<<

DIET, , AND DENTAL

MORPHOLOGY IN TERRESTRIAL

Sílvia Pineda-Munoz, MSc Department of Biological Sciences Macquarie University Sydney, Australia

Principal Supervisor: Dr. John Alroy Co-Supervisor(s): Dr Alistair R. Evans Dr Glenn A. Brock

This thesis is submitted for the degree of Doctor of Philosophy April 2016

2

To my Little Bean; and her future siblings and cousins

Al meu Fessolet;

I als seus futurs germans i cosins

i

ii

STATEMENT OF CANDIDATE

I certify that the work in this thesis entitled “Diet, ecology and dental morphology in terrestrial mammals” has not previously been submitted for a degree nor has in been submitted as part or requirements for a degree to any other university or institution other than Macquarie University.

I also certify that this thesis is an original piece of research and that has been written by me. Any collaboration, help or assistance has been appropriately acknowledged. No

Ethics Committee approval was required.

Sílvia Pineda-Munoz, MSc

MQID: 42622409

iii

iv Diet, ecology, and dental morphology in terrestrial mammals – Silvia Pineda-Munoz – November 2015

ABSTRACT

Dietary inferences are a key foundation for paleoecological, ecomorphological and macroevolutionary studies because they inform us about the direct relationships between the components of an . Thus, the first part of my thesis involved creating a statistically based diet classification based on a literature compilation of stomach content data for 139 terrestrial mammals. I observed that diet is far more complex than a traditional -- classification, which masks important feeding specializations. To solve this problem I proposed a new classification scheme that emphasizes the primary resource in a given diet (Chapter 3).

This new classification was then contrasted with body mass (Chapter 4). I observed that there is a specific optimum body mass range for every dietary specialization, with the medium size range mostly composed of frugivorous that inhabit tropical and subtropical rainforests. Thus, the near absence of medium-sized mammals in open environments can be linked to the decreasing density of fruit trees needed to support a pure frugivorous diet all year round.

I then evaluated previous dietary proxies and observed that a relevant time scale needs to be determined before choosing a dietary proxy (Chapter 5).

The main goal of my PhD research was to design quantitative and phylogeny-free method to infer the typical diet of each species. I therefore designed Multi-Proxy Dental

Morphology Analysis (MPDMA) (Chapter 6). I three dimensionally scanned the of 138 extant mammals (28 marsupials and 110 placentals) and qualitatively classified their diets. Multiple variables were estimated from the 3D scans (i.e., orientation patch count, slope diversity, and relief index) and multivariate statistical analyses were used to test for discrimination power across dietary specializations

(Chapter 7). MPDMA demonstrates significant morphological differences across diets

(P < 0.05) and correctly discriminates diet for up to 82% of the specimens in the

v

dataset. Most remarkably, marsupials and placentals with the same dietary specializations overlap strongly in ecomorphospace, which suggests convergent phenotypic across both clades.

vi

ACKNOWLEDGEMENTS

I would like to thank the Higher Degree Research team at Macquarie University and the

Office of Fellowships at the Smithsonian Institution for financial support. Additionally,

I would like to thank all of the Paleobiology Database Workshop in Analytical Methods instructors for bringing me the opportunity to learn so much in such a short time.

This thesis would have never materialized without the support I received from colleagues and friends from all around the world.

Thank you to Kayla Friedman and Malcolm Morgan of the Centre for Sustainable

Development, University of Cambridge, UK for producing the Microsoft Word thesis template used to produce this document.

To my advisors:

I would first like to thank John Alroy for believing in me when I first wrote you more than 4 years ago. For being available to chat any time I knocked your door. For teaching me how to approach science and for your invaluable advice during the whole journey of my Ph.D. For all the constructive comments on my manuscripts and for challenging me to keep improving my English skills. For saying “you go and figure it out” when I asked you how to write in R. For helping me when I first arrived in Australia. I have told you this so many times, but you will always be my advisor.

I am also incredibly grateful to my co-advisor Alistair Evans. For opening the door to your lab when I first asked for your help, and teaching me how to think out of the box.

vii

For being patient with my curiosity and brainstorming and making me feel like home every time I visited Melbourne. For the long conversations about science and life.

John and Alistair: I can’t think of a better duo of advisors, and I will always remember the time you both gave me advice about my Romer Session presentation at SVP in

2015. Siting next to each other, commenting on every single slide. That was the best criticism I have ever received.

Thanks Glenn Brock. Although my research was so different from yours, you were always there for support and advice. I will always think of our lab as another home to me.

My thesis was greatly improved by the opportunity Kay Behrensmeyer and Kate Lyons created for me by awarding me a pre-doc Research Fellowship at the National Museum of Natural History Smithsonian Institution. Thank you Kay for introducing me to so many researchers and believing in my work. For including me to your family plans for all the holidays – I can’t wait to run another race together! Thank you Kate for the long conversations and for making me see the importance of my work. But mainly, for trusting me and offering me such a rewarding post-doctoral position with the ETE program.

During my Ph.D. program I was incredibly lucky to work in three amazing labs:

The Paleobiology lab in Macquarie University. Thanks Graeme, Sarah, Bryony, Luke,

Marissa, Sarah, Patrick, David, James, Julieta, Nick, Christian, Gabrielle and Matt. For the lab meetings and the paper reading clubs. You inspired me to keep learning about new topics every day. Special thanks to the Genes to Geosciences team.

The Evans lab at Monash University. Thanks Matt, David, Travis, Angi, Lap, Alana and

Kathleen. It was wonderful to hang out and discuss science.

viii

The National Museum of Natural History Smithsonian Institution. Thanks Matt,

Andrew, Aniko, Advait, Amelia, Carlos, Andrew, Danielle, Laura, Arden, Scott, Rich,

René, Thomas, and everybody who helped me during my stay at a wonderful institution.

To my master's thesis advisor Isaac Casanovas-Vilar. I will soon lose my Padawan status, Mestre. And Mikael Fortelius, the first person in palaeontology to give me an opportunity.

To my lab mates at the Institut Català de Paleontologia Miquel Crusafont. Special thanks to Daniel de Miguel, who taught me how to mould and cast. To my dear Maria: I miss our coffee breaks. Fortu, you’ve inspired me so much with your missal. Albert, your support has been crucial in so many ways. David, Marta, Miriam, Gretel, Salvador,

Alba, Novella.

To my friends at the University of Helsinki: Juha, Allu, Ellodie and Pierre.

To my family:

Thank you Mama. For encouraging me to give my best every day, and being there to share my laughs and wipe my tears. For our late night conversations and for picking up the phone at 3 a.m. pretending you were awake. For keeping me in touch with my most side, for keeping my secrets and being my best friend in so many ways.

Thank you Ferran (my dad). You used to tell me: “Aim for the sun, and you’ll reach the moon”. One of the last times I saw you, you told me: “I told you to aim for the sun so you could reach the moon; but you’ve gone beyond the sun already”. You made me so happy. But also thank you for the hikes in the mountains, for teaching me how to love nature.

Thank you Tata-capu (my sister). For being there to cheer me up when I needed it the most, for being proud of me despite our having such different lives. But also, for giving me what I love the most in this world, my little bean, my niece Èlia, to whom I dedicate this thesis.

ix

Thank you to my grandmas and grandpas, uncles and aunties, and cousins. You are always there to support me although you sometimes struggle to understand my life as a scientist. I miss you tito Jose.

To all my friends:

Who have always been there to cheer me up and share the good and bad times. Thank you Zoraida, for being there to read my messages from Catalunya, the Czech Republic or Galicia. Julieta, for becoming my “sister” overseas and for our adventures together that have just started. Laura, for you becoming someone so important in my life in

Australia. Peri, for so many important days together, for being my first climbing partner. Aniko, for your incredible friendship. Leo, for belaying me on the wall and in my personal life. Yuri, for mutual inspiration. Nick, the best lab-mate I can think of.

Graeme, for dealing so well with all my quirkiness. Pere and Anna, you were so much like a little piece of home overseas.

Thanks to my advisors’ families: Heather, Adi and Linus; Gudrun, Luke, David and

Zoe; Bill, Sarah and Kristina; Kieran and Pete. For making me feel like home every time I visited.

To my friends back in Catalunya who are always there to chat no matter what the time.

Special thanks to Tere, Cristian and Nuri, for having made it so far to visit me. Nica, for taking my auntie duties over and becoming part of the family. To my friends in

Australia, with a special mention to the Catalan crew for always being there to celebrate with me, and to my roommates in Shirley Road. Also James,Vashi, Pato, Laura and

Koos for housing me during my last few weeks in Australia. Thanks to friends in DC:

Laís, Sasha, Bel. And to my awesome roommates Eboni, Spencer and Katherine, you have all inspired me in so many ways. Thanks Nacho for being there on the last stages of my thesis.

x

A most special thanks to everyone who has ever climbed a wall or hiked a mountain with me. You remind me of the effort and sacrifice to keep moving up and the satisfaction once you make it! And to all my travel-mates.

xi

CONTENTS

1 PROLOGUE ...... XXX

2 INTRODUCTION ...... 32

3 DIETARY CLASSIFICATION OF TERRESTRIAL MAMMALS ...... 38

3.1 ABSTRACT ...... 39

3.2 INTRODUCTION ...... 39

3.3 MATERIAL AND METHODS ...... 41

3.4 RESULTS ...... 44

3.5 DISCUSSION ...... 50

3.5.1 Proposed categories ...... 50

3.5.2 Earlier classification schemes ...... 51

3.5.3 Frequency of dietary specialization ...... 54

3.5.4 Implications for research ...... 55

3.6 ACKNOWLEDGMENTS ...... 55

3.7 AUTHOR’S CONTRIBUTIONS ...... 55

4 THE RELATIONSHIP BETWEEN DIET AND BODY MASS IN

TERRESTRIAL MAMMALS ...... 56

4.1 ABSTRACT ...... 57

4.2 INTRODUCTION ...... 58

4.3 METHODS ...... 60

4.4 RESULTS ...... 64

4.5 DISCUSSION ...... 69

4.5.1 Body mass and the degree of diet specialization ...... 69

4.5.2 Frugivory and body mass ...... 71

4.5.3 Frugivory and the medium-sized gap ...... 75

4.5.4 The consequences of improper diet classifications ...... 76 xii

4.6 ACKNOWLEDGEMENTS ...... 77

4.7 AUTHOR’S CONTRIBUTIONS ...... 77

5 THE TEMPORAL SCALE OF DIET AND DIETARY PROXIES ...... 78

5.1 ABSTRACT ...... 79

5.2 INTRODUCTION ...... 79

5.3 DISCUSSION ...... 83

5.3.1 Diet changes over time ...... 83

5.3.2 Cross-scale studies offer a richer view of diet ...... 91

5.3.3 Towards a cross-scale dietary framework ...... 93

5.4 ACKNOWLEDGEMENTS ...... 95

5.5 AUTHOR’S CONTRIBUTIONS ...... 95

6 DENTAL MORPHOLOGY PREDICTS DIET ACROSS MARSUPIALS AND

PLACENTALS ...... 96

6.1 ABSTRACT ...... 97

6.2 MAIN TEXT ...... 97

6.3 MATERIALS AND METHODS: ...... 104

6.3.1 Specimens ...... 104

6.3.2 Data collection ...... 104

6.3.3 Dietary classification ...... 105

6.3.4 Data analysis ...... 106

6.4 ACKNOWLEDGMENTS: ...... 106

6.5 AUTHOR’S CONTRIBUTIONS ...... 107

7 DENTAL MORPHOLOGY VARIABILITY IN RELATION TO DIET IN

TERRESTRIAL MAMMALS ...... 108

7.1 ABSTRACT ...... 109

7.2 INTRODUCTION ...... 110

xiii

7.3 MATERIALS AND METHODS ...... 112

7.4 DENTAL MORPHOLOGY PROXIES: RESULTS AND DISCUSSION ...... 114

7.4.1 Orientation Patch Count (OPCR) ...... 114

7.4.2 Relief Index (RI) ...... 120

7.4.3 Dental Slope ...... 122

7.4.4 MPDMA ...... 124

7.5 DISCUSSION ...... 127

7.5.1 Relationship between upper and lower tooth rows ...... 127

7.5.2 The benefits of mixing different dietary proxies ...... 129

7.5.3 Diet categories and dental morphology ...... 130

7.6 ACKNOWLEDGMENTS ...... 130

7.7 AUTHOR’S CONTRIBUTIONS ...... 131

8 CONCLUSIONS ...... 132

9 FUTURE WORK ...... 138

10 REFERENCES ...... 142

11 APPENDICES ...... 166

11.1 SUPPLEMENTARY INFORMATION 1 ...... 167

11.2 SUPPLEMENTARY INFORMATION 2 ...... 182

11.3 SUPPLEMENTARY INFORMATION 3 ...... 189

11.3.1 Proxies measure diet at different temporal scales ...... 189

11.3.2 Temporal dataset preparation ...... 192

11.4 SUPPLEMENTARY INFORMATION 4 ...... 201

11.4.1 Dietary classification ...... 201

11.4.2 MPDMA data ...... 201

11.4.3 Figures ...... 212

11.4.4 PCA stats ...... 217 xiv

LIST OF TABLES

TABLE 3.5-1: FEEDING CATEGORIES BASED ON THE NEW CLASSIFICATION CRITERIA AND

ON THE CLASSIC CRITERIA (ASTERISKS INDICATE THOSE ALSO

IDENTIFIED BY EISENBERG (1989))...... 51

TABLE 4.4-1: P-VALUES OF PAIRWISE COMPARISONS BASED ON THE WILCOXON RANK-

SUM TEST FOR BODY MASS DISTRIBUTIONS IN DIET CATEGORIES AS DESCRIBED BY

PINEDA-MUNOZ & ALROY (2014) FOR (A) THE WHOLE DATASET AND (B) ONLY

RODENTS...... 65

TABLE 4.4-2: PROPORTION OF ANIMALS IN EACH DIETARY CATEGORY FOR EACH BODY

MASS RANGE AS DISCUSSED IN THE TEXT...... 67

TABLE 4.4-3: AVERAGE, STANDARD DEVIATION, AND MAXIMUM DIET DIVERSITY VALUES

FOR EACH BODY MASS RANGE AS DISCUSSED IN THE TEXT...... 69

TABLE 6.2-1: MANOVA AND DISCRIMINANT ANALYSIS RESULTS FOR SEVEN DATASETS.

BOTH TESTS WERE CARRIED OUT BY USING EIGHT DIETARY CATEGORIES: HERBIVORY,

CARNIVORY, FRUGIVORY, GRANIVORY, INSECTIVORY, FUNGIVORY, GUMIVORY AND

GENERALIZED. MANOVA WAS PERFORMED USING THE TEST PILLAI. DISCRIMINANT

ANALYSIS VALUES INDICATE THE PROPORTION OF SPECIES PROPERLY ASSIGNED IN

EACH DATASET...... 101

TABLE 7.4-1: P-VALUES OF PAIRWISE COMPARISONS BASED ON THE WILCOXON RANK-

SUM TEST FOR OVERALL OPCR IN DIET CATEGORIES AS DESCRIBED BY PINEDA-

MUNOZ & ALROY (2014) FOR (A) LOWER TOOTH ROWS AND (B) UPPER TOOTH ROWS

...... 115

TABLE 7.4-2: P-VALUES OF PAIRWISE COMPARISONS BASED ON THE WILCOXON RANK-

SUM TEST FOR AVERAGE OPCR BETWEEN INDIVIDUAL TEETH IN DIET CATEGORIES AS

xv

DESCRIBED BY PINEDA-MUNOZ & ALROY (2014) FOR (A) LOWER TOOTH ROWS AND

(B) UPPER TOOTH ROWS ...... 117

TABLE 7.4-3: P-VALUES OF PAIRWISE COMPARISONS BASED ON THE WILCOXON RANK-

SUM TEST FOR STANDARD DEVIATION OF OPCR BETWEEN INDIVIDUAL TEETH IN DIET

CATEGORIES AS DESCRIBED BY PINEDA-MUNOZ & ALROY (2014) FOR (A) LOWER

TOOTH ROWS AND (B) UPPER TOOTH ROWS ...... 119

TABLE 7.4-4: P-VALUES OF PAIRWISE COMPARISONS BASED ON THE WILCOXON RANK-

SUM TEST FOR RELIEF INDEX IN DIET CATEGORIES AS DESCRIBED BY PINEDA-MUNOZ

& ALROY (2014) FOR (A) LOWER TOOTH ROWS AND (B) UPPER TOOTH ROWS ...... 121

TABLE 7.4-5: P-VALUES OF PAIRWISE COMPARISONS BASED ON THE WILCOXON RANK-

SUM TEST FOR AVERAGE SLOPE IN DIET CATEGORIES AS DESCRIBED BY PINEDA-

MUNOZ & ALROY (2014) FOR (A) LOWER TOOTH ROWS AND (B) UPPER TOOTH ROWS.

...... 122

TABLE 7.4-6: P-VALUES OF PAIRWISE COMPARISONS BASED ON THE WILCOXON RANK-

SUM TEST FOR STANDARD DEVIATION OF THE SLOPE IN DIET CATEGORIES AS

DESCRIBED BY PINEDA-MUNOZ & ALROY (2014) FOR (A) LOWER TOOTH ROWS AND

(B) UPPER TOOTH ROWS ...... 124

TABLE 7.4-7: MANOVA AND DISCRIMINANT ANALYSIS RESULTS FOR SEVEN DATASETS.

BOTH TESTS WERE CARRIED OUT BY USING EIGHT DIETARY CATEGORIES: HERBIVORY,

CARNIVORY, FRUGIVORY, GRANIVORY, INSECTIVORY, FUNGIVORY, GUMIVORY AND

GENERALIZATION. MANOVA WAS PERFORMED USING THE TEST PILLAI.

DISCRIMINANT ANALYSIS VALUES INDICATE THE PROPORTION OF SPECIMENS

PROPERLY ASSIGNED IN EACH DATASET...... 127

2 TABLE 7.5-1: LINEAR CORRELATION (R ) BETWEEN UPPER AND LOWER TOOTH ROWS FOR

THE DENTAL MORPHOLOGY VARIABLES STUDIED IN THE PRESENT WORK...... 128

xvi

TABLE 11.1-1: SUMMARY OF THE LITERATURE RESOURCES FOR THE STOMACH CONTENT

DATA FOR THE 139 MAMMALIAN SPECIES ANALYSED IN THIS STUDY...... 167

TABLE 11.1-2: SUMMARY OF STOMACH CONTENT DATA FOR THE 139 MAMMALIAN

SPECIES ANALYSED IN THIS STUDY. U=UNSPECIFIED; A= % SEEDS, B= %

INVERTEBRATES; C= % FUNGI; D= % ; E= % FLOWERS AND GUMS; F=

% ROOTS AND TUBERS; G= % GREEN PLANTS, VEGETATION; H= % FRUIT; N=

SAMPLE SIZE...... 176

TABLE 11.1-3: FEEDING RESOURCES INCLUDED IN EACH OF THE FEEDING CATEGORIES . 180

TABLE 11.1-4: IMPORTANCE OF THE COMPONENTS AND LOADINGS OF THE VARIABLES

BASED ON A PRINCIPAL COMPONENT ANALYSIS OF THE STOMACH CONTENT DATA FOR

THE 139 SPECIES IN THIS STUDY...... 180

TABLE 11.2-1: DATASET SUMMARISING BODY MASS AND DIET DIVERSITY DATA FOR THE

139 SPECIES IN THE STUDY. DIET DATA AS IN TABLE 11.1.1...... 182

TABLE 11.2-2: P-VALUES OF PAIRWISE COMPARISONS BASED ON THE WILCOXON RANK-

SUM TEST FOR BODY MASS DISTRIBUTIONS IN DIET CATEGORIES FROM WILMAN ET

AL. (2014) DATA...... 185

TABLE 11.2-3: IMPORTANCE OF THE COMPONENTS AND LOADINGS OF THE VARIABLES

BASED ON A PRINCIPAL COMPONENT ANALYSIS OF THE STOMACH CONTENT DATA FOR

THE 139 SPECIES IN THIS STUDY...... 185

TABLE 11.2-4: IMPORTANCE OF THE COMPONENTS AND LOADINGS OF THE VARIABLES

BASED ON A PRINCIPAL COMPONENT ANALYSIS FOR WILMAN ET AL. (2014) DATA.186

TABLE 11.3-1: FEASIBILITY AND TIME AVERAGING FOR POPULAR DIET PROXIES.

*NONLETHAL METHODS HAVE BEEN DEVELOPED TO RETRIEVE STOMACH CONTENTS

IN CERTAIN SPECIES (KRONFELD AND DAYAN 1998). †METHODS EXIST FOR

DETAILED EXAMINATION OF DENTAL MICRO- AND MESOWEAR IN LIVING INDIVIDUALS

(BARNES AND LONGHURST 1960, TEAFORD AND OYEN 1989)...... 200

xvii

TABLE 11.4-1: MPDMA ANALYSIS DATA FOR LOWER TOOTH ROWS OF 138 MAMMALIAN

SPECIMENS. A=OVERALL OPCR; B=AVERAGE OPCR BETWEEN INDIVIDUAL TEETH

IN A TOOTH ROW; C=STANDARD DEVIATION OF THE OPCR BETWEEN INDIVIDUAL

TEETH IN A TOOTH ROW; D= RELIEF INDEX; E=AVERAGE SLOPE; F=STANDARD

DEVIATION OF THE SLOPE. INFORMATION ON EACH VARIABLE CAN BE FOUND IN THE

MAIN TEXT. VALUES WERE CALCULATED USING SURFER MANIPULATOR (EVANS

2011). EACH SPECIMEN IS LISTED WITH ITS MUSEUM COLLECTION ID...... 201

TABLE 11.4-2: MPDMA ANALYSIS DATA FOR UPPER TOOTH ROWS OF 138 MAMMALIAN

SPECIMENS. A=OVERALL OPCR; B=AVERAGE OPCR BETWEEN INDIVIDUAL TEETH

IN A TOOTH ROW; C=STANDARD DEVIATION OF THE OPCR BETWEEN INDIVIDUAL

TEETH IN A TOOTH ROW; D= RELIEF INDEX; E=AVERAGE SLOPE; F=STANDARD

DEVIATION OF THE SLOPE. INFORMATION ON EACH VARIABLE CAN BE FOUND IN THE

MAIN TEXT. VALUES WERE CALCULATED USING SURFER MANIPULATOR (EVANS

2011). EACH SPECIMEN IS LISTED WITH ITS MUSEUM COLLECTION ID ...... 206

TABLE 11.4-3: PRINCIPAL COMPONENT ANALYSIS LOADINGS AND IMPORTANCE OF THE

COMPONENTS BASED ON MPDMA DATA FOR 138 MAMMALIAN DENTITIONS

BELONGING TO THE ORDERS RODENTIA, CARNIVORA, AND

PRIMATES. DATA WERE SQUARE-ROOT TRANSFORMED BEFORE PCA...... 217

TABLE 11.4-4: PRINCIPAL COMPONENT ANALYSIS LOADINGS AND IMPORTANCE OF THE

COMPONENTS BASED ON MPDMA DATA FOR 74 MAMMALIAN DENTITIONS

BELONGING TO THE ORDERS CARNIVORA, DIPROTODONTIA AND . DATA

WERE SQUARE-ROOT TRANSFORMED BEFORE PCA...... 217

TABLE 11.4-5: PRINCIPAL COMPONENT ANALYSIS LOADINGS AND IMPORTANCE OF THE

COMPONENTS BASED ON MPDMA DATA FOR 64 MAMMALIAN DENTITIONS

BELONGING TO THE ORDER RODENTIA. DATA WERE SQUARE-ROOT TRANSFORMED

BEFORE PCA...... 218 xviii

TABLE 11.4-6: PRINCIPAL COMPONENT ANALYSIS LOADINGS AND IMPORTANCE OF THE

COMPONENTS BASED ON MPDMA DATA FOR 32 MAMMALIAN DENTITIONS

BELONGING TO THE ORDER CARNIVORA. DATA WERE SQUARE-ROOT TRANSFORMED

BEFORE PCA...... 218

TABLE 11.4-7: PRINCIPAL COMPONENT ANALYSIS LOADINGS AND IMPORTANCE OF THE

COMPONENTS BASED ON MPDMA DATA FOR 28 MAMMALIAN DENTITIONS

BELONGING TO THE ORDER DIPROTODONTIA. DATA WERE SQUARE-ROOT

TRANSFORMED BEFORE PCA...... 219

TABLE 11.4-8: PRINCIPAL COMPONENT ANALYSIS LOADINGS AND IMPORTANCE OF THE

COMPONENTS BASED ON MPDMA DATA FOR 16 MAMMALIAN DENTITIONS

BELONGING TO THE ORDER PRIMATES. DATA WERE SQUARE-ROOT TRANSFORMED

BEFORE PCA...... 219

TABLE 11.4-9: SPECIMENS FOR WHICH DISCRIMINANT ANALYSIS INFERRED DIETARY

CLASSIFICATIONS THAT MISMATCHED OBSERVED DIETS EXTRACTED FROM THE

LITERATURE (NOWAK 1999, WILMAN ET AL. 2014) FOR (A) THE WHOLE DATASET IN

TABLES 1 AND 2 AND FOR (B) THE WHOLE DATASET EXCLUDING RODENTS...... 219

xix

LIST OF FIGURES

FIGURE 3.4-1: SCORES OF THE TWO FIRST COMPONENTS OF THE PRINCIPAL COMPONENT

ANALYSIS OF DIETARY DATA FOR ALL THE 139 SPECIES IN THE DATASET. RAW DATA

ARE SQUARE-ROOT TRANSFORMED PERCENTAGES OF FOOD ITEMS BASED ON

STOMACH CONTENT EXAMINATIONS. ARROWS INDICATE THE LOADINGS OF THE TWO

FIRST COMPONENTS OF THE ANALYSIS. SYMBOLS REPRESENT TAXONOMIC ORDERS.

CROSSES INDICATE ORDERS EACH REPRESENTED BY LESS THAN THREE SPECIES...... 45

FIGURE 3.4-2: UPGMA CLUSTER ANALYSIS BASED ON EUCLIDEAN DISTANCES OF

DIETARY DATA FOR THE 139 SPECIES IN THE DATASET. RAW DATA ARE AS IN FIG. 1.

NUMBERS REPRESENT CLUSTERS IDENTIFIED IN TABLE 1. * BRANCHES CONTAINING

MIXED-FEEDING SPECIES (SEE TEXT)...... 49

FIGURE 3.5-1: DENDROGRAMS SHOWING A) THE PROPORTIONS OF PLANT AND ANIMAL

RESOURCES IN EACH SPECIES' DIET AND B) FEEDING CLASSIFICATIONS BASED ON

CLASSIC TROPHIC RELATIONSHIPS (HERBIVORE-OMNIVORE-CARNIVORE). THE 139

SPECIES IN THE DATASET ARE SORTED BY DESCENDING PROPORTIONS OF PLANT

RESOURCES IN THEIR DIETS...... 52

FIGURE 4.4-1: BOXPLOT ILLUSTRATING THE BODY MASS RANGE FOR EACH DIETARY

CATEGORY OF THE 139 SPECIES IN THE DATASET. CATEGORIES ARE AS

DESCRIBED BY PINEDA-MUNOZ AND ALROY (2014)...... 65

FIGURE 4.4-2: SCORES OF THE TWO FIRST AXES OF THE PRINCIPAL COMPONENT ANALYSIS

OF DIETARY DATA FOR ALL THE 139 SPECIES IN THE DATASET (ADAPTED FROM

PINEDA-MUNOZ AND ALROY [2014]). RAW DATA ARE PERCENTAGE OF STOMACH

CONTENT EXAMINATIONS RESCALED TO Z-SCORES. ARROWS INDICATE LOADINGS ON

THE TWO FIRST COMPONENTS. COLOURS AND SIZES INDICATE THE LOG10 BODY MASS

(G) OF EACH SPECIES AS SHOWN IN THE LEGEND...... 67

xx

FIGURE 4.4-3: RELATIONSHIP BETWEEN LOG10 BODY MASS (G) AND DIET DIVERSITY

INDICES CALCULATED BY APPLYING INVERSE SIMPSON INDICES TO STOMACH

CONTENT PERCENTAGES. COLOURS REPRESENT DIFFERENT DIETARY

SPECIALIZATIONS. THE LINE CONNECTS THE MAXIMUM DIETARY DIVERSITY VALUE

FOR EACH HALF-LOG10 UNIT IN ORDER TO VISUALIZE THE MAXIMUM DEGREE OF

FOOD MIXING. URSUS ARCTOS HAS BEEN EXCLUDED FROM THE MAXIMUM DIETARY

DIVERSITY CURVE IN ORDER TO BETTER VISUALIZE THE GENERAL PATTERN AND FOR

REASONS MENTIONED IN THE TEXT...... 68

FIGURE 4.5-1: GEOGRAPHICAL DISTRIBUTION OF TROPICAL AND SUB-TROPICAL MOIST

BROADLEAF FORESTS (GREEN) (ADAPTED FROM OLSON ET AL. (2001)) AND OF PURE

FRUGIVORE (BLUE CONTOUR LINE) AND MIXED SPECIES (RED CONTOUR

LINE) IN OUR DATASET (EXTRACTED FROM MAP OF LIFE (MOL))...... 72

FIGURE 5.2-1: (A) A PROXY’S RESOLUTION IS GIVEN BY ITS TEMPORAL GRAIN. TEMPORAL

EXTENT IS THE RANGE OF TIME OVER WHICH A PROXY CAN BE USED. PROXIES

MARKED BY A POINTED END HAVE RANGES THAT EXTEND PAST THE GRAPH LIMITS.

DOUBLE POINTED ENDS INDICATE THAT PROXIES CAN BE USED IN EXCEPTIONAL

FOSSIL CASES. (B) THE PERCEIVED DIET OF THE AFRICAN BUSH ELEPHANT

(LOXODONTA AFRICANA) DEPENDS ON THE TIMESPAN OVER WHICH IT IS MEASURED.

BLUE SHADED AREA REPRESENTS DIETARY LIMITS OF ELEPHANTS AT DIFFERENT

SCALES ESTIMATED FROM OBSERVATIONAL AND FOSSIL EVIDENCE. GREY LINES

SHOW ACTUAL DIETS AT DIFFERENT SCALES COMPUTED FROM THE 6 YEAR ISOTOPIC

RECORD OF (CERLING ET AL. 2009). FOUR LINES ARE HIGHLIGHTED IN COLOUR TO

SHOW HOW PERCEIVED DIET CHANGES AS IT IS AVERAGED FOR LONGER PERIODS OF

TIME. BOTH GRAPHS SHARE THE SAME LOGGED X AXIS GIVEN IN YEARS ABOVE AND

COMMON CALENDAR UNITS BELOW...... 82

xxi

FIGURE 5.3-1: DIET CHANGES OVER TIME. (A) SCORES OF THE TWO FIRST COMPONENTS OF

THE PRINCIPAL COMPONENT ANALYSIS OF DIETARY DATA FOR ALL THE 139 SPECIES

IN PINEDA-MUNOZ AND ALROY (2014). RAW DATA ARE SQUARE-ROOT

TRANSFORMED PERCENTAGES OF FOOD ITEMS BASED ON STOMACH CONTENT

EXAMINATIONS. SHADED AREAS REPRESENT 99% CONFIDENCE ELLIPSES AROUND

THOSE SPECIES THAT CONSUME 50% OR MORE OF A PARTICULAR DIETARY CATEGORY

AS SHOWN IN THE LEGEND. (B–F) PROJECTED DIET TIME SERIES INTO THE PCA SPACE

IN FIGURE SECTION A FOR DIFFERENT DIETARY STUDIES: (B) DIET OF SHASTA

GROUND SLOTH (NOTHROTHERIOPS SHASTENSIS) FROM DUNG DEPOSITS IN RAMPART

CAVE, ARIZONA (HANSEN 1978). DAILY (C) AND WEEKLY (D) DIET OF CHIMPANZEES

(PAN TROGLODYTES) OBSERVED IN KIGALI NATIONAL PARK, UGANDA BY DAVID

WATTS (PERSONAL COMMUNICATION). MONTHLY (E) AND YEARLY (F) DIET OF

BROWN BEARS (URSUS ARCTOS) IN YELLOWSTONE NATIONAL PARK, WYOMING

BASED ON SCATS (MATTSON ET AL. 1991). AXES ARE IDENTICAL FOR ALL PLOTS

EXCEPT MONTHLY BEAR DIET (E). THE MULTI-COLOURED LINES (B–F) SHOW WHERE

SPECIES ARE IN THE DIETARY SPACE AT SAMPLED TIMES...... 86

FIGURE 5.3-2: DIAGRAMMATIC EXAMPLES OF HOW MISMATCHING TEMPORAL GRAIN (A)

AND EXTENT (B) BETWEEN QUESTIONS AND PROXIES CAN LEAD TO ERRONEOUS

INFERENCES. TEMPORAL GRAIN EXAMPLE (A) BASED ON (REYNOLDS-HOGLAND &

MITCHELL 2007). TEMPORAL EXTENT EXAMPLE (B) BASED ON (MARTÍNEZ DEL RIO

ET AL. 2009). SEE FIG. S6 FOR SIMILAR PALEOECOLOGICAL EXAMPLES...... 92

FIGURE 6.2-1: DENTAL AND DIETARY DIVERSITY AND CONVERGENCE IN MARSUPIALS AND

PLACENTALS. THREE-DIMENSIONAL OCCLUSAL RECONSTRUCTIONS OF TWO

MARSUPIAL UPPER TOOTH ROWS (TOP LEFT, COMMON BRUSHTAIL POSSUM,

TRICHOSURUS VULPECULA; BOTTOM LEFT, STRIPPED POSSUM, DACTYLOPSIA

TRIVIRGATA) AND TWO PLACENTAL UPPER TOOTH ROWS (TOP RIGHT, WESTERN xxii

GORILLA, GORILLA; BOTTOM RIGHT, SERRA DO MAR GRASS MOUSE,

AKODON SERRENSIS). SURFACE ORIENTATION (AS INDICATED IN EACH COLOUR

WHEEL) AND SLOPE VALUES (GREEN-YELLOW-RED GRADIENT FROM 0 TO 90º AS IN

SLOPE CHART) ARE TWO OF THE VARIABLES OBTAINED FOR EACH SPECIMEN AND ARE

COMPARED BETWEEN HERBIVOROUS AND INSECTIVOROUS DIETS. EACH DIET

CATEGORY HAS ITS OWN SLOPE CHART. UPPER LEFT TOOTH ROWS FOR ALL SPECIES

EXCEPT AKODON SERRENSIS, WHICH IS A REFLECTED UPPER RIGHT TOOTH ROW;

ANTERIOR TOWARDS THE LEFT. SCALE BARS, 1 MM...... 100

FIGURE 6.2-2: DIVERSITY IN DENTAL MORPHOLOGY USING MPDMA. PLOTS OF THE TWO

FIRST COMPONENTS OF THE PRINCIPAL COMPONENT ANALYSIS OF MPDMA DATA

FOR 83 PLACENTAL AND MARSUPIAL SPECIES (A-B) AND FOR 53 SPECIES INCLUDING

ONLY DIETARY SPECIALIZATIONS COMMON TO BOTH MARSUPIALS (OPEN CIRCLES)

AND PLACENTALS (CLOSED CIRCLES) (C-D). PLOTS ILLUSTRATE DIFFERENCES IN

PRIMARY DIET AND PHYLOGENY (A, C); AND PHYLOGENY AT THE INFRAORDER

LEVEL (PLACENTALS VS. MARSUPIALS) (B, D). POLYGON COLOURS AS SHOWN IN

LEGENDS...... 103

FIGURE 7.4-1: BOXPLOT ILLUSTRATING THE RANGE OF OVERALL OPCR FOR EACH

DIETARY CATEGORY FOR LOWER AND UPPER TOOTH ROWS OF THE 138 MAMMAL

SPECIES IN THE DATASET. DIET CATEGORIES ARE AS DESCRIBED BY PINEDA-MUNOZ

AND ALROY (2014)...... 115

FIGURE 7.4-2: BOXPLOT ILLUSTRATING THE RANGE OF AVERAGE OPCR BETWEEN

INDIVIDUAL TEETH FOR EACH DIETARY CATEGORY FOR LOWER AND UPPER TOOTH

ROWS OF THE 138 MAMMAL SPECIES IN THE DATASET. DIET CATEGORIES ARE AS

DESCRIBED BY PINEDA-MUNOZ AND ALROY (2014)...... 116

FIGURE 7.4-3: BOXPLOT ILLUSTRATING THE RANGE OF STANDARD DEVIATION OF OPCR

BETWEEN INDIVIDUAL TEETH FOR EACH DIETARY CATEGORY FOR LOWER AND UPPER

xxiii

TOOTH ROWS OF THE 138 MAMMAL SPECIES IN THE DATASET. DIET CATEGORIES ARE

AS DESCRIBED BY PINEDA-MUNOZ AND ALROY (2014)...... 119

FIGURE 7.4-4: BOXPLOT ILLUSTRATING THE RANGE OF RELIEF INDEX FOR EACH DIETARY

CATEGORY FOR LOWER AND UPPER TOOTH ROWS OF THE 138 MAMMAL SPECIES IN

THE DATASET. DIET CATEGORIES ARE AS DESCRIBED BY PINEDA-MUNOZ AND ALROY

(2014)...... 120

FIGURE 7.4-5: BOXPLOT ILLUSTRATING THE RANGE OF AVERAGE SLOPE FOR EACH

DIETARY CATEGORY FOR LOWER AND UPPER TOOTH ROWS OF THE 138 MAMMAL

SPECIES IN THE DATASET. DIET CATEGORIES ARE AS DESCRIBED BY PINEDA-MUNOZ

AND ALROY (2014)...... 122

FIGURE 7.4-6: BOXPLOT ILLUSTRATING THE RANGE OF STANDARD DEVIATION OF THE

SLOPE FOR EACH DIETARY CATEGORY FOR LOWER AND UPPER TOOTH ROWS OF THE

138 MAMMAL SPECIES IN THE DATASET. DIET CATEGORIES ARE AS DESCRIBED BY

PINEDA-MUNOZ AND ALROY (2014)...... 123

FIGURE 7.4-7: DIVERSITY IN DENTAL MORPHOLOGY USING MPDMA. PLOTS OF THE TWO

FIRST COMPONENTS OF THE PRINCIPAL COMPONENT ANALYSIS OF MPDMA DATA

FOR 83 PLACENTAL AND MARSUPIAL SPECIES (A-B) AND FOR 53 SPECIES INCLUDING

ONLY DIETARY SPECIALIZATIONS COMMON TO BOTH MARSUPIALS (OPEN CIRCLES)

AND PLACENTALS (CLOSED CIRCLES) (C-D). PLOTS ILLUSTRATE DIFFERENCES IN

PRIMARY DIET AND PHYLOGENY (A, C); AND PHYLOGENY AT THE INFRAORDER

LEVEL (PLACENTALS VS. MARSUPIALS) (B, D). POLYGON COLOURS AS SHOWN IN

LEGENDS...... 126

FIGURE 7.4-8: ...... 127

FIGURE 7.7-1: HYPOTHETICAL FAUNAL TURNOVER EVENT. (A) PRE-PERTURBATION

MAMMAL ASSEMBLAGE. (B) AND (C) TWO HYPOTHETICAL POST-PERTURBATION

MAMMAL ASSEMBLAGES. EACH ELLIPSE COVERS THE ECOMORPHOLOGICAL xxiv

VARIABILITY OF A TAXON. OVERLAPPING ELLIPSES REPRESENT TAXA WITH THE SAME

MORPHOLOGICAL ADAPTATIONS...... 140

FIGURE 7.7-2: HYPOTHETICAL FAUNAL TURNOVER EVENT. A) PRE-PERTURBATION

MAMMAL ASSEMBLAGE. B AND C) TWO HYPOTHETICAL POST-PERTURBATION

MAMMAL ASSEMBLAGES. EACH ELLIPSE COVERS THE ECOMORPHOLOGICAL

VARIABILITY OF A TAXON. OVERLAPPING ELLIPSES REPRESENT TAXA WITH THE SAME

MORPHOLOGICAL ADAPTATIONS...... 140

FIGURE 11.2-1: CORRELATION BETWEEN THE PERCENTAGE OF FRUIT IN THE DIET OF THE

FRUGIVORE SPECIES IN PINEDA-MUNOZ AND ALROY (2014) AND THE VALUES FOR

THE SAME SPECIES PROVIDED BY WILMAN ET AL. (2014). CORRELATION IN THE TOP-

RIGHT CORNER OF THE GRAPH...... 187

FIGURE 11.2-2: BOXPLOT ILLUSTRATING THE BODY MASS RANGE FOR EACH DIETARY

CATEGORY IN WILMAN ET AL. (2014) DATA. DIET CATEGORIES ARE AS DESCRIBED BY

PINEDA-MUNOZ AND ALROY (2014)...... 187

FIGURE 11.2-3: SCORES OF THE TWO FIRST AXES OF THE PRINCIPAL COMPONENT ANALYSIS

OF DIETARY DATA FOR THE SPECIES IN WILMAN ET AL. (2014). RAW DATA ARE

PERCENTAGE OF FOOD RESOURCE RESCALED TO Z-SCORES. ARROWS INDICATE

LOADINGS ON THE TWO FIRST COMPONENTS. COLOURS AND SIZES INDICATE THE

LOG10 BODY MASS (G) OF EACH SPECIES AS SHOWN IN THE LEGEND...... 188

FIGURE 11.2-4: RELATIONSHIP BETWEEN LOG10 BODY MASS (G) AND DIET DIVERSITY

INDICES CALCULATED BY APPLYING INVERSE SIMPSON INDICES TO STOMACH

CONTENT PERCENTAGES FOR WILMAN ET AL. (2014) DATA. COLOURS REPRESENT

DIFFERENT DIETARY SPECIALIZATIONS...... 188

FIGURE 11.3-1: DIET OF CHIMPANZEES (PAN TROGLODYTES) OBSERVED IN KIGALI

NATIONAL PARK, UGANDA FROM APRIL 7 TO JULY 28, 2013 BY DAVID WATTS

(PERSONAL COMMUNICATION) PROJECTED INTO THE PCA MULTIDIMENSIONAL

xxv

DIETARY SPACE OF ELTONTRAITS 1.0 (WILMAN ET AL. 2014). COMPONENTS 1 AND 2

ACCOUNT FOR 20% AND 13% OF DIETARY VARIANCE, RESPECTIVELY. SHADED

AREAS REPRESENT THE 99% CONFIDENCE ELLIPSES AROUND THOSE SPECIES THAT

CONSUME 50% OR MORE OF A PARTICULAR DIETARY CATEGORY. SEE WILMAN ET AL.

(2014) FOR DEFINI- TIONS OF DIETARY CATEGORIES USED. THE MULTICOLORED LINE

SHOWS WHERE CHIMPANZEES ARE IN DIETARY SPACE AT SAMPLED TIMES. (A) DAILY

DIET AND (B) SEVEN DAY AVERAGE DIET...... 195

FIGURE 11.3-2: DIET OF BROWN BEARS (URSUS ARCTOS) IN YELLOWSTONE NATIONAL

PARK, WYOMING FROM 1977-1987 OBTAINED BY OPPORTUNISTIC SCAT SAMPLING

(MATTSON ET AL. 1991). (A) MONTHLY AVERAGE DIET AND (B) YEARLY AVERAGE

DIET. FIGURE FOLLOWS CONVENTIONS OF FIGURE 11.3-1...... 196

FIGURE 11.3-3: DIET OF SHASTA GROUND SLOTH (NOTHROTHERIOPS SHASTENSIS)

OBTAINED FROM DUNG DEPOSITS IN RAMPART CAVE, ARIZONA (HANSEN 1978).

FIGURE FOLLOWS CONVENTIONS OF FIGURE 11.3-1...... 197

FIGURE 11.3-4: INTRASPECIFIC TEMPORAL VARIATION IN DIET CAN BE LARGE. FIGURE

SHOWS EUCLIDEAN DISTANCES BETWEEN SPECIES’ UNTRANSFORMED PERCENTAGE

DIETS USING THE CLASSIFICATION AND DATA IN ELTONTRAITS 1.0 (WILMAN ET AL.

