rica ...... 4 1.2 The capuchin monkey ...... 7 1.2.1 Species overview: The white-faced capuchin monkey...... 7 1.2.2 Study subjects and data collection ...... 9 1.2.2.1 Data collection schedule ...... 9 1.2.2.2 Female study subjects ...... 11 1.2.2.3 Data collection ...... 13 1.3 Dissertation outline ...... 13

CHAPTER TWO: DIETARY PROFILE, FOOD COMPOSITION, AND NUTRITIONAL INTAKE ...... 16 2.1 Introduction ...... 16 2.2 Methods ...... 21 2.2.1 Study Site ...... 21 2.2.2 Study Subjects ...... 21 2.2.3 Reproductive state ...... 22 2.2.4 Estimating capuchin nutritional requirements ...... 23 2.2.5 Behavioral Data Collection ...... 24 2.2.5.1 Focal follows ...... 24 2.2.5.2 Activity budgets ...... 24 2.2.5.3 Ingestion rates: food items, energy, and macronutrients ...... 25 2.2.6 Ecological Data Collection ...... 25 2.2.6.1 Nutritional sample collection ...... 25 2.2.6.2 Laboratory nutritional sample analysis ...... 28 2.2.6.3 Calculating the energy density of food items ...... 28 2.2.7 Fruit abundance: fruit biomass (kg/ha) and ripe fruit energy density (kJ/ha) .31 2.2.8 Statistical analyses ...... 33 2.3 Results ...... 34 2.3.1 General activity and foraging budgets ...... 34 2.3.2 Dietary profile ...... 36 2.3.3 Nutritional composition of different capuchin foods ...... 37 2.3.4 Dietary composition, intake rates and nutritional profitability ...... 43 2.3.5 Estimated nutritional demands for female capuchins ...... 44 2.3.6 Ripe fruit energy density ...... 44 2.3.7 Differences in feeding behavior and intake according to fruit abundance ...... 45 2.3.8 Differences in macronutrient intake according to fruit abundance ...... 49 2.4 Discussion ...... 51

xi CHAPTER THREE: THE NUTRITIONAL IMPORTANCE OF INVERTEBRATES ...58 3.1 Introduction ...... 58 3.2 Methods ...... 62 3.2.1 Study site ...... 62 3.2.2 Study subjects ...... 63 3.2.3 Behavioral data collection ...... 63 3.2.4 Nutritional sampling and analysis ...... 64 3.2.5 Ecological sampling and fruit availability ...... 64 3.2.6 Invertebrate abundance ...... 65 3.2.7 Statistical analysis ...... 66 3.3 Results ...... 68 3.3.1 Group-level variation in fruit availability and ripe fruit energy density ...... 68 3.3.2 Fruit energy density and invertebrate consumption ...... 68 3.3.3 Dietary profile of invertebrate prey captures ...... 72 3.3.4 Annual energy contribution from invertebrate consumption ...... 74 3.3.5 Seasonal variation in invertebrate consumption ...... 77 3.4 Discussion ...... 84 3.4.1 Variation in ripe fruit energy density across the study groups’ home ranges and the relationship between fruit and invertebrate consumption ...... 84 3.4.2 Intra- and inter-annual variation in ripe fruit availability and energy density.85 3.4.3 The invertebrate role in the dietary profile of female capuchins ...... 86 3.4.4 The energetic and seasonal importance of invertebrates ...... 87 3.4.5 The nutritional role of invertebrates in relation to the capuchin’s staple diet of seasonally variable fruit ...... 88

CHAPTER FOUR: USING URINARY PARAMETERS TO ESTIMATE SEASONAL VARIATION IN THE RELATIVE MUSCLE MASS OF FEMALES ...... 92 4.1 Introduction ...... 92 4.1.1 The role of creatinine in muscle mass estimation as a measure of physical condition ...... 93 4.2 Methods ...... 97 4.2.1 Sample collection ...... 97 4.2.2 Urine analysis ...... 97 4.2.2.1 Specific gravity ...... 97 4.2.2.2 Creatinine ...... 97 4.2.3 Ecological and nutritional sampling ...... 99 4.2.4 Statistical analysis ...... 99 4.3 Results ...... 101 4.3.1 Specific gravity: lab versus field measures ...... 101 4.3.2 Relative muscle mass ...... 102 4.4 Discussion ...... 105

CHAPTER FIVE: ASSESSMENT OF ENERGY BALANCE...... 110 5.1 Introduction ...... 110 5.1.1 C-peptide: a biomarker for energy balance ...... 112 5.1.2 Application of UCP in non-human primate studies ...... 113 5.1.3 White-faced capuchins ...... 114

xii 5.1.4 Research Questions and Predictions ...... 115 5.2 Methods ...... 117 5.2.1 Study Site ...... 117 5.2.2 Reproductive state ...... 117 5.2.2.1 Rank ...... 118 5.2.2.2 Energy intake ...... 119 5.2.2.3 Energy expenditure ...... 119 5.2.3 Ecological Data Collection ...... 123 5.2.4 Urine collection ...... 123 5.2.5 Laboratory measurement of C-peptide ...... 124 5.2.6 C-peptide standardization ...... 126 5.2.7 Data Analysis ...... 128 5.3 Results ...... 129 5.3.1 The relationship between calculated energy balance and UCP ...... 129 5.3.3 Ecological and social predictors of urinary C-peptide ...... 131 5.4 Discussion ...... 133 5.4.1 Calculated energy balance versus UCP ...... 134 5.4.2 UCP as a predictor of ketone production ...... 136 5.4.3 Predictors of energy balance (UCP) ...... 138

CHAPTER SIX: DISCUSSION ...... 143 6.1 Summary and synthesis ...... 143 6.2 Broader application and future directions ...... 149 6.2.1 Within-species inter-annual comparison and the stability of ecological and social environments ...... 149 6.2.2 Inter-population and inter-species comparison ...... 151 6.2.3 Urine as a biological medium ...... 153 6.3 Conclusion ...... 154

REFERENCES ...... 155

APPENDIX A: DATA COLLECTION SCHEDULE AND GROUP CONTACT...... 177

APPENDIX B: GROUP DEMOGRAPHICS ...... 181

APPENDIX C: FOCAL ANIMAL SAMPLE TOTALS ...... 185

APPENDIX D: BEHAVIORAL ETHOGRAM ...... 186

APPENDIX E: URINALYSIS TEST PARAMETERS ...... 191

APPENDIX F: LABORATORY METHODS USED IN NUTRITIONAL ANALYSES192

APPENDIX G: NUTRITIONAL COMPOSITION OF VEGETATIVE FOOD ITEMS193

APPENDIX H: NUTRITIONAL COMPOSITION OF INVERTEBRATE FOOD ITEMS196

xiii APPENDIX I: FEEDING, ENERGY INTAKE AND MACRONUTRIENT INTAKE RATES ...... 197

APPENDIX J: ENERGY CONSTANTS FOR ENERGY EXPENDITURE DURING ACTIVITY ...... 200

xiv List of Tables

Table 1.1 Details of female study subjects. Name, female ID, group membership and date of birth, as well as age class, parity and rank are listed by study period (i.e., field season 1, 2 and 3)...... 12

Table 2.1 Contribution of different food types to the annual foraging profile. Values are displayed as mean  SE for female white-faced capuchins at Sector Santa Rosa, Costa Rica...... 37

Table 2.2 Summary of macronutrient content and energy density per food type for food items eaten by adult females in the study groups...... 39

Table 2.4 Summary of linear mixed effects models analyzing the effect of monthly ripe fruit energy density (kJ/ha) on variation in mean feeding time and nutrient intake...... 47

Table 3.2 Linear regression of ripe fruit energy density and the percentage of energy intake from invertebrates. Two outliers (> 2SD from mean) were removed for analysis...... 72

Table 3.3 Taxonomic composition of invertebrate prey consumed by females...... 74

Table 3.5 Results of directional analysis of seasonality in invertebrate consumption...... 77

Table 4.1 Mixed effects model results predicting the relationship between creatinine and specific gravity (Model 1) and the effect of ripe fruit availability and group on the relationship between creatinine and specific gravity (Model 2)...... 104

Table 5.1 Generalized estimating equation results predicting the effect of urinary C-peptide on the presence of urinary ketones...... 130

Table 5.2 Model set and model selection results for urinary C-peptide...... 132

xv List of Figures and Illustrations

Figure 1.1 Diagram depicting the link among foraging behavior, energy balance and reproductive success...... 2

Figure 1.3 Sectors of the Área de Conservaciόn Guanacaste, Costa Rica. Image courtesy of the ACG website, 2012...... 5

Figure 1.4 Weather data collected at Sector Santa Rosa, Costa Rica. Mean minimum temperature, maximum temperature and rainfall are shown by month during the study period (Jan-Apr 2011, May-Aug 2010, Sep-Dec 2009)...... 6

Figure 1.5 Illustrations of the home range locations, sizes and overlap among LV, CP and GN groups during the 2009-2011 study periods. Individual group home ranges (right) are scaled from least intensively used (white/yellow) to most intensively used (red). The administration area is located where the home ranges of all three groups overlap. Figure courtesy of Fernando Campos...... 10

Figure 2.1 Annual foraging budget. The pie chart depicts the mean percentage of time spent foraging on invertebrate, fruit, pith, flower, vertebrate, water/other food types, as well as visually foraging by all three study groups...... 35

Figure 2.2 Monthly foraging budget. The graph depicts the mean  SE percentage of time spent foraging on the four most common food types by all three study groups...... 35

Figure 2.3 Box-and-whisker plots of the nutritional composition of different food types eaten by adult female study subjects. Graphs depict: a) percentage of moisture, b) crude protein, c) crude fat, d) water soluble carbohydrates (WSC), e) neutral detergent fiber (NDF), f) energy (kilojoules/gram dry matter), g) energy (kilojoules/gram wet mass), h) energy (kilojoules/item). Open circles and asterisks represent outliers > 1.5 times the IQR and > 3 times the IQR, respectively...... 42

Figure 2.4 Energy density from ripe fruit (kJ/ha) based on 30 fruiting species important to the diet of white-faced capuchins at Sector Santa Rosa, Costa Rica. Data collection periods for this study are highlighted in yellow...... 45

Figure 2.5 Monthly variation in feeding time and intake by females. Mean ( 95% Confidence Interval) monthly values for a) feeding time, b) the proportion of energy intake from fruit (kJ/hr), c) mass ingested (grams dry matter per hour), and d) energy intake (kJ/hr) are plotted against monthly ripe fruit energy density for 25 white-faced capuchin females at Sector Santa Rosa, Costa Rica. The dotted line depicts the direction of the relationship...... 48

Figure 2.6 Monthly variation in macronutrient intake by females. Mean ( 95% Confidence Interval) monthly intake, reported in grams intake per hour, for a) crude protein, b) crude fat, c) water soluble carbohydrates (WSC), and d) neutral detergent fiber (NDF) are plotted against monthly ripe fruit energy density for 25 female white-faced

xvi capuchins at Sector Santa Rosa, Costa Rica. The dotted line depicts the direction of the relationship...... 50

Figure 3.1 Ripe fruit energy density (kJ/ha) available in the home ranges of the three study groups. Calculations are based on ranging, transect and phenology data for the 2009- 2011 study period. The dotted line represents the group’s annual mean ripe fruit energy density...... 70

Figure 3.2 Monthly ripe fruit energy density versus the monthly percentage of energy intake from invertebrates...... 71

Figure 3.3 Group-level correlation between ripe fruit energy density and the percentage of energy intake from invertebrates (log transformed). One outlier (> 2SD from mean) was removed before analysis...... 71

Figure 3.4 Rose diagrams depicting the seasonality of invertebrate energy intake from the four most important invertebrate orders: a) Lepidoptera, b) Orthoptera, c) Hemiptera, and d) Hymenoptera. Months are assigned as angles and axes are labeled as the number of follows (of the standardized 120) that include each category of invertebrate consumption...... 78

Figure 3.5 Cartesian graphs of mean monthly energy intake rates for different food types. Energy intake rates were measured per hour of focal observation time for a) plants, b) lepidopterans, c) orthopterans, d) hemipterans, e) hymenopterans and f) other invertebrates. Bars represent mean intake rate  standard error...... 81

Figure 3.6 Percentage of monthly energy intake (left column) and protein intake (right column) from plants and top invertebrate categories. Data are shown for females in CP group (top row, a and b), GN group (middle row, c and d), and LV group (bottom row, e and f). Other than plant foods, energy and protein intake from lepidopterans from May – August comprises the majority of intake...... 82

Figure 3.7 Mean contributions of fruit and invertebrates to the a) estimated daily energy intake and b) protein intake per metabolic body weight per day. The dotted red line indicates monthly estimated a) energy intake based on the estimated 1000 kJ/day and b) protein requirements based on the 1.8 g/metabolic kg/day. The dotted black line is adjusted by multiplying minimum requirements by the corresponding energy coefficients (1.00, 1.25 and 1.50) of the mean monthly reproductive state (cycling, gestating and lactating, respectively) (Ausman et al. 1986; Ausman and Hegsted 1980; Key and Ross 1999)...... 83

Figure 4.1 Linear relationship between specific gravity (SG) measured using urinalysis tests in the field versus a refractometer in the laboratory for urine samples (N = 486). Filled circles and the associated regression slope (short-dash line) indicate lower SG readings (N = 168) and open circles and the associated regression slope (long-dash line) indicate higher SG readings (N = 318). The solid regression line includes all data points (both filled and open circles)...... 102

xvii Figure 4.2 Scatterplot displaying the relationship between specific gravity (minus 1) and creatinine (mg/ml) of urine samples (N = 703)...... 103

Figure 4.3 Comparison of relative muscle mass ( SE) across groups and between months with high (grey diamonds) versus low (black circles) ripe fruit availability. Shown on the y-axis is the model-fitted estimated mean residuals. Data are plotted per group and by season...... 105

Figure 5.1 Monthly mean ( SE) urinary C-peptide (black line) and monthly mean ( SE) daily energy balance (grey dashed line). Grey shading indicates months with urinary ketone production...... 129

Figure 5.2 Monthly mean ( SE) ripe fruit energy density (kJ/ha) (black bars) and monthly mean ( SE) urinary C-peptide (black line). Ripe fruit energy density significantly predicted UCPSG concentration (intercept = 2.375, estimate = 0.034, SE = 0.009, df = 208, t = 3.575, p < 0.001, 95% CI = 0.015 to 0.052)...... 133

xviii

Chapter One: General Introduction

Energy obtained through diet is required for animals to maintain life and stimulate tissue maintenance, growth and reproduction. Carbohydrates, proteins and lipids are the macronutrients that provide the backbone of these processes. Mammals, including primates, obtain these nutrients through a wide variety of means influenced by morphology and behavior

(National Research Council 2003; Stevens and Hume 1996). Among mammals, metabolic demands scale according to body mass (Kleiber 1961). Individual needs and subsequent nutritional intake and foraging behavior may also be affected by other intrinsic factors such as sex, age, reproductive state and digestive processes or by social factors such as dominance rank and associated resource competition (Coelho 1974; Coelho et al. 1976; Durnin and Passmore

1967; Leonard and Robertson 1997; Marshall et al. 2012; National Research Council 2003;

Oftedal et al. 1991; Taylor et al. 1982; Taylor et al. 1970) . The differential and combined effect of each of these factors influences diet and contributes to a large amount of inter-species and inter-individual variation in energetics, and ultimately, reproductive success (e.g., baboons,

Altmann 1998).

A female’s reproductive success is closely tied to her ability to access food resources, as the energetic costs of reproduction (including the production of nutrient-rich ova and energy expenditure during gestation, lactation and infant care) are often higher for females than males in mammalian species (Bateman 1948; Trivers 1972; Wade and Schneider 1992). Consequently, a female’s foraging strategies are expected to maximize the amount of energy obtained per unit time (Schoener 1971). Spatial or temporal variation in the availability of energy and nutrients resulting from ecological changes and population/group dynamics complicate this process. A

female may modify foraging behavior in terms of time spent searching for food, the type of food eaten, the nutrients she seeks, her rate of food intake, or the compounds she avoids in attempt to minimize metabolic costs (Bronson 1985; Clutton-Brock and Harvey 1977; Freeland and Janzen

1974; Simpson and Raubenheimer 1993). The success of foraging strategies directly affects female energy balance. The latter reflects energy intake minus the energy expended through basal metabolic processes and the added metabolic demands of activity and reproductive state.

In turn, a female’s energy balance plays an important role in reproductive processes that influence ovulation (Ellison and Valeggia 2003; Willis et al. 1996), and thus provides an important clue to understanding the relationship between the behavior and reproductive success of females.

Figure 1.1 Diagram depicting the link among foraging behavior, energy balance and reproductive success.

Primate research has identified a number of ecological variables that are central to our understanding of foraging and feeding behavior, including spatial and temporal variation in food resources (Oates 1987). However, the implementation of methods to assess nutrition and energetic status in studies of wild primates is relatively new (Chapman et al. 2003; Felton et al.

2009a; Lambert 2011; Oftedal et al. 1991; Rothman et al. 2012). Quantification of physiological processes in wild populations is logistically difficult, but possible using nutritional methods and physiological biomarkers that reflect metabolic stress, including glucocorticoids, ketones, creatinine, stable isotope ratios and C-peptide (Chapman et al. 2006; Conklin-Brittain et al. 2006;

Deschner et al. 2012; Emery Thompson et al. 2012; Foerster and Monfort 2010; Knott 1998;

Sherry and Ellison 2007). Of these, C-peptide, a polypeptide segment of the proinsulin molecule that is cleaved in an equal relationship to insulin during its production, has been of particular interest in recent primate research because it serves as a metabolic biomarker and has successfully been shown to reflect the energy balance of individuals in many species of apes and

Old World monkeys [chimpanzees (Pan troglodytes) and orangutans (Pongo pygmaeus): Sherry and Ellison 2007; chimpanzees: Emery Thompson et al. 2009, Georgiev 2012; orangutans:

Emery Thompson and Knott 2008; bonobos (Pan paniscus): Deschner et al. 2008; mountain gorillas (Gorilla beringei beringei): Grueter et al. 2014; rhesus macaques (Macaca mulatta):

Girard-Buttoz et al. 2011; black and white colobus (Colobus guereza): Harris et al. 2010; and baboons (Papio hamadryas anubis): Lodge 2012]. The broad application of nutritional methods to assess diet and the utilization of recent methodological advances to measure the physiology of wild primates will together help us to gain a better understanding of the proximate variables affecting the physical condition of individuals. Overall, this approach will elucidate the relationships among behavior, the environment, social dynamics and reproductive success.

This dissertation describes a 12-month field study investigating the effect of seasonal variation on the behavior, nutrition and physical condition of female white-faced capuchins

(Cebus capucinus) living in three habituated social groups within a tropical dry forest in Sector

Santa Rosa (SSR), Área de Conservaciόn Guanacaste (ACG), northwest Costa Rica. I use non-

invasive methods and an inter-disciplinary approach to collect data on ecological variation (i.e., climate and food abundance), capuchin feeding and social behavior, the nutritional composition of capuchin foods, and urinary parameters that help to quantify the physical condition of adult females. The overarching goal of this project is to better understand how a seasonal environment affects the nutrition and physiology of female capuchins. Examining the relationships among these variables in this long-term study population will provide specific insight into the nutritional and energetic consequences of seasonal changes in climate and resource abundance. I also hope to place these consequences into a broader framework for understanding how proximate ecological and social pressures in this species affect female reproductive success.

1.1 Study site: Sector Santa Rosa, Área de Conservaciόn Guanacaste, Costa Rica

This study was conducted in Sector Santa Rosa (SSR; 10º50ʹ30ʹʹN, 85º37ʹ0ʹʹW) within the Área de Conservaciόn Guanacaste (ACG), Costa Rica (Figure 1.2). SSR is defined by over 100 km2 of regenerating tropical dry forest and was founded in 1971 as the first protected area of the current 18 sectors within the greater ACG (Figure 1.3). The ACG spans approximately 163,000 hectares and includes four types of tropical ecosystems (marine, dry forest, cloud forest and rainforest) traversing Costa Rica from the coast of the Pacific Ocean, through the volcanic terrain of the Guanacaste Volcanic Range (Orosí, Cacao and Rincón de la Vieja), and to the Caribbean lowlands.

Figure 1.2 Location of Sector Santa Rosa and the Área de Conservaciόn Guanacaste in northwestern Costa Rica. Dark gray shading denotes the area encompassed by the ACG and the white shading denotes Santa Rosa. Image courtesy of Melanie Meeking.

Figure 1.3 Sectors of the Área de Conservaciόn Guanacaste, Costa Rica. Image courtesy of the ACG website, 2012.

The tropical dry forest in SSR is highly seasonal. A distinct and severe dry season occurs from mid-December to mid-May, during which deciduous trees drop their leaves and standing water evaporates. A delimited rainy season ensues from mid-May to mid-November (average rainfall: 1,792 mm) (Fedigan and Jack 2012). Annual temperatures recorded by various members of the Fedigan and Jack research teams from 1980 to 2012 range from an average minimum of 17.0 ºC (s = 3.2, SE = 0.5) to a maximum of 37.5 ºC (s = 3.5, SE = 0.6). Variations in temperature (minimum = 22.5 – 26.0 and maximum = 27.8 – 33.3) and rainfall (0.0 – 413.4 mm/mo) recorded during this study are shown in Figure 1.4. Climatic seasonality creates considerable variation in the availability of food and water that may affect the behavior and physiology of species inhabiting this region.

900 34

800

700 32

600 30

500 28 400 26 300

24 Mean Rainfall (mm) Rainfall Mean

200 ºC Temperature Mean

100 22

0 20 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month

Rainfall Max Temperature Min Temperature

The ACG was named a UNESCO World Heritage Site in 1999 as one of the largest tropical forest preservation initiatives containing 2.6% of the world’s biodiversity (UNESCO

World Heritage Committee 1999). An estimated 250 bird species and 115 species of mammals inhabit SSR (ACG website 2012), including three sympatric primates: the mantled howler monkey (Alouatta palliata), the black-handed spider monkey (Ateles geoffroyi) and the white- faced capuchin monkey (Cebus capucinus). Of these, the white-faced capuchins have been studied most extensively by Dr. Linda Fedigan and her colleagues and students since 1983.

1.2 The capuchin monkey

Capuchin monkeys are neotropical primates found across a broad geographic range that spans from S.E. Honduras, Central America, to the southern regions of the Atlantic Forest in S.E.

Brazil, South America (Fragaszy et al. 2004). In 2012, the capuchin genus Cebus was divided into two clades based on genetic evidence that robust “tufted” capuchin species (Sapajus apella,

S. flavius, S. libidinosus, S. nigritus, and S. xanthosternos) diverged from gracile “untufted” capuchin species (Cebus capucinus, C. albifrons, C. kaapori, C. olivaceus) approximately 6.2

Ma (Lynch Alfaro et al. 2012a). This taxonomic distinction was also supported by morphological and behavioral evidence (Lynch Alfaro et al. 2012b).

1.2.1 Species overview: The white-faced capuchin monkey

White-faced capuchins are a good primate model for addressing questions regarding variation in nutrition and energetics in relation to ecological and social pressures. They form cohesive female philopatric groups composed of multiple males and females (Fedigan 1993). As frugivore-insectivores (Fragaszy et al. 2004), they compete directly over fruit (Vogel 2005), a high-quality, clumped and monopolizable food resource (Isbell 1991), but also show dietary

flexibility by altering foraging behavior to exploit alternate food resources. Their habitats are often seasonal (i.e., distinct rainy and dry seasons) creating fluctuation in abundance, distribution and nutritional quality of resources.

The SSR population inhabits a highly seasonal environment and shows considerable variation in diet and ranging patterns during an annual cycle (Campos et al. 2014; Chapman and

Fedigan 1990; McCabe 2005; Melin et al. 2014b; Mosdossy et al. In Review). This temporal resource variation may affect the level of direct competition (i.e., win/loss contests) and the type of dominance relationships formed among females (Vogel et al. 2007). Female dominance rank may influence food acquisition and energy balance. Females in SSR exhibit linear and stable dominance hierarchies but show both competitive and cooperative social behaviors (Bergstrom and Fedigan 2010; Bergstrom and Fedigan 2013). The extent of rank-related differences in nutritional intake and physiological correlates is unknown. Reproductive events fluctuate with seasonal variation in food availability and females have been classified as moderately seasonal

Income II breeders whereby the timing of births occur prior to the mean peak in food abundance

(Carnegie et al. 2011a; Di Bitetti and Janson 2001; Janson and Verdolin 2005), but the role of nutrition and energy balance in the timing and success of reproduction is unknown. These characteristics make capuchins an ideal species in which to investigate how proximate ecological and social factors affect physiological processes and ultimately reproductive success among females. Results may be compared among Cebus and Sapajus species to better understand the relationship between diet and the evolution of morphological and functional differences.

1.2.2 Study subjects and data collection

1.2.2.1 Data collection schedule

Over a period of 20 months between August 2009 and May 2011, I observed three social groups of white-faced capuchins: Los Valles (LV), Cerco de Piedras (CP) and Guanacaste (GN). This period was divided into three 4-month field seasons (Season 1: September – December 2009;

Season 2: May – August 2010; Season 3: January – April 2011) totaling 12 months of data collection. This schedule allowed me to sample during each season (late rainy, early rainy and dry, respectively) and each month of the annual cycle to ensure the data encompassed all seasonal variation for comparative purposes.

The average annual home range size for capuchin groups in this study population is

197.768  12.518 hectares (Campos et al. 2014). Figure 1.5 depicts the home ranges of each study group, which were constructed using a movement-based kernel method (Benhamou 2011) and location points collected at 30-minute intervals during this study. The study groups are located within a few kilometers of the SSR park administration area. An extensive trail system has been established and is maintained within the home ranges of each study group to aid in access to each group by researchers.

With the aid of 1-2 field assistants per season, I followed each group for 4-6 day rotations per month (depending on group size) from dawn until dusk (5:00 am to 6:00 pm), when possible.

We collected data during 188 contact days across all seasons (see Appendix A for group contact information). I focused data collection on four key components: 1) measurement of ecological factors (fruit abundance and distribution), 2) behavioral observation (foraging, ranging and dominance behaviors), 3) nutritional composition of food items consumed, and 4) collection of urine samples for measurement of C-peptide.

The size of the study groups ranged from 20 to 37 individuals (LV: N = 20-23; CP: N =

26-33; GN: N = 33-37). Details of the demographics for study subjects within each of the three groups are shown in Appendix B. Individuals were identified based on physical features such as facial structure, hair coloration, skin pigmentation and scarring.

1.2.2.2 Female study subjects

I focused data collection on adult females (Appendix C) because they remain in their natal groups, allowing for long-term analysis, and because their reproductive success is more affected by access to food than that of males (Wrangham 1980). To determine which females had reached reproductive maturation, I used a subset of primiparous and multiparous females with known birthdates (N = 23). I calculated the average age at first conception (푥̅ = 6.10 years, s =

0.58, SE = 0.12) based on female age at first birth (푥̅ = 6.53 years, s = 0.58, SE = 0.12) and the length of female gestation (푥̅ = 157.83 days  8.13) (Carnegie et al. 2011b). Based on these findings, at the start of each season I included as study subjects all females that had already or would reached 6 years of age during that study season as well as females that had conceived at an earlier age. This method of classification includes females as subjects at least one month prior to the average age at first conception. Females were not included as study subjects if they disappeared or died before one 4-month field season was complete. The total number of adult females included in my analyses was 25, but the number of study subjects ranged from 24 to 25 during the study period (LV: N = 5; CP: N = 10; GN: N = 9-10). Table 1.1 lists pertinent information for all female study subjects.

Table 1.1 Details of female study subjects. Name, female ID, group membership and date of birth, as well as age class, parity and rank are listed by study period (i.e., field season 1, 2 and 3).

Age Class Parity Rank Female Group D.O.B 1 2 3 1 2 3 1 2 3 Kathy Lee LV 4/1/1989a A A A MP MP MP 1 1 1 Dos Leches LV 5/29/1991a A A A MP MP MP 2 2 2 Salsa LV 3/1/1996a A A A MP MP MP 3 3 3 Chutney LV 8/2/1999a A A A MP MP MP 4 4 4 Pickles LV 4/15/2003a A A A PP PP PP 5 5 5 Seria CP 6/10/1989a A A A MP MP MP 8 8 8 Timone CP 5/16/1996a A A A MP MP MP 6 6 6 Simba CP 8/5/1998a A A A MP MP MP 1 1 1 Zazu CP 2/2/1999a A A A MP MP MP 9 9 9 Ed CP 5/14/2000a A A A MP MP MP 3 4 4 Sarabi CP 1/1/2001a A A A MP MP MP 2 2 2 Kiara CP 4/29/2002a A A A PP PP PP 5 5 5 Shanti CP 3/5/2003a A A A PP PP PP 4 3 3 Baloo CP 5/3/2003a A A A PP PP PP 7 7 7 Nemo CP 3/11/2004a A A A NP PP PP 10 10 10 Minerva GN 1/1/1987c A A A MP MP MP 3 4 4 Luna Lovegood GN 1/1/1993c A A A MP MP MP 4 5 5 Mrs Weasley GN 1/1/1995c A A A MP MP MP 7 8 8 Lily GN 1/1/1997c A A A MP MP MP 1 1 1 Petunia GN 6/23/1999b A A A MP MP MP 2 3 3 Fleur GN 09/21/1999b A A A MP MP MP 6 7 7 Rita Skeeter GN 1/19/2001b A A A PP PP PP 5 6 6 Lavender GN 12/12/2001b A A A MP MP MP 9 10 10 Padma GN 2/11/2003b A A A PP PP PP 8 9 9 Cho Chang GN 1/26/2004b J A A -- NP PP -- 2 2 Age: A = Adult; J = Juvenile Parity: NP = nulliparous; PP = primiparous; MP = multiparous -- Indicates that the female was excluded as a study subject during this period because she was not yet reproductively mature. Superscript: a) Known from birth/infancy, error < 1 yr; b) Known prior to first birth and estimated based on mean age at first birth, error < 1 yr; c) Known from adulthood and estimated based on size/physical features, error 1-5 yrs.

1.2.2.3 Data collection

I collected behavioral data in the form of 10-minute focal animal samples (Altmann 1974; see ethogram, Appendix D). These data were recorded by two people to ensure accuracy of data collection given quickly occurring social interactions; one person dictated behaviors while the other recorded data on a handheld computer (Psion Workabout MX). I also recorded ad libitum behavioral interactions (e.g., intergroup encounters and female-female agonism). I opportunistically collected urine samples from all female study subjects. I conducted in-field urinalysis tests to measure urinary analytes (Siemens Multistix 10 SG Reagent Strips, Appendix

E) and later ran analyses to measure specific gravity, creatinine and C-peptide at the Hominoid

Reproductive Ecology Laboratory, University of New Mexico, under the supervision of Dr.

Melissa Emery Thompson. I recorded daily temperature using a Kestrel weather meter and daily rainfall using a standard cylindrical rain gauge. Each month I recorded the coverage and maturity of leaves, flowers and fruits for a subset of trees located along a phenology route that is part of long-term data collection at this field site (Melin et al. 2014a). Finally, I collected vegetative and invertebrate food items for nutritional analysis at Dairy One Forage Laboratory,

NY (Appendix F). Details regarding the measurement and calculation of each variable are explained in the respective methods and appendix sections.

1.3 Dissertation outline

The overarching goal of this dissertation is to use an interdisciplinary approach to investigate the relationships among behavior, nutrition and energetics in female white-faced capuchin monkeys living in a highly seasonal habitat. Below, I outline the specific aims of each chapter.

In Chapter 2, I document the foraging behavior and dietary profile of female white-faced capuchins from three habituated study groups within Sector Santa Rosa. I quantify the nutritional composition of different types of capuchin foods and assess the relationship between foraging behavior and nutritional intake. I compare the profitability of different food types based on energy and macronutrient intake rates. I estimate the nutritional requirements of female capuchins based on laboratory research and primate-wide nutritional standards to determine whether females are meeting these requirements in the face of seasonal variation in fruit abundance.

In Chapter 3, I assess group-level differences in monthly feeding behavior and nutritional intake with respect to home range quality and fruit abundance. Then, I explore the role of invertebrate food resources in the diet of female capuchins in greater detail. First, I identify which invertebrate groups contribute most heavily to the dietary profile in terms of items consumed and energy intake. Second, I assess the temporal importance of different invertebrate food types. Finally, I describe how seasonal variation in feeding corresponds to overall metabolic and macronutrient requirements specifically related to the dietary role of the most important invertebrate food sources.

In Chapter 4, I measure urinary creatinine and specific gravity to assess variation in the relative muscle mass of females. Creatinine is derived in muscle, and the relationship between urinary creatinine excretion and urine density (specific gravity) has been used as an estimate of relative muscle mass in wild primates (Emery Thompson et al. 2012). I confirm the relationship between creatinine and specific gravity (i.e., the ratio of the density of the urine relative to the density of water) in capuchin urine. I also evaluate whether semi-quantitative specific gravity readings from urinalysis test strips are equivalent to specific gravity determinations by a

refractometer. As the focus, I assess the effect of seasonal changes in ripe fruit abundance and group membership on the relative muscle mass of females. This chapter evaluates the application to capuchins of a recently proposed method for estimating broad-scale changes in the physical condition of females using urine as a biological medium and relatively inexpensive measurement of the urinary analytes, creatinine and specific gravity.

In Chapter 5, I investigate variation in the physical condition of females on a finer scale by measuring variation in female energy balance. The first goal of this chapter is to assess the strength of urinary C-peptide, a physiological biomarker for energy balance, as an analytical tool in the study of wild white-faced capuchin monkeys. I calculate the difference between the energy intake and energy expenditure of each female using behavioral observation and nutritional analysis, and compare these calculated measures to urinary C-peptide values. I also determine if variation in C-peptide predicts the production of urinary ketones (i.e., ketonuria).

Ketonuria is an indication that an individual is catabolizing fat stores, which is a more crude indication of negative energy balance (Laffel 1999; Soskin and Levine 1944). The second goal of this chapter is to determine which ecological and social variables affect energy balance.

Specifically, I use statistical modeling to assess the strength of climate, fruit abundance, reproductive state and dominance rank as ecological, intrinsic and social predictors of variation in energy balance.

In sum, the broad aim of this dissertation is to better understand how capuchins cope with seasonal variation in food abundance, focusing on the dietary changes and metabolic consequences associated with low food periods. Measuring the diet and physical condition of females will help us to assess in greater detail the relationships among diet, nutritional intake and physiological responses that lead to long-term variation in reproductive success.

Chapter Two: Dietary profile, food composition, and nutritional intake

2.1 Introduction

Variation in the spatial and temporal abundance of food resources presents a number of challenges related to energy and nutrient availability that may ultimately affect the reproductive success of mammals living in seasonal habitats (Bronson 1985; Clutton-Brock et al. 1982;

Clutton-Brock and Harvey 1977). Gaining access to food resources necessary for important biological processes such as growth, maintenance of body condition and reproduction greatly influences the survival and reproductive success of mammals, including primates (Altmann

1998). To meet their nutritional goals, primate species facing seasonal challenges may therefore attempt to: maximize overall energy intake; maximize specific nutrient intake (e.g., protein); avoid secondary plant metabolites (e.g., tannins); minimize the intake of factors that inhibit efficient digestion (e.g., fiber); and/or balance the intake of nutrients or modify behavior to reduce energy expenditure (reviewed by Felton et al. 2009). If unsuccessful, consequences include weight loss, decelerated growth and development, decreased reproductive output, and increased mortality or population-level changes in density and distribution (Altmann 1991;

Altmann 1998; Kay et al. 1997; van Schaik et al. 1993).

In addition to ecological variation in nutrient abundance, there is considerable variation in nutritional demands related to morphology that impact the growth and development of individuals as well as the care of offspring and reproductive success across an individual’s lifespan. Males and females often have dietary differences due to added metabolic demands as a consequence of the larger body size of males in sexually dimorphic species, and of energetically demanding reproductive states (i.e., late gestation and early lactation) for females (Bell 1971;

Clutton-Brock et al. 1982). Nutritional intake and body condition can significantly affect female lifetime reproductive success by influencing the timing of reproductive maturation as well as the persistence of ovulatory cycling at regular intervals. Previous studies of human and non-human primates have shown that females who consume a higher quality diet and maintain better body condition, or time reproduction with food abundance, give birth at an earlier age and/or exhibit higher rates of reproduction and infant survival than do other females [e.g., reviewed in humans and apes: (Bentley 1999); reviewed in humans: (Voland 1998); marmosets: (Tardif and Jaquish

1997); tamarins: (Miller et al. 2006); mangabeys: McCabe and Emery Thompson, 2013].

Maternal nutrient intake during gestation impacts fetal development (reviewed by Hinde and

Milligan, 2011), with consequences as extreme as impairment of fetal cerebral development due to maternal nutrient restriction [baboons, Papio spp.: (Antonow-Schlorke et al. 2011)].

Nutritional intake by primates during infancy can influence the secretion of hormones and growth factors, also leading to long-term effects on growth, metabolism and susceptibility to disease (Lucas 1998; Mott et al. 1990; Mott et al. 1991), which may ultimately impact reproductive success (Altmann 1991).

I aim to investigate seasonal variation in foraging behavior, the dietary profile and nutritional intake in a wild population of white-faced capuchin monkeys (Cebus capucinus).

Capuchins are arboreal monkeys that make use of multiple forest strata to exploit a wide diversity of resources including fruit (including seeds, grasses, bromeliads and arils), invertebrates, flowers, pith and vertebrates (Chapman and Fedigan 1990; Fragaszy et al. 2004).

They are broadly categorized as omnivorous in that they consume foods from multiple trophic levels, and they are specifically categorized as frugivore-faunivores since fruit and invertebrates make up the majority of their diet (Fragaszy et al. 2004). Capuchins are sexually dimorphic in

body mass (Cebus spp. males = 3.1 kg, females = 2.3 kg; Ford and Davis 1992), and sex has been shown to be a strong predictor of differences in foraging activities and of the types of foods ingested. In some capuchin species, males have been reported to forage more on animal matter, and females to focus more of their foraging efforts on plant-based food items (e.g., wedge- capped capuchins, C. olivaceus, Fragaszy and Boinski, 1995). In other capuchin species, males and females differ in foraging strategies; for example, white-faced capuchin males spent more time foraging for invertebrates on the ground, whereas females spent more time foraging for embedded invertebrates (Melin et al. 2010).

To better understand the relationship between foraging behavior and diet, it is important to estimate species-specific nutritional demands for maintenance, reproduction and growth.

Unfortunately, species-specific values and information on metabolizable energy (i.e., gross energy per food item minus the undigested energy lost in the feces, urine and through combustible gases) are unknown for most non-human primate species (National Research

Council 2003). The National Research Council (2003) reported mean adequate macronutrient concentrations for primate foods based on a large body of research. These values are based on data collected from many primate taxa, the diets of a variety of mammalian species, and studies of maintenance, reproduction and growth. As a percent concentration of the total diet, nutrient requirements for non-human primates are reported as 15-22% crude protein and 10-30% neutral detergent fiber (National Research Council 2003). Protein requirements for adult maintenance of body weight were determined for captive adult male C. albifrons weighing 2.0 kg as 1.8 g/kg/day protein, or 7.5% of the calories consumed and 7.1% of the dry matter consumed (Ausman et al.

1986; Ausman and Hegsted 1980). However, that study was based on high-quality protein

(lactalbumin), and likely underestimates the amount needed from naturally occurring and less metabolizable plant and animal proteins.

Protein requirements for gestating and lactating females have not been determined for most non-human primates (National Research Council 2003; Oftedal et al. 1991). In humans, protein requirements for adult females during pregnancy may be as high as those for the developing fetus due to the expansion of reproductive organs and her blood supply as well as the amino acids required for fetal development and growth (Pellett 1990). The requirement during pregnancy is approximately 10 g/day or 1.2 times higher than cycling females, whereas the requirement for lactating females is approximately 20 g/day or 1.4 times higher than cycling females (Pellett 1990). The caloric content of human milk (0.71 kcal/g) is lower than the mean for the milk of Old World monkeys (0.79 kcal/g), similar to the mean for New World monkeys

(0.71 kcal/g), but about 1.25 times lower than estimates for captive Cebus apella (0.89 kcal/g)

(Hinde and Milligan 2011; Milligan 2010). Overall, protein requirements are likely higher for wild adult Cebus capucinus based on the larger mean body size for adult female C. capucinus in the wild (2.54 kg; Smith and Jungers, 1997), and the added protein demands during gestation and lactation, including producing a milk that is more caloric than human milk (Hinde and Milligan

2011). Additionally, naturally occurring protein sources are of lower digestibility compared to man-made (purified and semi-purified) protein sources on which captive requirements are based.

By estimating nutritional requirements and determining species-specific patterns of temporal variation in feeding behavior and nutritional intake, we can better understand variation in reproductive success in primates living in seasonal habitats. However, a number of factors complicate a simple presentation of how primates use food resources to meet nutritional needs, including seasonality of resources, variation in dietary food types, variation in intake rates as a

result of their distribution and required processing, and differences in the energy and macronutrient availability across food types (Oftedal et al. 1991). It is therefore important to consider these confounding factors and directly assess the relationships among diet composition, nutritional intake and behavior rather than making assumptions regarding diet based on feeding and foraging behavior alone. Here, I investigate the nutritional ecology of white-faced capuchin monkeys in Sector Santa Rosa, Costa Rica, by measuring the foraging behavior, dietary profile, nutritional composition of food items and nutrient intake with respect to their estimated nutritional requirements. I do so to address the following series of research questions:

1) What percentage of total foraging time do females spend foraging for different types of

food (fruits, flowers, invertebrates, pith and vertebrates)? What is the frequency of

consumption of different food types? Is foraging time representative of energy intake?

2) Does the energy density and macronutrient composition of capuchin foods differ across

food types (i.e., fruit, seeds, flowers, caterpillars and other invertebrates)?

3) Do females consume energy and macronutrients at different rates depending on the type

of food eaten?