2014) AS THE BLACK BOXPLOT. COLORED BOXPLOTS REPRESENT THE DISTRIBUTION

OF INTRASPECIFIC DIETARY DISTANCES BETWEEN ALL SAMPLED POINTS IN EACH

TEMPORAL DATASET. COLORED CIRCLES REPRESENT ONLY THOSE DISTANCES TAKEN

BETWEEN CONSECUTIVE TIMEPOINTS LIKE DAY 1 TO DAY 2, JUNE TO JULY, ETC. THE

DIETARY DISTANCE BETWEEN ELEPHANTS AND OTHER SPECIES IN ELTONTRAITS 1.0

(WILMAN ET AL. 2014) IS GIVEN FOR REFERENCE...... 198

FIGURE 11.3-5: INTRASPECIFIC TEMPORAL VARIATION IN DIET CAN BE LARGE. FIGURE

SHOWS EUCLIDEAN DISTANCES BETWEEN SPECIES’ UNTRANSFORMED PERCENTAGE

DIETS USING THE CLASSIFICATION AND DATA IN PINEDA-MUNOZ AND ALROY (2014) xxvi

AS THE BLACK BOXPLOT. COLORED BOXPLOTS REPRESENT THE DISTRIBUTION OF

INTRASPECIFIC DIETARY DISTANCES BETWEEN ALL SAMPLED POINTS IN EACH

TEMPORAL DATASET. COLORED CIRCLES REPRESENT ONLY THOSE DISTANCES TAKEN

BETWEEN CONSECUTIVE TIMEPOINTS LIKE DAY 1 TO DAY 2, JUNE TO JULY, ETC. THE

DIETARY DISTANCE BETWEEN ELEPHANTS AND OTHER SPECIES IN PINEDA-MUNOZ

AND ALROY (2014) IS GIVEN FOR REFERENCE. BECAUSE ELTONTRAITS 1.0 (WILMAN

ET AL. 2014) HAS A DIFFERENT CLASSIFICATION SCHEME AND USES QUANTITATIVE

ESTIMATES OF DIET BASED ON QUALITATIVE DATA, THE DISTANCES BETWEEN

ELEPHANTS AND OTHER REFERENCE SPECIES IN FIGURE 11.4-4 ARE DIFFERENT THAN

THOSE IN THIS FIGURE, WHICH ARE BASED ON PINEDA-MUNOZ AND ALROY (2014).

...... 199

FIGURE 11.3-6: DIAGRAMMATIC EXAMPLES OF HOW MISMATCHING TEMPORAL GRAIN (A)

AND EXTENT (B) BETWEEN QUESTIONS AND PROXIES CAN LEAD TO ERRONEOUS

INFERENCES IN A PALEOECOLOGICAL CONTEXT. TEMPORAL GRAIN EXAMPLE (A)

BASED ON (FAITH 2011). TEMPORAL EXTENT EXAMPLE (B) BASED ON (EMSLIE &

PATTERSON 2007)...... 200

FIGURE 11.4-1: PLACENTALS AND MARSUPIALS OVERLAP IN DENTAL ECOMORPHOSPACE.

PLOT OF THE TWO FIRST COMPONENTS OF THE PRINCIPAL COMPONENT ANALYSIS OF

MPDMA DATA FOR 138 PLACENTAL AND MARSUPIAL SPECIMENS INCLUDING

RODENTS. POINTS ARE PLOTTED ACCORDING TO PRIMARY DIET (A); PRIMARY DIET

AND PHYLOGENY (B); AND PHYLOGENY AT THE INFRACLASS LEVEL (PLACENTALS VS.

MARSUPIALS) (C). COLOURS AND POLYGONS ARE AS SHOWN IN LEGENDS. DATA

WERE SQUARE-ROOT TRANSFORMED BEFORE PCA...... 212

FIGURE 11.4-2: DIETARY ECOMORPHOSPACE OF THE ORDER RODENTIA. PLOT OF THE TWO

FIRST COMPONENTS OF THE PRINCIPAL COMPONENT ANALYSIS OF MPDMA DATA

FOR 64 RODENT SPECIMENS. POINTS ARE PLOTTED ACCORDING TO PRIMARY DIET.

xxvii

COLOURS OF THE POLYGONS ARE AS SHOWN IN LEGENDS. DATA WERE SQUARE-ROOT

TRANSFORMED BEFORE PCA...... 213

FIGURE 11.4-3: DIETARY ECOMORPHOSPACE OF THE ORDER CARNIVORA. PLOT OF THE

TWO FIRST COMPONENTS OF THE PRINCIPAL COMPONENT ANALYSIS OF MPDMA

DATA FOR 32 CARNIVORAN SPECIMENS. POINTS ARE PLOTTED ACCORDING TO

PRIMARY DIET. COLOURS OF THE POLYGONS ARE AS SHOWN IN LEGENDS. DATA WERE

SQUARE-ROOT TRANSFORMED BEFORE PCA...... 214

FIGURE 11.4-4: DIETARY ECOMORPHOSPACE OF THE ORDER DIPROTODONTIA. PLOT OF

THE TWO FIRST COMPONENTS OF THE PRINCIPAL COMPONENT ANALYSIS OF MPDMA

DATA FOR 32 DIPROTODONTID SPECIMENS. POINTS ARE PLOTTED ACCORDING TO

PRIMARY DIET. COLOURS OF THE POLYGONS ARE AS SHOWN IN LEGENDS. DATA WERE

SQUARE-ROOT TRANSFORMED BEFORE PCA...... 215

FIGURE 11.4-5: DIETARY ECOMORPHOSPACES OF THE ORDER PRIMATES. PLOT OF THE

TWO FIRST COMPONENTS OF THE PRINCIPAL COMPONENT ANALYSIS OF MPDMA

DATA FOR 19 SPECIMENS. POINTS ARE PLOTTED ACCORDING TO PRIMARY

DIET. COLOURS OF THE POLYGONS ARE AS SHOWN IN LEGENDS. DATA WERE SQUARE-

ROOT TRANSFORMED BEFORE PCA...... 216

xxviii

xxix

1 PROLOGUE

PhD thesis projects can become impersonal. We can write about what we found and

how we did everything and then outline some conclusions, and although it will look

finished, it won't really be complete.

I still remember one of my 7th grade teachers. He came into the class and gave us a 40-

minute talk about how to “find the X”. I felt nothing but confused. Yes, everything he

said was logical and made a lot of sense, but when he asked us if we had a question, I

raised my hand: “I don’t understand it”. He attempted to explain me the same thing

again, but I stopped him: “Yes, I get it, I get the mechanism. What I don’t understand is

why are you telling us about it, what is it for?”. And so I will begin by explaining why I

carried out my thesis. We shouldn’t do science in order to get a PhD: a PhD should be

another step toward learning how to do science.

I was always curious about biology. My family was a really outdoorsy and I spent most

of my childhood hiking in the Pyrenees. I started watching documentaries and

reading books about animals at a really early age. The more I read, the more questions

popped into my head. As clichéd as this might sound, it was a documentary about

Darwin that led the way to being interested in evolutionary biology and .

xxx

After I finished watching it, I sat in silence for a while before approaching my mother:

“Tell dad I now know the answer, the egg was first!”.

I was finishing my master’s thesis when I found out that John Alroy was looking for

PhD students. He asked me to write a short research proposal. I had to be honest with myself. I was fighting for something I might end up doing for four years, so it had better be something I loved.

I put all my ideas down on a piece of paper, all the questions my curiosity was driving me towards. My biggest interest was to be able to draw, describe, and understand all sorts of : worldwide, and at any time in the past. And I wrote about it. I can’t say this thesis answers exactly what I wrote in that proposal. It was more of a life-time project than anything feasible in 4 years time. Because before solving the bigger picture,

I needed to solve the little puzzles I encountered on the way and make sure this work formed the basis for my future research.

This is what my thesis is about: creating a foundation for my life-time science project. I think a thesis shouldn’t be a one-book story with a definite ending, but the beginning of a researcher’s series.

“Never stop questioning, curiosity has its own reasons for existing.”

– Albert Einstein

xxxi Diet, ecology, and dental morphology in terrestrial mammals

2 INTRODUCTION

Paleoecological reconstructions are as old as fossil discoveries. Naturalists and

philosophers from ancient times did not only write about fossils, but also provided their

own environmental and ecological interpretations. As proposed by Xenophanes of

Colophon (570-480 BC), finding a marine organism in a terrestrial outcrop could mean

that land was once covered in water (McKirahan 2011); a petrified bamboo fossil forest

in a rather dry area in China could be evidence of climate change away from more

humid environments, suggested Shen Kuo (1031–1095) (Chan et al. 2001).

Indeed, the fossil record provides a unique opportunity to understand life on Earth and

the biotic and abiotic processes that have led to present ecosystems (Kitchell 1985). The

ultimate goal for much of paleontological research is describing past ecosystems.

Ideally, one would do so in sufficient detail to imagine travelling back in time and

performing the same experiments we carry out in modern environments. Although we

can only obtain limited data from the fossil record, applying the knowledge of many

disciplines— and specifically combining what we know about life on Earth at the

present time— can bring us closer to the real picture of the past. Stratigraphy and

32 Silvia Pineda-Munoz - April 2016

Chapter 2: Introduction

sedimentology provide accurate environmental reconstructions, while taphonomy informs us about the processes organisms went through before we collected them as fossils (Behrensmeyer and Hill 1988). Similarly, diversity curves informs about the taxonomic composition of the ecosystems and can trace out patterns of radiation and extinction (Sepkoski Jr 1984). Additionally, reconstructing the trophic relationships between organisms in an ecosystem can provide direct information about its functioning: from structure to and environmental conditions (Dalerum and Angerbjörn 2005, Feranec 2004).

The main goal of my Ph.D. thesis was to design a quantitative method that could be applied to dietary reconstruction as well as to the study of macroevolutionary patterns in mammalian ecosystems. In order to evaluate the reliability of such a method it is necessary to test it on extant species whose diets are already known. However, obtaining data from modern ecosystems is not a straightforward task. The literature can provide information that is far from methodologically consistent. Many different approaches can be used, from making visual observations to analysing scat contents.

However, sampling error can be very high (Dickman and Huang 1988, Kohfeld and

Harrison 2000, McInnis et al. 1983, Shrestha and Wegge 2006). Visual observations will not provide volumetric values of the foods being consumed, while digestive processes will differentially affect the items found in scats. Additionally, some studies provide a highly detailed list of consumed foods down to the species level while others give broad categories that might not be consistent across studies.

Chapter 3 addresses all of these questions. After analysing stomach content data from the literature for 139 terrestrial mammals, I propose a statistically grounded dietary

Silvia Pineda-Munoz - April 2016 33

Diet, ecology, and dental morphology in terrestrial mammals

classification that accurately predicts major feeding specializations. Previous dietary classifications led to important losses of information, as is also discussed in the chapter.

Body mass is one of the most-studied morphological variables in the ecological and paleoecological literature. It is relatively easy to estimate and infer from modern and fossil populations and it has historically provided important information about ecosystem dynamics and macroevolution (Alroy 1998, Burness et al. 2001, Smith et al.

2010, Smith and Lyons 2011). The dietary classification proposed in Chapter 3 was based on what can be found in the literature. However, it needed to be contrasted with other ecological variables in order to test whether it was consistent with ecomorphological specializations. Chapter 4 studies the relationship between diet and body mass in terrestrial mammals and suggests that the combination of these two variables can help to make strong paleoclimatological inferences.

During the last few decades, palaeontologists have proposed many different proxies for inferring diet in fossil species: stable isotope values, dental microwear, general morphology, and so forth (Ungar 2010). However, most studies arbitrarily choose a methodology without considering either methodological limitations or the bigger research questions that need to be answered. Each of the dietary proxies will reflect a different temporal scale and a particular time span of averaging, and therefore will be more suitable to answering only particular questions. For example, dental microwear informs us of the diet of an individual during the last three to four weeks of its life

(DeMiguel et al. 2011). Therefore, an animal living in a seasonal environment, switching between diets through the year, will display a different microwear pattern depending on the season of death. Dental morphology however will not inform us of seasonality, but will be a better indicator of the evolutionary pressures bearing on

34 Silvia Pineda-Munoz - April 2016

Chapter 2: Introduction

dietary specialization (Polly 2015). Chapter 5 evaluates the limitations and temporal scale of dietary proxies used by palaeontologists and neontologists.

The idea that dental morphology predicts diet is attributed to Georges Cuvier, who said

“Montrez-moi vos dents et je vous dirai qui vous êtes”, meaning, “show me your teeth and I’ll tell you who you are” (Cuvier 1825). Nearly two centuries later, Andrews et al.

(1979) used ecological diversity indices to study the relationship between community structure and in modern and fossil mammalian localities. With respect to the criteria for classifying mammals according to their feeding specializations they stated:

“The feeding classes […] were chosen mainly on the basis of primary dietary adaptation as indicated by tooth morphology, this being the evidence available from fossils”.

Although dental morphology has been traditionally used to infer diet in the fossil record, it was only during the last decades that researchers started approaching this topic from a more quantitative point of view. The idea behind using -free dental morphology methods as dietary predictors is that diet drives the evolution of dental morphology towards functional optima. Thus, distantly related mammal orders with the same dietary specializations should display similar morphological features on their tooth surfaces. Chapter 6 studies this hypothesis, suggesting that there is phenotypic convergence evolution across placental and marsupial dentitions, the two clades having split during the Late Jurassic or Early . For this study, I designed a new methodology called multi-proxy dental morphology analysis (MPDMA). The method's innovation is that combines different metrics obtained from three-dimensional scans of mammal dentitions in order to approximate the observed multidimensionality of diet specialization (Chapter 3). In Chapter 7 I evaluate the effectiveness of individual dental

Silvia Pineda-Munoz - April 2016 35

Diet, ecology, and dental morphology in terrestrial mammals

morphology metrics as dietary predictors in contrast with the improved discrimination power of MPDMA.

Together, these five chapters provide a new perspective on the study of mammalian diet in past and present ecosystems. Additionally, they settle the basis of my future research in paleoecology.

36 Silvia Pineda-Munoz - April 2016

Chapter 2: Introduction

Silvia Pineda-Munoz - April 2016 37

Diet, ecology, and dental morphology in terrestrial mammals

3 DIETARY CLASSIFICATION OF TERRESTRIAL MAMMALS

Published in Proceeding of the Royal Society B as: Pineda-Munoz, S., and

J. Alroy. 2014. Dietary characterization of terrestrial mammals.

Proceedings of the Royal Society B: Biological Sciences 281(1789).

38 Silvia Pineda-Munoz - April 2016

Chapter 3: Dietary classification of terrestrial mammals

3.1 Abstract Understanding the feeding behaviour of the species that make up any ecosystem is essential for designing further research. Mammals have been studied intensively, but the criteria used for classifying their diets are far from being standardized. We built a database summarizing the dietary preferences of terrestrial mammals using published data regarding their stomach contents. We performed multivariate analyses in order to set up a standardized classification scheme. Ideally, food consumption percentages should be used instead of qualitative classifications. However, when highly detailed information is not available we propose classifying animals based on their main feeding resources. They should be classified as generalists when none of the feeding resources constitute over 50% of the diet. The term “omnivore” should be avoided because it does not communicate all the complexity inherent to food choice. Moreover, so-called omnivore diets actually involve several distinctive adaptations. Our dataset shows that terrestrial mammals are generally highly specialized and that some degree of food mixing may even be required for most species.

Keywords: Mammal ecology, mammal paleoecology, dietary specialization, ecomorphology

3.2 Introduction Reconstructing the nutritional requirements of an ecosystem's components is essential for understanding its function and for designing further biological studies. Accurate information about feeding behaviour is also a pre-requisite for research on ecomorphology; making inferences about the fossil record (Evans et al. 2007, Palmqvist et al. 2003, Ungar 2010); inferring the climate and ecological context of fossil localities,

Silvia Pineda-Munoz - April 2016 39

Diet, ecology, and dental morphology in terrestrial mammals

which enables tracking global climate change throughout the geological time scale

(Andrews et al. 1979, Fortelius et al. 2002b, Liu et al. 2012); and modelling food webs

(Kondoh 2003, Petchey et al. 2008, Pimm and Lawton 1977, 1978).

Contemporary mammals are extraordinarily diverse, having adapted to fill almost all available ecological niches (Eisenberg 1981, Ungar 2010, Wilson and Reeder 2005).

They also play a key role in the dynamics of the ecosystems in which they live (Jones and Safi 2011, Sinclair 2003). Therefore, it is crucial to reconstruct the trophic relationships between mammals and the other components of their ecological communities (Kronfeld and Dayan 1998). Many researchers have attempted to analyse the morphology, ecology and paleoecology of mammals based on their dietary preferences (e.g. (Andrews et al. 1979, Mendoza et al. 2005, Reed 1998)). Some have used basic feeding classifications equating with classic trophic levels – , , and , plus a few variations (Evans et al. 2007, Reed 1998,

Schoener 1989, Ungar 2010). Others distinguished multiple food resources within diets

(Andrews and Evans 1979, Mendoza et al. 2005, Palmqvist et al. 2003). However, none of the previous workers performed multivariate analyses to support their dietary categorizations. This situation hinders comparing results from different ecomorphological studies.

Here we offer a new classification scheme of the feeding preferences of modern mammals. This scheme is intended to provide a standardized foundation for research concerning ecomorphology and the functioning of global terrestrial ecosystems.

40 Silvia Pineda-Munoz - April 2016

Chapter 3: Dietary classification of terrestrial mammals

3.3 Material and Methods We built a database summarizing the dietary preferences of terrestrial mammals using published data. Data were compiled from primary resources (see Table 11.1-1) that were identified using academic search engines and databases. Quantitative information on diet can be generated using any of three major sampling methodologies: (1) time spent each food resource; (2) faeces composition; and (3) stomach contents. The analysis of time budgets is far from replicable as it is subject to observer error and personal biases (McInnis et al. 1983, Shrestha and Wegge 2006). Moreover, food resources have variable processing and/or acquisition times that are not necessarily proportional to the volume of food consumed, and it is impossible to sample some mammal species in this way (Shrestha and Wegge 2006). Faces contents are easier to evaluate because less time is required and there is no direct interaction with the animals

(Balestrieri et al. 2011, Dickman and Huang 1988). However, different kinds of food have different responses to digestive processes and these responses are highly variable among species and individuals. Thus, these problems may cause important sampling biases (Dickman and Huang 1988, Kohfeld and Harrison 2000). Analysis of stomach/gut contents usually demands sacrificing a large number of animals, which makes it difficult to work with threatened species. Thus, fewer studies use this sampling methodology (Balestrieri et al. 2011, Dickman and Huang 1988, Kohfeld and Harrison

2000). However, this method provides direct information, with all ingested food potentially being found and degradation from digestive processes being minimal

(Balestrieri et al. 2011, Kohfeld and Harrison 2000).

We define “diet” as the average variety of food ingested over the entire lifetime of the individuals of a species. However, data of this exact kind do not exist because

Silvia Pineda-Munoz - April 2016 41

Diet, ecology, and dental morphology in terrestrial mammals

observations of feeding behaviour over short intervals, of scats, or of gut contents only apply to short-term diets of a few individuals in one or a few places. We believe this problem is not insurmountable because relatively minor variation is seen within particular data sets and because we have taken steps in order to compile a dataset as close to our dietary definition as possible.

Because we were able to retrieve a significant amount of data of volumetric percentages of stomach contents we restricted our analysis to data of this nature. Unfortunately, some of the references did not report the sample size of the analysed species. We nonetheless included all collected data in our analysis because we felt that sample sizes were likely to be reasonable regardless of whether they were reported. Additionally, we utilized gross averages for each of the species independently of the season and locality in order to approximate the average diet for the species. Literature reported the average across all seasons for 16% of the species in the dataset; seasonal breakdowns were available for 31% of the data records, and in these cases we calculated the average across all seasons; 53% of the species lacked any data regarding seasonal changes in diet, which meant that only a single record could be employed. There was information for 8% of the species regarding multiple localities, and in those cases we similarly used the average across all the localities. We ultimately found data on percentage stomach content volume for 139 mammalian species (Table 11.1-1).

We used the taxonomic classification of Wilson and Reeder (Wilson and Reeder 2005) to standardize the nomenclature. Classifying food resources was a complex task. Some authors reported the taxonomic attribution of the consumed items within the stomach contents, while others classified them within broad categories. Moreover, resource classifications varied significantly among authors. For example, some distinguished

42 Silvia Pineda-Munoz - April 2016

Chapter 3: Dietary classification of terrestrial mammals

between grass, forbs, leaves and branches, while others included all of the above in a single food category (vegetation). Previous literature focused on classifying herbivore diets in order to infer and behaviours (Gagnon and Chew 2000).

However, this kind of information was generally not available in the stomach content data sets we recovered. Because we were focusing on the diet of whole terrestrial mammalian faunas and wanted to maximize the extent of our survey, we therefore felt comfortable with not breaking down herbivores into browsers and grazers.

We established eight feeding resource groups easily recognizable in all terrestrial ecosystems: seeds, invertebrates, vertebrates, , flowers and gum, roots and tubers, green plants and fruit (Table 11.1-2 gives details). We excluded values pertaining to miscellaneous undefined food resources to make the data comparable among species.

We used the R environment (Ihaka and Gentleman 1996) for statistical analysis and construction of tables and figures (packages “psych” and “stats”). We carried out principal components analysis (PCA) to identify the variables that best summarize the different dietary specializations. Factor analysis was also applied but is not discussed further because it provided similar results. Unfortunately, raw percentage values are non-independent because they must add up to 100, which make them computationally unsuitable for PCA. More importantly, PCA tends to underweight variables if the percentages are consistently low because the lower bound of zero compresses potential variance. This property masks the contribution of rare dietary preferences. We solved the problems of non-independence and underweighting by rescaling each variable as a z-score before carrying out the PCA. We then undertook UPGMA cluster analysis based on Euclidean distances among the PCA scores. Compared to more arbitrary manipulations such as taking square roots of the percentages, the z-score transformation

Silvia Pineda-Munoz - April 2016 43

Diet, ecology, and dental morphology in terrestrial mammals

had no qualitative impact on the PCA and UPGMA results apart from drawing more attention to five unusual species: Scapanus townsendii, which eats substantially more roots and tubers than any other species in the data set; Euoticus elegantulus, which is the only true gumnivore; and three heavily fungivorous sciurids. The clustering results were analysed visually in order to determine the best criteria for establishing quantitative feeding specializations. We performed K-means cluster analysis (Hartigan

1979) of the transformed percentages in order to test our classification scheme. Species were grouped into Linnean orders in order to evaluate the degree to which taxonomic attribution is related to diet.

We attempted to test previous feeding classifications based on classic trophic levels. We assigned each feeding resource to the categories of "plant" and "animal." Previous workers using this approach excluded fungi from their data sets (Evans et al. 2007,

Reed 1998, Schoener 1989, Ungar 2010). However, we recorded many species that consume a significant amount of fungi. Due to need for standardization, "plants" included any non-animal food resource and therefore fungi and . "Animals" included both vertebrates and invertebrates.

3.4 Results The first four components of the PCA together express 65.76% of the variance (Table

11.1-3). The first component (19.4% of the variance) mainly discriminates between animals that feed on invertebrates or roots and tubers and the ones feeding on green plants. The second component (17.53% of the variance) discriminates mainly between fruit-eaters and seed-eaters. Therefore, the two first components alone identify the four major feeding groups (Figure 3.4-1). The third component (15.19%) discriminates

44 Silvia Pineda-Munoz - April 2016

Chapter 3: Dietary classification of terrestrial mammals

between many of the food resources, with specially high values for fungi and flower- gum eaters. The -feeding variable loads strongly only on the fourth component (13.64% of the variance).

Figure 3.4-1: Scores of the two first components of the principal component analysis of dietary data for all the 139 species in the dataset. Raw data are z-scores transformed percentages of food items based on stomach content examinations. Arrows indicate the loadings of the two first components of the analysis. Symbols represent taxonomic orders. Crosses indicate orders each represented by less than three species.

The four first components correspond to five major groups of feeding resources: green plants, fruit, seeds, invertebrates, and vertebrates (Figure 3.4-1). It is reasonable to define a species as a pure generalist if none of its feeding resources make up more than

50% of its diet. Just 20 species (14% of the total) have a diet of this kind (Figure 3.4-2).

Silvia Pineda-Munoz - April 2016 45

Diet, ecology, and dental morphology in terrestrial mammals

We might expect a more widespread feeding spectrum or less differentiated feeding groups if most or at least many of the species were truly generalists.

Figure 3.4-2 presents a cluster analysis of the 139 species on the dataset. Most of the feeding classifications form consistent and homogeneous clusters. In particular, seven discrete clusters can be differentiated easily. We interpret them as the main dietary specializations: carnivory, insectivory, fungivory, herbivory, granivory, gumivory, and frugivory. The first cluster separates an species with a high amount of roots and tubers on its diet, the talpid Scapanus townsendii. The second cluster identifies the only truly gumnivore species, the primate Euoticus elegantulus. The third cluster represents carnivores. Then, a series of minor clusters containing other mammals with peculiar diets separates before the major feeding cluster becomes evident. This is a result of the z-score transformation of percentage data, which draws more attention to unusual diet specializations.

Cluster groupings strongly correlate with the results of the PCA, the most readily apparent clusters being the ones formed by carnivores, , herbivores, granivores and . Most mammalian species appear to choose between one of these five feeding strategies even if they complement their diet with other resources.

The fungi and flowers-gum dietary specializations could be also defined as a feeding category, but they are represented by very few species and present in few environments.

Examples include the chipmunks and Euoticus elegantulus.

Special attention must be paid to branches marked with an asterisk (Figure 3.4-2).

Species on these branches are generalists (as defined above). The cluster analysis allocates them based on whatever makes up the plurality of the diet. Similarly, some green plant eaters have been classified within the seed-eater or invertebrate-eater

46 Silvia Pineda-Munoz - April 2016

Chapter 3: Dietary classification of terrestrial mammals

clusters despite having a greater amount of green plants in their diets. Cluster analysis allocates them into these groups because their diets are composed of the same resources, but in different proportions. This results in species consuming relatively low proportions of seeds or invertebrates being plotted into seed- or invertebrate-feeding clusters instead of the green plants cluster.

Table 3.5-1 represents the different feeding categories identified in Figure 3.4-2. The results of the K-means analysis were strongly correlated with the resultant branches of the cluster and with our classification scheme.

The orders Diprotodontia, Macroscelidea, Paucituberculata, Pholidota, Proboscidea,

Scadentia, and Xenanthra are each represented by less than three species and hence are all plotted using the same symbol (Figure 3.4-1). The artiodactyls in our dataset are divided between the fruit and green plant areas in Figure 3.4-1. Most of them belong to the family Bovidae, which consists almost entirely of herbivore and frugivore species

(Gagnon and Chew 2000). Other bunodont artiodactyl families that are traditionally described as omnivores (e.g., the Suidae) were not considered in the present analysis because stomach content data were unavailable. Carnivorans appear to be distributed across most of the feeding spectrum as has already been discussed by Evans et al.

(Evans et al. 2007). Eulipotyphla are all placed in the invertebrate region even though some of them complement their diet with other feeding resources. The talpid Scapanus townsendii can be identified at the top-left corner of the chart because of the high amount of roots and tubers in its diet. Primates are distributed between the fruit, invertebrate, and green plant resource areas. Rodentia is the most well-represented order in our database with dietary habits covering all major food resource regions. Moreover,

90% of the species in our dataset that fit into our description of generalists belong to the

Silvia Pineda-Munoz - April 2016 47

Diet, ecology, and dental morphology in terrestrial mammals

order Rodentia. The highly diverse feeding adaptations of rodents as well as their unspecialized dietary behaviour has been considered to be a primary factor driving their highly successful adaptive radiation (Landry 1970).

In summary, Figure 3.5-1 shows that classifying the diet of terrestrial mammals within trophic levels would lead to an important loss of information. Very few species have a diet based in a single food category (i.e., are pure herbivores or pure carnivores).

Furthermore, establishing simple boundaries between herbivores, omnivores and carnivores would present a challenging task.

48 Silvia Pineda-Munoz - April 2016

Chapter 3: Dietary classification of terrestrial mammals

Figure 3.4-2: UPGMA cluster analysis based on Euclidean distances of dietary data for the 139 species in the dataset. Raw data are as in Fig. 1. Numbers represent clusters identified in Table 1. * Branches containing mixed-feeding species (see text).

Silvia Pineda-Munoz - April 2016 49

Diet, ecology, and dental morphology in terrestrial mammals

3.5 Discussion

3.5.1 Proposed categories The main goal of the present study is to create a solid platform for research in ecology and paleoecology concerning dietary adaptations. Ideally, when detailed information is available one would want to perform analyses using food consumption percentages instead of qualitative classifications. In such cases one might opt for performing multivariate analyses such as those presented here (Figure 3.4-2 and 3.5-1). Although the analyses would be more complex, they would prevent an important loss of information. However, information regarding specific percentages of consumed food resources is simply unavailable or unclear for the vast majority of species. This problem is even more severe when it comes to fossil species because it is impossible to obtain precise information regarding their dietary habits.

Thus, we really have no choice but to develop a standardized classification scheme for mammalian diets based on what we can observe in real ecosystems. This explains why we have used the major branches of the clustering dendrogram (Figure 3.4-2) to establish a novel qualitative classification.

To cut to the chase, we propose classifying diets based on the most frequently consumed food resource. We suggest classifying a species as a dietary specialist if a single food resource makes up 50% or more of the diet. Feeding resources consumed with a frequency between 20% and 50% should be also used to categorize the diet. For example, a species consuming 70% green plants and 30% fruits should be classified as an herbivore-frugivore. We classify a species as a “generalist” when none of the feeding resource accounts for at least 50% of the diet.

50 Silvia Pineda-Munoz - April 2016

Chapter 3: Dietary classification of terrestrial mammals

Table 3.5-1: Feeding categories based on the new classification criteria and on the classic trophic level criteria (asterisks indicate those also identified by Eisenberg

(1989)).

Classic Main food Secondary food resource Cluster in trophic resource (20-50%) Figure 2 levels (>50%) classification Carnivore --- * 1 Carnivore Carnivore Frugivore 2 Omnivore Insectivore --- * 3 Carnivore Insectivore Carnivore 4 Carnivore Insectivore Granivore 5 Omnivore Insectivore Herbivore * 6 Omnivore Insectivore Fungivore 7 Omnivore Fungivore --- 8 Herbivore Fungivore Herbivore * 9 Herbivore Herbivore --- * 10 Herbivore Herbivore mixer 11 Herbivore Herbivore Frugivore * 12 Herbivore Herbivore Granivore 13 Herbivore Herbivore Insectivore 14 Omnivore Granivore --- 15 Herbivore Granivore Herbivore 16 Herbivore Granivore Insectivore 17 Carnivore Gumivore --- * 18 Herbivore Frugivore --- 19 Herbivore Frugivore Gumnivore * 20 Herbivore Frugivore Herbivore * 21 Herbivore Frugivore Insectivore 22 Omnivore Generalists star Omnivore

3.5.2 Earlier classification schemes Polis (Polis 1991) classified the terrestrial vertebrates of a sand community in the

Coachella Valley desert in order to assess trophic interactions in a real . He proposed a similar approach to ours in that he classified mammalian diets according to primary and secondary feeding categories. However, he considered any food resource

Silvia Pineda-Munoz - April 2016 51

Diet, ecology, and dental morphology in terrestrial mammals

that made up over the 10% of the diet to be a “primary resource”; while any other food resource was considered to be a “secondary resource” no matter how small the percentage. This classification worked with this particular community because the number of species and the number of interactions were both relatively limited.

However, it is hardly applicable in our dataset because there is so much more variation in diets, and as a result more feeding resources make up at least 10% of the diet in multiple species.

Figure 3.5-1: Dendrograms showing a) the proportions of plant and animal resources in each species' diet and b) feeding classifications based on classic trophic relationships (herbivore- omnivore-carnivore). The 139 species in the dataset are sorted by descending proportions of plant resources in their diets.

Some researchers have instead worked with feeding classifications based on classic trophic relationships (herbivores, carnivores and omnivores plus a few variations)

(Evans et al. 2007, Reed 1998, Schoener 1989, Ungar 2010). Figure 3.5-1 shows the

52 Silvia Pineda-Munoz - April 2016

Chapter 3: Dietary classification of terrestrial mammals

vagueness and oversimplification of this classification criterion. Based on this pattern it seems self-evident that we should avoid applying the three-way dietary classification whenever possible unless addressing broad trophic level questions. Doing so would not only lead to an overall loss of information but would specifically fail because species traditionally described as “omnivores” (see Table 3.5-1) display highly varied diets and are placed here in completely divergent clusters (Figure 3.4-2).

More realistic classifications should be based on the physical, nutritive, and ecological characteristics of food items. Some earlier researchers did distinguish multiple food resources for the purpose of establishing classifications (Andrews et al. 1979, Eisenberg

1981, Mendoza et al. 2005, Palmqvist et al. 2003). Among them, Eisenberg (Eisenberg

1981) presented the most similar classification to ours. He proposed 16 qualitative categories for mammals, some of which can be identified in our results (Table 3.5-1).

The most appreciable difference between his classification and ours involves the degree of resolution: he split up some of our feeding resources (i.e. green plants, invertebrates, and vertebrates) into subcategories in order to correlate dietary specializations with substrate utilization. He also overlooked some feeding specializations that we identify in our dataset (i.e. pure granivores or generalists). Finally, Eisenberg (Eisenberg 1981) described a group that mostly fed on fruits and seeds. This group of species can be identified in Figure 3.4-2 but they also feed on a considerable amount of green plants, so we have classified them as generalists.

Mendoza et al. (Mendoza et al. 2005) proposed an intermediate classification scheme for mammalian diets in order to evaluate ecological patterns in the trophic and body size structure of large mammal communities. Carnivores and herbivores were subclassified in more detailed categories that are suitable for studying highly dietary specialized

Silvia Pineda-Munoz - April 2016 53

Diet, ecology, and dental morphology in terrestrial mammals

mammalian orders such as carnivorans and artiodactyls. However, they included frugivores, , and gumnivores in a single feeding category and overlooked granivores. Thus, their feeding categories classified in detail some mammalian orders

(i.e. carnivorans and artiodactyls) while others were presumed to display no intra- ordinal diet variability (i.e. primates).

Andrews et al. (Andrews and Evans 1979) also differentiated between insectivores and carnivores and between frugivores and herbivores. In doing so, they also took into account differences in tooth morphology. However, their scheme is only incrementally more detailed than a traditional three-way trophic categorization.

3.5.3 Frequency of dietary specialization Our results suggest that terrestrial mammals are in general highly specialized. However, some degree of food mixing may be required for most mammals because only 23.7% of the records indicate that a single food resource constitutes over 90% of the diet. Singer and Bernays (Singer and Bernays 2003) evaluated this topic and suggested that nutritive and non-nutritive factors such as parasite and predator avoidance or ecological restrictions (e.g. or food web dynamics) may be responsible for food mixing. Single resource feeders are truly exceptional, and they are usually at the high end of the body mass spectrum. In our dataset only carnivores, insectivores, and herbivores display single resource diets, while frugivores, granivores, fungivores, and gumnivores seem to require a higher degree of food mixing. Schoener (Schoener 1989) has already noted that the main plant matter in omnivore diets consisted of fruits and seeds. Such diets require less physiological specialization than cellulose-rich herbivore diets.

54 Silvia Pineda-Munoz - April 2016

Chapter 3: Dietary classification of terrestrial mammals

3.5.4 Implications for research Standardized dietary classifications should serve as a cornerstone for research that aims to reconstruct past and present mammalian food webs. Here we propose an empirically tested scheme that has been erected only after quantitatively analysing dietary data drawn from real ecosystems. That said, our proposal should be considered a work in progress rather than a permanently fixed classification because the diets of so many species remain to be described in detail. In particular, our criteria do allow adding new feeding groups or even establishing more detailed subcategories when dealing with highly dietary specialized mammalian orders (e.g. artiodactyls or carnivorans).

Regardless of the details, however, continuing to work instead with untested dietary classifications would generate unstandardized results and thereby prevent synergy between intellectually overlapping ecological and paleoecological research programs.

3.6 Acknowledgments The senior author's work was supported by Macquarie University's HDR Project

Support Funds. The junior author is the recipient of an Australian Research Council

Future Fellowship (project number FT0992161).

3.7 Author’s contributions Both authors design the study. SP-M Collected and analysed the data, and drafted the manuscript. JA assisted data analysis and interpretation and draft the manuscript. All authors gave final approval for publication.

Silvia Pineda-Munoz - April 2016 55

Diet, ecology, and dental morphology in terrestrial mammals

4 THE RELATIONSHIP BETWEEN DIET AND BODY MASS IN TERRESTRIAL MAMMALS

Accepted in Paleobiology as: Pineda-Munoz S., A.R. Evans and J. Alroy. The relationship between diet and body mass in terrestrial mammals.

56 Silvia Pineda-Munoz - April 2016

Chapter 4: The relationship between diet and body mass in terrestrial mammals

4.1 Abstract Diet and body mass are highly important factors in mammalian ecology, and they have also have proven to be powerful paleoecological indicators. Our previous research has proposed a new classification scheme for mammals with more dietary divisions that emphasizes the primary resource in a given diet. We analysed a database summarizing the dietary preferences of 139 species of marsupial and placental terrestrial mammals

(including 14 orders) and their average body masses in order to explore whether this new classification better highlights ecomorphological differences between species.

Additionally, the dietary diversity of every species in the dataset was quantified by applying the inverse Simpson index to stomach content percentages. We observed a decrease in maximum dietary diversity with increasing body mass. Having lower requirements for energy and nutrients per unit of body weight or ecological advantages such as larger home ranges allows larger mammals to feed on less nutritive feeding resources (i.e., structural plant material). Our results also suggest that body size ranges are different across dietary specializations. Smaller mammals (<1 kg) are mainly insectivores, granivores, or mixed feeders while bigger animals (>30 kg) are usually either carnivores or herbivores that feed specifically on grasses and leaves. The medium size range (1-30 kg) is mostly composed of frugivorous species that inhabit tropical and subtropical rainforests. Thus, the near absence of medium-sized mammals in open environments such as savannahs can be linked to the decreasing density of fruit trees needed to support a pure frugivorous diet all around the year. In other words, seasonality of precipitation prevents species from specializing on a totally frugivorous diet. Our results suggest that this new classification scheme correlates well with body mass, one of the most studied morphological variables in paleoecology and

Silvia Pineda-Munoz - April 2016 57

Diet, ecology, and dental morphology in terrestrial mammals

ecomorphology. Therefore, the classification should serve as a useful basis for future paleoclimatological studies.

4.2 Introduction Contemporary mammals are extraordinarily diverse, having adapted to fill most available ecological niches (Eisenberg 1981, Ungar 2010, Wilson and Reeder 2005). In order to understand this successful radiation, special attention has been paid to the ecological diversity of mammalian communities and its relationship with climate

(Andrews et al. 1979, Eisenberg 1981, Legendre 1986, Petchey et al. 2008). For example, various proxies such as dental morphology and stable isotopes have been used to explore dietary diversity in past and present ecosystems (Demes and Creel 1988,

Palmqvist et al. 2003, Wilson et al. 2012). Reconstructing the interaction between ecological diversity and climate can help track global climate changes throughout the geological time scale by inferring the climate and ecological context of fossil localities

(Andrews and Evans 1979, Fortelius et al. 2002a, Liu et al. 2012).

Body mass is a crucial factor in the dynamics of mammalian evolution (Alroy 1998,

Burness et al. 2001, Smith and Lyons 2011). For example, similar body-mass distributions can be observed across different continents and geological time periods

(Brown and Nicoletto 1991, Fernández-Hernández et al. 2006, Smith et al. 2004, Smith and Lyons 2011, Travouillon and Legendre 2009). However, the factors that may constrain body mass distributions in fossil and modern mammal communities are still open to investigation. Previous work has suggested physiological and mechanical constraints, phylogenetic constraints, correlations with home range area, or the effects of predator-prey relationships (Gingerich 1989, Siemann and Brown 1999, Smith and

58 Silvia Pineda-Munoz - April 2016

Chapter 4: The relationship between diet and body mass in terrestrial mammals

Lyons 2011). Since we can observe really strong patterns across continents in the evolution of body size distribution, it is important to test whether these trends might relate to different climates.

Andrews (1979) analysed both diet and body mass together with locomotion and taxonomy in order to explore differences between ecosystems. He observed that mammals living in environments with similar climates displayed similar ecological diversity and thus similar ecomorphospace occupation. However, a direct relationship between diet and body mass was not explicitly quantified. Later, Fernández-Hernández et al. (2006) evaluated the power of Andrew’s variables for inferring environments and concluded that only body mass was significantly correlated with climate. However, the dietary classification used in these two studies was generalized and not statistically grounded.

Pineda-Munoz and Alroy (2014) proposed a statistically-based classification scheme that emphasized major feeding resources. The categories were herbivory, carnivory, frugivory, granivory, insectivory, fungivory, gumivory and generalised. They argued for abandoning the broadly-used three-way herbivore-omnivore-carnivore categorization because it grouped together species with markedly different dietary specializations.

Previous research has evaluated mammalian body size in relation to other ecological variables such as , ecology or life history (Andrews et al. 1979, Demment and Van Soest 1985, Eisenberg 1981, Legendre 1986). However, their dietary classification might have hindered some evolutionary and ecological patterns. In the present work we will use the same dataset in order to show how this more detailed classification scheme discriminates between ecomorphological specialization. In particular, we will link the near absence of medium sized mammals (1-30 kg) in open

Silvia Pineda-Munoz - April 2016 59

Diet, ecology, and dental morphology in terrestrial mammals

landscapes to the lack of fruit trees needed to support a pure frugivore diet all year round.

4.3 Methods We used a database compiled by Pineda-Munoz and Alroy (2014) that summarizes the dietary preferences of 139 species of terrestrial mammals. We augmented this information with published body mass values (Smith et al. 2003) (See Table 11.2-1).

Each species was classified as a dietary specialist if a single food resource made up 50% or more of the diet (Pineda-Munoz and Alroy 2014). Dietary data were compiled from primary sources presenting volumetric percentages of stomach contents. Despite the fact that stomach contents analyses are not numerous, they provide direct feeding information with minimum degradation from digestive processes and more potentially for identifying ingested foods. Thus, we restricted the main analysis to the species in that study. Dietary classifications included herbivory, carnivory, frugivory, granivory,

insectivory, fungivory, gumivory and generalization. Other researchers have classified diet based on trophic relationships (herbivores, carnivores and omnivores plus a few variations) (Reed 1998, Schoener 1989). Although this can be a good background for some ecological studies, Pineda-Munoz and Alroy (2014) showed how species described as omnivores could display very distinctive dietary specializations (e.g. carnivore-herbivore and insectivore-granivore).

In order to test whether the same ecomorphological patterns could be observed in a wider scale, we also analysed the mammal data of Wilman et al. (2014), a database that includes quantitative percent estimates of lifelong diet for 5,400 mammal species. To make comparisons possible we restricted the analysis to terrestrial non-volant species.

60 Silvia Pineda-Munoz - April 2016

Chapter 4: The relationship between diet and body mass in terrestrial mammals

Their dietary classification divided the items found in a given diet in 10 categories:

Birds, reptiles, fish, unknown vertebrates, scavenge-carrion, fruit, nectar, seeds and plants. All vertebrate feeding resources (mammals, , fish, reptiles and scavenge- carrion) were put in the same feeding category. Unfortunately, we were unable to discriminate fungus and root feeding categories because these were included in the vegetation category by Wilman et al. (2014). Dietary classification otherwise followed the same criteria as in our main dataset. The resulting data set included 2,835 mammal species.

Additionally, we correlated the percentage of fruit in the diet of the frugivore species in our dataset with the values for the same species provided by Wilman et al. (2014). The correlation was poor (r = 0.27) (See Figure 11.2-1). All of our data come from stomach content studies, which suggests that mixing different methodologies might have biased the Wilman et al. (2014) dataset and therefore that adding information from that dataset would bias our analysis.

We used the R statistical environment (R Core Team 2013) to perform analyses and construct tables and figures (packages “psych”, “stats” and “vegan”). A Shapiro-Wilk test for normality showed that the body mass distribution of the species in some dietary categories was non-normal. We therefore performed Kruskal-Wallis tests in order to show whether differences in body mass existed among the feeding categories. We applied a pairwise Wilcoxon rank-sum test to the whole dataset and to the rodent dataset in order to show whether body mass differences existed between mammals with different diet specializations. We carried out principal components analysis (PCA) to explore the relationship between dietary specialization and body mass. PCA was based on covariance matrices instead of correlation matrices, as is standard practice (Bro and

Silvia Pineda-Munoz - April 2016 61

Diet, ecology, and dental morphology in terrestrial mammals

Smilde 2014). Unfortunately, raw percentage values are non-independent because they must add up to 100, which make them computationally unsuitable for PCA. More importantly, PCA tends to underweight variables if the percentages are consistently low because the lower bound of zero compresses potential variance. This property masks the contribution of rare dietary preferences. We solved the problems of non-independence and underweighting by rescaling the percentage values as z-scores before carrying out the PCA. Factor analysis was also applied, but the results were similar and so are not discussed further.

In order to test the relationship between degree of food mixing and body mass we calculated a dietary diversity index. We applied the inverse of the Simpson index

(Simpson 1949) to the resource percentage data extracted from Pineda-Munoz and

Alroy (2014). In this way, the percentage contribution of every resource in the diet of a species was treated as analogous to the percentage of a given species in an ecosystem. Similar methods were proposed in earlier decades to infer dietary diversity and niche breadth (MacArthur and Pianka 1966, Schwartz and Ellis 1981).

We plotted the correlation between body mass and dietary diversity to visually evaluate ecological patterns in feeding behaviour. Additionally, the species were classified into

14 ranked body mass categories using a base-10 logarithm scale (from log10 of body mass = 0.5 to log10 of body mass = 7 with 0.5 unit increments). The maximum degree of dietary diversity for every category was then plotted. The brown bear Ursus arctos was excluded from this particular analysis for reasons discussed below.

All the same analyses (Kruskal-Wallis tests, PCA and comparisons of diet diversity with body mass) were applied to the modified data of Wilman et al. (2014) using the same parameters.

62 Silvia Pineda-Munoz - April 2016

Chapter 4: The relationship between diet and body mass in terrestrial mammals

We evaluated frugivory further because it was one of the few dietary specializations observable in the medium body size range. Frugivore species were classified as either pure frugivores or mixed frugivores following Pineda-Munoz and Alroy (2014). Fruit constitutes 50-80% of the diet of a mixed frugivore's diet and more than 80% of a pure frugivore's diet. Geographical distribution maps for mixed frugivore and pure frugivore species were extracted from Map of Life (MOL) (http:// mol.org). The geographical distribution of tropical and sub-tropical moist broadleaf forests was adapted from Olson et al. (2001). The intersection between these maps and their areas were calculated using

ArcMap (ArcGIS, Esri, USA). Because of the poor correlation between our frugivore data and those given by Wilman et al. (2014) this analysis was not performed using their data.

Our comparisons of body mass and dietary categories assume that data points representing species are statistically independent. We recognise that phylogenetic autocorrelation could cause these comparisons to reflect shared inheritance instead of direct causal relationships. We have, however, not taken an approach such as using phylogenetic contrasts because we have restricted our discussion to describing broad patterns and to offering hypotheses about possible mechanisms as a basis for future research. In other words, we wish to establish the basic patterns prior to engaging in a detailed analysis of evolutionary processes. Furthermore, we believe that body mass evolution is so labile and diet exhibits such rampant that concerns about phylogenetic autocorrelation are likely to be unfounded.

Silvia Pineda-Munoz - April 2016 63

Diet, ecology, and dental morphology in terrestrial mammals

4.4 Results The Kruskal-Wallis analysis of variance for the species in our dataset shows a significant relationship between body mass and dietary specialization in mammals (p- value < 0.001). Table 4.4-1 summarizes the results of the pairwise Wilcoxon tests for the whole dataset and demonstrates statistical differences between some dietary specializations. Frugivory is the most distinctive dietary specialization, with frugivores having significantly different body masses when compared with granivores, insectivores, and generalists in the whole dataset. This interpretation is supported visually by the boxplot in Figure 4.4-1. Small mammals (10-999 g) forage on invertebrates, seeds, or fungi or display opportunistic generalist diets while medium- sized mammals (1-30 kg) have carnivorous or frugivorous diets. Gummivores are represented by a single species in our dataset and this category also fits in the medium- sized body mass range. Herbivore diets cover the entire body size range beyond the size of the smallest mammals (< 10 g). The same patterns can be observed in the data from

Wilman et al. (2014) (See Table 11.2-3, Table 11.2-4 and Figure 11.2-2).

The first four components of the PCA together express 65.76% of the variance and correspond to five major groups of feeding resources: green plants, fruit, seeds, invertebrates, and vertebrates (Table 11.2-3). The first component (19.4% of the variance) mainly discriminates between a group of small-sized mammals that feed on invertebrates or roots and tubers, to the left of the chart, and the ones mainly feeding on green plants, to the right of the chart (Figure 4.4-2). The second component (17.53% of the variance) discriminates mainly between medium-sized fruit-eaters in the bottom of the chart and seed-eaters on the top right of the chart (Figure 4.4-2). Therefore, the two first components alone identify the four major feeding groups. The third component

64 Silvia Pineda-Munoz - April 2016

Chapter 4: The relationship between diet and body mass in terrestrial mammals

(15.19%) discriminates between many of the food resources, with especially high values for fungus and flower-gum eaters. The vertebrate-feeding variable loads strongly only on the fourth component (13.64% of the variance). Some degree of overlap can be observed between the groupings (Figure 4.4-2), which could be related to the amount of food mixing displayed by some species. For example, many of the species in our dataset mix seeds and vegetation in their diet. The same patterns can be observed in the data from Wilman et al. (2014) (See Table 11.2-4 and Figure 11.2-3).

Figure 4.4-1: Boxplot illustrating the body mass range for each dietary category of the 139 mammal species in the dataset. Categories are as described by Pineda-Munoz and Alroy (2014).

Table 4.4-1: P-values of pairwise comparisons based on the Wilcoxon rank-sum test for body mass distributions in diet categories as described by Pineda-Munoz & Alroy (2014) for (A) the whole dataset and (B) only rodents.

Silvia Pineda-Munoz - April 2016 65

Diet, ecology, and dental morphology in terrestrial mammals

a) Whole dataset Carnivore Frugivore Fungivore Granivore Gumivore Herbivore Insectivore Frugivore 1 ------Fungivore 0.6071 0.1946 - - - - - Granivore 0.0271 <0.0001 1 - - - - Gumivore 1 1 1 1 - - - Herbivore 1 0.251 1 0.1946 1 - - Insectivore 0.0338 <0.0001 1 1 1 0.0073 - Generalists 0.0038 <0.0001 1 1 1 0.1048 1

b) Rodents Frugivore Fungivore Granivore Herbivore Insectivore Fungivore 0.76364 - - - - Granivore 0.01074 0.77941 - - - Herbivore 0.03231 1 0.84345 - - Insectivore 0.02424 0.76364 1 0.24684 - Generalists 0.00013 0.18195 1 0.76364 0.86508 Figure 4.4-3 shows that species with extreme body sizes (micro- and megamammals) have more specialized and less diverse diets on average as based on the inverse-

Simpson index (Table 4.4-3). Micromammals (< 10 g) are represented by 3 lipotyphlans with pure insectivore diets and 2 rodents having herbivore and generalist diets. Small mammals (10–999 g) display the most diverse diets (Figure 4.4-3 and Table 4.4-3).

They mostly belong to two taxonomic orders: Rodentia (75%) and Lipotyphla (10.9%).

Medium-sized mammals (1-30 kg) mainly have frugivorous (51.5%), herbivorous

(15.1%), carnivorous (12.1%), or insectivorous (12.1%) diets and they mainly belong to the orders Primates (36.4%), Carnivora (30.3%), and Artiodactyla (18.2%). Large mammals (> 30 kg) are only represented by six artiodactylan species, two carnivorans

(Ursus arctos and Canis lupus), and a proboscidean (Loxodonta africana). The same patterns are seen in the data from Wilman et al. (2014) (See Figure 11.2-4).

66 Silvia Pineda-Munoz - April 2016

Chapter 4: The relationship between diet and body mass in terrestrial mammals

Figure 4.4-2: Scores of the two first axes of the principal component analysis of dietary data for all the 139 species in the dataset (adapted from Pineda-Munoz and Alroy [2014]). Raw data are percentage of stomach content examinations rescaled to z-scores. Arrows indicate loadings on the two first components. Colours and sizes indicate the log10 body mass (g) of each species as shown in the legend.

Table 4.4-2: Proportion of animals in each dietary category for each body mass range as discussed in the text.

Medium-sized Micromammals Small mammals Large mammals mammals (1-30 (<10 g) (10–999 g) (>30 kg) kg) Carnivore 0 0 4 (12.12%) 1 (11.11%) Frugivore 0 9 (9.78%) 17 (51.52%) 2 (22.22%) Fungivore 0 3 (3.26%) 0 0 Granivore 0 15 (16.30%) 1 (3.03%) 0 Gumivore 0 1 (1.02%) 0 0 Herbivore 1 (25%) 18 (19.57%) 5 (15.15%) 6 (66.67%) Insectivore 3 (75%) 29 (31.53%) 4 (12.12%) 0 Generalists 1 (25%) 17 (18.49%) 2 (6.06%) 0

Silvia Pineda-Munoz - April 2016 67

Diet, ecology, and dental morphology in terrestrial mammals

Figure 4.4-3: Relationship between log10 body mass (g) and diet diversity indices calculated by applying inverse Simpson indices to stomach content percentages. Colours represent different dietary specializations. The line connects the maximum dietary diversity value for each half- log10 unit in order to visualize the maximum degree of food mixing. Ursus arctos has been excluded from the maximum dietary diversity curve in order to better visualize the general pattern and for reasons mentioned in the text.

Around 50% of the medium-sized mammals in our dataset have a frugivorous diet and

75% of the frugivorous species fit in the medium size range. 46.6% of the geographical distributions of the pure frugivores and 54.6% of the distributions of the mixed frugivores (Figure 4.5-1) overlap with the distribution of tropical and sub-tropical moist broadleaf forests as defined by Olson et al. (2001). Some differences can be observed in the geographic distribution patterns between South America and Africa. While South

American frugivores only occur within the tropical rainforest region, the ones in Africa show wider distributions, sometimes ranging into such biomes as tropical savannahs.

68 Silvia Pineda-Munoz - April 2016

Chapter 4: The relationship between diet and body mass in terrestrial mammals

Table 4.4-3: Average, standard deviation, and maximum diet diversity values for each body mass range as discussed in the text.

Medium- Small Large Micromammals sized mammals mammals (<10 g) mammals (10–999 g) (>30 kg) (1-30 kg) Average diet diversity (inv- 1.41 1.95 1.64 1.38 simpson index) Standard deviation diet 0.71 0.84 0.48 0.6 diversity (inv- simpson index) Max value for diet diversity (inv- 2.64 4.58 2.92 2.7 simpson index)

4.5 Discussion

4.5.1 Body mass and the degree of diet specialization Our results show how most dietary specializations are restricted to certain body size classes in mammals. In a general way, this sort of pattern has been previously related to phylogenetic, ecologic, and energetic constraints (Eisenberg 1981, Gittleman 1985,

Price and Hopkins 2015). Smaller mammals in our dataset, having higher daily energy and protein requirements, generally require highly digestible food resources (i.e., grains, fungi, and ) (Clauss et al. 2013). They require less dietary specialization because these particular food resources are usually abundant enough to support large populations. Some small mammals also display an herbivorous diet, which has been suggested to be a consequence of ecological opportunity rather than being related to any physiological advantage (Clauss et al. 2013).

In contrast, large mammals are not able to forage on high nutritive rapidly digestible foods because these resources are too rare to support their populations. Instead, they specialise their feeding on vegetation, a much more abundant food resource. Our results

Silvia Pineda-Munoz - April 2016 69

Diet, ecology, and dental morphology in terrestrial mammals

support this idea and show how dietary diversity decreases with increased body mass as a general trend. Having lower requirements for energy and nutrients per unit of body weight allows large mammals to feed on less nutritive feeding resources (i.e., structural plant material) and get most nutrients from a very specialised diet (Clauss et al. 2007,

Clauss et al. 2013, Demment and Van Soest 1985). Increased body size allows herbivores to evolve gut structures that increase volume and retention time of the ingesta and so they are capable of extracting a higher fraction of nutrients from low- energetic plant materials (i.e., leaves and grasses). However, it has been observed that long retention times are not characteristic of very large mammals. In those cases, other ecological advantages such as larger home ranges, predator avoidance, or resource competition are a potential benefit of large body size (Clauss et al. 2007, Clauss et al.

2013, Steuer et al. 2014).

Most micromammals (<10g) in the dataset display a rather specialized insectivorous diet. Their high metabolic costs require them to feed on food resources with substantial energy content such as insects (Peters 1986). Additionally, their small size might mechanically restrict their diets to small invertebrates (Fisher and Dickman 1993).

Thus, it could be postulated that extreme body sizes require higher levels of specialization and that the optimum for a very diverse diet therefore must lie in the small range (10-999 g) (Raia et al. 2012), as it does.

Among predators (pure insectivores and carnivores), the maximum body mass for the insectivores in our dataset is 8.5 kg, with larger predators being carnivores. Similarly,

Carbone et. al. (1999) estimated a maximum sustainable mass of 21.5 kg for invertebrate diets, although some exceptionally larger insectivore mammals such as the (Orycteropus afer) (52 kg) do exist. Additionally, the average size of the pure

70 Silvia Pineda-Munoz - April 2016

Chapter 4: The relationship between diet and body mass in terrestrial mammals

insectivore mammals in Wilman et al. (2014) --- the biggest dataset known so far--- is

914g. Previous research suggested that myrmecophagous mammals from Afrotropical forests seemed to have the lowest population densities. Thus, resource availability might be limiting the maximum body mass of insectivores.

As mentioned, the maximum degree of food mixing tends to decrease with increasing body mass. However, the brown bear U. arctos falls outside of this pattern according to the dietary diversity statistic. Surviving hibernation requires storing energy by consuming highly energetic food resources and increasing body fat (Humphries et al.

2003). In early spring, when more energetic and protein foods are less abundant, U. arctos feeds on vegetation. However, it switches to a more nutritious diet in late summer. This mixed diet would be beneficial for supporting a seasonal higher demand of nutrients prior to hibernating periods (Beeman and Pelton 1980, McLellan 2011).

Thus, U. arctos and other generalist hibernating ursids feed on a diverse, unspecialized diet despite their high body mass.

A few of the generalist species in the dataset show a less diverse diet than some classified as specialists, which could be explained as a mathematical artifact arising in unusual circumstances. For example, an animal feeding on 60% vegetation, 10% insects, 10% fruit, 10% fungi and 10% seeds would be classified as an herbivore, but its diet will be more diverse (inverse Simpson index = 1.23) than that of an animal eating

45% vegetation, 45% fruit and 10% fungi (inverse Simpson index = 0.95).

4.5.2 Frugivory and body mass Most dietary specializations have an optimum body mass range and few dietary specializations occur in the medium-sized range: the only common ones are herbivory, frugivory, and carnivory. Most frugivorous mammal species in our dataset have a body

Silvia Pineda-Munoz - April 2016 71

Diet, ecology, and dental morphology in terrestrial mammals

mass between 500 g and 30 kg. Previous studies found similar patterns, suggesting a peak in frugivory in the medium size range for Neotropical primates (Hawes and Peres

2014, Kay 1984, Robinson and Redford 1986). A diet with a very high proportion of fruit has been proposed to constrain body size due to mechanical, locomotional, ecological and metabolic factors (Hawes and Peres 2014, Milton and May 1976,

Robinson and Redford 1986).

Figure 4.5-1: Geographical distribution of tropical and sub-tropical moist broadleaf forests (green) (adapted from Olson et al. (2001)) and of pure frugivore (blue contour line) and mixed frugivore species (red contour line) in our dataset (extracted from Map of Life (MOL)).

Hawes and Peres (2014) performed an exhaustive study on the frugivory of Neotropical primates with special attention paid to the relationship with body mass. They observed higher rates of frugivory in medium-sized primate species (2-3 kg). The proportion of fruit in the diet of smaller species was much lower, with a higher intake of seeds and insects; while the largest ones foraged on an increasingly higher amount of foliage, a pattern also observed in previous studies (Kay 1984).

72 Silvia Pineda-Munoz - April 2016

Chapter 4: The relationship between diet and body mass in terrestrial mammals

Frugivore species in our dataset are distributed around tropical and subtropical moist broadleaf forests. Thus, species with a higher percentage of fruit in their diet have a more restricted geographical distribution. They can also be found in forested patches in the surrounding areas where they have to complement their diets with other food resources such as vegetation or insects, as shown in our resource utilization dataset.

We recorded frugivore species in tropical environments in South America and Africa but not in Southeast Asia and the Indo-Pacific (Figure 4.5-1). This fact could be related to the nature of our dietary data. Pineda-Munoz & Alroy (2014) limited their study to the stomach content literature in order to standardize data and avoid sampling bias.

Unfortunately, very little stomach content research has been performed in the latter two regions. However, there are some examples of medium-sized tropical frugivores in these regions such as the Indian giant squirrel Ratufa indica, with an average adult size of 1.5-2 kg; the liontail Macaca silenus, with an average size of 3-10 kg

(Ganesh and Davidar 1999); and the binturong Arctictis binturong, with an average size of 9-20 kg (Colon and Campos-Arceiz 2013).

Similarly, the more restricted geographical distribution of South American frugivore species as compared to African ones could be related to sampling biases and a higher level of endemism in the South American region. Stomach content studies have been mainly carried out near the Amazon River. However, South America seems to have a high diversity of frugivore species such as primates, which ultimately results in allopatric, closely related species having more restricted geographic distributions

(Wilson et al. 1988). Pure frugivores such as the black uacari (Cacajao melanocephalus) or the Colombian woolly monkey (Lagothrix lugens) can be found in

Silvia Pineda-Munoz - April 2016 73

Diet, ecology, and dental morphology in terrestrial mammals

areas surrounding those that provided most of the South American stomach content data.

We hypothesize that wholly frugivorous diet is only suitable in tropical rainforests or similar biomes where fruit resources are available all year round. Seasonality of precipitation or outright aridity prevents species from specializing on an entirely frugivorous diet in more open environments (Ganesh and Davidar 1999). Thus, based on our results we predict that medium-sized mammals should be less frequent in open environments. These observations are consistent with Rodríguez (1999), who quantitatively documented a decrease in the density of mid-sized mammals in increasingly open landscapes.

Plant species composition in open environments is highly heterogeneous, with fruiting trees widely dispersed across space. As a consequence, frugivores in such environments would have to rely on a food resource that is rather unstable and patchily distributed, which would therefore require them to have unrealistically large home ranges (Ganesh and Davidar 1999, Milton and May 1976).

According to optimal theory, diet choice is conditioned by the need to maximize energy intake per unit time spent on the foraging activity (Bartumeus and

Catalan 2009, MacArthur and Pianka 1966). An animal relying on a patchily-distributed resource will then be forced to face a trade-off between the nature of resources and the energy required to move from patch to patch (Bartumeus and Catalan 2009, Pyke et al.

1977). Thus, the and the of fruit trees in tropical rainforests play an important role in the evolution of the relationship between diet specialization and body mass in tropical mammalian species. This trade-off explains why the optimum body mass for a frugivore diet would fit around the medium size

74 Silvia Pineda-Munoz - April 2016

Chapter 4: The relationship between diet and body mass in terrestrial mammals

range (1-30 kg). The energy invested in foraging activity – moving across patches or climbing trees – is cost-ineffective for smaller and for bigger species (Bartumeus and

Catalan 2009, Pyke et al. 1977). A pure frugivore diet is therefore restricted to the medium size range within which foraging efficiency reaches its maximum.

4.5.3 Frugivory and the medium-sized gap Many ecological analyses have pointed out the decline or absence of medium-sized mammals (500 g-30 kg) in open-environment mammalian communities. Some interpretations in the literature include predator-prey relationships, trophic, physiological and mechanical constraints, taxonomic limitations or predator avoidance

(Gingerich 1989, Legendre 1986, Smith and Lyons 2011, Valverde 1967).

Alroy et al. (2000) hypothesized that the opening of a medium-sized gap in North

America during the middle of the Cenozoic could be linked to increased seasonality.

The middle-sized range was emptied during the Middle Eocene (about 46 million years ago) as ecosystems became more arid and seasonal. They also recognized a decrease in ecomorphological diversity after this period, as arboreal frugivorous species were replaced by large terrestrial herbivores. This interpretation supports our idea of a relationship between the opening of a medium-sized gap and the decrease or disappearance of frugivores in open environments.

Additionally, Alroy (1998) recognized a general evolutionary trend in North American mammals towards increased body mass during the Cenozoic. However, he observed an upper size limit to the evolution of small taxa around 500 g, where the mid-size gap starts in North American mammals. This limit would have been established when vegetation structure changed towards more open landscapes, reducing the number of niches left to explore. In parallel, medium-sized tree-dwelling mammals evolved

Silvia Pineda-Munoz - April 2016 75

Diet, ecology, and dental morphology in terrestrial mammals

towards increased body size (Alroy 1998) or were replaced by open-environment herbivores. Interestingly, and as pointed out by Smith and Lyons (2011), the upper limit of the medium-size gap coincides approximately with the lower limit observed for ruminant herbivores (5–10 kg) (Demment and Van Soest 1985).

4.5.4 The consequences of improper diet classifications All paleoecological methods make assumptions. In particular, biological parameters need to be categorized properly in order to statistically test their relationships with ecological and climatological variables. The present results suggest that the dietary categories of Pineda-Munoz and Alroy (2014) strongly correlate with body mass, which suggests that the classification is useful. Additionally, recent research has shown that this classification correlates well with dental morphology (Pineda-Munoz 2015).

Fernández-Hernández (2006) statistically tested some paleoecological methods in order to evaluate their power as climate and paleoclimates estimators: ecological diversity analysis (Andrews et al. 1979) and cenograms (Legendre 1986). The significance of the individual paleoecological variables used in these methodologies was also examined.

The variables were taxonomic affiliation, trophic relationship, locomotion, and body size. The results suggested that body size was the best ecological variable for inferring climate of those tested. However, the dietary classification was statistically untested and ambiguous because frugivorous and granivorous species were put together in a single category. Our dataset shows that frugivore and granivore mammals have statistically different body size ranges (p-value < 0.05, see Table 4.4-1). Thus, including them in a single category masked some ecological signals, which may have caused diet to be undervalued as a climatological indicator (Fernández-Hernández et al. 2006).

Furthermore, the geographical distribution of the frugivorous species in our dataset

76 Silvia Pineda-Munoz - April 2016

Chapter 4: The relationship between diet and body mass in terrestrial mammals

shows a strong correlation between frugivory and tropical and subtropical climates.

Thus, if frugivores had been separated from granivores, diet would have better discriminated between rainforests and open savannahs. This observation reinforces the idea that a more comprehensive dietary categorization can more greatly empower paleoecology.

4.6 Acknowledgements We thank M. McCurry and Xuan Zhu for assistance with GIS analysis. We also thank

Nick Chan, David M. Alba, colleges at Macquarie University and National Museum of

Natural History Smithsonian Institution and two anonymous reviewers for comments and suggestions. SP-M was supported by Macquarie University's HDR Project Support

Funds. AE acknowledges the support of the Australian Research Council and Monash

University. This is the Evolution of Terrestrial Ecosystems Program publication number

339.

4.7 Author’s contributions SP-M designed the study, collected and analysed the data, and drafted the manuscript.

AE created the GIS maps and helped interpret the data and draft the manuscript. JA assisted data analysis and interpretation and drafting of the manuscript. All authors gave final approval for publication.

Silvia Pineda-Munoz - April 2016 77

Diet, ecology, and dental morphology in terrestrial mammals

5 THE TEMPORAL SCALE OF DIET AND DIETARY PROXIES

Accepted in Ecology and Evolution as: Davis M, S. Pineda-Munoz. The temporal

scale of diet and dietary proxies.

78 Silvia Pineda-Munoz - April 2016

Chapter 5: The temporal scale of diet and dietary proxies

5.1 Abstract Diets estimated from different proxies such as stable isotopes, stomach contents, and dental microwear often disagree, leading to nominally well-supported but greatly differing estimates of diet that complicate our understanding of a species’ ecology and natural history. Using several high-resolution fossil and modern mammal datasets, we show that these perceived diet incongruences can be caused by proxies recording the natural variability of diet on vastly different timescales. Field observations reveal a diet averaged over minutes or hours whereas dental morphology may reflect the diet of a lineage over millions of years of evolution. Failing to explicitly consider the scale of proxies and the potentially large temporal variability in diet can cause erroneous predictions in any downstream analyses like conservation planning or paleohabitat reconstructions. We demonstrate how large the temporal variability in diet can be and propose a cross-scale framework for conceptualizing diet suitable for both modern ecologists and palaeontologists. Treating diet in this temporally explicit framework will lead to a clearer understanding of ecological and evolutionary processes as long as we match the scale of our questions with the scale of our data.

Keywords: diet, isotopes, microwear, mesowear, faecal, stomach, scaling, time averaging, dietary proxies, temporal ecology

5.2 Introduction Diet is a fundamentally important biological trait that widely influences physiology and morphology, tempo of trophic evolution and its impact on mammalian diversification

(Price et al. 2012). Diet reconstructions are crucial for properly managing species, constructing food webs, studying niche theory, examining evolutionary changes in

Silvia Pineda-Munoz - April 2016 79

Diet, ecology, and dental morphology in terrestrial mammals

function, and inferring ancient climates and (Dalerum and Angerbjörn 2005,

Feranec 2004, Pineda-Munoz and Alroy 2014). To infer natural diets, both neontologists and palaeontologists have developed a series of proxies that depend on behaviour, dental wear, stable isotopes, gut contents, and skeletal morphometrics (See

Table 11.3-1 and Supplementary Information 3 [Section 11.3]). However, agreement among proxies and test diets is often poor (Kessler et al. 1981, McCann and Hastings

1997, Schubert et al. 2006, Shrestha and Wegge 2006). Three proxies may show strong support for three different dietary reconstructions, confounding analysis of both fossil

(Figueirido et al. 2010, Mendoza et al. 2002, Schubert et al. 2006) and extant species

(Gogarten and Grine 2013, Shrestha and Wegge 2006).

This is partly because the operational meaning of “diet” is rarely explicitly defined

(Gagnon and Chew 2000, Hyslop 1980). Is food ranked by the volume or mass consumed, its caloric value, or the feeding time necessary to manipulate it (Cumberland et al. 2001, Hyslop 1980)? Each proxy also records a slightly different aspect of diet.

Feeding observations reveal what food enters an animal’s mouth, stable isotopes record those nutrients that actually contribute proteins to growing tissues, and faecal analysis technically only measures food that passes through the gastrointestinal tract with minimal digestion (Shrestha and Wegge 2006, West et al. 2006).

Most importantly, different dietary proxies record diet across a large range of temporal scales (Figure 5.2-1-A, Table 11.3-1) (Dalerum and Angerbjörn 2005, Del Rio et al.

2009, Fortelius and Solounias 2000, Kaiser et al. 2013, Kurle and Worthy 2002,

Loffredo and DeSantis 2014, Münzel et al. 2014, Schubert et al. 2006). Stomach contents might average together up to a week’s worth of meals (Kararli 1995) whereas tissues like hair record an isotopic signature of diet for as long as they are growing and

80 Silvia Pineda-Munoz - April 2016

Chapter 5: The temporal scale of diet and dietary proxies

could represent years or even decades of an animal’s diet (West et al. 2006). Because diet can change significantly over ontogenetic (Kurle and Worthy 2002), ecological

(Hobson et al. 1999, Munro et al. 2006), and evolutionary timescales (Cerling et al.

2009, Rivals et al. 2012, Rivals and Semprebon 2011), diet reconstructions can also change depending on the temporal extent of measurement (Dalerum and Angerbjörn

2005, Del Rio et al. 2009). Although the different temporal scales of proxies are widely recognized, the common failure to explicitly consider them leads to perceived incongruences in diet and questionable inferences about any related concepts like paleohabitat reconstructions or foraging theory (Dalerum and Angerbjörn 2005, Feranec

2004, Reynolds-Hogland and Mitchell 2007, Schubert et al. 2006).

Here, we reanalyse several modern and fossil dietary datasets, demonstrating quantitatively that ignoring the temporal scaling of diet (Figure 5.3-1, Figure 11.3-1-

Figure 11.3-5) and dietary proxies (Figure 5.2-1a) can lead to sizable incongruences

(Figure 5.2-1b) resulting in spurious inferences in downstream analyses (Figure 5.3-2,

Figure 11.3-6). We propose a cross-scale framework of diet suitable for conservation biologists as well as palaeoecologists, and outline areas where future research is most needed. This framework stresses an explicit alignment of dietary data with the scale of evolutionary or ecological questions (Figure 5.3-2, Figure 11.3-6) and allows us to roughly infer diet through almost 15 orders of temporal magnitude from seconds of an organism’s life to millions of years of a lineage’s evolutionary history (Figure 5.2-1a).

Silvia Pineda-Munoz - April 2016 81

Diet, ecology, and dental morphology in terrestrial mammals

Figure 5.2-1: (a) A proxy’s resolution is given by its temporal grain. Temporal extent is the range of time over which a proxy can be used. Proxies marked by a pointed end have ranges that extend past the graph limits. Double pointed ends indicate that proxies can be used in exceptional fossil cases. (b) The perceived diet of the African bush elephant (Loxodonta africana) depends on the timespan over which it is measured. Blue shaded area represents dietary limits of elephants at different scales estimated from observational and fossil evidence. Grey lines show actual diets at different scales computed from the 6 year isotopic record of (Cerling et al. 2009). Four lines are highlighted in colour to show how perceived diet changes as it is averaged for longer periods of time. Both graphs share the same logged x axis given in years above and common calendar units below.

82 Silvia Pineda-Munoz - April 2016

Chapter 5: The temporal scale of diet and dietary proxies

5.3 Discussion

5.3.1 Diet changes over time Despite a broad acceptance of the temporal variability of diet (Dalerum and Angerbjörn

2005), most models implicitly assume diet stationarity, assigning a single qualitative variable like “browser” or “omnivore” whether examining trophic guilds over millions of years or a small forest plot through one day (Wilman et al. 2014). Although criticized for having arbitrary boundaries (Feranec 2004, Pineda-Munoz and Alroy

2014), these rough diet categorizations are comparable across taxonomic groups and studies and are simpler to incorporate into complex statistical models (Price et al. 2012).

They are also often the only dietary information available.

Collecting data from proxies like stomach contents is difficult and prohibitively expensive (Kessler et al. 1981, McInnis et al. 1983). Even compiling data from the literature is complicated, as experts often don’t express diet quantitatively or use classification schemes that are difficult to compare, e.g., percent seeds vs. leaves and percent graminoids vs. forbs (Gagnon and Chew 2000, Wilman et al. 2014). Most studies considering more than a few species must rely on previously compiled datasets that synthesize information from many proxies and sources into rough measures of diet

(Cerling et al. 2009, Gagnon and Chew 2000, Wilman et al. 2014). These datasets often don’t list how long diet was measured for, so temporal scaling is rarely accounted for or even acknowledged in downstream analyses (Jones and Safi 2011). But dismissing intraspecific temporal variation in diet may introduce severe errors into comparative analyses (Dalerum and Angerbjörn 2005, Losos 2008) .

To show how large this intraspecific temporal dietary variability can be, we reanalysed three datasets measuring diet over time for chimpanzees (David Watts, personal

Silvia Pineda-Munoz - April 2016 83

Diet, ecology, and dental morphology in terrestrial mammals

communication), bears (Mattson et al. 1991), and extinct ground sloths (Hansen 1978)

(Figure 5.3-1) (see Supplementary Information 3 for more detailed descriptions of source data). Pineda-Munoz and Alroy (2014) compiled detailed stomach content data for a taxonomically diverse group of 139 species of mammals from primary sources by calculating the relative volume of food consumed in each of the following categories: seeds, invertebrates, vertebrates, fungi, flowers and gum, roots and tubers, green plants, and fruit. We created a multidimensional dietary space for those 139 mammal species by calculating the Euclidean distance between the percent diets of each species pair and then performing a principal component analysis (PCA). To compensate for the non- independence of percentage categories, raw percentages in each category were z-score transformed by subtracting the mean of that category and then dividing by its standard deviation. All diet categories were weighted equally when calculating distances.

For simplicity, we plotted only the first two principal components (19% and 18% of variance, respectively) and drew 99% confidence interval ellipses around all those species that had 50% or more of their diet in one category (Figure 5.3-1a). This procedure creates a continuous dietary space for mammals with regions that represent traditional dietary categorizations like “frugivore” or “carnivore.” To show how diets can change over time, we projected the chimpanzee, bear, and ground sloth diet time series into this PCA space by z-score transforming their percentage diets (using the same means and standard deviations as in the original dataset) and multiplying the resultant values by the loadings for each diet category (Figure 5.3-1b-f).

The stomach data used by Pineda-Munoz and Alroy (2014) mostly represent yearlong estimates of diet, but data from other scales were also included, making this compilation comparable to other popular databases (Jones and Safi 2011, Price et al. 2012, Wilman

84 Silvia Pineda-Munoz - April 2016

Chapter 5: The temporal scale of diet and dietary proxies

et al. 2014). To ensure that the classification scheme and species in Pineda-Munoz and

Alroy (2014) were not driving any of the patterns discussed below, we also reran analyses using the most complete dietary database available, EltonTraits 1.0 (Wilman et al. 2014) which partitions carnivory more finely than Pineda-Munoz and Alroy (2014) and includes quantitative percent estimates of lifelong diet for 5,400 mammal species

(Figure 11.3-1- Figure 11.3-4). We performed all analyses in the statistical software R, version 3.0.1 (R Development Core Team 2011).

Silvia Pineda-Munoz - April 2016 85

Diet, ecology, and dental morphology in terrestrial mammals

Figure 5.3-1: Diet changes over time. (a) Scores of the two first components of the principal component analysis of dietary data for all the 139 species in Pineda-Munoz and Alroy (2014).

86 Silvia Pineda-Munoz - April 2016

Chapter 5: The temporal scale of diet and dietary proxies

Raw data are square-root transformed percentages of food items based on stomach content examinations. Shaded areas represent 99% confidence ellipses around those species that consume 50% or more of a particular dietary category as shown in the legend. (b–f) Projected diet time series into the PCA space in figure section a for different dietary studies: (b) Diet of Shasta ground sloth (Nothrotheriops shastensis) from dung deposits in Rampart Cave, Arizona (Hansen 1978). Daily (c) and weekly (d) diet of chimpanzees (Pan troglodytes) observed in Kigali National Park, Uganda by David Watts (personal communication). Monthly (e) and yearly (f) diet of brown bears (Ursus arctos) in Yellowstone National Park, Wyoming based on scats (Mattson et al. 1991). Axes are identical for all plots except monthly bear diet (e). The multi-coloured lines (b–f) show where species are in the dietary space at sampled times.

Watts (personal communication) and his field assistants recorded the amount of time per day that chimpanzees (Pan troglodytes) spent feeding on different resources at the Ngogo research site in Kigali National Park, Uganda from April 7 to July 28, 2013.

Plotting the daily diets of chimpanzees on the underlying mammalian dietary space shows how quickly chimpanzees switch back and forth between a frugivorous and herbivorous diet (Figure 5.3-1c). Despite this high daily variability, a seven day average (the same approximate scale as stomach content analysis) shows a clear loop through dietary space as the summer progresses and the chimpanzees switch from a diet heavy in fig fruit to one of mostly leaves and seeds (Figure 5.3-1d). No researcher would ever confuse an African bush elephant (Loxodonta africana) with a European pine marten (Martes martes) in a diet-based niche model of the Serengeti, but we commit a similar error if we ignore the intraspecific temporal variation in chimpanzee diets. The chimpanzee’s diet can be as different from one day to another as an elephant’s diet is from a pine marten’s (Figure 11.3-5).

High daily variability in diet is expected, but brown bears (Ursus arctos), the quintessential mammalian omnivores, reveal how large variability can be even at monthly and yearly scales (Figure 5.3-1e-f). Mattson et al. (1991) opportunistically

Silvia Pineda-Munoz - April 2016 87

Diet, ecology, and dental morphology in terrestrial mammals

collected scat samples while tracking 96 radio collared brown bears in Yellowstone

National Park, USA from 1977 to 1987. Bears show extremely high month to month variation in diet (Figure 5.3-1e) going well outside the dietary space of all other species several times during fall months when root consumption is very high (Mattson et al.

1991). Over half of the differences in bear diet between any two months are larger than the dietary distance that separates bears from elephants (Figure 11.3-5). Even yearly averages of bear diet (Figure 5.3-1f) show little sign of reaching equilibrium over time

(Mattson et al. 1991). Bears can differ more in diet between consecutive years than they differ from elephants (Figure 11.3-5).

Relatively specialized, obligate herbivores can also show long term changes in diet. Hansen (1978) reported a >30,000 year long record of diet for the Shasta ground sloth (Nothrotheriops shastensis) from dung collected in Rampart Cave, Arizona.

Although possessing only a small fraction of the variability of chimps and bears, the ground sloth’s fruit consumption oscillates measurably over thousands of years (Figure

5.3-1b, Figure 11.3-5) corroborating multiple studies showing that even in perceived specialists, diet stationarity is a poor assumption over long timespans (Cerling et al.

2015, Faith 2011).

Both generalists and presumed specialists can show substantial variation in diet over short and long timespans (5.3-1b-f; for similar results using the EltonTraits 1.0 database see Figure 11.3-1- Figure 11.3-4). Admittedly, both bears and chimpanzees are recognized generalists and the high dietary variability they display is likely near the upper bound expected in mammals. But even the temporal differences in ground sloth diet (Figure 5.3-1b, Figure 11.3-5) are large enough that we need to consider how static measures of diet could affect ecological and evolutionary comparative analyses.

88 Silvia Pineda-Munoz - April 2016

Chapter 5: The temporal scale of diet and dietary proxies

Proxy scaling and temporal variability can create incongruences in diet reconstructions

African bush elephants present a practical example of how ignoring the temporal scales of proxies can lead to large and seemingly inexplicable incongruences in diet reconstructions. Elephants are a threatened species and major ecosystem engineers

(Johnson et al. 1999), so knowing their actual diet is an important part of conservation efforts as well as paleohabitat reconstructions (Cerling et al. 2015, Feranec 2004,

Schubert et al. 2006). Elephants posses large teeth with grooved ridges that are continually replaced in a conveyor belt fashion (Laursen and Bekoff 1978), features thought to be adaptations to a diet of massive quantities of grass (Cerling et al. 1999).

However, isotopic work on modern elephants shows that they are browsers, consuming little graze (Cerling et al. 1999). Descriptions of elephant diet based on stomach contents and field observations further complicate matters as they vary widely, even when considering elephants in the same region (Cerling et al. 1999).

To show how proxies recording different temporal scales can change the perceived diet of an elephant, we reanalysed seasonality data from Cerling et al. (2009) who used stable isotope analysis of tail hair serially sampled every 5 mm to produce a 6 year dietary record with near weekly resolution of percentage C4 (tropical grasses) vs. C3

(trees, shrubs, and forbs) plants consumed by one family group of African elephants in northern Kenya (Supplementary Information 3). Starting at the last date measured by

Cerling et al. (2009), January 11, 2006, we computed successive cumulative averages representing diet for ~ 1 week, 2 weeks, 3 weeks, and so on until the dietary signal was averaged for the full timespan of the study, almost 6 years. This shows how much diet can change based only on how long it is measured. But this line of successively larger cumulative averages is heavily influenced by starting conditions and would change

Silvia Pineda-Munoz - April 2016 89

Diet, ecology, and dental morphology in terrestrial mammals

depending on whether we began measuring an elephant’s diet during the rainy season when it was consuming more grass or during the dry season when it was consuming more browse. We therefore repeated this process for each measured time point in

Cerling et al. (2009) producing 367 lines of decreasing length that show how an individual’s diet could change if it were averaged over longer and longer timespans

(Figure 5.2-1b).

For example, stable isotopes from a 5 mm clipping of tail hair would reveal that for the week around May 18, 2002, an elephant ate 67% C4 plants, making it a grass-dominated mixed feeder in most classifications (Cerling et al. 2009). But if we sampled the stable isotopes in the collagen of a 5 mm radius transverse slice from that same elephant’s tusk, we would find that it ate only 24% C4, making it a browser (Codron et al. 2012).

This large incongruence exists only because we are averaging diet for a single week with one proxy and for 70 weeks with the other; both record the same diet and both are correct.

The best dataset available (Cerling et al. 2009) only covers a small portion of the range of scales over which we might investigate diet. But it would be trivial to find an elephant eating 100% browse or graze during any given minute (Cerling et al. 1999), so at shorter timespans, we would expect estimates to cover the full range of possibilities between grazer and browser (Figure 5.2-1b). Isotopes from fossils (Cerling et al. 2015,

Cerling et al. 1999) show that elephants ate a much higher percentage of C4 plants in the past than they do now, so we would also expect an upswing in C4 plant consumption if diets were averaged for millions of years (Figure 5.2-1b). This is why the morphology of modern elephant teeth, a very coarse-grained proxy, suggests a grazing diet whereas

90 Silvia Pineda-Munoz - April 2016

Chapter 5: The temporal scale of diet and dietary proxies

isotopes of those same teeth suggest browsing affinities. Each proxy measures diet over a very different timespan.

5.3.2 Cross-scale studies offer a richer view of diet The richest understanding of diet comes not from one particular scale but from comparing diet across several scales (Dalerum and Angerbjörn 2005, Del Rio et al.

2009, Schubert et al. 2006). Repeated sampling of a proxy over time, comparing proxies with different temporal scales, and serially subsampling a tissue that integrates diet over time all generate cross-scale estimates of diet (Dalerum and Angerbjörn 2005).

Sampling one proxy repeatedly over time is the most common form of cross scale dietary analysis (Dalerum and Angerbjörn 2005) and measures seasonal and yearly variation (Andelt et al. 1987) or major trends over millions of years (Kimura et al. 2013) depending on whether the proxy has a short temporal extent like faecal contents or a long extent like dental isotopes.

Comparing differently scaled proxies in the same species is the least common method of constructing cross scale descriptions among modern ecologists (Dalerum and

Angerbjörn 2005) but it should be used more often, especially in conservation studies, because a wide range of dietary timescales can be sampled non-lethally during just one encounter (Del Rio et al. 2009). This type of multi-proxy analysis is already common in paleontological studies and is increasingly used in a temporally explicit framework to understand diet from ecological to evolutionary timescales (Feranec 2004, Loffredo and

DeSantis 2014, Louys et al. 2012, Rivals et al. 2007, Schubert et al. 2006, Semprebon and Rivals 2010, Tütken et al. 2013).

Serially sampling tissues that have progressive growth like hair, baleen, tusks, and feathers has the greatest potential to create long term, high-resolution dietary records

Silvia Pineda-Munoz - April 2016 91

Diet, ecology, and dental morphology in terrestrial mammals

(Cerling et al. 2009) as some tissues could record decades of diet with almost weekly resolution (Dalerum and Angerbjörn 2005, Schell 2000). If isotopic baseline issues are accounted for (Casey and Post 2011), high resolution dietary records tracking anthropogenic climate change through decades or even centuries could be created by linking together tusk or baleen data from multiple museum specimens (Schell 2000).

Even with a cross-scale description of diet, the temporal scales of proxies must be explicit. The exact timespan over which proxies average diet can vary by tissue, age, and taxon with relatively few species having parameters confirmed by diet switch experiments (Crawford et al. 2008, Dalerum and Angerbjörn 2005, Rio and Carleton

2012). Still, care should be taken to avoid “scale jumping” by grossly mismatching the extent and grain of an ecological or evolutionary question with the extent and grain of data (Behrensmeyer 2006, Crawford et al. 2008, Reynolds-Hogland and Mitchell 2007)

(Figs. 3, S6). Even an order of magnitude estimate of a proxy's temporal scale should improve understanding and reduce potential error compared to a scale-blind approach.

Figure 5.3-2: Diagrammatic examples of how mismatching temporal grain (a) and extent (b) between questions and proxies can lead to erroneous inferences. Temporal grain example (a)

92 Silvia Pineda-Munoz - April 2016

Chapter 5: The temporal scale of diet and dietary proxies

based on (Reynolds-Hogland & Mitchell 2007). Temporal extent example (b) based on (Martínez del Rio et al. 2009). See Fig. S6 for similar paleoecological examples.

5.3.3 Towards a cross-scale dietary framework Asking what an elephant eats without specifying when and for how long its diet is measured makes no more sense than asking what the length of the coastline of Great

Britain is without specifying the length of the ruler used (Mandelbrot 1967). Just as spatial ecologists recognize that biological patterns vary across spatial scales (Belmaker and Jetz 2011), we need to change how we conceptualize traits like diet that vary across temporal scales (Wolkovich et al. 2014). Differences in diet reconstructions are only problematic when we fail to consider this scale dependency. Examining differently scaled proxies gives us a much richer understanding of an organism’s biology compared to just one annual average (Dalerum and Angerbjörn 2005, Schubert et al. 2006).

However, as a field, we don’t have enough data to adopt rigorous multiscale approaches in all analyses. Although some taxa like elephants are well studied with multiple proxies (Cerling et al. 1999), many species lack any kind of dietary data (Penone et al.

2014). Price et al. (2012) were only able to find primary reports of diet obtained from direct observation or stomach and faecal contents for 1,530 out of the 5,020 mammal species (30%) they considered. PanTHERIA (Jones et al. 2009), a larger effort using additional proxies, could only assemble data for 2,161 of the 5,416 mammal species

(40%) included. Even using more numerous qualitative sources of data, EltonTraits 1.0

(Wilman et al. 2014) found diets for at most, 4,352 of its 5,400 mammal species (81%).