4) Do female capuchins meet estimated energy intake and macronutrient requirements?

How does feeding behavior, the contribution of different food types to the diet, and

energy and macronutrient intake differ according to variation in fruit abundance?

By increasing our understanding of the feeding behavior and nutritional composition of foods eaten by white-faced capuchins, we can determine whether they are meeting nutritional requirements, gain species-specific insight into the range of dietary flexibility exhibited by females, and identify limiting factors that may influence how they meet their nutritional goals.

Knowledge of feeding behavior and nutrition in wild populations may also help to inform captive

management practices. More broadly, these data are valuable for making cross-species comparisons linking feeding ecology to reproductive success and patterns of sociality.

2.2 Methods

2.2.1 Study Site

I conducted this study in the highly seasonal tropical dry forest of Sector Santa Rosa (SSR), in the Área de Conservaciόn Guanacaste, northwestern Costa Rica. The rainy season extends from mid-May until mid-November. The mean annual rainfall based on daily records collected during the study period (2009-2011) was 2304 mm (SD = 737) and the temperature ranged from a mean minimum of 22.9º C (SD = 0.1) in the wet season to a mean maximum of 29.6º C (SD = 0.8) during the dry season.

2.2.2 Study Subjects

I observed three study groups, LV, CP and GN, over a period of 20 months between September

2009 and May 2011. I collected data during three 4-month periods (Sep – Dec 2009, May – Aug

2010, and Jan – Apr 2011) to account for seasonal variation. I followed study groups on a rotational basis for four to six days per group, totaling three weeks of data per month (575.14 focal hours (LV = 120.09, CP = 227.06, GN = 227.99, Appendix C); 2,124.43 hours of observational contact; Appendix A). I report the details of the groups’ compositions in

Appendix B. I focused data collection on adult females. I classified females ≥ 6 years old as adults according to the average age at first conception (Carnegie et al. 2011b), unless they conceived before that time, in which case they were included as an adult at the beginning of the data collection season during which they gave birth to their first infant. I did not include females that disappeared or died during the study (N = 3) as study subjects due to lack of data. The

number of female study subjects ranged from 24 to 25 (LV: N = 5; CP: N = 10; GN: N = 9-10) per season due to the maturation and inclusion of a female in GN group for the 2010 and 2011 field seasons (see Table 1.1).

2.2.3 Reproductive state

I broadly categorized female reproductive state per month (cycling, gestating and lactating).

Based on the average length of female gestation for this study population, I classified females as

“pregnant” beginning 158 days prior to the date their infant was born until parturition (Carnegie et al. 2011b). Although females can be confidently classified as “cycling” one month preceding conception, it is difficult to determine whether a female is cycling during prior months without data on reproductive hormones. Furthermore, during my study I occasionally observed infants nursing from females assumed to be cycling (one month prior to the date of conception) or that I knew to be gestating, making the cut-off between the presence and absence of nursing an unreliable reference for inferring reproductive state. Therefore, I classified females as either

“lactating” or “cycling” based on observed drops in infant nursing rates under the assumption that females are able to cycle and gestate while receiving a low level of infant nursing.

Specifically, I classified a female as “lactating” from parturition until the infant stopped nursing completely, or infant nursing rates consistently dropped below one bout per hour based on the observation that when nursing did occur during pregnancy, it was below this rate. I classified a female as “cycling” during the period between lactation and gestation. It is important to note that, using this method of classification, I assumed that a female resumed cycling as soon as lactation dropped below one bout per hour, and thus I likely included a small proportion of females that were not gestating, lactating or cycling (Carnegie et al. 2005). Finally, I assigned a monthly energy coefficient of 1.00 for cycling females, 1.25 for pregnant females and 1.50 for

lactating females based on the estimated energy demands of each reproductive state. I averaged monthly values across females to estimate the mean reproductive demand for females during this study.

2.2.4 Estimating capuchin nutritional requirements

I estimated the nutritional requirements of capuchins based on the metabolic demands determined for mammals as well as published studies of the nutritional requirements for maintenance and growth of captive capuchins. Mean energy intake requirements for maintenance and growth can be estimated for animals by measuring energy intake, energy expenditure and changes in body mass, and have been documented for captive primates

(reviewed by the National Research Council 2003). The energy intake for maintenance in adult male capuchins (Cebus albifrons, body weight 1.5 – 3.0 kg) was estimated as 397.5 kJ/kg/day

(range 251 – 523 kJ/kg/day) (Ausman and Hegsted 1980). Accordingly, the estimated intake requirement for Cebus capucinus females is roughly 1000 kJ per day, if intake is proportional to the change in body size (Ausman and Hegsted 1980). These overall energy intake requirements were estimated to be 1.25 times higher in pregnant females (1250 kJ/day) and 1.50 times higher in lactating females (1500 kJ/day), due to the additional energetic demands of these reproductive states (Key and Ross 1999). I estimated that protein requirements for female capuchins are 1.8 g/kg/day for cycling females, 2.25 g/kg/day for gestating females and 2.7 g/kg/day for lactating females. These values were based on the requirements reported for captive C. albifrons

(Ausman et al. 1986; Ausman and Hegsted 1980) to which I applied the same coefficients for added reproductive demands as was done for energy intake (1.25 and 1.50, respectively) using the estimated reproductive requirements from human studies plus the added demands of producing higher quality milk as justification.

2.2.5 Behavioral Data Collection

2.2.5.1 Focal follows

I collected behavioral data in the form of 10-minute focal animal follows and ad libitum sampling (Altmann 1974). Two people often recorded behavioral data to ensure accuracy of data collection given quickly occurring social interactions; I dictated behaviors while a field assistant recorded data on a handheld computer (PSION Workabout hand-held computer). To observe focal females I followed a random rotation outlined in Bergstrom and Fedigan (2010) and following Perry (1996). During follows, I continuously recorded the general state behaviors of the focal female (e.g., travel, forage, rest, feed, social and solitary) and detailed social and feeding event behaviors (e.g., aggression, submission, affiliation, sexual and foraging behavior,

Appendix D).

2.2.5.2 Activity budgets

I used information gathered during focal follows to calculate time budgets as well as interaction and ingestion rates for each female. Using the continuous state behavior data to construct general time budgets, I grouped together specific behavioral states recorded during focal follows to calculate the mean percentage of time all females spent in six broad behavioral states [forage, social, rest, travel and other (which included self-direct, vigilant and allospecific association)].

Accordingly, I summed the amount of time spent in each category, divided by the total amount of time that all females were observed across the entire study period, and then multiplied this value by 100. To construct foraging time budgets, I determined the percentage of time females spent foraging on different types of resources (invertebrates, fruit, flowers, pith, small vertebrates and water/other) as well as visually foraging (whereby females visually searched trees without continuously handling substrates such as bark or leaves) by summing the total amount of time

spent in each category, dividing by the total amount of time observed foraging, and multiplying by 100.

2.2.5.3 Ingestion rates: food items, energy, and macronutrients

I calculated ingestion rates as the number of food items eaten per minute of observed feeding time for the two most commonly consumed food types, fruit (including seeds) and invertebrates

(McCabe and Fedigan 2007; O'Malley and Fedigan 2005). I only included feeding bouts during which specific food items were targeted in these calculations. Fruit species or invertebrate groups for which less than 5 items, 2 bouts or 10 minutes of total feeding time were observed were excluded from the dataset as these items were eaten too infrequently to accurately quantify ingestion rates. Food characteristics (e.g., size and abundance) greatly differ across species and affect variation in the rate at which each species can be ingested. For example, larger-sized food items take longer to ingest, on average, and are consequently eaten at slower rates, and species that are more spatially clumped and abundant may be located and ingested more rapidly.

Accordingly, I calculated energy and macronutrient intake rates to compare the importance of different food types in terms of the nutrients ingested.

2.2.6 Ecological Data Collection

2.2.6.1 Nutritional sample collection

I collected samples from plant species (including fruit, seeds, flowers and pith) and broad categories of invertebrates (single species or groups of related species) consumed by female capuchins in the field, and I processed and dried them in the laboratory at the field station.

Sample size per species ranged from five to thousands of specimens, depending on size, to achieve a total dry weight of at least 16 grams for analysis. If sample collection was not possible on the day the species was observed to be eaten, my assistants or I collected samples later ( 3

days for fruit and flowers) from as many of the locations in which foraging was observed as possible. Invertebrate samples were an exception; I collected those samples during a season in

May 2013 dedicated to this analysis. Capuchins often consume specific parts of food items

(shell, flesh and seed); therefore, when processing fruits with multiple parts, I separated the shell from the flesh and cut away the pulp from the seed. I only analyzed the part of the food item that was consumed by the capuchins for nutritional content unless separation of components resulted in loss of the item’s integrity and composition (e.g., water content). Subsequently, I dehydrated the samples at 30 °C using a food dehydrator (Nesco American Harvest Gardenmaster Pro,

Model FD-1020) and stored them in airtight waterproof bags with silica until exported for analysis.

I could not always taxonomically identify invertebrates on a fine scale. Consequently, I grouped invertebrate samples into the following broader categories for nutritional analysis: peppered roaches (Archimandrita tesselata), cicadas (Fidicina mannifera), shield bugs

(Pentatomidae), ants (Hymenoptera), Satellite sphinx caterpillars (Eumorpha satellitia, mean wet mass = 1.47 g), medium-sized noctuid caterpillars (Euscirrhopterus poeyi and Gerra

Hallowach01, mean wet mass = 0.26 g), medium-sized caterpillars from various families

(Lepidoptera), small-sized caterpillars from various families (Lepidoptera, mean wet mass = 0.05 g), jumping bean moth larvae (Cydia deshaisiana), crickets (Gryllidae), grasshoppers and katydids (Caelifera), wasp larvae (Polistes), scorpions (Centruroides limbatus), and a bulk category that included unidentified small non-caterpillar invertebrates. Similarly to plant items, I collected, flash-froze, and dehydrated the samples at a slightly higher temperature of 42 °C using a hot air oven for 2-6 days, depending upon size and density, to avoid molding that may occur

when drying animal matter at slower rates (DSO-3000DF, Digisystem Laboratory Instruments

Inc.).

I stored individual samples in Ziplock bags labeled with the food species name, food part consumed, collection date, number of items, wet weight, and dry weight. For each data collection period, I combined all dried samples with silica gel in airtight, waterproof storage bags until analyzed. Samples were consistently monitored for contamination; any food items showing signs of fermentation or molding were discarded and new samples were collected and dried, since both of these processes change the chemical composition of the food (Conklin-Brittain et al. 2006; Harborne 1984). Due to the higher risk of contamination associated with dehydrated animal matter, I stored invertebrate samples in a temperature and humidity controlled room until

I exported the samples for analysis. All samples were transported to Dairy One Forage

Laboratory, New York, USA to measure the macronutrient composition (e.g., crude protein, crude fat, water soluble carbohydrates, neutral detergent fiber, total ash and organic matter) of each species and food type combination.

There may be intraspecific differences in the nutritional composition of plant foods across space and time (Chapman et al. 2003). As a result, the nutrient values of food found in some trees in which the capuchins foraged may have been different from others. There may also have been nutritional differences in food across forest types and group ranges, as well as inter-annual variation in food quality. I collected samples from multiple trees and the collection of these plant samples spanned the period in which each species was eaten. Thus, values should be more representative of a species average for the period of this study, rather than tied to a specific time and location. I consider the nutritional values subsequently obtained as applicable to this study, but caution should be used when using these values for broad application.

2.2.6.2 Laboratory nutritional sample analysis

Specifics regarding analytical procedures performed by Dairy One can be found in Appendix F.

Briefly, additional drying was done to accurately calculate total dry mass. Dry matter (DM) was determined by drying a subsample at 135 °C for two hours and weighing the sample after cooling in a desiccator (AOAC International 2012d). Moisture content was determined by first subtracting the final dry matter weight (grams) from the initial weight of the wet fruit sample collected in the field and then dividing that value by the total wet weight to obtain an overall percentage (Conklin-Brittain et al. 2006). Crude protein (CP) was determined by measuring total nitrogen using combustion analysis and a Leco FP-528 Nitrogen/Protein Analyzer and then multiplying by the nitrogen to protein conversion factor of 6.25 (AOAC International 2012b;

AOAC International 2012c; AOAC International 2012e; Leco Corporation ; Milton and Dintzis

1981). Crude fat (CF) was gravimetrically determined by ether extraction using anhydrous diethyl ether and a Soxtec HT6 System (AOAC International 2012a). Water soluble sugars

(WSC) were first extracted with water, followed by acid hydrolysis with sulfuric acid and a colorimetric reaction with potassium ferricyanide, and then determined using a Thermo

Scientific Genesys 10S Vis Spectrophotometer (Hall et al. 1999). Neutral detergent fiber (NDF) was determined using the detergent system of fiber analysis (Van Soest et al. 1991).

2.2.6.3 Calculating the energy density of food items

I used the per item macronutrient values determined by the laboratory nutritional analyses to calculate the total energy density of food items as described below. Energy density is an estimate of the total amount of digestible energy (kJ) per gram of dry matter. Mean gross energy concentrations for macronutrients determined using bomb calorimetry and published by the

National Research Council (2003) are known: 4.1 kcal for carbohydrates, 5.6 kcal for protein,

and 9.4 kcal for fat. In attempt to approximate the digestible portion of gross energy, physiologically available energy conversion factors assigned based on early studies of food digestibility (Merrill and Watt 1955) have been published for humans as 4 kcal for carbohydrates, 4 kcal for protein and 9 kcal for fat. Unfortunately, research on primates has not been done to quantify requirements for each nutritional component, as based on metabolizable energy (gross energy minus the undigested energy lost in fecal matter), which has been done for a number of domesticated animals (NRC 2003). I converted the values reported for humans from kilocalories to kilojoules (kJ) using the conversion factor of 4.184 (16.74, 16.74 and 37.66, respectively). I used Formula 1 to calculate the energy density (kJ per gram dry matter) of each food item, where CP is the proportion of dry mass as crude protein, WSC is the proportion of dry mass as water-soluble carbohydrates, and CF is the proportion of dry mass as crude fat (Janson

1985; National Research Council 2003). I multiplied energy density by the dry mass per item to calculate the energy per food item (kJ/item).

Formula 1

퐸푛푒푟𝑔푦 (푘퐽) = (16.74 × (퐶푃 + 푊푆퐶)) + (37.66 × 퐶퐹)

I did not include fiber (measured in this study as neutral detergent fiber, NDF), in this calculation for a number of reasons. First, given their frugivorous-insectivorous diet, capuchin monkeys have a simple gastrointestinal system and a relatively short gut transit time of approximately 3.5 hours, which is comparable to other frugivorous platyrrhines (e.g., Ateles, 4.4 hours) but much shorter than that of folivorous platyrrhines (e.g., Alouatta, 20.4 hours) (Milton

1981). Because capuchins lack a specialized digestive system, it is unlikely that they are able to extract significant amounts of energy from difficult to digest plant materials such as fiber.

Regarding invertebrate consumption, crude protein is likely a more accurate estimate of the energy available from chitin than is fiber for species, like capuchins, that produce chitinase

(Finke 2007).

I collected and analyzed a total of 53 plant-based food items and 10 invertebrate categories during this project. For species that I could not collect for nutritional analyses, I used published data on specimens collected from Sector Santa Rosa, Costa Rica, by McCabe (2005), collected from the nearby site of Lomas Barbudal Biological Reserve, Costa Rica, by Vogel (unpublished,

2004, 2005), and information for one fruit species published by the USDA (2014). I estimated energy values for unknown plant species or plant species for which I was unable to conduct nutritional analyses or use literature values. When possible, I used the energy values for congeners of species with unknown values (N = 1) or species with similar size and composition

(N = 2), as denoted in Appendices G and H. When using conger species was not possible, I used the median energy value (kJ/item) for fruit, the mean energy value for flowers and an assigned energy value for pith (1 kJ per1 inch unit). Similarly, I calculated energy and macronutrient values per item for two groups of invertebrates (shield bugs and crickets), which were not analyzed by Dairy One, using the mean macronutrient values for the same order of invertebrates

(Hemiptera and Orthoptera, respectively) multiplied by the dry mass per item (obtained during sample collection for this study). Overall, 76.4% of the ingested energy was measured using food items collected during this study and approximately 23.6% of the ingested energy was estimated using values from published studies, congener species or mean and median values for the broader food category. Although estimated values are likely closer to true values than if these food items had been excluded from calculations (i.e., assigned a value of zero), it is

important to note that these are not true nutritional values, as would have been reflected if they were collected during this study from exact locations and times at which they were consumed.

2.2.7 Fruit abundance: fruit biomass (kg/ha) and ripe fruit energy density (kJ/ha)

To determine the energy density of ripe fruit (kJ/ha) in the study area per monthly round of data collection, I first calculated the fruit abundance in terms of ripe fruit biomass (kg/ha) using phenological data in combination with biological transects, and then applied energy values

(kJ/gram wet mass) per fruit species. Monthly phenological data is collected as part of a long- term and on-going collaborative data collection project at SSR. I assessed the monthly fruit coverage and maturity values for data collected from January 2009 through December 2011 for approximately 8 individual trees (푥̅ = 7.93, SD = 2.46) for each of 30 fruit species important to the capuchin diet. I used a five-point index (0 = absent, 0-25% = 1, 25-50% = 2, 50-75% = 3, and 75-100% = 4) to assess the score for the percentage of fruit coverage (C) and the score for the percentage of mature fruit (M) (Melin Meachem 2011). These species represented 44%

(9,394 of 21,347) of the fruit ingestion events recorded during focal observations. Many of the capuchin food species missing from this list were either wind dispersed species (e.g., Luehea candida and L. speciosa), for which the Peters (1988) equation (Formula 3) is not likely to provide an accurate fruit biomass estimate, or species for which it is difficult to gain accurate phenological information such as vines, shrubs, palms or bromeliads. I calculated a combined index score (CI) for each tree (i) by multiplying the proportion coverage index (C/4), by the proportion maturity index (M/4) (Campos et al. 2014). Using this combined index, trees were assigned the minimum score of 0 when the coverage or maturity indices were 0, and trees

received the maximum score of 1 when the coverage and maturity was considered 100%

(Formula 2).

Formula 2

C M Tree combined index score (CI ) = ( × ) 푖 4 4

Then, I calculated the mean monthly index (MI) for each species (s) as the mean combined index score for that species in that month. Tree abundance data were obtained from 151 botanical transects covering a total area of 3.02 hectares and distributed across all group home ranges (for details see Melin et al., 2014a). Following Peters (1988), for each tree in the transect, the biomass was calculated using Formula 3, where F is the estimated grams of fruit produced by a tree of a given DBH.

Formula 3

Tree fruit biomass (F) = 47 × DBH1.9

The total ripe fruit biomass (Bs) in kilograms per hectare (kg/ha) per month for each species was calculated as the sum of F divided by the sampled area, multiplied by 1000, to convert grams to kilograms, and then multiplied by the MIs to obtain a fruit biomass score for each month

(Formula 4).

Formula 4

∑푛 퐹 Species ripe fruit biomass (B ) = MI × 1000 × ( 푖=1 푖) 푠 푠 3.02

The monthly total energy availability from ripe fruit (EA) in kilojoules per hectare (kJ/ha) was calculated as the sum of the species-specific fruit biomass score (Bs) multiplied by the species-

specific energy content Es (kJ/kg wet weight) for all fruit species (Formula 5). I categorized data collection months as high and low energy density based on these calculations.

Formula 5

푛

Total ripe fruit energy availability (EA) = ∑ B푠 × E푠 푠=1

2.2.8 Statistical analyses

To compare the dry and wet mass between ripe fruit and invertebrate food items, I performed independent-samples Mann-Whitney U tests ( = 0.05, 2-tailed) because group sizes were unequal and Kolmogorov-Smirnov tests confirmed that the data were not normally distributed. I performed Kruskal-Wallis tests with Dunn’s post-hoc tests to compare the nutritional composition among five food types (fruit, seed, flower, non-caterpillar invertebrate and caterpillar). This non-parametric test was used due to small sample size for some food categories and because the data were not normally distributed, even after transformation. I reported the standardized test statistic and adjusted significance (multiple comparisons) for post-hoc tests. To examine the effects of seasonal variation in fruit abundance on a number of feeding and intake variables, I ran linear mixed effects models, whereby I included ripe fruit energy density as the fixed effect (standardized as a unitless Z-score) and female ID as the random effect. I also included monthly means per female of a) the proportion of time spent foraging, b) the proportion of energy intake from fruit, c) mass intake rate (gDM/hr), d) energy intake (kJ/hr), e) protein intake (g/hr), f) fat intake (g/hr), g) water soluble carbohydrates (g/hr), and h) neutral detergent fiber (g/hr) as response variables, each in separate models. All statistical analyses were performed using SPSS 21.0 statistical software (SPSS Inc., Chicago, IL, USA).

2.3 Results

2.3.1 General activity and foraging budgets

The mean annual time budget for all general activity indicates that females in LV, CP and GN groups spent the majority of their time searching for or actively consuming food (60.4%) followed by resting (18.1%) and socializing (12.7%). There was considerable variation in general activity patterns across study months. Although females spent the most time foraging, versus performing other activities, for the majority of the annual cycle, they spent roughly the same amount of time foraging as resting in April, the month with the highest maximum temperature (33.6 °C) and high fruit abundance.

I also divided the total foraging budget into the percentage of time spent foraging (i.e., handling and ingesting) for the following foods: fruit, invertebrates, flowers, pith, small vertebrates, water/other (i.e., drinking water and geophagy) and visually foraging. Foraging on fruit (20.0%) and invertebrates (70.3%) represented the majority (90.3% total) of the average time spent foraging over the entire study period. The annual mean percentage of time spent foraging on each food type is shown in Figure 2.1. Both the percentage of total time spent foraging as well as the contribution of fruit and invertebrates to the foraging budget varied across months (Figure 2.2). Females spent the greatest percentage of time foraging on invertabrates during the early rainy season months of June and July, which coincided with a flush of caterpillars, whereas the peak in the percentage of time spent foraging on fruit occurred during the late dry season in March and April, which corresponded to a peak in fruit abundance.

Figure 2.2 Monthly foraging budget. The graph depicts the mean  SE percentage of time spent foraging on the four most common food types by all three study groups.

2.3.2 Dietary profile

The female white-faced capuchins in Sector Santa Rosa consumed fruit (including arils, bromeliads, seeds and grasses), flowers, pith, invertebrates, vertebrates and a small amount of other items (e.g., dirt) during the study. I identified 88 plant food items (fruit from multiple ripeness stages, flowers and pith) from 41 families, including fruit, seeds, bromeliads and grasses from 64 species, flowers from seven species and pith from three species. A small proportion of fruit, flowers and pith ingested were unidentified (N species = 2 fruit, 1 flower, 1 pith). It is estimated that these species comprised less than 1% of the energy consumed during this study.

The capuchins also consumed 29 identifiable types of invertebrates from 10 orders, including

Araneae, Blattodea, Coleoptera, Hemiptera, Hymenoptera, Lepidoptera, Odonata, Orthoptera,

Phasmatodea, and Scorpiones. Based on the subset of species analyzed for nutritional content

(Appendices G and H), fruit items were significantly larger (N = 55, range = 0.01 to 19.24 gDM, median = 0.30, 푥̅ = 1.51  0.46, mean rank = 36.89) than invertebrates (N = 13, range = 4.72 

10-3 to 2.29 gDM, median = 0.05, 푥̅ = 0.38  0.18, mean rank = 24.38) on a dry matter basis (Z =

-2.051, p = 0.040). The wet mass of ripe fruit was also greater (N = 55, range = 0.02 to 99.00 g, median = 0.84, 푥̅ = 6.52  2.39, mean rank = 36.64) than that of invertebrates (N = 13, range =

0.01 to 7.29 g, median = 0.26, 푥̅ = 1.29  0.56, mean rank = 25.46) (Z = -1.833, p = 0.067).

In Table 2.1, I present the importance of the top four food types in terms of the percentage a) time females spent foraging on each food type; b) contribution in terms of the number of items ingested; c) contribution in terms of dry matter ingested; and d) contribution in terms of the energy (kJ) to the total observed intake across all female study subjects in the three study groups. Females spent a large percentage of their time foraging on invertebrates.

Invertebrates also comprised the largest percentage of the annual diet in terms of the number of 36

items ingested followed by fruit, flowers, and pith. However, fruit made a notably higher contribution to the percentage of dry matter ingested and the overall energy intake than did invertebrates.

Table 2.1 Contribution of different food types to the annual foraging profile. Values are displayed as mean  SE for female white-faced capuchins at Sector Santa Rosa, Costa Rica. Dry matter Energy ingested Food Type Foraging time (%) Items ingested (%) ingested (%) (%) Fruit 19.96  0.20 30.19 ± 1.11 61.69 ± 1.93 57.58 ± 2.01 Invertebrate 70.34  0.22 66.56 ± 1.04 35.77 ± 1.87 39.06 ± 1.90 Flower 0.74  0.10 1.84 ± 0.36 1.20 ± 0.36 0.80 ± 0.28 Pith 1.12  0.15 1.49 ± 0.29 na 1.23 ± 0.27 The calculation of percentage dry matter ingested does not include pith, as I was unable to successfully measure dry mass per unit. Total energy ingested is only the sum of categories for which nutritional processing and estimation were possible, and excludes vertebrates and miscellaneous items.

2.3.3 Nutritional composition of different capuchin foods

Nutritional composition varied greatly among food types; values are summarized in Table 2.2.

Results described below are visually depicted in Figures 2.3 a, b, c, d, e, f, g and h. Water content significantly differed by food type (N = 81, H = 11.787, df = 4, p = 0.019). Flowers (푥̅ =

80.24%  3.76) and caterpillars (푥̅ = 79.84%  5.38) contained the most moisture and seeds contained the least amount of moisture (푥̅ = 40.83%  13.35). The difference in water content between caterpillars and seeds was significant (t = 2.957, p = 0.031). The percentage of crude protein significantly differed with respect to food type (N = 81, H = 43.704, df = 4, p <0.001).

The crude protein content of caterpillars (t = 3.790, p = 0.002) and non-caterpillar invertebrates

(t = -5.650, p <0.001) was significantly higher than fruit. Caterpillars contained higher levels of crude fat than did the other food types (푥̅ = 21.94%  10.13), and flowers contained the lowest levels (푥̅ = 3.94  2.22), although the difference in fat content across categories was not significant (N = 81, H= 7.719, df = 4, p = 0.102). The water-soluble carbohydrates (WSC)

significantly differed with respect to food type (N = 81, H = 33.833, df = 4, p < 0.001). Fruits contained significantly higher concentrations of water-soluble carbohydrates than did non- caterpillar invertebrates (t = 5.006, p < 0.001) or caterpillars (t = -2.911, p = 0.036). There were no significant differences in the percentage of neutral detergent fiber (NDF) across food types (N

= 80, H = 5.778, df = 4, p = 0.216), although caterpillars contained the lowest amount of fiber (푥̅

= 14.06  2.22), and both fruit (푥̅ = 30.66  2.50) and seeds (푥̅ = 30.62  8.09) contained the highest amount of fiber.

Finally, food types differed in their total energy content (kJ per gram dry matter) (N = 81,

H = 18.705, df = 4, p = 0.001). Non-caterpillar invertebrates contained significantly more energy per gram dry matter than did fruits (t = -2.983, p = 0.029). Although not significant, non- caterpillar invertebrates also contained more energy per gram dry matter than did flowers (t = -

2.687, p = 0.072), and caterpillars contained more energy on a dry matter basis than fruit (t =

2.746, p = 0.060) and flowers (t = 2.805, p = 0.050). When energy was assessed per gram wet mass to account for differences in water content, there were significant differences across food types (N = 81, H = 12.330, df = 4, p = 0.015). Non-caterpillar invertebrates contained more energy per gram wet mass than did flowers and fruit; however, these differences were not significant once corrections for multiple post-hoc tests were made (flowers: t = -2.678, p = 0.074 and fruit: t = -2.555, p = 0.106). When energy was assessed per item to account for differences in water content and size, there was a significant difference across food types (N = 81, H =

16.234, df = 4, p = 0.003). Specifically, fruit contained significantly more energy (kJ) per item than did seeds (t = 3.093, p = 0.020).

Table 2.2 Summary of macronutrient content and energy density per food type for food items eaten by adult females in the study groups. N Energy Food item species % H2O % CP % CF % WSC % NDF (kJ/gDM) Ripe fruit, seeds and grasses 58 66.21 ± 2.47 8.86 ± 0.72 12.46 ± 2.13 35.85 ± 3.20 30.46 ± 2.47 12.18 ± 0.71 Ripe fruit 53 68.60 ± 2.18 8.11 ± 0.64 12.06 ± 2.29 38.21 ± 3.30 30.45 ± 2.62 12.30 ± 0.75 Seeds and grasses 5 40.83 ± 13.35 16.89 ± 3.65 16.64 ± 4.95 10.79 ± 3.52 30.62 ± 8.09 10.90 ± 2.30 Flowers 3 80.24 ± 3.76 13.64 ± 3.13 3.94 ± 2.22 28.61 ± 8.65 21.25 ± 4.07 8.56 ± 0.84 Invertebrates 13 72.99 ± 2.73 64.92 ± 4.27 16.17 ± 4.01 5.79 ± 1.67 21.65 ± 2.55 17.93 ± 1.02 Non-caterpillar invertebrates 8 68.71 ± 1.89 67.25 ± 5.16 12.56 ± 1.82 5.50 ± 2.63 26.40 ± 2.82 16.91 ± 0.66 Caterpillars 5 79.84 ± 5.38 61.18 ± 7.87 21.94 ± 10.13 6.26 ± 1.55 14.06 ± 2.22 19.55 ± 2.41 CP, Crude protein; CF, Crude fat; WSC, water-soluble carbohydrates; NDF, neutral detergent fiber. Mean  SE macronutrient and energy values are listed as the percentage of dry mass. Species with estimated nutritional composition are excluded.

2.3.4 Dietary composition, intake rates and nutritional profitability

The mean annual diet of female capuchins as a percentage of dry matter ingested, including all food types, was comprised of 27.15% (SE = 1.00) crude protein, 9.93% (SE = 0.27) crude fat,

35.75% (SE = 1.48) water-soluble carbohydrates and 25.54% (SE = 0.07) neutral detergent fiber.

As a percentage of metabolizable energy, protein comprised 30.15% (SE = 1.11), fat comprised

25.32% (SE = 0.60) and water-soluble sugar comprised 43.36% (SE = 1.52). Item intake rates, energy intake rates and macronutrient intake rates for two food types eaten by female white- faced capuchins (fruit and seeds compared to invertebrates) are shown in Table 2.3 (for item- specific rates see Appendix I). Although the item intake rate for fruit was nearly 1.25 times higher than for invertebrates, the difference was not significant. However, the energy intake rate

(kJ/min) was significantly higher for fruits than invertebrates (Z = -3.868, p < 0.001). The difference in energy consumption was a result of the significantly higher nutrient intake rate

(g/hr) of water-soluble sugar (Z = -4.031, p < 0.001) and fat (Z = -2.949, p = 0.003) during fruit versus invertebrate consumption, as the difference in the intake of protein between fruit and invertebrates was not significant (Z = -1.136, p = 0.256).

Table 2.3 Intake rate and nutritional profitability of two food types eaten by females: fruit and seeds, and invertebrates. Food items included in calculations are a subset of the larger nutritional dataset for which the targeting of specific species while foraging allowed for measurement of foraging bout length. Macronutrient intake (protein, fat, water-soluble sugar and fiber) is measured in grams per hour. Food Type Variable Fruit and seeds Invertebrates Species (N) 44 8 Item intake rate (N/min) 4.82 ± 0.77 3.89 ± 1.15 Energy intake (kJ/min) 10.70 ± 1.36 1.75 ± 0.57 Protein (CP) 5.71 ± 0.85 3.58 ± 1.12 Fat (CF) 7.18 ± 1.34 0.84 ± 0.25 Sugar (WSC) 28.89 ± 5.13 0.77 ± 0.56 Fiber (NDF) 29.13 ± 5.68 1.27 ± 0.52

2.3.5 Estimated nutritional demands for female capuchins

Based on the mean number of females that were gestating and lactating during each month, the mean monthly reproductive demand for females during this study was 1.40 (SE = 0.01) times the requirements for a non-cycling female. Thus, the mean daily energy requirement for females, who spend a disproportionate amount of the annual cycle in a lactational state (mean weaning completion age = 20.5 mo., Fragaszy et al. 2004; mean interbirth interval = 26.36 mo., Fedigan and Rose 1995), was estimated as 1,400 kJ/day and the mean daily protein requirement was estimated as 2.52 g/kg/day.

2.3.6 Ripe fruit energy density

During this study, biannual peaks in energy density (kJ/ha) occurred from February – March and

September – October (Figure 2.4). Although fruiting patterns are variable inter-annually, these peaks coincided with the mean annual variation in fruit biomass collected between 2007 and

2013 at this field site (Campos et al. 2014). Mean ripe fruit energy density (kJ/ha) during the study period was 557,763  95,514 kJ/ha. The highest energy density occurred during February

2011 (1,066,184 kJ/ha), whereas the lowest occurred during December 2009 (74,030 kJ/ha).

1600000

1400000

1200000

1000000

800000

600000

400000

200000

Ripe Fruit Energy Density (kJ/ha) Density Energy Fruit Ripe 0

Jul-2009 Jul-2010 Jul-2011

Jan-2009 Jan-2010 Jan-2011

Jun-2009 Jun-2010 Jun-2011

Oct-2009 Oct-2010 Oct-2011

Sep-2011 Feb-2009 Sep-2009 Feb-2010 Sep-2010 Feb-2011

Apr-2009 Apr-2010 Apr-2011

Dec-2009 Dec-2010 Dec-2011

Mar-2009 Mar-2010 Mar-2011

Aug-2009 Nov-2009 Aug-2010 Nov-2010 Aug-2011 Nov-2011

May-2009 May-2010 May-2011 Month

2.3.7 Differences in feeding behavior and intake according to fruit abundance

Ripe fruit energy density was a significant predictor of the mean proportion of time spent feeding across study months (Table 2.4). There was a negative relationship between the two variables, whereby females spent a lower proportion of their total activity budget foraging when fruit abundance was high and a higher proportion of time feeding when fruit abundance was low

(Figure 2.5a, range = 0.139 – 0.956, mean monthly range = 0.343 – 0.778). As expected, ripe fruit energy density significantly predicted the proportion of energy intake from fruit (Table 2.4), whereby the more fruit that was available, the higher the proportion of fruit-based energy intake by females (Figure 2.5b, range = 0.000 – 0.990). Females consumed the highest average proportion of energy from fruit during April 2011 (푥̅ = 0.861  0.036) and the lowest proportion

of energy intake from fruit during May 2010 (푥̅ = 0.132  0.042). Neither the rate of food intake

(grams of dry matter per hour) nor the rate of total energy intake was significantly predicted by ripe fruit abundance (Table 2.4, Figure 2.5c and 2.5d). However, the lowest mean monthly food intake rate (푥̅ = 3.251 g/hr  0.433) and energy intake rate (푥̅ = 53.097 kJ/hr  6.270) did occur during June 2010, a month with very low ripe fruit abundance (256,209 kJ/ha). The highest food intake rate (18.451  3.336) and energy intake rate (233.329  39.669) occurred during a month with higher than the annual mean ripe fruit abundance in January 2011 (657,730 kJ/ha). Based on these data, mean daily (12-hour) energy intake is estimated to have reached as high as 2800 kJ/day and as low as 637 kJ/day. When placed in the context of the minimum estimated requirement of 1000 kJ/day, and the requirement of white-faced capuchin females adjusted for the reproductive demands of 1400 kJ/day (inferred from studies of captive capuchins by Ausman and Hegsted 1980), there is the potential for female capuchins to far exceed or fall short of daily energy intake requirements at this field site.

Table 2.4 Summary of linear mixed effects models analyzing the effect of monthly ripe fruit energy density (kJ/ha) on variation in mean feeding time and nutrient intake.

95% Confidence Interval Standard p Response variable Fixed effect Intercept Estimate Error df t value Lower Upper Feeding time (proportion) Energy density 0.573 -0.075 0.009 270.565 -8.065 <0.000 -0.094 -0.057 Proportion of intake from fruit Energy density 0.583 0.167 0.018 270.546 9.184 <0.000 0.131 0.202 Intake (gDM/hr) Energy density 10.856 0.768 0.573 294.000 1.340 0.181 -0.360 1.897 Energy intake (kJ/hr) Energy density 142.121 7.637 6.446 270.368 1.185 0.237 -5.054 20.329 Crude protein (g/hr) Energy density 1.832 -0.323 0.050 269.815 -6.475 <0.000 -0.421 -0.225 Crude fat (g/hr) Energy density 0.851 -0.001 0.046 294.000 -0.020 0.984 -0.091 0.090 Water soluble carbohydrates (g/hr) Energy density 4.656 0.834 0.300 270.375 2.775 0.006 0.242 1.425 Neutral detergent fiber (g/hr) Energy density 2.820 0.331 0.180 294.000 1.842 0.067 -0.023 0.685

a) b)

c) d)

2.3.8 Differences in macronutrient intake according to fruit abundance

Macronutrient intake varied substantially across study months (Table 2.4, Figure 2.6).

Specifically, ripe fruit energy density significantly predicted protein intake rate (g/hr, Table 2.4).

There was a negative relationship between the two variables, whereby females consumed protein at lower rates when fruit abundance was high and at higher rates when fruit abundance was low

(Figure 2.6a, mean monthly range = 1.015 – 2.511 g/hr). Ripe fruit energy density did not predict variation in fat intake or fiber intake (Table 2.4, Figure 2.6b and 2.6d). Finally, ripe fruit energy density significantly predicted the rate of water-soluble sugar intake (Table 2.4), whereby females consumed sugar at higher rates during months with higher ripe fruit energy density

(Figure 2.6c, mean monthly range = 0.681 – 9.300).

a) b)

c) d)

2.4 Discussion

The objectives of this study were to characterize the capuchin diet through assessment of the female foraging budget, dietary profile, and nutritional composition of different food types, as well as to evaluate temporal variation in feeding behavior, energy intake and macronutrient intake in light of nutritional requirements.

In answer to question 1 regarding activity and foraging budgets, I found that females spent over half of their time foraging and feeding, during which they consumed a diversity of food items from two key dietary categories, plants and invertebrates. Despite inter-annual variation in food abundance and diet, the food item diversity is comparable to previous studies with documented feeding behavior conducted at this field site (Chapman and Fedigan 1990;

MacKinnon 2006). The foraging budget is also comparable to previous studies at this site (e.g., females: 53%, Rose 1994a; all individuals: 53% Melin et al. 2009), as well as across Cebus and

Sapajus species (e.g., C. olivaceus: 43%, Fragaszy, 1990; C. albifrons: 54%, Matthews 2009, female median: 54.8% and 53.1%, van Schaik and van Noordwijk 1989; S. apella: female median: 48.2%, van Schaik and van Noordwijk 1989; S. libidinosus: 41.9%, Moura 2004 and

46% Izar et al. 2012; S. nigritis: 58%, Izar et al. 2012).

Although there is variation in diet across the annual cycle, fruit and invertebrates make the most significant contribution to capuchin foraging time, number of items ingested and overall energy gain across all months, as has been previously found in several capuchin studies

(Chapman and Fedigan 1990; Fragaszy et al. 2004; McCabe and Fedigan 2007; Melin et al.

2014b; Rose 1994). Invertebrates are arguably a more important food source when considering the amount of time that females spent foraging on them and the number of items ingested.

Females spent an average of 70% of their foraging time actively searching for and ingesting

invertebrates versus only 20% for fruit. These foraging budgets are comparable to other capuchin species (C. albifrons and C. apella) as well as more insectivorous common squirrel monkeys (Saimiri sciureus) and emperor tamarins (Saguinus imperator) (Terborgh 1983).

However, fruit contributes substantially more to the annual dry matter ingested than do invertebrates. Fruit also comprises a larger percentage of the total annual energy consumed by females than do invertebrates. Differences between foraging time and energetic contribution highlight the importance of including nutritional analyses in assessments and comparisons of diet and behavior, as well as considering the nutritional values of food items with respect to a food item’s size and moisture content rather than making comparisons of nutrients on a dry matter basis alone (Rothman et al. 2014b). While this insight seems rather obvious, due to lack of available nutritional data, a number of studies have based data analysis on the assumption that the amount of food eaten and/or the energy ingested is proportional to the time spent eating or the number of items ingested. This point is particularly important when making comparisons across food types with the largest size disparities (e.g., between fruit and invertebrates), since distribution, size, handling time and food processing all may affect energy and macronutrient intake rates (Hladik 1977; Milton 1984).

To address question 2, there was variation in the nutritional composition among different food types analyzed in this study (fruit, seed, flower, pith, non-caterpillar invertebrate, and caterpillar). Seeds contained significantly less moisture than other food types. On a dry matter basis, caterpillars and non-caterpillar invertebrates were higher in protein and overall energy (kJ) than fruit and seeds, although fruit and seeds contained higher levels of water-soluble sugars.

This pattern persisted when invertebrate foods and plant foods were compared as two broad categories. These values are expected given that invertebrates are primarily composed of

protein, fat and chitin (Rothman et al. 2014b). They are also comparable to other studies that have analyzed the nutritional composition of primate foods including those specific to platyrrhines (reviewed by Norconk et al. 2009).

The energy and nutritional composition of invertebrates may be considerably offset by their significantly smaller size when compared to fruits on a dry matter basis (Rothman et al.

2014b). The more dispersed distribution of invertebrates in the environment relative to the generally more highly clumped distribution of fruits produced by angiosperms may also greatly affect intake rate and profitability in terms of nutritional gain per unit time foraging. Although I was unable to quantify differences in food distribution across food types, energy and macronutrient intake rates give insight into the profitability of fruit versus invertebrates.