Many proxies are used out of operational necessity not for a preferred temporal scale of investigation (Reynolds-Hogland and Mitchell 2007) e.g. non-invasive scat sampling is indispensible for conservation biologists but coprolites are relatively rare to the

Silvia Pineda-Munoz - April 2016 93

Diet, ecology, and dental morphology in terrestrial mammals

palaeontologist. These operational limitations create large, non-random, and sometimes unacknowledged temporal biases between subfields that hamper our understanding of ecological and evolutionary patterns. Using mostly coarse-grained proxies, palaeontologists can underestimate dietary variability at shorter timescales

(Supplementary Information 3, Figure 11.3-6), leading to erroneous paleohabitat reconstructions based only on a portion of a species’ diet. Neontologists, using mostly proxies with shorter extents, can underestimate variability at longer timescales (Figure

11.3-6b), producing models that may not accurately predict ranges under climate change.

As long as we recognize that these diet estimates are limited and correspond to a certain temporal scale, we can ensure that our questions do not exceed the scale of our data

(Behrensmeyer 2006, Crawford et al. 2008, Reynolds-Hogland and Mitchell 2007)

(Figure 5.3-1, Figure 11.3-6). But whenever possible, we should consider multiple proxies and a cross-scale framework (Dalerum and Angerbjörn 2005, Schubert et al.

2006). We need more basic research on the range of intraspecific dietary variability that can be found in cross-scale analyses. A single museum specimen could provide observational data in the form of field notes, stomach contents, dental microwear and mesowear, tooth isotopes, bone collagen isotopes, serially sectioned hair isotopes, and tooth morphology all linked to a spatially and temporally explicit location Dalerum and

Angerbjörn 2005). Examining these modern “incongruences” in a temporally explicit way should allow us to estimate the diets of both modern and extinct species more accurately (Schubert et al. 2006) and lead us to a richer understanding of ecological and evolutionary processes (Wolkovich et al. 2014).

94 Silvia Pineda-Munoz - April 2016

Chapter 5: The temporal scale of diet and dietary proxies

5.4 Acknowledgements We thank Thure E. Cerling and David Watts for generously sharing data and John

Alroy, Alistair Evans, and Walter Jetz for helpful comments and suggestions on earlier versions of the manuscript. Research was supported by grants from the Geological

Society of America, the American Society of Mammalogists, the Yale Institute for

Biospheric Studies, the International Palaeontological Association and Congress,

Macquarie University's HDR Project Support Funds, and two Smithsonian Institution

Predoctoral Fellowships.

5.5 Author’s contributions Both authors conceived the project, performed analyses and wrote sections for the manuscript. MD compiled the first draft of the manuscript and both authors contributed to revisions.

Silvia Pineda-Munoz - April 2016 95

Diet, ecology, and dental morphology in terrestrial mammals

6 DENTAL MORPHOLOGY PREDICTS DIET ACROSS MARSUPIALS AND PLACENTALS

Authors: Pineda-Munoz S, J. Alroy, and A.R. Evans.

96 Silvia Pineda-Munoz - April 2016

Chapter 6: dental morphology predicts diet across marsupials and placentals

Abstract

Marsupials and placentals have experienced extraordinary adaptive radiations (O'Leary et al. 2013). Here we provide the first comparisons between these groups using a taxonomy-free multidimensional methodology that assesses whether their dental phenotypes have convergently evolved effective solutions for the wide range of diets in each group. Using three-dimensional scans of 32 marsupial and 115 placental dentitions, we calculated orientation patch count, slope diversity and the relief index (Boyer 2008,

Evans et al. 2007, Lazzari et al. 2008b), and used multivariate statistical analyses to test whether dental phenotype could discriminate diets of species. There were significant morphological differences across diets in a dataset including all the species and for each order separately, other than rodents. Diet was correctly discriminated for up to 82% of the species. Marsupials and placentals with the same dietary specializations overlap in each ecomorphospace, suggesting strong convergent phenotypic evolution, and gives promise for a common framework for the prediction of diet in all fossil mammals.

6.1 Main text Recent studies place the divergence between eutherians (including placentals) and metatherians (including marsupials) within the Early Cretaceous or even earlier

(O'Leary et al. 2013). Both clades have experienced an extraordinary radiation spanning body sizes from a few grams to about 3 tonnes in marsupials and more than 10 tonnes in terrestrial placentals (Saarinen et al. 2014, Ungar 2010, Wroe et al. 2004).

Additionally, they occupy most terrestrial habitats from alpine mountains to tropical rainforests and deserts (Lillegraven et al. 1979, Ungar 2010) and have independently evolved to forage on almost the entire spectrum of feeding resources available. This

Silvia Pineda-Munoz - April 2016 97

Diet, ecology, and dental morphology in terrestrial mammals

remarkable variety makes them ideal for testing for convergent evolution in one of the fundamental feeding systems of mammals, teeth.

Teeth adapt readily to cope with the mechanical needs of dietary specialization (Evans and Sanson 2003, Lucas 2004, Polly 2015, Ungar 2010). Thus, mammals from independent evolutionary lineages that have the same diet preferences are expected to evolve morphologically-similar dentitions. Although this idea has been widely accepted for decades, the relationship between morphology and diet has not been tested both broadly and quantitatively. Since adapting to an available food resource will improve fitness, we expect the physical properties of food to drive dental evolution towards

“ideal” morphologies (Evans and Sanson 2003, Lucas 2004, Polly 2015, Ungar 2010).

However, structural and genetic limitations – such as phylogenetic history — may prevent the evolution of functionally-perfect dentitions.

A major obstacle is the fact that many taxonomy-free methods previously used in the study of dental morphology discriminate along a particular diet dimension. To cite some examples, hypsodonty correlates with the abrasiveness of diets (e.g. browsing or grazing) in usually open and dry landscapes (Fortelius and Solounias 2000); while

OPCR, a dental complexity index, discriminates well between diets with high or low mechanical processing requirements, such as herbivory and carnivory respectively

(Evans et al. 2007). It is crucial to choose a dietary proxy that covers at least as many dimensions as diet categories we need to discriminate. Thus, a single dietary proxy can be used to reconstruct the diet of highly specialized animal clades such as bovids, which are mostly browsers, grazers or intermediate feeders (Gagnon and Chew 2000), but a more complex methodology will be required to discriminate between more variable diets such as those of primates (Hawes and Peres 2014). A recent study (Pineda-Munoz

98 Silvia Pineda-Munoz - April 2016

Chapter 6: dental morphology predicts diet across marsupials and placentals

and Alroy 2014) showed that the diet of terrestrial mammals is far more complex than a traditional herbivore-omnivore-carnivore classification. Thus, a multidimensional morphological method is required to capture the variability of diet across marsupials and placentals.

We have designed an approach called multi-proxy dental morphology analysis

(MPDMA) to capture the phenotypic variability of mammalian dentitions. This method can be used to draw dietary inferences or answer macroevolutionary questions such as how morphology evolves in response to environmental perturbations. For functional consistency, we chose only adult specimens with a wear level that optimizes chewing processing (See methods and Supplementary Information 4, Dietary classification)

(Evans et al. 2007). The sample included 138 species: 28 marsupials (order

Diprotodontia) and 110 placentals (32 Carnivora, 14 Primates and 64 Rodentia). We qualitatively classified the diets of the species in the dataset following a classification scheme that emphasizes the primary resource in a given diet. The sample represented eight different feeding strategies: herbivory, carnivory, frugivory, granivory, insectivory, fungivory, gumivory and generalized.

Upper and lower post-canine teeth were digitized using either high-resolution laser or micro-CT scanners. Because we hypothesized that dentitions evolve to add extra features in the remaining teeth to compensate for tooth loss, all post-canine teeth were studied instead of only one or a few. Following previous studies (Evans et al. 2007), all tooth rows were scaled to the same number of data rows. Upper and lower rows were converted to GIS topographic maps and six variables were measured: OPCR of the overall tooth row; mean and standard deviation of OPCR between the individual teeth in a tooth row; relief index; and mean and standard deviation of topographical slope (see

Silvia Pineda-Munoz - April 2016 99

Diet, ecology, and dental morphology in terrestrial mammals

Methods and Figure 6.2-1). The combination of upper and lower tooth rows provides the overall set of tools for mastication and processing of the foods (Evans and Sanson

2003, Polly 2004, Polly 2015). Thus, we recorded six variables for both upper and lower tooth rows per specimen. Multivariate statistical analyses were performed in order to define ecomorphospaces and test for discriminatory power.

Figure 6.1-1: Dental and dietary diversity and convergence in marsupials and placentals. Three-dimensional occlusal reconstructions of two marsupial upper tooth rows (top left, common brushtail possum, Trichosurus vulpecula; bottom left, stripped possum, Dactylopsia trivirgata) and two placental upper tooth rows (top right, western gorilla, Gorilla gorilla; bottom right, Serra do Mar grass mouse, Akodon serrensis). Surface orientation (as indicated in each colour wheel) and slope values (green-yellow-red gradient from 0 to 90º as in slope chart) are two of the variables obtained for each specimen and are compared between herbivorous and insectivorous diets. Each diet category has its own slope chart. Upper left tooth rows for all species except Akodon serrensis, which is a reflected upper right tooth row; anterior towards the left. Scale bars, 1 mm.

100 Silvia Pineda-Munoz - April 2016

Chapter 6: dental morphology predicts diet across marsupials and placentals

The first two principal components in the PCA plot for all the species in the dataset differentiate well between herbivory, carnivory and gumivory (Table 11.4-3).

MANOVA analysis suggests significant morphological differences between diet categories (P < 0.05: see Table 6.2-1). Additionally, 66% of the species are correctly classified in the discriminant analysis (Table 6.2-1). When applying MPDMA to individual orders (Carnivora, Diprotodontia, Primates and Rodentia) we also found significant morphological differences across diets (P < 0.05) in all the datasets with the exception of rodents (P = 0.321; see Table 6.2-1 and Figure 11.4-2). Additionally, rodents are responsible for 60% of the misclassifications in the discriminant analysis.

Table 6.1-1: MANOVA and discriminant analysis results for seven datasets. Both tests were carried out by using eight dietary categories: herbivory, carnivory, frugivory, granivory, insectivory, fungivory, gumivory and generalized. MANOVA was performed using the test Pillai. Discriminant analysis values indicate the proportion of species properly assigned in each dataset.

MANOVA Discriminant Dataset analysis (P) analysis All data including rodents <0.05 0.630 All data excluding rodents <0.05 0.811 Common diets <0.05 0.796 Rodents 0.321 0.500 Carnivorans <0.05 0.969 Diprotodontia <0.05 1.000 Primates <0.05 0.857

Thus, we performed the same analysis excluding rodents from the dataset. Doing so greatly increases discrimination power (81% correct classification for discriminant analysis; Table 6.2-1), with carnivory, herbivory, insectivory, frugivory and gumivory much better differentiated from other diet groups (Figure 11.4-2). Additionally, rodents are the only order in this study for which dental morphology is not significantly

Silvia Pineda-Munoz - April 2016 101

Diet, ecology, and dental morphology in terrestrial mammals

different across dietary specializations (see Table 6.2-1 and Figure 11.4-2), which could be related to their generalist feeding nature. Retaining an unspecialized might allow rodents to quickly adapt to new niches and could be the reason for their successful diversification. Further research on the order Rodentia will help to elucidate the reasons for this pattern.

In order to test for convergent phenotypic evolution, we performed the same analysis on a dataset including only species with common dietary specializations across placentals

(excluding rodents) and marsupials (27 and 31 species respectively). These were herbivory, frugivory, gumivory, insectivory and generalized. MPDMA predicted diet with 78% of the species correctly classified. The most remarkable result, however, is the overlap in ecomorphospaces observed between these two distantly-related mammalian infraclasses, which suggests that they have both evolved similar suites of morphological adaptations independently (Figure 6.2-2). Although morphology seems to be a good dietary predictor, a few species were misclassified across the analyses (see

Table 11-4.6). Historical constraints could be the reason behind these improper classifications (Evans and Sanson 2003, Polly 2015, Smith et al. 1985), such as belonging to a highly specialized order or feeding on a generalized diet while displaying a rather specialized phenotype (Liem’s Paradox (Liem 1980)). To cite a few examples, the slender mongoose (Galerella sanguinea), a carnivoran that mainly feed on insects

(Nowak 2005), is classified as carnivorous when all the dietary classifications are analysed because of its morphological similarity to highly specialized carnivores

(Goswami and Friscia 2010). Additionally, this species plots away from the general ecomorphospace in the convergent evolution test with common dietary specializations

(Figure 6.2-2). The giant panda (Ailuropoda melanoleuca), a carnivoran with an

102 Silvia Pineda-Munoz - April 2016

Chapter 6: dental morphology predicts diet across marsupials and placentals

extremely specialized herbivorous diet (Nowak 1999), also occupies a singular position in the ecomorphospace. This suggests that historical and functional constraints might have forced giant pandas to cope with the mechanical needs of a pure bamboo diet by using a different strategy than those employed by other herbivores. Future work combining MPDMA with phylogenetic analysis could provide a deeper understanding of the evolution of dental morphology in mammals.

Figure 6.1-2: Diversity in dental morphology using MPDMA. Plots of the two first components of the principal component analysis of MPDMA data for 83 placental and marsupial species (A- B) and for 53 species including only dietary specializations common to both marsupials (open circles) and placentals (closed circles) (C-D). Plots illustrate differences in primary diet and

Silvia Pineda-Munoz - April 2016 103

Diet, ecology, and dental morphology in terrestrial mammals

phylogeny (A, C); and phylogeny at the infraorder level (placentals vs. marsupials) (B, D). Polygon colours as shown in legends.

Despite marsupials and placentals having split during the mid- to late Mesozoic

(O'Leary et al. 2013), phylogenetic and biomechanical differences (e.g., non- homologous dental structures and even different chewing directions) have not prevented them from converging in the MPDMA ecomorphospace (Figure 6.2-2). Thus, our results agree with the hypothesis that diet specialization drives dental morphology evolution towards functional optima, an idea that has been discussed for more than

2,000 years (Aristotle 1983, Cuvier 1825, Owen 1845).

6.2 Materials and Methods:

6.2.1 Specimens The 138 specimens (28 Diprotodontia, 32 Carnivora, 14 Primates and 64 Rodentia) were obtained from Monash University Zoology Research Collection (MUZ), Museum

Victoria (NMV), Harvard Museum of Comparative Zoology (MCZ) and Finnish

Museum of Natural History (LUOMUS). Only carnivoran specimens with light wear were chosen. Primates and rodents were chosen to display light to moderate wear

(Evans et al. 2007). Some diprotodontians have extremely abrasive diets that sometimes wear their most anterior teeth down before the last molar has occluded. Thus, we chose diprotodontian specimens that displayed light to moderate wear, with all molars at the occluding stage.

6.2.2 Data collection Specimens were scanned using either a laser scanner (Laser Design DS 2025 3D scanner with a RPS-120 probe (Laser Design Inc., Minneapolis, MN) or Nextec Hawk

104 Silvia Pineda-Munoz - April 2016

Chapter 6: dental morphology predicts diet across marsupials and placentals

(Evans et al. 2007) (Nextec details) or a micro-CT (Skyscan 1174 Compact Micro-Ct

(Bruker micro-CT) or XRA-002 X-Tek MicroCT). Previous work suggested that the exact nature of the machine used for three-dimensional scans does not affect morphological results (Wilson et al. 2012). Depending on the size of the tooth row, scanning resolution was set between 10 and 50 µm. Tooth crown orientation followed

Evans et al. (2007).

Three-dimensional files were processed using either Surfer for Windows (Golden

Software, Inc.) or Geomagic 12 (Geomagic Inc., North Carolina, USA 2010) in order to obtain topographic maps. Data standardization followed Evans et al. (2007) for whole tooth rows. Individual teeth were scaled to 50 data rows. The six variables mentioned previously were measured from the 3D scans for each tooth row using Surfer

Manipulator (Evans 2011). OPCR is described in Evans et al. (2007). Av OPCR and sd

OPCR correspond respectively to the average and standard deviation of the number of patches of each individual tooth in a tooth row. The relief index estimates the ratio of three-dimensional surface area to the two-dimensional base area (Boyer 2008, Lazzari et al. 2008b). For the slope variables we measured the steepness of the tooth surface by computing the slope of every grid point on the tooth surface and from this obtained a mean and standard deviation value for the whole tooth row.

6.2.3 Dietary classification We qualitatively classified the diets of all species in the dataset following the classification scheme proposed by Pineda-Munoz and Alroy (2014), which emphasizes the primary resource in a given diet. Thus, a species was classified as a specialist if a single food resource made up 50% or more of its diet. The specializations in our dataset were browsing (mainly feeding on leaves), grazing (mainly feeding on grasses), mixed

Silvia Pineda-Munoz - April 2016 105

Diet, ecology, and dental morphology in terrestrial mammals

feeding (a combination of both), carnivory, frugivory, granivory, insectivory, fungivory, gumivory and generalisation (i.e., no food resource makes up 50% or more of the diet).

Dietary data was extracted from primary literature (Nowak 1999, Wilman et al. 2014)

(Table 11.3-1 and 11.3-2).

6.2.4 Data analysis The R environment (R Core Team 2013) was used for statistical analysis and construction of tables and figures (packages “psych”, “mass” and “stats”). Each data point consisted of 12 variables, 6 for the upper tooth row and 6 for the lower. Data were square-root transformed because variables covered different orders of magnitudes and did not follow a normal distribution in some cases. We performed principal component analyses and discriminant analyses using seven different datasets: (a) all the specimens together, (b) all the specimens excluding rodents, (c) common diets between marsupial and placental species in the dataset and (d-e) each individual order. MANOVA analysis was performed on the first 5 principal components, which together explain 87% of the variation (Table 6.2-1, Table 11.4-3- Table 11.4-8).

6.3 Acknowledgments: We thank Lap Chieu for assistance with 3D scanning. We also thank Kay

Behrensmeyer, Graeme Lloyd, Kate Lyons, Julieta Martinelli, Yuri Kimura and colleagues from Macquarie University and National Museum of Natural History

(Smithsonian Institution) for discussion, comments and suggestions. SP-M was supported by a Macquarie University HDR Project Support Fund and a Smithsonian

Predoctoral Fellowship. AE acknowledges the support of the Australian Research

Council FT130100968 and Monash University.

106 Silvia Pineda-Munoz - April 2016

Chapter 6: dental morphology predicts diet across marsupials and placentals

6.4 Author’s contributions All authors designed the study. SP-M and AE collected the data. SP-M analysed the data, and drafted the manuscript. AE created the GIS maps and helped interpret the data and draft the manuscript. JA assisted data analysis and interpretation and drafting of the manuscript. All authors gave final approval for publication.

Silvia Pineda-Munoz - April 2016 107

Diet, ecology, and dental morphology in terrestrial mammals

7 DENTAL MORPHOLOGY VA RIA BILITY IN RELATION TO DIET IN TERRESTRIAL MAMMALS

Authors: Pineda-Munoz S, J. Alroy and A.R. Evans.

108 Silvia Pineda-Munoz - April 2016

Chapter 7: Dental morphology variability in relation to diet in terrestrial mammals

7.1 Abstract Dietary inferences are a key foundation for palaeoecological, ecomorphological and macroevolutionary studies because they inform us about the direct relationships between the components of an ecosystem. However, we need to consider the range of dietary variation we want to investigate and characterize before choosing a proxy. The goal of the present work is to evaluate the differences in dietary discrimination power between unidimensional dental morphology proxies such as orientation patch count

(OPCR), relief index (RI) or slope and the recently described multidimensional multi- proxy dental morphology analysis (MPDMA). In order to do that, we three- dimensionally scanned the dentitions of 138 extant mammals including 28 marsupials

(order Diprotodontia) and 110 placentals (orders Carnivora, Primates and Rodentia). A recent study observed that diet is far more complex than a traditional herbivore- omnivore-carnivore classification and proposed a new classification scheme that emphasizes the primary resource in a given diet. Unidimensional proxies significantly discriminate (P < 0.05) between one or two diet categories on one hand and the rest on the other. For example, OPCR discriminates well between carnivorous and non- carnivorous species. However, none of the individual proxies discriminate all categories. MPDMA demonstrates significant morphological differences across diets (P

< 0.05) and correctly discriminates diet for 66 to 82% of the specimens in the dataset including and excluding rodents respectively. Combining different morphological variables makes it possible to draw better dietary inferences and fully represent the multidimensional nature of dietary specializations.

Silvia Pineda-Munoz - April 2016 109

Diet, ecology, and dental morphology in terrestrial mammals

Keywords: Mammal diet, ecomorphology, dental morphology, orientation patch count, relief index, dental slope, multi-proxy dental morphology analysis

7.2 Introduction Diet is one of the most studied variables in ecological and paleoecological research.

Reconstructing the trophic relationships between the components of an ecosystem helps to elucidate the functioning of present and past ecosystems (Dalerum and Angerbjörn

2005, Feranec 2004). Many proxies for diet have been described during the last few decades: dental microwear and mesowear, isotope analysis, dental and skull morphology, etc. In order to evaluate their reliability, the different methodologies have been tested in modern species having well-known dietary habits (Fortelius and

Solounias 2000, Merwe 1982, Teaford and Oyen 1989, Ungar 2010).

Previous research has recognized that each proxy is informative about a different temporal scale. Methods used by palaeontologists usually span many temporal scales: from weeks to millions of years. Davis and Pineda-Munoz (in review) approached this question and showed the importance of making sure that the scale of the proxy being used matches the scale of the question being asked. For example, dental microwear could be informative about the diet of the last three to four weeks of an animal’s lifespan while dental morphology could illustrate the diet a species has adapted to eating over millions of years of evolution (Fortelius and Solounias 2000, Levin 1995,

Sponheimer et al. 2003). Thus, dental microwear studies could for example reveal seasonal changes in the environment (DeMiguel et al. 2011) and dental morphology could be used to study the effects of long-term climate change (Davis and Pineda-

Munoz in review).

110 Silvia Pineda-Munoz - April 2016

Chapter 7: Dental morphology variability in relation to diet in terrestrial mammals

Meanwhile, each proxy will discriminate within a particular dimension of diet. To cite some examples, dental mesowear reflects the abrasiveness of the items being consumed

(Fortelius and Solounias 2000) while isotope analysis discriminates between the proportion of C3 and C4 plants in the diet (Cerling et al. 2009). Thus, it is important to choose a dietary proxy with at least as many dimensions as diets we need to discriminate. Additionally, the parameters of our studies will change depending on the taxon being studied. We can use a single dietary proxy when reconstructing the diet of highly specialized animal clades (e.g. members of family Bovidae being mostly browsers or grazers (Gagnon and Chew 2000)]) but a more complex methodology will be required when discriminating more variable diets (e.g. order Primates [(Hawes and

Peres 2014)]). Pineda-Munoz and Alroy (2014) showed that the diet of terrestrial mammals is far more complex than a traditional herbivore- omnivore-carnivore classification, which also suggests that a multidimensional ecometric method will be necessary to capture the variability of diet and dietary morphospaces in the whole mammalian clade.

Previous literature showed that mammals belonging to independent evolutionary lineages that have the same diet preferences have evolved morphologically similar dentitions (Evans et al. 2007). Our main goal here is to design a quantitative, phylogeny-free method to infer the typical diet of a species that can also be applied to macroevolutionary questions such as how morphology responds evolutionarily to environmental perturbations on geological time scales. Therefore, we developed an approach called multi-proxy dental morphology analysis (MPDMA) (Pineda-Munoz et al. In review). MPDMA combines 6 morphological variables extracted from 3D scans of tooth rows and offers enough multidimensionality to discriminate the range of dietary

Silvia Pineda-Munoz - April 2016 111

Diet, ecology, and dental morphology in terrestrial mammals

specializations observed in mammals. Additionally, the method showed how marsupials and placentals with the same dietary specializations converge in the same region of ecomorphospace despite having diverged at least during the late Jurassic-early

Cretaceous (O'Leary et al. 2013).

In the present work we more deeply examine MPDMA, evaluating the discriminant power of each individual variable and the benefits of combining them in a single multivariate analysis.

7.3 Materials and Methods The data include upper and lower post-canine teeth from 138 specimens: 28 marsupials

(order Diprotodontia) and 110 placentals (14 Primates, 32 Carnivora and 64 Rodentia).

Only specimens with light wear were chosen when it came to carnivorans. Primates and rodents were chosen to display light to moderate wear (Evans et al. 2007). Some diprotodontids have extremely abrasive diets that sometimes wear the most anterior teeth down completely before the last molar has occluded. Thus, we choose diprotodontian specimens that displayed light to moderate wear with all molars at the occluding stage.

We qualitatively classified the diets of all species in the dataset following the classification scheme proposed by Pineda-Munoz and Alroy (2014), which emphasizes the primary resource in a given diet. Thus, a species was classified as a specialist if a single food resource made up 50% or more of its diet. The specializations in our dataset were herbivory, carnivory, frugivory, granivory, insectivory, fungivory, gumivory and generalisation (i.e., no food resource makes up 50% or more of the diet). Dietary data

112 Silvia Pineda-Munoz - April 2016

Chapter 7: Dental morphology variability in relation to diet in terrestrial mammals

were extracted from primary literature (Nowak 1999, Wilman et al. 2014) (Table 11.4-1 and 11.4-2).

Specimens were scanned using either a laser scanner (Laser Design DS 2025 3D scanner with a RPS-120 probe (Laser Design Inc., Minneapolis, MN) or Nextec Hawk

(Evans et al. 2007) (Nextec details) or a micro-CT (Skyscan 1174 Compact Micro-Ct

[Bruker micro-CT] or XRA-002 X-Tek MicroCT). Previous work suggested that the exact nature of the machine used for three-dimensional scans does not affect morphological results (Wilson et al. 2012). Depending on the size of the tooth row, scanning resolution was set between 10 and 50 µm. Tooth crown orientation followed

Evans et al. (2007).

Three-dimensional files were processed using either Surfer for Windows (Golden

Software, Inc.) or Geomagic 12 (Geomagic Inc., North Carolina, USA 2010) in order to obtain topographic maps. Data standardization followed Evans et al. (2007): whole tooth rows were scaled to 50xN data rows, where N is the number of teeth in the tooth row. Therefore, a tooth row with 3 teeth would be scaled to 150 data rows. Individual teeth were scaled to 50 data rows. Six morphological variables were measured from the

3D scans for each tooth row using Surfer Manipulator (Evans 2011). They will be described in more detail in the following sections. The R environment (R Core Team

2013) was used for statistical analysis and construction of tables and figures (packages

“psych”, “mass” and “stats”). We used Surfer Manipulator software (Evans 2011) to estimate the variables of the study. In order to test for the discrimination power of each individual variable, ANOVA analysis and pairwise Wilcox testing in relation to diet specialization was performed for each one. The correlation between upper and lower tooth rows in an individual was calculated for each variable.

Silvia Pineda-Munoz - April 2016 113

Diet, ecology, and dental morphology in terrestrial mammals

7.4 Dental morphology proxies: results and discussion

7.4.1 Orientation Patch Count (OPCR) Orientation patch count (OPCR) is a homology-free quantitative ecometric proxy that quantifies dental complexity (Evans et al. 2007). It is based on the idea that animals feeding on more complex foods will evolve more complex dentitions. In the process, teeth are 3-dimensionally scanned and converted into surface maps. Topographic maps are created by representing the orientation of every grid point as being one in eight compass directions. Contiguous points with the same orientation form a patch and the number of patches is proportional to the complexity of the molar surface. This method has proven to be a good taxon-free proxy for inferring diet and the phenotypic evolution of extant and extinct species, and even computer simulated tooth models (Evans et al.

2007, Salazar-Ciudad and Marin-Riera 2013, Wilson et al. 2012).

We computed OPCR for all post-canine teeth as a whole and for each individual tooth independently for both upper and lower tooth rows. This allowed calculating the average and standard deviation of the number of patches of each individual tooth in a tooth row. All variables (overall OPCR, average OPCR between individual teeth

[avOPCR] and standard deviation of OPCR between individual teeth [sdOPCR]) were tested against dietary specializations.

114 Silvia Pineda-Munoz - April 2016

Chapter 7: Dental morphology variability in relation to diet in terrestrial mammals

Figure 7.4-1: Boxplot illustrating the range of overall OPCR for each dietary category for lower and upper tooth rows of the 138 mammal species in the dataset. Diet categories are as described by Pineda-Munoz and Alroy (2014).

Table 7.4-1: P-values of pairwise comparisons based on the Wilcoxon rank-sum test for overall OPCR in diet categories as described by Pineda-Munoz & Alroy (2014) for (a) lower tooth rows and (b) upper tooth rows

a) Lower tooth rows Carnivore Frugivore Fungivore Generalist Granivore Gumivore Herbivore Frugivore <0.05 ------Fungivore 1 1 - - - - - Generalist <0.05 1 1 - - - - Granivore <0.05 1 1 1 - - - Gumivore <0.05 1 1 1 1 - - Herbivore <0.05 1 1 1 1 1 - Insectivore <0.05 0.05648 1 1 1 1 <0.05

b) Upper tooth rows

Carnivore Frugivore Fungivore Generalist Granivore Gumivore Herbivore Frugivore <0.05 ------Fungivore 1 1 - - - - - Generalist <0.05 1 1 - - - - Granivore <0.05 1 1 1 - - - Gumivore <0.05 1 1 1 1 - - Herbivore <0.05 1 1 0.3582 1 1 - Insectivore <0.05 1 1 1 1 1 <0.05

Silvia Pineda-Munoz - April 2016 115

Diet, ecology, and dental morphology in terrestrial mammals

The carnivoran Acinonyx jubatus displays the lowest OPCR values for both upper and lower tooth rows, while the primates Alouatta seniculus and Gorilla gorilla have the highest values for upper and lower tooth rows respectively (Figure 7.4-1, Tables 7.4-1,

11.1-1 and 11.1-2). Overall, OPCR significantly discriminates between carnivores and all other dietary specializations except for generalization in both upper and lower tooth rows. Carnivores have the lower OPCR values, as pointed out in previous literature

(Evans et al. 2007), which could be related to the less complex morphology of secodont dentitions (Ungar 2010). An almost conical morphology displays the lowest OPCR values and is much more efficient for puncturing flesh and tearing it apart. On the other hand, herbivore species require more complex dentitions. Having extra “tools” on the tooth surface facilitates crushing leaves and other highly fibrous plant structures. Since these foods are not easily digested, greater mechanical processing prior to swallowing accelerates the metabolic breakdown of cellulose (Ungar 2010).

Figure 7.4-2: Boxplot illustrating the range of average OPCR between individual teeth for each dietary category for lower and upper tooth rows of the 138 mammal species in the dataset. Diet categories are as described by Pineda-Munoz and Alroy (2014).

116 Silvia Pineda-Munoz - April 2016

Chapter 7: Dental morphology variability in relation to diet in terrestrial mammals

Table 7.4-2: P-values of pairwise comparisons based on the Wilcoxon rank-sum test for average OPCR between individual teeth in diet categories as described by Pineda-Munoz & Alroy (2014) for (a) lower tooth rows and (b) upper tooth rows

a) Lower tooth rows

Carnivore Frugivore Fungivore Generalist Granivore Gumivore Herbivore

Frugivore <0.05 ------Fungivore 1 1 - - - - - Generalist <0.05 0.95151 1 - - - - Granivore <0.05 0.11212 1 1 - - - Gumivore 0.74975 0.08338 1 <0.05 0.08858 - - Herbivore <0.05 1 1 1 0.74975 <0.05 - Insectivore <0.05 1 1 1 0.18967 <0.05 1

b) Upper tooth rows

Carnivore Frugivore Fungivore Generalist Granivore Gumivore Herbivore

Frugivore <0.05 ------Fungivore 1 1 - - - - - Generalist <0.05 1 1 - - - - Granivore <0.05 0.1308 1 1 - - - Gumivore 1 0.47986 1 0.30272 0.15984 - - Herbivore <0.05 0.32057 1 1 1 <0.05 - Insectivore <0.05 1 1 1 0.70686 0.11033 1

The minimum and maximum values for the average OPCR between individual teeth are both found in carnivoran species (Figure 7.4-2 and Table 7.4-2). The mustelid Mustela erminea and the felid Panthera leo display the minimum values of average OPCR between individual teeth for the lower and upper tooth rows respectively (see Tables

11.1-1 and 11.1-2). The ursid Ailuropoda melanoleuca and the rodent Calomys sp. have the highest values of average OPCR for lower and upper tooth rows respectively.

Average OPCR significantly differentiates between carnivore and gumivore dentitions, with the lowest values, and the rest of dietary specializations, with generally higher values. Therefore, herbivores and granivores display the most complex dentitions,

Silvia Pineda-Munoz - April 2016 117

Diet, ecology, and dental morphology in terrestrial mammals

which could be related to higher mechanical processing requirements for cellulose breakdown or crushing hard seeds.

The standard deviation of OPCR among individual teeth discriminates well between carnivorous species and frugivores, generalist feeders, gumivores and herbivores with respect to the lower tooth row (Figure 7.4-3 and Table 7.4-3). The minimum values are found in such as Crocuta crocuta and Panthera leo. This fact can be related to the oversimplification of their dentitions. By contrast, the herbivore

Ailuropoda melanoleuca has the highest values, where complexity is higher in its most posterior teeth (M1 = 46, M2 = 98, M3 = 127, Table 11.1-1), which leads to the standard deviation being higher. This pattern could be related to the abrasiveness of a pure herbivore diet. By the time the last molar (M3) has fully occluded, meaning when the dentition is optimal for morphological studies, the first molars will already be worn down. Abrasiveness will therefore reduce complexity while flattening molar surfaces over time. This phenomenon can be also observed in other herbivore species such as kangaroos (Ungar 2010).

118 Silvia Pineda-Munoz - April 2016

Chapter 7: Dental morphology variability in relation to diet in terrestrial mammals

Figure 7.4-3: Boxplot illustrating the range of standard deviation of OPCR between individual teeth for each dietary category for lower and upper tooth rows of the 138 mammal species in the dataset. Diet categories are as described by Pineda-Munoz and Alroy (2014).

Table 7.4-3: P-values of pairwise comparisons based on the Wilcoxon rank-sum test for standard deviation of OPCR between individual teeth in diet categories as described by Pineda- Munoz & Alroy (2014) for (a) lower tooth rows and (b) upper tooth rows

a) Lower tooth rows

Carnivore Frugivore Fungivore Generalist Granivore Gumivore Herbivore Frugivore <0.05 ------Fungivore 1 1 - - - - - Generalist <0.05 1 1 - - - - Granivore 0.23957 1 1 1 - - - Gumivore <0.05 1 1 1 1 - - Herbivore <0.05 1 1 1 1 1 - Insectivore 0.15558 0.36392 1 0.13285 1 0.16426 <0.05

b) Upper tooth rows

Carnivore Frugivore Fungivore Generalist Granivore Gumivore Herbivore Frugivore 1 ------Fungivore 1 1 - - - - - Generalist 1 1 1 - - - - Granivore 1 1 1 1 - - - Gumivore 0.216 <0.05 1 0.446 0.216 - - Herbivore 1 0.991 1 1 1 0.342 - Insectivore 1 1 1 1 1 0.287 1

The standard deviation between individual teeth OPCR for the upper tooth row seems to provide very little discriminant power (ANOVA = 0.105). Gumivores and frugivores are the only two dietary specializations that are significantly discriminated (Figure 7.4-3 and Table 7.4-3). Higher and lower values are also found in and herbivore species respectively (Ailuropoda melanoleuca and Acinonyx jubatus).

Silvia Pineda-Munoz - April 2016 119

Diet, ecology, and dental morphology in terrestrial mammals

7.4.2 Relief Index (RI) The relief index (RI) is another homology-free quantitative ecometric proxy that estimates the steepness of the tooth surface (Boyer 2008, Lazzari et al. 2008b). It calculates the ratio of the three-dimensional surface area to the two-dimensional base area, with the prediction being that steeper dentitions will display higher ratios. Surface area is calculated from 3-dimensional scans of all post-canine teeth.

In the lower tooth rows, the lowest RI values are found in the gumivore primate alleni and the highest values in the hypercarnivore Panthera leo (Table 11.1 1). The carnivoran Mustela erminea and the granivore rodent Phyllotis sp. display the highest and lowest values respectively in the upper tooth rows (Table 11.1 2). Lower tooth rows discriminate well between gumivore species and all other dietary categories except for frugivores. In addition, carnivore species are also significantly discriminated in the upper tooth rows (Figure 7.4-4 and Table 7.4-4).

Figure 7.4-4: Boxplot illustrating the range of relief index for each dietary category for lower and upper tooth rows of the 138 mammal species in the dataset. Diet categories are as described by Pineda-Munoz and Alroy (2014).

120 Silvia Pineda-Munoz - April 2016

Chapter 7: Dental morphology variability in relation to diet in terrestrial mammals

Table 7.4-4: P-values of pairwise comparisons based on the Wilcoxon rank-sum test for relief index in diet categories as described by Pineda-Munoz & Alroy (2014) for (a) lower tooth rows and (b) upper tooth rows

a) Lower tooth rows

Carnivore Frugivore Fungivore Generalist Granivore Gumivore Herbivore Frugivore 0.8247 ------Fungivore 1 1 - - - - - Generalist 1 0.4599 1 - - - - Granivore 1 0.6146 1 1 - - - Gumivore <0.05 0.299 1 <0.05 <0.05 - - Herbivore 1 1 1 1 1 <0.05 - Insectivore 1 1 1 1 1 <0.05 1

b) Upper tooth rows

Carnivore Frugivore Fungivore Generalist Granivore Gumivore Herbivore Frugivore <0.05 ------Fungivore 1 1 - - - - - Generalist <0.05 1 1 - - - - Granivore <0.05 1 1 1 - - - Gumivore 1 0.10122 1 0.08754 0.09324 - - Herbivore <0.05 1 1 1 1 <0.05 - Insectivore <0.05 1 1 1 1 <0.05 1

The RI could be related to the number of surfaces required to mechanically process different foods. Gumivores have low values because soft plant tissues such as gum and nectar do not require much mastication. To the contrary, herbivore and granivore species have higher values because they require more surface area to mechanically process more fibrous or harder tissues such as cellulose or seeds. The differences observed between upper and lower tooth rows will be discussed in the following sections.

Silvia Pineda-Munoz - April 2016 121

Diet, ecology, and dental morphology in terrestrial mammals

7.4.3 Dental Slope For the slope variables we measured the steepness of the tooth surface by computing the slope of every grid point on the tooth surface from 0 to 90˚. We then obtained the mean

(avSlope) and standard deviation (sdSlope) value for the whole tooth row.

Figure 7.4-5: Boxplot illustrating the range of average slope for each dietary category for lower and upper tooth rows of the 138 mammal species in the dataset. Diet categories are as described by Pineda-Munoz and Alroy (2014).

Table 7.4-5: P-values of pairwise comparisons based on the Wilcoxon rank-sum test for average slope in diet categories as described by Pineda-Munoz & Alroy (2014) for (a) lower tooth rows and (b) upper tooth rows.

a) Lower tooth rows

Carnivore Frugivore Fungivore Generalist Granivore Gumivore Herbivore Frugivore <0.05 ------Fungivore 1 1 - - - - - Generalist <0.05 1 1 - - - - Granivore <0.05 0.3146 1 1 - - - Gumivore <0.05 1 1 1 1 - - Herbivore <0.05 1 1 1 1 1 - Insectivore <0.05 1 1 1 1 1 1

b) Upper tooth rows

122 Silvia Pineda-Munoz - April 2016

Chapter 7: Dental morphology variability in relation to diet in terrestrial mammals

Carnivore Frugivore Fungivore Generalist Granivore Gumivore Herbivore Frugivore <0.05 ------Fungivore 1 1 - - - - - Generalist <0.05 1 1 - - - - Granivore <0.05 1 1 1 - - - Gumivore 0.062 1 1 1 1 - - Herbivore <0.05 1 1 1 1 1 - Insectivore <0.05 1 1 1 1 1 1

The rodent Anisomys imitator has the lowest values of avSlope for both upper and lower tooth rows (Figure 7.4-5 and Table 7.4-5). As in the case of the RI, a gumivore diet requires less “tools” on the tooth surface, where teeth can be almost flat without affecting the effectiveness of digesting gums and nectar. The hypercarnivores Panthera leo and Crocuta crocuta have the higher values for avSlope for lower and upper tooth rows respectively. Additionally, avSlope mainly discriminates between carnivore dentitions and all the rest of the dietary specializations. Secodont carnivore dentitions will therefore display steeper teeth, ideal for puncturing and shearing flesh.

Figure 7.4-6: Boxplot illustrating the range of standard deviation of the slope for each dietary category for lower and upper tooth rows of the 138 mammal species in the dataset. Diet categories are as described by Pineda-Munoz and Alroy (2014).

Silvia Pineda-Munoz - April 2016 123

Diet, ecology, and dental morphology in terrestrial mammals

Table 7.4-6: P-values of pairwise comparisons based on the Wilcoxon rank-sum test for standard deviation of the slope in diet categories as described by Pineda-Munoz & Alroy (2014) for (a) lower tooth rows and (b) upper tooth rows

a) Lower tooth rows

Carnivore Frugivore Fungivore Generalist Granivore Gumivore Herbivore Frugivore 1 ------Fungivore 1 1 - - - - - Generalist 1 1 1 - - - - Granivore 1 1 1 1 - - - Gumivore 1 1 1 1 1 - - Herbivore 0.59 1 1 1 1 1 - Insectivore 1 1 1 1 1 1 1

b) Upper tooth rows

Carnivore Frugivore Fungivore Generalist Granivore Gumivore Herbivore Frugivore 1 ------Fungivore 1 1 - - - - - Generalist 1 1 1 - - - - Granivore 1 1 1 1 - - - Gumivore <0.05 <0.05 1 <0.05 1 - - Herbivore 1 1 1 1 1 <0.05 - Insectivore 1 1 1 1 1 <0.05 1 The ANOVA shows that no significant differences exist across diet specializations for sdSlope for lower tooth rows (p=0.49) (Figure 7.4-6 and Table 7.4-6). In the upper tooth rows, however, gumivores appear to be significantly different from all dietary specializations except for granivores. Gumivore species have the lower values on average, which suggests more homogeneous and simplified tooth morphologies along the upper tooth row.

7.4.4 MPDMA In order to better capture the variability of diet we developed an approach called multi- proxy dental morphology analysis (MPDMA). It is also designed to capture the phenotypic variability of mammalian dentitions in general (Pineda-Munoz et al. In review). To implement this method we created a dataset including all the variables

124 Silvia Pineda-Munoz - April 2016

Chapter 7: Dental morphology variability in relation to diet in terrestrial mammals

described above for the 138 species in the study: 12 variables, 6 for the upper tooth row and 6 for the lower. Data were square-root transformed because variables covered different orders of magnitudes and did not follow a normal distribution in some cases.

We performed MANOVA, principal component analyses, and discriminant analyses using seven different datasets: (1) all the specimens together, (2) all the specimens excluding rodents, (3) diets common to both marsupial and placental species and (4-7) each individual order (Table 1).

The first two principal components for all the species in the dataset differentiate well between herbivory, carnivory, and gumivory (Table 11.1-3 and Figure 11.1-1).

MANOVA analysis suggests that there are significant morphological differences between diet categories (P < 0.05: see Table 7.4-7). Additionally, 63% of the species are correctly classified in the discriminant analysis (Table 7.4-7). When applying

MPDMA to individual orders (Carnivora, Diprotodontia, Primates and Rodentia) we also found significant morphological differences across diets (P < 0.05) in all the datasets with the exception of rodents (P = 0.321; see Table 7.4-7 and Figure 11.1-2).

Additionally, rodents were responsible for 60% of the misclassifications in the discriminant analysis. Thus, we performed the same analysis excluding rodents from the dataset. Doing so greatly increases discrimination power (81% of classifications are correct: Table 7.4-7), with carnivory, herbivory, insectivory, frugivory, and gumivory much better differentiated from other diet groups (Figure 7.4-7).

Silvia Pineda-Munoz - April 2016 125

Diet, ecology, and dental morphology in terrestrial mammals

Figure 7.4-7: Diversity in dental morphology using MPDMA. Plots of the two first components of the principal component analysis of MPDMA data for 83 placental and marsupial species (A- B) and for 53 species including only dietary specializations common to both marsupials (open circles) and placentals (closed circles) (C-D). Plots illustrate differences in primary diet and phylogeny (A, C); and phylogeny at the infraorder level (placentals vs. marsupials) (B, D). Polygon colours as shown in legends.

126 Silvia Pineda-Munoz - April 2016

Chapter 7: Dental morphology variability in relation to diet in terrestrial mammals

Table 7.4-7: MANOVA and discriminant analysis results for seven datasets. Both tests were carried out by using eight dietary categories: herbivory, carnivory, frugivory, granivory, insectivory, fungivory, gumivory and generalization. MANOVA was performed using the test Pillai. Discriminant analysis values indicate the proportion of specimens properly assigned in each dataset.

MANOVA Discriminant Dataset analysis analysis (P) All data including rodents <0.05 0.63 All data excluding rodents <0.05 0.811 Common diets <0.05 0.796 Rodents 0.321 0.5 Carnivorans <0.05 0.969 Diprotodontia <0.05 1 Primates <0.05 0.857

Figure 7.4-8:

7.5 Discussion

7.5.1 Relationship between upper and lower tooth rows Overall OPCR, average OPCR, and average and standard deviation of the slope show high correlations between upper and lower tooth rows (Table 7.5-1). However, this is not the case for sdOPCR and RI, where the correlations are lower than 0.5. Although overall OPCR for upper and lower tooth rows discriminates between the same diet categories as RI for the upper ones, the correlation for RI between uppers and lowers is much lower (Table 7.5-1). This fact suggests that each variable is informative about different aspects of dental morphology.

Silvia Pineda-Munoz - April 2016 127

Diet, ecology, and dental morphology in terrestrial mammals

Table 7.5-1: Linear correlation (R2) between upper and lower tooth rows for the dental morphology variables studied in the present work.

Overall Average Average sd OPCR Relief index sd Slope OPCR OPCR slope 0.874 0.863 0.321 0.482 0.927 0.735

The variables in this study generally infer two main properties: the tools required for chewing and trade-offs in occlusion. Upper and lower dentitions need to evolve in parallel. Failing to do so leads to malocclusion and therefore to reduced fitness (Polly

2015, Polly et al. 2005). Overall and average OPCR variables estimate the number of mechanical tools on the tooth surface. The high correlation observed between these two variables suggests that upper and lower tooth rows avoid malocclusion by developing dental structures almost simultaneously. This is also the case for slope variables.

Dentitions adapted to mechanically crush food items such us grains become flatter, creating more area for food processing on both upper and lower tooth surfaces.

However, this is not the case for RI. In carnivorous species, lower tooth rows have the highest RI values while upper tooth rows have the lowest. This could be related to the shearing morphology required for cutting flesh. Lower teeth require more surface to accommodate bigger, knife-like puncturing structures on upper teeth. Having two big occluding structures would prevent the mouth from closing. Additionally, lower dentitions also become the surface cross which meat is cut. This relationship provides a good example of the trade-offs that might constrain the evolution of dental morphology

(Polly 2015).

128 Silvia Pineda-Munoz - April 2016

Chapter 7: Dental morphology variability in relation to diet in terrestrial mammals

7.5.2 The benefits of mixing different dietary proxies As discussed above, individual variables discriminate mostly between carnivore and gumivore species and all other dietary specializations, and also between herbivores and insectivores (Figures 7.4-1 to 7.4-6). Each individual variable will be better at discriminating between a particular pair of diets. For example, gumivores and frugivores display significantly different values of sdOPCR but frugivory is the only diet category that is not significantly different from gumivory according to RI values.

Thus, simply by mixing two variables in the same study one could add some important discriminatory power that would be otherwise lost.

Some of the dietary categories are not properly discriminated by any of the individual proxies. However, the principal component analysis (Figure 7.4-7) and discriminant analysis (Table 7.4-7) show how combining 12 different morphological variables highlights morphological differences that would otherwise be overlooked.

Ideally, each dietary category will have its own combination of variables that make it distinctive. However, some species were misclassified in the MPDMA analyses.

Phylogenetic constraints such as belonging to a mammalian order that historically specialized on feeding on a rather specialized resource could be the reason behind most of the misclassifications (Evans and Sanson 2003, Pineda-Munoz et al. In review, Polly

2015, Smith et al. 1985). Additionally, some species might be feeding on foods that are more easily available in a given environment, while having morphological adaptations for a different diet specialization (Liem’s Paradox (Liem 1980)).

Silvia Pineda-Munoz - April 2016 129

Diet, ecology, and dental morphology in terrestrial mammals

7.5.3 Diet categories and dental morphology Because omnivore species can be very distinct (e.g. if they feed on meat and vegetation instead of seeds and insects), Pineda-Munoz and Alroy (2014) proposed avoiding using the classic trophic classification "herbivore-omnivore-carnivore" whenever possible.

This new classification scheme was later tested against body-mass (Pineda-Munoz et al. in press). The proposed classification strongly correlated with body mass in mammals, suggesting that it was robust and conformed well with other ecological variables

(Pineda-Munoz et al. in press). Additionally, it was shown how including (for example) granivores and frugivores in a single dietary category masked the potential of diet as a climatological indicator (Fernández-Hernández et al. 2006). In the present work the same dietary classification scheme has been used to test the relationship between diet and dental morphology, and the results that this dietary classification proposed in correlates well with ecomorphological specializations seen across many mammalian orders.

7.6 Acknowledgments We thank Lap Chieu for assistance with 3D scanning. We also thank Kay

Behrensmeyer, Yuri Kimura, Graeme Lloyd, Kate Lyons, Julieta Martinelli and colleagues from Macquarie University and National Museum of Natural History

(Smithsonian Institution) for discussion, comments, and suggestions. SP-M was supported by a Macquarie University HDR Project Support Fund and a Smithsonian

Predoctoral Fellowship. AE acknowledges the support of the Australian Research

Council FT130100968 and Monash University.

130 Silvia Pineda-Munoz - April 2016

Chapter 7: Dental morphology variability in relation to diet in terrestrial mammals

7.7 Author’s contributions All authors designed the study. SP-M and AE collected the data. SP-M analysed the data, and drafted the manuscript. AE created the GIS maps and helped interpret the data and draft the manuscript. JA assisted data analysis and interpretation and drafting of the manuscript. All authors gave final approval for publication.

Silvia Pineda-Munoz - April 2016 131

Diet, ecology, and dental morphology in terrestrial mammals

8 CONCLUSIONS

“The laws of history are as absolute as the laws of physics, and if the probabilities of

error are greater, it is only because history does not deal with as many as

physics does atoms, so that individual variations count for more.”

― Isaac Asimov, Foundation and Empire

Nature is complex. Our work as biologists, ecologists or palaeontologists is to

understand natural systems and predict future outcomes. Experimental design is always

based on assumptions such as the parameters of the study will stay stable. Because

choosing the wrong assumptions will lead to making less accurate predictions and

generating higher experimental error, it is crucial to carefully evaluate them. And this

matter is what the present thesis is about: evaluating paleoecological and

ecomorphological research protocols to build a stronger foundation for future research

on trophic specialization.

Inferring diet requires working with a solid dietary classification scheme. Diet

preferences of living animals were traditionally used to study mammal morphology,

132 Silvia Pineda-Munoz - April 2016

Chapter 8: Conclusions

ecology, and palaeoecology (Andrews et al. 1979, Mendoza et al. 2005, Reed 1998).

However, the criteria used for classifying diet were far from being standardized. In

Chapter 3 I approached this problem by building and analysing a database summarising the dietary preferences of 139 mammal species on the basis of published stomach content data. The resulting analyses defined a statistically grounded classification scheme based on the main food resource utilised by each species.

Previous research used basic feeding classifications that equated with classic trophic levels – herbivore, carnivore, and omnivore, plus a few variations (Evans et al. 2007,

Reed 1998, Schoener 1989, Ungar 2010). I proposed avoiding the term “omnivore” because it fails to communicate all the complexity inherent to food choice. Furthermore, species traditionally classified in this feeding category can display highly varied diets.

While no one would claim that an animal feeding on meat and vegetation should be classified in the same category as one feeding on seeds and insects, they would still be described as “omnivores” according to the traditional definition. Another important observation stemming from my research was that terrestrial mammals are generally highly specialized. However, some degree of food mixing may be required for most species.

Body mass is a well-studied ecological variable that has been used to explain macroevolutionary and geographical distribution patterns (Alroy 1998, Burness et al.

2001, Smith et al. 2010, Smith and Lyons 2011). Therefore, it provides a unique opportunity to test whether my proposed dietary classification scheme relates in a sensible way to mammal ecomorphological specializations. In order to do so, I compiled the body mass of the 139 mammal species studied in Chapter 3. Overall, I observe that dietary specialisations occupy distinctive body size ranges. Smaller

Silvia Pineda-Munoz - April 2016 133

Diet, ecology, and dental morphology in terrestrial mammals

mammals (<1 kg) are mainly insectivores, granivores, or mixed feeders while bigger animals (>30 kg) are usually either carnivores or herbivores that feed specifically on grasses and leaves.

The most innovative pattern I observed is that the medium size range (1-30 kg) is mostly composed of tropical and subtropical, rainforest-dwelling frugivorous species.

This observation allows to link the near absence of medium-sized mammals in open environments such as savannahs to the lower density of fruit trees required to support the nutritional requirements of pure frugivores. The results in Chapter 4 therefore provide another explanation for the mid-sized gap, a widely discussed topic in palaeoecological literature (Alroy et al. 2000, Gingerich 1989, Legendre 1986, Smith and Lyons 2011, Valverde 1967). Unfortunately, using untested dietary classifications— and specifically classifying granivore and frugivore species in a single category— prevented previous literature from recognising this pattern (Fernández-

Hernández et al. 2006).

Up to this point in my thesis, I had established the diet categories a good dietary proxy should discriminate. I then had to choose a proxy that allowed me to answer my future research questions. To solve this problem, together with my colleague Matthew Davis in Yale University I created a dataset summarising the timespan over which proxies used by palaeontologists and neonatologists (e.g. stable isotopes, stomach contents, and dental microwear) average diet. In Chapter 5 we argued that proxies recording the natural variability of diet on vastly different timescales could explain incongruence in diet inferences. Additionally, using fine-scaled field studies we demonstrated how the diet of a species could be described differently depending on the temporal scale we chose to consider.

134 Silvia Pineda-Munoz - April 2016

Chapter 8: Conclusions

The main goal of my Ph.D. research was to design a method that could be applied to drawing dietary inferences as well as studying the dynamics of morphological evolution. The results in Chapter 5 suggested that morphological studies would better address my research questions because they represent the most relevant temporal scale.

However, taxonomy-free methods previously used in the study of dental morphology only discriminate along a single, particular diet dimension (e.g. hypsodonty and abrasiveness of foods consumed or OPCR and mechanical processing needs) (Evans et al. 2007, Fortelius and Solounias 2000). Thus, I designed a new approach called multi- proxy dental morphology analysis (MPDMA). I three-dimensionally scanned 138 mammal species (32 marsupials and 115 placentals) and qualitatively classified their diets according to my novel dietary classification scheme (Chapter 3, Pineda-Munoz and Alroy, 2014). I then calculated orientation patch count (OPCR), slope diversity, and the relief index from the dental 3D scans (Boyer 2008, Evans et al. 2007, Lazzari et al.

2008a) and applied multivariate statistical analyses. MPDMA captures the variability of diet and morphospaces. Significant morphological differences across diets are observed

(P < 0.05) in a dataset including all species and in separate datasets including all individual orders save rodents (P = 0.321). Additionally, MPDMA correctly discriminates diet for between 66 and 82% of the specimens in the dataset, respectively including and excluding rodents.

In Chapter 6 I used MPDMA to quantitatively test whether diet drives the evolution of dental morphology towards dental optima. My dataset included a good representation of marsupials and placentals, two mammalian clades that split during the Late Jurassic or

Early Cretaceous and that have since undergone extraordinary adaptive radiations

(O'Leary et al. 2013). Animals from both clades having the same dietary specializations

Silvia Pineda-Munoz - April 2016 135

Diet, ecology, and dental morphology in terrestrial mammals

overlap strongly in MPDMA ecomorphospaces, which suggests convergent phenotypic evolution. Thus, MPDMA does not only work well as a dietary predictor but provides a useful tool for the study of morphological evolution.

Finally, in Chapter 7 I evaluated the differences in dietary discrimination power between unidimensional dental morphology proxies such as orientation patch count

(OPCR), relief index (RI) or slope and MPDMA using the same 138 specimens examined in Chapter 6. Unidimensional proxies significantly discriminate (P < 0.05) between one or two diet categories on one hand and the rest on the other. For example,

OPCR discriminates well between carnivorous and non-carnivorous species. However, none of the individual proxies discriminate all categories. This contrasts with the up to

81% of correctly discriminated species achieved by combining different morphological variables in MPDMA.

136 Silvia Pineda-Munoz - April 2016

Chapter 8: Conclusions

Silvia Pineda-Munoz - April 2016 137

Diet, ecology, and dental morphology in terrestrial mammals

9 FUTURE WORK

During my Ph.D. research I found that it is possible to draw accurate dietary inferences

from dental morphology and that the resultant ecomorphospaces correlate well with

major dietary specialisations. Additionally, morphology constitutes a good proxy for

evolutionary change within a phylogenetic clade and highlights processes of

diversification. Thus, the method I designed provides a solid foundation for future

research. I would like to study morphological evolution across time in the context of

environmental perturbations such as climate change or mass extinctions. The main

points of my future research are:

1) To discover whether mammals from different taxonomic groups

evolve similar or different suites of morphological adaptations in pre-

and post-perturbation ecosystems.

2) To test whether particular morphological traits favoured survival or

diversification of modern mammal clades during major extinction

events.

138 Silvia Pineda-Munoz - April 2016

Chapter 9: Future work

3) To examine the larger question of whether perturbations have driven

the evolution of new morphological traits or whether organisms that

already have traits suited for new environmental condition are the

ones to succeed.

4) To explore whether mammal ecomorphological specialisations have

changed across time in response to environmental perturbation events.

I would like to study fossil collections covering time series spanning well-known environmental perturbation events, such as the Paleocene-Eocene thermal maximum

(PETM) in North America. I will three-dimensionally scan mammal dentitions and document their first and last appearance datums. Then, using MPDMA I will calculate both the morphological disparity of clades and the ecomorphological specialisations at each time interval through stratigraphic time.

To exemplify how my research questions will be interpreted I modelled and analysed hypothetical pre- and post-perturbation event populations.

During environmental perturbations such as the PETM, mammal assemblages experienced a major taxonomical turnover (Alroy et al. 2000, Clyde and Gingerich

1998, Gingerich 2003, Gingerich 2006, Secord et al. 2012). Did pre- and post- perturbation ecosystems evolve similar morphological adaptations?

Silvia Pineda-Munoz - April 2016 139

Diet, ecology, and dental morphology in terrestrial mammals

Figure 7.7-1: Hypothetical faunal turnover event. (a) Pre-perturbation mammal assemblage. (b) and (c) Two hypothetical post-perturbation mammal assemblages. Each ellipse covers the ecomorphological variability of a taxon. Overlapping ellipses represent taxa with the same morphological adaptations.

Figure 7.7-1b represents a hypothetical faunal turnover event in which

new taxa, with different morphological adaptations, can be found after the

perturbation. Meanwhile, the new mammal assemblage in Figure 7.7-1c evolves

the same suites of morphological adaptations as the original population despite

having different taxonomic composition. Quantifying patterns such as these will

help me to answer my research question 1 – whether mammals from different

taxonomic groups evolve similar or different suites of morphological adaptations

in pre- and post-perturbation ecosystems.

Meanwhile, during perturbations some clades go extinct even while others

diversify. Additionally, several orders of mammals may appear together. Do

these successful taxa have particular ecomorphological traits?

Figure 7.7-2: Hypothetical faunal turnover event. a) Pre-perturbation mammal assemblage. b and c) Two hypothetical post-perturbation mammal assemblages. Each ellipse covers the ecomorphological variability of a taxon. Overlapping ellipses represent taxa with the same morphological adaptations.

Figure 7.7-2b represents a post-perturbation event assemblage in which particular morphological adaptations favour the survival and diversification of taxa A and C and

140 Silvia Pineda-Munoz - April 2016

Chapter 9: Future work

even the appearance of a new taxon (Taxon E). In contrast, no particular morphological adaptations favour survival in Figure 7.7-2c, where taxa go extinct randomly and a new taxon is not linked to any pre-existent morphological adaptations (Taxon F). Thus, it can be proposed that perturbations were the driver for the evolution of new morphological traits in the post-perturbation assemblage in Figure 7.7-2c. These analyses could answer my research questions 2 and 3, about whether particular morphological traits favour survival or diversification and about the dynamics of evolution of new morphological traits.

Similarly, by inferring the diet of the species present in each time interval, we can answer research question 4: have environmental perturbations such as mass extinction events changed the ecomorphological specializations of mammal assemblages?

Because the fossil record can inform us about the effects of biotic and abiotic perturbations on the stability of ecosystems via responses of their component taxa, it serves as a predictor of future outcomes and aftermaths and it records patterns of ecosystem recovery (DiMichele et al. 2004, Secord et al. 2012). Studying morphological evolutionary responses across well-studied perturbation events will provide a unique opportunity to understand the processes that moulded the diversification of modern mammals.

Silvia Pineda-Munoz - April 2016 141

Diet, ecology, and dental morphology in terrestrial mammals

10 REFERENCES

Alroy, J. 1998. Cope's rule and the dynamics of body mass evolution in North American

fossil mammals. Science 280(5364):731-734.

Alroy, J., P. L. Koch, and J. C. Zachos. 2000. Global climate change and North

American mammalian evolution. Paleobiology 26:259-288.

Andelt, W. F., J. G. Kie, F. F. Knowlton, and K. Cardwell. 1987. Variation in coyote

diets associated with season and successional changes in vegetation. The Journal

of Wildlife Management:273-277.

Anderson, P. S., M. Friedman, M. D. Brazeau, and E. J. Rayfield. 2011. Initial radiation

of jaws demonstrated stability despite faunal and environmental change. Nature

476(7359):206-209.

Andrews, P., and E. N. Evans. 1979. The environment of ramapithecus in Africa.

Paleobiology 5(1):22-30.

Andrews, P., J. M. Lord, and E. M. N. Evans. 1979. Patterns of ecological diversity in

fossil and modern mammalian faunas. Biological Journal of the Linnean Society

11(2):177-205.

142 Silvia Pineda-Munoz - April 2016

Chapter 10: References

Aristotle. 1983. Parts of Animals. Harvard University Press.

Balestrieri, A., L. Remonti, and C. Prigioni. 2011. Assessing carnivore diet by faecal

samples and stomach contents: a case study with Alpine red foxes. Central

European Journal of Biology 6(2):283-292.

Barnes, R. D., and W. M. Longhurst. 1960. Techniques for dental impressions,

restraining and embedding markers in live-trapped deer. The Journal of Wildlife

Management:224-226.

Bartumeus, F., and J. Catalan. 2009. Optimal search behavior and classic foraging

theory. Journal of Physics A: Mathematical and Theoretical 42(43):434002.

Beeman, L. E., and M. R. Pelton. 1980. Seasonal foods and feeding ecology of black

bears in the Smoky Mountains. Bears: Their Biology and Management 4:141-

147.

Behrensmeyer, A. K. 2006. Climate change and human evolution. Science(Washington)

311(5760):476-478.

Behrensmeyer, A. K., and A. P. Hill. 1988. Fossils in the making: Vertebrate

taphonomy and paleoecology. University of Chicago Press.

Belmaker, J., and W. Jetz. 2011. Cross‐scale variation in –environment

associations. Global Ecology and 20(3):464-474.

Boyer, D. M. 2008. Relief index of second mandibular molars is a correlate of diet

among prosimian primates and other euarchontan mammals. Journal of Human

Evolution 55(6):1118-1137.

Bro, R., and A. K. Smilde. 2014. Principal component analysis. Analytical Methods

6(9):2812-2831.

Silvia Pineda-Munoz - April 2016 143

Diet, ecology, and dental morphology in terrestrial mammals

Brown, J. H., and P. F. Nicoletto. 1991. Spatial scaling of species composition: body

masses of North American land mammals. American Naturalist:1478-1512.

Burness, G. P., J. Diamond, and T. Flannery. 2001. Dinosaurs, dragons, and dwarfs:

The evolution of maximal body size. Proceedings of the National Academy of

Sciences 98(25):14518-14523.

Carbone, C., G. M. Mace, S. C. Roberts, and D. W. Macdonald. 1999. Energetic

constraints on the diet of terrestrial carnivores. Nature 402(6759):286-288.

Casey, M. M., and D. M. Post. 2011. The problem of isotopic baseline: reconstructing

the diet and trophic position of fossil animals. Earth-Science Reviews

106(1):131-148.

Cerling, T., J. Harris, J. Hart, P. Kaleme, H. Klingel, M. Leakey, N. Levin, R. Lewison,

and B. Passey. 2008. Stable isotope ecology of the common hippopotamus.

Journal of Zoology 276(2):204-212.

Cerling, T. E., S. A. Andanje, S. A. Blumenthal, F. H. Brown, K. L. Chritz, J. M.

Harris, J. A. Hart, F. M. Kirera, P. Kaleme, and L. N. Leakey. 2015. Dietary

changes of large herbivores in the Turkana Basin, Kenya from 4 to 1 Ma.

Proceedings of the National Academy of Sciences 112(37):11467-11472.

Cerling, T. E., J. M. Harris, and M. G. Leakey. 1999. Browsing and grazing in

elephants: the isotope record of modern and fossil proboscideans. Oecologia

120(3):364-374.

Cerling, T. E., J. M. Harris, and B. H. Passey. 2003. Diets of East African Bovidae

based on stable isotope analysis. Journal of Mammalogy 84(2):456-470.

Cerling, T. E., G. Wittemyer, J. R. Ehleringer, C. H. Remien, and I. Douglas-Hamilton.

2009. History of Animals using Isotope Records (HAIR): A 6-year dietary

144 Silvia Pineda-Munoz - April 2016

Chapter 10: References

history of one family of African elephants. Proceedings of the National

Academy of Sciences 106(20):8093-8100.

Chan, A. K., G. K. Clancey, and H. C. Loy. 2001. Historical perspectives on East Asian

science, technology, and medicine. Singapore University Press.

Clauss, M., A. Schwarm, S. Ortmann, W. J. Streich, and J. Hummel. 2007. A case of

non-scaling in mammalian physiology? Body size, digestive capacity, food

intake, and ingesta passage in mammalian herbivores. Comparative

Biochemistry and Physiology Part A: Molecular & Integrative Physiology

148(2):249-265.

Clauss, M., P. Steuer, D. W. H. Müller, D. Codron, and J. Hummel. 2013. Herbivory

and body size: Allometries of diet quality and gastrointestinal physiology, and

implications for herbivore ecology and dinosaur gigantism. PLoS One

8(10):e68714.

Clyde, W. C., and P. D. Gingerich. 1998. Mammalian community response to the latest

Paleocene thermal maximum: An isotaphonomic study in the northern Bighorn

Basin, Wyoming. Geology 26(11):1011-1014.

Codron, J., D. Codron, M. Sponheimer, K. Kirkman, K. J. Duffy, E. J. Raubenheimer,

J.-L. Mélice, R. Grant, M. Clauss, and J. A. Lee-Thorp. 2012. Stable isotope

series from elephant ivory reveal lifetime histories of a true dietary generalist.

Proceedings of the Royal Society of London B: Biological Sciences

279(1737):2433-2441.

Colon, C. P., and A. Campos-Arceiz. 2013. The impact of gut passage by binturongs

(Arctictis binturong) on seed germination. The Raffles Bulletin of Zoology

61(1):417-421.

Silvia Pineda-Munoz - April 2016 145

Diet, ecology, and dental morphology in terrestrial mammals

Crawford, K., R. A. Mcdonald, and S. Bearhop. 2008. Applications of stable isotope

techniques to the ecology of mammals. Mammal Review 38(1):87-107.

Cumberland, R. E., J. A. Dempsey, and G. J. Forbes. 2001. Should diet be based on

? Importance of larger prey to the American marten. Wildlife Society

Bulletin:1125-1130.

Cuvier, G. 1825. Discours sur les révolutions de la surface du globe: et sur les

changemens qu'elles ont produits dans le règne animal. G. Dufour et Ed.

D'Ocagne.

Dalerum, F., and A. Angerbjörn. 2005. Resolving temporal variation in vertebrate diets

using naturally occurring stable isotopes. Oecologia 144(4):647-658.

Davis, M., and S. Pineda-Munoz. in review. The temporal scale of diet and dietary

proxies.

Del Rio, C. M., P. Sabat, R. Anderson-Sprecher, and S. P. Gonzalez. 2009. Dietary and

isotopic specialization: the isotopic niche of three Cinclodes ovenbirds.

Oecologia 161(1):149-159.

Demes, B., and N. Creel. 1988. Bite force, diet, and cranial morphology of fossil

hominids. Journal of Human Evolution 17(7):657-670.

DeMiguel, D., B. Azanza, and J. Morales. 2011. Paleoenvironments and paleoclimate of

the Middle Miocene of central Spain: A reconstruction from dental wear of

ruminants. Palaeogeography, Palaeoclimatology, Palaeoecology

302(3–4):452-463.

Demment, M. W., and P. J. Van Soest. 1985. A Nutritional explanation for body-size

patterns of ruminant and nonruminant herbivores. The American Naturalist

125(5):641-672.

146 Silvia Pineda-Munoz - April 2016

Chapter 10: References

Dickman, C. R., and C. Huang. 1988. The reliability of fecal analysis as a method for

determining the diet of insectivorous mammals. Journal of Mammalogy

69(1):108-113.

DiMichele, W. A., A. K. Behrensmeyer, T. D. Olszewski, C. C. Labandeira, J. M.

Pandolfi, S. L. Wing, and R. Bobe. 2004. Long-term stasis in ecological

assemblages: Evidence from the fossil record. Annual Review of Ecology,

Evolution & Systematics 35(1):285-322.

Eisenberg, J. F. 1981. The mammalian radiations: An analysis of trends in evolution,

adaptation, and behavior. University of Chicago Press.

Ellis-Felege, S. N., J. S. Burnam, W. E. Palmer, D. C. Sisson, S. D. Wellendorf, R. P.

Thornton, H. L. Stribling, and J. P. Carroll. 2008. Cameras identify white-tailed

deer depredating Northern Bobwhite nests. Southeastern Naturalist 7(3):562-

564.

Emslie, S. D., and W. P. Patterson. 2007. Abrupt recent shift in δ13C and δ15N values

in Adélie penguin eggshell in Antarctica. Proceedings of the National Academy

of Sciences 104(28):11666-11669.

Evans, A. R. 2011. Surfer Manipulator.

http://users.monash.edu.au/~arevans/software.html.

Evans, A. R., and G. D. Sanson. 2003. The tooth of perfection: functional and spatial

constraints on mammalian tooth shape. Biological Journal of the Linnean

Society 78(2):173-191.

Evans, A. R., G. P. Wilson, M. Fortelius, and J. Jernvall. 2007. High-level similarity of

dentitions in carnivorans and rodents. Nature 445(7123):78-81.

Silvia Pineda-Munoz - April 2016 147

Diet, ecology, and dental morphology in terrestrial mammals

Faith, J. T. 2011. Late Quaternary dietary shifts of the Cape grysbok (Raphicerus

melanotis) in southern Africa. Quaternary Research 75(1):159-165.

Feranec, R. S. 2004. Geographic variation in the diet of hypsodont herbivores from the

Rancholabrean of Florida. Palaeogeography, Palaeoclimatology, Palaeoecology

207(3–4):359-369.

Fernández-Hernández, M., M. T. Alberdi, B. Azanza, P. Montoya, J. Morales, M. Nieto,

and P. Peláez-Campomanes. 2006. Identification problems of arid environments

in the Neogene-Quaternary mammal record of Spain. Journal of Arid

Environments 66(3):585-608.

Figueirido, B., J. A. Pérez-Claros, V. Torregrosa, A. Martín-Serra, and P. Palmqvist.

2010. Demythologizing Arctodus simus, the ‘short-faced’long-legged and

predaceous bear that never was. Journal of Vertebrate Paleontology 30(1):262-

275.

Fisher, D. O., and C. R. Dickman. 1993. Diets of insectivorous marsupials in arid

Australia: selection for prey type, size or hardness? Journal of Arid

Environments 25(4):397-410.

Fortelius, M., J. Eronen, J. Jernvall, L. Liu, D. Pushkina, J. Rinne, A. Tesakov, and I.

Vislobokova. 2002a. Fossil mammals resolve regional patterns of Eurasian

climate change during 20 million years. Research

(4):1005-1016.

Fortelius, M., J. Eronen, J. Jernvall, L. Liu, D. Pushkina, J. Rinne, A. Tesakov, and I.

Vislobokova. 2002b. Fossil mammals resolve regional patterns of Eurasian

climate change during 20 million years. Evolutionary Ecology Research 4:1005-

1016.

148 Silvia Pineda-Munoz - April 2016

Chapter 10: References

Fortelius, M., and N. Solounias. 2000. Functional characterization of ungulate molars

using the abrasion-attrition wear gradient: A new method for reconstructing

paleodiets. American Museum of Natural History.

Gagnon, M., and A. E. Chew. 2000. Dietary preferences in extant african Bovidae.

Journal of Mammalogy 81(2):490-511.

Ganesh, T., and P. Davidar. 1999. Fruit biomass and relative abundance of frugivores in

a rain forest of southern Western Ghats, India. Journal of Tropical Ecology

15(04):399-413.

Gingerich, P. D. 1989. New earliest Wasatchian mammalian fauna from the Eocene of

northwestern Wyoming: composition and diversity in a rarely sampled high-

floodplain assemblage. Museum of Paleontology, University of Michigan.

Gingerich, P. D. 2003. Mammalian responses to climate change at the Paleocene-

Eocene boundary: Polecat Bench record in the northern Bighorn Basin,

Wyoming. Geological Society of America Special Papers 369:463-478.

Gingerich, P. D. 2006. Environment and evolution through the Paleocene–Eocene

thermal maximum. Trends in Ecology & Evolution 21(5):246-253.

Gittleman, J. 1985. Carnivore body size: Ecological and taxonomic correlates.

Oecologia 67(4):540-554.

Gogarten, J. F., and F. E. Grine. 2013. Seasonal mortality patterns in primates:

implications for the interpretation of dental microwear. Evolutionary

Anthropology: Issues, News, and Reviews 22(1):9-19.

Goswami, A., and A. Friscia. 2010. Carnivoran evolution: New views on phylogeny,

form and function. Cambridge University Press.

Silvia Pineda-Munoz - April 2016 149

Diet, ecology, and dental morphology in terrestrial mammals

Guthrie, R., E. Vrba, and G. Schaller. 2000. Paleolithic art as a resource in artiodactyl

paleobiology. Antelopes, deer, and relatives: fossil record, ,

systematics, and conservation (E. Vrba and GB Schaller, eds.). Yale University

Press, New Haven, Connecticut:96-127.

Hansen, R. M. 1978. Shasta ground sloth food habits, Rampart Cave, Arizona.

Paleobiology:302-319.

Hartigan, J. A. 1979. Algorithm AS 136: A k-means clustering algorithm. Journal of the

Royal Statistical Society 28(1):100.

Hawes, J. E., and C. A. Peres. 2014. Ecological correlates of trophic status and

frugivory in neotropical primates. Oikos 123(3):365-377.

Hobson, K. A., M. C. Drever, and G. W. Kaiser. 1999. Norway rats as predators of

burrow-nesting seabirds: insights from stable isotope analyses. The Journal of

Wildlife Management:14-25.

Hoppe, K. A., S. M. Stover, J. R. Pascoe, and R. Amundson. 2004. Tooth enamel

biomineralization in extant horses: implications for isotopic microsampling.

Palaeogeography, Palaeoclimatology, Palaeoecology 206(3):355-365.

Hummel, J., E. Findeisen, K.-H. Südekum, I. Ruf, T. M. Kaiser, M. Bucher, M. Clauss,

and D. Codron. 2011. Another one bites the dust: faecal silica levels in large

herbivores correlate with high-crowned teeth. Proceedings of the Royal Society

of London B: Biological Sciences 278(1712):1742-1747.

Humphries, M. M., D. W. Thomas, and D. L. Kramer. 2003. The role of energy

availability in mammalian hibernation: A cost-benefit approach. Physiological

and Biochemical Zoology 76(2):165-179.

150 Silvia Pineda-Munoz - April 2016

Chapter 10: References

Hutson, J. M., C. C. Burke, and G. Haynes. 2013. Osteophagia and bone modifications

by giraffe and other large ungulates. Journal of Archaeological Science

40(12):4139-4149.

Huwaldt, J. A., and S. Steinhorst. 2001. Plot Digitizer.

Hyslop, E. 1980. Stomach contents analysis-a review of methods and their application.

Journal of fish biology 17(4):411-429.

Ihaka, R., and R. Gentleman. 1996. R: A language for data analysis and graphics.

Journal of Computational and Graphical Statistics 5(3):299-314.

J, O. r. H., and A. C. Kitchener. 2005. The effects of captivity on the morphology of

captive, domesticated and feral mammals. Mammal Review 35(3‐4):215-230.

Johnson, C. F., R. M. Cowling, and P. B. Phillipson. 1999. The flora of the Addo

Elephant National Park, South Africa: are threatened species vulnerable to

elephant damage? & Conservation 8(11):1447-1456.

Jones, K. E., J. Bielby, M. Cardillo, S. A. Fritz, J. O'Dell, C. D. L. Orme, K. Safi, W.

Sechrest, E. H. Boakes, and C. Carbone. 2009. PanTHERIA: a species-level

database of life history, ecology, and geography of extant and recently extinct

mammals: Ecological Archives E090-184. Ecology 90(9):2648-2648.

Jones, K. E., and K. Safi. 2011. Ecology and evolution of mammalian biodiversity.

Philosophical Transactions of the Royal Society B: Biological Sciences

366(1577):2451-2461.

Jones, R., M. Ludlow, J. Troughton, and C. Blunt. 1981. Changes in the natural carbon

isotope ratios of the hair from steers fed diets of C4, C3 and C4 species in

sequence. Search (Australia).

Silvia Pineda-Munoz - April 2016 151

Diet, ecology, and dental morphology in terrestrial mammals

Kaiser, T. M., D. W. Müller, M. Fortelius, E. Schulz, D. Codron, and M. Clauss. 2013.

Hypsodonty and tooth facet development in relation to diet and habitat in

herbivorous ungulates: implications for understanding tooth wear. Mammal

Review 43(1):34-46.

Kararli, T. T. 1995. Comparison of the gastrointestinal anatomy, physiology, and

biochemistry of humans and commonly used laboratory animals.

Biopharmaceutics & drug disposition 16(5):351-380.

Kay, R. F. 1984. On the use of anatomical features to infer foraging behavior in extinct

primates. Pp. 21-53.

Kessler, W. B., W. F. Kasworm, and W. L. Bodie. 1981. Three methods compared for

analysis of pronghorn diets. The Journal of Wildlife Management:612-619.

Kimura, Y., L. L. Jacobs, T. E. Cerling, K. T. Uno, K. M. Ferguson, L. J. Flynn, and R.

Patnaik. 2013. Fossil mice and rats show isotopic evidence of niche partitioning

and change in dental ecomorphology related to dietary shift in late Miocene of

Pakistan. PLoS One 8(8):e69308.

Kitchell, J. A. 1985. Evolutionary paleocology: Recent contributions to evolutionary

theory. Paleobiology 11(1):91-104.

Klare, U., J. F. Kamler, and D. W. Macdonald. 2011. The -eared fox: A dietary

specialist? Mammalian Biology - Mammalian Biology-Zeitschrift für

Säugetierkunde 76(5):646-650.

Kohfeld, K. E., and S. P. Harrison. 2000. How well can we simulate past climates?

Evaluating the models using global palaeoenvironmental datasets. Quaternary

Science Reviews 19(1‚ 2000):321-346.

152 Silvia Pineda-Munoz - April 2016

Chapter 10: References

Kondoh, M. 2003. Foraging adaptation and the relationship between food-web

complexity and stability. Science 299(5611):1388-1391.

Kriwet, J., F. Witzmann, S. Klug, and U. H. Heidtke. 2008. First direct evidence of a

vertebrate three-level trophic chain in the fossil record. Proceedings of the Royal

Society of London B: Biological Sciences 275(1631):181-186.

Kronfeld, N., and T. Dayan. 1998. A new method of determining diets of rodents.

Journal of Mammalogy 79(4):1198-1202.

Kurle, C. M., and G. A. Worthy. 2002. Stable nitrogen and carbon isotope ratios in

multiple tissues of the northern fur seal Callorhinus ursinus: implications for

dietary and migratory reconstructions. Marine Ecology Progress Series 236:289-

300.

Landry, S. O., Jr. 1970. The Rodentia as omnivores. The Quarterly Review of Biology

45(4):351-372.

Laursen, L., and M. Bekoff. 1978. Loxodonta africana. Mammalian Species:1-8.

Lazzari, V., P. Tafforeau, J.-P. Aguilar, and J. Michaux. 2008a. Topographic Maps

Applied to Comparative Molar Morphology: The Case of Murine and Cricetine

Dental Plans (Rodentia, Muroidea). Paleobiology 34(1):46-64.

Lazzari, V., P. Tafforeau, J.-P. Aguilar, and J. Michaux. 2008b. Topographic maps

applied to comparative molar morphology: the case of murine and cricetine

dental plans (Rodentia, Muroidea). Paleobiology 34(01):46-64.

Legendre, S. 1986. Analysis of mammalian communities from the late Eocene and

Oligocene of southern France. Laboratoire de Paléontologie de l'Université des

Sciences et Techniques du Languedoc.

Silvia Pineda-Munoz - April 2016 153

Diet, ecology, and dental morphology in terrestrial mammals

Levin, S. A. 1995. The problem of pattern and scale in ecology. Pp. 277-326. Ecological

Time Series. Springer.

Lieberman, D. E., G. E. Krovitz, F. W. Yates, M. Devlin, and M. S. Claire. 2004.

Effects of food processing on masticatory strain and craniofacial growth in a

retrognathic face. Journal of Human Evolution 46(6):655-677.

Liem, K. F. 1980. Adaptive significance of intra-and interspecific differences in the

feeding repertoires of cichlid fishes. American Zoologist 20(1):295-314.

Lillegraven, J. A., Z. Kielan-Jaworowska, and W. A. Clemens. 1979. Mesozoic

mammals: The first two-thirds of mammalian history. University of California

Press.

Liu, L., K. Puolamäki, J. T. Eronen, M. M. Ataabadi, E. Hernesniemi, and M. Fortelius.

2012. Dental functional traits of mammals resolve productivity in terrestrial

ecosystems past and present. Proceedings of the Royal Society B: Biological

Sciences 279(1739):2793-2799.

Loffredo, L. F., and L. R. DeSantis. 2014. Cautionary lessons from assessing dental

mesowear observer variability and integrating paleoecological proxies of an

extreme generalist Cormohipparion emsliei. Palaeogeography,

Palaeoclimatology, Palaeoecology 395:42-52.

Losos, J. B. 2008. Phylogenetic niche conservatism, phylogenetic signal and the

relationship between phylogenetic relatedness and ecological similarity among

species. Ecology Letters 11(10):995-1003.

Louys, J., P. Ditchfield, C. Meloro, S. Elton, and L. C. Bishop. 2012. Stable isotopes

provide independent support for the use of mesowear variables for inferring

154 Silvia Pineda-Munoz - April 2016

Chapter 10: References

diets in African antelopes. Proceedings of the Royal Society of London B:

Biological Sciences 279(1746):4441-4446.

Lucas, P. W. 2004. Dental functional morphology: How teeth work. Cambridge

University Press.

MacArthur, R. H., and E. R. Pianka. 1966. On optimal use of a patchy environment.

The American Naturalist 100(916):603-609.

Mandelbrot, B. B. 1967. How long is the coast of Britain. Science 156(3775):636-638.

Mattson, D. J., B. M. Blanchard, and R. R. Knight. 1991. Food habits of Yellowstone

grizzly bears, 1977-1987. Canadian Journal of Zoology 69(6):1619-1629.

McCann, K., and A. Hastings. 1997. Re-evaluating the omnivory-stability relationship

in food webs. Proceedings: Biological Sciences 264(1385):1249-1254.

McInnis, M. L., M. Vavra, and W. C. Krueger. 1983. A comparison of four methods

used to determine the diets of large herbivores. Journal of Range Management

36(3):302-306.

McKirahan, R. D. 2011. Philosophy before socrates (Second Edition): An introduction

with texts and commentary. Hackett Publishing.

McLellan, B. N. 2011. Implications of a high-energy and low-protein diet on the body

composition, fitness, and competitive abilities of black (Ursus americanus) and

grizzly (Ursus arctos) bears. Canadian Journal of Zoology 89(6):546-558.

Mendoza, M., C. M. Janis, and P. Palmqvist. 2002. Characterizing complex

craniodental patterns related to feeding behaviour in ungulates: a multivariate

approach. Journal of Zoology 258(02):223-246.

Silvia Pineda-Munoz - April 2016 155

Diet, ecology, and dental morphology in terrestrial mammals

Mendoza, M., C. M. Janis, and P. Palmqvist. 2005. Ecological patterns in the trophic-

size structure of large mammal communities: a “taxon-free” characterization.

Evolutionary Ecology, 7(4):505-530.

Merwe, N. J. v. d. 1982. Carbon isotopes, , and archaeology: Different

pathways of photosynthesis cause characteristic changes in carbon isotope ratios

that make possible the study of prehistoric human diets. American Scientist

70(6):596-606.

Milton, K., and M. L. May. 1976. Body weight, diet and home range area in primates.

Nature 259(5543):459-462.

MOL. Map of Life. http://www.mol.org/.

Munro, R., S. Nielsen, M. Price, G. Stenhouse, and M. Boyce. 2006. Seasonal and diel

patterns of grizzly bear diet and activity in west-central Alberta. Journal of

Mammalogy 87(6):1112-1121.

Münzel, S. C., F. Rivals, M. Pacher, D. Döppes, G. Rabeder, N. J. Conard, and H.

Bocherens. 2014. Behavioural ecology of Late Pleistocene bears (Ursus

spelaeus, Ursus ingressus): Insight from stable isotopes (C, N, O) and tooth

microwear. Quaternary International 339:148-163.

Nowak, R. M. 1999. Walker's Mammals of the World. Baltimore.

Nowak, R. M. 2005. Walker's Carnivores of the World. Johns Hopkins University

Press.

O'Leary, M. A., J. I. Bloch, J. J. Flynn, T. J. Gaudin, A. Giallombardo, N. P. Giannini,

S. L. Goldberg, B. P. Kraatz, Z.-X. Luo, J. Meng, X. Ni, M. J. Novacek, F. A.

Perini, Z. S. Randall, G. W. Rougier, E. J. Sargis, M. T. Silcox, N. B. Simmons,

M. Spaulding, P. M. Velazco, M. Weksler, J. R. Wible, and A. L. Cirranello.

156 Silvia Pineda-Munoz - April 2016

Chapter 10: References

2013. The placental mammal ancestor and the post–K-Pg radiation of placentals.

Science 339(6120):662-667.

Olson, D. M., E. Dinerstein, E. D. Wikramanayake, N. D. Burgess, G. V. Powell, E. C.

Underwood, J. A. D'amico, I. Itoua, H. E. Strand, and J. C. Morrison. 2001.

Terrestrial Ecoregions of the World: A New Map of Life on Earth A new global

map of terrestrial ecoregions provides an innovative tool for conserving

biodiversity. BioScience 51(11):933-938.

Owen, R. 1845. Odontography, or a treatise on the comparative anatomy of the teeth,

their physiological relations, mode of development and microscopic structure in

the vertebrate animals. Bailliere, London, Bailliere.

Palmqvist, P., D. R. Gröcke, A. Arribas, and R. A. Fariña. 2003. Paleoecological

reconstruction of a lower Pleistocene large mammal community using

biogeochemical (δ13C, δ15N, δ18O, Sr:Zn) and ecomorphological approaches.

Paleobiology 29(2):205-229.

Penone, C., A. D. Davidson, K. T. Shoemaker, M. Di Marco, C. Rondinini, T. M.

Brooks, B. E. Young, C. H. Graham, and G. C. Costa. 2014. Imputation of

missing data in life‐history trait datasets: which approach performs the best?

Methods in Ecology and Evolution 5(9):961-970.

Petchey, O. L., A. P. Beckerman, J. O. Riede, and P. H. Warren. 2008. Size, foraging,

and food web structure. Proceedings of the National Academy of Sciences

105(11):4191-4196.

Peters, R. H. 1986. The Ecological Implications of Body Size. Cambridge University

Press.

Silvia Pineda-Munoz - April 2016 157

Diet, ecology, and dental morphology in terrestrial mammals

Pimm, S. L., and J. H. Lawton. 1977. Number of trophic levels in ecological

communities. Nature 268(5618):329-331.

Pimm, S. L., and J. H. Lawton. 1978. On feeding on more than one trophic level. Nature

275(5680):542-544.

Pineda-Munoz. 2015. Mutli-proxy dental morphology analysis: a new approach for

inferring diet. SVP 75th Annual Meeting. Society of Vertebrate Paleontology

Dallas, TX, USA.