In answer to question 3 regarding variation in nutrient consumption rates by food type, females consumed fruit at a slightly higher rate than invertebrates on a per item basis; however, the energy intake rate (kJ/min) as well as water-soluble sugar and protein intake rates (g/hr), were significantly higher during fruit versus invertebrate foraging. It is therefore important to emphasize that the assumption that time spent foraging accurately reflects intake is not well justified, at least for capuchins, due to the large degree of variation in size and nutrient content across food items. Generally speaking, neotropical fruit may provide sufficient levels of dietary protein during high fruit abundance months based on nutritional composition. A recent study showed that the nitrogen content is significantly higher in neotropical fruits than fruits found in

Madagascar (Ganzhorn et al. 2009).

Finally, in answer to question 4 regarding whether females met estimated nutritional requirements, water-soluble carbohydrate, fat and fiber composition seemed comparable to other

New World primates, but protein intake was higher than other studies. For example, protein

intake was slightly higher than the range (15-22%) proposed for all primates by the National

Research Council (2003). However, protein intake greatly exceeded both the requirements determined for weight maintenance in captive male Cebus albifrons (7.1% dry matter and 7.5% metabolizable energy) and my estimated 9.0%, adjusted based on the assumption that plant- based proteins (versus lactalbumin used in the captive study) are of lower digestibility (Ausman et al. 1986; Ausman and Hegsted 1980). The protein composition was also higher than what has been reported for folivorous howler monkeys (Alouatta palliata, 9.6% DM, Hladik et al. 1971) and frugivorous spider monkeys (Ateles geoffroyi, 7.4% DM, Hladik et al. 1971; Ateles chamek,

12.5%, Felton et al. 2009b) but comparable to the much smaller-bodied and more insectivorous- gumnivorous tamarin (Saguinus geoffroyi, 20.6% DM, Hladik et al. 1971). I measured crude protein during this study, which likely represents more protein than what is metabolically available to the capuchins since non-protein nitrogen that may be bound in the form of secondary compounds is also included (Rothman et al. 2012); the difference between crude and available protein may account for some of the differences of protein intake described in this study. Fat intake was also most comparable to tamarins (S. geoffroyi, 9.1% DM) (Hladik et al. 1971). Fiber intake by females fell within the recommended range of 10-30% (National Research Council

2003) but was higher than the 19% (DM cellulose  2.5) neutral detergent fiber intake reported for another population of C. capucinus (Hladik et al. 1971); both sugar and fiber intake were most comparable to spider monkeys (A. geoffroyi,, 33.7% and 27.5% DM, respectively) (Hladik et al. 1971; National Research Council 2003). Given the dietary similarities to tamarins, it would be interesting to compare the capuchin diet to the more closely related squirrel monkey (genus

Saimiri); unfortunately, to my knowledge studies of the dietary composition of wild squirrel monkeys that quantify nutritional intake have not been done.

Like many other free-living animals, capuchins in Sector Santa Rosa face temporal variation in the abundance of food resources and exhibit variation in both foraging and intake patterns across study months. As expected, ripe fruit abundance significantly predicted both the overall proportion of time spent foraging and the proportion of energy intake from ripe fruit. As fruit abundance increased, females decreased the proportion of time they spent foraging and the amount of energy consumed from fruit increased. Although females appeared to meet estimated energy intake requirements (1,400 kJ/day, adjusted for the mean reproductive state of females) during some months, it is important to note that they did not meet estimated requirements during the month with the lowest intake rate. In Chapter 3, I assess variation in energy and macronutrient intake in more detail to advance our understanding of dietary variation when the abundance of fruit is low. This reduction in intake during low fruit months suggests that the smaller size and more widely distributed nature of invertebrate foods, as well as the need to travel further to find fruit, may have constrained the ability of capuchins to consume food at high rates during periods of low fruit abundance. Despite a reduction in the dry mass and overall energy ingested, the mean rate of protein intake was higher during months with low fruit abundance due to the significantly higher protein concentration of invertebrates compared to other food types. Based on these data, daily (12-hour) protein intake is estimated as 9.401 grams per body mass per day during high fruit months and 12.468 g/kg/day during low fruit months.

Again, this is much higher than the estimates for captive male C. albifrons (1.8 g/kg/day) required for weight maintenance, even once adjusted for the added estimated demands of gestation and lactation, regardless of fruit abundance. Fat intake remained relatively constant across months, and both water-soluble sugar and fiber intake decreased with fruit abundance.

Analysis of plant toxins was out of the scope of this project and therefore, I did not test the hypothesis that capuchin nutritional goals could be driven by the avoidance of secondary plant metabolites. Females did not seem to minimize energy expenditure as fruit abundance decreased, but instead increased the proportion of time spent foraging. When fruit abundance was high, females seemed to maximize water-soluble carbohydrates based on intake rates, but also seemingly exceeded protein requirements. In contrast, during periods of fruit scarcity, females increased the foraging of invertebrates and “over-consumed” protein to meet overall energy requirements. These patterns suggest that females are neither attempting to balance nutrient intake or protein-regulate but instead are attempting to maximize caloric intake in response to a decrease in fruit abundance. This pattern is similar to that seen in mountain gorillas (Gorilla beringei), whereby the gorillas balance the intake of protein and non-protein energy during periods when fruit is available, but over-ingest protein during time periods when leaves are the dominant source of energy and nutrients (Rothman et al. 2011).

This study provides important information regarding how seasonality affects the behavior and diet of female white-faced capuchins. There are a number of methodological considerations that may allow for a more refined assessment of nutritional intake and energy requirements in future studies. For example, I extrapolated time budgets and intake values from 10-minute focal animal follows. All-day follows would provide a complete measure of daily intake, as there may be variation in feeding behavior (e.g., bout length and rate) throughout the day. I also pooled many samples collected for each food item across the study period to measure nutritional composition or I used published values if I was unable to collect samples. Collection and separate nutritional analysis of many samples for each food item that represent all periods and locations in which consumption occurred would also increase the accuracy of intake values, as

this method would help to minimize the error introduced due to intraspecific variation in food composition across space and time (Chapman et al. 2003; Rothman et al. 2014a; Rothman et al.

2012).

In summary, I found that female white-faced capuchins at Sector Santa Rosa, Costa Rica, focused foraging efforts on fruit and invertebrate food items. Fruit contributed the most to the overall energy gain despite the greater proportion of time devoted to searching for and consuming invertebrates. Although the nutritional composition of food types is variable, fruits were the most important source of water-soluble sugar, whereas high proportions of protein intake came from invertebrates, particularly as fruit abundance decreased. Females were able to consume macronutrients at a much higher rate while foraging fruit, likely due to the larger size of food items compared to invertebrates. There was temporal variation in the types of foods consumed and in the ability of female capuchins to meet energy requirements, warranting a more detailed investigation into capuchin foraging patterns, variation in the utilization of other food types and nutritional intake during this study.

Chapter Three: The nutritional importance of invertebrates

3.1 Introduction

The negative relationship between primate body size and dietary specialization on high-quality resources that are easily digestible and low in secondary compounds is well established and reflects both the higher metabolic needs and lower ability to ferment low-quality foods exhibited by smaller animals (Gaulin 1979; Sailer et al. 1985). Conversely, large animals have absolutely higher nutritional needs and must consume large volumes of food (Kleiber 1961; Sailer et al.

1985). Invertebrate prey are generally rich in protein, fat and easily digested, but small in size, and require intensive effort to procure (Rothman et al. 2014b; Terborgh 1983). As a result, invertebrates do not provide enough mass and nutritional value per unit of consumer body weight and foraging effort to fully sustain primates with larger body sizes. While small-bodied primates are able to maintain a highly insectivorous diet, and medium-bodied primates show a range of dietary patterns that often incorporate a combination of fruit, animal matter and/or leaves, large- bodied primates are able to maintain highly folivorous diets and many possess anatomical specializations to process structural carbohydrates and detoxify plant secondary compounds

(Clutton-Brock and Harvey 1977; Gaulin 1979; Sailer et al. 1985). Despite differences in digestibility, both invertebrates and leaves provide an important source of protein for primates on the smaller and larger ends of the body size spectrum, respectively (Gaulin 1979; Kay 1975;

Richard 1985).

For primate species with a mean body mass that is too large to subsist on a completely insectivorous diet, invertebrate prey may still provide an important dietary component by complementing, augmenting or replacing other items in the diet with sources of protein, fat,

vitamins or minerals (reviewed by Rothman et al. 2014). It is possible that invertebrates play more than one nutritional role for primate species that live in tropical habitats since seasonal shifts in resource availability require temporal changes in foraging behavior as well as dietary modifications (Hemmingway and Bynum 2005). For example, during insect outbreaks (i.e., temporary increases in the abundance of a specific species) abundance may be high enough to increase consumption rates to a profitable level for larger-bodied primates that otherwise rely upon plant-based food resources (Rothman et al. 2014b). A seasonal landscape therefore provides a good environment in which to examine patterns of invertebrate prey consumption to better understand dietary flexibility and how omnivorous primates meet their nutritional requirements. Here, I use behavioral data on white-faced capuchins (Cebus capucinus), a neotropical monkey living in the highly seasonal tropical dry forest of Sector Santa Rosa, Costa

Rica, to assess variation in invertebrate consumption and seasonal importance with respect to their nutritional needs.

A recent comparative review of the protein concentrations of neotropical fruits versus those in Madagascar shows that the relatively higher protein concentrations in neotropical fruit may provide sufficient levels of protein to meet dietary requirements without the need to consume supplementary resources (Ganzhorn et al. 2009). Nutritional analyses performed on the dietary items consumed by the Santa Rosa capuchins (Chapter 2) show that on a dry matter basis, fruits eaten by capuchins are significantly lower in protein (range = 2.2 to 24.2%) than are invertebrates (range = 32.3 to 86.7%), although there is a considerable variation for both food types. This difference is reduced when fruit and invertebrates are compared per gram wet mass.

Among fruits, figs may be an especially important source of protein compared to the fruits of other neotropical angiosperms, as they often containing wasp larvae in addition to a high protein

versus other macronutrient ratio (Hladik et al. 1971; Valenta and Melin 2012). Fruit (including arils, bromeliads, seeds and grasses) makes up a large proportion of the capuchin monkey’s omnivorous diet, but on an annual basis, invertebrate consumption is important as well

(Fragaszy et al. 2004; McCabe and Fedigan 2007; Melin et al. 2007; Young 2005). Whether or not fruit provides sufficient protein to meet capuchin nutritional needs may depend on seasonal availability.

Based on the relationships between body size, metabolism and diet, and using anatomical evidence from fossil adapines, a threshold of 500 g (“Kay’s threshold”) has been used to predict the upper limit of body weight for primate species to maintain a primarily insectivorous diet

(Fleagle 1999; Gingerich 1979; Kay 1975; Kay 1980; Kay 1984). Based on average capuchin body mass (mean female = 2.3 kg and male = 3.1 kg, Ford and Davis 1992), an insectivorous diet alone would not provide enough energy to meet nutritional requirements. While capuchins are not exclusively insectivorous, previous studies have indicated that capuchin species consume a higher amount of insects than expected based on body mass and that invertebrates provide an important source of protein (Janson and Boinski 1992; Norconk et al. 2009; Terborgh 1983).

This, in part, may reflect the energetic demands of their proportionally large brains (Armstrong

1983; Fragaszy et al. 2004), and capuchin-specific adaptations, including modified gastrointestinal morphology as well as the presence of chitinase, allow them to efficiently extract energy from invertebrate prey (Ullrey et al. 2003).

Like many New World monkeys, capuchins glean invertebrates from substrates. They are also well-known for their unique cognitive ability and motor dexterity, which allows them to find and extract invertebrates embedded in plant and invertebrate-made structures that are not readily available to other sympatric insect predators (Janson and Boinski 1992; Melin et al. 2007).

Melin et al. (2014b) found that the extraction of embedded foods was highly seasonal and exploitation of this type of invertebrate resource by white-faced capuchins was negatively correlated to ripe fruit abundance, supporting the hypothesis that embedded insects are not only a key fallback food, but one that has likely played a role in driving the evolution of sensorimotor intelligence in capuchins.

Despite the greater prevalence of protein-rich fruit in the neotropics, in highly seasonal environments there may be considerable variation in the temporal availability and intake of important macronutrients. The timing of this availability in correspondence with energetically demanding reproductive states may affect how females meet their nutritional needs. The current study builds on past research on food consumption and seasonality by investigating the nutritional role of invertebrates as compared to fruit, as well as the importance of different orders of invertebrates to meeting energy and macronutrient requirements across the annual cycle in a seasonal habitat. These analyses will help to document the range of capuchin dietary flexibility as well as how they respond to seasonal changes in resources in the face of metabolic and reproductive demands.

In Chapter 2, I found that females spent 60% of their annual activity budget foraging, of which 20% was focused on fruit and 70% was focused on invertebrates. However, fruit comprised 58% and invertebrates comprised 40% of the total energy ingested. This difference was largely due to the higher energy and macronutrient intake rates that are driven by the significantly larger size of fruits. Fruits were ingested at a rate of 4.82  0.77 items/min, whereas invertebrates were ingested at a rate of 3.89  1.15 items/min. Females met, and sometimes exceeded, energy and macronutrient requirements on an annual basis, but showed differences between months with high versus low fruit availability in time spent foraging, dry

mass intake, energy intake and macronutrient intake. In this chapter, I address the following research questions to assess the nutritional importance and seasonal role of invertebrates to the capuchin diet in greater detail:

1) What is the intra-annual variation in fruit availability and fruit energy density in the home

ranges of the three capuchin study groups? In addition, what is the relationship between

ripe fruit energy density (kJ/ha) and invertebrate consumption?

2) Which invertebrate groups contribute to the dietary profile of female capuchins in terms of

items consumed?

3) Which invertebrate orders are most energetically important to the capuchin diet, and when

are they temporally important (i.e., is their consumption seasonal)?

4) What role do invertebrates play in relation to the capuchin’s staple diet of seasonally

variable fruit?

3.2 Methods

3.2.1 Study site

I studied three groups of white-faced capuchins living in Sector Santa Rosa (SSR), Área de

Conservaciόn Guanacaste, Costa Rica. SSR is comprised of highly seasonal tropical dry forest that experiences rainfall from mid-May until mid-November, and a dry season (free of rain) from mid-November through mid-May. Rainfall and temperature data were recorded daily during the study period (2009-2011). Peak temperatures occur during March and April, whereas peak rainfall occurs during September and October. Weather data collected during this study is presented in Chapter 1, Figure 1.4.

3.2.2 Study subjects

I collected focal data on 25 adult females (≥ 6 years old, unless reproductive maturation occurred earlier) from three habituated social groups (LV, CP and GN) during 2,124.43 hours of observational contact (575.14 focal hours; LV = 120.09, CP = 227.06, GN = 227.99). I followed groups on a rotational basis for 4-6 days per month during three 4-month periods (Sep – Dec

2009, May – Aug 2010, and Jan – Apr 2011), which collectively represent all seasonal variation at this site. As described in Chapter 2, I based energy (1000 kJ/day) and protein (1.8 g/kg/day) requirements for maintenance of body weight on a study of captive male Cebus albifrons

(Ausman et al. 1986; Ausman and Hegsted 1980), and then I adjusted these values based on the mean weight of female C. capucinus (2.54 kg) to determine minimum requirements (Smith and

Jungers 1997). I multiplied the minimum requirements by coefficients associated with the additional demands of reproduction (cycling = 1.00, gestating = 1.25 and lactating = 1.50) to calculate a monthly adjusted requirement for both energy (as estimated by Key and Ross 1999) and protein. I used the same coefficients that were used in the adjustment of energy requirements due to lack of data on increased protein demands for gestating and lactating females in the primate literature (See Chapter 2 for methodological details).

3.2.3 Behavioral data collection

I collected data in the form of 10-minute focal animal follows (Altmann 1974). Details of focal samples are described in Chapter 2. During follows, I recorded all event behaviors performed or received by the focal females as well as all state behaviors in which the focal female was engaged (e.g., forage, social, rest, travel and other, as outlined in Appendix D). I recorded ranging data points at 30-minute intervals to determine home range size using Kernel Density estimates calculated with Home Range Tools for ArcGIS® V. 2.0.0004 in ArcGIS® 10 (ESRI

2011; Rodgers et al. 2012). I calculated estimates for the 50% and 95% isopleths for each group using a fixed kernel and reference bandwidth. Points within the 95% isopleths were later used in assigning transects to group ranges for calculation of fruit availability and fruit energy density.

3.2.4 Nutritional sampling and analysis

I collected and processed food items during this study to quantify their nutritional composition

(details are described in Chapter 2). When collection was not possible, I used values from previous studies that quantified the nutritional composition of food items eaten at this site

(McCabe 2005) and neighboring site, Lomas Barbudal Biological Reserve, Costa Rica (Vogel

2005). In a few cases where nutritional processing was not possible, I derived values from congener species or species with similar size and composition. Due to the large diversity of invertebrate fauna present in Sector Santa Rosa (e.g., > 3000 spp. lepidopterans, Janzen 1988), as well as the fast pace at which the capuchins found and consumed invertebrate prey, I was unable to identify all invertebrate items eaten at the genus and species level. When identification was feasible, I recorded the scientific name to the finest level of detail possible. Invertebrate species were broadly categorized based on size, and sometimes family, as described in Chapter 2 and

Appendix H. I calculated the energy density (kilojoules per gram dry matter) of each invertebrate food item using the formula (0.1674  (% CP + % WSC) + (0.3766  % CF))

(Janson 1985), where CP is crude protein, WSC is water-soluble carbohydrates and CF is crude fat (National Research Council 2003).

3.2.5 Ecological sampling and fruit availability

Ripe fruit availability (kg/ha) and ripe fruit energy density (kJ/ha) were calculated using phenological and nutritional data for 30 species commonly eaten by white-faced capuchins collected from 2009-2011, and transect data collected by Melin and colleagues (Melin et al.

2014a) (see Chapter 2 for detailed methodology). Data on the seasonal variation in fruit availability and energy density are presented in Chapter 2, Figure 2.4, for the entire study area.

Here, I followed the same methodology for calculating both variables at the group level using a subset of the total number of transects per group, based on transect overlap with ranging points taken at 30-minute intervals during observation days. If a ranging point fell within 250 meters of a transect, the transect was included in the calculations for that group. I used this as an inclusive/liberal approach to assigning transects as a means of maximizing the transect area used to extrapolate availability data. I based this decision on the assumptions that ranging data points were only taken at 30-minute intervals and so did not include all travel areas, the capuchin group had access to resources in all locations through which they travelled, and because ranging points did not account for the entire area in which the group was spread. I assigned transects to multiple groups’ ranges in areas where ranging points overlapped. Fruit availability and ripe fruit energy density are positively correlated at this site (Chapter 2). In this study, I used energy density for comparative analysis. I calculated the availability and energy densities of ripe fruit using a subset of fruiting species previously determined to be important to the capuchin diet based on long-term data. The transect area represented a small proportion of the groups’ home ranges (0.41% for CP and LV groups and 0.32% for GN group).

3.2.6 Invertebrate abundance

I did not directly measure insect abundance, which is logistically difficult and time consuming.

Mosdossy (2013) and Mosdossy et al. (In Review) provided broad estimations of non-caterpillar insect abundance of 8 identifiable invertebrate orders (including orders not consumed by capuchins) from samples collected using canopy and terrestrial malaise traps baited with 90% ethanol. These estimations showed a consistent presence of invertebrates across the annual cycle

with a seasonal peak in mean monthly count during April and May. I referred to the extensive long-term research performed by Dr. Daniel Janzen in Sector Santa Rosa since the early 1980’s for information on Lepidopteran larvae (i.e., caterpillar abundance and seasonality). At this field site, there is considerable intra-annual variation in caterpillar species composition and biomass as well as inter-annual variation in the intensity of caterpillar abundance. This variation is driven by intrinsic fluctuation as well as abiotic fluctuation in rainfall, temperature and soil patterns impacting deciduous leaf biomass on which the caterpillars feed (Janzen 1993). Despite variation in species diversity and intensity, there is a consistent seasonal pattern for the majority of the lepidopteran species at Santa Rosa (but see exceptions that exhibit high host-specificity: e.g., Eulepidotis), beginning with either migration to the forest, or the cueing of pupal eclosion during the early rainy season in response to a drop in temperature (late May). Immediate mating and subsequent oviposition follow pupal eclosion 3-6 weeks later, leading to larval eclosion and a noticeable increase in caterpillar abundance during June, July and August. Abundance subsequently declines in the late rainy season when many species pupate or migrate. It is worth mentioning that while the former temporal pattern is the most prominent, Janzen (1993) noted that the existence of a short dry season during late July (locally termed “veranillo”) could lead to a strong enough drop in temperature when the September rains begin to trigger a second phase of pupal and larval eclosion during the late rainy season (October through November).

3.2.7 Statistical analysis

Using IBM SPSS Statistics 21.0.0.2 (IBM Corp. Released 2012), I performed a one-way

ANOVA with Bonferroni post-hoc comparisons to compare the mean annual ripe fruit energy density across study groups. I used linear regression analyses to assess the strength of the relationship between monthly ripe fruit energy density and the percentage of energy consumption

from invertebrates for each group. The percentage of energy consumption was log transformed prior to statistical analysis to conform to the requirements of the linear regression. One outlier (>

2 standard deviations from the mean) from GN group (November) occurred due to our inability to measure the abundance of bromeliad fruits using the phenological methodology; thus, fruit abundance appears to have been low in periods when Bromelia plumieri comprised over 80% of their total energy intake; I removed this outlier prior to analysis.

I conducted circular statistics to test for seasonality in the intake of four of the most important invertebrate orders using Oriana 4.0 statistical software (Kovach 2011). I used

Rayleigh’s Uniformity test to analyze whether the distribution of data was significantly different from a von Mises distribution, a uniform linear distribution, which in this analysis indicates lack of seasonality (Batchelet 1981). In this directional analysis, the mean angle represents the sample mean, and the mean vector length represents the degree in which samples are clustered around the sample mean, whereby r ranges from 0 (uniform) to 1 (highly clustered) (Janson and

Verdolin 2005). The concentration () measures the degree to which the sample distribution is uniform (von Mises) versus skewed. Months of the year were entered as angles of a circle ‘’. I used the frequency of focal follows with energy intake from each invertebrate order; count data were calculated from 120 randomly selected follows with invertebrate consumption per month

(the minimum amount of monthly focal follows collected, equal to 20 hours of observation) to standardize sampling across months and to conform to the ordinal data requirements for this test.

3.3 Results

3.3.1 Group-level variation in fruit availability and ripe fruit energy density

There were differences in home range size, fruit availability and ripe fruit energy density among the three study groups (Table 3.1). The home range of GN group was nearly twice the size of the home ranges of CP and LV groups. The mean annual ripe fruit availability in CP group’s home range was less than half of the availability in the home ranges of LV and GN groups. The energy density from ripe fruit was significantly different across groups (F = 3.540, df = 2, p = 0.040); specifically, energy density was lower in CP group compared to GN group (p = 0.021) and LV group (p = 0.039). Despite group-level differences in annual ripe fruit energy density, the temporal pattern of variation in energy density remained consistent across groups; energy density was above each group’s annual mean in January, February, March, April, September and

October, and below the annual mean in May, June, July, August, November and December

(Figure 3.1). The energy density of CP group remained lower than GN and LV group across all months.

Table 3.1 Home range size, fruit availability and energy density for the three study groups. Values represent data collected during three field seasons (Sep-Dec 2009, May-Aug 2010, Jan-Apr 2011). Home range size Annual fruit Mean monthly Annual fruit Mean monthly (ha), availability fruit availability energy density fruit energy Group core 50%/ 95% (kg/ha) (kg/ha) (kJ/ha) density (kJ/ha) CP 58.74/219.76 672.59 56.05 2,630,070.59 219,172.55 GN 100.79/409.39 1762.93 146.91 6,310,326.90 525,860.58 LV 51.96/272.16 1666.28 138.86 5,883,888.59 490,324.05

3.3.2 Fruit energy density and invertebrate consumption

There was a negative logarithmic relationship between the energy density of ripe fruit (kJ/ha) and the percentage of energy intake from invertebrate food items (Figure 3.2, percent energy

2 intake from invertebrates = 253.42 – 17.26  log (fruit energy density), R = 0.431, F1,34 =

25.800, p < 0.001). There was a significant negative linear relationship between the log percentage of energy intake from invertebrates and the ripe fruit energy density, whereby the percentage of energy from invertebrates decreased when more fruit energy was available in the environment. This relationship was significant for all groups, and strongest in CP, when groups were analyzed separately (Table 3.2, Figure 3.3).

Figure 3.1 Ripe fruit energy density (kJ/ha) available in the home ranges of the three study groups. Calculations are based on ranging, transect and phenology data for the 2009-2011 study period. The dotted line represents the group’s annual mean ripe fruit energy density.

Figure 3.2 Monthly ripe fruit energy density versus the monthly percentage of energy intake from invertebrates.

Table 3.2 Linear regression of ripe fruit energy density and the percentage of energy intake from invertebrates. Two outliers (> 2SD from mean) were removed for analysis. Group N months R2 Test statistic Significance All 35 0.634 F1,33 = 57.210 p < 0.001 CP 12 0.760 F1,10 = 31.631 p < 0.001 GN 11 0.658 F1,9 = 17.350 p = 0.002 LV 12 0.626 F1,1 0= 16.742 p = 0.002

3.3.3 Dietary profile of invertebrate prey captures

For data collected during focal follows, I categorized invertebrate prey into 21 groups based on the accuracy with which I could reliably identify them as they were consumed. I did not identify a small proportion (12.84%) of the invertebrates consumed. The majority of the unidentified invertebrates were too small and consumed too quickly for accurate identification. The contribution of each category to the total number of invertebrate prey consumed is shown in

Table 3.3. Six categories of invertebrates, both gleaned from plant surfaces and extracted from substrates, each contributed to greater than 1% of the items ingested (ants, caterpillars, shield bugs, jumping bean moth larvae, wasp larvae and acacia ant larvae). Ants and caterpillars made the most significant annual contribution (> 74%) to the total number of invertebrates captured and ingested.

Female capuchins do not seem highly selective in their consumption of caterpillars, with the exception of the embedded Cydia deshaisiana moth larvae, which they actively find and extract from Sebasiana pavoniana fruit when the larvae are annually available in May. Females in CP and GN groups spent approximately one quarter to one third of their foraging budget in

May extractively foraging these larvae (34.61%  7.20 and 22.28%  4.04, respectively). Based on transect data, the host tree species (S. pavoniana) was not abundant in LV’s home range.

Cydia deshaisiana was an important dietary resource for CP and GN groups, comprising 15.40%

(SE = 3.53) of the energy intake, 11.95% (SE = 3.46) of the protein intake, and 31.93% (SE =

5.42) of the fat intake during May, a month with low fruit energy availability. Females were observed to ingest other species of caterpillars of various sizes and instar stages from many families, including but not limited to, Noctuidae, Geometridae, Pieridae, Nymphalidae,

Sphingidae and Saturniidae. Most species were gleaned directly from branches or leaves but some were processed before they were eaten. Processing behavior ranged from minor tasks like biting the head off large caterpillars (e.g. Eumorpha satellitia) to drain the guts from the abdomen before ingesting the remainder of the body, to more time-consuming caterpillar defense-avoidance strategies. For example, the monkeys quickly grabbed saturniid caterpillars, or the leaves surrounding the caterpillars, and scrubbed them vigorously against a branch before ingestion to remove stinging setae or spines. On a few occasions, individuals were observed to use these caterpillars as anointing material between processing and ingestion, as has been observed in other C. capucinus populations (e.g., Lomas Barbudal Biological Reserve, Costa

Rica) and the robust capuchin species, S. nigritis and S. libidinosus (reviewed in Lynch Alfaro et al. 2012) . Shield bugs gleaned from plant surfaces, embedded Pseudomyrmex ant larvae extracted from Vachellia plant thorns, and Polistes wasp larvae extracted from the cells of procured nests comprised the remaining top contributors to the total number of prey items consumed. Taken together, the orthoperans (grasshoppers, katydids and crickets) gleaned from plant surfaces and the forest floor were important contributors as well.

Table 3.3 Taxonomic composition of invertebrate prey consumed by females.

Total N ingested 25,434 N identified 22,169 (87.16%) N items % items Order: Family: Genus species Common name ingested ingested Araneae Spider and spider nest 29 0.114 Blattodea: Blaberidae: Archimandrita tesselata Cockroach 12 0.047 Blattodea: Termitidae: Nasutitermes Termite 13 0.051 Coleoptera Beetle 1 0.004 Hemiptera: Cicadidae: Fidicina mannifera Cicada 65 0.256 Hemiptera: Pentatomidae Shield bug 871 3.425 Hemiptera: Pentatomidae Shield bug egg 19 0.075 Hymenoptera: Formicidae Ant 11,210 44.075 Hymenoptera: Formicidae: Pseudomyrmex Acacia ant larvae 313 1.231 Hymenoptera: Vespidae: Polistes Wasp larvae 333 1.309 Lepidoptera Moth 1 0.004 Lepidoptera Caterpillar pupae 195 0.767 Lepidoptera Caterpillar 7,717 30.341 Lepidoptera: Tortricidae: Cydia deshaisiana Jumping bean moth larvae 819 3.220 Odonata Dragonfly 4 0.016 Orthoptera Grasshopper (sm) 217 0.853 Orthoptera Grasshopper, katydid (lg) 76 0.299 Orthoptera: Gryllidae Cricket 220 0.865 Phasmatodea Walking stick 16 0.063 Scorpiones: Buthidae: Centruroides limbatus Scorpion 5 0.020 various (e.g., Hymenoptera, Coleoptera) Embedded larvae 33 0.130 With the exception of caterpillars, which are lumped into one category, scientific names are listed by the level of specificity at which they were distinguished during observation of feeding behavior.

3.3.4 Annual energy contribution from invertebrate consumption

Along with fruit (a category including arils, bromeliads, seeds and grasses), females also consumed flowers, pith, invertebrates, vertebrates and a small amount of other items (e.g., dirt) during the study. However fruit and invertebrates comprised over 96% of the estimated annual energy intake (57.58% and 39.06%, respectively, excluding vertebrates, which were procured opportunistically and rarely). Table 3.4 lists the top invertebrate orders consumed by white-

faced capuchin females based on their total annual energy contribution to the diet. Four groups

(Lepidoptera, Orthoptera, Hemiptera, and Hymenoptera) comprised the majority (75%) of annual energy consumed from invertebrates as quantified during this study. On an annual basis, lepidopterans comprised the most energy gained from invertebrate consumption, followed by orthopterans, hemipterans and hymenopterans. Lepidopterans ranked fifth overall (including all plant foods) in terms of the mean percentage of energy contribution to the diet. Orthoptera

(grasshoppers, katydids and crickets) was the only other invertebrate category that contributed greater than 1% to the total annual energy consumption including all food types across all three groups.

Table 3.4 Energetic contribution of important invertebrate prey. Invertebrate orders are ranked based on the percentage of the total energy consumed from invertebrates. Rank is listed for each category and categories are listed in order of energetic contribution. Total % energy Invertebrate Order N % intake energy Family: Genus species Common name Code ingested ingested kJ/N (kJ) intake Rank Lepidoptera Caterpillar 8535 33.56 7299.50 41.67 1 various medium-sized, various families CATG 5896 23.18 1.06 6223.87 35.53 Sphingidae: Eumorpha satellitia large satellite sphinx larvae CATF 163 0.64 3.19 520.22 2.97 Tortricidae: Cydia deshaisiana jumping bean moth larvae SPAV 819 3.22 0.45 365.44 2.09 various small-sized, various families CATL 1621 6.37 0.10 168.97 0.96 Noctuidae: Euscirrhopterus, Gerra medium-sized noctuid CATA 36 0.14 0.58 21.01 0.12 Orthoptera Grasshopper, katydid, cricket 513 2.017 2463.22 14.06 2 various small grasshopper GHOP 217 0.85 4.34 942.08 5.38 various large grasshopper or katydid KATY 76 0.30 11.74 892.16 5.09 Gryllidae crickets CRIC 220 0.86 2.86 628.98 3.59 Hemiptera Cicada, shield bug and eggs 955 3.755 1819.59 10.83 3 Cicadidae cicada CICA 65 0.256 14.31 930.11 5.31 Pentatomidae shield bug ACID 871 3.42 1.00 868.52 5.40 Pentatomidae shield bug egg CPEG 19 0.075 1.10 20.96 0.12 Hymenoptera Ant, wasp 11856 46.61 1390.47 7.94 4 Formicidae ant ANTS 11210 44.07 0.10 1137.10 6.49 Vespidae: Polistes wasp larvae WASP 333 1.31 0.54 181.00 1.03 Formicidae: Pseudomyrmex acacia ant larvae AANT 313 1.23 0.23 72.37 0.41 Orders are broken down into smaller categories based on family and size to distinguish their nutritional contribution in greater detail. The number of acacia ant larvae (AANT) consumed per thorn, rather than a single larva, is used as the unit of measure (N). A small handful of shield bug eggs (CPEG) rather than single eggs were used as the unit of measure (N) and the nutritional value was estimated using the combined insects category, as analysis on this food item was not completed.

3.3.5 Seasonal variation in invertebrate consumption

Directional analyses of the number of focal follows with invertebrate consumption showed significant seasonality in consumption of all four orders (Table 3.5). These analyses confirm strong seasonal consumption of lepidopterans during the early rainy season, followed closely in seasonal intensity by consumption of orthopterans during the dry season. The consumption of embedded hymenopterans was weakly seasonal, as indicated by the vector length and concentration, and also occurred during the dry season. Although the mean vector for hemipterans occurred during January, the smaller vector length and concentration indicate a significant aseasonal pattern that is characteristic of a more consistent level of consumption across months. Consumption of hemipterans occurred in all months except the early rainy season (May – Aug), when lepidopteran consumption was high (Figure 3.4). Figure 3.5 displays the mean monthly intake rate for plants, the top four invertebrate orders and all other invertebrates (grouped together).

Table 3.5 Results of directional analysis of seasonality in invertebrate consumption. Lepidoptera Orthoptera Hemiptera Hymenoptera Subset focal follows with intake (N) 579 197 361 517 Mean vector month July February January February Vector length (r) 0.491 0.466 0.156 0.264 Concentration () 1.124 1.051 0.315 0.548 Rayleigh’s test (Z) 139.596 42.766 8.762 36.056 Significance (p) < 0.001 < 0.001 < 0.001 < 0.001

Figure 3.6 (left column) shows the percentage of total energy intake from plant foods, the four most important invertebrate orders, and all other invertebrates. When evaluated as a percentage of total monthly energy consumption, lepidopterans made the largest invertebrate contribution to the capuchin diet. The energy contribution by lepidopterans corresponded to the

a) December January b) December January 120 120

November 90 February November 90 February

60 60

October 30 March October 30 March

120 90 60 30 30 60 90 120 120 90 60 30 30 60 90 120

30 30 September April September April

60 60

90 90 August May August May 120 120

July June July June

c) December January d) December January 120 120

November 90 February November 90 February

60 60

October 30 March October 30 March

120 90 60 30 30 60 90 120 120 90 60 30 30 60 90 120

30 30 September April September April

60 60

90 90 August May August May 120 120 July June July June

highly seasonal pattern of intake. Lepidopteran consumption predominately occurred during the early rainy season from May through August. During peak consumption (June and July), lepidopterans comprised the majority (> 50%) of all energy intake in all three groups (CP =

77.07%  4.94, GN = 74.89%  4.50, LV = 51.41%  7.40). Orthoptera (range = 0 – 11.67%),

Hemiptera (range = 0 – 7.44%), Hymenoptera (range = 0.04 – 5.10%), and other invertebrates such as scorpions, cockroaches and beetles (range = 0.25 – 36.45%) made up a lower percentage of monthly intake. Other than Lepidoptera, a single category never comprised greater than approximately one third (36.45%) of the total energy consumed.

Figure 3.6 (right column) shows the monthly percentage of protein intake from plant foods and the four most important invertebrate orders. On an annual basis, plants contributed the highest annual percentage of protein (푥̅ = 37.74%  3.78) across groups, with the highest plant protein intake occurring during the dry season (Jan – Apr). Hymenoptera (CP: 푥̅ = 5.53% 

1.66, GN: 푥̅ = 3.91%  0.78, LV: 푥̅ = 2.80%  0.52), Hemiptera (CP: 푥̅ = 9.76%  2.82, GN: 푥̅ =

5.44%  0.87, LV: 푥̅ = 7.54%  1.48) and Orthoptera (CP: 푥̅ = 10.39%  3.17, GN: 푥̅ = 8.01% 

2.62, LV: 푥̅ = 11.89%  4.39) comprised a small percentage of the monthly protein intake.

Lepidopterans made the largest invertebrate contribution to total protein intake (21.98%  0.84) and 88.87% (SE = 2.74) of the annual lepidopteran intake occurred during the early rainy season

(May – Aug). During the early rainy season the mean lepidopteran protein consumption across groups was 64.5% (SE = 1.31) of the total intake (range = 38.29% to 84.38%). Like energy intake, protein intake from lepidopterans corresponded to the highly seasonal pattern of consumption of this invertebrate order. The “other” invertebrate category, which included

Araneae, Blattodea, Coleoptera, Odonata, Phasmatodea and insects too small to be identified,

comprised a large percentage (푥̅ = 49.06%  1.12) of protein intake during the late rainy season

(Oct – Dec).

Figure 3.7 shows the monthly energy intake from fruit and invertebrates relative to the estimated energy and protein requirements for females (minimum energy = 1000 kJ and minimum protein = 1.8 g/kg/day; minimum requirements were multiplied by 1.25 for gestating females and 1.50 for lactating females; see Key and Ross 1999 and Chapter 2 for methodological details). Even before accounting for the additional reproductive demands of gestation and lactation, females did not meet the estimated minimum intake requirements (1000 kJ/day) during the month of June (Figure 3.7a). Although minimum requirements were met during May, July and December based on intake from both food types, females would not have met requirements based on the intake of plant foods alone during these months. Once the minimum requirement was adjusted for additional demands based on the mean reproductive state of the females, requirements were not met May, June or July, even when considering the energy intake from both fruit and invertebrates together. The energy deficit was the largest in June, when mean energy intake was estimated as 638 kJ/day based on hourly rates of intake. On average, females met estimated minimum and adjusted protein intake requirements based on the intake of plant protein alone, and when accounting for both invertebrate and plant protein, greatly exceeded requirements (Figure 3.7b). Intake rates during June and July were exceptions to this pattern. In these months, females would not have met minimum protein requirements without protein intake from invertebrate foods.

a) b)

c) d)

e) f)

Energy Intake (kJ) Protein Intake (g) a) b)

c) d)

e) f)

a) 3000

2500

2000

1500

1000

500

Mean Energy Intake (kJ/day)Energy Mean J F M A M J J A S O N D

Fruit Invertebrate Minimum Energy Requirement Adjusted Energy Requirement

b) 16 14 12 10 8 6 4 2 0

Protein Intake(g/kg/day) Protein J F M A M J J A S O N D Month

Fruit Invertebrate Minimum Protein Requirement Adjusted Protein Requirement

3.4 Discussion

Together with many omnivorous Old World primates, capuchins have been classified as occasional insectivores based on their opportunistic consumption of invertebrates (McGrew

2001). The aims of this study were to better understand how seasonal variation in fruit, the white-faced capuchins’ staple food resource, corresponded to the consumption and dietary role of different orders of invertebrates as well as to determine the energetic importance of invertebrates to females. I found that while many orders of invertebrates seemed to complement nutritional intake in terms of energy and protein consumption throughout the annual cycle, lepidopterans played a particularly important role. This was especially true for females in the group whose home range contained lower ripe fruit energy density, as well as during periods of lower fruit energy availability. In these cases, lepidopterans served as a replacement for a considerable proportion of the energy and protein intake that was obtained from fruit in other study months.

3.4.1 Variation in ripe fruit energy density across the study groups’ home ranges and the relationship between fruit and invertebrate consumption

If the primary staple food resource of white-faced capuchins is fruit, then home range “quality” may be best described in terms of the ripe fruit energy density (kJ/ha), which is strongly correlated to fruit availability (kg/ha) at this site (Chapter 2). There was a negative relationship between the quality of a capuchin home range and the percentage of energy intake from invertebrates. This negative relationship was the strongest in CP group, who inhabited the poorest quality home range. However, the negative relationship was also observed in the other two groups (GN and LV), who shared similar quality home ranges despite differences in home range size. The variation in home range size and quality provides a good explanation for the

relatively stronger reliance upon invertebrate foods for energy by CP females than by females in the other study groups.

3.4.2 Intra- and inter-annual variation in ripe fruit availability and energy density

The conclusions drawn in this study regarding seasonal variation in ripe fruit availability draw attention to a dip in fruit abundance in the early-mid rainy season that was drastic in this study but also evident – yet not appreciated or discussed – in previously published studies at this field site. Specifically, I found that ripe fruit availability and energy density were low during June and

July. These months have been included as part of the peak fruit period previously, because 1) they lie at the middle of two peaks in fruit abundance and thus may be masked by the peaks in fruit abundance lying on each side of this two month dip, and 2) visits to fruit patches are frequent during this time (Carnegie et al. 2011a; Melin et al. 2014b; Mosdossy et al. In Review).

To address the first point, this bi-modal trend had not been previously identified because it is not detectable using directional statistics, as the peaks are not 180 degrees offset from each other.

However, my results demonstrate this pattern and it is clearly visible in a recent paper by

Campos et al. (2014), summarizing ecological variation across several years at this site. The high frequency of fruit patch visits – which are recorded independent of their duration – may be explicable in that the monkeys are visiting small, quickly depleted fruit patches in these months, which does indeed seem often to be the case (Melin et al. 2014b). It is important to note that in addition to intra-annual seasonality, there is also likely to be considerable inter-annual variation in fruit availability in response to changes in climate, which can be highly variable across years at this field site (Blanco Segura 2014), and may explain the severity of the dip during my study.