Pineda-Munoz, S., and J. Alroy. 2014. Dietary characterization of terrestrial mammals.

Proceedings of the Royal Society B: Biological Sciences 281(1789).

Pineda-Munoz, S., J. Alroy, and A. R. Evans. In review. Dental morphology predicts

diet across marsupials and placentals

Pineda-Munoz, S., A. Evans, and J. Alroy. in press. The relationship between diet and

body mass in mammals. Paleobiology.

Polis, G. A. 1991. Complex trophic interactions in deserts: An empirical critique of

food-web theory. The American Naturalist 138(1):123-155.

Polly, P. D. 2004. On the simulation of the evolution of morphological shape:

multivariate shape under selection and drift. Palaeontologia Electronica 7(2):1-

28.

Polly, P. D. 2015. Gene networks, occlusal clocks, and functional patches: new

understanding of pattern and process in the evolution of the dentition.

Odontology 103(2):117-125.

Polly, P. D., S. Le Comber, and T. Burland. 2005. On the occlusal fit of tribosphenic

molars: Are we underestimating species diversity in the Mesozoic? Journal of

Mammalian Evolution 12(1-2):283-299.

158 Silvia Pineda-Munoz - April 2016

Chapter 10: References

Price, S. A., S. S. Hopkins, K. K. Smith, and V. L. Roth. 2012. Tempo of trophic

evolution and its impact on mammalian diversification. Proceedings of the

National Academy of Sciences 109(18):7008-7012.

Price, S. A., and S. S. B. Hopkins. 2015. The macroevolutionary relationship between

diet and body mass across mammals. Biological Journal of the Linnean Society

115(1):173-184.

Pyke, G. H., H. R. Pulliam, and E. Charnov. 1977. Optimal foraging: a selective review

of theory and tests. Quarterly Review of Biology 52:137-154.

R Core Team. 2013. R: A Language and environment for statistical computing. Vienna,

Austria.

Raia, P., F. Carotenuto, F. Passaro, D. Fulgione, and M. Fortelius. 2012. Ecological

specialization in fossil mammals explains Cope’s rule. The American Naturalist

179(3):328-337.

Rawn-Schatzinger, V. 1992. The scimitar cat, Homotherium serum cope: osteology,

functional morphology, and predatory behavior. Illinois State Museum.

Reed, K. E. 1998. Using large mammal communities to examine ecological and

taxonomic structure and predict vegetation in extant and extinct assemblages.

Paleobiology 24(3):384-408.

Reynolds-Hogland, M. J., and M. S. Mitchell. 2007. Three axes of ecological studies.

Pp. 174-194. Temporal Dimensions of . Springer.

Rio, C. M. d., and S. A. Carleton. 2012. How fast and how faithful: the dynamics of

isotopic incorporation into animal tissues. Journal of Mammalogy 93(2):353-

359.

Silvia Pineda-Munoz - April 2016 159

Diet, ecology, and dental morphology in terrestrial mammals

Rivals, F., G. Semprebon, and A. Lister. 2012. An examination of dietary diversity

patterns in Pleistocene proboscideans (Mammuthus, Palaeoloxodon, and

Mammut) from Europe and North America as revealed by dental microwear.

Quaternary International 255:188-195.

Rivals, F., and G. M. Semprebon. 2011. Dietary plasticity in ungulates: insight from

tooth microwear analysis. Quaternary International 245(2):279-284.

Rivals, F., N. Solounias, and M. C. Mihlbachler. 2007. Evidence for geographic

variation in the diets of late Pleistocene and early Holocene Bison in North

America, and differences from the diets of recent Bison. Quaternary Research

68(3):338-346.

Robinson, J. G., and K. H. Redford. 1986. Body size, diet, and population density of

Neotropical forest mammals. The American Naturalist 128(5):665-680.

Saarinen, J. J., A. G. Boyer, J. H. Brown, D. P. Costa, S. K. M. Ernest, A. R. Evans, M.

Fortelius, J. L. Gittleman, M. J. Hamilton, L. E. Harding, K. Lintulaakso, S. K.

Lyons, J. G. Okie, R. M. Sibly, P. R. Stephens, J. Theodor, M. D. Uhen, and F.

A. Smith. 2014. Patterns of maximum body size evolution in Cenozoic land

mammals: eco-evolutionary processes and abiotic forcing. Proceedings of the

Royal Society B 281(1784).

Salazar-Ciudad, I., and M. Marin-Riera. 2013. Adaptive dynamics under development-

based genotype-phenotype maps. Nature 497(7449):361-364.

Salter, R., and R. Hudson. 1979. Feeding ecology of feral horses in western Alberta.

Journal of Range Management:221-225.

Schell, D. M. 2000. Declining in the Bering Sea: isotopic evidence

from whale baleen. Limnology and Oceanography 45(2):459-462.

160 Silvia Pineda-Munoz - April 2016

Chapter 10: References

Schoener, T. W. 1989. Food webs from the small to the large: The Robert H.

MacArthur Award Lecture. Ecology 70(6):1559-1589.

Schubert, B. W., P. S. Ungar, M. Sponheimer, and K. E. Reed. 2006. Microwear

evidence for Plio–Pleistocene bovid diets from Makapansgat Limeworks Cave,

South Africa. Palaeogeography, Palaeoclimatology, Palaeoecology 241(2):301-

319.

Schwartz, C. C., and J. E. Ellis. 1981. Feeding ecology and niche separation in some

native and domestic ungulates on the Shortgrass Prairie. Journal of Applied

Ecology 18(2):343-353.

Secord, R., J. I. Bloch, S. G. B. Chester, D. M. Boyer, A. R. Wood, S. L. Wing, M. J.

Kraus, F. A. McInerney, and J. Krigbaum. 2012. Evolution of the earliest horses

driven by climate change in the Paleocene-Eocene Thermal Maximum. Science

335(6071):959-962.

Semprebon, G. M., and F. Rivals. 2010. Trends in the paleodietary habits of fossil

camels from the Tertiary and Quaternary of North America. Palaeogeography,

Palaeoclimatology, Palaeoecology 295(1):131-145.

Sepkoski Jr, J. J. 1984. A kinetic model of Phanerozoic taxonomic diversity. III. Post-

Paleozoic families and mass extinctions. Paleobiology:246-267.

Shrestha, R., and P. Wegge. 2006. Determining the composition of herbivore diets in

the Trans-Himalayan Rangelands: A comparison of field methods. Rangeland

Ecology & Management 59(5):512-518.

Siemann, E., and J. H. Brown. 1999. Gaps in mammalian body size distributions

reexamined. Ecology 80(8):2788-2792.

Simpson, E. H. 1949. Measurement of diversity. Nature 163(4148):688-688.

Silvia Pineda-Munoz - April 2016 161

Diet, ecology, and dental morphology in terrestrial mammals

Sinclair, A. R. E. 2003. Mammal population regulation, keystone processes and

ecosystem dynamics. Philosophical Transactions of the Royal Society of

London. Series B: Biological Sciences 358(1438):1729-1740.

Singer, M. S., and E. A. Bernays. 2003. Understanding omnivory needs a behavioral

perspective. Ecology 84(10):2532-2537.

Smith, F. A., A. G. Boyer, J. H. Brown, D. P. Costa, T. Dayan, S. K. M. Ernest, A. R.

Evans, M. Fortelius, J. L. Gittleman, M. J. Hamilton, L. E. Harding, K.

Lintulaakso, S. K. Lyons, C. McCain, J. G. Okie, J. J. Saarinen, R. M. Sibly, P.

R. Stephens, J. Theodor, and M. D. Uhen. 2010. The evolution of maximum

body size of terrestrial mammals. Science 330(6008):1216-1219.

Smith, F. A., J. H. Brown, J. P. Haskell, S. K. Lyons, J. Alroy, E. L. Charnov, T. Dayan,

B. J. Enquist, S. M. Ernest, and E. A. Hadly. 2004. Similarity of mammalian

body size across the taxonomic hierarchy and across space and time. The

American Naturalist 163(5):672-691.

Smith, F. A., and S. K. Lyons. 2011. How big should a mammal be? A macroecological

look at mammalian body size over space and time. Philosophical Transactions of

the Royal Society B: Biological Sciences 366(1576):2364-2378.

Smith, F. A., S. K. Lyons, S. K. M. Ernest, K. E. Jones, D. M. Kaufman, T. Dayan, P.

A. Marquet, J. H. Brown, and J. P. Haskell. 2003. Body mass of late quaternary

mammals. Ecology 84(12):3403-3403.

Smith, J. M., R. Burian, S. Kauffman, P. Alberch, J. Campbell, B. Goodwin, R. Lande,

D. Raup, and L. Wolpert. 1985. Developmental constraints and evolution: a

perspective from the Mountain Lake conference on development and evolution.

Quarterly Review of Biology:265-287.

162 Silvia Pineda-Munoz - April 2016

Chapter 10: References

Sponheimer, M., T. Robinson, L. Ayliffe, B. Passey, B. Roeder, L. Shipley, E. Lopez,

T. Cerling, D. Dearing, and J. Ehleringer. 2003. An experimental study of

carbon-isotope fractionation between diet, hair, and feces of mammalian

herbivores. Canadian Journal of Zoology 81(5):871-876.

Steuer, P., K.-H. Südekum, T. Tütken, D. W. H. Müller, J. Kaandorp, M. Bucher, M.

Clauss, and J. Hummel. 2014. Does body mass convey a digestive advantage for

large herbivores? 28(5):1127-1134.

Teaford, M. F., and O. J. Oyen. 1989. In vivo and in vitro turnover in dental microwear.

American journal of physical anthropology 80(4):447-460.

Travouillon, K. J., and S. Legendre. 2009. Using cenograms to investigate gaps in

mammalian body mass distributions in Australian mammals. Palaeogeography,

Palaeoclimatology, Palaeoecology 272(1–2):69-84.

Tütken, T., T. M. Kaiser, T. Vennemann, and G. Merceron. 2013. Opportunistic feeding

strategy for the earliest old world hypsodont equids: Evidence from stable

isotope and dental wear proxies. PLoS One 8(9):e74463.

Ungar, P. S. 2010. Mammal teeth: Origin, evolution, and diversity. Johns Hopkins

University Press.

Valverde, J. A. 1967. Estructura de una comunidad mediterránea de vertebrados

terrestres. Consejo Superior de Investigaciones Científicas.

Van Valkenburgh, B. 2009. Costs of carnivory: tooth fracture in Pleistocene and Recent

carnivorans. Biological Journal of the Linnean Society 96(1):68-81.

West, J. B., G. J. Bowen, T. E. Cerling, and J. R. Ehleringer. 2006. Stable isotopes as

one of nature's ecological recorders. Trends in Ecology & Evolution 21(7):408-

414.

Silvia Pineda-Munoz - April 2016 163

Diet, ecology, and dental morphology in terrestrial mammals

Willerslev, E., J. Davison, M. Moora, M. Zobel, E. Coissac, M. E. Edwards, E. D.

Lorenzen, M. Vestergård, G. Gussarova, and J. Haile. 2014. Fifty thousand years

of Arctic vegetation and megafaunal diet. Nature 506(7486):47-51.

Wilman, H., J. Belmaker, J. Simpson, C. de la Rosa, M. M. Rivadeneira, and W. Jetz.

2014. EltonTraits 1.0: Species-level foraging attributes of the world's birds and

mammals. Ecology 95(7):2027-2027.

Wilson, D. E., and D. A. M. Reeder. 2005. Mammal species of the world: A taxonomic

and geographic reference. Johns Hopkins University Press.

Wilson, E. O., H. University, C. L. Sciences, D. E. L. Studies, and N. A. S. S.

Institution. 1988. Biodiversity. National Academies Press.

Wilson, G. P., A. R. Evans, I. J. Corfe, P. D. Smits, M. Fortelius, and J. Jernvall. 2012.

Adaptive radiation of multituberculate mammals before the extinction of

dinosaurs. Nature 483(7390):457-460.

Wolkovich, E., B. Cook, K. McLauchlan, and T. Davies. 2014. Temporal ecology in the

Anthropocene. Ecology Letters 17(11):1365-1379.

Wroe, S., M. Crowther, J. Dortch, and J. Chong. 2004. The size of the largest marsupial

and why it matters. Proceedings of the Royal Society of London B: Biological

Sciences 271(Suppl 3):S34-S36.

Zazzo, A., A. Mariotti, C. Lécuyer, and E. Heintz. 2002. Intra-tooth isotope variations

in late Miocene bovid enamel from Afghanistan: paleobiological, taphonomic,

and climatic implications. Palaeogeography, Palaeoclimatology, Palaeoecology

186(1):145-161.

Zeileis, A., and G. Grothendieck. 2005. zoo: S3 infrastructure for regular and irregular

time series. arXiv preprint math/0505527.

164 Silvia Pineda-Munoz - April 2016

Chapter 10: References

Silvia Pineda-Munoz - April 2016 165

Diet, ecology, and dental morphology in terrestrial mammals

11 APPENDICES

166 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

11.1 Supplementary information 1 Table 11.1-1: Summary of the literature resources for the stomach content data for the 139 mammalian species analysed in this study.

Order Family Species References

Artiodactyla Bovidae Aepyceros melampus Dunham 1980

Artiodactyla Bovidae Ammotragus lervia Gray and Simpson 1980

Artiodactyla Bovidae Antidorcas marsupialis Davies et al. 1986

Artiodactyla Bovidae Cephalophus callipygus Gautier-Hion et al. 1980

Artiodactyla Bovidae Cephalophus dorsalis Gautier-Hion et al. 1980

Artiodactyla Bovidae Cephalophus leucogaster Gautier-Hion et al. 1980

Artiodactyla Bovidae Cephalophus nigrifrons Gautier-Hion et al. 1980

Artiodactyla Bovidae Cephalophus silvicultor Gautier-Hion et al. 1980

Artiodactyla Bovidae Hyemoschus aquaticus Gautier-Hion et al. 1980

Artiodactyla Bovidae Philantomba monticola Gautier-Hion et al. 1980

Artiodactyla Bovidae Tragelaphus spekii Gautier-Hion et al. 1980

Artiodactyla Tayassuidae Tayassu pecari Sowls 1997

Carnivora Canidae Canis lupus dingo Newsome et al. 1983

Carnivora Canidae Canis mesomelas Bothma 1966 Carnivora Canidae Otocyon megalotis Bothma 1966, Klare et al. 2011

Carnivora Canidae Urocyon cinereoargenteus Hockman and Chapman 1983

Carnivora Canidae Vulpes lagopus Anthony et al. 2000

Carnivora Canidae Vulpes vulpes Hockman and Chapman 1983

Carnivora Herpestidae Helogale hirtula Hemming 1972

Carnivora Mustelidae Galictis vittata Bisbal E. 1986

Carnivora mustelidae Martes martes Ruiz-Olmo 1996

Carnivora Mustelidae Meles meles Cleary et al. 2011

Carnivora Nandinidae Nandinia binotata McNab 1995

Carnivora Procyonidae Procyon lotor Baker et al. 1945

Carnivora Ursidae Ursus arctos Sato et al. 2005

Diprotodontia Burramyidae Cercartetus caudatus Flannery and Schouten 1994

Diprotodontia Macropodidae Thylogale stigmatica Vernes 1995 Fielden et al. 1990, Eulipotyphla Chrysochloridae Eremitalpa granti Perrin and Fielden 1999

Eulipotyphla Soricidae Blarina brevicauda Hahus and Smith 1990

Eulipotyphla Soricidae Crocidura cyanea Monadjem 1997

Eulipotyphla Soricidae Crocidura flavecens Monadjem 1997

Eulipotyphla Soricidae Crocidura mariquensis Monadjem 1997

Eulipotyphla Soricidae Myosorex cafer Monadjem 1997

Eulipotyphla Soricidae Myosorex varius Monadjem 1997

Eulipotyphla Soricidae Sorex fumeus Owen 1984

Eulipotyphla Soricidae Sorex hoyi Long 1974

Silvia Pineda-Munoz - April 2016 167

Diet, ecology, and dental morphology in terrestrial mammals

Eulipotyphla Soricidae Sorex palustris Beneski and Stinson 1987

Eulipotyphla Parascalops breweri Hallett 1978

Eulipotyphla Talpidae Scapanus townsendii Carraway et al. 1993

Eulipotyphla Talpidae Talpa europaea Funmilayo 1979 Elephantulus Koontz and Roeper 1983, Leirs et al. Macroscelidea Macroscelidae brachyrhynchus 1995

Macroscelidea Macroscelidae Elephantulus rufescens Hemming 1972

Macroscelidea Macroscelidae Macroscelides proboscideus Kerley 1992

Microbiotheria Microbiotheriidae Dromiciops gliroides Meserve et al. 1988

Paucituberculata Caenolestidae Caenolestes fuliginosus Barkley and Whitaker 1984

Paucituberculata Caenolestidae Rhyncholestes raphanurus Meserve 1981

Pholidota Manidae Manis crassicaudata Heath 1995

Primates Atelidae Alouatta seniculus Guillotin et al. 1994

Primates Atelidae Ateles paniscus Guillotin et al. 1994

Primates Cebidae Cebus apella Guillotin et al. 1994

Primates Cercopithecidae Cercocebus agilis Gautier-Hion et al. 1980

Primates Cercopithecidae Cercopithecus cephus Gautier-Hion et al. 1980

Primates Cercopithecidae Cercopithecus neglectus Gautier-Hion et al. 1980

Primates Cercopithecidae Cercopithecus nictitans Gautier-Hion et al. 1980

Primates Cercopithecidae Cercopithecus pogonias Gautier-Hion et al. 1980

Primates Cercopithecidae Colobus guereza Gautier-Hion et al. 1980

Primates Cercopithecidae Mandrillus sphinx Gautier-Hion et al. 1980

Primates Cercopithecidae Miopithecus talapoin Gautier-Hion et al. 1980

Primates Galagidae Euoticus elegantulus Charles-Dominique 1974

Primates Galagidae Galago alleni Charles-Dominique 1974

Primates Galagidae Galago demidoff Charles-Dominique 1974

Primates Lorisidae Arctocebus calabarensis Charles-Dominique 1974

Primates Lorisidae Perodictus potto Charles-Dominique 1974

Proboscidea Elephantidae Loxodonta africana Buss 1961

Rodentia Heteromyidae Chaetodipus baileyi Reichman 1975

Rodentia Heteromyidae Chaetodipus intermedius Reichman 1975

Rodentia Heteromyidae Dipodomys merriami Reichman 1975

Rodentia Heteromyidae Perognathus amplus Reichman 1975

Rodentia Muridae Abrothrix longipilis Meserve et al. 1988

Rodentia Muridae Abrothrix sanborni Meserve et al. 1988

Rodentia Muridae Acomys cahirinus Hemming 1972

Rodentia Muridae Acomys dimidiatus Varty 1990

Rodentia Muridae Aethomys crisophilus Monadjem 1997

Rodentia Muridae Akodon azarae Ellis et al. 1998

Rodentia Muridae Arvicanthis niloticus Hemming 1972

Rodentia Muridae Calomys laucha Ellis et al. 1998

Rodentia Muridae Calomys musculinus Ellis et al. 1998

Rodentia Muridae Geoxus valdivianus Meserve et al. 1988

168 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Rodentia Muridae Gerbilliscus leucogaster Monadjem 1997

Rodentia Muridae Gerbilliscus robustus Hemming 1972

Rodentia Muridae Gerbillurus paeba Perrin et al. 1999

Rodentia Muridae Grammomys dolichurus Varty 1990

Rodentia Muridae Irenomys tarsalis Meserve et al. 1988

Rodentia Muridae Lemniscomys rosalia Monadjem 1997

Rodentia Muridae Lorentzimys nouhuysi Jackson 1992

Rodentia Muridae Loxodontomys micropus Meserve et al. 1988 Rodentia Muridae Mastomys natalensis Langham 1983, Monadjem 1997

Rodentia Muridae Maxomys surifer Monadjem 1997

Rodentia Muridae Micaelamys namaquensis Langham 1983

Rodentia Muridae Microtus ochrogaster Hahus and Smith 1990 Rodentia Muridae Mus minutoides Langham 1983, Kerley 1992

Rodentia Muridae Myotomys unisulcatus Kerley 1992

Rodentia Muridae Necromys obscurus Ellis et al. 1998

Rodentia Muridae Oligoryzomys flavescens Ellis et al. 1998

Rodentia Muridae Oryzomys longicaudatus Meserve et al. 1988

Rodentia Muridae Otomys angoniensis Langham 1983

Rodentia Muridae Otomys irroratus Langham 1983

Rodentia Muridae Peromyscus attwateri Schmidly 1974 Rodentia Muridae Peromyscus leucopus Whitaker 1966, Hahus and Smith 1990

Rodentia Muridae Phyllotis darwini Meserve 1981

Rodentia Muridae Pseudohydromys ellermani Jackson 1992

Rodentia Muridae Pseudohydromys fuscus Jackson 1992

Rodentia Muridae Pseudohydromys murinus Jackson 1992 Pseudohydromys

Rodentia Muridae Jackson 1992 occidentalis

Rodentia Muridae Rattus rattus Langham 1983

Rodentia Muridae Rhabdomys pumilo Kerley 1992

Rodentia Muridae Taterillus harringtoni Hemming 1972

Rodentia Nesomydae Dendromus mesomelas Langham 1983

Rodentia Nesomydae Dendromus mystacalis Langham 1983

Rodentia Nesomydae Malacothrix typica Kerley 1992

Rodentia Nesomydae Saccostomus campestris Langham 1983

Rodentia Nesomydae Saccostomus mearnsi Varty 1990

Rodentia Nesomydae Steatomys pratensis Langham 1983

Rodentia Octodontidae Ctenomys mendocinus Torres-Mura et al. 1989

Rodentia Octodontidae Octodon degus Meserve et al. 1983

Rodentia Octodontidae Tympanoctomys barrerae Torres-Mura et al. 1989

Rodentia Sciuridae Callosciurus melanogaster Whitten 1981

Rodentia Sciuridae Citellus lateralis Tevis 1953

Rodentia Sciuridae Cynomys ludovicianus Wydeven and Dahlgren 1982 Rodentia Sciuridae Epixerus ebii Emmons 1980, Gautier-Hion et al. 1980

Silvia Pineda-Munoz - April 2016 169

Diet, ecology, and dental morphology in terrestrial mammals

Rodentia Sciuridae Funisciurus anerythrus Emmons 1980

Rodentia Sciuridae Funisciurus isabella Emmons 1980

Rodentia Sciuridae Funisciurus lemniscatus Emmons 1980

Rodentia Sciuridae Funisciurus pyrropus Emmons 1980

Rodentia Sciuridae Heliosciurus rufobrachium Emmons 1980

Rodentia Sciuridae Lariscus obscurus Whitten 1981 Rodentia Sciuridae Myosciurus pumilio Emmons 1980, Gautier-Hion et al. 1980

Rodentia Sciuridae Neotamias townsendi Tevis 1953 Rodentia Sciuridae Paraxeros poensis Emmons 1980, Gautier-Hion et al. 1980 Rodentia Sciuridae Protoxerus stangeri Emmons 1980, Gautier-Hion et al. 1980

Rodentia Sciuridae Sundasciurus lowii Whitten 1981

Rodentia Sciuridae Tamias amoenus Tevis 1953

Rodentia Sciuridae Tamias quadrimaculatus Tevis 1953

Rodentia Sciuridae Tamias speciosus Tevis 1953 Rodentia Sciuridae Xerus rutilus Hemming 1972, O'Shea 1991

Rodentia Zapodidae Napaeozapus insignis Whitaker 1963

Scandentia Tupiidae glis Langham 1983

Xenarthra Dasypodidae Chaetophractus vellerosus Greegor 1980

Xenarthra Dasypodidae Dasypus novemcinctus Whitaker 1963

Dataset references Anthony, R. M., N. L. Barten, and P. E. Seiser. 2000. Foods of Arctic foxes Alopex

lagopus during winter and spring in western Alaska. Journal of Mammalogy

81:820-828.

Baker, R. H., C. C. Newman, and F. Wilke. 1945. Food habits of the raccoon in eastern

Texas. Journal of Wildlife Management 9:45-48.

Barkley, L. J. and J. O. Whitaker, Jr. 1984. Confirmation of Caenolestes in Peru with

Information on Diet. Journal of Mammalogy 65:328-330.

Beneski, J. T., Jr. and D. W. Stinson. 1987. Sorex palustris. Mammalian Species 296:1-

6.

Bisbal E., F. J. 1986. Food habits of some neotropical carnivores in Venezuela

(Mammalia, Carnivora). Mammalia 50:329-340.

170 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Bothma, J. D. P. 1966. Notes on the stomach contents of certain Carnivora (Mammalia)

from the Kalahari Gemsbok Park. Koedoe 9:37–39.

Buss, I. O. 1961. Some observations on food habits and behaviour of the African

Elephant. Journal of Wildlife Management 25:131-148.

Carraway, L. N., L. F. Alexander, and B. J. Verts. 1993. Scapanus townsendii.

Mammalian Species 434:1-7.

Charles-Dominique, P. 1974 Ecology and feeding behaviour of five sympatric lorisids

in Gabon. Pages 131-150 in G. A. D. R. D. Martin, and A. C. Walker, editor.

Prosimian Biology. University of Pittsburgh Press, Pittsburgh, PA.

Cleary, G. P., L. A. L. Corner, J. O’Keeffe, and N. M. Marples. 2011. Diet of the

European badger Meles meles in the Republic of Ireland: A comparison of results

from an analysis of stomach contents and rectal faeces. Mammalian Biology -

Zeitschrift für Säugetierkunde 76:470-475.

Davies, R. A. G., P. Botha, and J. D. Skinner. 1986. Diet selected by springbok

Antidorcas marsupialis and Merino sheep Ovis aries during a Karoo drought.

Transactions of the Royal Society of South Africa 46:165-176.

Dunham, K. M. 1980. The diet of impala Aepyceros melampus in the Sengwa Wildlife

Research Area, Rhodesia. Journal of Zoology 192:41-57.

Ellis, B. A., J. N. Mills, G. E. Glass, K. T. McKee, Jr., D. A. Enria, and J. E. Childs.

1998. Dietary habits of the common rodents in an agroecosystem in Argentina.

Journal of Mammalogy 79:1203-1220.

Emmons, L. H. 1980. Ecology and resource partitioning among nine species of African

rain forest squirrels. Ecological Monographs 50:31-54.

Silvia Pineda-Munoz - April 2016 171

Diet, ecology, and dental morphology in terrestrial mammals

Fielden, L. J., M. R. Perrin, and G. C. Hickman. 1990. Feeding ecology and foraging

behaviour of the Namib Desert golden mole, Eremitalpa granti namibensis

Chrysochloridae. Journal of Zoology 220:367-389.

Flannery, T. F. and P. Schouten. 1994. Possums of the world: a monograph of the

Phalangeroidea. GEO Productions in association with the Australian Museum.

Funmilayo, O. 1979. Food consumption preferences and storage in the mole Talpa

europaea. Acta Theriologica 24:379-390.

Gautier-Hion, A., L. H. Emmons, and G. Dubost. 1980. A comparison of the diets of

three major groups of primary consumers of Gabon primates, squirrels and

ruminants. Oecologia 45:182-189.

Gray, G. G. and C. D. Simpson. 1980. Ammotragus lervia. Mammalian Species 144:1-

7.

Greegor, D. H., Jr. 1980. Diet of the little hairy , Chaetophractus vellerosus,

of northwestern Argentina. Journal of Mammalogy 61:331-334.

Guillotin, M., G. Dubost, and D. Sabatier. 1994. Food choice and food competition

among the three major primate species of French Guiana. Journal of Zoology

233:551-579.

Hahus, S. C. and K. G. Smith. 1990. Food habits of Blarina, Peromyscus, and Microtus

in relation to an emergence of periodical cicadas Magicicada. Journal of

Mammalogy 71:249-252.

Hallett, J. G. 1978. Parascalops breweri. Mammalian Species 98:1-4.

Heath, M. E. 1995. Manis crassicaudata. Mammalian Species 414:1-4.

172 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Hemming, C. F. 1972. The South Turkana Expedition: scientific papers. VIII. The

ecology of South Turkana: a reconnaissance classification. Geographical Journal

138:15-40.

Hockman, J. G. and J. A. Chapman. 1983. Comparative feeding habits of red foxes

Vulpes vulpes and fray foxes Urocyon cinereoargenteus in Maryland. American

Midland Naturalist 110:276-285.

Jackson, K. L. a. W., P. A. 1992. The diet of five species of New Guinean rodents.

Science in New Guinea 19:77–86.

Kerley, G. I. H. 1992. Trophic status of small mammals in the semi-arid Karoo, South

Africa. Journal of Zoology 226:563-572.

Klare, U., J. F. Kamler, and D. W. Macdonald. 2011. The bat-eared fox: a dietary

specialist? Mammalian Biology - Zeitschrift für Säugetierkunde 76:646-650.

Koontz, F. W. and N. J. Roeper. 1983. Elephantulus rufescens. Mammalian Species

204:1-5.

Langham, N. 1983. Distribution and ecology of small mammals in three rain forest

localities of Peninsula Malaysia with particular references to Kedah Peak.

Biotropica 15:199-206.

Leirs, H., R. Verhagen, W. Verheyen, and M. R. Perrin. 1995. The biology of

Elephantulus brachyrhynchus in natural miombo woodland in Tanzania. Mammal

Review 25:45-49.

Long, C. A. 1974. Microsorex hoyi and Microsorex thompsoni. Mammalian Species

33:1-4.

McNab, B. K. 1995. Energy expenditure and conservation in frugivorous and mixed-

diet carnivorans. Journal of Mammalogy 76:206-222.

Silvia Pineda-Munoz - April 2016 173

Diet, ecology, and dental morphology in terrestrial mammals

Meserve, P. L. 1981. Trophic Relationships among small mammals in a Chilean

semiarid thorn scrub community. Journal of Mammalogy 62:304-314.

Meserve, P. L., B. K. Lang, and B. D. Patterson. 1988. Trophic relationships of small

mammals in a Chilean temperate rainforest. Journal of Mammalogy 69:721-730.

Meserve, P. L., R. E. Martin, and J. Rodriguez. 1983. Feeding ecology of two Chilean

caviomorphs in a central Mediterranean savanna. Journal of Mammalogy 64:322-

325.

Monadjem, A. 1997. Stomach contents of 19 species of small mammals from

Swaziland. South African Journal of Zoology 32:23.

Newsome, A., L. Corbett, P. Catling, and R. Burt. 1983. The feeding ecology of the

dingo. 1. Stomach contents from trapping in south-eastern Australia, and the non-

target wildlife also caught in dingo traps. Wildlife Research 10:477-486.

O'Shea, T. J. 1991. Xerus rutilus. Mammalian Species 370:1-5.

Owen, J. G. 1984. Sorex fumeus. Mammalian Species 215:1-8.

Perrin, M. R., E. R. Dempster, and C. T. Downs. 1999. Gerbillurus paeba. Mammalian

Species 606:1-6.

Perrin, M. R., and L. J. Fielden. 1999. Eremitalpa granti. Mammalian Species 629:1-4.

Reichman, O. J. 1975. Relation of desert rodent diets to available resources. Journal of

Mammalogy 56:731-751.

Ruiz-Olmo, J., López-Martín, J. M. . 1996 Seasonal food of pine marten Martes martes

L., 1758 in a fir forest of Pyrenean mountains Northeastern Spain In: Mathias,

M. L., Santos-Reis, M., Amori, G., Libois, R., Mitchell-Jones, A., Saint-Girons,

M.C. Eds.. Proceedings of the I European Congress of Mammalogy. Lisboa ,

Portugal: 189-196.

174 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Sato, Y., T. Mano, and S. Takatsuki. 2005. Stomach contents of brown bears Ursus

arctos in Hokkaido, Japan. Wildlife Biology 11: 133-144.

Schmidly, D. J. 1974. Peromyscus attwateri. Mammalian Species 48:1-3.

Sowls, L. K. 1997. Javelinas and Other Peccaries: Their Biology, Management, and

Use. Texas A&M University Press.

Tevis, L., Jr. 1953. Stomach contents of chipmunks and mantled squirrels in

northeastern California. Journal of Mammalogy 34: 316-324.

Torres-Mura, J. C., M. L. Lemus, and L. C. Contreras. 1989. Herbivorous specialization

of the South American desert rodent Tympanoctomys barrerae. Journal of

Mammalogy 70:646-648.

Varty, N. 1990. Ecology of the small mammals in the riverine forests of the Jubba

Valley, Southern Somalia. Journal of Tropical Ecology 6:179-189.

Vernes, K. 1995. The diet of the red-legged pademelon Thylogale stigmatica Gould

(Marsupialia : Macropodidae) in fragmented tropical rainforest, north Queensland,

Australia. Mammalia 59: 517-526.

Whitaker, J. O., Jr. 1963. Food, habitat and parasites of the woodland jumping mouse in

central New York. Journal of Mammalogy 44: 316-321.

Whitaker, J. O., Jr. 1966. Food of Mus musculus, Peromyscus maniculatus bairdi and

Peromyscus leucopus in Vigo County, Indiana. Journal of Mammalogy 47: 473-

486.

Whitten, J. E. J. 1981. Ecological separation of three diurnal squirrels in tropical

rainforest on Siberut Island, Indonesia. Journal of Zoology 193: 405-420.

Silvia Pineda-Munoz - April 2016 175

Diet, ecology, and dental morphology in terrestrial mammals

Wydeven, P. R. and R. B. Dahlgren. 1982. A comparison of prairie dog stomach contents and feces using a microhistological technique. Journal of Wildlife Management

46: 1104-1108.

Table 11.1-2: Summary of stomach content data for the 139 mammalian species analysed in this study. U=unspecified; A= % seeds, B= % invertebrates; C= % fungi; D= % vertebrates; E= % flowers and gums; F= % roots and tubers; G= % green plants, vegetation; H= % fruit; N= sample size.

Species A B C D E F G H N

Aepyceros melampus 0 0 0 0 0 0 100 0 1737

Ammotragus lervia 0 0 0 0 0 0 100 0 109

Antidorcas marsupialis 0 0 0 0 0 0 98.5 1.5 U

Cephalophus callipygus 0 0 0.1 0 0.9 0 16.3 82.7 14

Cephalophus dorsalis 0 0.1 0.1 0 0.4 0 26.57 72.83 8

Cephalophus leucogaster 0 0.1 0.1 0 2.2 0 24.7 72.9 9

Cephalophus nigrifrons 0 0.1 0.4 0 0.2 0 27.7 71.6 7

Cephalophus silvicultor 0 0.1 0 0 0 0 28.6 71.3 4

Hyemoschus aquaticus 0 0.2 0.7 0 0.1 0 30.37 68.63 19

Philantomba monticola 0 0.5 0.2 0 0.6 0 20.28 78.42 16

Tragelaphus spekii 0 0 0 0 0 0 99.9 0.1 3

Tayassu pecari 0 0 0 0 0 0 39 61 34

Canis lupus dingo 0 0 0 100 0 0 0 0 230

Canis mesomelas 0 68.4 0 29.2 0 0 1.8 0.6 11

Otocyon megalotis 0 56.72 0 0.9 0 0 26.82 15.55 185

Urocyon cinereoargenteus 31.11 3.65 0 45.93 0 0 1.25 18.06 63

Vulpes lagopus 0 0 0 100 0 0 0 0 100

Vulpes vulpes 1.36 0.94 0 84.34 0 0 0 13.36 128

Helogale hirtula 0 100 0 0 0 0 0 0 5

Galictis vittata 0 0 0 100 0 0 0 0 U

Martes martes 0 0.1 0 53.9 0 0 0 46 42

Meles meles 4.47 17.1 0 7.06 0 0 70.87 0.49 281

Nandinia binotata 0 10 0 0 0 0 0 90 U

Procyon lotor 67 25 0 4 0 0 0 4 23

176 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Ursus arctos 13.22 6.81 0 7.85 0 0.18 56.32 15.62 532

Cercartetus caudatus 0 100 0 0 0 0 0 0 U

Thylogale stigmatica 0 0 0 0 0 0 100 0 16

Eremitalpa granti 0 93 0 0 0 7 0 0 19

Blarina brevicauda 0.65 96.8 0 0 0 0 2.55 0

Crocidura cyanea 0 100 0 0 0 0 0 0 23

Crocidura flavecens 0 93.3 0 0 0 0 6.7 0 5

Crocidura mariquensis 0 100 0 0 0 0 0 0 U

Myosorex cafer 0 74.8 0 0 0 0 25.2 0 7

Myosorex varius 0 97.1 0 0 0 0 2.9 0 U

Sorex fumeus 0 99 0 0 0 0 1 0 4

Sorex hoyi 0 100 0 0 0 0 0 0 10

Sorex palustris 0 87 0 0 0 0 13 0 U

Parascalops breweri 0 84 0 0 0 2 14 0 U

Scapanus townsendii 1.91 72.46 0 0.59 0 18.41 6.64 0 100

Talpa europaea 0 100 0 0 0 0 0 0 252

Elephantulus 0 100 0 0 0 0 0 0 U brachyrhynchus

Elephantulus rufescens 0.3 63 0 0 0 0 36.7 0 23

Macroscelides proboscideus 0 98 0 0 0 0 1 1 7

Dromiciops gliroides 4.96 71.63 2.23 0 0.1 0 18.64 2.43 71

Caenolestes fuliginosus 0 96.9 0 2.5 0 0 0.6 0 38

Rhyncholestes raphanurus 3.84 54.6 7.99 0 0.51 0 31.75 1.31 11

Manis crassicaudata 0 100 0 0 0 0 0 0 31

Alouatta seniculus 0 0.11 0.12 0 0.35 0 53.67 45.75 U

Ateles paniscus 0 0.08 0 0.08 0 0 9.6 90.24 60

Cebus apella 0 26.74 0 0.19 0.04 0 4.75 68.28 44

Cercocebus agilis 0 6.1 0 0 5.1 0 6.1 82.7 97

Cercopithecus cephus 0 12.6 0 0 0 0 6.1 81.3 15

Cercopithecus neglectus 0 4.9 5.4 0 3 0 9.4 77.3 62

Cercopithecus nictitans 0 9.6 0.4 0 1 0 17 72 9

Cercopithecus pogonias 0 16.1 0.1 0 0.1 0 1.2 82.5 100

Colobus guereza 0 0.1 0 0 0 0 50.8 49.1 52

Silvia Pineda-Munoz - April 2016 177

Diet, ecology, and dental morphology in terrestrial mammals

Mandrillus sphinx 0 4 1 0 0 0 3 92 5

Miopithecus talapoin 0 36.2 0 0 1.5 0 2 60.3 U

Euoticus elegantulus 0 20 0 0 75 0 0 5 9

Galago alleni 0 23.47 0 2.04 0 0 0 74.49 52

Galago demidoff 0 70.71 0 0 10.1 0 0 19.19 12

Arctocebus calabarensis 0 85.86 0 0 0 0 0 14.14 55

Perodictus potto 0 10.42 0 0 21.88 0 0 67.71 14

Loxodonta africana 0 0 0 0 0 0 96 4 41

Chaetodipus baileyi 88.1 8.9 0 0 0 0 3 0 71

Chaetodipus intermedius 82.4 16.2 0 0 0 0 1.4 0 330

Dipodomys merriami 78.4 15.5 0 0 0 0 6.1 0 783

Perognathus amplus 94.3 3.7 0 0 0 0 2 0 1054

Abrothrix longipilis 16.1 51.27 1.86 1.12 0 2.49 27.16 0 714

Abrothrix sanborni 4.62 37.1 25.69 0 0.62 0 29.39 2.57 5

Acomys cahirinus 42.44 7.01 0 0 0 0 32.23 18.32 55

Acomys dimidiatus 53 13 0 0 3 0 31 0 17

Aethomys crisophilus 58.3 4.2 0 0 0 0 37.5 0 30

Akodon azarae 33.92 53.18 0 0 0 0 12.9 0 10

Arvicanthis niloticus 0 0.1 0 0 0 0 99.9 0 81

Calomys laucha 59.62 27.84 0 0 0 0 12.54 0 22

Calomys musculinus 57.08 27.32 0 0 0 0 15.6 0 139

Geoxus valdivianus 1.91 55.87 8.12 0 0.2 0 33.9 0 141

Gerbilliscus leucogaster 25.5 24.9 0 0 0 0 49.6 0 18

Gerbilliscus robustus 25.93 14.21 0 0 6.41 0 27.33 26.13 5

Gerbillurus paeba 0 8.41 0 0 0 0 91.59 0 4

Grammomys dolichurus 56 2 7 0 0 0 35 0 42

Irenomys tarsalis 28.5 2.38 3.21 0 7.56 0 40.83 17.51 8

Lemniscomys rosalia 15.4 0 0 0 0 0 84.6 0 10

Lorentzimys nouhuysi 0 38.88 21.28 0 0 0 39.85 0 9

Loxodontomys micropus 19.37 2.75 5.81 0 19.47 0 24.57 28.03 58

Mastomys natalensis 47 9.5 0 0 0 0 43.5 0 16

Maxomys surifer 0 40.94 0 0 0 0 52.57 6.49 59

Micaelamys namaquensis 40.2 0 0 0 0 0 59.8 0 45

Microtus ochrogaster 1.15 2.65 0 0 0 0 96.2 0 6

178 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Mus minutoides 7.61 8.81 0 0 0 0 83.58 0 57

Myotomys unisulcatus 0 0 0 0 0 0 100 0 69

Necromys obscurus 36.09 43.07 0 0 0 0 20.84 0 16

Oligoryzomys flavescens 47.07 41.57 0 0 0 0 11.36 0 43

Oryzomys longicaudatus 39.02 14.27 2.44 0 5 0 26.46 12.8 26

Otomys angoniensis 0 0 0 0 0 0 100 0 37

Otomys irroratus 0 0 0 0 0 0 100 0 U

Peromyscus attwateri 60.36 31.03 0 0 0 2.6 1.6 4.4 5

Peromyscus leucopus 40.02 37.31 0.23 0.4 0 0 22.04 0 U

Phyllotis darwini 41.65 2.38 0.03 3.1 0 0 52.85 0 27

Pseudohydromys ellermani 0 61.5 15.3 0 0 0 23.2 0 147

Pseudohydromys fuscus 0 52.4 21.9 0 0 0 25.7 0 9

Pseudohydromys murinus 0 66.3 23.1 0 0 0 10.6 0 7

Pseudohydromys 0 99.5 0 0 0 0 0.5 0 24 occidentalis

Rattus rattus 61.2 0.4 0 0 0 0 38.4 0 U

Rhabdomys pumilo 32.83 5.46 0 0 0 0 61.7 0 U

Taterillus harringtoni 24.92 2.3 0 0 3.45 0 26.98 42.34 101

Dendromus mesomelas 87 1.3 0 0 0 0 11.7 0 27

Dendromus mystacalis 39.7 16.2 0 0 0 0 44.1 0 5

Malacothrix typica 16 11.8 0 0 0 0 72.2 0 6

Saccostomus campestris 66.7 0 0 0 0 0 33.3 0 6

Saccostomus mearnsi 44 3 0 0 0 0 53 0 U

Steatomys pratensis 38.8 0 0 0 0 0 61.2 0 U

Ctenomys mendocinus 0 0 0 0 0 0 100 0 U

Octodon degus 34.07 0.42 0 3.39 0 0 62.11 0 U

Tympanoctomys barrerae 0 0 0 0 0 0 100 0 74

Callosciurus melanogaster 23.71 63.92 0 0 0 0 12.37 0 U

Citellus lateralis 4.33 6 61.33 2 3.33 1 21.67 0.33 3

Cynomys ludovicianus 0 0 0 0 0 0 100 0 273

Epixerus ebii 0 1.6 0 0 0 0 0 98.4 U

Funisciurus anerythrus 0 19.5 0.6 0 0 0 3 76.9 8

Funisciurus isabella 0 5.53 1.64 0 0 0 9.93 82.91 15

Funisciurus lemniscatus 0 36.76 0.1 0 0 0 3.16 59.98 14

Silvia Pineda-Munoz - April 2016 179

Diet, ecology, and dental morphology in terrestrial mammals

Funisciurus pyrropus 0 13.1 1.8 0 1.8 0 0.7 82.6 15

Heliosciurus rufobrachium 0 4.53 0 0 0.2 0 6.24 89.03 12

Lariscus obscurus 71.29 21.78 0 0 0 0 6.93 0 15

Myosciurus pumilio 0 36.8 0 0 0 0 29.9 33.3 3

Neotamias townsendi 8 8 72.33 0 10 0 0 1.67 6

Paraxeros poensis 0 11.15 0 0 0 0 0 88.85 48

Protoxerus stangeri 0 0.31 3.3 0 0 0 8.94 87.44 16

Sundasciurus lowii 0 40.21 0 0 0 0 59.79 0 13

Tamias amoenus 41.33 20.67 27.67 0.17 2.67 0.33 7 0.17 7

Tamias quadrimaculatus 14.33 6.67 66 0.33 3.67 0 7.33 1.67 170

Tamias speciosus 29.33 15.67 31.67 0.33 21.67 0 1 0.33 165

Xerus rutilus 53.51 5.01 0 0 7.61 0 22.86 11.02 126

Napaeozapus insignis 24 19 37 0 0 0 8 12 6

Tupaia glis 0 66.5 0 0 0 0 20.9 12.6 103

Chaetophractus vellerosus 0 40.68 0 21.35 0 0 37.96 0 24

Dasypus novemcinctus 0 88.93 0 8.58 0 0 0 2.5 84

Table 11.1-3: Feeding resources included in each of the feeding categories

Feeding categories Feeding resources Seeds Seeds and their tissues Invertebrates , worms and molluscs Vertebrates Vertebrate animals Fungi Fungi and lichens Flowers & Gum Flowers and gum

Roots & Tubers Roots and tubers

Green Plants Angiosperm leaves and branches and

Fruit Fruits and their tissues

Table 11.1-4: Importance of the components and loadings of the variables based on a principal component analysis of the stomach content data for the 139 species in this study.