Furthermore, the phenological records represent a subset of important fruit species for which we have nutritional information, but do not include all species consumed during all study months –

for example, we do not include very small species that have had high visitations in July during other study periods (e.g. Erythroxylum havanense and Cordia guanacastensis). Adding these species to calculations of fruit biomass might shift our calculations slightly, but their small size and fruit crops should not make a large difference. Finally, we are working to refine calculations from past studies. Specifically, here I excluded fruit biomass contributions from a species for which capuchins primarily consume invertebrates from within the hollow fruit (Sebastiana pavoniana), as I believe it is best treated as an invertebrate source. I did not include the biomass of species with small wind-dispersed seeds (e.g., Luehea candida and L. speciosa), as biomass would not be accurately represented in the calculations. I also excluded species that were rarely eaten and for which we do not have nutritional records (Guazuma ulmifolia, Cecropia peltata,

Randia thurberi, Sapium glandulossum). Most notably, a phenology index that incorporates both fruit coverage and maturity, transect samples, and nutritional data were used together in the analyses for this study, and the biomass and density sums across species were used in statistical comparison rather than counts of species in peak fruit. These refined analyses have been used in recent studies (Campos et al. 2014) and will be implemented in future studies to assess the extent of inter-annual variation in fruit availability experienced by capuchins.

3.4.3 The invertebrate role in the dietary profile of female capuchins

Invertebrates that were quickly gleaned from plant surfaces, as well as invertebrates that required extractive foraging techniques, were exploited by females throughout the year, as indicated by the frequency of invertebrate items captured across orders. Ants and caterpillars comprised the highest frequency of invertebrate item captures. A recent study at this field site indicated that the majority of non-caterpillar invertebrates (including those not consumed by capuchins) are present throughout the year within the home ranges of the study groups, with a peak in

abundance during April and May (Mosdossy 2013). I expected the consumption of non- caterpillar invertebrates to track their abundance pattern. However, the number of observed consumption events of the top non-caterpillar orders only partially tracked previously measured annual abundance patterns, as peak consumption did not occur during these months. Invertebrate abundance may vary considerably among years due to biological variation as well as in response to climatic and ecological variables (Janzen 1993; Uvarov 1931). Because I did not measure invertebrate abundance, it is unclear whether the foraging patterns I observed tracked the abundance of invertebrates during this study period.

3.4.4 The energetic and seasonal importance of invertebrates

The consumption of hemipterans and hymenopterans lacked a strong seasonal pattern and was high in months with both high and low fruit energy density. The consumption of orthopterans was strongly seasonal and centralized during the dry season (Jan – Apr) when the ripe fruit energy density was also high, although there was a low level of consumption during other months as well. Monthly energy intake followed the same pattern as per-item intake for these three orders of invertebrates; although the magnitude of invertebrate energy intake was considerably less than energy intake from plant foods (predominately fruit). The mean monthly energy intake for these three orders of invertebrates comprised between 5% and 15% of the total intake across groups. Based on these patterns, the nutritional role of non-caterpillar insects may be best characterized as foods that are complementary to the staple diet, providing additional protein and fat to that gained from fruit. From these data, I draw similar conclusions to those made by Melin et al. (2014) and Mosdossy et al. (In Review) based on their patterns of consumption, that non-caterpillar invertebrates as a whole are therefore not likely to be staple fallback foods, but are more similar to high-quality filler fallback foods. Although the majority

of captures occurred aseasonally, the seasonal captures during the late rainy season of invertebrates categorized as “other” included embedded insects and are consistent with findings that these are important during November and December when fruit availability is at a low to moderate level (Melin et al. 2014).

In contrast to non-caterpillar invertebrates, statistical analysis indicated that the consumption of lepidopterans was strongly seasonal. Female capuchins consumed large amounts of jumping bean moth larvae (Tortricidae: Cydia deshaisiana) during May in the two groups in which this resource was available. Females also consumed caterpillars from various families during the predictable seasonal caterpillar outbreaks that occur in Sector Santa Rosa from May through August, most heavily in CP group, which had the lowest ripe fruit energy density. The timing of the caterpillar outbreak coincided with a four-month period of low fruit availability and energy density in all three groups. Caterpillars therefore stand out as a highly seasonal and energetically important resource for female capuchins, comprising over 40% of the annual energy intake from invertebrates. Caterpillars represented at least half, and in CP group almost all of the mean energy intake during June and July, which corresponded to the early rainy season.

Other than the Cydia deshaisiana moth larvae extracted from Sebastiania pavoniana fruit during

May, all other caterpillars were gleaned from the surfaces of plant substrates, making them an easily accessible resource.

3.4.5 The nutritional role of invertebrates in relation to the capuchin’s staple diet of seasonally variable fruit

Data collected during this study indicate that females would not have met energy intake requirements from fruit intake alone during the early rainy season months. Even with high caterpillar consumption, it is likely that females were utilizing metabolic energy stores at this

time since they did not meet estimated energy requirements. It is possible that 10-minute focal animal follows did not capture the full range of behavior and the patterning of nutritional intake for each study month – full-day follows would provide a more accurate picture of total energy and nutrient intake. Additionally, intraspecific variability in the nutritional composition of food items across space and time and the use of published values for some food items could have led to the over- or under-estimation of intake during some months. However, subsequent analyses of metabolic biomarkers (Chapters 4 and 5) provide a more detailed picture of the metabolic consequences associated with seasonality in resource availability and support the conclusions drawn based on patterns of seasonality in this chapter. Opportunistic consumption of caterpillars during high-quantity outbreaks was necessary to meet protein requirements and ameliorate the deficit in meeting energy requirements during the outbreak period. Although data on the availability and distribution of Lepidoptera larvae are limited in our long-term dataset, most tropical lepidopteran species show host specificity during larval stages (Dyer et al. 2007), and therefore, spatial distribution is likely more similar to the clumped nature of fruit resources.

Accordingly, caterpillars may be a more nutritionally profitable resource for capuchins to locate and exploit compared to alternatives during this season.

The white-faced capuchin is not the only species that relies heavily upon seasonal invertebrate outbreaks (reviewed by Rothman et al. 2014). Lepidoptera also comprised a large percentage (11%) of the annual foraging budget of the Ka’apor capuchin (Cebus kaapori) in the

Brazilian Amazon, and in contrast to other months during which caterpillars were a more minor dietary resource, they comprised the majority of the June feeding scans (91.8%) (de Oliveira et al. 2014). Corroborating the idea that caterpillars are a clumped resource, capuchins were observed to directly and aggressively compete with sympatric squirrel monkeys (Saimiri

sciureus collinsi) over access to caterpillars during the seasonal outbreak at that field site.

Additionally, bearded sakis (Chiropotes satanas chiropotes) in the Brazilian Amazon are highly frugivorous; but during August, a period when lepidopteran larvae were abundant and other resources were suspected to be scarce based on low rainfall and an increase in the diversity of saki food resources, sakis spent a large proportion of their feeding time (56%) targeting primarily lepidopteran invertebrates (Frazäo 1991).

Overall, the high energy consumption of caterpillars by female capuchins may be due to a shift in diet that results from several factors: a relative dip in available fruit energy; the high quantity of caterpillars available during the outbreak; the clumped distribution of these species due to their dietary preference for the leaves of certain tree species; or a combination of all of these. It is clear though, that Lepidoptera is the most important invertebrate food source to capuchins based on its nutritional role as replacement energy and protein (to that gained from fruit) during periods of low ripe fruit energy density (kg/ha). Variation in climatic factors and fruit abundance may affect the inter-annual dietary importance of invertebrates to capuchins. If the seasonal dietary value of caterpillars is consistently high across years, ecological changes that disrupt the caterpillar life-cycle and abundance at this site could result in negative metabolic consequences for capuchins, especially if the nature (in terms of distribution, accessibility and nutritional value) of other invertebrate orders deem caterpillars a relatively more profitable resource. Those consequences may affect the reproductive success of females if they align with reproductively demanding periods such as early lactation or during the weaning of offspring, the latter of which is a time period when females have been observed at this site to increase the rate of consumption of insects, possibly to supplement protein intake (McCabe and Fedigan 2007).

More severely, a drop in caterpillar abundance during this low fruit period may lead to an

increase in capuchin mortality. A similar phenomenon was observed on Barro Colorado Island, whereby high capuchin mortality followed a climate-associated drop in arthropod abundance during an already nutritionally stressful period (Milton and Giacalone 2014). More detailed records (i.e., all occurrences) of daily and monthly food consumption and further analyses of invertebrate abundance would be extremely valuable. These data may shed light upon the variation in inter-annual availability of different invertebrate orders, whether capuchins are actively selecting certain food types based on profitability, and how variation in food availability corresponds to capuchin foraging behavior and nutritional intake.

Chapter Four: Using urinary parameters to estimate seasonal variation in the relative muscle mass of females

4.1 Introduction

Measuring aspects of physical condition is important for gaining insight into the health of humans and animals. The reproductive success of female primates is closely tied to the quality and abundance of food resources (Trivers 1972; Wrangham 1980) since a minimum level of energy balance is necessary to resume ovarian cycling (Ellison and Valeggia 2003). Insulin is a hormone that acts to regulate energy balance, but also plays an important role in ovarian function

(Willis et al. 1996). Body size and lean muscle mass have also been positively correlated to fertility, possibly because relatively better body condition reflects the ability to meet energetic demands and build expensive tissue (Albon et al. 1986). Nutritional intake and body condition can affect the timing of reproductive events such as maturation (Bentley 1999; Bercovitch and

Strum 1993; Eveleth and Tanner 1990; Voland 1998), the persistence of ovulatory cycling at regular intervals, and seasonality of births (Brockman and van Schaik 2005; Carnegie et al.

2011a; Drent and Daan 1980; Houston et al. 2007; Janson and Verdolin 2005; Lewis and

Kappeler 2005a; Richard et al. 2000). Specifically, some females that experience a decrease in energy balance (i.e., intake minus expenditure) and the loss of body weight due to demands of lactation (Chao 1987) or poor environmental conditions (e.g., food shortage), exhibit a delay in weaning infants and the subsequent onset of ovarian cycling (Koenig et al. 1997; Lewis and

Kappeler 2005b).

While many studies have shown that resource scarcity has a negative effect on broad measures such as growth, reproductive output and overall survival, few studies have been able to

measure more direct short-term consequences of resource scarcity on health such as variation in body condition. The latter would give us a more detailed assessment of physiological decline and the associated reproductive consequences on a shorter time scale. Accurate quantification of body condition is logistically challenging in non-invasive studies of wild animals, particularly since measurement often requires baiting and trapping methods that interfere with the variables under investigation in physiological studies, such as foraging behavior and nutritional status.

However, non-invasively collected data on body condition are very useful for drawing accurate links between the behavior of animals and the physiological processes that ultimately affect fitness. More in-depth monitoring of changes in physiological condition will help us to identify how individuals respond to a variety of pressures, and how body condition corresponds to life- history characteristics such as the timing of reproductive events and survival of offspring.

Measurement of relative muscle mass using urinary parameters is one way in which researchers can non-invasively monitor the physical condition of wild animals.

4.1.1 The role of creatinine in muscle mass estimation as a measure of physical condition

Creatine and phosphocreatine are molecules present in the body that play an essential role in supplying energy to muscle cells through their role in ATP regeneration; they dehydrate to produce the by-product creatinine, which is excreted in the urine (Heymsfield et al. 1983; Hunter

1928). The majority (98%) of creatine and phosphocreatine is contained in muscle (Burger

1919). The 24-hour creatinine method, whereby the amount of creatinine excreted in the urine in a 24-hour period is used to estimate the volume of muscle tissue present in the body, was tested in a variety of mammalian species and led to the development of a non-invasive and inexpensive means to measure muscle mass in humans (Myers and Fine 1913; Palladin and Wallenburger

1915). Myers and Fine (1913) found that the within-species concentration of creatine (as a

percentage of muscle tissue) is constant, and that the relationship between total body creatine concentration and the daily excretion of urinary creatinine is also constant. These are important findings because they demonstrate that the amount of urinary creatinine produced within a given time period reflects the total muscle mass of an individual. Individuals with more muscle mass produce and excrete more creatinine. This method has been widely used to monitor muscle mass and renal function as part of human health care (Adebisi et al. 2001; Heymsfield et al. 1983;

Kasiske and Keane 2000).

It is logistically impossible to collect 24-hour urine yields from wild animals. An alternate method of assessing the relative amount of creatinine produced by the body is to compare the ratio of creatinine concentration in single urine samples to the specific gravity of the urine (i.e., the ratio of the density of the urine relative to the density of water), since specific gravity is not affected by variation in muscle mass. This method was developed and tested using urine samples from chimpanzees (Pan troglodytes) (Emery Thompson et al. 2012). Emery

Thompson et al. (2012) found that multiple urine samples spot-collected over an extended time period could be used in regression analyses to generate a “specific gravity – creatinine slope”

(i.e., SG-Cr slope) representative of the relative muscle mass of individual chimpanzees. Here, I employ this method to assess changes in the relative muscle mass of female white-faced capuchin monkeys (Cebus capucinus) as a measure of their body condition.

White-faced capuchins in Sector Santa Rosa, Área de Conservaciόn Guanacaste, Costa

Rica, live in a highly seasonal tropical dry forest. Their diet is omnivorous, and fruit and invertebrates comprise the majority of their caloric intake (Chapter 2). I have measured multiple ecological, behavioral and physiological variables related to seasonal energy availability, which is advantageous when assessing the physical condition of females in this study population. Both

fruit and invertebrate foods show seasonal variation in abundance, and consequently, there is considerable intra-annual variation in caloric intake (Chapter 2; Chapter 3; Melin et al. 2014;

Mosdossy et al. In Review). Behavioral observation of food intake rates compared with daily energy intake requirements (1000-1500 kJ/day), estimated based on body size and a captive capuchin study (Ausman and Hegsted 1980) and energy coefficients associated with reproductive state (Key and Ross 1999), suggest that female capuchins may not always meet requirements during periods of low fruit availability (Chapter 3). A failure to meet nutritional requirements during certain seasons may result in fluctuations in body condition throughout the year.

The purpose of this study is to use urinary parameters to infer variation in the relative muscle mass of female white-faced capuchins as an indication of changes in their physical condition. Specifically, I address the following research questions:

1) Are there strong positive correlations between specific gravity and creatinine, as in

reports from chimpanzees (Emery Thompson et al. 2012; Anestis et al. 2009) and humans

(Carrieri et al. 2000; Haddow et al. 1994)? In addressing this question, I also evaluate

whether semi-quantitative specific gravity readings from urinalysis test strips are

equivalent to specific gravity determinations by refractometer.

2) Does female body condition, as measured by variation in creatinine excretion, vary across

social groups, and between periods of high and low resource abundance (measured as

ripe fruit energy density (kJ/ha)?

To address these questions I collected urine from females within the Sector Santa Rosa capuchin population to measure the specific gravity and creatinine concentration (mg/ml). First,

I compared specific gravity measured using urinalysis test strips under field conditions to values

obtained using a refractometer in a laboratory setting. Then, I used mixed effects models to assess group-level and seasonal changes in the relative muscle mass of females, as indicated by the relationship between specific gravity and creatinine (i.e., SG-Cr slope). If the physical condition of females declines with seasonal reduction in the abundance of resources and associated energy intake by females (H1), I predict that during periods of low fruit abundance females will show a relative reduction in muscle mass, as indicated by the relationship between urinary creatinine and specific gravity (P1a).

I also predict that the physical condition of females will differ by group, whereby groups in experiencing lower fruit abundance will have lower relative muscle mass (P1b). There is documented variation across our study groups in home range size and quality in terms of resource abundance (Chapter 3; Campos et al. 2014) . Finally, I predict that because variation in fruit abundance can differ across home ranges, that there may be an interaction between group and fruit abundance period, whereby females in some groups will show greater changes in muscle mass between periods of high and low fruit abundance than females in other groups

(P1c). On a broad scale, changes in body condition may also influence the timing of reproductive events and highlight the consequences of potential mismatches between resource abundance and the timing of energetically costly states for either the mother (lactation) or infant

(weaning) (Carnegie et al. 2011a). Such mismatches may affect the physical condition and potentially, the survival of both parties.

4.2 Methods

4.2.1 Sample collection

I collected urine samples from individually identifiable female white-faced capuchins in Sector

Santa Rosa, Costa Rica, during three data collection periods: September – December 2009, May

– August 2010, and January – April 2011. Field assistants and I opportunistically caught urine samples on a plastic substrate or pipetted samples from vegetation. At the time of collection, I performed urinalysis tests using reagent strips (Seimens Multistix 10 SG, Appendix E) to measure a number of urinary parameters, including ketones and specific gravity. I then placed samples on ice and froze them within 6-8 hours of collection. The samples were transported on dry ice to the Hominoid Reproductive Ecology Laboratory at the University of New Mexico for laboratory testing. I assumed that at the time of this study, the female capuchins were healthy and were not suffering from physiological or pathological conditions (e.g., kidney disease) that might affect normal excretion of urinary parameters.

4.2.2 Urine analysis

4.2.2.1 Specific gravity

I assessed the specific gravity of 100 l aliquots of urine from 734 samples using a handheld refractometer with a resolution of 0.001(Atago PAL-10S). I cleaned the refractometer between samples and checked the calibration of the unit using double-distilled water between sets of 10 samples. Specific gravity values ranged from 1.001 to 1.045.

4.2.2.2 Creatinine

Creatinine was measured in 885 urine samples via the Jaffe reaction, whereby urinary creatinine concentrations are measured based on colorimetric determination following a reaction with picric acid in an alkaline solution (Bonsnes and Taussky 1945). The color change that occurs from this

reaction is directly proportional to the concentration of creatinine in the sample. Following the

Jaffe reaction protocol, I diluted sets of samples with water using a 1:10 dilution ratio. I pipetted

40 µl of urine sample followed by 360 µl ddH20 into a 96-well tube setup, and vortexed the tubes. I pipetted 100 µl of the diluted samples in triplicate into 96-well polystyrene Costar reaction plates. Standards (Sigma-Aldrich, 0.01, 0.03, and 0.10 mg/ml concentrations) were also pipetted in triplicate in the first row of the plate. I added 50 µl of 0.75 M sodium hydroxide

(NaOH) followed by 50 µl of 0.02 N picric acid to all wells. Plates were transferred onto a shaker for 15 minutes and then the absorbance was read at 492 and 620 nm using a spectrophotometer (Thermo MultiskanTM GO Microplate Spectrophotometer). The final absorbance reading (620 nm) from each sample was averaged for each set of triplicates and subtracted from the mean of the initial absorbance readings (492 nm) to calculate a corrected absorbance value (e.g., to correct for background reflectance). I calculated creatinine concentrations by applying the corrected absorbance values for each sample to the regression equation obtained from the standard curve [i.e., plot of corrected absorbance against final creatinine concentration (mg/ml)]:

Formula

mg sample absorbance − yintercept Creatinine ( ) = [ ]  sample dilution ml slope

I also calculated the coefficient of variation (CV) for each sample. An intra-assay reliability of 1.89% was determined by calculating the mean of all sample CVs for each assay.

Because low intra-assay CVs suggest high reliability of determinations, I ran a single replicate of those samples with insufficient volume to assay in triplicate. Valid results (i.e., measurement with Cr > 0.000 and CV < 10%) were obtained from 746 samples. Creatinine ranged from 0.021

to 1.583. In total, these analyses yielded 692 samples from which I obtained both a creatinine and specific gravity value for muscle mass analysis.

4.2.3 Ecological and nutritional sampling

The methods I used to measure ripe fruit abundance (kg/ha) are reported in Chapter 2, and the distribution of fruit abundance by group home range (broadly calculated using Kernel Density estimates) is described in Chapter 3. I calculated ripe fruit energy density as the sum of each species’ fruit abundance multiplied by its nutritional value (kJ/kg) (Chapters 2 and 3).

4.2.4 Statistical analysis

I performed a Spearman’s rho correlation to determine the relationship between the specific gravity measured using urinalysis test strips (SG field) versus values obtained using a refractometer (SG lab). Because the refractometer measure was truly quantitative, and the two methods only weakly correlated, I conducted all future analyses using only the SG lab measure.

SG and creatinine are expected to produce highly correlated values, as each is a measure of relative water content in urine. However, as creatinine excretion is dependent on relative muscle mass, and specific gravity is not, the the slope from the regression of the two measures can be used to estimate relative muscle mass (Emery Thompson et al. 2012; Forbes and Bruining 1976).

In their validation of this method for evaluating muscle mass, Emery Thompson et al.

(2012) determined that, when comparing SG-Cr slopes between individual chimpanzees, it was desirable to have 25 data points (i.e., samples) per regression. As this level of individual sampling was infeasible for each individual in each season during my study, I instead modeled the effect of specific gravity on creatinine concentration for the full set of urine samples and then used the residuals of this relationship as data points in an analysis of group and seasonal effects.

This was a mixed effects model with restricted maximum likelihood estimation. Because water

has a specific gravity of 1 and a creatinine concentration of 0, I first subtracted 1 from the specific gravity measures and ran the model through the origin rather than computing intercepts

(Emery Thompson et al. 2012). I included both linear and quadratic terms for specific gravity, as the relationship between creatinine and specific gravity is slightly curvilinear. Before fitting the model, I created orthogonal terms to standardize both the linear and quadratic specific gravity to reduce multicollinearity so both could be included. A Spearman’s rank-order correlation confirmed that the model residuals were uncorrelated to specific gravity (rs = -0.091, p = 0.014).

Having removed the effect of urine density on creatinine concentrations, residuals of this model should capture individual and temporal variation in muscle mass.

To examine the effects of season and group on relative muscle mass, I generated a new mixed effects model using maximum likelihood estimation and the residuals from the first model as the dependent variable. I included monthly ripe fruit abundance (high: n = 6 mo., 216 samples and low: n = 6 mo., 476 samples) and group identity (LV: n = 141 samples, CP: n = 307 samples and GN: n = 244 samples) as fixed effects based on my predictions that females would have higher relative muscle mass in months where fruit abundance was high, and that this would vary by group based on home range quality. I also included the interaction between group and fruit abundance as a fixed effect based on my prediction that females in some groups might respond differently to changes in fruit abundance within their home range than others. I included female identity as a random effect to control for individual variation and unequal sampling. I conducted all statistical analyses using IBM SPSS Statistics 21.0.0.2 software (IBM Corp. Released 2012).

100

4.3 Results

4.3.1 Specific gravity: lab versus field measures

There was a significant but weak positive correlation between the semi-quantitative specific gravity measurement taken using urinalysis strips in the field and those determined using the refractometer in the laboratory (rs = 0.306, N = 486, p < 0.001). Visual inspection of the data indicated that while values obtained with the refractometer, which has a wider range, were generally higher than those obtained from the urinalysis strips, samples with high SG from field urinalysis did not yield correspondingly high SG using the refractometer (Figure 4.1). I split values into low (SG  1.015) and high (SG  1.020) categories based on the urinalysis results and conducted separate correlation analyses. When separated, the correlation between low SG field urinalysis values and laboratory results was significant and stronger (rs = 0.568, N = 168, p

< 0.001) but there was not a significant correlation between high SG field urinalysis values and laboratory results (rs = 0.092, N = 318, p = 0.103). Based on these results, and because the refractometer generated a continuous, quantitative measure of specific gravity, I only used samples in all further analyses for which I was able to measure SG using the refractometer.

101

4.3.2 Relative muscle mass

Both linear and quadratic specific gravity terms significantly predicted creatinine, as expected

(Figure 4.2, Table 4.1, Model 1). In Model 2, season was a significant predictor of the creatinine residuals, such that relative muscle mass was estimated to be greater during months with high ripe fruit availability versus low fruit availability (Table 4.1, Model 2). Although the effect was smaller, group identity also significantly predicted relative muscle mass, whereby females in CP group had significantly lower relative muscle mass than females in GN group and LV group

(Table 4.1, Model 2). There was also a significant interaction effect between group and fruit

102

abundance, whereby females in LV group showed less variation in relative muscle mass than females in CP group across periods of high and low fruit abundance. The mean residuals from this model are displayed by group and season in Figure 4.3.

Figure 4.2 Scatterplot displaying the relationship between specific gravity (minus 1) and creatinine (mg/ml) of urine samples (N = 703).

103

95% Confidence

Model 1: Creatinine Parameter Standard Degrees of interval (CI) Fixed effects estimate (Est) error (SE) freedom (df) t statistic p value Lower Upper Specific gravitya (linear) 0.180 0.019 690 9.272 < 0.001 0.142 0.218 Specific gravitya (quadratic) 0.042 0.019 690 2.139 0.033 0.003 0.080 Model 2: Model 1 Cr Residuals Parameter 95% CI Fixed effects estimate (Est) Lower Upper Intercept 0.296 0.018 31.299 16.521 < 0.001 0.259 0.332 Fruit Availability = High 0.345 0.027 690.333 12.789 < 0.001 0.292 0.398 Fruit Availability = Low 0b 0.000 Group = LV 0.099 0.032 36.334 3.085 0.004 0.033 0.163 Group = GN 0.076 0.027 39.804 2.828 0.007 0.022 0.130 Group = CP 0b 0.000 Fruit Availability*Group LV -0.114 0.046 681.279 -2.479 0.013 -0.205 -0.024 Fruit Availability*Group GN -0.038 0.039 689.950 -0.842 0.400 -0.109 0.044 Fruit Availability*Group CP 0b Random effects Est SE Wald Z p value CI Female identity 0.001 0.001 1.534 0.125 0.000 0.004 a) Specific gravity minus 1 standardized as linear and quadratic orthogonal terms; b) Reference parameter, set to 0.

104

.800

.700

.600

.500

.400

.300

.200

Mean Creatinine Residuals Creatinine Mean .100

.000 LV GN CP Group Fruit Availability: High Low

4.4 Discussion

Comparison of specific gravity values obtained using urinalysis strips versus a refractometer showed that on average, the values obtained using the test strips were lower than those obtained using the refractometer. In addition, values at the lower end of the specific gravity spectrum corresponded more closely to refractometer readings than values at the higher end of the spectrum. The discrepancy at higher field specific gravity measurements may have resulted because urine that is more concentrated is often very dark in color, which may stain or obscure the color pad on the reagent strips. It is also possible that there were errors in discerning colors on the test strip due to variable lighting conditions in the forest, or loss of sensitivity in the strips, which are sensitive to moisture. Regardless, it appears that urinalysis strips are not well suited for determining the specific gravity of capuchin urine samples under these field conditions, and 105

readings should be interpreted with caution if they are not validated using comparative methodology.

The relationship between urinary creatinine excretion and the specific gravity of urine samples indicated that variation in muscle mass among females exists across seasons and social groups as predicted based on fruit abundance and home range quality. This suggests that the method employed is effective in measuring variation in the relative muscle mass of adult female white-faced capuchins. Specifically, these results supported my first prediction (P1a), that females had higher relative percentages of muscle mass during months with high versus low ripe fruit energy density (mean high = 840,951 kJ/ha  54,928 versus mean low = 274,576 kJ/ha 

64,881, Chapters 2 and 3). The results also supported my second prediction (P1b) that there would be group-level differences in relative muscle mass. Groups with higher mean annual ripe fruit energy density in their home ranges (GN and LV) also showed a stronger positive relationship between specific gravity and creatinine. This indicates that on average, resident females in these two groups maintained higher percentages of muscle mass compared to females in CP group, whose home range has approximately half the available ripe fruit energy density than did the ranges of the other two groups. The magnitude of change in relative muscle mass across seasons was also larger for females in CP group (P1c).

The change in the percentage of relative muscle mass of females between months with high and low ripe fruit energy density supports theoretical predictions that females maintain relatively higher percentages of muscle mass during months when the energy density from ripe fruit is higher. Furthermore, group variation in muscle mass is supported by group-specific estimates of habitat quality (i.e., ripe fruit energy density) in that groups with better home range quality (GN and LV) had relatively higher percentages of muscle mass. In addition, females in

106

LV group showed less variation in muscle mass between months with high and low fruit abundance than those in CP group (Chapter 3). However, body size and body mass have not been measured in this population of white-faced capuchins. As a result, I was not able to run correlational analyses with muscle mass estimates and true body mass to support these data.

Despite this, these findings provide strong preliminary support for the implementation of future research investigating the relationship between ecological and social factors that may affect body condition in both this population and species of capuchin.

Studies have shown that “creatinine equivalence” (kg muscle mass/g urinary creatinine) is not completely stable since factors other than muscle mass may produce intra-individual variation in urinary creatinine excretion (Heymsfield et al. 1978, p. 478). For example, in humans, there is normal, unexplained variation in creatinine excretion of 4-8% (Cryer and Sode

1970; Greenblatt et al. 1976). An increase in metabolism from strenuous exercise or stress is associated with increases in creatinine output (Hobson 1939; Scrimshaw et al. 1966; Srivastava et al. 1957) and other factors may also influence creatinine excretion. However, most studies have not been able to conclude that changes in output associated with outside factors are not also influenced by associated changes in muscle mass (reviewed by Heymsfield et al. 1983).

Unfortunately, there is no way to have controlled for natural intra-individual variation in urinary creatinine excretion by white-faced capuchin females during this study, and I assumed that variation not directly correlated with changes in muscle mass was minimal. For example, increased protein intake may have increased creatinine levels (Lykken et al. 1980), but protein intake is closely tied to muscle mass and it would be impossible to separate these variables under field conditions. Creatinine excretion may have increased during the ovarian cycle of females

(Smith 1942), but based on approximate gestation length and interbirth intervals from long-term

107

data collected at this field site, the number of cycles females experience between weaning and their subsequent pregnancy is minimal (Carnegie et al. 2011b; Fragaszy et al. 2004; Sargeant

2014). Furthermore, reproductive state could affect the physical condition of females beyond the effects of food availability due to metabolic demands, and may also alter the proportional deposition and use of muscle versus fat (Butte and Hopkinson 1998; Kramer et al. 1993). It is also important to reiterate that these results were interpreted on a broad scale, using a number of urine samples collected over an extended period, to generate a slope based on the relationship between urinary creatinine and specific gravity. Ideally, researchers would collect enough samples to generate SG-Cr slopes per individual per time period (i.e., 25 samples per female per month) for comparative analysis.

Despite these minor shortcomings, this study was successful in measuring variation in the physical condition of wild female white-faced capuchins. These data also provide further evidence supporting the link between direct intergroup competition, home range quality and primate health in terms of body condition. Relative muscle mass estimation has also been successfully used in recent primate studies to show inter-individual variation in physical condition. In chimpanzees, variation in relative muscle mass corresponded to known differences in body size between males and females, and across different age groups (Emery Thompson et al. 2012). In the same population, the testosterone level of adult males was a significant predictor of relative muscle mass using the SG-Cr method, which is expected based on the role of testosterone in muscle growth and maintenance (Emery Thompson et al. 2012). Additionally, among redtail (Cercopithecus ascanius) and blue monkeys (C. mitis) in Kibale National Park,

Uganda, males in their early tenure had relatively higher muscle mass than late-tenure males, indicating that they were in better physical condition during periods closer in time to takeover

108

events (Brown et al. In Review). In the same study, urinary C-peptide data reflecting the energy balance of males corroborated the muscle mass results, in that late-tenure males also showed relatively lower energetic condition than did early-tenure males.

Given the invasiveness and logistical challenges associated with measuring body size, body weight and the muscle mass of wild animals, non-invasive measurement of specific gravity and creatinine to estimate muscle mass is a valuable tool for assessment of their physical condition. This is especially true when a study requires preservation of natural behavioral and physiological processes since invasive measures may disrupt normal activity patterns and induce stress responses. Non-invasive assessment of muscle mass provides key information to help better understand the range of intra-and inter-annual variation in the physical condition of individuals as well as the relationship among ecological variables (i.e., climate and food availability), social variables (i.e., inter- and intra-group competition and dominance relationships), physiological conditions (i.e., reproductive states and disease), and survival and reproductive success. It will also help to support results obtained through assessment of physiological factors via other means, such as the measurement of energetic condition and stress.

109

Chapter Five: Assessment of energy balance

5.1 Introduction

A key goal of behavioral ecology is to better understand the behavioral and physiological mechanisms underlying variation in reproductive fitness. While a number of social and ecological variables may contribute to an individual’s reproductive success, the proximate access to resources, subsequent nutritional gain, energetic output and resulting energy balance certainly play a central role in reproductive ability (Ellison and Valeggia 2003; Willis et al. 1996). Energy balance is an individual’s energy intake minus their metabolic expenditure including the added energetic demands of activity and reproduction. Measurement of energy balance provides key physiological information that may help to disentangle the influences of many factors on the overall fitness of individuals. These factors include ecological variables such as the seasonal availability and consumption of food resources (Emery Thompson et al. 2009b; Goldizen et al.

1988; Knott 1998; Koenig et al. 1997), or more extreme climate changes such as drought

(Richard et al. 2000), aspects of sociality such as dominance rank (Higham and Maestripieri

2014; Koenig and Borries 2006), and factors related to the physiological condition of individuals such as reproductive state (Georgiev 2012; McCabe and Emery Thompson 2013a) and illness

(Emery Thompson et al. 2009b).

Studies aiming to accurately measure the social behavior and health of animals avoid capturing and sedating techniques, as they interfere with normal activity and physiological processes. However, calculating energy balance noninvasively through measurement of energy intake and energy expenditure is very labor intensive and tends to provide a crude estimation due to the number of variables that may affect energy balance. Accordingly, a number of non-

110

invasive methodological approaches to measure the physical condition of wild animals have been introduced to the field of behavioral ecology. Direct quantification, such as baiting or provisioning scales to obtain body weight measurements while avoiding trapping and capturing animals, is accurate, yet also artificially affects the nutritional intake and physiology of the study animals. Qualitative visual assessment measures such as rating physical condition along a scale from meager to fat based on the visibility of bones (Berman and Schwartz 1988; Johnson and

Kapsalis 1995; Koenig and Borries 2006), or the use of photography and a visual scale, or parallel lasers, to estimate body size or condition (Kurita et al. 2012; Rothman et al. 2008b), provide estimations of changes in physical condition. However, these measures may not be applicable to all species based on species-specific morphology, body position, and the proximity of the researcher to the study animals. A highly refined measure of total energy expenditure can be obtained using the doubly labeled water technique, whereby water is labeled with hydrogen and oxygen isotopes and elimination rates are measured (Roberts 1989; Speakman 1998).

However, in using this technique it is essential to administer an exact dose to study subjects and obtain multiple subsequent samples (blood, saliva, or urine) in order to accurately calculate energy expenditure, which has been successful in captivity or when using more invasive methods

(Drack et al. 1999; Nagy and Milton 1979; Pontzer et al. 2014; Pontzer et al. 2010; Rosetta et al.

2011; Schmid and Speakman 2000; Simmen et al. 2010), but is logistically difficult to do in non- invasive studies of wild animals. Recently, energetics research has gained momentum in studies of wild animals, and includes the implementation of non-invasive methodology to directly measure the physiology of individuals using biomarkers. Specifically, the use of urinary C- peptide as a biomarker for insulin production in primate research has proven successful as an accurate indicator of energy balance (Deschner et al. 2008; Emery Thompson and Knott 2008;

111

Emery Thompson et al. 2009b; Girard-Buttoz et al. 2011; Grueter et al. 2014; Harris et al. 2010;

Lodge 2012; McCabe and Emery Thompson 2013a; Sherry and Ellison 2007).

5.1.1 C-peptide: a biomarker for energy balance

Insulin is a peptide hormone that controls metabolic activity primarily through the regulation of glucose, and is produced in the pancreas in the β-cells of the Islets of Langerhans. Glucose intake is widely known to promote the secretion of insulin to aid in glycolysis (i.e., the conversion of glucose to pyruvate for use as energy by cells) and storage of excess glucose as glucagon. However, ingestion of other dietary components such as amino acids and fatty acids promote insulin secretion to varying degrees as well, since they play an important role in energy homeostasis and the modulation of glucose via insulin secretion (Floyd et al. 1966a; Floyd et al.

1966b; Krezowski et al. 1986; Nuttall et al. 1985). Thus, insulin is an important factor in the absorption of amino acids and synthesis to protein in addition to its role in the metabolism of simple sugars.

During insulin production, C-peptide (i.e., connecting peptide) is cleaved from proinsulin and released in equimolar (i.e., equal) quantities to insulin from the pancreas in response to metabolic demands (Hoogwerf and Goetz 1983; Steiner 1984). While insulin undergoes hepatic and extra-renal clearance, the concentrations of C-peptide excreted in the urine are highly correlated to the insulin produced by the β-cells in both humans and chimpanzees (Meistas et al.

1982; Meistas et al. 1981; Sherry and Ellison 2007). In this regard, C-peptide has been an important biomarker for insulin production in clinical studies of human disease, such as diabetes.

Since insulin is one of the key physiological hormones regulating metabolism and energy storage, measurement of insulin, and thus C-peptide production, indicates the balance between energy intake and expenditure in relation to body condition (Havel 2001).

112

5.1.2 Application of UCP in non-human primate studies

Measurement of urinary C-peptide (UCP) provides a non-invasive means to infer insulin production and to obtain an accurate measure of energy balance (i.e., the relationship between energy intake and energy expenditure) in wild animals. Many primate studies have validated the use of UCP as an accurate measure of energy balance using human assay kits (Emery Thompson and Knott 2008; Girard-Buttoz et al. 2011; Sherry and Ellison 2007) because of the close homology between humans and monkeys. For example, C-peptide differs from humans by only one amino acid in green monkeys (Chlorocebus sabaeus; Peterson et al. 1972) and rhesus macaques (Macaca mulatta; Naithani et al. 1984). In captive great apes and Old World monkeys serum C-peptide levels have been confirmed to positively correlate with urinary C-peptide levels

(Sherry and Ellison 2007; Wolden-Hanson et al. 1993).

Low UCP also corresponds with periods of urinary ketone production, a more crude measure of energy shortage. Ketones are produced and expected to be present in urine when individuals are metabolically stressed and metabolizing fat stores, most specifically during nutritional deficit and severe carbohydrate shortages (Laffel 1999; Soskin and Levine 1944).

Therefore, individuals should produce ketones when they are experiencing negative energy balance. Indeed, UCP levels positively correlate with energy intake and negatively correlate with the presence of urinary ketones, showing on a broad level that UCP reflects energy balance in primates (i.e., evidence of metabolic stress and fat metabolism; orangutans: Emery Thompson and Knott 2008; chimpanzees and orangutans: Sherry and Ellison 2007; bonobos: Deschner et al.

2008; and evidence of illness in chimpanzees: Emery Thompson et al. 2009a) .

C-peptide has been used to show the relationship between variation in energy balance and a variety of intrinsic, ecological and social factors in numerous primate species. UCP

113

concentrations have been positively correlated with ecological measures of energy availability, including habitat quality, food availability and dietary quality (chimpanzees: Emery Thompson et al. 2009a; orangutans: Emery Thompson and Knott 2008; chimpanzees and orangutans: Sherry and Ellison 2007; bonobos: Deschner et al. 2008; olive baboons: Lodge 2012; mountain gorillas:

Grueter et al. 2014). UCP also indicates inter-individual variation in metabolic demands. In a study of black and white colobus monkeys, Harris et al. (2010) used C-peptide as a biomarker to show lower energy balance (and thus metabolic stress) in lactating females during periods of food scarcity. C-peptide has also been used to measure changes in energy balance across reproductive states to investigate flexibility in the dietary strategies of Sanje mangabeys

(Cercocebus sanjei) (McCabe and Emery Thompson 2013a). Finally, C-peptide has been used to indicate variation in energy balance related to social variables such as dominance rank and male-male competition in macaques and chimpanzees (Georgiev 2012; Higham and Maestripieri

2014; Higham et al. 2011b). Indeed, C-peptide has proven to be a useful biomarker for measuring the energetics of wild primates and is a promising tool to use in monitoring the nutritional and energetic state of primates living in their natural environments. However, the validity of its use has yet to be tested across the Order Primates, and has not been examined in wild New World primate species. The objective of this study is to determine if UCP tracks seasonal variation in energy balance among female white-faced capuchins (Cebus capucinus) living in the highly seasonal tropical dry forest in Sector Santa Rosa, Área de Conservación

Guanacaste, Costa Rica.

5.1.3 White-faced capuchins

White-faced capuchins live in cohesive social groups and their average group size is 16 individuals (Fragaszy et al. 2004). Male capuchins disperse, but females are largely philopatric.

114

Their distribution ranges from Honduras southward to the northern region of Ecuador. They live in a wide range of habitats and maintain an omnivorous diet comprised of fruit, seeds, invertebrates, flowers and opportunistically procured vertebrates (Fragaszy et al. 2004). In the tropical dry forests of Sector Santa Rosa, Costa Rica, fruit and invertebrates comprise the majority of the capuchin diet in terms of foraging time and caloric intake (Chapter 2). Due to extreme seasonal fluctuations in temperature and rainfall in Santa Rosa, these resources exhibit seasonal variation in availability, which affects the dietary composition and energy intake of capuchins across the annual cycle (Chapters 2 and 3). Ranging behavior, in terms of daily distance traveled and home range size, also shows temporal variation at this field site (Campos et al. 2014). These patterns suggest that there is considerable temporal variation in energy intake and energy expenditure that may lead to changes in the energy balance of individuals over the course of the year. A broad comparison of the physical condition of females using urinary creatinine excretion indicated that significant differences in relative muscle mass exist among groups and between months with high versus low fruit availability (Chapter 4). These findings indicate that there is great potential for the application of UCP to provide a more refined measure of temporal changes in energy balance.

5.1.4 Research Questions and Predictions

In this study I examine urinary C-peptide in female white-faced capuchins living in a Costa

Rican tropical dry forest. My aims are to determine whether UCP accurately reflects energy balance and to establish which factors influence variation in energy balance. Specifically, I investigate the following research questions:

1) Are calculated energy balance and UCP correlated?

115

2) Does variation in UCP predict the production of urinary ketones, a more crude measure

of energy balance?

3) Which ecological and social variables most strongly predict energy balance?