Importance of components: Comp.1 Comp.2 Comp.3 Comp.4 Comp.5 Comp.6 Comp.7 Comp.8

180 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Standard 1.241 1.180 1.098 1.041 0.991 0.945 0.919 0.000 deviation Proportion of 0.194 0.175 0.152 0.136 0.124 0.112 0.106 0.000 Variance Cumulative 0.194 0.369 0.521 0.658 0.781 0.894 1.000 1.000 Proportion

Loadings: Feeding resources Comp.1 Comp.2 Comp.3 Comp.4 Comp.5 Comp.6 Comp.7 Comp.8

Seeds 0.241 0.436 -0.328 -0.33 0.626 -0.369

Invertebrates -0.691 0.235 0.174 0.122 0.35 0.148 -0.527

Fungi -0.611 0.145 -0.438 -0.609 -0.174

Vertebrate -0.17 -0.828 -0.422 -0.15 -0.267

Flowers and gum -0.166 -0.604 0.168 -0.217 0.715 -0.11

Roots and tubers -0.339 0.179 0.163 0.117 -0.888 -0.139

Green plants 0.579 0.272 0.296 0.324 -0.365 0.129 -0.488

Fruit -0.767 0.152 0.309 -0.214 -0.483

Silvia Pineda-Munoz - April 2016 181

Diet, ecology, and dental morphology in terrestrial mammals

11.2 Supplementary information 2 Table 11.2-1: Dataset summarising body mass and diet diversity data for the 139 species in the study. Diet data as in Table 11.1.1.

Diet diversity Order Family Species Diet Body mass (g) (Inv simpson)

Afrosoricida Chrysochloridae Eremitalpa granti Insectivore 26.1 0.253638947 Artiodactyla Bovidae Aepyceros melampus Herbivore 52500.084 0 Artiodactyla Bovidae Ammotragus lervia Herbivore 48000 0 Artiodactyla Bovidae Antidorcas marsupialis Herbivore 31500 0.077882509 Artiodactyla Bovidae Cephalophus callipygus Frugivore 18249.879 0.502074492 Artiodactyla Bovidae Cephalophus dorsalis Frugivore 20000 0.618958648 Artiodactyla Bovidae Cephalophus leucogaster Frugivore 12699.891 0.673603275 Artiodactyla Bovidae Cephalophus nigrifrons Frugivore 13899.846 0.636215958 Artiodactyla Bovidae Cephalophus silvicultor Frugivore 72500.333 0.606101368 Artiodactyla Bovidae Hyemoschus aquaticus Frugivore 10850.007 0.674344771 Artiodactyla Bovidae Philantomba monticola Frugivore 6250 0.583823388 Artiodactyla Bovidae Tragelaphus spekii Herbivore 77999.173 0.007907255 Artiodactyla Tayassuidae Tayassu pecari Frugivore 32233.688 0.668748087 Carnivora Canidae Canis lupus dingo Carnivore 32183.333 0 Carnivora Canidae Canis mesomelas Insectivore 8500.021 0.722242705 Carnivora Canidae Otocyon megalotis Insectivore 4150.018 1.006382099 Mixed Carnivora Canidae Urocyon cinereoargenteus feeder 3833.717 1.205310317 Carnivora Canidae Vulpes lagopus Carnivore 4867.67 0 Carnivora Canidae Vulpes vulpes Carnivore 5476.174 0.514885637 Carnivora Herpestidae Helogale hirtula Insectivore 289.001 0 Carnivora Mustelidae Galictis vittata Carnivore 3200 0 Carnivora Mustelidae Martes martes Carnivore 1300 0.697234401 Carnivora Mustelidae Meles meles Herbivore 13000 0.898132982 Carnivora Nandinidae Nandinia binotata Frugivore 2000 0.325082973 Carnivora Procyonidae Procyon lotor Granivore 5524.971 0.872403626 Carnivora Ursidae Ursus arctos Herbivore 180520.422 1.274948149 Mixed Dasypodidae Chaetophractus vellerosus feeder 1030.007 1.0632599 Cingulata Dasypodidae Dasypus novemcinctus Insectivore 4203.78 0.407316679 Diprotodontia Burramyidae Cercartetus caudatus Insectivore 25 0 Diprotodontia Macropodidae Thylogale stigmatica Herbivore 4600 0 Elephantulus Macroscelidea Macroscelidae brachyrhynchus Insectivore 47.5 0 Macroscelidea Macroscelidae Elephantulus rufescens Insectivore 37.5 0.676388158 Macroscelidea Macroscelidae Macroscelides proboscideus Insectivore 38.2 0.111902057 Microbiotheria Microbiotheriidae Dromiciops gliroides Insectivore 22.3 0.883147243 Paucituberculata Caenolestidae Caenolestes fuliginosus Insectivore 27.8 0.153432418

182 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Paucituberculata Caenolestidae Rhyncholestes raphanurus Insectivore 21 1.105453867 Pholidota Manidae Manis crassicaudata Insectivore 3900 0 Primates Atelidae Alouatta seniculus Herbivore 6145.539 0.727108796 Primates Atelidae Ateles paniscus Frugivore 7900.053 0.329050651 Primates Cebidae Cebus apella Frugivore 2500 0.772996301 Primates Cercopithecidae Cercocebus agilis Frugivore 8000 0.650081077 Primates Cercopithecidae Cercopithecus cephus Frugivore 3484.978 0.599926061 Primates Cercopithecidae Cercopithecus neglectus Frugivore 5479.993 0.83187961 Primates Cercopithecidae Cercopithecus nictitans Frugivore 5803.345 0.830860217 Primates Cercopithecidae Cercopithecus pogonias Frugivore 3764.957 0.519639003 Primates Cercopithecidae Colobus guereza Herbivore 10600 0.700216637 Primates Cercopithecidae Mandrillus sphinx Frugivore 18249.879 0.356714552 Primates Cercopithecidae Miopithecus talapoin Frugivore 1250 0.814088606 Primates Galagidae Euoticus elegantulus Gumivore 300 0.687435751 Primates Galagidae Galago alleni Frugivore 260 0.638963544 Primates Galagidae Galago demidoff Insectivore 61 0.793409928 Primates Lorisidae Arctocebus calabarensis Insectivore 210 0.407496774 Primates Lorisidae Perodicticus potto Frugivore 1099.993 0.832172906 Proboscidea Elephantidae Loxodonta africana Herbivore 3940034.276 0.167944148 Rodentia Abrothrix longipilis Insectivore 37.55 1.206944279 Mixed Rodentia Cricetidae Abrothrix sanborni feeder 24.7 1.344595536 Rodentia Cricetidae Akodon azarae Insectivore 25 0.966741138 Rodentia Cricetidae Calomys laucha Granivore 14 0.924692572 Rodentia Cricetidae Calomys musculinus Granivore 20.1 0.964380178 Mixed Rodentia Cricetidae Irenomys tarsalis feeder 43.15 1.423199571 Mixed Rodentia Cricetidae Loxodontomys micropus feeder 72.699 1.602073909 Rodentia Cricetidae Microtus ochrogaster Herbivore 38.013 0.184832051 Mixed Rodentia Cricetidae Necromys obscurus feeder 40.7 1.057443031 Rodentia Cricetidae Peromyscus attwateri Granivore 27.9 0.966328821 Mixed Rodentia Cricetidae Peromyscus leucopus feeder 21.175 1.103713452 Rodentia Cricetidae Phyllotis darwini Herbivore 50.833 0.900926705 Rodentia Ctenomyidae Ctenomys mendocinus Herbivore 163.05 0 Rodentia Heteromyidae Chaetodipus baileyi Granivore 26.35 0.432118952 Rodentia Heteromyidae Chaetodipus intermedius Granivore 16.5 0.514141353 Rodentia Heteromyidae Dipodomys merriami Granivore 42 0.650364408 Rodentia Heteromyidae Perognathus amplus Granivore 11.7 0.255567166 Mixed Rodentia Muridae Acomys cahirinus feeder 52.454 1.225912549 Rodentia Muridae Acomys dimidiatus Granivore 52.454 1.069977653 Rodentia Muridae Aethomys chrysophilus Granivore 74.85 0.815522766 Rodentia Muridae Arvicanthis niloticus Herbivore 106.9 0.007907255 Rodentia Muridae Geoxus valdivianus Insectivore 31.5 0.983866592

Silvia Pineda-Munoz - April 2016 183

Diet, ecology, and dental morphology in terrestrial mammals

Mixed Rodentia Muridae Gerbilliscus leucogaster feeder 72.833 1.042425644 Mixed Rodentia Muridae Gerbilliscus robustus feeder 96 1.508523812 Rodentia Muridae Gerbillurus paeba Herbivore 23.3 0.288670533 Rodentia Muridae Grammomys dolichurus Granivore 41.175 0.956524764 Rodentia Muridae Lemniscomys rosalia Herbivore 59.3 0.4295852 Mixed Rodentia Muridae Lorentzimys nouhuysi feeder 14.2 1.063215533 Mixed Rodentia Muridae Mastomys natalensis feeder 54.54 0.940577084 Rodentia Muridae Maxomys surifer Herbivore 150.365 0.881153355 Rodentia Muridae Micaelamys namaquensis Herbivore 50.625 0.673814268 Rodentia Muridae Mus minutoides Herbivore 7.011 0.559945157 Rodentia Muridae Myotomys unisulcatus Herbivore 124.5 0 Mixed Rodentia Muridae Oligoryzomys flavescens feeder 23.767 0.966674661 Oligoryzomys Mixed Rodentia Muridae longicaudatus feeder 27 1.500420131 Rodentia Muridae Otomys angoniensis Herbivore 121.9 0 Rodentia Muridae Otomys irroratus Herbivore 118.017 0 Rodentia Muridae Pseudohydromys ellermani Insectivore 22.5 0.925157512 Rodentia Muridae Pseudohydromys fuscus Insectivore 18.75 1.020414374 Rodentia Muridae Pseudohydromys murinus Insectivore 16.8 0.848870425 Pseudohydromys Rodentia Muridae occidentalis Insectivore 21.9 0.031479066 Rodentia Muridae Rattus rattus Granivore 178.25 0.690123204 Rodentia Muridae Rhabdomys pumilio Herbivore 41.117 0.822354201 Mixed Rodentia Muridae Taterillus harringtoni feeder 52 1.266550469 Rodentia Nesomydae Dendromus mesomelas Granivore 11.12 0.428647493 Mixed Rodentia Nesomydae Dendromus mystacalis feeder 7.717 1.022673179 Rodentia Nesomydae Malacothrix typica Herbivore 13.4 0.780564533 Rodentia Nesomydae Saccostomus campestris Granivore 51.225 0.636282869 Rodentia Nesomydae Saccostomus mearnsi Herbivore 45.5 0.802913664 Rodentia Nesomydae Steatomys pratensis Herbivore 48 0.66784505 Rodentia Octodontidae Octodon degus Herbivore 210 0.800351185 Rodentia Octodontidae Tympanoctomys barrerae Herbivore 70 0 Rodentia Sciuridae Callosciurus melanogaster Insectivore 278 0.885838047 Rodentia Sciuridae Cynomys ludovicianus Herbivore 1364 0 Rodentia Sciuridae Epixerus ebii Frugivore 591.998 0.082033977 Rodentia Sciuridae Funisciurus anerythrus Frugivore 218.002 0.656658931 Rodentia Sciuridae Funisciurus isabella Frugivore 109.001 0.612272685 Rodentia Sciuridae Funisciurus lemniscatus Frugivore 141 0.790545568 Rodentia Sciuridae Funisciurus pyrropus Frugivore 297.002 0.603522392 Rodentia Sciuridae Heliosciurus rufobrachium Frugivore 371.501 0.429167184 Rodentia Sciuridae Lariscus obscurus Granivore 200 0.758204619 Mixed Rodentia Sciuridae Myosciurus pumilio feeder 164.999 1.09503668

184 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Rodentia Sciuridae Paraxerus poensis Frugivore 124.5 0.349640003 Rodentia Sciuridae Protoxerus stangeri Frugivore 621.498 0.463652217 Rodentia Sciuridae Spermophilus lateralis Fungivore 190.999 1.192439315 Rodentia Sciuridae Sundasciurus lowii Herbivore 87.5 0.673853962 Mixed Rodentia Sciuridae Tamias amoenus feeder 50.5 1.369939804 Rodentia Sciuridae Tamias quadrimaculatus Fungivore 85.249 1.133276898 Mixed Rodentia Sciuridae Tamias speciosus feeder 62 1.429472458 Rodentia Sciuridae Tamias townsendii Fungivore 74.776 0.937016721 Rodentia Sciuridae Xerus rutilus Granivore 317.497 1.260975373 Mixed Rodentia Zapodidae Napaeozapus insignis feeder 22.3 1.482410112 Scandentia Tupiidae Tupaia glis Insectivore 169.833 0.859477518 Soricomorpha Soricidae Blarina brevicauda Insectivore 28 0.157777604 Soricomorpha Soricidae Crocidura cyanea Insectivore 9.35 0 Soricomorpha Soricidae Crocidura flavescens Insectivore 28.49 0.245808821 Soricomorpha Soricidae Crocidura mariquensis Insectivore 11.25 0 Soricomorpha Soricidae Myosorex cafer Insectivore 13.4 0.564521721 Soricomorpha Soricidae Myosorex varius Insectivore 13.2 0.131248699 Soricomorpha Soricidae Sorex fumeus Insectivore 7.7 0.056001534 Soricomorpha Soricidae Sorex hoyi Insectivore 2.575 0 Soricomorpha Soricidae Sorex palustris Insectivore 13.45 0.386386706 Soricomorpha Talpidae Parascalops breweri Insectivore 51 0.499953105 Soricomorpha Talpidae Scapanus townsendii Insectivore 141.7 0.830947623 Soricomorpha Talpidae Talpa europaea Insectivore 77 0

Table 11.2-2: P-values of pairwise comparisons based on the Wilcoxon rank-sum test for body mass distributions in diet categories from Wilman et al. (2014) data.

Carnivore Frugivore Generalist Granivore Gumivore Herbivore Frugivore <0.05 - - - - - Generalist <0.05 <0.05 - - - - Granivore <0.05 <0.05 <0.05 - - - Gumivore <0.05 <0.05 0.085 <0.05 - - Herbivore <0.05 0.615 <0.05 <0.05 0.142 - Insectivore <0.05 <0.05 <0.05 <0.05 <0.05 <0.05

Table 11.2-3: Importance of the components and loadings of the variables based on a principal component analysis of the stomach content data for the 139 species in this study.

Importance of components:

Silvia Pineda-Munoz - April 2016 185

Diet, ecology, and dental morphology in terrestrial mammals

Comp.1 Comp.2 Comp.3 Comp.4 Comp.5 Comp.6 Comp.7 Comp.8

Standard deviation 1.241 1.180 1.098 1.041 0.991 0.945 0.919 0.000

Proportion of 0.194 0.175 0.152 0.136 0.124 0.112 0.106 0.000 Variance Cumulative 0.194 0.369 0.521 0.658 0.781 0.894 1.000 1.000 Proportion

Loadings: Feeding resources Comp.1 Comp.2 Comp.3 Comp.4 Comp.5 Comp.6 Comp.7 Comp.8

Seeds 0.241 0.436 -0.328 -0.33 0.626 -0.369

Invertebrates -0.691 0.235 0.174 0.122 0.35 0.148 -0.527

Fungi -0.611 0.145 -0.438 -0.609 -0.174

Vertebrate -0.17 -0.828 -0.422 -0.15 -0.267

Flowers and gum -0.166 -0.604 0.168 -0.217 0.715 -0.11

Roots and tubers -0.339 0.179 0.163 0.117 -0.888 -0.139

Green plants 0.579 0.272 0.296 0.324 -0.365 0.129 -0.488

Fruit -0.767 0.152 0.309 -0.214 -0.483

Table 11.2-4: Importance of the components and loadings of the variables based on a principal component analysis for Wilman et al. (2014) data.

Comp.1 Comp.2 Comp.3 Comp.4 Comp.5 Comp.6 Seeds 0.229 0.553 -0.320 -0.321 0.582 -0.312 Invertebrate -0.595 -0.055 -0.010 -0.543 -0.292 -0.512 Vertebrate -0.461 -0.138 -0.202 0.672 0.373 -0.370 Flowers.Gum -0.074 0.104 0.917 -0.020 0.356 -0.125 Green.Plants 0.597 -0.506 0.044 0.010 -0.057 -0.618 Fruit 0.136 0.636 0.119 0.388 -0.554 -0.324

Importance of components: Comp.1 Comp.2 Comp.3 Comp.4 Comp.5 Comp.6 Standard deviation 1.347 1.166 1.016 0.975 0.917 0.000 Proportion of Variance 0.302 0.227 0.172 0.159 0.140 0.000 Cumulative Proportion 0.302 0.529 0.701 0.860 1.000 1.000

186 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Figure 11.2-1: Correlation between the percentage of fruit in the diet of the frugivore species in Pineda-Munoz and Alroy (2014) and the values for the same species provided by Wilman et al. (2014). Correlation in the top-right corner of the graph.

Figure 11.2-2: Boxplot illustrating the body mass range for each dietary category in Wilman et al. (2014) data. Diet categories are as described by Pineda-Munoz and Alroy (2014).

Silvia Pineda-Munoz - April 2016 187

Diet, ecology, and dental morphology in terrestrial mammals

Figure 11.2-3: Scores of the two first axes of the principal component analysis of dietary data for the species in Wilman et al. (2014). Raw data are percentage of food resource rescaled to z- scores. Arrows indicate loadings on the two first components. Colours and sizes indicate the log10 body mass (g) of each species as shown in the legend.

Figure 11.2-4: Relationship between log10 body mass (g) and diet diversity indices calculated by applying inverse Simpson indices to stomach content percentages for Wilman et al. (2014) data. Colours represent different dietary specializations.

188 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

11.3 Supplementary information 3

11.3.1 Proxies measure diet at different temporal scales The utility of different proxies depends on the nature of the research question and the temporal scale examined (Kronfeld and Dayan 1998) but operational limitations need to be considered (Reynolds-Hogland and Mitchell 2007). Most dietary sampling is unfortunately opportunistic and we often do not have a choice of which proxy we will use. While feeding observations are commonly employed to study extant animals

(Table S1), the proxy is impossible to apply to fossil species excepting what can be gleaned from prehistoric cave art (Guthrie et al. 2000). Proxies like feeding site analysis (Rawn-Schatzinger 1992) and scat (Hansen 1978) and stomach contents

(Kriwet et al. 2008) can be used with extinct species but methods involving the morphological and/or isotopic examination of dentition have a much larger potential sample size relying only on the robust fossil record of isolated teeth (Kimura et al.

2013). While all the dietary proxies discussed here are feasible to use with extant animals (Table 11.3-1), certain proxies require destructive sampling and would be unsuitable for use on rare or endangered species (Pineda-Munoz and Alroy 2014). This is why most conservation biologists study diet through non-invasive methods like feeding observations or sampling dung and shed hair (Del Rio et al. 2009). No one could justify killing an endangered species to take isotopic samples of bone collagen but for a fossil species known only by fragmentary postcrania, bone collagen is the only proxy possible. These different operational limitations force researchers to favour certain proxies over others creating a nonrandom bias in which temporal scales of diet are considered by each disciplinary subfield (Table 11.3-1). If these biases remain unacknowledged, they hamper our understanding of general ecological and evolutionary

Silvia Pineda-Munoz - April 2016 189

Diet, ecology, and dental morphology in terrestrial mammals

patterns by reducing meaningful communication between subfields. Even if we use a proxy because it is the only method available, it is crucial to understand the scale at which that proxy records diet so that we can ensure that the scale of our question does not exceed that of our data (Reynolds-Hogland and Mitchell 2007).

Direct observation of feeding habits is the only way to record diet at small temporal scales less than an hour. This may be the only scale that captures perceived dietary aberrations like the significant, albeit small by volume, consumption of meat

(Ellis-Felege et al. 2008) and bone (Hutson et al. 2013) by otherwise herbivorous artiodactyls. Observations and feeding site investigations are also probably the only proxies that can measure cyclical diel changes in diet despite the difficulty of close monitoring at night (Munro et al. 2006, Salter and Hudson 1979).

Stomach and scat content analyses using species specific correction factors to adjust for differential digestibility of food (Klare et al. 2011), provide the most detailed reconstructions of diet as the material consumed can be identified by genetics

(Willerslev et al. 2014), isotopes (Sponheimer et al. 2003), or morphology (Shrestha and Wegge 2006). Although the contents of a single stomach or scat likely record an average diet of hours to days (Kararli 1995), many samples are often collected throughout a year and averaged together to generate qualitative diet assignments

(Cerling et al. 2003, Gagnon and Chew 2000, Mattson et al. 1991). Small scratches and pits formed on teeth called dental microwear also record about a week of diet and like stomach contents, they record the diet during the last week the animal was alive

(Münzel et al. 2014, Teaford and Oyen 1989). This could lead to a taphonomic megabias in dietary reconstruction where fossil microwear is recording an atypical diet of starvation but initial analysis of primates shows that this is unlikely as mortality rates

190 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

are not systematically higher during periods of resource stress (Gogarten and Grine

2013). Unlike the quickly overwritten patterns of microwear, tooth fractures (Van

Valkenburgh 2009) and “mesowear” that visibly changes tooth cusp shapes (Fortelius and Solounias 2000) probably reflect the rigors of diet over a substantial portion of an animal’s lifespan and may operate at temporal scales similar to many isotopic proxies.

Ever growing tissues like tusks and hair can be sampled from living animals, museum specimens, and exceptionally preserved fossils (Sponheimer et al. 2003).

These tissues record carbon, nitrogen, and sulphur stable isotopic signatures on a wide range of temporal scales that can track changes in diet associated with seasonality, ontogeny, and long term migrations (Cerling et al. 2008, Hoppe et al. 2004). Unlike most other sources of stable isotopes, these tissues are not remodelled throughout the animal’s lifetime so they incorporate an isotopic signal of diet only with new growth and therefore can be subsampled to show finer temporal resolutions (West et al. 2006).

The time period integrated by the tusks or hair is as long as that structure has been growing, sometimes the animal’s entire lifespan (Cerling et al. 2008) and the resolution is limited only by how fast the tissue grows with daily resolutions achievable (Hoppe et al. 2004). Teeth that aren’t ever growing can also be serially sectioned but they will only record an isotopic signal during the few years they are forming, which may represent a juvenile diet depending on the species (Zazzo et al. 2002). However, if multiple teeth are sampled from the same specimen, their isotopic curves can be lined up to form a longer, multiyear signal, even if the exact eruption sequence for the species is unknown (J and Kitchener 2005, Lieberman et al. 2004).

Although alterations in the hardness and nutrition of diet can cause significant morphological changes over an animals lifetime (J and Kitchener 2005, Lieberman et al.

Silvia Pineda-Munoz - April 2016 191

Diet, ecology, and dental morphology in terrestrial mammals

2004), sometimes in periods as short as several months (Hummel et al. 2011, Kaiser et al. 2013), major differences in masticatory morphology like degree of hypsodonty (how high crowned teeth are) are likely the result of many thousands or millions of years of evolution (Dickman and Huang 1988, Jones et al. 1981). Unlike other proxies such as scat and isotopes, whose temporal scales can be elucidated relatively easily with diet switch experiments (Gogarten and Grine 2013), it is harder to constrain which specific scales a certain morphological feature represents. Without relatively complete, well calibrated fossil sequences, it is difficult to determine whether a plastic morphological feature is the result of an ancient phenotypic bottleneck that occurred over a relatively short period of time or whether it represents morphology adapted over millions of years to an average diet (Anderson et al. 2011). Using any morphological feature as a dietary proxy requires the difficult burden of proof that the feature is a result of long term adaption rather than a short term effect or vice versa. Gross dental or jaw morphology may be appropriate proxies for studies examining major evolutionary trends or adaptive radiations that took place over millions of years (Anderson et al. 2011) but they are probably best used in conjunction with other proxies as the upper limit on a multi-scale analysis. A strong congruence between diet observed in the field and diet inferred from morphology may represent a morphological optimum in functionality or a diet that has remained relatively constant over millions of years.

(References in main References section)

11.3.2 Temporal dataset preparation David Watts (personal communication) and his field assistants recorded the amount of time per day that chimpanzees (Pan troglodytes) spent feeding on different resources at the Ngogo research site in Kigali National Park, Uganda from April 7 to

192 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

July 28, 2013. We calculated percentage diets by dividing the time spent feeding on a category by the total time spent feeding that day. We used the na.approx() function in the R package zoo (Zeileis and Grothendieck 2005) to linearly interpolate values for 7 days that lacked observational data.

Mattson et al. (1991) opportunistically collected 3,423 scat samples while tracking 96 radio collared brown bears (Ursus arctos horribilis) in Yellowstone

National Park, USA from 1977-1987. We extracted mean monthly percentage volume diets from Figure 2 of Mattson et al. (1991) with Plot Digitizer 2.6.4 (Huwaldt and

Steinhorst 2001) and recalculated percentages after removing categories for garbage, debris, and unknowns when using the classification of Pineda-Munoz and Alroy

(2014). Percentages were recalculated without removing garbage when using the

EltonTraits 1.0 (Wilman et al. 2014) classification which includes scavenge and garbage. Mid month values April/May and September/October were assigned to April and October respectively. Sample size per month was not explicitly stated but at most,

19 of the 68 months illustrated had less than 20 scat samples although unquantified extra collections probably make the number of insufficiently sampled months much smaller (Mattson et al. 1991).

We reanalyzed the > 30,000 year long dung record for the Shasta ground sloth

(Nothrotheriops shastensis) from Rampart Cave, USA described by (Hansen 1978), recalculating percent volume of food in dung after removing trace species without exact percentages. The average time elapsed between binned dung samples was 3,188 years with a range from 320 to 10,400 years. The oldest two dung samples were beyond the range of radiocarbon dating so they were arbitrarily assigned to 40,000 ybp and 40,500 ybp respectively for graphing purposes.

Silvia Pineda-Munoz - April 2016 193

Diet, ecology, and dental morphology in terrestrial mammals

To show how proxies recording different temporal scales can change the range of perceived dietary variability in a herbivore, we used data from (Cerling et al. 2009).

Cerling and colleagues (2009) used stable isotope analysis of tail hair serially sampled every 5 mm to produce a 6 year dietary record with near weekly resolution for one family group of African bush elephants (Loxodonta africana) in northern Kenya. Each individual tail hair covered a median of 18 months of diet so Cerling and colleagues

(2009) combined the isotopic signals recorded in 7 different tail hairs from 3 female elephants to produce the final 6 year dietary record of percentage C4 (tropical grasses) vs. C3 (trees, shrubs, and forbs) plants consumed. It is reasonable to combine hairs from the 3 individuals into one record because they were part of the same family group and

GPS tracking revealed that they spent >80% of their time in direct proximity with each other (Cerling et al. 2009). Two points in the original data with negative percentages of

C4 grass were set to 0 because negative percentages are impossible in this scenario and an artifact of the mixing model used (Cerling et al. 2009).

(References in main References section)

194 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Figure 11.3-1: Diet of chimpanzees (Pan troglodytes) observed in Kigali National Park, Uganda from April 7 to July 28, 2013 by David Watts (personal communication) projected into the PCA multidimensional dietary space of EltonTraits 1.0 (Wilman et al. 2014). Components 1 and 2 account for 20% and 13% of dietary variance, respectively. Shaded areas represent the 99% confidence ellipses around those species that consume 50% or more of a particular dietary category. See Wilman et al. (2014) for defini- tions of dietary categories used. The multicolored

Silvia Pineda-Munoz - April 2016 195

Diet, ecology, and dental morphology in terrestrial mammals

line shows where chimpanzees are in dietary space at sampled times. (a) Daily diet and (b) seven day average diet.

Figure 11.3-2: Diet of brown bears (Ursus arctos) in Yellowstone National Park, Wyoming from 1977-1987 obtained by opportunistic scat sampling (Mattson et al. 1991). (a) Monthly average diet and (b) yearly average diet. Figure follows conventions of Figure 11.3-1.

196 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Figure 11.3-3: Diet of Shasta ground sloth (Nothrotheriops shastensis) obtained from dung deposits in Rampart Cave, Arizona (Hansen 1978). Figure follows conventions of Figure 11.3- 1.

Silvia Pineda-Munoz - April 2016 197

Diet, ecology, and dental morphology in terrestrial mammals

Figure 11.3-4: Intraspecific temporal variation in diet can be large. Figure shows Euclidean distances between species’ untransformed percentage diets using the classification and data in EltonTraits 1.0 (Wilman et al. 2014) as the black boxplot. Colored boxplots represent the distribution of intraspecific dietary distances between all sampled points in each temporal dataset. Colored circles represent only those distances taken between consecutive timepoints like Day 1 to Day 2, June to July, etc. The dietary distance between elephants and other species in EltonTraits 1.0 (Wilman et al. 2014) is given for reference.

198 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Figure 11.3-5: Intraspecific temporal variation in diet can be large. Figure shows Euclidean distances between species’ untransformed percentage diets using the classification and data in Pineda-Munoz and Alroy (2014) as the black boxplot. Colored boxplots represent the distribution of intraspecific dietary distances between all sampled points in each temporal dataset. Colored circles represent only those distances taken between consecutive timepoints like Day 1 to Day 2, June to July, etc. The dietary distance between elephants and other species in Pineda-Munoz and Alroy (2014) is given for reference. Because EltonTraits 1.0 (Wilman et al. 2014) has a different classification scheme and uses quantitative estimates of diet based on

Silvia Pineda-Munoz - April 2016 199

Diet, ecology, and dental morphology in terrestrial mammals

qualitative data, the distances between elephants and other reference species in Figure 11.4-4 are different than those in this figure, which are based on Pineda-Munoz and Alroy (2014).

Figure 11.3-6: Diagrammatic examples of how mismatching temporal grain (a) and extent (b) between questions and proxies can lead to erroneous inferences in a paleoecological context. Temporal grain example (a) based on (Faith 2011). Temporal extent example (b) based on (Emslie & Patterson 2007).

Table 11.3-1: Feasibility and time averaging for popular diet proxies. *Nonlethal methods have been developed to retrieve stomach contents in certain species (Kronfeld and Dayan 1998). †Methods exist for detailed examination of dental micro- and mesowear in living individuals (Barnes and Longhurst 1960, Teaford and Oyen 1989).

Invasiveness Feasibility Method Proxy Time averaging for live for fossils animals Feeding observations/bite Observations Seconds to minutes Low None counts Observations Feeding site analysis Hours None Very low Contents Fistulation Hours High None Contents Stomach contents Hours to days Lethal* Low Contents Fecal contents Days None Low Isotopic Hair/baleen/feather isotopes Days to decades Low Low Isotopic Whole blood isotopes Months to years Medium None Isotopic Tooth enamel isotopes Years to decades Lethal High Isotopic Bone collagen isotopes Years to decades Lethal Medium Wear Microwear Weeks Medium† High

200 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Wear Mesowear Years to decades Medium† High Morphology Hypsodonty Millions of years Medium High Months to millions of Morphology Dental or jaw morphology Lethal Medium years

11.4 Supplementary information 4

11.4.1 Dietary classification We qualitatively classified the diets of all species in the dataset following the classification scheme proposed by Pineda-Munoz and Alroy (2014), which emphasizes the primary resource in a given diet. Thus, a species was classified as a specialist if a single food resource made up 50% or more of its diet. The specializations in our dataset were browsing (mainly feeding on leaves), grazing (mainly feeding on grasses), mixed feeding (a combination of both), carnivory, frugivory, granivory, insectivory, fungivory, gumivory and generalisation (i.e., no food resource makes up 50% or more of the diet).

Dietary data was extracted from primary literature (Nowak 1999, Wilman et al. 2014)

(Table 11.4-1 and 11.4-2).

11.4.2 MPDMA data Table 11.4-1: MPDMA analysis data for lower tooth rows of 138 mammalian specimens. A=Overall OPCR; B=average OPCR between individual teeth in a tooth row; C=standard deviation of the OPCR between individual teeth in a tooth row; D= relief index; E=average slope; F=standard deviation of the slope. Information on each variable can be found in the main text. Values were calculated using Surfer Manipulator (Evans 2011). Each specimen is listed with its museum collection ID.

Order Specimen A B C D E F Diet

Carnivora Acinonyx_jubatus-HELU31 17 17.00 0.00 7.11 59.91 17.35 Carnivore

Carnivora Ailuropoda_melanoleuca-USNM259400 258 90.33 41.04 4.13 45.07 17.34 Herbivore Carnivora Ailurus_fulgens-1.537 128 64.50 2.12 6.35 50.02 16.52 Herbivore

Silvia Pineda-Munoz - April 2016 201

Diet, ecology, and dental morphology in terrestrial mammals

Carnivora Alopex_lagopus-A583655 94 29.00 9.85 5.80 52.89 18.81 Carnivore

Carnivora Canis_aureus-ZMB52447 128 36.33 11.68 4.42 53.13 18.30 Carnivore

Carnivora Canis_lupus-A20025013 101 35.00 4.36 3.53 53.19 17.07 Carnivore

Carnivora Crocuta_crocuta-HEL30.196 21 21.00 0.00 8.10 59.89 13.99 Carnivore Carnivora Felis_silvestris-1.780 30 17.00 0.00 8.92 59.51 17.81 Carnivore

Carnivora Galerella_sanguinea-HEL678 82 42.50 0.71 3.98 48.90 19.44 Insectivore

Carnivora Genetta_genetta-A58042 96 52.00 4.24 4.00 48.04 18.91 Carnivore Carnivora Gulo_gulo-3619 38 20.50 3.54 2.78 51.73 14.82 Carnivore

Carnivora Herpestes_ichneumon-ZMB83028 71 33.50 0.71 3.12 51.64 17.97 Carnivore Carnivora Lutra_lutra-33.853 66 30.00 7.07 3.33 47.93 17.75 Carnivore Carnivora Lynx_lynx-31.929 21 21.00 0.00 7.66 54.16 16.71 Carnivore Carnivora Martes_foina-1.781 58 31.00 4.24 3.98 49.49 17.22 Carnivore Carnivora Martes_martes-31.542 60 27.00 7.07 5.84 47.14 17.67 Carnivore

Carnivora Meles_meles-A587441 88 41.50 4.95 4.57 43.64 19.12 Generalist

Carnivora Mustela_erminea-A580133 41 14.50 6.36 4.61 52.47 15.66 Carnivore

Carnivora Mustela_eversmannii-A581155 37 18.00 4.24 5.02 56.71 15.34 Carnivore

Carnivora Mustela_lutreola-ZMB94308 49 29.50 6.36 5.90 54.08 17.00 Carnivore

Carnivora Mustela_nivalis-A710101 47 24.50 3.54 4.41 52.40 17.51 Carnivore

Carnivora Mustela_putorius-A595407 45 23.50 2.12 4.94 55.61 16.17 Carnivore

Carnivora Otocyon_megalotis-ZMB65785 201 48.00 8.91 3.48 48.22 16.91 Insectivore

Carnivora Panthera_leo-ZMB55282 17 17.00 0.00 9.36 60.10 13.92 Carnivore

Carnivora Paradoxurus_hermaphroditus-U229 99 44.00 9.90 3.14 32.79 16.23 Generalist

Carnivora Procyon_lotor-HEL885 109 53.50 4.95 5.37 42.57 17.26 Granivore

Carnivora Ursus_americanus-HEL995 192 62.67 14.98 3.46 36.81 17.80 Generalist

Carnivora Ursus_arctos-A585456 183 58.67 9.02 3.84 39.03 16.27 Herbivore

Carnivora Ursus_maritimus-HELU201 201 67.33 21.03 2.88 39.27 17.25 Carnivore

Carnivora Viverra_zibetha-HEL390 115 50.50 3.54 4.76 49.23 19.53 Carnivore

Carnivora Vormela_peregusna-A915107 47 24.00 2.83 5.23 54.84 17.55 Carnivore

Carnivora Vulpes_vulpes-A584977 122 39.67 12.06 3.40 54.05 19.23 Carnivore

Diprotodontia Acrobates_pygmaeus-NMVC4005 196 40.20 13.37 3.96 55.36 19.98 Insectivore

Diprotodontia Bettongia_gaimardi_NMV C11375 214 40.60 18.06 2.71 40.41 17.72 Fungivore

Diprotodontia Dactylopsila_trivirgata-NMVC4375_av 174.5 45.00 12.34 4.56 38.77 17.49 Insectivore

Diprotodontia Dendrolagus_bennetianus_NMVC7116 202 40.40 15.79 3.78 46.45 18.53 Frugivore

Diprotodontia Dendrolagus_lumholtzi-NMVc7111 162 33.00 9.25 2.38 41.47 19.09 Frugivore

202 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Diprotodontia Hemibelideus_lemuroides_NMV_C3961 253 48.40 12.80 2.68 47.83 17.34 Herbivore

Hypsiprymnodon_moschatus- Diprotodontia NMVC6784_av 209.5 40.70 11.91 3.12 37.90 20.93 Insectivore

Diprotodontia Macropus_agilis-NMVC6437 246 43.20 20.71 7.44 46.63 18.86 Herbivore

Diprotodontia Macropus_antilopinus_NMVDTC134 205 34.20 16.57 2.74 45.99 18.00 Herbivore

Diprotodontia Macropus_dorsalis_NMV_C10729 218 37.00 21.15 3.10 44.41 19.69 Herbivore

Diprotodontia Macropus_fuliginosus_NMV_c6375 230 49.75 17.78 5.97 45.39 18.89 Herbivore

Diprotodontia Macropus_parma_NMV_C10719 228 37.80 16.45 2.60 47.84 19.92 Herbivore

Diprotodontia Macropus_parryi_MZ815 266 47.80 24.06 4.30 47.97 19.35 Herbivore

Diprotodontia Macropus_rufogriseus_MZ2854 268 49.60 22.52 3.09 45.05 19.51 Herbivore

Diprotodontia Onychogalea_fraenata_NMV_C6394 221 40.40 14.48 4.89 49.25 19.00 Herbivore

Diprotodontia Petauroides_volans-NMVC6815_av 214.5 39.40 11.82 1.47 41.55 16.48 Herbivore

Diprotodontia Petauroides_volans-NMVC6816_av 235 44.70 9.24 2.58 47.78 16.71 Herbivore

Diprotodontia Petaurus_australis-NMVC3728_av 171 31.20 11.86 2.19 37.62 17.22 Gumivore

Diprotodontia Petaurus_australis-NMVC3729_av 159 31.30 12.76 2.55 33.42 17.07 Gumivore

Diprotodontia Petaurus_australis-NMVC3730 159 31.20 15.47 1.97 34.02 17.07 Gumivore

Diprotodontia Petaurus_breviceps-NMVC22460_av 182 35.60 16.71 2.21 34.16 17.18 Gumivore

Diprotodontia Petaurus_breviceps-NMVC29890_av 158 29.50 13.21 1.74 36.21 17.29 Gumivore

Diprotodontia Petaurus_gracilis_NMV_35952 203 39.40 16.53 3.16 41.03 17.38 Gumivore

Diprotodontia Petaurus_norfolcenssis_NMV_C18896 173 36.00 12.94 2.24 37.47 19.51 Gumivore

Diprotodontia Petrogale_brachyotis_MNV_DTC200 222 40.40 19.45 2.81 47.35 19.67 Herbivore

Diprotodontia Petrogale_vemustula_MZ4415 199 34.40 20.31 3.61 40.80 20.38 Herbivore

Diprotodontia Petrogale_xanthopus_NMV_C6606 271 50.00 19.56 4.21 41.28 19.42 Herbivore

Diprotodontia Potorous_tridactylus_NMV_C33011 196 39.00 14.93 2.73 41.62 18.63 Herbivore

Diprotodontia Thylogale_sticmatica_MNV_C8283 220 42.80 16.24 1.94 38.45 18.44 Herbivore

Diprotodontia Trichosurus_vulpecula_NMV_c3234 216 40.60 12.24 3.91 44.03 17.79 Herbivore

Diprotodontia Vombatus_ursinus_NMV_C6652 205 33.80 9.42 3.12 38.24 19.03 Herbivore

Diprotodontia Wallabia_bicolor_MZ6416 205 38.40 17.95 3.26 38.08 19.71 Herbivore

Primates Alouatta_seniculus_MCZ30432_av 347 57.00 5.34 2.67 38.35 18.97 Herbivore

Primates Alouatta_seniculus_MCZ8296_av 356 54.00 7.91 3.38 42.70 18.94 Herbivore

Primates Callithrix_argentata_MZC30585_av 206 39.40 6.44 3.41 45.86 18.95 Frugivore

Primates Callithrix_argentata_MZC32166 235 43.40 9.91 3.48 46.99 18.49 Frugivore

Silvia Pineda-Munoz - April 2016 203

Diet, ecology, and dental morphology in terrestrial mammals

Primates Cebus_apella_MCZ30726 243 39.67 10.11 4.74 41.36 17.41 Generalist

Primates Cebus_apella_MCZ32049 297 47.67 10.63 4.98 42.00 17.81 Generalist

Primates Cercopithecus_mitis_MCZ27174_av 249.5 47.80 12.98 2.71 45.47 18.15 Frugivore

Primates Cercopithecus_mitis_MZC39390_av 236.5 47.40 10.45 2.64 46.88 17.75 Frugivore

Primates Cercopithecus_nictitans_MCZ14726 236 47.60 17.27 2.56 48.70 16.25 Frugivore

Primates Cercopithecus_nictitans_MCZ17691_av 257.5 49.20 13.07 2.88 46.19 18.21 Frugivore

Primates Chlorocebus_aethiops_MZC31976 249 49.00 17.87 3.35 47.27 19.16 Generalist

Primates Euoticus_elegantulus_MCZ14657 195 28.83 13.11 3.19 50.62 18.56 Gumivore

Primates Euoticus_elegantulus_MCZ17591 190 29.33 12.68 2.79 52.07 17.49 Gumivore

Primates Galago_alleni_MCZ17589 237 37.83 13.42 1.64 45.18 17.86 Gumivore

Primates Gorilla_gorilla_MCZ20039 353 65.60 26.82 3.61 43.17 17.37 Herbivore

Primates Miopithecus_talapoin_MCZ34264_av 256.5 49.20 15.52 4.05 45.38 18.37 Frugivore

Primates Perodictus_potto_MCZ31720 290 45.17 12.75 2.28 44.14 18.03 Frugivore

Primates Sanguinus_midas_MZC30589_av 217.5 42.40 6.40 3.39 44.89 17.82 Frugivore

Primates Sanguinus_midas_MZC30597_av 221 42.40 10.91 2.82 42.67 17.75 Frugivore

Rodentia Aethomys_hindei-A584096 225 64.00 3.61 5.19 33.16 16.53 Herbivore

Rodentia Akodon_serrensis-A581588 117 39.33 11.06 4.76 35.59 18.97 Insectivore

Rodentia Anisomys_imitator-ZMB39700 231 68.33 6.43 4.20 21.66 14.47 Gumivore

Rodentia Apodemus_agrarius-A925142 191 55.67 14.74 5.94 45.11 18.50 Herbivore

Rodentia Arvicanthis_niloticus-A584072 150 48.33 10.26 4.67 39.22 15.36 Herbivore

Rodentia Bandicota_indica-A591499 181 55.33 22.59 4.83 38.34 18.21 Herbivore

Rodentia Berylmys_bowersi-ZMB45364 194 61.67 14.50 4.41 31.80 18.44 Herbivore

Rodentia Bunomys_coelestis-ZMB46071 182 56.00 10.15 4.55 34.85 16.66 Insectivore Rodentia Calomys_sp-A586361 220 65.00 5.00 6.12 37.28 18.75 Granivore