To achieve my research goals, I use behavioral observations of energy intake and expenditure, and compare calculated values of energy balance to values of C-peptide. I make the following hypotheses and predictions. If C-peptide is an accurate measure of energy balance, as indicated by previous primate research (H1), I predict that C-peptide levels will be positively correlated with energy balance, calculated as energy intake minus energy expenditure (i.e., basal metabolic costs plus the costs of activity and reproductive state) (P1).

Next, I determine whether low energy balance, as indicated by UCP, correlates to the presence of urinary ketones. If the presence of urinary ketones indicates negative energy balance in white-faced capuchins (H2), I predict that a) capuchin urine samples will contain ketones during months in which they are experiencing negative calculated energy balance (P2a) and b)

UCP levels will be lower in samples that test positive for ketone bodies, thus reflecting an overall negative energy balance (P2b).

Finally, I assess the role of ecological (climate and fruit availability), intrinsic

(reproductive state), and social (dominance rank) variables as predictors of variation in the energy balance of female capuchins. Specifically, I predict that there will be a positive relationship between UCP and monthly ripe fruit availability (P3) as well as dominance rank

(P4). I also predict that UCP will vary across reproductive states, with the highest values occurring during the pre-conceptive stage when energetic demands are expected to be the lowest, and the lowest values occurring during late gestation and early lactation when energetic demands associated with pregnancy and infant care are highest (P5; McCabe and Fedigan 2007).

116

5.2 Methods

5.2.1 Study Site

I collected data from three groups (LV, CP and GN) of wild white-faced capuchin monkeys living in the highly seasonal tropical dry forest of Santa Rosa Sector (SSR), Área de

Conservaciόn Guanacaste, Costa Rica, over three 4-month periods between September 2009 and

May 2011. Study groups are part of a long-term study population habituated to researcher presence (CP group since 1983, LV since 1990 and GN group since 2005). Ages are known for individuals in LV and CP groups; in GN group, ages were estimated based on physical features, reproductive parity, and comparison to females with known birthdates. I collected behavioral and ecological data as well as urine samples to measure C-peptide and other urinary parameters.

The number of adult female study subjects sampled per month ranged from 24-25 (LV = 5, CP =

10, GN = 9-10). I excluded nulliparous females, and females that disappeared or died during the study, from the dataset.

5.2.2 Reproductive state

I divided reproductive state into five periods. Pregnancy was inferred based on infant birth dates and mean gestation time (푥̅ = 158 days) for females in this population (Carnegie et al. 2011b).

First, I divided the mean gestation time into two periods of equal length: early gestation (EG) and late gestation (LG), the latter of which is expected to entail higher metabolic demands based on the relatively larger size of the fetus. Then, I determined the duration of lactation based on infant nursing rates. Based on the nursing patterns and estimated conception dates during this study, females were able to conceive once nursing rates fell below one bout per hour, at which point I considered them “pre-conceptive” versus “lactating”. I divided lactation into two periods.

During the first period, early lactation (EL, birth to 14 weeks), the metabolic demand of infant

117

care is estimated to be higher. This estimate is based on higher nursing rates and longer nursing durations, the larger proportions of time mothers spend carrying their infants, and because the mothers’ energy intake rates remain at or above those at the time of conception (McCabe and

Fedigan 2007; Sargeant 2014). During the second period, late gestation (LL, > 14 weeks), the mothers’ rates of nursing, infant carrying and energy intake drop dramatically (McCabe and

Fedigan 2007; Sargeant 2014). Finally, I considered pre-conceptive females as those who were neither pregnant (based on the day of parturition and the mean length of gestation) or lactating

(based on behavioral observation of the frequency of nursing bouts). This period, prior to conception during which females were neither pregnant nor lactating, ranged from one to four months during this study, but I was unable to calculate the mean length based on the small sample size and gaps between data collection periods. In sum, the five reproductive periods I included were pre-conception (PC), early gestation (EG), late gestation (LG), early lactation

(EL) and late lactation (LL).

5.2.2.1 Rank

I used agonistic behavior collected during 10-minute focal animal samples (Altmann 1974) in conjunction with data collected ad libitum to construct dominance hierarchies using the I&SI method in MatMan 1.1 (Noldus Information Technology 1998). I calculated a standardized rank

(Ranks) that ranges from zero, which represents the lowest-ranking female, to one, which represents the highest-ranking female. Here, I used Formula 1, whereby R is the numerical rank and N is the number of females in the group:

Formula 1

푅 − 1 푅푎푛푘 = | − 1| 푠 푁 − 1

118

5.2.2.2 Energy intake

Details of nutritional analyses are provided in Chapters 2 and 3. I calculated monthly energy intake rates per female using intake data recorded during focal animal samples and Formula 2.

In Formula 2, EIr refers to the energy intake rate for one female, Ni is the number of items ingested for one food type during the focal animal sampling period, Vi is the nutritional value

(kJ) of that food item, and t is the amount of time for which I sampled the female.

Formula 2

푘퐽 ∑푛 푁 × 푉 퐸퐼 ( ) = 푖=1 푖 푖 푟 ℎ푟 푡

5.2.2.3 Energy expenditure

I calculated energy expenditure as the estimated metabolic costs to females based on basal metabolic rate and body size plus the added costs of performing different activities, including the added demands of daily travel; the specific steps of these calculations are outlined below. Basal metabolic rate (BMR) is the amount of energy required by an organism in an inactive and thermoneutral state within a 24-hour period (Durnin and Passmore 1967). I calculated a daily

12-hour BMR for females as 294.63 kJ using the average female body weight published by

Smith and Jungers (1997) of 2.54 kg for white-faced capuchins and the Kleiber (1961) equation converted to kJ/12 hours, where M is body mass in kilograms (Formula 3).

Formula 3

0.75 퐵푀푅푑 = 146.44 × 푀

I also calculated the durations of different activity states (i.e., rest, stationary feed, active feed/forage, travel, low-intensity social and high-intensity social), which were recorded

119

continuously during focal animal samples. Activity-specific energy constants adjusted to primate metabolic demands and derived from human metabolic research have been used in numerous studies of mammals and primates to estimate the energetic cost (i.e., energy expenditure) of different types of activities (Coelho 1974; Coelho et al. 1976; Key and Ross

1999; Leonard and Robertson 1997; Taylor et al. 1970). To calculate the energetic costs of different activities for this study, I assigned energy constants to the capuchin activity state categories noted above based on the estimated energetic demand of each activity using a combination of these literature values (Appendix J). Following Leonard and Robertson’s (1997) adaptation of values published by Coelho (1974) and Coelho et al., (1976), I assigned the energy constant 1.25 to resting and 1.38 to stationary feeding state behaviors that require limited movement for processing such as foraging pith and vertebrates. I assigned the higher energy constant of 1.7 to calculate the basal cost of traveling, but see Formulas 5 and 6 for the calculation of additional energetic travel costs (Taylor et al. 1970). I also assigned the value of

1.7 to active feeding and foraging behaviors that involve considerable movement between feeding sites such as foraging fruit, insects, flowers and visually foraging. I assigned the energy constant 1.38 to the low-intensity social state of grooming, as this behavioural state comprises a minimal level of hand movement and travel between social partners (akin to stationary feeding).

I assigned high-intensity social behaviors including affiliation, agonism, play, sex and fur- rubbing the value of 2.35 proposed by Leonard and Robertson (1997) and adapted from Coelho

(1974) and Coelho et al., (1976). I used energy constants in conjunction with activity budgets measured from state changes recorded during focal follows (proportion of time spent in the activity state) to calculate the energetic cost of all activity states (kJ/12 hours) using Formula 4.

120

Formula 4

푬푬풔풕풂풕풆 = 퐵푀푅푑  [ퟏ. ퟐퟓ × (푻풓풆풔풕) + ퟏ. ퟑퟖ × (푻푺풇풆풆풅 + 푇퐿푠표푐푖푎푙) + ퟏ. ퟕ × (푻푨풇풆풆풅 + 푻풕풓풂풗풆풍)

+ ퟐ. ퟑퟓ × (푻푯풔풐풄풊풂풍)]

I carried a Garmin GPS 76 handheld receiver and recorded systematic waypoints at 30- minute intervals to measure group location. I turned the unit on upon making contact with the group and I recorded a final point at the end of each observation day. Using these data, I calculated the daily path length for each group. When the group was not followed for a full day,

1) I used the mean rate of travel during that day to infer the missing distance, or if more than two hours of data was missing, 2) I used the mean travel distance for the other observation days in that month of sampling. I used the same daily values for all females within the same study group. Since white-faced capuchins maintain a high level of cohesion (group spread diameter of approximately 80 meters, Janson and Vogel 2006), and because observers maintain a distance of

5-10 meters from the animals so not to alter their behavior, error across females is likely to be minimal in the context of the total distance traveled per day. I calculated the energetic cost of travelling (quadrupedal locomotion) beyond the basal travel expenditure using two formulas, following Campos et al. (In Prep). Formula 5 calculates the “incremental cost of transport”

(ICT) to account for the energy expended during movement as a function of body mass, beyond basal travel costs, as 0.02 kJ/m for females (Altmann 1987; Taylor et al. 1982).

Formula 5

퐼퐶푇 = 0.0105 × 푀0.702

Using Formula 6, I multiplied ICT by the group day range in meters (RD) to account for the daily distance traveled (Altmann 1987; Taylor et al. 1982). Following Janson (1988), I then 121

multiplied this value by 5.64 to account for vertical movement within the canopy resulting from arboreal travel (twice the day range since the original formula was based on terrestrial travel) and the estimated distance moved by the individual in excess of the group’s day range (2.82 times the center of the group for Cebus apella).

Formula 6

퐸퐸푡푟푎푣푒푙 = 퐼퐶푇 × 푅퐷 × 5.64

Finally, to obtain the total daily energy expenditure in kJ/12 hours, I summed the energy expenditure for activity states (EEstate) and travel (EEtravel) and multiplied the result by the reproductive scaling factors (srep) of 1.00 for cycling females, 1.25 for gestating females and 1.50 for lactating females. I used these scaling factors to account for the increased metabolic demands associated with each reproductive state in Formula 7 (Key and Ross 1999). I divided this value by 12 to get estimated hourly energy expenditure per day, and then calculated an average for all days per round (i.e., sampling period per month) of data collection. Finally, I calculated mean monthly energy balance as the mean energy intake rate minus the mean energy expenditure rate.

Formula 7

퐸퐸푡표푡푎푙 = (퐸퐸푠푡푎푡푒 + 퐸퐸푡푟푎푣푒푙)

It is important to note that since the energy constants I used are broad estimates extrapolated from human studies that are likely inaccurate measures of true energy expenditure by female capuchins. Additionally, behaviors were lumped into broad categories and more finite behaviors within each category may differ slightly in energetic costs. However, because the

122

same values are applied to all females, these methods should provide unbiased estimates of inter- and intra-individual variation in energy expenditure within this population.

5.2.3 Ecological Data Collection

I measured monthly ripe fruit availability using phenological data on the coverage and maturity of ripe fruits (N = 30 species) collected during this study, in combination with biological transects used to record the density of tree species (Melin et al. 2014a), to calculate the sum of fruit biomass (kg) per species divided by the transect area (ha) (Chapter 2). I calculated the ripe fruit energy density per group home range by multiplying the ripe fruit biomass by the species- specific values for energy (kJ per gram of fresh fruit), summing all fruit species, and dividing by the transect area (Chapters 2 and 3).

5.2.4 Urine collection

I collected urine samples opportunistically from all female study subjects either by catching urine on plastic bags attached to extended hoop structures or by pipetting it from vegetation

(Emery Thompson and Knott 2008). I collected an average of 1-3 urine samples per female per monthly collection round, spaced as equally as possible over each 4-6 day observation period, yielding 825 urine samples that contained sufficient volume for C-peptide analysis. I determined the presence/absence of ketone bodies in all samples for which sufficient volume was collected to perform a urinalysis test as well as further C-peptide analysis (N = 562). Ketones were measured in each urine sample using urinalysis reagent strips either immediately upon sample collection or after data collection finished the same day the sample was collected (Siemens

Multistix 10 SG, Appendix E). A small amount of urine was pipetted onto the test site of the reagent strip to measure the level of ketone bodies present as a broad indication of the degree of ketosis (produced in the process of fat metabolism).

123

5.2.5 Laboratory measurement of C-peptide

I analyzed UCP concentration in the Hominoid Reproductive Ecology Laboratory at the

University of New Mexico under the training and supervision of Dr. Melissa Emery Thompson. I measured C-peptide using commercially available human radioimmunoassay kits with a sensitivity range of 100 pg/ml to 5,000 pg/ml (RIA Human C-peptide Kit, Millipore Corporation,

Billerica, MA) (Emery Thompson and Knott 2008; Emery Thompson et al. 2009b; Sherry and

Ellison 2007).

I set up the C-peptide assays over two days following the Millipore kit procedure. On day one, I pipetted 100 µl of each urine sample, 100 µl of each standard (100, 200, 500, 1000,

2000 and 5000 pg/ml purified human C-peptide in assay buffer), and Controls I and II (low and high concentrations of purified human C-peptide in assay buffer) in duplicate into numbered tubes. If the urine sample volume was too low to run using 100 µl in duplicate, I diluted the sample and ran 50 µl of urine and 50 µl of assay buffer in duplicate (1:2 ratio) and subsequently corrected the calculated concentration for this dilution. I prepared non-specific binding controls, which act as zero standards without antibody, by pipetting 200 µl assay buffer (0.05M phosphosaline pH 7.4 containing 0.025M EDTA, 0.08% sodium azide, 1% RIA grade BSA) in duplicate. To assess total counts, I prepared tubes with 100 µl assay buffer and 100 µl 125I-

Human C-peptide tracer in duplicate. Using the repeater pipette, I pipetted 100 µl 125I-Human C- peptide tracer into all standard, control and sample tubes. Next, using the repeater pipette, I pipetted 100 µl of human C-peptide antibody (guinea pig anti-human C-peptide antibody in assay buffer) into all standard, control and sample tubes (excluding the non-specific binding control and reference tubes). I vortexed all tubes, covered the tray with foil, and incubated the samples at 4 ºC overnight. At this stage, the 125I-labeled human C-peptide antigen in the tracer

124

competes for antibody binding sites with the unlabeled C-peptide antigen present in the capuchin urine sample.

On day two, using the repeater pipette, I pipetted 1 ml of anti-rabbit precipitating reagent

(goat anti-guinea pig IgG serum, 3% PEG and 0.05% Triton X-100 in 0.05M phosphosaline,

0.025M EDTA, 0.08% sodium azide) into all tubes except the reference tubes. The precipitating reagent is a second antibody that binds to the original antibody (capturing all bound C-peptide).

I vortexed all tubes and incubated the samples for 20-25 minutes at 4 °C. Next, I centrifuged samples at 4 ºC and 3000 xg for 60 minutes to form a pellet containing the precipitating reagent with bound C-peptide. Using a vacuum, I immediately suctioned the supernatant from the pellet.

All tubes were counted in a gamma counter for two 1-minute counts to obtain a mean Iodine-125 count value. The count reflects the amount of 125I-human C-peptide present in the pellet. Thus, there is an inverse relationship between the Iodine-125 radioisotope count value and the capuchin C-peptide concentration, whereby the greater the concentration of C-peptide in the capuchin sample, the more competitively it binds to the antibody, consequently decreasing the ability of 125I-human C-peptide to bind to the antibody sites. The reference standards were used to calibrate the counts to estimations of C-peptide concentration in pg/ml.

I calculated the coefficient of variation for each C-peptide sample duplicate as the standard deviation of C-peptide concentrations from the two duplicate tubes divided by the mean determination, multiplied by 100. Samples were re-assayed if the CV exceeded 15% for sample concentrations > 250 pg/ml and 25% for sample concentrations  250 pg/ml due to the smaller margin of error at lower values. I tested for intra-assay reliability by calculating the mean of all sample CVs for each assay. The intra-assay coefficients of variation were 11.9% for low ( 250 pg/ml) and 7.1% for high (> 250 pg/ml) samples, respectively. I tested for inter-assay reliability

125

using Control I and Control II concentrations by calculating the average concentration for each control per plate and then calculating the percentage CV across all plates per control. The inter- assay coefficients of variation were 10.8% and 9.4% for Controls I and II, respectively. In some cases of samples with very low C-peptide, only one replicate of the sample produced a result, while the other replicate was below the sensitivity of the standard curve. While I could not calculate a CV for such values, I inspected the counts of the two vials to be sure that they were similar; because these values were already at the low end of the range of variation, small deviations had little potential to skew the analysis. To avoid artificially excluding samples with low C-peptide values, if a sample value fell below assay sensitivity but had a sufficient creatinine level (Cr  0.100, indicating that the sample was not low because it was too dilute), it was assigned the value of minimum assay sensitivity of 100 pg/ml (Deschner et al. 2008; Girard-

Buttoz et al. 2011).

5.2.6 C-peptide standardization

Urinary analytes are most commonly standardized using creatinine (Cr), whereby the analyte concentration, in this case UCP (pg/ml), is divided by creatinine concentration (mg/ml) to control for variance in the time between urination events, the dilution of urine samples from water consumed by the animal, or possible contamination by rain water (Taussky 1954).

However, previous analysis of urinary parameters in this population indicated that creatinine concentration, which fluctuates according to the relative proportion of skeletal muscle per individual, showed significant seasonal variation that was independent of urine concentration as indicated by specific gravity (which is free from muscle mass effects). Therefore, the UCP standardization using creatinine may have inflated UCP values during periods of decreased muscle mass, during which time creatinine excretion was reduced. Instead, I standardized the

126

UCP assay values to account for the dilution of the urine using the sample’s specific gravity, the ratio of the density of the urine sample compared to the density of water. Specific gravity is prevalent in clinical research and is a good and cost-efficient alternative measurement to creatinine to standardize analytes in endocrine research (Haddow et al. 1994). I used Formula 8 to standardize UCP (Miller et al. 2004), whereby UCPASSAY is the raw value obtained from the

C-peptide assay, SGPOP is the population mean specific gravity value, and SGSAMPLE is the specific gravity value obtained from each specific sample.

Formula 8

푆퐺푃푂푃 − 1 푈퐶푃푆퐺 = 푈퐶푃퐴푆푆퐴푌  푆퐺푆퐴푀푃퐿퐸 − 1

I measured the specific gravity of 100 l aliquots of urine using a refractometer (Atago

PAL-10S) with a resolution of 0.001. Specific gravity values ranged from 1.001 to 1.045 and the population mean was 1.023 (N = 734 samples). Standardized C-peptide concentration (N = 535) ranged from 90.24 pg/ml to 1134.25 pg/ml (푥̅ = 241.08 pg/ml). Since the C-peptide concentration in each urine sample is likely to reflect the energy balance of individuals very near to the time of collection, C-peptide is expected to fluctuate throughout the day based on the feeding patterns of individuals. This pattern of diurnal variation according to food ingestion was noted in the Kanyawara chimpanzee population (Georgiev 2012). Unfortunately, I was not able to collect enough urine samples per female per day to test this; however, I used the monthly mean UCPSG per female in all analyses in an attempt to average the effects of variation caused by the proximity of sample collection to feeding time. There did not appear to be any systematic bias in the time of sample collection that might alter the results with respect to the variables used in analysis.

127

5.2.7 Data Analysis

To compare calculated versus biomarker values of energy balance, I ran a non-parametric

Spearman’s rho rank correlation to measure the strength of the relationship between mean monthly calculated energy balance and mean monthly UCPSG. To determine whether UCPSG measures were predictive of urinary ketone presence, I grouped urine samples into those that tested ketone “positive” versus “negative” and ran a generalized estimating equation with urine samples (n = 409; n = 372 ketone-negative, n= 37 ketone-positive, range = 7 to 25 samples per female) as the unit of analysis, ketone presence as a binary logit response variable, UCPSG

(log10UCPSG, log transformed for normality) as a predictor variable, and Female ID (N = 25) as a repeated subject to account for individual variation and repeated sampling. These statistical analyses were conducted using IBM SPSS Statistics 21.0.0.2 software.

Using R Project Software version 3.1.2 for Windows (R Development Core Team 2014) and R package lme4 (Bates et al. 2014), I fit linear mixed-effects models with maximum likelihood estimation to test whether variation in UCPSG related to each predictor variable. Data points represented monthly means and I included the mean standardized UCP value per female per month (N = 210) as the response variable, which was log transformed to normalize the distribution (log10UCPSG). I included 1) monthly ripe fruit energy density (kJ/ha); 2) mean monthly maximum temperature (°C); 3) reproductive state (PC, EG, LG, EL, LL); and 4) standardized dominance rank (0 – 1) as fixed effects. Although it was not the focus of this paper,

I included monthly maximum temperature as a climatic variable because it may influence energy expenditure in terms of thermoregulation and travel. I included Female ID and Group ID as random effects to control for individual variation, group differences and repeated sampling.

Before fitting the models, I transformed the continuous independent variables into unitless Z

128

scores to improve model convergence as well as to aid in the direct comparison of model estimates and the relative importance of each predictor. I also used the R package MuMIn

(Bartoń 2014) to perform automated model selection based on corrected Akaike’s Information

Criterion scores (AICc); the null model included only the response variable and random effects.

5.3 Results

5.3.1 The relationship between calculated energy balance and UCP

The Spearman's rank-order correlation indicated that there was a significant positive correlation between mean calculated energy balance and mean monthly UCPSG concentration (rs =

0.248, p < 0.001,  = 0.01, two-tailed). Figure 5.1 shows variation in these two measures by month.

Figure 5.1 Monthly mean ( SE) urinary C-peptide (black line) and monthly mean ( SE) daily energy balance (grey dashed line). Grey shading indicates months with urinary ketone production.

129

5.3.2 A comparison of UCP and ketone production

Ketones were only present in urine samples during the June and July study months. These were two of the three study months during which females experienced the lowest mean calculated energy balance and UCPSG concentrations (Figure 5.1). Of the samples collected during June and July, 24.06% (N = 45 of 187) were ketone-positive. Forty-four percent of the female study subjects (11 of 25: SH, ED, KI, TI, BO, ZA, SE, NE, FL, PD, MW) produced urine that contained ketones (June: 31 of 97 samples, 31.96%, ten females; July: 14 of 90 samples, 15.56%, eight females) and all females were members of CP and GN groups (eight and three females, respectively). Additionally, all females that produced ketones were lactating. According to the generalized estimating equation, there was a statistical trend for samples with lower C-peptide concentrations to also be ketone-positive (Table 5.1).

Table 5.1 Generalized estimating equation results predicting the effect of urinary C-peptide on the presence of urinary ketones. 95% Wald Confidence Interval Hypothesis Test B Standard Error (SE) Lower Upper Wald 2 df p value (Intercept) 3.283 3.0357 -2.667 9.233 1.170 1 .279 log10UCPSG -2.393 1.3015 -4.944 .158 3.381 1 .066

130

5.3.3 Ecological and social predictors of urinary C-peptide

Only one model produced results that were better than the null model, where only the random effects of Female ID and Group ID were included (Table 5.2). The selected model contained the predictor variable of monthly ripe fruit energy density and the random effects of Female ID and

Group ID. Monthly ripe fruit energy density had a positive effect on UCPSG (intercept = 2.375, estimate = 0.034, SE = 0.009, df = 208, t = 3.575, p < 0.001, 95% CI = 0.015 to 0.052, n = 210, female = 25, group = 3). Variation in mean ripe fruit energy density and mean urinary C- peptide are plotted by month in Figure 5.2.

131

Table 5.2 Model set and model selection results for urinary C-peptide.

Estimates Rank Model Intercept RS ED Rank T df LLV AICc AIC w (AIC) max i 1 Null Model + ED 2.375 0.03383 5 111.028 -211.8 0 0.760 2 Null Model 2.375 4 108.539 -208.9 2.88 0.180 3 Null Model + Tmax 2.375 0.02422 5 107.941 -205.6 6.18 0.035 4 Null Model + ED + Tmax 2.375 0.02878 0.01155 6 107.995 -203.6 8.19 0.013 5 Null Model + ED + Rank 2.375 0.03352 -0.00880 6 107.721 -203.0 8.73 0.010 6 Null Model + Rank 2.375 -0.01001 5 105.354 -200.4 11.35 0.003 7 Null Model + Rank + Tmax 2.375 -0.00856 0.02370 6 104.614 -196.8 14.95 0 8 Null model + ED + Rank + Tmax 2.375 0.02869 -0.00830 0.01109 7 104.641 -194.7 17.04 0 9 Null Model + RS + ED 2.343 + 0.03946 9 102.837 -186.8 24.99 0 10 Null Model + RS 2.379 + 8 99.419 -182.1 29.64 0 11 Null Model + RS + Tmax 2.361 + 0.02583 9 99.043 -179.2 32.58 0 12 Null Model + RS + ED + Tmax 2.339 + 0.03398 0.01316 10 99.977 -178.8 32.92 0 13 Null Model + RS + ED + Rank 2.340 + 0.03871 -0.00986 10 99.633 -178.2 33.6 0 14 Null Model + RS+ Rank 2.375 + -0.01117 9 96.349 -173.8 37.96 0 15 Null Model + RS+ Rank + Tmax 2.358 + -0.00941 0.02521 10 95.802 -170.5 41.26 0 16 Null model + RS + ED + Rank + Tmax 2.337 + 0.03357 -0.00915 0.01258 11 96.707 -170.1 41.68 0

Fixed effects were converted to Z scores. ED: ripe fruit energy density (kJ/ha); Tmax: maximum temperature; Rank: standardized dominance rank; RS: reproductive state (PC, EG, LG, EL, LL); LLV: log likelihood value; compared to the best model; wi(AIC): Akaike weight. The null model included the intercept and the random effects of Female ID and GroupID.

132

5.4 Discussion

Urinary C-peptide has been an effective biomarker for non-invasive assessment of energy balance in wild primates. The ability to monitor variation in UCP has helped to track how groups respond to temporal and spatial variation in ecological factors such as seasonal fluctuation in food availability (Emery Thompson and Knott 2008; Emery Thompson et al.

2009b; Sherry and Ellison 2007) and differences in habitat quality (Grueter et al. 2014; Lodge

2012). It has also helped to measure the physiological consequences of reproduction (Harris et al. 2010; McCabe and Emery Thompson 2013a) and the differential effects of resource

133

competition (Georgiev 2012; Higham and Maestripieri 2014; Higham et al. 2011b). My goal in this study was to compare UCP concentrations with energy balance (as calculated based on behavioral observations) to determine whether it is feasible to apply this methodology to the study of wild New World primates, and to determine how various intrinsic, ecological and social variables predicted variation in UCP. My results show significant relationships between UCP and calculated energy balance, ketone production, and energy availability, indicating its potential strength as an analytical tool in this species. This is the first empirical study to utilize UCP to investigate variation in the energy balance of capuchin monkeys and one of the first to study energy balance in New World primates.

5.4.1 Calculated energy balance versus UCP

As predicted, there was a significant positive correlation between mean monthly energy balance

(calculated as energy intake minus energy expenditure) and mean monthly UCP per female (P1), indicating that measurement of energy gained from food intake minus the estimated energy expenditure during activity, including travel, aligns with insulin physiology in capuchins.

Although the correlation was slightly lower than expected given that both are measures of energy balance, there are a number of methodological factors that may provide sufficient explanation for this finding. Here, I note factors that may have influenced the strength of the correlation between these two variables and which further support my assertion that the use of a physiological biomarker (rather than estimation based on behavioral observation alone) minimizes the potential error in measuring energy balance.

First, the calculation of energy balance is an estimated versus true value. Since all-day follows of each female were not feasible due to time constraints and the need to sample all females, I was unable to measure the exact energy intake and the proportion of time spent in

134

activity states per day per female per sampling period. Rather, I measured these variables using

10-minute focal animal samples and used these data to extrapolate mean daily energy intake and expenditure values. Next, the activity-specific energy constants assigned to activity states and used to calculate energy expenditure are not specific to capuchins, have not been determined more broadly for non-human primates, and thus were derived from human metabolic research and adjusted to primate metabolic demands (Coelho 1974; Coelho et al. 1976; Leonard and

Robertson 1997; Taylor et al. 1970). Also, although I included a correction factor for reproductive state in the calculation of energy balance, there is a lack of empirical data regarding the metabolic demands of gestation and lactation in the primate literature, and so these correction factors are only estimates (Key and Ross 1999; National Research Council 2003). Finally, there are a number of metabolic factors related to energy expenditure that cannot be included in the calculation, including the metabolic costs of digestion, thermoregulation and illness. Any differences in estimated energy expenditure from the true values may explain differences between calculated and biomarker values of energy balance.

Potential error in biomarker energy balance values is likely closely tied to sampling frequency and total sample size. During this study, urine samples were spot-collected from females whenever possible, but there was a large amount of variation in the number of samples I obtained per female per month as well as the yield per sample. This variation was due to factors such as seasonal fluctuation in the availability of water and the hydration of the females, the canopy height from which urination events occurred, and the presence of urine washing behavior

(i.e., a behavior whereby capuchins rub the urine on the palmar surfaces of their hands and feet;

Campos and Fedigan 2013), all of which reduced the volume of urine that I was able to collect.

Next, the UCP concentration was sensitive to short-term effects of food consumption in

135

chimpanzees, suggesting that repeated sampling is necessary to accurately capture long-term energy balance (Georgiev 2012). Therefore, the monthly mean UCP values for capuchin females in the months during which I was unable to collect a large number of samples may not be as reflective of the mean concentration as other months, based on the proximity in time during which the samples were collected in relation to feeding events. Finally, because I used C-peptide assay kits developed for humans, it is possible that there was a relatively lower affinity for capuchin C-peptide, despite its inferred similarity to the green monkey amino acid sequence

(Peterson et al. 1972) and the high affinity found in great apes (Sherry and Ellison 2007).

The list above highlights potential sources of error, the difficulty in calculating energy balance using behavioral data, and the importance of using a large set of urine samples to accurately quantify UCP and measure individual variation in energy balance. Overall, it is important to remember that the significant correlation found between calculated and biomarker values despite these sources of error, highlights the strength of UCP as an analytical tool. This physiological biomarker reflects, in one variable, the metabolic outcome of the behavioral measures (e.g., energy intake, basal metabolic rate, and metabolic costs of various activities including travel) and others that may affect energy balance but are difficult to quantify, including thermoregulation and the metabolic costs of illness.

5.4.2 UCP as a predictor of ketone production

In support of my second prediction (P2a), ketone-positive samples were found during two of the twelve study months, June and July 2010, during which females experienced negative energy balance according to both the calculated and biomarker energy balance values. Although UCP was not a significant predictor of urinary ketones in individual samples, the statistical trend showing a decrease in the likelihood that a sample was ketone-positive as UCP concentration

136

increased, suggests a negative relationship between the two variables (P2b). The most probable explanation for the lack of significant results is simply that ketones are not as sensitive a measure of energy shortage as that reflected by C-peptide, and may be specific to the restriction of certain nutrients such as carbohydrates (Weinhouse 1952; Werk et al. 1955). However, other sources of error cannot be ruled out. Given the small ketone-positive sample size, there may not have been enough statistical power to provide a clear result. Also, although it is unlikely that the urinalysis strips produced false positives, since the results were supported by multiple samples from the same females, it may be that there were false negatives, as both pH and humid conditions can affect the sensitivity of the test sites (Siemens Multistix 10 SG product insert). Furthermore, because the ketone test does not take the concentration of the urine sample into account, it is possible that many of the urine samples were too dilute to react on the test site. It is also possible for acetone to affect the specific gravity of urine samples, potentially making them appear lighter

(acetone is less dense than water) (George 2001; Miller et al. 2004). If this were the case, because acetone is a ketone, this may have confounded the true relationship between ketone production and C-peptide excretion, which was standardized using specific gravity.

Although these potential sources of error are possible, other data collected during this study support the pattern of ketone production found during June and July. First, analysis of the relationship between creatinine production and the specific gravity (i.e., density) of urine indicates that females experienced a decrease in relative muscle mass during low fruit months, a subset which included the June and July study months, providing further evidence that females were potentially metabolizing fat stores during this period of negative energy balance (Chapter

4). Next, under normal metabolic conditions, most ketones are metabolized and do not appear in the urine; it is under hypoglycemic conditions that the body switches to fat stores and produces

137

ketones in such high concentrations that some are excreted in the urine (Laffel 1999).

Interestingly, at this site, this time period coincides with a drop in fruit abundance and a seasonal outbreak of Lepidopteran larvae (i.e., caterpillars), which are high in fat and protein but low in carbohydrates per unit dry mass. The ripe fruit energy density (푥̅ = 274,982  18,772 kJ/ha) and calculated female energy balance (x = -399 kJ/day  60) were predictably lower during June and

July than most other study months (9, 674,352 kJ/ha  95,900 and 704  94 kJ/day, respectively), supporting the production of ketones when carbohydrate intake may have been low

(this chapter and Chapter 3). Data collected during the month of December contradicts this pattern. Levels of ripe fruit energy density and UCP were similar in December to levels during

June and July, yet ketones were not produced. This may be because in December, the capuchins exploited alternate sources of carbohydrates in flowers (Hogan 2015), and may not have reached the hypoglycemic conditions that they potentially did in June and July when the only alternate resource was carbohydrate-poor caterpillars.

5.4.3 Predictors of energy balance (UCP)

Of the intrinsic, ecological and social variables tested, the mean monthly ripe fruit energy density was a significant predictor of energy balance. This supported my prediction (P3) that fruit availability would be an important predictor of variation in energy balance since a large proportion of the capuchin diet is comprised of fruit. Indeed, energy balance was relatively higher among females during months with high ripe fruit energy density and energy balance dropped during the low fruit months of June – July and November – December. It is likely that this drop represented both a reduction in caloric intake as well as an increase in energy expenditure through search effort and travel to find the more limited fruit sources among other food types. It is important to note that the capuchins are frugivore-faunivores, and although fruit

138

is arguably most important, other food sources such as invertebrates also comprise a large proportion of the capuchin diet in terms of energy and macronutrient intake. The dietary importance varies seasonally according to the availability of fruit and the biological life cycles of invertebrates (Chapter 3). Therefore, a measure of food abundance that included all food types

(e.g., fruit, invertebrates, flowers and pith) would likely be an even stronger predictor of energy balance in this species. The effect of the consumption of other types of food on energy balance may explain any weakness in ripe fruit energy density as a predictor variable in this study.

Using both calculated and biomarker measures of energy balance, I identified two periods during which female capuchins were vulnerable to and experienced negative energy balance – during the early rainy season (June and July) and during the transition between the late rainy and dry season (November and December). This vulnerability coincides with periods of low fruit abundance. Although there is likely considerable inter-annual variation in fruit abundance at this field site, multi-year data (2007 – 2013, N = 81 months) on fruit phenology show the same bimodal pattern in the fruiting cycle, and similar timing (Campos et al. 2014). During the June and July drop in ripe fruit abundance, there is a predictable emergence of Lepidopteran larvae

(i.e., caterpillars) that the capuchins use as their primary food source. Caterpillars were heavily eaten at this time during this study, and yet females still experienced very low energy balance and produced urinary ketones. It is possible that there are years in which fruit and/or caterpillar abundance is higher, during which energy balance would also be relatively higher. However, if during this period either the fruit or caterpillar resources drop below the level measured during this study, there is the potential for females to experience severe metabolic consequences that may affect reproductive success and potentially their own survival. During the later annual drop in fruit abundance in November and December, caterpillars are not as abundant, but the

139

capuchins supplement their energy intake with flowers and pith (Hogan 2015) and they increase the amount of time spent extractively foraging invertebrates (Melin et al. 2014b). Thus, variation in the abundance of these alternate resources also has the potential to drop energy balance to critical levels and jeopardize the health and survival of capuchins in this population.

Contrary to predictions, dominance rank (P4) was not a significant predictor of energy balance. Female white-faced capuchins in the Sector Santa Rosa study population exhibit matrilineal, linear and stable hierarchies, and a large proportion of dominance interactions occur over food resources (Bergstrom and Fedigan 2010). Accordingly, I expected that dominant females would have higher energy balance because they may gain access to higher quality resources or maintain control over optimal foraging positions within food trees for longer periods of time. Although dominant females tend to remain more central within the social group (Hall and Fedigan 1997), mid- and low-ranking females could maintain the same level of energy balance by traveling in front of the group to exploit resources before the group arrives. It is also possible that their more intermediate and peripheral positions include access to additional food trees or to resources that are different from those accessed by individuals at the center of the group (e.g., mid- to low-ranking females exploit more low-quality resources than high-ranking females to achieve the same energy balance). This finding that energy balance was not significantly better in high-ranking than low-ranking females may help to explain why dominant females have not been found to exhibit higher fecundity and infant survival rates in this population in previous studies (Fedigan et al. 2008). As a next step, I plan to assess in greater detail how different behavioural variables such as a female’s position within the group, the type of foods that she consumes, and the amount of time spent foraging versus socializing, may lead to the relatively equal energy balance exhibited across females of varying dominance ranks.

140

Also contrary to predictions, reproductive state (P5) was not a significant predictor of energy balance. Metabolic costs are expected to be highest in mammals during early lactation when an infant is still growing and is highly dependent on the mother for energy, and lowest during the pre-conceptive stage when females are free from the energetic demands of offspring

(Clutton-Brock et al. 1989; Key and Ross 1999). I expected energy balance to reflect metabolic costs, whereby females in early stages of lactation would have the lowest energy balance and pre-conceptive females would have the highest energy balance, but there was not a significant effect of reproductive state on variation in energy balance. It is possible that reproductive state was not a significant predictor of energy balance in my study due to my small sample size.

However, a previous study of the feeding behaviour and nutritional intake of females in this population indicated that lactating females consumed significantly more energy per unit time by eating food at faster rates than other females (McCabe and Fedigan 2007). Thus, it is also possible that modifications to feeding and foraging behavior account for the lack of differences in energy balance among reproductive categories. Finally, it is likely that changes in insulin sensitivity accompanying pregnancy and lactation make it difficult to directly compare energy balance among females of different reproductive states (Bell and Bauman 1997; Catalano et al.

1991). For example, if lactating females are more metabolically stressed but also somewhat insulin resistant (which helps make sure energy remains available for milk production instead of storage), these effects can balance out (Bell and Bauman 1997).

In summary, this study shows that urinary C-peptide can be used to successfully monitor the energy balance of white-faced capuchins, a New World primate. Here, I compared this biomarker to calculated energy balance and ketone production, and I documented the range of variation in energy balance experienced by females across an annual cycle. Variation in ripe

141

fruit energy density was the strongest predictor of variation in UCP, which indicated that females experienced lower energy balance during low fruit months. This suggests that UCP may be a valuable tool for assessing the energetic patterns that lead to the reproductive patterns observed among capuchins, including the timing of conception, births and weaning as well as the effect of seasonality and energetics on capuchin health and survival. With continued research, a larger sample size would allow for a more thorough investigation of inter-individual variation, which may help to identify differences due to social factors not identified in this study. UCP has also helped us to understand the role of energetics in the outcome of male-male competition in other primate species (Brown et al. In Review; Georgiev 2012; Girard-Buttoz et al. 2014; Higham et al. 2011b). White-faced capuchins live in multi-male, multi-female groups that experience a regular cycle of male takeovers (Fedigan and Jack 2012) – accordingly, UCP may help to quantify differences in male quality that influence the timing of takeover events and successful outcomes.

142

Chapter Six: Discussion

6.1 Summary and synthesis

This 12-month study is the first to use physiological biomarkers to investigate variation in the energy balance of female white-faced capuchin monkeys (Cebus capucinus). I employed a variety of data collection methods, including behavioral observation, phenological surveys of fruit abundance, nutritional analyses of capuchin foods, and laboratory analysis of urinary parameters, to measure physical condition and energy balance. My project objectives were to 1) document the dietary profile of females and the nutritional composition of foods consumed, and examine how seasonal variation in diet and nutritional intake affect the ability of females to meet estimated nutritional requirements; 2) measure seasonal variation in the nutritional importance of invertebrate foods; 3) determine how seasonal variation in the availability of foods affects the physical condition of females in terms of relative muscle mass; and 4) quantify energy balance and identify the most important ecological and social predictors of seasonal variation, employing urinary C-peptide as an analytical tool. The use of multiple ecological and behavioral measures revealed consistent patterns among female capuchins and is advantageous in better understanding the relationships among ecology, sociality and health. In this chapter, I summarize and integrate my results and suggest how these findings may be applied to future research on the nutritional ecology and energetics of capuchin monkeys and across primate taxa.

In Chapter 2, I presented data on variation in the dietary profiles, foraging behavior and nutritional intake of female white-faced capuchins in response to temporal variation in the abundance of food resources. Many studies have looked at primate foraging behavior as a measure of diet, but nutritional ecology is a relatively new area of research in primatology

143

(Chapman et al. 2003; Knott 1998; Rothman et al. 2008a; Rothman et al. 2012). My results supported data from previous studies that female white-faced capuchins focus foraging efforts on fruit and invertebrate food items (Chapman and Fedigan 1990; Fragaszy et al. 2004; McCabe

2005; Melin et al. 2010; Rose 1994). Fruit contributed the most to the overall energy gain despite females devoting a greater proportion their time to searching for and consuming invertebrates. On a dry matter basis, fruits provided the most important source of sugar, whereas high proportions of protein intake came from invertebrates, particularly when fruit availability was low. However, the larger size of fruit items compared to invertebrates on a dry matter basis likely contributed to females’ ability to consume macronutrients at higher rates while foraging on fruit.

Future research that expands the nutritional dataset to include: 1) vine fruits (which were not quantified in our phenological data collection); 2) fruits that are available but not consumed by capuchins; 3) invertebrate abundance; 4) mineral concentrations; and 5) the presence of secondary plant metabolites such as toxins, would help to better determine nutritional factors important to food selection, and how selection may vary in relation to overall nutrient and energy availability. Whether or not a resource is “important” in terms of energy, macronutrient and mineral intake, and how these factors affect ranging patterns and group defense, may not only relate to a food’s nutritional composition but also to the abundance of alternative resources. On a broad scale, during this study females consumed energy and macronutrients beyond estimated requirements when fruit abundance was high. As fruit abundance decreased, female capuchins showed behavioral flexibility during foraging by changing their diet: fruit made up a smaller proportion of energy intake, females spent more time foraging, and they consumed fewer carbohydrates and more protein. These patterns suggest that females are either attempting to

144

maximize the proportion of fruit in their diet in the face of declining abundance or are simply trying to maximize caloric intake in general when food abundance is low. Despite these changes, females did not meet estimated nutritional requirements during months with the lowest fruit abundance. Further analyses, including inter-annual assessment of seasonal variation in nutritional intake as well as assessment of the intake geometry of various nutritional components, may help to elucidate patterns of capuchin foraging and the role of specific capuchin foods (Raubenheimer 2011; Raubenheimer and Simpson 1993; Simpson and

Raubenheimer 1993), as has been successfully implemented in other primate studies (Felton et al. 2009b; Raubenheimer et al. 2014; Rothman et al. 2011).

In Chapter 3, I looked more closely at group-level differences in nutritional intake and how the nutritional role of invertebrates changed as fruit abundance fluctuated. Annual fruit production in Sector Santa Rosa is bimodal with distinct peaks in the dry season and late rainy season (Campos et al. 2014). Seasonal and intergroup variation in home range quality (in terms of the energy density from ripe fruits that constitute the capuchin diet) followed this bimodal pattern. As the pattern of carbohydrates and protein intake during low fruit months suggested

(Chapter 2), invertebrates constituted an increasingly larger proportion of total energy intake as fruit abundance decreased. While there is a negative correlation between ripe fruit abundance and invertebrate consumption, implying that invertebrates are a “fallback food” by Marshall and

Wrangham’s (2007) definition, it is important to recognize that there are many types of invertebrates that vary in size, distribution, nutritional composition and seasonal importance that may serve different functions as dietary resources for capuchins. Although many orders of invertebrates contributed to nutritional intake in terms of energy and protein consumption throughout the annual cycle, this study highlighted the important dietary role of lepidopteran

145

larvae (i.e., caterpillars), since their seasonal emergence coincided with a four-month period of low fruit abundance. Although lower in carbohydrates than fruit, caterpillars and fruit share many resource characteristics: due to host specificity they are clumped in distribution, they are easily gleaned from plant substrates by the capuchins, they require minimal processing, and they are of relatively high quality in terms of protein and fat content. Therefore, they may be considered a high-quality and preferred resource. However, it is difficult to measure the preference of lepidopterans compared to fruit as a capuchin food source since detailed measures of caterpillar abundance within each group’s home range are not available. The primary period of caterpillar emergence occurs only during a period of low fruit abundance (rather than during both low and high fruit abundance periods), and the reproductive demands of female capuchins change throughout the year. Temporal data on caterpillar abundance would be needed across multiple years to draw firm conclusions. During this study, despite the considerable contribution of lepidopterans to capuchin energy and protein intake, females in some periods did not appear to meet total monthly estimated energy requirements, after accounting for estimated reproductive demands. Thus, it is likely that females were utilizing metabolic energy stores during two low fruit months. The analysis of urinary parameters (Chapter 4) and the metabolic biomarker C- peptide (Chapter 5) provided a more detailed picture and support for the metabolic consequences associated with these seasonal changes in resource abundance.

In Chapter 4, I used the urinary parameters of creatinine and specific gravity to measure the relative muscle mass of females. Creatinine excretion is positively related to muscle mass, and specific gravity is a measure of urine density relative to water that is free from muscle mass effects (Burger 1919; Myers and Fine 1913; Palladin and Wallenburger 1915). Using statistical modeling, I determined that season significantly predicted the relationship between creatinine

146

and specific gravity, whereby relative muscle mass was greater during months with high fruit abundance versus low fruit abundance. The number of urine samples I was able to collect constrained my ability to assess changes in relative muscle mass per individual, and on a smaller time scale than “season”. Although the effect was smaller, group identity also significantly predicted relative muscle mass, such that females in CP group, whose home range quality in terms of fruit availability was nearly half that of the other two groups during this study, had significantly lower relative muscle mass than did females in GN and LV groups. These data on inferred muscle mass supported the findings regarding variation in home range quality and large- scale seasonal variation in energy intake (Chapter 3), and indicated that females experienced physiological consequences of this variation in terms of physical condition.

The approach of measuring nutritional intake and physical condition used in this study may also shed light on research on intergroup dynamics and home range quality. Intergroup encounters, whereby groups participate in direct competition over resources, are common in this species (Campos et al. 2014; Childers 2008; Crowfoot 2013; Crowfoot et al. 2008). Previous research indicates that white-faced capuchins may show short-term energetic costs resulting from the consequences of intergroup competition, in terms of increased travel time and distance and a decrease in the quality of resources obtained (Crowfoot 2013). Further investigation into the nutritional value of contested resources and seasonal variation in the rate of occurrence of intergroup encounters may better inform us about how these short-term physiological consequences translate into the formation of home ranges and long-term fitness consequences related to intergroup dynamics and ranging behavior. It may also help us to better estimate ideal group size in terms of balancing the number of group members needed to create an advantage in intergroup competition while minimizing intragroup competition for resources.

147

In Chapter 5, I quantified the physical condition of females in greater detail using monthly variation in energy balance. I confirmed that energy balance, as calculated based on energy intake minus energy expenditure, was positively correlated to the energetic biomarker urinary C- peptide. Females showed broad indications of negative energy balance through the production of urinary ketones, a signal of fat catabolism, during the low fruit and energy balance months of

June and July, and there was a trend for C-peptide to coincide with this measure. Ripe fruit energy density (kJ/ha) was the single most important predictor of the energy balance of females.

Contrary to my prediction, my results indicated that rank was not an important predictor of variation in energy balance. It is possible that I did not collect urine samples on a fine enough scale to detect this. However, assuming this information is correct, as a next step, I plan to investigate rank-related differences in feeding behavior (including food type, time spent foraging and group location) to determine whether individuals adopt behavioral strategies to compensate for the potential nutritional deficits that may result from resource competition.

I also predicted that C-peptide would reflect the energetic demands of reproduction, but analysis did not reveal significant variation in energy balance across reproductive states. During data exploration, it appeared that although there was considerable inter-individual variation, pre- conceptive females showed the highest energy balance and females in the early stages of lactation experienced the lowest energy balance. Perhaps a larger dataset would increase the statistical power of this analysis, or a more detailed approach to analysis may reveal a pattern, although changes in insulin resistance across reproductive states may mask differences in energy balance. As a next step, I hope to examine the relationship between energy balance and reproduction in greater detail. If the patterns observed during this study are relatively consistent across years, the addition of energetics data to the analysis of long-term life-history patterns may

148

allow for a more nuanced analysis of our previous interpretation of the causes and consequences of reproductive seasonality. Using a 25-year dataset, Carnegie et al. (2011a) found that 44% of infant births in this capuchin population occurred from May through July (mean = May), which classifies them as moderately seasonal breeders (van Schaik et al. 1999). The timing of births corresponds to one of the bimodal periods of low fruit abundance, during which females experienced negative energy balance in this study. The inferred drop in muscle mass at this time

(Chapter 4) and production of urinary ketones (Chapter 5) suggest that females are losing weight and utilizing fat reserves (Knott et al. 2009). In light of more recent data on energetics, further investigation into breeding strategies that focus on female energy balance at the time of conception, birth, lactation and weaning may shed considerably more insight into patterns of reproductive success. Using this approach, McCabe and Emery Thompson (2013b) were able to conclude that those females who timed conception to periods of high fruit abundance experienced a considerably higher rate of infant survival. A mismatch in the timing of reproductive events in the highly seasonal habitat in Sector Santa Rosa may show similar consequences in terms of the reproductive success of females.

6.2 Broader application and future directions

6.2.1 Within-species inter-annual comparison and the stability of ecological and social environments

While this study provided a thorough assessment of the seasonal changes in the diet and physical condition of capuchins over one annual cycle, it by no means represents the complete range of variation possible for this species. Inter-annual assessment and comparison is important, as one annual cycle could not possibly reflect the behavioral and physiological responses to

149

environments that undergo inter-annual variation in climate, which in turn affect ecological (e.g., tree production) and biological cycles (invertebrate abundance). These factors may differentially affect the nutritional intake and energy balance of females from year to year.

Instability in demographics and the social environment may also affect nutrition and energy balance. For example, since this study was conducted, the females in CP group fissioned along matrilineal lines into two separate social groups, each of which maintains different home ranges than the formerly united group’s home range (unpublished). The physiological consequences of this split have yet to be determined, but the changes in home ranges likely resulted in changes to each group’s access to resources.

Male takeover events, female disappearances and hierarchical instability affect the generalized stress response (measured via cortisol) in other primate species, much of which is focused on but not limited to baboons (Anestis 2010; Beehner et al. 2005; Crockford et al. 2008;

Engh et al. 2006; Sapolsky 1983; Sapolsky 2005). Physiology of the stress response is intertwined with metabolic processes and health (Sapolsky 1992). Another direction I would like to take with future research is to investigate variation in the stress response of females in this long-term study population using cortisol as a stress biomarker. While studies of stress using cortisol have aided in identifying categories of stressors across a diversity of primate species, a major limitation of its use is its generalized response. Nutritional (e.g., food quality) and psychosocial factors (e.g., aggression and social support) (Sapolsky 2005) both impact cortisol production and are conflated when analyzing fluctuation in its levels. Simultaneous measurement of cortisol with a biomarker for energy balance (i.e., C-peptide) allows for independent assessment of the energetic stress component and thus, differentiation from social stressors. This approach will allow me to focus on seasonal and inter-individual variation in

150

stress levels while disentangling the effects of metabolic versus psychosocial stress. Examining the differential stress responses of females will allow me to look more closely at possible behavioral strategies employed by females at varying reproductive states and social ranks in response to proximate ecological and social pressures. Nutritional and energetic studies will also help to better monitor population health in response to ecological changes and anthropogenic influences as well as inform captive management practices by documenting the range of natural variation experienced by individuals in this species.

6.2.2 Inter-population and inter-species comparison

In the context of the nutritional ecology of capuchin monkeys, this study adds to the documentation of behavioral responses to variation in fruit availability and the nutritional properties of food. However, these data must be put into a larger multi-year framework that is compared across field sites and species in order to determine the consistency of these responses, how they are linked to the fitness of individuals, and then how they were shaped by evolutionary pressures. For example, within this capuchin population, Melin and colleagues (2014b) recently suggested that extractive faunivory during low fruit periods may have driven the evolution of sensorimotor intelligence in Cebus and Sapajus and suggest that a quantitative cross-species analysis of manual dexterity would provide better support for their ideas related to diet and functional morphology.

C-peptide is an effective tool for making energetic comparisons across populations. For example, Emery Thompson et al. (2009b) discovered that chimpanzees in the Kanyawara population at the Kibale Forest National Park, Uganda, had significantly lower C-peptide than

Ngogo chimpanzees who inhabited a richer neighboring habitat (higher fruit tree density and lower chimpanzee density), suggesting that habitat quality and food availability impact energy

151

balance. The existence of multiple long-term capuchin research sites creates the potential to make cross-site comparisons. However, there are methodological issues related to the standardization of C-peptide that must first be resolved before making these comparisons.

Raw C-peptide assay values must be corrected for the dilution of the urine sample using a measure of urine density, since more concentrated urine will logically contain higher concentrations of analytes. This is most commonly accomplished for C-peptide and other urinary hormones by dividing the raw value by urinary creatinine concentration (Emery

Thompson and Knott 2008; Emery Thompson et al. 2010; Georgiev 2012; Girard-Buttoz et al.

2011; Grueter et al. 2014; Harris et al. 2010; Lodge 2012; Sherry and Ellison 2007). However, in my study, variation in creatinine-corrected C-peptide did not correspond to variation in the physical condition of females as indicated using intake and expenditure measures. This may be because creatinine excretion fluctuates with muscle mass, and this population of capuchins shows substantial variation in energy intake and physical condition as indicated by relative muscle mass estimation and urinary ketone production. Consequently, standardizing C-peptide concentration with creatinine did not seem to accurately reflect energy balance. In fact, when C- peptide was standardized using creatinine, it suggested a pattern of energy balance that completely contradicted what was predicted based on energy intake, energy expenditure, relative muscle mass and ketone production. The specific gravity of urine is not sensitive to changes in muscle mass and has been used as an alternative for normalizing urinary hormone concentrations

(Miller et al. 2004). When I used this method of standardizing, variation in C-peptide followed a similar annual pattern of variation to that of calculated energy balance. The key drawback to using specific gravity to standardize C-peptide is that standardization uses the population mean,

152

and thus the standardization coefficient differs across populations leading to difficulties in making comparisons.

6.2.3 Urine as a biological medium

My study provides further support that the use of urine as a biological medium can serve as an invaluable non-invasive tool for studying physiological variation in wild animal populations, and provides evidence that it is difficult, yet feasible, to collect urine from arboreal New World monkeys. Prior to use, researchers must thoroughly evaluate the habitat and other limitations of field conditions of potential study species (Strier and Ziegler 2005). Collection may not be feasible from species that move very rapidly or that are too high in the canopy or too low to the ground, due to dispersion or obstruction of urine streams originating from excessive heights, or absorption and contamination issues if animals are too low and samples hit the ground (Girard-

Buttoz et al. 2011; Higham et al. 2011a). Urine was not collected as frequently as I had hoped due to these factors and my own energetic, time and resource constraints. However, calculating monthly averages per sampling period should have provided an accurate estimate of temporal fluctuations in C-peptide levels, since mean concentrations likely muted the small-scale fluctuations resulting from proximate changes in dietary intake. The range of capuchin C- peptide values fell on the lower end of the standard curve and I was unable to obtain a result from a subset of samples. This may have been due to the assay affinity of capuchin C-peptide or the concentration of the urine samples. Based on the range of variation in specific gravity measurements, urine from capuchins at this field site seems to be, on average, more dilute than that of other species [e.g., chimpanzees (Pan troglodytes), Emery Thompson et al. 2012; redtail monkeys (Cercopithecus ascanius) and blue monkeys (C. mitis), M. Brown unpublished). Since values fell so low on the standard curve and I did not further dilute the urine samples before

153

running the assays, as was done in studies of other primate species [e.g., rhesus macaques

(Macaca mulatta), Girard-Buttoz et al. 2011], urine dilution may be a limiting factor for the quantification of “low” capuchin C-peptide levels. Researchers should be aware of urine dilution and associated analyte concentration as a potential species-specific limitation, and also assure that they are able to obtain sufficient sample volumes for analysis (which may be limited by the environment or size of the animal) before deciding to collect urine for research use.

6.3 Conclusion

Incorporating the collection of nutritional data into more studies of wild non-human primates will help us to better understand their nutritional requirements in natural environments. It will allow researchers to address how these requirements and associated behaviors change in the face of reproductive demands and with respect to social dynamics. The use of urinary parameters in this study and others has advanced the knowledge gathered via observation and nutritional analysis to a new level of understanding and brought to light the importance of quantifying physiological status to understand the range of variation in energy balance experienced by individuals. There are many confounding variables that influence health in studies conducted within natural settings. However, implementing the use of non-invasive field methodology, collecting and analyzing multiple biomarkers, and utilizing statistical analyses that account for multiple variables, will help link proximate processes with ultimate fitness consequences. Multi- year and cross-species comparison of intra- and inter-individual variation will also help to identify the most important behavioral and social predictors of variation in nutritional intake and energy balance, and determine the limits at which negative energy balance begins to negatively affect reproductive output and long-term reproductive fitness.

154

References

Adebisi SA, Adekunle BA, and Etu AK. 2001. Creatinine clearance: alternative approach to traditional 24-hour urine collection in normal individuals. African Journal of Medicine and Medical Sciences 30(1-2):27-30.

Albon SD, Mitchell B, Huby BJ, and Brown D. 1986. Fertility in female red deer (Cervus elaphus): the effects of body composition, age and reproductive status. Journal of Zoology 209(3):447-460.

Altmann J. 1974. Observational study of behavior: sampling methods. Behaviour 49:227-265.

Altmann S. 1991. Diets of yearling female primates (Papio cynocephalus) predict lifetime fitness. Proceedings of the National Academy of Sciences of the United States of America 88(2):420-423.

Altmann S. 1998. Foraging for survival: yearling baboons in Africa. Chicago: University of Chicago Press.

Altmann SA. 1987. The impact of locomotor energetics on mammalian foraging. The Journal of Zoology 211(2):215-225.

Anestis SF. 2010. Hormones and social behavior in primates. Evolutionary Anthropology 19:66- 78.

ANKOM Technology. 2011. Method 6, 4/13/2011: Neutral detergent fiber in feeds - filter bag technique (for A200 and A200I). Macedon, NY: Ankom Technology.

Antonow-Schlorke I, Schwab M, Cox LA, Li C, Stuchlik K, Witte OW, Nathanielsz PW, and McDonald TJ. 2011. Vulnerability of the fetal primate brain to moderate reduction in maternal global nutrient availability. Proceedings of the National Academy of Sciences of the United States of America 108(7):3011-3016.

AOAC International. 2012a. Crude fat in feeds, cereal grains, and forages. In: Latimer G, editor. Official Methods of Analysis of AOAC International. 19 ed. Gaithersburg, Maryland, USA: AOAC International, Official Method 2003.05.

AOAC International. 2012b. Crude protein in cereal grains and oilseeds: generic combustion method. In: Latimer G, editor. Official Methods of Analysis of AOAC International. 19 ed. Gaithersburg, Maryland, USA: AOAC International, Official Method 992.23.

155

AOAC International. 2012c. Crude protein in meat and meat products including pet foods. In: Latimer G, editor. Official Methods of Analysis of AOAC International. 19 ed. Gaithersburg, Maryland, USA: AOAC International, Official Method 992.15.

AOAC International. 2012d. Loss on drying (moisture) for feeds (at 135 for 2 hours), dry matter on oven drying for feeds (at 135 for 2 hours). In: Latimer G, editor. Official Methods of Analysis of AOAC International. 19 ed. Gaithersburg, MD, USA: AOAC International, Official Method 930.15.

AOAC International. 2012e. Protein (crude) in animal feed: combustion method. In: Latimer G, editor. Official Methods of Analysis of AOAC International. 19 ed. Gaithersburg, Maryland, USA: AOAC International, Official Method 990.03.

Armstrong E. 1983. Relative brain size and metabolism in mammals. Science 220(4603):1302- 1304.

Ausman LM, Gallina DL, Hayes KC, and Hegsted DM. 1986. Comparative assessment of soy and milk protein quality in infant cebus monkeys. The American Journal of Clinical Nutrition 43(1):112-127.

Ausman LM, and Hegsted DM. 1980. Protein requirements of adult cebus monkeys (Cebus albifrons). The American Journal of Clinical Nutrition 33(12):2551-2558.

Bartoń K. 2014. MuMIn: Multi-Model Inference. p R package version 1.12.11. http://CRAN.R- project.org/package=MuMIn.

Batchelet E. 1981. Circular statistics in biology. London: Academic Press.

Bateman AJ. 1948. Intra-sexual selection in Drosophila. Heredity 2:349-368.

Bates D, Maechler M, Bolker B, and Walker S. 2014. lme4: Linear mixed-effects models using Eigen and S4.

Beehner JC, Bergman TJ, Cheney DL, Seyfarth RM, and Whitten PL. 2005. The effect of new alpha males on female stress in free-ranging baboons. Animal Behaviour 69:1211-1221.

Bell AW, and Bauman DE. 1997. Adaptations of glucose metabolism during pregnancy and lactation. Journal of Mammary Gland Biology and Neoplasia 2(3):265-278.

Bell RHV. 1971. A grazing ecosystem in the Serengeti. Scientific American 225:86-93.

Benhamou S. 2011. Dynamic approach to space and habitat use based on biased random bridges. PLoS One 6(1):e14592.

156

Bentley GR. 1999. Aping our ancestors: comparative aspects of reproductive ecology. Evolutionary Anthropology 7(5):175-185.

Bercovitch FB, and Strum SC. 1993. Dominance rank, resource availability and reproductive maturation in female savanna baboons. Behavioral Ecology and Sociobiology 33:313- 318.

Bergstrom ML. 2009. Dominance among female white-faced capuchins (Cebus capucinus) at Santa Rosa National Park, Costa Rica. Calgary: University of Calgary. 127 p.

Bergstrom ML, and Fedigan LM. 2010. Dominance among female white-faced capuchin monkeys (Cebus capucinus): hierarchical linearity, nepotism, strength and stability. Behaviour 147(7):899-931.

Bergstrom ML, and Fedigan LM. 2013. Dominance style of female white-faced capuchins. American Journal of Physical Anthropology 150(4):591-601.

Berman CM, and Schwartz SM. 1988. A nonintrusive method for determining relative body fat in free-ranging monkeys. American Journal of Primatology 14(1):53-64.

Blanco Segura R. 2014. Gráficos de precipitación para Santa Rosa 1979 a 2013. Área de Conservación Guanacaste. www.acguanacaste.ac.cr.

Bonsnes RW, and Taussky HH. 1945. On the colorimetric determination of creatinine by the Jaffe reaction. The Journal of Biological Chemistry 158:581-591.

Brockman DK, and van Schaik CP. 2005. Seasonality and reproductive function. In: Brockman DK, and van Schaik CP, editors. Seasonality in primates: studies of living and extinct human and non-human primates. Cambridge: Cambridge University Press. p 269-306.

Bronson FH. 1985. Mammalian reproduction: an ecological perspective. Biology of Reproduction 32(1):1-26.

Brown M, Jones R, Angedakin S, and Emery Thompson M. In Review. Reduced male body condition and elevated female energy balance correspond with group takeovers in redtails (Cercopithecus ascanius) and blue monkeys (C. mitis). American Journal of Primatology.

Burger MZ. 1919. Die Bedeutung des Kreatininkoefizienten fur die quantitative bewertund der muskulatur als korpergewichtskomponente. Zeitschrift Fur Die Gesamte Experimentelle Medizin Einschliesslich Experimenteller Chirurgie 9:361-399.

Butte NF, and Hopkinson JM. 1998. Body composition changes during lactation are highly variable among women. The Journal of Nutrition 128(2):381S-385S.

157

Campos FA, Bergstrom ML, Childers A, Hogan JD, Jack KM, Melin AD, Mosdossy KN, Myers MS, Parr NA, Sargeant E, Schoof VAM, and Fedigan LM. 2014. Drivers of home range characteristics across spatiotemporal scales in a Neotropical primate, Cebus capucinus. Animal Behaviour 91:93-109.

Campos FA, Bergstrom ML, and Fedigan L. In Prep. Energy returns on foraging in white-faced capuchins, Cebus capucinus. Prepared for Submission to Behavioral Ecology and Sociobiology.

Campos FA, and Fedigan LM. 2013. Urine-washing in white-faced capuchins: a new look at an old puzzle. Behaviour 150(7):763-798.

Carnegie SD, Fedigan LM, and Melin AD. 2011a. Reproductive seasonality in female capuchins (Cebus capucinus) in Santa Rosa (Area de Conservación Guanacaste), Costa Rica. International Journal of Primatology 32(5):1076-1090.

Carnegie SD, Fedigan LM, and Ziegler TE. 2005. Behavioral indicators of ovarian phase in white-faced capuchins (Cebus capucinus). American Journal of Primatology 67:51-68.

Carnegie SD, Fedigan LM, and Ziegler TE. 2011b. Social and environmental factors affecting fecal glucocorticoids in wild, female white-faced capuchins (Cebus capucinus). American Journal of Primatology 73(9):861-869.

Carrieri M, Trevisan A, and Bartolucci GB. 2000. Adjustment to concentration-dilution of spot urine samples: correlation between specific gravity and creatinine. International Archives of Occupational and Environmental Health 74(1):63-67.

Catalano PM, Tyzbir ED, Roman NM, Amini SB, and Sims EAH. 1991. Longitudinal changes in insulin release and insulin resistance in nonobese pregnant women. American Journal of Obstetrics and Gynecology 165(6, Part 1):1667-1672.

Chao S. 1987. The effect of lactation on ovulation and fertility. Clinics in perinatology 14(1):39- 50.

Chapman CA, Chapman LJ, Rode KD, Hauck EM, and McDowell LR. 2003. Variation in the nutritional value of primate foods: among trees, time periods, and areas. International Journal of Primatology 24(2):317-333.

Chapman CA, and Fedigan LM. 1990. Dietary differences between neighboring Cebus capucinus groups: local traditions, food availability or responses to food profitability? Folia Primatologica 54:177-186.

158

Chapman CA, Wasserman MD, Gillespie TR, Speirs M, Lawes MJ, Saj TL, and Ziegler TE. 2006. Do food availability, parasitism, and stress have synergistic effects on red colobus populations living in forest fragments? American Journal of Physical Anthropology 131(4):525-534.

Childers A. 2008. Spatial ecology of Costa Rican white-faced capuchins: socioecological and cognitive implications. New Orleans, LA: Tulane University.

Clutton-Brock TH, Albon SD, and Guinness FE. 1989. Fitness costs of gestation and lactation in wild mammals. Nature 337(6204):260-262.

Clutton-Brock TH, Guinness FE, and Albon SD. 1982. Red deer: behavior and ecology of two sexes. Edinburgh, U.K.: Edinburgh University Press.

Clutton-Brock TH, and Harvey PH. 1977. Species differences in feeding and ranging behavior in primates. In: Clutton-Brock TH, editor. Primate ecology. London: Academic Press. p 557-584.

Coelho Jr. AM. 1974. Socio-bioenergetics and sexual dimorphism in primates. Primates 15(2- 3):263-269.

Coelho Jr. AM, Bramblett CA, Quick LB, and Bramblett SS. 1976. Resource availability and population density in primates. A sociobioenergetic analysis of energy budgets of Guatemalan howler and spider monkeys. Primates 17(1):63-80.

Conklin-Brittain NL, Knott CD, and Wrangham RW. 2006. Energy intake by wild chimpanzees and orangutans: methodological considerations and a preliminary comparison. In: Hohmann G, Robbins MM, and Boesch C, editors. Feeding Ecology in Apes and Other Primates Ecological, Physical and Behavioral Aspects. Cambridge: Cambridge University Press. p 445-471.

Crockford C, Wittig RM, Whitten PL, Seyfarth RM, and Cheney DL. 2008. Social stressors and coping mechanisms in wild female baboons (Papio hamadryas ursinus). Hormones and Behavior 53:254-265.

Crowfoot MC. 2013. The cost of defeat: capuchin groups travel further, faster and later after losing conflicts with neighbors. American Journal of Physical Anthropology 152(1):79- 85.

Crowfoot MC, Gilby IC, Wikelski M, and Kays RW. 2008. Interaction location outweighs the competitive advantage of numerical superiority in Cebus capucinus intergroup contests. Proceedings of the National Academy of Sciences 105(2):577-581.

Cryer PE, and Sode J. 1970. Variation in urinary creatinine excretion and its relationship to measurement of urinary 17-hydroxy-corticosteroids. Clinical Chemistry 16:1012-1015.

159

de Oliveira SG, Lynch Alfaro JW, and Veiga LM. 2014. Activity budget, diet, and habitat use in the critically endangered Ka'apor capuchin monkey (Cebus kaapori) in Pará State, Brazil: A preliminary comparison to other capuchin monkeys. American Journal of Primatology.

Deschner T, Fuller BT, Oelze VM, Boesch C, Hublin J-J, Mundry R, Richards MP, Ortmann S, and Hohmann G. 2012. Identification of energy consumption and nutritional stress by isotopic and elemental analysis of urine in bonobos (Pan paniscus). Rapid Communications in Mass Spectrometry 26(1):69-77.

Deschner T, Kratzsch J, and Hohmann G. 2008. Urinary C-peptide as a method for monitoring body mass changes in captive bonobos (Pan paniscus)`. Hormones and Behavior 54:620- 626.

Di Bitetti MS, and Janson CH. 2001. When will the stork arrive? Patterns of birth seasonality in neotropical primates. American Journal of Primatology 50:109-130.

Drack S, Ortmann S, Bührmann N, Schmid J, Heldmaier R, Heldmaier G, and Ganzhorn J. 1999. Field metabolic rate and the cost of ranging of the red-tailed sportive lemur (Lepilemur ruficaudatus). In: Rakotosamimanana B, Rasamimanana H, Ganzhorn J, and Goodman S, editors. New Directions in Lemur Studies: Springer US. p 83-91.

Drent RH, and Daan S. 1980. The prudent parent: energetic adjustments in avian breeding. In: Klomp H, and Woldendrop JW, editors. The integrated study of bird populations. New York: North Holland. p 225-252.

Durnin JVGA, and Passmore R. 1967. Energy, work and leisure. London: Heinemann.

Dyer LA, Singer MS, Lill JT, Stireman JO, Gentry GL, Marquis RJ, Ricklefs RE, Greeney HF, Wagner DL, Morais HC, Diniz IR, Kursar TA, and Coley PD. 2007. Host specificity of Lepidoptera in tropical and temperate forests. Nature 448(7154):696-699.

Ellison PT, and Valeggia CR. 2003. C-peptide levels and the duration of lactational amenorrhea. Fertility and Sterility 80(5):1279-1280.

Emery Thompson M, Kahlenberg S, Muller MN, and Wrangham RW. 2009a. The dynamics of stress in female chimpanzees: interactions of social and ecological factors. American Journal of Physical Anthropology Suppl 48:124.

Emery Thompson M, and Knott CD. 2008. Urinary C-peptide of insulin as a non-invasive marker of energy balance in wild orangutans. Hormones and Behavior 53:526-535.

Emery Thompson M, Muller MN, Kahlenberg S, and Wrangham RW. 2010. Dynamics of social and energetic stress in wild female chimpanzees. Hormones and Behavior 58(3):440-449.

160

Emery Thompson M, Muller MN, and Wrangham RW. 2012. Variation in muscle mass in wild chimpanzees: application of a modified urinary creatinine method. American Journal of Physical Anthropology 149(4):622-627.

Emery Thompson M, Muller MN, Wrangham RW, Lwanga JS, and Potts KB. 2009b. Urinary C- peptide tracks seasonal and individual variation in energy balance in wild chimpanzees. Hormones and Behavior 55(2):299-305.

Engh AL, Beehner JC, Berman TJ, Whitten PL, Hoffmeier RR, Seyfarth RM, and Cheney DL. 2006. Female hierarchy instability, male immigration and infanticide increase glucocorticoid levels in female chacma baboons. Animal Behavior 71(1227-1237).

ESRI. 2011. ArcGIS Desktop. 10 ed. Redlands, CA: Environmental Systems Research Institute.

Eveleth PB, and Tanner JM. 1990. Worldwide varation in human growth. Cambridge: Cambridge University Press.

Fedigan LM. 1993. Sex differences and intersexual relations in adult white-faced capuchins (Cebus capucinus). International Journal of Primatology 14(6):853-877.

Fedigan LM, Carnegie SD, and Jack KM. 2008. Predictors of reproductive success in female white-faced capuchins (Cebus capucinus). American Journal of Physical Anthropology 137:82-90.

Fedigan LM, and Jack KM. 2012. Tracking monkeys in Santa Rosa: lessons from a regenerating tropical dry forest. In: Kappeler P, and Watts D, editors. Long-term field studies of primates: Springer Press. p 165-184.

Fedigan LM, and Rose LM. 1995. Interbirth interval variation in three sympatric species of neotropical monkey. American Journal of Primatology 37(1):9-24.

Felton AM, Felton A, Lindenmayer DB, and Foley WJ. 2009a. Nutritional goals of wild primates. Functional Ecology 23(1):70-78.

Felton AM, Felton A, Raubenheimer D, Simpson SJ, Foley WJ, Wood JT, Wallis IR, and Lindenmayer DB. 2009b. Protein content of diets dictates the daily energy intake of a free-ranging primate. Behavioral Ecology 20(4):685-690.

Finke MD. 2007. Estimate of chitin in raw whole insects. Zoo Biology 26(2):105-115.

Fleagle JG. 1999. Primate adaptation and evolution. New York: Academic Press.

Floyd Jr. JC, Fajans SF, Conn JW, Knopf RF, and Rull J. 1966a. Insulin secretion in response to protein ingestion. Journal of Clinical Investigation 45(9):1479-1486.

161

Floyd Jr. JC, Fajans SF, Conn JW, Knopf RF, and Rull J. 1966b. Stimulation of insulin secretion by amino acids. Journal of Clinical Investigation 45(9):1487-1502.

Foerster S, and Monfort SL. 2010. Fecal glucocorticoids as indicators of metabolic stress in female Sykes' monkeys (Cercopithecus mitis albogularis). Hormones and Behavior 58(4):685-697.

Forbes GB, and Bruining GJ. 1976. Urinary creatinine excretion and lean body mass. The American Journal of Clinical Nutrition 29(12):1359-1366.

Ford SM, and Davis LC. 1992. Systematics and body size: Implications for feeding adaptations in New World monkeys. American Journal of Physical Anthropology 88(4):415-468.

Fragaszy D, Visalberghi E, and Fedigan LM. 2004. The complete capuchin: the biology of the genus Cebus. Cambridge: Cambridge University Press.

Fragaszy DM. 1990. Sex and age differences in the organization of behavior in wedge-capped capuchins, Cebus olivaceus. Behavioral Ecology 1(1):81-94.

Fragaszy DM, and Boinski S. 1995. Patterns of individual diet choice and efficiency of foraging in wedge-capped capuchin monkeys (Cebus olivaceus). Journal of Comparative Psychology 109(4):339-348.

Frazäo ER. 1991. Insectivory in free-ranging bearded saki (Chiropotes satanas chiropotes). Primates 32(2):243-245.

Freeland WJ, and Janzen DH. 1974. Strategies in herbivory by mammals - role of plant secondary compounds. American Naturalist 108:269-289.

Ganzhorn JU, Arrigo-Nelson S, Boinski S, Bollen A, Carrai V, Derby A, Donati G, Koenig A, Kowalewski M, Lahann P, Norscia I, Polowinsky SY, Schwitzer C, Stevenson PR, Talebi MG, Tan C, Vogel ER, and Wright PC. 2009. Possible fruit protein effects on primate communities in Madagascar and the Neotropics. PLoS ONE 4(12):e8253.

Gaulin SJC. 1979. A Jarman/Bell model of primate feeding niches. Human Ecology 7(1):1-20.

George JW. 2001. The usefulness and limitations of hand-held refractometers in veterinary laboratory medicine: an historical and technical review. Veterinary Clinical Pathology 30(4):201-210.

Georgiev AV. 2012. Energetic costs of reproductive effort in male chimpanzees. Cambridge, MA: Harvard University. 210 p.

Gingerich PD. 1979. Phylogeny of middle Eocene Adapidae (Mammalia, Primates) in North America: Smilodectes and Notharctus. Journal of Paleontology 53(1):153-163.

162

Girard-Buttoz C, Heistermann M, Rahmi E, Marzec A, Agil M, Fauzan P, and Engelhardt A. 2014. Mate-guarding constrains feeding activity but not energetic status of wild male long-tailed macaques (Macaca fascicularis). Behavioral Ecology and Sociobiology 68(4):583-595.

Girard-Buttoz C, Higham JP, Heistermann M, Wedegartner S, Maestripieri D, and Engelhardt A. 2011. Urinary C-peptide measurement as a marker of nutritional status in macaques. PLOS One 6(3):e18042.

Goldizen AW, Terborgh J, Cornejo F, Porras DT, and Evans R. 1988. Seasonal food shortage, weight loss, and the timing of births in saddle-back tamarins (Saguinus fuscicollis). Journal of Animal Ecology 57(3):893-901.

Greenblatt DJ, Rausil BJ, Harmatz JS, Smith TW, Duhme DW, and Koch-Weser J. 1976. Variability of 24-hour urinary creatinine excretion by normal subjects. The Journal of Clinical Pharmacology 16:321-328.

Grueter CC, Deschner T, Behringer V, Fawcett K, and Robbins MM. 2014. Socioecological correlates of energy balance using urinary C-peptide measurements in wild female mountain gorillas. Physiology and Behavior 127:13-19.

Haddow JE, Knight GJ, Palomaki GE, Neveux LM, and Chilmonczyk BA. 1994. Replacing creatinine measurements with specific gravity values to adjust urine cotinine concentrations. Clinical Chemistry 40(4):562-564.

Hall CL, and Fedigan LM. 1997. Spacial benefits afforded by high rank in white-faced capuchins. Animal Behaviour 53(5):1069-1082.

Hall MB, Hoover WH, Jennings JP, and Miller Webster TK. 1999. A method for partitioning neutral detergent-soluble carbohydrates. Journal of the Science of Food and Agriculture 79:2079-2086.

Harborne JB. 1984. Phytochemical Methods: A Guide to Modern Techniques of Plant Analysis. London: Chapman and Hall.

Harris TR, Chapman CA, and Monfort SL. 2010. Small folivorous primate groups exhibit behavioral and physiological effects of food scarcity. Behavioral Ecology 21(1):46-56.

Havel PJ. 2001. Peripheral signals conveying metabolic information to the brain: short-term and long-term regulation of food intake and energy homeostasis. Proceedings of the Society for Experimental Biology and Medicine 226(11):963-977.

Hemmingway CA, and Bynum N. 2005. The influence of seasonality on primate diet and ranging. In: Brockman DK, and van Schaik CP, editors. Seasonality in primates. Cambridge: Cambridge University Press. p 57-104.

163

Heymsfield SB, Arteaga C, McManus C, Smith J, and Moffitt S. 1983. Measurement of muscle mass in humans: validity of the 24-hour urinary creatinine method. American Journal of Clinical Nutrition 37:478-494.

Higham J, and Maestripieri D. 2014. The costs of reproductive success in male rhesus macaques (Macaca mulatta) on Cayo Santiago. International Journal of Primatology 35(3-4):661- 676.

Higham JP, Girard-Buttoz C, Engelhardt A, and Heistermann M. 2011a. Urinary C-peptide of insulin as a non-invasive marker of nutritional status: some practicalities. PLoS One 6(7):e22398.

Higham JP, Heistermann M, and Maestripieri D. 2011b. The energetics of male–male endurance rivalry in free-ranging rhesus macaques, Macaca mulatta. Animal Behaviour 81(5):1001- 1007.

Hinde K, and Milligan LA. 2011. Primate milk: Proximate mechanisms and ultimate perspectives. Evolutionary Anthropology 20:9-23.

Hladik CM. 1977. A comparative study of the feeding strategies of two sympatric species of leaf monkeys: Presbytis senex and Presbytis entellus. In: Clutton-Brock TH, editor. Primate ecology: studies of feeding and ranging behavior in lemurs, monkeys and apes. London: Academic Press. p 324-353.

Hladik CM, Hladik A, Bousset J, Valdebouze P, Viroben G, and Delort-Laval J. 1971. Le régime alimentaire des primates de l'île de Barro-Colorado (Panama). Folia Primatologica 16:85-122.

Hobson W. 1939. Urinary output of creatine and creatinine associated with physical exertion and its relationship to carbohydrate metabolism. Biochemical Journal 33(9):1425-1431.

Hogan JD. 2015. Florivory by Cebus capucinus: How variation in food abundance and colour vision affect foraging strategies. Calgary: University of Calgary.

Hoogwerf BJ, and Goetz C. 1983. Urinary C-peptide: a simple measure of integrated insulin production with emphasis on the effects of body size, diet, and corticosteroids. The Journal of Clinical Endocrinology & Metabolism 56(1):60-67.

Houston AI, Stephens PA, Boyd IL, Harding KC, and McNamara JM. 2007. Capital or income breeding? A theoretical model of female reproductive strategies. Behavioral Ecology 18(1):241-250.

Hunter A. 1928. Creatine and creatinine, monographs on biochemistry. New York: Longmans, Green and Co., Ltd.

164

IBM Corp. Released 2012. IBM SPSS Statistics for Windows. 21.0 ed. Armonk, NY: IBM Corp.

Isbell LA. 1991. Contest and scramble competition: patterns of aggression and ranging behavior among primates. Behavioral Ecology 2:143-155.

Izar P, Verderane MP, Peternelli-dos-Santos L, Mendonça-Furtado O, Presotto A, Tokuda M, Visalberghi E, and Fragaszy D. 2012. Flexible and conservative features of social systems in tufted capuchin monkeys: comparing the socioecology of Sapajus libidinosus and Sapajus nigritus. American Journal of Primatology 74(4):315-331.

Janson CH. 1985. Aggressive competition and individual food consumption in wild brown capuchin monkeys (Cebus apella). Behavioral Ecology and Sociobiology 18:125-138.

Janson CH. 1988. Food competition in brown capuchin monkeys (Cebus apella): Quantitative effects of group size and tree productivity. Behaviour 105(1):53-76.

Janson CH, and Boinski S. 1992. Morphological and behavioral adaptations for foraging in generalist primates: The case of the cebines. American Journal of Physical Anthropology 88(4):483-498.

Janson CH, and Verdolin J. 2005. Seasonality of primate births in relation to climate. In: Brockman DK, and van Schaik CP, editors. Seasonality in primates: studies of living and extinct human and non-human primates. Cambridge: Cambridge University Press. p 307- 350.

Janson CH, and Vogel ER. 2006. Estimating the effects of hunger on primate social ecology. In: Hohmann G, Robbins AM, and Boesch C, editors. Feeding ecology of apes and other primates. Cambridge: Cambridge University Press. p 285-312.

Janzen DH. 1988. Ecological characterization of a Costa Rican dry forest caterpillar fauna. Biotropica 20(2):120-135.

Janzen DH. 1993. Caterpillar seasonality in a Costa Rican dry forest. In: Stamp NE, and Casey TM, editors. Caterpillars: ecological and evolutionary constraints on foraging New York, NY: Chapman & Hall. p 448-477.

Johnson RL, and Kapsalis E. 1995. Ageing, infecundity and reproductive senescence in free- ranging female rhesus macaques. Journal of Reproduction and Fertility 105:271-278.

Kasiske BL, and Keane WF. 2000. Laboratory assessment of renal disease: clearance, urinalysis, and renal biopsy. In: Brenner BM, editor. The Kidney. Philadelphia: WB Saunders Company. p 1129-1170.

Kay RF. 1975. The functional adaptations of primate molar teeth. American Journal of Physical Anthropology 43(2):195-215.

165

Kay RF. 1980. Platyrrhine origins: a reappraisal of the dental evidence. In: Ciochon RL, and Chiarelli AB, editors. Evolutionary biology of the New World monkeys and continental drift. New York: Plenum Press. p 159-188.

Kay RF. 1984. On the use of anatomical features to infer foraging behavior in extinct primates. In: Rodman PS, and Cant JGH, editors. Adaptations for foraging in nonhuman primates: contributions to an organismal biology of prosimians, monkeys, and apes. New York: Columbia University Press. p 21-53.

Kay RF, Madden RH, Van Schaik C, and Higdon D. 1997. Primate species richness is determined by plant productivity: Implications for conservation. Proceedings of the National Academy of Sciences 94(24):13023-13027.

Key C, and Ross C. 1999. Sex differences in energy expenditure in non-human primates. Proceedings of the Royal Society London, Series B 266(1437):2479–2485.

Kleiber M. 1961. The fire of life: an introduction to animal energetics. Huntington, NY: Krieger.

Knott CD. 1998. Changes in orangutan caloric intake, energy balance, and ketones in response to fluctuating fruit availability. International Journal of Primatology 19(6):1061-1079.

Knott CD, Emery Thompson M, and Wich SA. 2009. The ecology of female reproduction in wild orangutans. In: Wich SA, Utami Atmoko SS, Setia TM, and van Schaik CP, editors. Orangutans: geographic variation in behavioral ecology and conservation. New York: Oxford University Press. p 171-188.

Koenig A, and Borries C. 2006. The predictive power of socioecological models: a reconsideration of resource characteristics, agonism, and dominance hierarchies. In: Hohmann G, Robbins MM, and Boesch C, editors. Feeding Ecology in Apes and Other Primates. Cambridge: Cambridge University Press. p 261-282.

Koenig A, Borries C, Chalise MK, and Winkler P. 1997. Ecology, nutrition and timing of reproductive events in an Asian primate, the Hanuman langur (Presbytis entellus). Journal of Zoology 243:215-235.

Kovach WL. 2011. Oriana - Circular Statistics for Windows. 4 ed. Pentraeth, Wales, U.K.: Kovach Computing Services.

Kramer FM, Liebschutz J, Marshall KA, McKinney S, and Stunkard AJ. 1993. Breast-feeding reduces maternal lower-body fat. Journal of the American Dietetic Association 93(4):429+.

Krezowski PA, Nuttall FQ, Gannon MC, and Bartosh MI. 1986. The effect of protein ingestion on the metabolic response to oral glucose in normal individuals. American Journal of Clinical Nutrition 44:847-856.

166

Kurita H, Suzumura T, Kanchi F, and Hamada Y. 2012. A photogrammetric method to evaluate nutritional status without capture in habituated free-ranging Japanese macaques (Macaca fuscata): a pilot study. Primates 53(1):7-11.

Laffel L. 1999. Ketone bodies: a review of physiology, pathophysiology and application of monitoring to diabetes. Diabetes/Metabolism Research and Reviews 15(6):412-426.

Lambert JE. 2011. Primate nutritional ecology: feeding biology and diet at ecological and evolutionary scales. In: Campbell B, Fuentes A, MacKinnon KC, Bearder SK, and Stumpf RM, editors. Primates in perspective. 2nd ed. Oxford: Oxford University Press. p 512-534.

Leco Corporation. Nitrogen/protein in animal feeds.

Leonard WR, and Robertson ML. 1997. Comparative primate energetics and hominid evolution. American Journal of Physical Anthropology 102(2):265-281.

Lewis RJ, and Kappeler PM. 2005a. Are Kirindy sifaka capital or income breeders? It depends. American Journal of Primatology 67(3):365-369.

Lewis RJ, and Kappeler PM. 2005b. Seasonality, body condition, and timing of reproduction in Propithecus verreauxi verreauxi in the Kirindy Forest. American Journal of Primatology 67(3):347-364.

Lodge E. 2012. Energetics and life-history of olive baboons (Papio hamadryas anubis) in the Gashaka Gumti National Park [Ph.D.]. London: University of Roehampton.

Lucas A. 1998. Programming by early nutrtion: an experimental approach. The Journal of Nutrition 128(2):401S-406S.

Lykken GI, Jacob RA, Munoz JM, and Sandstead HH. 1980. A mathematical model of creatine metabolism in normal males--comparison between theory and experiment. The American Journal of Clinical Nutrition 33(12):2674-2685.

Lynch Alfaro J, Boubli JP, Olson LE, Di Fiore A, Wilson B, Gutierrez-Espeleta GA, Chiou KL, Schulte M, Neitzel S, Ross V, Schwochow D, Nguyen MTT, Farias I, Janson CH, and Alfaro ME. 2012a. Explosive Pleistocene range expansion leads to widespread Amazonian sympatry between robust and gracile capuchin monkeys. Journal of Biogeography 39(2):272-288.

Lynch Alfaro J, Silva Jr. JS, and Rylands AB. 2012b. How different are robust and gracile capuchin monkeys? An argument for the use of Sapajus and Cebus. American Journal of Primatology 4:273-286.

167

Lynch Alfaro JW, Matthews L, Boyette AH, Macfarlan SJ, Phillips KA, Falótico T, Ottoni E, Verderane M, Izar P, Schulte M, Melin A, Fedigan L, Janson C, and Alfaro ME. 2012c. Anointing variation across wild capuchin populations: a review of material preferences, bout frequency and anointing sociality in Cebus and Sapajus. American Journal of Primatology 74(4):299-314.

MacKinnon KC. 2006. Food choice by juvenile capuchin monkeys (Cebus capucinus) in a tropical dry forest. In: Estrada A, Pavelka M, Garber P, and Luecke L, editors. New perspectives in the study of Mesoamerican primates: distribution, ecology, behavior, and conservation. New York: Springer Press. p 349-366.

Marshall A, and Wrangham R. 2007. Evolutionary consequences of fallback foods. International Journal of Primatology 28(6):1219-1235.

Marshall HH, Carter AJ, Coulson T, Rowcliffe JM, and Cowlishaw G. 2012. Exploring Foraging Decisions in a Social Primate Using Discrete-Choice Models. The American Naturalist 180(4):481-495.

Matthews L. 2009. Activity patterns, home range size, and intergroup encounters in Cebus albifrons support existing models of capuchin socioecology. International Journal of Primatology 30(5):709-728.

McCabe GM. 2005. Diet and nutrition in white-faced capuchins (Cebus capucinus): effects of group, sex and reproductive state. Calgary: University of Calgary. 119 p.

McCabe GM, and Emery Thompson M. 2013a. Ecology of reproduction in Sanje mangabeys (Cercocebus sanjei): dietary strategies and energetic condition during a high fruit period. American Journal of Primatology 75(12):1196-1208.

McCabe GM, and Emery Thompson M. 2013b. Reproductive seasonality in wild Sanje mangabeys (Cercocebus sanjei), Tanzania: Relationship between the capital breeding strategy and infant survival. Behaviour 150(12):1399-1429.

McCabe GM, and Fedigan LM. 2007. Effects of reproductive status on energy intake, ingestion rates, and dietary composition of female Cebus capucinus at Santa Rosa, Costa Rica. International Journal of Primatology 28:837-851.

McGrew WC. 2001. The other faunivory: primate insectivory and early human diet. In: Stanford CB, and Bunn HT, editors. Meat-eating and human evolution. Oxford: Oxford University Press. p 160-178.

Meistas MT, Rendell M, Margolis S, and Kowarski AA. 1982. Estimation of the secretion rate of insulin from the urinary excretion rate of C-peptide: study in obese and diabetic subjects. Diabetes 31:449-453.

168

Meistas MT, Zadik Z, Margolis S, and Kowarski AA. 1981. Correlation of urinary excretion of C-peptide with the integrated concentration and secretion rate of insulin. Diabetes 30:639-643.

Melin A, Fedigan L, Hiramatsu C, Hiwatashi T, Parr N, and Kawamura S. 2009. Fig foraging by dichromatic and trichromatic Cebus capucinus in a tropical dry forest. International Journal of Primatology 30(6):753-775.

Melin AD, Fedigan LM, Hiramatsu C, Sendall CL, and Kawamura S. 2007. Effects of colour vision phenotype on insect capture by a free-ranging population of white-faced capuchins, Cebus capucinus. Animal Behaviour 73(1):205-214.

Melin AD, Fedigan LM, Young HC, and Kawamura S. 2010. Can color vision variation explain sex differences in invertebrate foraging by capuchin monkeys? Current Zoology 56(3):300-312.

Melin AD, Hiramatsu C, Parr NA, Matsushita Y, Kawamura S, and Fedigan LM. 2014a. The behavioral ecology of color vision: considering fruit conspicuity, detection distance, and dietary importance. International Journal of Primatology 35(1):258-287.

Melin AD, Young HC, Mosdossy KN, and Fedigan LM. 2014b. Seasonality, extractive foraging and the evolution of primate sensorimotor intelligence. Journal of Human Evolution 71(0):77-86.

Melin Meachem AD. 2011. Polymorphic colour vision and foraging in white-faced capuchins: insights from field research and simulations of monkey vision [Ph.D.]. Calgary: University of Calgary. 189 p.

Merrill A, L., and Watt BK. 1955. Derivation of current calorie factors. In: USDA ARS, editor. Agricultural Handbook: Energy value of foods, basis and derivation. Washington D.C.: U.S. Government Printing Office. p 24-43.

Miller KE, Bales K, L., Ramos JH, and Dietz JM. 2006. Energy intake, energy expenditure, and reproductive costs of female wild golden lion tamarins (Leontopithecus rosalia). American Journal of Primatology 68(11):1037-1053.

Miller RC, Brindle E, Holman DJ, Shofer J, Klein NA, Soules MR, and O'Connor KA. 2004. Comparison of specific gravity and creatinine for normalizing urinary reproductive hormone concentrations. Clinical Chemistry 50:924-932.

Milligan LA. 2010. Milk composition of captive tufted capuchins (Cebus apella). American Journal of Primatology 72:81-86.

Milton K. 1981. Food choice and digestive strategies of two sympatric primate species. The American Naturalist 117(4):496-505.

169

Milton K. 1984. The role of food-processing factors in primate food choice. In: Rodman PS, and Cant JGH, editors. Adaptations for foraging in nonhuman primates: contributions to an organismal biology of prosimians, monkeys and apes. New York: Columbia University Press. p 249-279.

Milton K, and Dintzis F. 1981. Nitrogen-to-protein conversion factors for tropical plant samples. Biotropica 13:177-181.

Milton K, and Giacalone J. 2014. Differential effects of unusual climatic stress on capuchin (Cebus capucinus) and howler monkey (Alouatta palliata) populations on Barro Colorado Island, Panama. American Journal of Primatology 76(3):249-261.

Mosdossy K. 2013. Insect abundance in relation to capuchin (Cebus capucinus) foraging: fallback foods in a Costa Rican tropical dry forest. Calgary: University of Calgary.

Mosdossy KN, Melin AD, and Fedigan LM. In Review. Quantifying seasonal fallback on invertebrates, pith and bromeliad leaves by white-faced capuchin monkeys (Cebus capucinus) in a tropical dry forest. American Journal of Physical Anthropology.

Mott GE, Jackson EM, McMahan CA, and McGill Jr. HC. 1990. Cholesterol metabolism in adult baboons is influenced by infant diet. The Journal of Nutrition 120(3):243-251.

Mott GE, Lewis DS, and McGill Jr. HC. 1991. Programming of cholesterol metabolism by breast or formula feeding. Chichester, U.K.: Wiley.

Moura ACdA. 2004. The capuchin monkey and the Caatinga dry forest: a hard life in a harsh habitat [Ph.D.]. Ann Arbor: University of Cambridge (United Kingdom).

Myers VC, and Fine MS. 1913. The creatinine content of muscle under normal conditions. Its relation to the urinary creatinine. The Journal of Biological Chemistry 14:9-26.

Nagy KA, and Milton K. 1979. Energy metabolism and food consumption by wild howler monkeys (Alouatta palliata). Ecology 60(3):475-480.

Naithani VK, Steffens GJ, Tager HS, Buse G, Rubenstein AH, and Steiner DF. 1984. Isolation and amino-acid sequence determination of monkey insulin and proinsulin. Hoppe- Seyler´s Zeitschrift für physiologische Chemie 365(1):571-576.

National Research Council. 2003. Nutrient requirements of nonhuman primates. Washington, DC: National Academy Press.

Noldus Information Technology. 1998. Reference manual. MatMan 11.

170

Norconk MA, Wright BW, Conklin-Brittain NL, and Vinyard CJ. 2009. Mechanical and nutritional properties of food as factors in platyrrhine dietary adaptations. In: Garber PA, Estrada A, Bicca-Marques JC, Heymann EW, and Strier KB, editors. South American primates: comparative perspectives in the study of behavior, ecology, and conservation. New York: Springer Science+Business Media, LLC. p 279-319.

Nuttall FQ, Gannon MC, Wald JL, and Ahmed M. 1985. Plasma glucose and insulin profiles in normal subjects ingesting diets of varying carbohydrate, fat and protein content. Journal of the American College of Nutrition 4:437-450.

O'Malley R, and Fedigan L. 2005. Variability in food-processing behaviour among white-faced capuchins (Cebus capucinus) in Santa Rosa National Park, Costa Rica. American Journal of Physical Anthropology 128:63-73.

Oates JF. 1987. Food distribution and foraging behavior. In: Smuts BB, Cheney DL, Seyfarth RM, Wrangham RW, and Struhsaker TT, editors. Primate Societies. Chicago: University of Chicago Press. p 197-209.

Oftedal OT, Whiten A, Southgate DAT, and Van Soest P. 1991. The nutritional consequences of foraging in primates: the relationship of nutrient intakes to nutrient requirements. Philosophical Transactions of the Royal Society London, Series B 334(1270):161-170.

Palladin A, and Wallenburger L. 1915. Contribution a l'etude de la formation de la creatine dans l'organisme animal. Comptes Rendus des Séances et Mémoires de la Société de Biologie 78:111-113.

Pellett PL. 1990. Protein requirements in humans. The American Journal of Clinical Nutrition 51(5):723-737.

Perry S. 1996. Female-female social relationships in wild white-faced capuchin monkeys, Cebus capucinus. American Journal of Primatology 40(2):167-182.

Peters RH, Cloutier S, Dubé D, Evans A, Hastings P, Kaiser H, Kohn D, and Sarwer-Foner B. 1988. The allometry of the weight of fruit trees and shrubs in Barbados. Oecologia 74(4):612-616.

Peterson JD, Nehrlich S, Oyer PE, and Steiner DF. 1972. Determination of the amino acid sequence of the monkey, sheep, and dog proinsulin C-peptides by a semi-micro Edman degradation procedure. The Journal of Biological Chemistry 247(15):4866-4871.

Pontzer H, Raichlen DA, Gordon AD, Schroepfer-Walker KK, Hare B, O’Neill MC, Muldoon KM, Dunsworth HM, Wood BM, Isler K, Burkart J, Irwin M, Shumaker RW, Lonsdorf EV, and Ross SR. 2014. Primate energy expenditure and life history. Proceedings of the National Academy of Sciences 111(4):1433-1437.

171

Pontzer H, Raichlen DA, Shumaker RW, Ocobock C, and Wich SA. 2010. Metabolic adaptation for low energy throughput in orangutans. Proceedings of the National Academy of Sciences 107(32):14048-14052.

R Development Core Team. 2014. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing.

Raubenheimer D. 2011. Toward a quantitative nutritional ecology: the right-angled mixture triangle. Ecological Monographs 81(3):407-427.

Raubenheimer D, Rothman JM, Pontzer H, and Simpson SJ. 2014. Macronutrient contributions of insects to the diets of hunter–gatherers: A geometric analysis. Journal of Human Evolution 71(0):70-76.

Raubenheimer D, and Simpson SJ. 1993. The geometry of compensatory feeding in the locust. Animal Behaviour 45:953-964.

Richard AF. 1985. Primates in nature. New York: W.H. Freeman.

Richard AF, Dewar RE, Schwartz M, and Ratsirarson J. 2000. Mass change, environmental variability and female fertility in wild Propithecus verreauxi. Journal of Human Evolution 39(4):381-391.

Roberts SB. 1989. Use of the doubly labeled water method for measurement of energy expenditure, total body water, water intake and metabolizable energy intake in humans and small mammals. Canadian Journal of Physiology and Pharmacology 67(10):1190- 1198.

Rodgers AR, Kie JG, Wright D, Beyer HL, Smith L, and Carr AP. 2012. HRT: Home range tools for ArcGIS. Ontario Ministry of Natural Resources, Center for Northern Forest Ecosystem Research. Thunder Bay, Ontario, Canada.

Rose LM. 1994. Sex differences in diet and foraging behavior in white-faced capuchins (Cebus capucinus). International Journal of Primatology 15(1):95-114.

Rosetta L, Lee PC, and Garcia C. 2011. Energetics during reproduction: A doubly labeled water study of lactating baboons. American Journal of Physical Anthropology 144(4):661-668.

Rothman J, Dierenfeld E, Hintz H, and Pell A. 2008a. Nutritional quality of gorilla diets: consequences of age, sex, and season. Oecologia 155(1):111-122.

Rothman JM, Chapman CA, Struhsaker TT, Raubenheimer D, Twinomugisha D, and Waterman PG. 2014a. Long-term declines in nutritional quality of tropical leaves. Ecology 96(3):873-878.

172

Rothman JM, Chapman CA, Twinomugisha D, Wasserman MD, Lambert JE, and Goldberg TL. 2008b. Measuring physical traits of primates remotely: the use of parallel lasers. American Journal of Primatology 70:1191-1195.

Rothman JM, Chapman CA, and Van Soest P. 2012. Methods in primate nutritional ecology: a user's guide. International Journal of Primatology 33:542-566.

Rothman JM, Raubenheimer D, Bryer MAH, Takahashi M, and Gilbert CC. 2014b. Nutritional contributions of insects to primate diets: implications for primate evolution. Journal of Human Evolution 71:59-69.

Rothman JM, Raubenheimer D, and Chapman CA. 2011. Nutritional geometry: gorillas prioritize non-protein energy while consuming surplus protein. Biology Letters 7:847- 849.

Sailer L, Gaulin SC, Boster J, and Kurland J. 1985. Measuring the relationship between dietary quality and body size in primates. Primates 26(1):14-27.

Sapolsky RM. 1983. Endocrine aspects of social instability in the olive baboon (Papio anubis). American Journal of Primatology 5:365-379.

Sapolsky RM. 1992. Endocrinology of the stress-response. In: Becker JB, Breedlove SM, Crews D, and McCarthy MM, editors. Behavioral Endocrinology. 2nd ed. Cambridge: Massachusetts Institute of Technology Press. p 409-450.

Sapolsky RM. 2005. The influence of social hierarchy on primate health. Science 308:648-652.

Sargeant E. 2014. Infant handling and allonursing in white-faced capuchins [M.A.]. Calgary: University of Calgary.

Schmid J, and Speakman JR. 2000. Daily energy expenditure of the grey mouse lemur (Microcebus murinus): a small primate that uses torpor. Journal of Comparative Physiology B 170(8):633-641.

Schoener TW. 1971. Theory of feeding strategies. Annual Review of Ecology and Systematics 2:369-404.

Scrimshaw NS, Habicht J-P, Piche ML, Cholakos B, and Arroyave G. 1966. Protein metabolism of young men during university examinations. The American Journal of Clinical Nutrition 18(5):321-324.

Sherry DS, and Ellison PT. 2007. Potential applications of urinary C-peptide of insulin for comparative energetics research. American Journal of Physical Anthropology 133:771- 778.

173

Simmen B, Bayart F, Rasamimanana H, Zahariev A, Blanc S, and Pasquet P. 2010. Total energy expenditure and body composition in two free-living sympatric lemurs. PLoS ONE 5(3):e9860.

Simpson SJ, and Raubenheimer D. 1993. A multi-level analysis of feeding behaviour: the geometry of nutritional decisions. Philosophical Transactions: Biological Sciences 342(1302):381-402.

Smith OW. 1942. Creatinine excretion in women: data collected in the course of urinalysis for female sex hormones. The Journal of Clinical Endocrinology & Metabolism 2(1):1-12.

Smith RJ, and Jungers WL. 1997. Body mass in comparative primatology. Journal of Human Evolution 32(6):523-559.

Soskin S, and Levine R. 1944. Physiological and clinical aspects of ketosis. The American Journal of Digestive Diseases 11(10):305-313.

Speakman JR. 1998. The history and theory of the double labeled water technique. American Journal of Clinical Nutrition 68(4):932S-938S.

Srivastava SS, Mani KV, Soni CM, and Bhati J. 1957. Effect of muscular exercises on urinary excretion of creatine and creatinine. Indian Journal of Medical Research 55(9):953-960.

Steiner DF. 1984. The biosynthesis of insulin: genetic, evolutionary and pathophysiologic aspects. The Harvey Lectures Series 78. New York: Academic Press. p 191-228.

Stevens CE, and Hume ID. 1996. Comparative physiology of the vertebrate digestive system. Cambridge: Cambridge University Press.

Strier KB, and Ziegler TE. 2005. Advances in field-based studies of primate behavioral endocrinology. American Journal of Primatology 67(1):1-4.

Tardif SD, and Jaquish CE. 1997. Number of ovulations in the marmoset monkey (Callithrix jacchus): relation to body weight, age and repeatability. American Journal of Primatology 42(4):323-329.

Taussky HH. 1954. A microcolorimetric determination of creatinine in urine by the Jaffe reaction. The Journal of Biological Chemistry 208:853-862.

Taylor CR, Heglund NC, and Maloiy GM. 1982. Energetics and mechanics of terrestrial locomotion. I. Metabolic energy consumption as a function of speed and body size in birds and mammals. Journal of Experimental Biology 97(1):1-21.

Taylor CR, Schmidt-Nielsen K, and Raab JL. 1970. Scaling of energetic cost of running to body size in mammals. American Journal of Physiology 219(4):1104-1107.

174

Terborgh J. 1983. Five New World primates. Princeton, New Jersey: Princeton University Press.

Trivers RL. 1972. Parental investment and sexual selection. In: Campbell B, editor. Sexual Selection and the Descent of Man. Chicago: Aldine-Atherton. p 136-179.

Ullrey D, Allen M, Ausman L, Conklin-Brittain N, Edwards M, Erwin J, Holick M, Hopkins D, Lewis S, Lonnerdal B, and Rudel L. 2003. Carbohydrates and fiber. In: Nutrition NRCCoA, editor. Nutritional requirements of non-human primates. Second Revised Edition ed. Washington D.C.: The National Academies Press. p 58-74.

UNESCO World Heritage Committee. 1999. 23COM VIII.A.1 Área de Conservación Guanacaste (Costa Rica). Convention concerning the protection of the world cultural and natural heritage. Marrakesh, Morocco: United Nations Educational, Scientific and Cultural Organization.

USDA N. 2014. The PLANTS Database (http://plants.usda.gov, 7 July 2014). Greensboro, NC 27401-4901 USA: National Plant Data Team.

Uvarov BP. 1931. Insects and climate. Transactions of the Royal Entomological Society of London 79(1):1-232.

Valenta K, and Melin AD. 2012. Protein limitation explains variation in primate colour vision phenotypes: a unified model for the evolution of primate trichromatic vision. In: Garcia M-D, editor. Zoology: In Tech. van Schaik CP, Terborgh JW, and Wright SJ. 1993. The phenology of tropical rorests: adaptive significance and consequences for primary consumers. Annual Review of Ecology and Systematics 24(1):353-377. van Schaik CP, and van Noordwijk MA. 1989. The special role of male cebus monkeys in predation avoidance and its effect on group composition. Behavioral Ecology and Sociobiology 24(5):265-276. van Schaik CP, van Noordwijk MA, and Nunn CL. 1999. Sex and social evolution in primates. In: Lee PC, editor. Comparative primate socioecology. Cambridge: Cambridge University Press. p 204-240.

Van Soest PJ, Robertson JB, and Lewis BA. 1991. Methods for dietary fiber, neutral deterent fiber, and nonstarch polysaccharides in relation to animal nutrition. Journal of Dairy Science 74(10):3583-3597.

Vogel ER. 2004. The ecological basis of aggression in white-faced capuchin monkeys, Cebus capucinus, in a Costa Rican dry forest. Stony Brook, NY: Stony Brook University.

175

Vogel ER. 2005. Rank differences in energy intake rates in white-faced capuchin monkeys, Cebus capucinus: the effects of contest competition. Behavioral Ecology and Sociobiology 58:333-344.

Vogel ER, Munch CH, and Janson CH. 2007. Understanding escalated aggression over food resources in white-faced capuchin monkeys. Animal Behaviour 74:71-80.

Voland E. 1998. Evolutionary ecology of human reproduction. Annual Review of Anthropology 27:347-374.

Wade GN, and Schneider JE. 1992. Metabolic fuels and reproduction in female mammals. Neuroscience & Biobehavioral Reviews 16(2):235-272.

Weinhouse S. 1952. Factors involved in the formation and utilization of ketone bodies. Major Metabolic Fuels: Brookhaven Symposia in Biology. p 201.

Werk EE, McPherson HT, Hamrick LW, Myers JD, and Engel FL. 1955. Studies on ketone metabolism in man. I. A method for the quantitative estimation of splanchnic ketone production. Journal of Clinical Investigation 34(8):1256-1267.

Willis D, Mason H, Gilling-Smith C, and Franks S. 1996. Modulation by insulin of follicle- stimulating hormone and luteinizing hormone actions in human granulosa cells of normal and polycystic ovaries. Journal of Clinical Endocrinology and Metabolism 81(1):302- 309.

Wolden-Hanson T, Davis GA, Baum ST, and Kemnitz JW. 1993. Insulin levels, physical activity, and urinary catecholamine excretion of obese and non-obese rhesus monkeys. Obesity Research 1(5-17).

Wrangham RW. 1980. An ecological model of female-bonded primate groups. Behaviour 75:262-300.

Young HC. 2005. Sharp spines and toxic stings: how white-faced capuchins (Cebus capucinus) overcome the defence mechanisms of invertebrates in Costa Rica [M.A.]. Calgary: University of Calgary.

176

APPENDIX A: DATA COLLECTION SCHEDULE AND GROUP CONTACT

Rotation # Group Date Start of contact End of contact Total contact (hrs) 1 LV 24-Aug-09 05:16 17:30 12.23 1 LV 25-Aug-09 05:00 17:42 12.70 1 LV 26-Aug-09 05:30 17:30 12.00 1 LV 27-Aug-09 08:29 13:00 4.52 1 CP 31-Aug-09 05:30 17:30 12.00 1 CP 01-Sep-09 05:30 17:02 11.53 1 CP 02-Sep-09 06:00 12:30 6.50 1 CP 05-Sep-09 06:14 15:13 8.98 1 CP 06-Sep-09 05:30 16:22 10.87 1 CP 07-Sep-09 05:30 14:42 9.20 1 GN 10-Sep-09 05:29 17:31 12.03 1 GN 11-Sep-09 05:26 14:22 8.93 1 GN 12-Sep-09 09:59 17:53 7.90 1 GN 13-Sep-09 05:22 18:00 12.63 1 GN 14-Sep-09 05:29 18:00 12.52 1 GN 15-Sep-09 07:58 15:00 7.03 2 LV 26-Sep-09 05:24 15:30 10.10 2 LV 27-Sep-09 07:12 17:52 10.67 2 LV 28-Sep-09 05:34 17:42 12.13 2 LV 29-Sep-09 05:09 16:01 10.87 2 CP 03-Oct-09 05:28 17:45 12.28 2 CP 04-Oct-09 06:29 17:32 11.05 2 CP 05-Oct-09 05:30 15:01 9.52 2 CP 07-Oct-09 05:00 17:30 12.50 2 CP 08-Oct-09 05:20 15:33 10.22 2 CP 09-Oct-09 05:00 15:20 10.33 2 GN 12-Oct-09 07:24 17:43 10.32 2 GN 13-Oct-09 06:51 17:43 10.87 2 GN 14-Oct-09 05:00 17:41 12.68 2 GN 15-Oct-09 05:08 17:42 12.57 2 GN 17-Oct-09 05:00 16:18 11.30 2 GN 18-Oct-09 05:23 14:21 8.97 3 LV 24-Oct-09 05:13 17:45 12.53 3 LV 25-Oct-09 04:56 17:54 12.97 3 LV 26-Oct-09 05:04 17:40 12.60 3 LV 27-Oct-09 06:26 16:00 9.57 3 CP 31-Oct-09 06:43 17:38 10.92 3 CP 01-Nov-09 05:18 18:00 12.70 3 CP 02-Nov-09 05:06 14:24 9.30 3 CP 04-Nov-09 07:00 17:55 10.92 3 CP 05-Nov-09 05:04 17:30 12.43 3 CP 06-Nov-09 04:54 15:11 10.28 3 GN 09-Nov-09 12:38 17:30 4.87 3 GN 10-Nov-09 07:00 17:30 10.50 3 GN 11-Nov-09 05:00 17:02 12.03

177

Rotation # Group Date Start of contact End of contact Total contact (hrs) 3 GN 13-Nov-09 07:00 17:30 10.50 3 GN 14-Nov-09 06:14 15:36 9.37 3 GN 15-Nov-09 04:58 14:11 9.22 4 LV 21-Nov-09 09:22 17:49 8.45 4 LV 22-Nov-09 06:50 16:00 9.17 4 LV 23-Nov-09 06:45 17:42 10.95 4 LV 24-Nov-09 05:03 15:14 10.18 4 CP 28-Nov-09 04:56 17:30 12.57 4 CP 29-Nov-09 04:49 16:19 11.50 4 CP 01-Dec-09 04:52 17:30 12.63 4 CP 02-Dec-09 04:55 17:30 12.58 4 CP 03-Dec-09 04:57 17:02 12.08 4 GN 07-Dec-09 05:03 17:53 12.83 4 GN 08-Dec-09 05:00 17:42 12.70 4 GN 09-Dec-09 04:52 17:30 12.63 4 GN 10-Dec-09 05:04 17:30 12.43 4 GN 11-Dec-09 05:10 14:30 9.33 5 LV 12-May-10 06:23 18:09 11.77 5 LV 13-May-10 06:44 18:14 11.50 5 LV 14-May-10 15:56 18:13 2.28 5 LV 15-May-10 04:35 14:19 9.73 5 CP 18-May-10 04:53 18:05 13.20 5 CP 19-May-10 05:00 18:00 13.00 5 CP 20-May-10 05:00 17:30 12.50 5 CP 22-May-10 05:30 18:10 12.67 5 CP 23-May-10 05:30 18:00 12.50 5 CP 24-May-10 06:29 11:34 5.08 5 GN 27-May-10 04:57 18:00 13.05 5 GN 28-May-10 04:59 18:14 13.25 5 GN 29-May-10 05:28 18:09 12.68 5 GN 30-May-10 04:58 18:08 13.17 5 GN 31-May-10 04:59 16:37 11.63 6 LV 08-Jun-10 06:29 18:06 11.62 6 LV 09-Jun-10 04:42 18:00 13.30 6 LV 10-Jun-10 04:35 18:11 13.60 6 LV 11-Jun-10 04:57 14:30 9.55 6 CP 14-Jun-10 06:47 18:08 11.35 6 CP 15-Jun-10 04:58 18:04 13.10 6 CP 16-Jun-10 04:59 18:05 13.10 6 CP 17-Jun-10 04:50 18:05 13.25 6 CP 18-Jun-10 04:59 18:05 13.10 6 CP 19-Jun-10 04:59 11:11 6.20 6 GN 22-Jun-10 05:00 18:00 13.00 6 GN 23-Jun-10 05:20 18:14 12.90 6 GN 24-Jun-10 04:57 18:00 13.05 6 GN 25-Jun-10 04:59 18:16 13.28 6 GN 26-Jun-10 05:10 18:20 13.17 6 GN 27-Jun-10 05:22 16:30 11.13 178

Rotation # Group Date Start of contact End of contact Total contact (hrs) 7 LV 04-Jul-10 04:52 18:14 13.37 7 LV 05-Jul-10 04:54 18:08 13.23 7 LV 06-Jul-10 04:55 17:30 12.58 7 LV 07-Jul-10 04:52 15:30 10.63 7 CP 10-Jul-10 04:52 18:00 13.13 7 CP 11-Jul-10 04:51 18:00 13.15 7 CP 12-Jul-10 04:50 18:04 13.23 7 CP 13-Jul-10 04:56 18:00 13.07 7 CP 14-Jul-10 04:52 18:14 13.37 7 CP 15-Jul-10 04:55 14:49 9.90 7 GN 18-Jul-10 04:57 18:00 13.05 7 GN 19-Jul-10 04:48 18:00 13.20 7 GN 20-Jul-10 04:58 18:00 13.03 7 GN 21-Jul-10 04:52 18:00 13.13 7 GN 22-Jul-10 04:46 11:40 6.90 8 LV 30-Jul-10 04:59 18:00 13.02 8 LV 31-Jul-10 04:50 18:00 13.17 8 LV 01-Aug-10 04:55 18:00 13.08 8 LV 02-Aug-10 04:53 15:50 10.95 8 CP 05-Aug-10 04:47 18:00 13.22 8 CP 06-Aug-10 04:49 18:00 13.18 8 CP 07-Aug-10 04:59 18:00 13.02 8 CP 08-Aug-10 04:50 18:00 13.17 8 CP 09-Aug-10 04:53 18:00 13.12 8 CP 10-Aug-10 04:58 14:30 9.53 8 GN 13-Aug-10 04:58 18:00 13.03 8 GN 14-Aug-10 04:53 18:00 13.12 8 GN 15-Aug-10 04:47 18:00 13.22 8 GN 16-Aug-10 04:54 18:00 13.10 8 GN 17-Aug-10 04:52 18:00 13.13 8 GN 18-Aug-10 04:59 15:30 10.52 9 LV 12-Jan-11 17:00 17:30 0.50 9 LV 13-Jan-11 05:14 18:05 12.85 9 LV 14-Jan-11 05:23 18:00 12.62 9 LV 15-Jan-11 05:17 18:00 12.72 9 LV 16-Jan-11 05:20 15:05 9.75 9 CP 19-Jan-11 07:00 18:00 11.00 9 CP 20-Jan-11 06:30 18:00 11.50 9 CP 21-Jan-11 05:30 17:55 12.42 9 CP 22-Jan-11 05:24 18:00 12.60 9 CP 23-Jan-11 05:30 18:00 12.50 9 CP 24-Jan-11 05:24 13:05 7.68 9 GN 27-Jan-11 05:11 18:00 12.82 9 GN 28-Jan-11 05:30 18:00 12.50 9 GN 29-Jan-11 05:09 18:00 12.85 9 GN 30-Jan-11 05:28 18:08 12.67 9 GN 31-Jan-11 05:25 18:00 12.58 9 GN 01-Feb-11 05:23 14:36 9.22 179

Rotation # Group Date Start of contact End of contact Total contact (hrs) 10 LV 08-Feb-11 05:19 18:00 12.68 10 LV 09-Feb-11 06:39 18:00 11.35 10 LV 10-Feb-11 06:54 18:00 11.10 10 LV 11-Feb-11 06:50 14:16 7.43 10 CP 14-Feb-11 06:49 17:30 10.68 10 CP 15-Feb-11 07:20 17:30 10.17 10 CP 16-Feb-11 07:30 18:00 10.50 10 CP 17-Feb-11 05:20 18:09 12.82 10 CP 18-Feb-11 05:32 18:12 12.67 10 CP 19-Feb-11 05:33 15:47 10.23 10 GN 22-Feb-11 05:25 18:10 12.75 10 GN 23-Feb-11 05:30 18:07 12.62 10 GN 24-Feb-11 05:31 18:01 12.50 10 GN 25-Feb-11 05:25 18:07 12.70 10 GN 26-Feb-11 05:26 18:00 12.57 10 GN 27-Feb-11 05:23 12:47 7.40 11 LV 07-Mar-11 05:19 18:02 12.72 11 LV 08-Mar-11 07:42 18:02 10.33 11 LV 09-Mar-11 05:30 18:12 12.70 11 LV 10-Mar-11 05:19 13:10 7.85 11 CP 13-Mar-11 05:24 18:07 12.72 11 CP 14-Mar-11 06:54 18:15 11.35 11 CP 15-Mar-11 06:30 17:00 10.50 11 CP 16-Mar-11 06:37 17:01 10.40 11 CP 17-Mar-11 06:30 18:07 11.62 11 CP 18-Mar-11 05:18 14:47 9.48 11 GN 21-Mar-11 05:18 18:12 12.90 11 GN 22-Mar-11 05:19 18:14 12.92 11 GN 23-Mar-11 05:30 18:11 12.68 11 GN 24-Mar-11 05:16 18:18 13.03 11 GN 25-Mar-11 06:00 17:39 11.65 11 GN 26-Mar-11 05:55 13:11 7.27 12 LV 03-Apr-11 08:55 18:11 9.27 12 LV 04-Apr-11 04:59 18:11 13.20 12 LV 05-Apr-11 06:00 17:32 11.53 12 LV 06-Apr-11 06:00 16:41 10.68 12 CP 09-Apr-11 17:30 17:44 0.23 12 CP 10-Apr-11 06:00 18:09 12.15 12 CP 11-Apr-11 06:00 18:07 12.12 12 CP 13-Apr-11 06:53 17:32 10.65 12 CP 14-Apr-11 06:30 18:09 11.65 12 CP 15-Apr-11 06:10 12:38 6.47 12 GN 17-Apr-11 07:18 18:14 10.93 12 GN 18-Apr-11 05:00 18:20 13.33 12 GN 19-Apr-11 04:53 18:28 13.58 12 GN 20-Apr-11 04:49 18:14 13.42 12 GN 21-Apr-11 04:55 13:07 8.20

180

APPENDIX B: GROUP DEMOGRAPHICS

Age Class by Season Entrance/Exit Name ID Group D.O.B MotherID Sex 1 2 3 Date Cayenne CY LV 1/1/96a -- M Adult Adult Adult Mostaza MZ LV 12/31/00 Kathy Lee M Subadult Subadult -- 5/28/10d Nutella NU LV/CP 5/14/02 Salsa M Subadult Subadult Subadult Blanquita BB LV 1/1/83a -- F Adult Adult Adult 10/23/09d Kathy Lee KL LV 4/1/89 Gringa F Adult Adult Adult Dos Leches DL LV 5/29/91 Gringa F Adult Adult Adult Salsa SL LV 3/1/96 Blanquita F Adult Adult Adult Chutney CH LV 8/2/99 Blanquita F Adult Adult Adult Pickles PI LV 4/15/03 Blanquita F Adult Adult Adult Charchere CE LV 11/1/04 Dos Leches F Juvenile Juvenile Adult Velveeta VV LV 11/1/04 Salsa F Juvenile Juvenile Adult 1/23/11d Oregano OR LV 4/11/05 Blanquita F Juvenile Juvenile Adult Toyo TY LV 12/1/04 Kathy Lee M Juvenile Juvenile Juvenile Nutmeg NT LV 10/23/06 Salsa M Juvenile Juvenile Juvenile Mousse ME LV 3/8/07 Blanquita M Juvenile Juvenile Juvenile Sassafras SF LV 5/17/07 Kathy Lee F Juvenile Juvenile Juvenile Crema CM LV 5/18/07 Dos Leches F Juvenile Juvenile Juvenile 2/25/10d Chai CI LV 2/5/08 Chutney M Juvenile Juvenile Juvenile Thyme TH LV 5/30/08 Salsa F Juvenile Juvenile Juvenile Canela CN LV 2/19/09 Kathy Lee F Infant Juvenile Juvenile Cassia CS LV 6/21/09 Dos Leches F Infant Inf/Juv Juvenile Poppy PY LV 7/2/09 Pickles M Infant Inf/Juv Juvenile Ketchup KC LV 11/24/09 Salsa M Infant Infant Juvenile Vanilla VN LV 5/4/10 Chutney F -- Infant Infant Sage SG LV 7/28/10 Kathy Lee F -- Infant Infant Legolas LE CP 1/1/93a -- M Adult Adult Adult Buzz BZ CP 1/1/96a -- M Adult Adult Adult Rafiki RF CP 6/1/96a -- M -- Adult Adult Jafar JF CP 1/1/00a -- M -- Subadult Adult 1/3/10e Razoul RZ CP 6/1/03a -- M -- Subadult -- 8/7/10d, 1/3/10e

181

Age Class by Season Entrance/Exit Name ID Group D.O.B MotherID Sex 1 2 3 Date Seria SE CP 6/10/89 Patch1 F Adult Adult Adult Timone TI CP 5/16/96 Seria F Adult Adult Adult Simba SI CP 8/5/98 Limp F Adult Adult Adult Zazu ZA CP 2/2/99 Seria F Adult Adult Adult Ed ED CP 5/14/00 Pumba F Adult Adult Adult Sarabi SA CP 1/1/01 Limp F Adult Adult Adult Kiara KI CP 4/29/02 Pumba F Adult Adult Adult Shanti SH CP 3/5/03 Nyla F Adult Adult Adult Baloo BO CP 5/3/03 Timone F Adult Adult Adult Nemo NE CP 3/11/04 Seria F Juvenile Adult Adult Abu AB CP 4/25/05 Timone F Juvenile Juvenile Juvenile Beauty BT CP 3/12/06 Seria F Juvenile Juvenile Juvenile Arial AR CP 6/8/06 Zazu F Juvenile Juvenile Juvenile Dory DY CP 12/19/06 Timone M Juvenile Juvenile Juvenile Lucifer LU CP 2/5/07 Simba F Juvenile Juvenile Juvenile Mad Madam Mim MI CP 4/16/08 Seria F Juvenile Juvenile Juvenile Thimble TB CP 7/6/08 Kiara M Juvenile Juvenile Juvenile Urchin UR CP 7/19/08 Zazu M Juvenile Juvenile Juvenile Fantasia FT CP 7/23/08 Sarabi F Juvenile Juvenile Juvenile Geppetto GP CP 9/2/08 Simba M Juvenile Juvenile Juvenile Figaro FG CP 2/18/09 Timone M Infant Juvenile Juvenile Lady LA CP 4/28/09 Ed F Infant Juvenile Juvenile Lambert LB CP 5/30/09 Shanti M Infant Juvenile Juvenile Toaster TS CP 7/22/09 Baloo F Infant Inf/Juv Juvenile Lampwick LW CP 12/8/09 Seria F -- Infant Juvenile Marlin ML CP 4/28/10 Nemo M -- Infant Infant Badger BD CP 8/20/10 Sarabi M -- -- Infant Perdita PR CP 8/23/10 Simba F -- -- Infant Baba Ghanoush BG GN 1/1/92a -- M Adult Adult Adult Marmite MM GN 1/1/92a -- M Adult Adult Adult Albus AD GN 1/1/92a -- M Adult -- -- 10/23/09d

182

Age Class by Season Entrance/Exit Name ID Group D.O.B MotherID Sex 1 2 3 Date Dumbledore Goyle GY GN 1/1/98a -- M Adult Adult Adult Malfoy MF GN 1/1/00a -- M Adult Adult Adult Mr Muggles MU GN 1/1/01a -- M Adult Adult Adult Crabbe CR GN 1/1/01a -- M Adult Adult Adult Lefty LF GN 1/1/02a -- M Subadult Subadult Subadult Barty Crouch BY GN 1/1/03a -- M Subadultc Subadult Subadult Maxine MX GN 1/1/90a -- F Adult -- -- 10/23/09d Luna Lovegood LL GN 1/1/93a -- F Adult Adult Adult Minerva MV GN 1/1/87a -- F Adult Adult Adult Mrs Weasley MW GN 1/1/95a -- F Adult Adult Adult Lily LY GN 1/1/97a -- F Adult Adult Adult Petunia PT GN 6/23/99a -- F Adult Adult Adult Fleur FL GN 9/21/99a -- F Adult Adult Adult Rita Skeeter RS GN 1/19/01a -- F Adult Adult Adult Lavender LV GN 12/12/01a -- F Adult Adult Adult Padma PD GN 12/11/03a -- F Adult Adult Adult Cho Chang CO GN 1/26/04a Maxineb F Juvenile Adult Adult Voldemort VM GN 1/1/04a -- M Juvenile Juvenile Subadult Pigwidgeon PW GN 1/1/06 Lily M Juvenile Juvenile Juvenile Moody MD GN 1/1/06 Lunab M Juvenile Juvenile Juvenile Krum KR GN 1/1/06 Rosamertab M Juvenile Juvenile Juvenile Hagrid HG GN 4/12/07 Minerva M Juvenile Juvenile Juvenile Quidditch QD GN 5/1/07 Maxine F Juvenile Juvenile Juvenile Luna Snitch SN GN 5/14/07 Lovegood M Juvenile Juvenile Juvenile Peeves PV GN 7/31/07 Rita Skeeter M Juvenile Juvenile Juvenile Dementor DM GN 4/25/08 Lily M Juvenile Juvenile Juvenile Winky WK GN 6/1/08 Mrs Weasley F Juvenile Juvenile Juvenile Phineas PH GN 9/1/08 Petunia M Juvenile Juvenile Juvenile Selkie SK GN 3/2/09 Fleur F Infant Juvenile Juvenile Hufflepuff HP GN 5/22/09 Lavender M Infant Juvenile Juvenile

183

Age Class by Season Entrance/Exit Name ID Group D.O.B MotherID Sex 1 2 3 Date Norbert NO GN 7/21/09 Minerva M Infant -- -- 3/6/10d Griselda GS GN 8/13/09 Maxine F Infant -- -- 10/23/09d Seamus SM GN 8/22/09 Padma M Infant Infant Juvenile Sprout SP GN 10/28/09 Rita Skeeter F Infant Infant Juvenile Neville NV GN 6/3/10 Mrs Weasley M -- Infant Infant C1 C1 GN 8/6/10 Cho Chang U -- Infant -- 8/24/10d Bellatrix BX GN 3/22/11 Petunia F -- -- Infant Amos Diggory DG GN 4/1/11 Minerva M -- -- Infant Luna Colleen Creevey CV GN 4/7/11 Lovegood F -- -- Infant a) First contact >1yr post-birth. Age estimated based on size, physical features, and date of first birth, when applicable. b) Estimated based on social behavior (i.e., proximity, grooming, alloparenting) c) Classified as subadult because the individual has immigrated to a non-natal group d) Denotes estimated date of disappearance e) Denotes estimated date of entry into group When overlap occurred, age class was categorized based on the predominant age category each season.

184

APPENDIX C: FOCAL ANIMAL SAMPLE TOTALS

Total Focal Follows (N) Total Focal Time (hours) Female ID Group 1 2 3 All 1 2 3 All Kathy Lee KL LV 35 42 66 143 5.87 7.04 11.03 23.95 Dos Leches DL LV 35 43 68 146 5.87 7.19 11.36 24.42 Salsa SL LV 36 40 67 143 6.02 6.69 11.18 23.89 Chutney CH LV 34 41 67 142 5.71 6.86 11.19 23.76 Pickles PI LV 35 41 68 144 5.89 6.85 11.35 24.08 Seria SE CP 34 40 64 138 5.69 6.69 10.69 23.08 Timone TI CP 32 39 63 134 5.36 6.52 10.53 22.40 Simba SI CP 33 41 66 140 5.55 6.86 11.04 23.45 Zazu ZA CP 31 40 63 134 5.20 6.69 10.52 22.41 Ed ED CP 33 40 63 136 5.53 6.69 10.54 22.76 Sarabi SA CP 32 41 66 139 5.36 6.85 11.04 23.25 Kiara KI CP 31 38 63 132 5.20 6.36 10.53 22.08 Shanti SH CP 32 40 62 134 5.36 6.70 10.36 22.42 Baloo BO CP 32 40 64 136 5.37 6.69 10.70 22.77 Nemo NE CP 31 42 61 134 5.21 7.03 10.20 22.43 Luna Lovegood LL GN 34 36 69 139 5.70 6.02 11.53 23.25 Minerva MV GN 34 35 71 140 5.69 5.85 11.86 23.40 Mrs Weasley MW GN 33 37 70 140 5.53 6.19 11.69 23.41 Lily LY GN 34 36 70 140 5.71 6.02 11.70 23.42 Petunia PT GN 34 36 70 140 5.70 6.03 11.71 23.44 Fleur FL GN 34 35 70 139 5.70 5.85 11.69 23.24 Rita Skeeter RS GN 35 36 70 141 5.87 6.02 11.69 23.59 Lavender LV GN 34 35 70 139 5.71 5.85 11.70 23.26 Padma PD GN 34 36 68 138 5.71 6.02 11.37 23.10 Cho Chang CO GN -- 36 71 107 -- 6.02 11.86 17.88 Totals are shown per female per season: 1) Sep-Dec 2009; 2) May-Aug 2010; 3) Jan-Apr 2011.

185

APPENDIX D: BEHAVIORAL ETHOGRAM

Codes and definitions of behaviors recorded during focal and scan samples of white-faced capuchin monkeys. Behaviors are adapted from the ethogram used for the Lomas Barbudal Capuchin Monkey Project by Susan Perry and Joseph Manson and the ethogram used during my MA project (Bergstrom 2009).

FOCAL SAMPLE: SOCIAL Collected continuously during focal animal sampling. Affiliative Code Definition contact FC friendly physical contact including hug, inspect, suck, palpate, sniff, and touch grope FG friendly fur rubbing with another individual using a food or plant object handsniffing start/end HS/HE holding another individuals hand to one's face; digits close to/inserted in the nose groom invite/request MI gesture to receive grooming by laying down or crouching near another individual grooming start/end MM/ME picking through the fur of another individual in attempt to remove parasites and insects play PL biting, chasing, hitting, bouncing, pushing, pulling, lunging in a friendly context Aggressive (contact) Code Definition bite GB forcefully closing mouth on an individual; contact with teeth chase GC pursuing another individual at a rapid rate hit GH using one's hand to make forceful, physical contact with another individual pull GL grabbing another individual and bringing them toward oneself in a forceful manner push GP forceful physical contact with motion to move an individual away from oneself lunge GQ an intense motion toward another individual without physical contact pounce GU forcefully jumping directly onto another individual wrestle GW intense physical contact involving tumbling/rolling and entwined limbs Aggressive (non- contact) Code Definition swipe GA motion to grab or hit at another individual tooth grind GG low friction noise from forceful clenching/movement of upper and lower teeth against each other bounce GJ jumping up and down while making eye contact with another individual nip GN un-forcefully closing of one's mouth on another individual; contact with teeth

186

snap at GO an attempt/warning gesture to bite or nip another individual glare GR directing one's gaze at another individual; often involves furrowing of the brow supplant GS taking the place of another individual (at a desired resource) without coming into contact open-mouth threat GT staring gaze towards another individual while opening one's mouth to expose the canines Coalitionary Code Definition head flag HF soliciting help from another individual by quickly looking at them and then toward an opponent embrace HI standing next to/putting one's arm around an individual and directing gazes toward an opponent cheek to cheek HK pressing one's cheek to another individual's and directing gazes toward an opponent overlord HO climbing on top of another individual and directing gazes toward an opponent hip grasp HY grabbing another individual by the hips and pulling them toward oneself to initiate an overlord back up HZ backing up toward another individual to initiate an overlord against an opponent Infant-related Code Definition bridge IB creating a self-infant pathway over broken terrain/canopy using one's body part dismount ID an infant leaves dorsal position fetch IF retrieving an infant from another individual invite II crouching to allow an infant to approach or mount mount IM allowing an infant to assume a dorsal position nurse IN suckling/feeding on milk from the nipple of another individual failed nursing attempt IR attempting to suckle/feed on milk from an individual who is not lactating or responds to deter tantrum IT flailing of infant's limbs/tail, often accompanied by screaming restrain IU using one's arm to keep an infant in a dorsal/mounted position wrestle IW attempting to remove an infant from a dorsal/mounted position Sexual Code Definition sex display SD duck face, pace, turn or pirouette often accompanied by sex squeeks or grunts end copulation/dismount SE removing oneself from a superior dorsal/ventral position over another individual ejaculate SJ excretion of semen from the male's penis, usually during copulation

187

mount SM obtaining a superior dorsal/ventral position over another individual copulate SX copulation between two individuals involving mounting, thrusting and intromission Submissive Code Definition avoid EA actively moves away from another individual cower EC motion to shudder or withdraw from another individual flee EF rapid movement away from another individual grin EG horizontal widening of mouth and tensing of lips; teeth are exposed and together Vocalization Code Definition human alarm VA alert vocalization in response to seeing an unknown human other monkey alarm VU alert vocalization in response to seeing an extra-group monkey bird alarm VB alert vocalization in response to seeing a predatory bird (high pitch/longer than terrestrial/snake) aggressive cough VC short airy and throaty vocalization made during an aggressive encounter terrestrial/snake alarm VD alert vocalization made in response to seeing a terrestrial predator or snake (high-pitch, short) intense vocal threat VI loud, intense pulses, roars, and shrieks; varies by individual lost call VL one to multiple syllable even-toned vocalizations to denote separation from the group mild vocal threat VO low intensity pulses, roars, shrieks compared to intense vocal threat; varies by individual scream VS cry of distress; depending on intensity can range in pitch, tone, and duration yelp VY short cry; usually in response to being startled FOCAL SAMPLE: NON-SOCIAL Collected continuously during focal animal sampling. Movement Code Definition enter tree TE entering a fruiting tree - followed by 4-letter tree species code leave tree TL leaving a fruiting tree - followed by 4-letter tree species code approach to 0.5m AA individual enters proximity to 1 body length of another approach to 2.5m AB individual enters proximity to 5 body lengths of another leave to outside of 0.5m LA individual moves outside the proximity of 1 body length of another leave to outside of 2.5m LB individual moves outside the proximity of 5 body lengths of another

188

Excretion Code Definition urinate UR urination event without sample collection urine collection URINE urination event with sample collection urine wash AU rubbing one's own urine on the body using the palm of the hand defecate PO defecation event without sample collection fecal collection FECAL defecation event with sample collection State behaviors, proximity categories and individual location, recorded for scan samples (2.5 minute FOCAL SCAN SAMPLE intervals) as well as general state changes during focal animal samples. State behavior Code Definition self directed AM focusing behavior toward one’s own body (e.g., self-grooming) forage drink FD drinking from a waterhole, treehole or provisioned source forage fruit FF ingesting or manipulating fruit forage insect FI ingesting or manipulating insects forage pith FP ingesting or manipulating pith forage vertebrate FV ingesting or manipulating vertebrate prey forage flower FW ingesting or manipulating a flower forage unknown FX ingesting or manipulating other food (not categorized or unknown) rest alone RR rest/sit/stand/sleeps outside 5 body lengths of another individual rest social RS rest/sit/stand/sleeps within 5 body lengths of another individual aggressive SA female is directing or receiving an aggressive behavior submissive SE Female is directing or receiving a fearful behavior friendly SF female is directing or receiving an affiliative behavior grope SG female is fur-rubbing with at least one other individual groom SM female is grooming another individual sexual SS female is directing or receiving sexual behavior triadic ST female is participating in coalitionary behavior

189

traveling TT locomoting visually forage VF scanning for food items vigilant VV alert and scanning Proximity Code Definitions 0.5 meters AA within 0.5 meters of another individual 2.5 meters BB between 0.5 and 2.5 meters in distance from another individual 5 meters CC between 2.5 and 5 meters in distance from another individual dorsal DC dorsal carrying an infant contact HH touching another individual nurse NR receive nursing from an infant

190

APPENDIX E: URINALYSIS TEST PARAMETERS

Urinalysis test results and units of measurement for each urinary parameter tested using urine reagent strips (Siemens Healthcare International, Multistix 10 SG Reagent Strips). Note: The primary purpose of urinalysis was to measure the level of ketone bodies present within each urine sample across seasons.

Urinary constituent Possible test results Units Glucose Negative, 1/10, 1/4, 1/2, 1, 2+ 75-125 mg/dL glucose

Bilirubin Negative, small +, moderate ++, large +++ 0.4-0.8 mg/dL bilirubin

Ketone Negative, trace, 5, 10, 15, 40, 80, 160 5-10 mg/dL acetoacetic acid

Specific gravity 1.000, 1.005, 1.010, 1.015, 1.020, 1.025, 1.030 no units

Blood Negative, trace, moderate, hemolyzed trace, small +, moderate ++, 0.015-0.062 mg/dL large +++ hemoglobin pH 5.0, 6.0, 6.5, 7.0, 7.5, 8.0, 8.5 pH units

Protein Negative, trace, 30, 100, 300, 2000+ 15-30 mg/dL albumin

Urobilinogen Normal 0.2, Normal 1, 2, 4, 8 0.2-8 mg/dL

Nitrate Negative, Positive 0.06-0.10 mg/dL nitrate ion

Leukocyte Negative, trace, small +, moderate ++, large +++ 5-15 white blood cells/hpf in urine Sensitivity ranges are published in the product insert

191

APPENDIX F: LABORATORY METHODS USED IN NUTRITIONAL ANALYSES

The following table defines each nutritional component analyzed, and summarizes the methodology used by Dairy One Forage Laboratory, Ithaca, NY, USA according to their reference manual: Analytical Procedures - Forage, Manure, Water June 2012, http://www.dairyone.com/Forage/Procedures/default.htm. Nutritional component Analysis procedure References Dry matter (DM) The subsample was dried at 135 °C for 2 hours and weighed after cooling in a (AOAC International desiccator. 2012d) Crude protein (CP) Nitrogen was determined by combustion analysis using a Leco FP-528 (AOAC International Nitrogen/Protein Analyzer and multiplying by 6.25. 2012b; AOAC International 2012c; AOAC International 2012e; Leco Corporation) Neutral Detergent Fiber (NDF) The percentage of NDF per gram of dry matter was determined using an (ANKOM Technology ANKOM A200 Fiber Analyzer and ANKOM Technology Method 6 – 2011; Van Soest et al. Neutral Detergent Fiber in Feeds - Filter Bag Technique for A200 (4-13-11): 1991) Samples were individually weighed at 0.5g into filter bags and digested for 75 minutes as a group of 24 in 2L of NDF solution in ANKOM A200 Digestion Unit. 4ml of Alpha Amylase and 20g sodium sulfite were added at the start of digestion. Samples were rinsed three times with boiling water for 5 minutes. Alpha Amylase was added to the first 2 rinses. Water rinses were followed by a 3-minute acetone soak and dried at 100 ºC for 2 hours.

Water Soluble Carbohydrates Samples were incubated with water in a 40 ºC bath for 1 hour to extract water (Hall et al. 1999) (WSC) soluble carbohydrates comprised of simple sugars and fructan. This was followed by acid hydrolysis with sulfuric acid and colorimetric reaction with potassium ferricyanide. Next, the percentage of sugar per gram dry matter was determined using a Thermo Scientific Genesys 10S Vis Spectrophotometer after water extraction.

Crude Fat (CF) The percentage of CF per gram dry matter was determined gravimetrically by (AOAC International ether extraction using anhydrous diethyl ether and a Soxtec HT6 System. 2012a)

192

APPENDIX G: NUTRITIONAL COMPOSITION OF VEGETATIVE FOOD ITEMS

Wet Dry Plant Mass Mass % % % % Energy Family Genus species part (g) (g) H2O % DM CP CF WSC % NDF (kJ/gDM) Source Anacardiaceae Spondias mombin RF 4.69 0.64 86.40 13.60 3.70 6.60 48.30 19.10 11.19 1 Anacardiaceae Spondias mombin P na na na na na na na na na 5 Anacardiaceae Spondias purpurea RF 4.28 0.45 89.37 10.63 9.70 2.10 56.40 8.30 11.86 1 Annonaceae Annona reticulata RF 87.00 19.24 77.89 22.11 9.80 6.20 4.80 59.00 4.78 1 Apocynaceae Stemmadenia obovata RF 9.73 4.06 58.33 41.67 16.60 38.30 3.20 38.00 17.74 1 Araliaceae Sciadodendron excelsum RF 0.41 0.09 76.70 23.30 7.60 8.80 54.60 33.40 13.73 1 Areacaceae Acrocomia aculeata SD 6.13 1.99 67.61 32.39 3.60 23.80 25.00 21.50 13.75 1 Asclepiadaceae Matelea quirosii RF 0.84 0.30 73.72 26.28 6.88 5.40 39.40 29.20 9.78 6 Bignoniaceae Tabebuia ochracea RF 4.15 0.67 83.95 16.05 12.88 1.96 33.09 43.50 8.43 3 Boraginaceae Cordia panamensis RF 0.44 0.13 69.56 30.44 7.30 9.20 43.20 49.80 11.92 1 Bromeliaceae Bromelia pinguin RFBR 6.08 1.90 68.74 31.26 8.10 0.80 54.10 12.90 10.71 1 Bromeliaceae Bromelia plumieri RFBR 9.23 1.62 82.44 17.56 4.60 0.70 93.60 7.30 16.70 1 Burseraceae Bursera simaruba RF 0.11 0.09 21.15 78.85 3.50 12.40 1.50 81.70 5.51 1 Burseraceae Bursera simaruba P na na na na na na na na na 5 Cecropiaceae Cecropia peltata RF 0.84 0.30 73.72 26.28 6.88 5.40 39.40 29.20 9.78 6 Chrysobalanaceae Hirtella racemosa RF 0.32 0.05 85.05 14.95 5.90 1.30 63.90 8.30 12.17 1 Cyperaceae Cyperus luzulae GR 0.01 0.01 30.91 69.09 12.80 14.70 2.40 55.90 8.08 9 LSOR Dilleniaceae Curatella americana RF 0.09 0.04 50.66 49.34 11.30 36.00 13.40 29.40 17.69 1 Dilleniaceae Davilla kunthi RF 0.06 0.03 45.30 54.70 8.40 27.60 7.40 54.80 13.04 1 Dilleniaceae Doliocarpus dentatus RF 0.19 0.06 70.25 29.75 8.00 17.70 29.60 45.80 12.96 1 Dilleniaceae Doliocarpus dentatus MF 0.21 0.05 75.60 24.40 7.20 17.10 34.50 42.20 13.42 1 Ebenaceae Diospyros salicifolia RF 1.53 0.55 63.87 36.13 5.70 1.40 19.50 75.60 4.75 1 Elaeocarpaceae Sloanea terniflora RF 0.11 0.05 48.36 51.64 4.60 67.40 8.70 7.10 27.61 1 Euphorbiaceae Margaritaria nobilis MF 0.41 0.09 76.94 23.06 6.00 8.80 8.60 67.50 5.76 1 Fabaceae Bauhinia ungulata FL 0.43 0.06 80.24 19.76 13.64 3.94 28.61 21.25 8.56 7 Fabaceae Centrosema macrocarpum FL 1.01 0.13 87.15 12.85 18.33 1.55 19.51 28.77 6.92 2 Fabaceae Diphysa americana FL 0.43 0.06 80.24 19.76 13.64 3.94 28.61 21.25 8.56 7 Fabaceae Vachellia collinsii RF 0.20 0.12 40.01 59.99 6.10 0.60 94.30 7.00 17.03 1 Fabaceae Vachellia collinsii MF 0.36 0.11 68.62 31.38 7.10 0.70 90.20 8.60 16.55 1 Fabaceae Vachellia cornigera RF 0.84 0.30 73.72 26.28 6.88 5.40 39.40 29.20 9.78 6 193

Wet Dry Plant Mass Mass % % % % Energy Family Genus species part (g) (g) H2O % DM CP CF WSC % NDF (kJ/gDM) Source Fagaceae Quercus oleoides SD 0.90 0.42 53.37 46.63 6.13 2.57 22.99 16.45 5.84 3 Flacourtiaceae Casearia arguta RF 0.60 0.18 70.64 29.36 15.20 24.90 42.00 29.00 18.95 1 Flacourtiaceae Casearia arguta MF 0.88 0.29 67.16 32.84 15.60 26.70 27.60 31.60 17.29 1 Flacourtiaceae Casearia sylvestris RF 0.02 0.01 52.67 47.33 19.50 45.80 10.90 23.50 22.34 1 Flacourtiaceae Prockia crucis RF 0.22 0.02 90.62 9.38 11.50 5.40 48.60 25.70 12.09 1 Flacourtiaceae Zuelania guidonia RF 2.99 1.30 56.53 43.47 8.40 57.70 7.30 6.50 24.36 1 Malpighiaceae Bunchosia ocellata RF 0.71 0.36 50.19 49.81 9.35 0.34 14.68 37.98 4.15 2 Malpighiaceae Byrsonima crassifolia RF 1.40 0.22 84.44 15.56 4.80 10.00 18.10 40.20 7.60 1 Malpighiaceae Byrsonima crassifolia MF 1.80 0.28 84.47 15.53 5.10 7.00 22.40 38.30 7.24 1 Malvaceae Malvaviscus arboreus RF 0.75 0.17 77.50 22.50 14.10 7.80 33.90 45.70 10.97 1 Malvaceae Malvaviscus arboreus FL 0.22 0.05 79.34 20.66 14.90 8.38 20.42 20.19 9.07 2 Melastomataceae Miconia argentea RF 0.04 0.01 72.59 27.41 10.30 7.10 39.80 30.30 11.06 1 Meliaceae Trichilia americana P na na na na na na na na na 5 Moraceae Brosimum alicastrum RF 0.84 0.20 75.79 24.21 13.25 1.07 31.43 27.88 7.88 3 Moraceae Castilla elastica RF 13.59 3.61 73.44 26.56 10.20 11.10 58.30 17.00 15.65 1 Moraceae Ficus bullenei RF 4.20 1.04 75.34 24.66 6.20 8.40 17.10 46.20 7.06 1 Moraceae Ficus cotinifolia RF 0.52 0.16 68.82 31.18 7.35 6.12 24.75 41.17 7.68 8 Ficus Moraceae Ficus hondurensis RF 0.82 0.14 83.22 16.78 6.40 5.10 46.30 32.20 10.74 1 Moraceae Ficus morazaniana RF 9.98 2.02 79.76 20.24 6.20 4.70 39.00 34.60 9.34 1 Moraceae Ficus ovalis RF 0.77 0.17 78.16 21.84 3.30 3.40 68.40 19.90 13.28 1 Moraceae Maclura tinctoria RF 3.99 0.70 82.47 17.53 11.20 15.00 43.50 24.00 14.81 1 Moraceae Trophis racemosa RF 0.25 0.07 72.24 27.76 7.50 3.40 58.70 10.90 12.36 1 Muntingiaceae Muntingia calabura RF 0.81 0.30 62.67 37.33 6.88 9.64 39.40 20.56 11.38 3 Myrsinaceae Ardisia revoluta RF 0.17 0.02 86.42 13.58 2.60 12.20 83.10 4.50 18.94 1 Myrtaceae Callistemon viminalis FL 0.43 0.06 80.24 19.76 13.64 3.94 28.61 21.25 8.56 7 Myrtaceae Eugenia salamensis RF 2.82 0.58 79.45 20.55 3.20 2.80 62.70 14.10 12.09 1 Myrtaceae Psidium guajava RF 28.30 5.39 80.94 19.06 4.10 2.90 26.20 59.10 6.16 1 Passifloraceae Passiflora platyloba RF 99.00 15.63 84.21 15.79 2.20 0.70 23.38 na 4.55 4 Poaceae Lasiacis sorghoidea GR 0.01 0.01 30.91 69.09 12.80 14.70 2.40 55.90 8.08 1 Polygonaceae Coccoloba sp. RF 0.19 0.07 64.29 35.71 9.00 0.38 38.24 24.53 8.05 3 Rhamnaceae Karwinskia calderoni RF 0.25 0.15 38.98 61.02 7.01 0.58 48.16 24.28 9.45 2 Rubiaceae Alibertia edulis RF 10.24 3.82 62.73 37.27 3.40 0.48 26.98 40.82 5.27 2

194

Wet Dry Plant Mass Mass % % % % Energy Family Genus species part (g) (g) H2O % DM CP CF WSC % NDF (kJ/gDM) Source Rubiaceae Chomelia spinosa RF 0.58 0.19 67.74 32.26 3.40 2.70 18.30 60.20 4.65 1 Rubiaceae Genipa americana RF 13.70 5.60 59.12 40.88 8.56 5.10 60.13 37.40 13.42 3 Rubiaceae Guettarda macrosperma RF 2.98 0.65 78.08 21.92 3.60 2.50 40.80 41.40 8.37 1 Rubiaceae Guettarda macrosperma MF 1.84 0.45 75.73 24.27 5.00 3.50 23.30 66.90 6.06 1 Rubiaceae Randia monantha RF 8.48 3.47 59.04 40.96 5.40 0.70 87.20 8.00 15.76 1 Rubiaceae Randia thurberi RF 0.84 0.30 73.72 26.28 6.88 5.40 39.40 29.20 9.78 6 Sapindaceae Allophylus occidentalis RF 0.29 0.07 76.77 23.23 19.30 22.90 5.20 37.60 12.73 1 Sapindaceae Dipterodendron costaricense RF 0.17 0.04 74.43 25.57 4.90 70.00 5.50 4.90 28.10 1 Sapindaceae Paullinia cururu RF 0.19 0.06 70.25 29.75 8.00 17.70 29.60 45.80 12.96 9 DDEN Sapotaceae Manilkara chicle RF 4.79 1.19 75.22 24.78 2.50 10.20 44.30 9.60 11.68 1 Sapotaceae Manilkara chicle FL 0.07 0.02 74.22 25.78 7.70 1.90 45.90 14.80 9.69 1 Simaroubaceae Simarouba glauca RF 1.25 0.33 73.72 26.28 24.20 0.90 68.60 6.20 15.87 1 Simaroubaceae Simarouba glauca UFSD 0.09 0.02 81.28 18.72 28.10 33.60 12.20 13.20 19.40 1 Sterculiaceae Guazuma ulmifolia RF 1.70 1.33 21.57 78.43 6.00 2.28 42.92 37.40 9.05 3 Theophrastaceae Jacquinia nervosa RF 2.29 1.05 53.86 46.14 6.70 2.40 40.20 38.30 8.75 1 Tiliaceae Luehea candida SD 0.03 0.02 38.60 61.40 19.60 15.55 10.76 25.27 10.94 2 Tiliaceae Luehea candida FL 0.43 0.06 80.24 19.76 13.64 3.94 28.61 21.25 8.56 7 Tiliaceae Luehea speciosa SD 0.02 0.02 0.00 100.00 17.81 16.77 5.60 42.27 10.23 3 Tiliaceae Luehea speciosa FL 0.43 0.06 80.24 19.76 13.64 3.94 28.61 21.25 8.56 7 Viscaceae Phoradendron quadrangulare RF 0.04 0.01 82.54 17.46 8.20 7.70 46.10 23.30 11.99 1 Unknown sp. RF 0.84 0.30 73.72 26.28 6.88 5.40 39.40 29.20 9.78 6 Unknown sp. FL 0.43 0.06 80.24 19.76 13.64 3.94 28.61 21.25 8.56 7 Unknown sp. P na na na na na na na na na 5 CP, Crude protein; CF, Crude fat; WSC, water-soluble carbohydrates; NDF, neutral detergent fiber. Macronutrient and energy values are listed as the percentage of dry mass. Plant Parts: RF, ripe fruit; MF, midripe fruit; UF, unripe fruit; SD, seed; UFSD,unripe fruit seed; RFBR, bromeliad ripe fruit; GR, grass seed; FL, flower; P, pith. Sources: 1) Bergstrom (this study); 2) McCabe (2005); 3) Vogel (unpublished, 2004, 2005); 4) USDA (2014); 5) Assigned a value of 1kJ/unit (1 inch) ingested; 6) Median RF values (N=55); 7) Mean FL values (N=3); 8) Assigned congener mean (congener species code listed as first letter of genus and first three letters of species); 9) Assigned value of species with similar size and composition (species code listed).

195

APPENDIX H: NUTRITIONAL COMPOSITION OF INVERTEBRATE FOOD ITEMS

Wet Order: Family: Mas Dry % % % % % % Energy Genus species Invertebrate category s (g) Mass (g) H2O DM CP CF WSC NDF (kJ/gDM) Source Blattodea: Blaberidae: Archimandrita tesselata peppered roach 7.29 2.29 68.54 31.46 86.70 8.90 4.10 34.10 18.55 1 Hemiptera: Cicadidae: Fidicina mannifera cicada 2.62 0.86 67.09 32.91 81.60 6.60 2.80 37.60 16.61 1 Hemiptera: Pentatomidae: various spp. shield bug 0.16 0.06 62.92 37.08 81.60 6.60 2.80 37.60 16.61 3 Hymenoptera: various ants 0.02 6.80x10-3 66.18 33.82 51.76 15.58 2.30 24.53 14.92 2 Hymenoptera: Formicidae: Pseudomyrmex spp. acacia ant larvae 0.07 1.55x10-3 77.95 22.05 51.76 15.58 2.30 11.84 14.92 2 Hymenoptera: Vespidae: Polistes spp. paper wasp larvae 0.10 0.03 66.74 33.26 48.90 10.60 23.70 24.40 16.15 1 Lepidoptera: Noctuidae: medium noctuid various spp. caterpillar 0.26 0.04 86.62 13.38 63.90 12.50 7.20 12.90 16.61 1 Lepidoptera: Noctuidae: various spp. small noctuid caterpillar 0.05 0.01 87.49 12.51 78.40 6.70 7.40 9.60 16.89 1 Lepidoptera: various medium green caterpillars 0.26 0.05 80.17 19.83 60.20 21.90 10.80 12.10 20.13 1 Lepidoptera: Sphingidae: Eumorpha satellitia satellite sphinx caterpillar 1.47 0.21 85.98 14.02 71.10 7.60 4.20 13.10 15.47 1 Lepidoptera: Tortricidae: Cydia deshaisiana jumping bean moth larvae 0.04 0.02 58.96 41.04 32.30 61.00 1.70 22.60 28.66 1 Orthoptera: various small grasshopper 0.85 0.28 67.42 32.58 74.50 6.50 4.00 30.00 15.59 1 large grasshoppers and Orthoptera: various katydids 2.37 0.59 74.91 25.09 70.48 20.72 0.80 27.43 19.74 2 Orthoptera: Gryllidae: various spp. cricket 0.67 0.18 72.71 27.29 74.50 6.50 4.00 30.00 15.59 3 Scorpiones: Buthidae: Centruroides limbatus bark scorpion 1.45 0.57 60.85 39.15 72.30 16.00 4.00 21.30 18.80 1 Various: various combined invertebrates 0.30 0.08 72.96 27.04 60.88 8.13 1.29 38.39 13.47 2 CP, Crude protein; CF, Crude fat; WSC, water-soluble carbohydrates; NDF, neutral detergent fiber. Macronutrient and energy values are listed as the percentage of dry mass. Sources: 1) Bergstrom (this study); 2) McCabe (2005); 3) Assigned values of Order  DM(g) (wet and dry mass and percentages were determined for this project).

196

APPENDIX I: FEEDING, ENERGY INTAKE AND MACRONUTRIENT INTAKE RATES

Only bouts during which specific food items were being targeted were included. Food sources with less than 5 items, 2 bouts or 10 minutes of total feeding time were excluded. F=fruit; S=seed; W=flower; I=invertebrate; pith was excluded because macronutrient information was not available.

Total Macronutrient Intake Rate (g/hr) Wet Dry Feeding Feeding Intake Energy Food Mass Mass Ingests Bouts Duration Rate intake Protein Fat Sugar Fiber Food Source Type (g) (g) (N) (N) (min) (N/min) (kJ/min) (CP) (CF) (WSC) (NDF) Alibertia edulis F 10.24 3.82 6 5 13.18 0.46 9.15 3.54 0.50 28.12 42.54 Annona reticulata F 87.00 19.24 9.1 17 53.20 0.17 15.73 19.35 12.24 9.48 116.50 Acrocomia aculeata F 6.13 1.99 11 6 36.22 0.30 8.30 1.30 8.62 9.05 7.78 Byrsonima crassifolia F 1.40 0.22 81 15 41.87 1.93 3.20 2.86 4.79 11.82 22.55 Bunchosia ocellata F 0.71 0.36 43 10 21.12 2.04 3.01 4.06 0.15 6.38 16.51 Bromelia pinguin F 6.08 1.90 9 5 10.68 0.84 17.15 7.78 0.77 51.97 12.39 Bromelia plumieri F 9.23 1.62 431.5 225 932.35 0.46 12.53 2.07 0.31 42.12 3.28 Bursera simaruba F 0.11 0.09 241 33 63.67 3.79 1.85 0.70 2.49 0.30 16.43 Curatella americana F 0.09 0.04 542 14 40.97 13.23 9.89 3.79 12.07 4.49 9.86 Casearia arguta F 0.60 0.18 191 19 38.18 5.00 16.80 21.61 36.38 46.26 42.81 Cordia panamensis F 0.44 0.13 625 19 66.20 9.44 15.16 5.57 7.02 32.98 38.01 Cecropia peltata F 0.81 0.31 31 12 22.45 1.38 4.25 1.96 1.85 10.38 7.57 Chomelia spinosa F 0.58 0.19 73 8 14.68 4.97 4.36 1.91 1.52 10.29 33.85 Casearia sylvestris F 0.02 0.01 658 21 44.95 14.64 3.23 1.69 3.98 0.95 2.04 Dipterodendron costaricense F 0.17 0.04 105 6 25.30 4.15 5.11 0.53 7.64 0.60 0.53 Doliocarpus dentatus F 0.19 0.06 4818 109 296.55 16.25 11.76 7.96 18.20 33.39 46.06 Davilla kunthi F 0.06 0.03 2218 50 166.05 13.36 6.02 2.33 7.64 2.05 15.17 Diospyros salicifolia F 1.53 0.55 33 11 13.32 2.48 6.48 9.34 2.29 31.95 123.88 Small-sized fig F 0.70 0.16 6206 168 609.10 10.19 19.17 13.54 11.96 165.68 73.15 Large-sized fig F 7.09 1.53 129 22 73.82 1.75 22.88 19.87 19.09 101.23 123.51 Genipa americana F 13.70 5.60 24.85 25 88.10 0.28 21.20 8.11 4.83 56.99 35.45 Guettarda macrosperma F 2.98 0.65 451.5 26 109.37 4.13 22.59 11.35 7.91 91.75 140.87 Karwinskia calderoni F 0.25 0.15 248 9 40.78 6.08 8.66 3.85 0.32 26.48 13.35 Luehea candida S 0.03 0.02 33 10 26.77 1.23 0.26 0.28 0.22 0.15 0.36 Lasiacis sorghoidea S 0.01 0.01 89 10 22.57 3.94 0.26 0.24 0.28 0.05 1.07

197

Total Macronutrient Intake Rate (g/hr) Wet Dry Feeding Feeding Intake Energy Food Mass Mass Ingests Bouts Duration Rate intake Protein Fat Sugar Fiber Food Source Type (g) (g) (N) (N) (min) (N/min) (kJ/min) (CP) (CF) (WSC) (NDF) Luehea speciosa S 0.02 0.02 146 46 118.65 1.23 0.22 0.70 0.66 0.22 1.67 Malvaviscus arboreus F 0.75 0.17 385 51 90.42 4.26 7.93 12.23 6.77 29.41 39.64 Miconia argentea F 0.04 0.01 347 9 18.55 18.71 2.51 1.40 0.97 5.43 4.13 Muntingia calabura F 0.81 0.30 36 6 13.42 2.68 9.19 6.67 9.35 38.20 19.93 Manilkara chicle F 4.79 1.19 596 65 199.92 2.98 41.35 5.31 21.67 94.13 20.40 Margaritaria nobilis F 0.41 0.09 107 6 19.70 5.43 2.94 1.84 2.69 2.63 20.67 Maclura tinctoria F 3.99 0.70 433 53 159.77 2.71 28.09 12.75 17.07 49.52 27.32 Prockia crucis F 0.22 0.02 72 2 10.57 6.81 1.72 0.98 0.46 4.15 2.20 Psidium guajava F 28.30 5.39 9 5 31.48 0.29 9.51 7.59 5.37 48.48 109.36 Phoradendron quadrangulare F 0.04 0.01 201 6 13.97 14.39 1.17 0.48 0.45 2.70 1.36 Randia monantha F 8.48 3.47 118.5 99 323.88 0.37 20.02 4.11 0.53 66.45 6.10 Sciadodendron excelsum F 0.41 0.09 332 4 23.00 14.43 18.70 6.21 7.19 44.62 27.29 Simarouba glauca F 1.25 0.33 115 30 56.93 2.02 10.54 10.24 1.08 27.60 2.75 Spondias mombin F 4.69 0.64 98 26 62.27 1.57 11.23 2.23 3.97 29.08 11.50 Stemmadenia obovata F 9.73 4.06 5.1 5 14.47 0.35 25.36 14.24 32.85 2.75 32.60 Sloanea terniflora F 0.11 0.05 77 29 99.30 0.78 1.17 0.23 3.44 0.44 0.36 Trophis racemosa F 0.25 0.07 542 18 67.77 8.00 6.99 2.55 1.15 19.92 3.70 Vachellia collinsii F 0.20 0.12 36 14 17.13 2.10 4.34 1.94 0.19 27.27 2.30 Zuelania guidonia F 2.99 1.30 10 9 16.62 0.60 19.04 3.94 27.06 3.42 3.05 Bauhinia ungulata W 0.43 0.06 320 38 69.10 4.63 2.28 2.43 0.70 5.10 3.79 Centrosema macrocarpum W 1.01 0.13 78 20 40.60 1.92 1.72 2.73 0.23 2.91 4.29 Diphysa americana W 0.43 0.06 381 4 15.82 24.09 11.87 12.66 3.66 26.55 19.72 Luehea speciosa W 0.43 0.06 73 21 37.08 1.97 0.97 1.03 0.30 2.17 1.61 Malvaviscus arboreus W 0.22 0.05 56 16 15.62 3.59 1.50 1.48 0.83 2.03 2.01 Hymenoptera: Formicidae: Pseudomyrmex spp. I 0.07 0.02 158 40 80.45 1.96 0.45 0.95 0.28 0.04 0.45 Hemiptera: Pentatomidae I 0.16 0.06 76 15 67.13 1.13 1.13 3.33 0.27 0.11 1.53 Hymenoptera: Formicidae I 0.02 0.01 9778 669 1576.57 6.20 0.63 1.31 0.39 0.06 0.62 Hymenoptera: Vespidae: Polistes spp. I 0.10 0.03 333 25 34.37 9.69 5.27 9.57 2.07 4.64 4.78 Lepidoptera: Sphingidae: Eumorpha satellitia I 1.47 0.21 42 19 53.78 0.78 2.49 6.87 0.73 0.41 1.27 198

Total Macronutrient Intake Rate (g/hr) Wet Dry Feeding Feeding Intake Energy Food Mass Mass Ingests Bouts Duration Rate intake Protein Fat Sugar Fiber Food Source Type (g) (g) (N) (N) (min) (N/min) (kJ/min) (CP) (CF) (WSC) (NDF) Lepidoptera: various (med) I 0.26 0.05 3763 501 1824.40 2.06 2.18 3.91 1.42 0.70 0.79 Lepidoptera: various (sm) I 0.05 0.01 1470 65 216.77 6.78 0.71 1.97 0.17 0.19 0.24 Lepidoptera: Tortricidae: Cydia deshaisiana I 0.04 0.02 795 79 320.83 2.48 1.11 0.75 1.41 0.04 0.52

199

APPENDIX J: ENERGY CONSTANTS FOR ENERGY EXPENDITURE DURING ACTIVITY

Activity state Energy constant Rest 1.25a,b Stationary feed 1.38a,b Active feed/forage 1.70* Travel 1.70c Low-intensity social 1.38** High-intensity social 2.35a,b a) Coelho 1974; b) Coelho et al. 1976; c) Taylor et al. 1970; *Assigned the energy constant for travel; **Assigned the energy constant for stationary feed

200