Rodentia Calomys_sp-ZMB3706 225 72.33 12.66 4.82 41.38 18.67 Granivore

Rodentia Chiropodomys_sp-ZMB48613 245 80.33 12.74 4.14 29.53 14.03 Herbivore

Rodentia Crateromys_schadenbergi-ZMB10173 230 70.33 19.14 6.11 46.39 19.12 Frugivore

Rodentia Dasymys_sp-A594208 233 73.33 21.13 3.98 34.08 18.04 Herbivore

Rodentia Delomys_dorsalis-A596361 267 85.67 15.95 5.15 29.42 17.83 Granivore

Rodentia Dephomys_sp-ZMBEmin 183 55.33 7.64 4.32 35.45 16.74 Frugivore

Rodentia Geoxus_valdivianus-ZMB66514 185 53.67 7.51 3.18 31.50 15.83 Insectivore

Rodentia Grammomys_dolichurus-A584059 229 70.33 21.73 5.66 39.45 16.37 Herbivore

204 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Rodentia Grammomys_rutilans-ZMB66337 224 73.67 12.90 5.62 43.74 16.77 Herbivore

Rodentia Grammomys_rutilans-ZMB66339 220 79.00 14.73 5.86 32.78 16.77 Herbivore

Rodentia Holochilus_brasiliensis-A592476 286 86.67 11.85 4.73 33.18 18.98 Herbivore

Rodentia Hybomys_univittatus-A584094 221 68.33 24.79 6.97 45.31 17.90 Generalist

Rodentia Hydromys_chrysogaster-A581958 131 41.00 8.49 8.32 44.15 21.39 Insectivore

Rodentia Hylomyscus_stella-A594301 191 60.33 8.33 4.70 38.48 15.28 Frugivore

Rodentia Hyomys_goliath-A587220 276 67.67 19.66 6.67 38.27 18.54 Herbivore

Rodentia Ichthyomys_hydrobates-A584738 136 52.33 17.56 3.78 25.87 10.72 Insectivore

Rodentia Ichthyomys_stolzmanni-A586091 180 53.33 9.07 6.45 37.78 19.83 Insectivore

Rodentia Lemniscomys_striatus-A594081 167 54.33 12.50 5.46 40.52 14.86 Herbivore

Rodentia Leopoldamys_sabanus-ZMB13870 219 61.00 7.21 4.04 38.00 18.83 Insectivore

Rodentia Leptomys_elegans-ZMB31782 148 46.00 1.73 4.86 41.98 15.82 Insectivore

Rodentia Lophuromys_medicaudatus-ZMB83983 167 55.00 4.58 5.03 36.59 15.46 Insectivore

Rodentia Lorentzimys_nouhuysii-ZMB60078 192 59.00 8.72 5.16 30.31 13.40 Herbivore

Rodentia Malacomys_sp-ZMB33402 180 54.67 11.93 4.58 36.67 19.18 Insectivore

Rodentia Mallomys_rothschildi-A587224 241 73.00 16.52 8.03 46.76 19.36 Herbivore

Rodentia Mastomys_natalensis-A604221 184 64.00 7.00 4.53 39.79 16.28 Insectivore

Rodentia Melomys_levipes-ZMB92283 177 54.33 4.93 4.57 35.42 15.66 Frugivore

Rodentia Micromys_minutus-A584052 182 58.67 10.69 5.30 39.66 18.82 Generalist

Rodentia Mus_musculus-HEL13.241 168 60.33 2.08 4.05 38.19 16.39 Herbivore

Rodentia Myomys_derooi-ZMB84481 177 60.67 4.62 4.05 38.44 16.35 Insectivore

Rodentia Nectomys_squamipes-A582359 231 71.67 9.07 7.50 37.09 20.55 Generalist

Rodentia Nesokia_indica-A593571 204 51.00 18.36 4.51 36.20 18.12 Herbivore

Rodentia Niviventer_rapit-A591531 188 60.67 15.04 5.05 36.34 17.40 Generalist

Rodentia Notomys_mitchellii-A592841 196 70.67 3.51 5.30 29.35 18.85 Generalist

Rodentia Oenomys_hypoxanthus-A584066 210 67.33 10.21 6.94 42.27 19.07 Herbivore

Rodentia Otomys_denti-A584268 146 41.00 10.82 6.20 35.93 17.11 Herbivore

Rodentia Otomys_irroratus-A584309 128 39.67 12.66 5.98 37.04 17.38 Herbivore

Rodentia Oxymycterus_sp-A591560 165 55.67 7.51 5.19 41.76 20.45 Insectivore

Rodentia Parotomys_littledalei-ZMB60122 179 46.33 18.88 5.11 31.46 19.51 Herbivore

Rodentia Pelomys_campanae-ZMB70012 66 56.33 13.43 4.73 39.49 15.10 Herbivore

Rodentia Peromyscus_maniculatus-A875259 219 67.00 5.00 5.55 41.82 19.62 Insectivore

Silvia Pineda-Munoz - April 2016 205

Diet, ecology, and dental morphology in terrestrial mammals

Rodentia Phloeomys_sp-ZMB44925 233 76.00 13.45 6.02 37.65 23.91 Herbivore Rodentia Phyllotis_sp-A598232 235 67.33 7.02 6.06 41.40 24.50 Herbivore

Rodentia Pogonomys_sp-A700091 282 82.00 16.64 6.46 43.75 17.59 Herbivore

Rodentia Praomys_jacksoni-A584294 177 52.00 10.58 4.73 37.79 16.89 Generalist

Rodentia Rattus_leucopus-ZMB91918 238 75.00 10.15 4.48 34.32 17.93 Generalist

Rodentia Reithrodon_auritus-A584311 238 63.67 10.97 5.89 38.44 21.54 Herbivore Reithrodontomys_mexicanus- Rodentia ZMB92937 257 83.67 21.50 5.51 27.73 14.50 Granivore

Rodentia Rhabdomys_pumilio-A584571 193 58.33 14.36 6.08 41.90 15.90 Generalist

Rodentia Rhipidomys_sp-A583312 213 66.00 17.00 5.01 37.72 16.81 Generalist

Rodentia Sigmodon_hispidus-ZMB43184 267 77.00 8.00 4.28 27.07 17.85 Insectivore

Rodentia Stochomys_longicaudatus-ZMB83974 263 69.33 12.90 5.86 35.00 17.67 Frugivore

Rodentia Sundamys_muelleri-ZMB39517 203 60.67 15.04 5.25 40.44 19.22 Generalist

Rodentia Uromys_caudimaculatus-ZMB39872 184 56.00 13.08 5.25 38.04 17.99 Frugivore

Rodentia Zelotomys_cf-ZMB44338 183 55.67 1.15 4.65 40.35 16.41 Insectivore

Rodentia Zygodontomys_sp-ZMB81389 178 59.00 8.00 3.80 33.51 16.56 Granivore

Table 11.4-2: MPDMA analysis data for upper tooth rows of 138 mammalian specimens. A=Overall OPCR; B=average OPCR between individual teeth in a tooth row; C=standard deviation of the OPCR between individual teeth in a tooth row; D= relief index; E=average slope; F=standard deviation of the slope. Information on each variable can be found in the main text. Values were calculated using Surfer Manipulator (Evans 2011). Each specimen is listed with its museum collection ID

Order Specimen A B C D E F Diet

Carnivora Acinonyx_jubatus-HELU31 40.00 20.00 1.41 3.38 57.27 16.82 Carnivore

Carnivora Ailuropoda_melanoleuca-USNM259400 342.00 92.33 39.32 2.67 42.62 15.55 Herbivore

Carnivora Ailurus_fulgens-1.537 270.00 71.33 9.81 3.43 50.37 15.35 Herbivore

Carnivora Alopex_lagopus-A583655 179.00 44.00 16.64 2.57 49.67 18.23 Carnivore

Carnivora Canis_aureus-ZMB52447 187.00 47.33 16.86 2.76 50.25 17.50 Carnivore

Carnivora Canis_lupus-A20025013 140.00 37.67 11.15 3.20 50.89 17.02 Carnivore

Carnivora Crocuta_crocuta-HEL30.196 51.00 26.50 4.95 3.46 58.08 14.73 Carnivore

206 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Carnivora Felis_silvestris-1.780 48.00 19.50 6.36 3.62 54.93 18.93 Carnivore

Carnivora Galerella_sanguinea-HEL678 143.00 34.00 1.73 2.12 48.19 19.40 Insectivore

Carnivora Genetta_genetta-A58042 159.00 31.33 2.31 1.98 47.82 19.03 Carnivore Carnivora Gulo_gulo-3619 86.00 37.50 12.02 1.82 47.38 19.11 Carnivore

Carnivora Herpestes_ichneumon-ZMB83028 153.00 38.00 2.65 2.29 49.03 19.67 Carnivore

Carnivora Lutra_lutra-33.853 103.00 44.50 3.54 3.41 44.51 17.97 Carnivore

Carnivora Lynx_lynx-31.929 46.00 21.00 7.07 4.74 53.10 16.83 Carnivore

Carnivora Martes_foina-1.781 101.00 38.50 19.09 1.88 44.14 17.22 Carnivore

Carnivora Martes_martes-31.542 94.00 39.50 16.26 2.08 41.87 17.82 Carnivore

Carnivora Meles_meles-A587441 127.00 57.00 24.04 2.86 40.79 16.81 Generalist

Carnivora Mustela_erminea-A580133 61.00 25.00 5.66 1.58 46.48 18.29 Carnivore

Carnivora Mustela_eversmannii-A581155 68.00 25.00 11.31 1.81 49.88 18.19 Carnivore

Carnivora Mustela_lutreola-ZMB94308 99.00 36.50 10.61 2.09 46.70 18.02 Carnivore

Carnivora Mustela_nivalis-A710101 79.00 32.00 11.31 1.82 48.75 18.37 Carnivore

Carnivora Mustela_putorius-A595407 71.00 28.50 13.44 1.76 49.44 18.63 Carnivore

Carnivora Otocyon_megalotis-ZMB65785 216.00 46.75 12.50 2.89 46.22 16.92 Insectivore

Carnivora Panthera_leo-ZMB55282 42.00 16.00 5.66 2.33 55.92 14.96 Carnivore

Carnivora Paradoxurus_hermaphroditus-U229 115.00 30.33 15.01 2.17 37.13 17.44 Generalist

Carnivora Procyon_lotor-HEL885 188.00 52.67 10.07 3.77 47.42 15.15 Granivore

Carnivora Ursus_americanus-HEL995 150.00 40.33 8.96 3.49 39.74 16.05 Generalist

Carnivora Ursus_arctos-A585456 175.00 46.33 15.31 3.04 39.65 15.86 Herbivore

Carnivora Ursus_maritimus-HELU201 135.00 42.33 10.69 2.78 42.79 16.11 Carnivore

Carnivora Viverra_zibetha-HEL390 206.00 49.33 12.22 2.29 49.45 18.08 Carnivore

Carnivora Vormela_peregusna-A915107 87.00 38.00 11.31 2.36 46.13 20.32 Carnivore

Carnivora Vulpes_vulpes-A584977 177.00 50.33 20.82 2.67 48.37 18.57 Carnivore

Diprotodontia Acrobates_pygmaeus-NMVC4005 199.00 38.25 22.37 3.97 52.16 21.20 Insectivore

Diprotodontia Bettongia_gaimardi_NMV C11375 213.00 44.00 16.14 3.50 42.92 17.78 Fungivore

Diprotodontia Dactylopsila_trivirgata-NMVC4375_av 237.50 43.60 20.22 2.45 40.44 16.17 Insectivore

Diprotodontia Dendrolagus_bennetianus_NMVC7116 254.00 48.40 11.89 4.01 44.91 17.92 Frugivore

Silvia Pineda-Munoz - April 2016 207

Diet, ecology, and dental morphology in terrestrial mammals

Diprotodontia Dendrolagus_lumholtzi-NMVc7111 197.00 36.00 7.31 2.46 38.95 17.15 Frugivore

Diprotodontia Hemibelideus_lemuroides_NMV_C3961 392.00 71.60 16.67 3.40 46.72 16.71 Herbivore

Hypsiprymnodon_moschatus- Diprotodontia NMVC6784_av 215.00 43.20 10.02 2.53 39.27 21.91 Insectivore

Diprotodontia Macropus_agilis-NMVC6437 243.00 48.00 14.34 4.95 49.45 18.67 Herbivore

Diprotodontia Macropus_antilopinus_NMVDTC134 241.00 43.60 20.67 2.59 44.06 20.10 Herbivore

Diprotodontia Macropus_dorsalis_NMV_C10729 282.00 48.60 13.09 3.01 43.77 20.22 Herbivore

Diprotodontia Macropus_fuliginosus_NMV_c6375 193.00 44.75 14.31 3.28 46.44 18.48 Herbivore

Diprotodontia Macropus_parma_NMV_C10719 215.00 39.80 16.66 2.31 50.04 19.68 Herbivore

Diprotodontia Macropus_parryi_MZ815 264.00 50.00 19.85 3.04 48.02 17.89 Herbivore

Diprotodontia Macropus_rufogriseus_MZ2854 312.00 54.00 12.02 2.71 46.94 19.76 Herbivore

Diprotodontia Onychogalea_fraenata_NMV_C6394 276.00 48.40 14.24 4.38 49.79 18.58 Herbivore

Diprotodontia Petauroides_volans-NMVC6815_av 406.50 56.58 22.07 2.00 32.40 16.23 Herbivore

Diprotodontia Petauroides_volans-NMVC6816_av 337.50 47.17 21.32 2.29 43.33 16.40 Herbivore

Diprotodontia Petaurus_australis-NMVC3728_av 218.50 39.70 17.96 2.50 37.14 16.17 Gumivore

Diprotodontia Petaurus_australis-NMVC3729_av 178.00 32.60 16.36 1.90 35.42 14.34 Gumivore

Diprotodontia Petaurus_australis-NMVC3730 164.00 31.00 11.25 2.09 34.23 15.17 Gumivore

Diprotodontia Petaurus_breviceps-NMVC22460_av 171.50 33.50 18.95 2.05 37.79 15.63 Gumivore

Diprotodontia Petaurus_breviceps-NMVC29890_av 159.00 31.60 16.50 1.98 37.68 14.67 Gumivore

Diprotodontia Petaurus_gracilis_NMV_35952 190.00 37.00 15.81 3.06 43.93 15.95 Gumivore

Diprotodontia Petaurus_norfolcenssis_NMV_C18896 146.00 28.80 14.94 1.65 38.25 14.42 Gumivore

Diprotodontia Petrogale_brachyotis_MNV_DTC200 252.00 45.40 12.34 3.71 47.79 18.49 Herbivore

Diprotodontia Petrogale_vemustula_MZ4415 236.00 44.60 13.97 2.90 46.34 18.72 Herbivore

Diprotodontia Petrogale_xanthopus_NMV_C6606 292.00 51.80 13.77 4.02 45.30 19.35 Herbivore

Diprotodontia Potorous_tridactylus_NMV_C33011 192.00 37.80 15.16 1.98 42.29 17.97 Herbivore

208 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Diprotodontia Thylogale_sticmatica_MNV_C8283 250.00 45.60 9.91 1.70 42.47 17.29 Herbivore

Diprotodontia Trichosurus_vulpecula_NMV_c3234 218.00 42.80 12.89 2.89 45.23 16.40 Herbivore

Diprotodontia Vombatus_ursinus_NMV_C6652 230.00 38.80 8.70 2.66 33.92 20.68 Herbivore

Diprotodontia Wallabia_bicolor_MZ6416 275.00 49.00 10.20 2.62 34.77 18.95 Herbivore

Primates Alouatta_seniculus_MCZ30432_av 418.00 60.92 12.99 3.34 41.91 19.58 Herbivore

Primates Alouatta_seniculus_MCZ8296_av 425.00 55.58 13.96 3.54 43.72 19.32 Herbivore

Primates Callithrix_argentata_MZC30585_av 212.00 35.60 6.48 2.56 48.21 17.48 Frugivore

Primates Callithrix_argentata_MZC32166 255.00 33.00 8.86 3.02 48.97 16.92 Frugivore

Primates Cebus_apella_MCZ30726 263.00 37.83 8.11 4.05 43.84 18.40 Generalist

Primates Cebus_apella_MCZ32049 270.00 38.00 10.08 2.92 44.10 16.70 Generalist

Primates Cercopithecus_mitis_MCZ27174_av 262.50 49.90 8.81 2.90 46.84 17.14 Frugivore

Primates Cercopithecus_mitis_MZC39390_av 247.50 47.20 10.39 2.83 46.57 17.96 Frugivore

Primates Cercopithecus_nictitans_MCZ14726 237.00 47.20 10.28 3.14 50.09 16.86 Frugivore

Primates Cercopithecus_nictitans_MCZ17691_av 266.00 48.60 8.31 2.50 47.04 17.18 Frugivore

Primates Chlorocebus_aethiops_MZC31976 247.00 49.00 14.75 2.87 48.60 18.25 Generalist

Primates Euoticus_elegantulus_MCZ14657 254.00 32.00 16.24 2.75 49.08 15.82 Gumivore

Primates Euoticus_elegantulus_MCZ17591 280.00 33.83 13.39 2.69 48.63 17.32 Gumivore

Primates Galago_alleni_MCZ17589 316.00 41.67 13.88 2.46 45.31 15.84 Gumivore

Primates Gorilla_gorilla_MCZ20039 291.00 51.20 12.87 4.68 46.39 17.43 Herbivore

Primates Miopithecus_talapoin_MCZ34264_av 239.00 46.00 13.73 3.37 47.59 17.99 Frugivore

Primates Perodictus_potto_MCZ31720 308.00 39.33 14.36 2.51 48.11 17.36 Frugivore

Primates Sanguinus_midas_MZC30589_av 262.00 38.60 6.67 3.09 46.00 17.42 Frugivore

Primates Sanguinus_midas_MZC30597_av 248.00 37.50 5.60 2.65 46.43 16.92 Frugivore

Rodentia Aethomys_hindei-A584096 230.00 72.00 14.00 3.97 31.46 16.26 Herbivore

Rodentia Akodon_serrensis-A581588 144.00 42.67 13.01 4.06 39.08 18.79 Insectivore

Rodentia Anisomys_imitator-ZMB39700 281.00 86.00 17.06 3.91 21.41 14.34 Gumivore

Rodentia Apodemus_agrarius-A925142 189.00 61.00 5.00 4.14 44.06 18.63 Herbivore

Silvia Pineda-Munoz - April 2016 209

Diet, ecology, and dental morphology in terrestrial mammals

Rodentia Arvicanthis_niloticus-A584072 183.00 55.33 8.14 3.80 38.58 14.92 Herbivore

Rodentia Bandicota_indica-A591499 239.00 72.00 15.59 5.03 36.24 19.83 Herbivore

Rodentia Berylmys_bowersi-ZMB45364 209.00 60.33 6.66 4.09 29.73 18.23 Herbivore

Rodentia Bunomys_coelestis-ZMB46071 221.00 65.67 7.77 3.42 33.62 17.85 Insectivore

Rodentia Calomys_sp-A586361 186.00 58.33 2.52 6.00 42.14 16.78 Granivore

Rodentia Calomys_sp-ZMB3706 282.00 95.67 14.50 3.74 39.45 21.55 Granivore

Rodentia Chiropodomys_sp-ZMB48613 263.00 83.00 9.85 3.52 30.36 15.95 Herbivore

Rodentia Crateromys_schadenbergi-ZMB10173 285.00 79.67 5.51 6.26 42.41 21.42 Frugivore

Rodentia Dasymys_sp-A594208 309.00 82.33 8.33 4.16 31.88 18.37 Herbivore

Rodentia Delomys_dorsalis-A596361 271.00 89.00 14.73 5.03 31.72 19.22 Granivore

Rodentia Dephomys_sp-ZMBEmin 177.00 60.33 9.07 3.94 33.78 18.07 Frugivore

Rodentia Geoxus_valdivianus-ZMB66514 192.00 63.33 11.93 4.54 32.84 17.69 Insectivore

Rodentia Grammomys_dolichurus-A584059 266.00 81.00 8.89 4.34 41.37 16.03 Herbivore

Rodentia Grammomys_rutilans-ZMB66337 246.00 77.00 5.20 4.51 44.18 17.35 Herbivore

Rodentia Grammomys_rutilans-ZMB66339 274.00 83.00 7.94 5.07 31.86 17.98 Herbivore

Rodentia Holochilus_brasiliensis-A592476 287.00 82.33 15.04 5.36 34.69 19.66 Herbivore

Rodentia Hybomys_univittatus-A584094 231.00 67.00 6.56 3.46 42.40 19.51 Generalist

Rodentia Hydromys_chrysogaster-A581958 140.00 46.00 9.90 7.55 43.56 21.20 Insectivore

Rodentia Hylomyscus_stella-A594301 177.00 51.00 5.57 3.28 41.17 16.06 Frugivore

Rodentia Hyomys_goliath-A587220 310.00 79.67 10.60 5.04 40.04 19.79 Herbivore

Rodentia Ichthyomys_hydrobates-A584738 142.00 51.67 6.43 2.95 28.93 13.23 Insectivore

Rodentia Ichthyomys_stolzmanni-A586091 194.00 55.33 17.56 5.68 38.98 20.12 Insectivore

Rodentia Lemniscomys_striatus-A594081 218.00 67.67 12.58 4.53 39.51 15.68 Herbivore

Rodentia Leopoldamys_sabanus-ZMB13870 229.00 66.00 16.52 5.23 37.02 19.21 Insectivore

Rodentia Leptomys_elegans-ZMB31782 134.00 54.00 16.46 4.04 42.33 16.79 Insectivore

Rodentia Lophuromys_medicaudatus-ZMB83983 258.00 81.33 5.69 3.50 38.49 18.22 Insectivore

Rodentia Lorentzimys_nouhuysii-ZMB60078 216.00 78.67 21.03 4.95 32.26 14.88 Herbivore

Rodentia Malacomys_sp-ZMB33402 207.00 67.67 11.37 4.16 32.28 18.89 Insectivore

Rodentia Mallomys_rothschildi-A587224 260.00 70.33 2.89 3.09 39.37 18.64 Herbivore

210 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Rodentia Mastomys_natalensis-A604221 202.00 62.33 7.37 5.48 42.02 17.16 Insectivore

Rodentia Melomys_levipes-ZMB92283 189.00 52.00 8.19 4.37 33.41 15.37 Frugivore

Rodentia Micromys_minutus-A584052 195.00 60.67 7.51 3.76 41.45 17.76 Generalist

Rodentia Mus_musculus-HEL13.241 181.00 64.33 8.62 3.59 36.65 16.78 Herbivore

Rodentia Myomys_derooi-ZMB84481 229.00 70.00 12.49 4.76 36.80 15.75 Insectivore

Rodentia Nectomys_squamipes-A582359 282.00 95.00 24.02 6.29 36.74 22.82 Generalist

Rodentia Nesokia_indica-A593571 234.00 74.00 21.07 5.05 37.05 19.18 Herbivore

Rodentia Niviventer_rapit-A591531 210.00 62.33 6.81 3.55 34.32 17.38 Generalist

Rodentia Notomys_mitchellii-A592841 234.00 73.67 15.37 4.63 31.18 19.13 Generalist

Rodentia Oenomys_hypoxanthus-A584066 206.00 63.00 7.81 3.95 41.32 17.67 Herbivore

Rodentia Otomys_denti-A584268 198.00 58.67 14.22 4.64 34.95 17.85 Herbivore

Rodentia Otomys_irroratus-A584309 175.00 55.00 15.00 6.27 33.68 18.46 Herbivore

Rodentia Oxymycterus_sp-A591560 152.00 46.00 11.36 5.04 43.90 19.72 Insectivore

Rodentia Parotomys_littledalei-ZMB60122 226.00 68.00 15.10 5.41 29.84 19.82 Herbivore

Rodentia Pelomys_campanae-ZMB70012 197.00 62.33 7.02 3.84 39.19 15.76 Herbivore

Rodentia Peromyscus_maniculatus-A875259 234.00 75.33 3.06 6.37 42.10 18.57 Insectivore

Rodentia Phloeomys_sp-ZMB44925 193.00 60.67 10.26 8.15 37.79 25.45 Herbivore

Rodentia Phyllotis_sp-A598232 212.00 59.67 3.21 8.31 44.52 25.23 Herbivore

Rodentia Pogonomys_sp-A700091 302.00 90.00 6.08 5.03 45.22 17.98 Herbivore

Rodentia Praomys_jacksoni-A584294 162.00 53.00 7.00 4.11 39.40 15.67 Generalist

Rodentia Rattus_leucopus-ZMB91918 236.00 64.00 5.20 3.85 32.21 17.37 Generalist

Rodentia Reithrodon_auritus-A584311 276.00 88.67 14.19 3.96 36.48 20.16 Herbivore

Reithrodontomys_mexicanus- Rodentia ZMB92937 279.00 76.67 1.53 5.49 27.52 15.34 Granivore

Rodentia Rhabdomys_pumilio-A584571 190.00 62.00 4.36 3.79 40.26 17.94 Generalist

Rodentia Rhipidomys_sp-A583312 209.00 68.33 10.26 4.80 39.49 16.59 Generalist

Rodentia Sigmodon_hispidus-ZMB43184 276.00 91.33 9.45 3.98 30.30 20.03 Insectivore

Rodentia Stochomys_longicaudatus-ZMB83974 266.00 81.67 7.09 4.13 33.83 18.56 Frugivore

Rodentia Sundamys_muelleri-ZMB39517 244.00 68.00 9.85 6.95 40.64 19.69 Generalist

Rodentia Uromys_caudimaculatus-ZMB39872 196.00 57.33 15.95 4.50 34.49 18.42 Frugivore

Silvia Pineda-Munoz - April 2016 211

Diet, ecology, and dental morphology in terrestrial mammals

Rodentia Zelotomys_cf-ZMB44338 239.00 64.67 8.39 5.22 40.38 18.01 Insectivore

Rodentia Zygodontomys_sp-ZMB81389 201.00 73.67 13.20 3.85 36.23 20.57 Granivore

11.4.3 Figures

Figure 11.4-1: Placentals and marsupials overlap in dental ecomorphospace. Plot of the two first components of the principal component analysis of MPDMA data for 138 placental and marsupial specimens including rodents. Points are plotted according to primary diet (a); primary diet and phylogeny (b); and phylogeny at the infraclass level (placentals vs. marsupials) (c). Colours and polygons are as shown in legends. Data were square-root transformed before PCA.

212 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Figure 11.4-2: Dietary ecomorphospace of the order Rodentia. Plot of the two first components of the principal component analysis of MPDMA data for 64 rodent specimens. Points are plotted according to primary diet. Colours of the polygons are as shown in legends. Data were square-root transformed before PCA.

Silvia Pineda-Munoz - April 2016 213

Diet, ecology, and dental morphology in terrestrial mammals

Figure 11.4-3: Dietary ecomorphospace of the order Carnivora. Plot of the two first components of the principal component analysis of MPDMA data for 32 carnivoran specimens. Points are plotted according to primary diet. Colours of the polygons are as shown in legends. Data were square-root transformed before PCA.

214 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Figure 11.4-4: Dietary ecomorphospace of the order Diprotodontia. Plot of the two first components of the principal component analysis of MPDMA data for 32 diprotodontid specimens. Points are plotted according to primary diet. Colours of the polygons are as shown in legends. Data were square-root transformed before PCA.

Silvia Pineda-Munoz - April 2016 215

Diet, ecology, and dental morphology in terrestrial mammals

Figure 11.4-5: Dietary ecomorphospaces of the order Primates. Plot of the two first components of the principal component analysis of MPDMA data for 19 primate specimens. Points are plotted according to primary diet. Colours of the polygons are as shown in legends. Data were square-root transformed before PCA.

216 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

11.4.4 PCA stats Table 11.4-3: Principal component analysis loadings and importance of the components based on MPDMA data for 138 mammalian dentitions belonging to the orders Rodentia, Carnivora, Diprotodontia and Primates. Data were square-root transformed before PCA.

Loadings: Comp.1 Comp.2 Comp.3 Comp.4 Comp.5 Comp.6 Comp.7 Comp.8 Comp.9 Comp.10 Comp.11 Comp.12 Total OPC -0.384 -0.124 -0.238 -0.187 -0.246 0.244 -0.108 0.211 0.561 0.465 -0.192 Average OPC -0.409 0.144 -0.186 0.135 -0.199 -0.284 -0.594 0.28 -0.433 -0.106 Lower SD OPC -0.251 -0.224 -0.299 -0.313 0.138 -0.768 -0.193 0.152 -0.148 tooth row Relief index 0.52 0.236 -0.378 0.366 -0.135 -0.138 -0.313 0.478 0.161 Mean slope 0.357 0.125 -0.286 -0.404 0.199 -0.143 0.139 -0.117 -0.126 -0.706 SD slope 0.334 -0.518 0.277 -0.612 -0.356 0.104 Total OPC -0.365 -0.163 -0.243 -0.229 -0.168 0.359 0.103 0.469 -0.319 -0.484 Average OPC -0.402 0.157 -0.117 0.175 0.327 -0.377 0.148 -0.133 -0.36 0.568 0.148 Upper SD OPC -0.283 -0.29 0.173 0.834 0.222 0.149 0.164 tooth row Relief index -0.269 0.444 0.101 0.119 0.701 0.358 -0.189 -0.17 Mean slope 0.324 -0.336 -0.47 0.187 0.136 -0.26 0.121 0.639 SD slope 0.461 -0.37 0.36 -0.379 0.477 0.339 -0.122

Importance of component s: Comp.1 Comp.2 Comp.3 Comp.4 Comp.5 Comp.6 Comp.7 Comp.8 Comp.9 Comp.10 Comp.11 Comp.12 Standard deviation2.1445901 1.4204871 1.4096024 0.96694921 0.91318716 0.70509213 0.6026539 0.49748605 0.4549933 0.37194815 0.21218251 0.20466548 Proportion of Variance0.3860698 0.169376 0.1667902 0.07848463 0.06999981 0.04173198 0.0304869 0.02077491 0.0173775 0.01161294 0.00377917 0.00351614 Cumulative Proportion0.3860698 0.5554458 0.722236 0.80072065 0.87072046 0.91245244 0.9429393 0.96371425 0.9810918 0.99270469 0.99648386 1

Table 11.4-4: Principal component analysis loadings and importance of the components based on MPDMA data for 74 mammalian dentitions belonging to the orders Carnivora, Diprotodontia and Primates. Data were square-root transformed before PCA.

Loadings: Comp.1 Comp.2 Comp.3 Comp.4 Comp.5 Comp.6 Comp.7 Comp.8 Comp.9 Comp.10 Comp.11 Comp.12 Total OPC -0.388 -0.132 -0.181 -0.271 0.239 0.283 0.317 0.108 0.683 Average OPC -0.341 -0.183 0.196 -0.122 -0.464 -0.199 -0.47 -0.161 0.432 0.229 -0.218 Lower SD OPC -0.353 -0.114 0.174 0.505 -0.571 -0.467 -0.166 tooth row Relief index 0.278 -0.392 0.557 0.662 Mean slope 0.316 -0.355 0.372 -0.307 -0.133 0.708 SD slope -0.234 -0.557 0.111 0.18 -0.724 0.208 Total OPC -0.375 -0.162 -0.369 0.107 0.378 0.461 -0.56 Average OPC -0.333 -0.23 0.141 0.219 -0.554 0.155 0.13 -0.485 -0.231 0.349 Upper SD OPC -0.233 0.274 0.684 0.223 0.264 0.292 -0.137 0.412 tooth row Relief index -0.564 -0.421 0.291 0.253 0.237 0.271 -0.443 Mean slope 0.247 -0.508 0.13 -0.435 -0.243 -0.142 -0.109 0.157 -0.595 SD slope -0.723 0.241 0.116 -0.143 0.519 -0.256 0.187

Importance of component s: Comp.1 Comp.2 Comp.3 Comp.4 Comp.5 Comp.6 Comp.7 Comp.8 Comp.9 Comp.10 Comp.11 Comp.12 Standard deviation2.2826851 1.3504052 1.2227343 0.988967 0.80805038 0.74243267 0.59762347 0.557989 0.45434862 0.40815569 0.2284346 0.18053918 Proportion of Variance0.4401692 0.1540479 0.1262966 0.08262115 0.05515749 0.04656309 0.03017053 0.0263014 0.01743838 0.01407276 0.0044081 0.00275341 Cumulative Proportion0.4401692 0.5942171 0.7205137 0.80313485 0.85829234 0.90485542 0.93502595 0.9613274 0.97876573 0.99283849 0.9972466 1

Silvia Pineda-Munoz - April 2016 217

Diet, ecology, and dental morphology in terrestrial mammals

Table 11.4-5: Principal component analysis loadings and importance of the components based on MPDMA data for 64 mammalian dentitions belonging to the order Rodentia. Data were square-root transformed before PCA.

Loadings: Comp.1 Comp.2 Comp.3 Comp.4 Comp.5 Comp.6 Comp.7 Comp.8 Comp.9 Comp.10 Comp.11 Comp.12 Total OPC -0.456 0.121 0.208 0.512 0.464 -0.449 0.137 0.135 Average OPC -0.422 0.172 0.171 0.293 -0.1 0.347 -0.577 0.443 Lower SD OPC -0.202 0.264 -0.784 -0.388 0.228 0.191 0.152 tooth row Relief index -0.204 -0.338 0.13 -0.276 -0.707 -0.465 -0.123 Mean slope -0.44 0.34 0.36 0.157 0.102 0.145 0.293 0.568 0.308 SD slope -0.243 -0.384 -0.277 -0.156 0.272 -0.334 0.29 0.441 -0.441 0.16 Total OPC -0.444 0.204 0.195 0.118 -0.4 0.367 0.202 -0.599 Average OPC -0.406 0.246 0.385 0.109 -0.375 -0.194 -0.147 -0.275 0.581 Upper SD OPC -0.594 -0.377 0.513 0.445 0.131 tooth row Relief index -0.176 -0.306 -0.346 0.107 -0.318 -0.396 0.677 0.107 Mean slope -0.439 0.308 0.107 0.347 0.116 0.317 0.211 -0.117 -0.251 -0.504 -0.302 SD slope -0.279 -0.324 -0.339 0.415 -0.166 -0.197 -0.373 -0.413 0.328 -0.189

Importance of component s: Comp.1 Comp.2 Comp.3 Comp.4 Comp.5 Comp.6 Comp.7 Comp.8 Comp.9 Comp.10 Comp.11 Comp.12 Standard deviation1.9062175 1.816854 1.2751423 0.95173049 0.87397931 0.72222473 0.6507447 0.5178094 0.38956599 0.32249271 0.25858739 0.21546153 Proportion of Variance0.3076119 0.2794462 0.1376498 0.07668071 0.06466369 0.04415734 0.0358492 0.02269855 0.01284755 0.00880436 0.00566074 0.00393005 Cumulative Proportion0.3076119 0.5870581 0.7247078 0.80138854 0.86605223 0.91020956 0.9460588 0.96875731 0.98160485 0.99040922 0.99606995 1

Table 11.4-6: Principal component analysis loadings and importance of the components based on MPDMA data for 32 mammalian dentitions belonging to the order Carnivora. Data were square-root transformed before PCA.

Loadings: Comp.1 Comp.2 Comp.3 Comp.4 Comp.5 Comp.6 Comp.7 Comp.8 Comp.9 Comp.10 Comp.11 Comp.12 Total OPC -0.378 -0.215 0.312 0.128 0.544 0.441 0.223 0.369 Average OPC -0.374 -0.138 0.149 -0.244 0.283 0.131 -0.546 0.543 -0.253 Lower SD OPC -0.336 -0.295 0.115 0.681 0.192 0.151 -0.438 -0.244 tooth row Relief index 0.252 -0.425 -0.125 -0.141 0.292 -0.215 0.587 0.453 0.108 0.12 0.115 Mean slope 0.31 -0.114 -0.368 -0.406 -0.159 0.188 -0.146 0.333 0.422 -0.464 SD slope -0.154 0.258 -0.64 0.302 0.342 0.323 -0.388 -0.157 0.103 Total OPC -0.35 -0.105 -0.309 -0.361 -0.237 0.156 0.227 -0.587 -0.393 Average OPC -0.351 -0.116 -0.239 -0.192 -0.381 -0.336 0.209 -0.459 0.191 0.193 0.425 Upper SD OPC -0.278 -0.559 0.521 -0.103 -0.165 -0.234 0.393 -0.3 tooth row Relief index -0.507 -0.174 0.483 0.353 0.251 -0.485 0.179 -0.113 Mean slope 0.293 -0.294 -0.376 -0.141 -0.353 0.194 -0.272 -0.446 -0.177 0.447 SD slope 0.11 0.58 -0.328 0.207 -0.139 0.606 0.225 -0.198 0.142

Importance of component s: Comp.1 Comp.2 Comp.3 Comp.4 Comp.5 Comp.6 Comp.7 Comp.8 Comp.9 Comp.10 Comp.11 Comp.12 Standard deviation2.4168466 1.4286142 1.0825692 1.03063075 0.7265877 0.5834783 0.53162169 0.39025854 0.31245773 0.23362408 0.2043606 0.10716763 Proportion of Variance0.5024643 0.1755646 0.1008134 0.09137202 0.04541331 0.02928576 0.02431154 0.01310122 0.00839827 0.00469507 0.00359254 0.00098795 Cumulative Proportion0.5024643 0.6780289 0.7788423 0.87021435 0.91562766 0.94491342 0.96922495 0.98232618 0.99072444 0.99541951 0.99901205 1

218 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Table 11.4-7: Principal component analysis loadings and importance of the components based on MPDMA data for 28 mammalian dentitions belonging to the order Diprotodontia. Data were square-root transformed before PCA.

Loadings: Comp.1 Comp.2 Comp.3 Comp.4 Comp.5 Comp.6 Comp.7 Comp.8 Comp.9 Comp.10 Comp.11 Comp.12 Total OPC -0.393 -0.223 -0.212 0.177 0.466 0.582 0.245 -0.3 Average OPC -0.34 -0.173 -0.141 0.207 0.607 -0.226 -0.354 -0.434 0.212 Lower SD OPC -0.242 0.174 0.358 -0.678 -0.377 0.258 0.106 -0.288 0.108 tooth row Relief index -0.266 0.166 0.271 0.488 0.385 0.284 0.207 -0.201 0.522 Mean slope -0.329 0.355 -0.377 -0.134 -0.339 0.226 -0.362 0.478 -0.26 SD slope -0.119 0.622 -0.199 -0.145 0.214 -0.122 -0.577 -0.268 0.237 0.11 Total OPC -0.298 -0.244 -0.43 -0.246 0.151 -0.171 0.122 -0.252 0.136 0.154 0.257 0.606 Average OPC -0.34 -0.202 -0.396 0.148 -0.108 0.17 -0.304 -0.153 -0.419 -0.573 Upper SD OPC -0.406 0.459 -0.242 -0.224 0.256 0.593 -0.159 -0.2 0.155 tooth row Relief index -0.327 0.216 0.256 0.408 -0.421 0.128 -0.621 Mean slope -0.352 0.36 -0.122 -0.194 -0.143 -0.376 0.103 0.235 -0.624 0.266 SD slope -0.197 0.489 -0.449 0.117 0.401 0.45 -0.256 -0.208 0.136

Importance of component s: Comp.1 Comp.2 Comp.3 Comp.4 Comp.5 Comp.6 Comp.7 Comp.8 Comp.9 Comp.10 Comp.11 Comp.12 Standard deviation2.1755712 1.3561803 1.2643564 0.97028318 0.91316461 0.77809715 0.66557327 0.49328256 0.3639886 0.30207504 0.27474138 0.18495139 Proportion of Variance0.4090342 0.1589454 0.1381504 0.08135983 0.07206281 0.05232156 0.03828289 0.02102832 0.01144955 0.00788575 0.00652321 0.00295616 Cumulative Proportion0.4090342 0.5679796 0.7061299 0.78748975 0.85955256 0.91187412 0.95015701 0.97118533 0.98263489 0.99052063 0.99704384 1

Table 11.4-8: Principal component analysis loadings and importance of the components based on MPDMA data for 16 mammalian dentitions belonging to the order Primates. Data were square-root transformed before PCA.

Loadings: Comp.1 Comp.2 Comp.3 Comp.4 Comp.5 Comp.6 Comp.7 Comp.8 Comp.9 Comp.10 Comp.11 Comp.12 Total OPC -0.402 -0.226 -0.188 0.379 0.497 -0.366 0.457 Average OPC -0.393 -0.176 0.392 -0.116 0.313 0.618 -0.394 Lower SD OPC -0.616 -0.263 -0.119 -0.645 -0.225 -0.159 -0.159 tooth row Relief index -0.103 0.177 -0.588 0.276 -0.381 -0.1 0.162 0.146 0.347 -0.285 -0.364 Mean slope 0.379 -0.224 0.463 -0.365 -0.376 0.498 0.106 0.207 SD slope 0.231 0.234 0.744 0.125 -0.328 -0.16 -0.411 Total OPC -0.268 0.487 -0.316 0.39 0.167 -0.346 -0.181 -0.493 Average OPC -0.37 -0.191 0.104 0.318 0.225 -0.593 0.28 0.263 -0.192 -0.354 Upper SD OPC -0.473 0.188 0.474 -0.406 0.15 0.361 0.266 0.297 0.147 tooth row Relief index -0.337 -0.401 -0.113 -0.229 -0.522 0.353 -0.375 0.204 0.276 Mean slope 0.323 -0.186 -0.216 0.183 0.486 0.163 0.46 -0.129 0.373 -0.191 -0.247 -0.229 SD slope -0.298 0.322 -0.159 0.275 0.141 0.619 0.267 0.164 -0.375 0.124 0.202

Importance of component s: Comp.1 Comp.2 Comp.3 Comp.4 Comp.5 Comp.6 Comp.7 Comp.8 Comp.9 Comp.10 Comp.11 Comp.12 Standard deviation2.1678581 1.4034672 1.3283737 1.0567891 0.87038703 0.59240119 0.48099071 0.39237451 0.2315878 0.17290782 0.08993197 0.08190187 Proportion of Variance0.4217597 0.1767698 0.1583595 0.1002259 0.06798737 0.03149454 0.02076237 0.01381672 0.00481321 0.00268307 0.00072582 0.00060199 Cumulative Proportion0.4217597 0.5985295 0.756889 0.8571149 0.92510227 0.95659681 0.97735918 0.9911759 0.99598911 0.99867218 0.99939801 1

Table 11.4-9: Specimens for which discriminant analysis inferred dietary classifications that mismatched observed diets extracted from the literature (Nowak 1999, Wilman et al. 2014) for (a) the whole dataset in Tables 1 and 2 and for (b) the whole dataset excluding rodents.

a) Discriminant Specimen Order Diet diet Galerella_sanguinea-HEL678 Carnivora Insectivore Carnivore Meles_meles-A587441 Carnivora Generalist Carnivore Otocyon_megalotis-ZMB65785 Carnivora Insectivore Frugivore Paradoxurus_hermaphroditus-U229 Carnivora Generalist Insectivore Procyon_lotor-HEL885 Carnivora Granivore Insectivore Ursus_arctos-A585456 Carnivora Herbivore Generalist

Silvia Pineda-Munoz - April 2016 219

Diet, ecology, and dental morphology in terrestrial mammals

Ursus_maritimus-HELU201 Carnivora Carnivore Generalist Acrobates_pygmaeus-NMVC4005 Diprotodontia Insectivore Carnivore Bettongia_gaimardi_NMV C11375 Diprotodontia Fungivore Herbivore Dactylopsila_trivirgata-NMVC4375_L Diprotodontia Insectivore Herbivore Dendrolagus_bennetianus_NMVC7116 Diprotodontia Frugivore Herbivore Dendrolagus_lumholtzi-NMVc7111 Diprotodontia Frugivore Gumivore Hypsiprymnodon_moschatus-NMVC6784_L Diprotodontia Insectivore Herbivore Petauroides_volans-NMVC6816_L Diprotodontia Herbivore Frugivore Petaurus_gracilis_NMV_35952 Diprotodontia Gumivore Herbivore Vombatus_ursinus_NMV_C6652 Diprotodontia Herbivore Frugivore Cebus_apella_MCZ30726R Primates Generalist Herbivore Chlorocebus_aethiops_MZC31976L Primates Generalist Herbivore Euoticus_elegantulus_MCZ17591L Primates Gumivore Frugivore Gorilla_gorilla_MCZ20039R Primates Herbivore Frugivore Miopithecus_talapoin_MCZ34264L Primates Frugivore Herbivore Miopithecus_talapoin_MCZ34264R Primates Frugivore Herbivore Bunomys_coelestis-ZMB46071 Rodentia Insectivore Herbivore Calomys_sp-A586361 Rodentia Granivore Insectivore Calomys_sp-ZMB3706 Rodentia Granivore Herbivore Crateromys_schadenbergi-ZMB10173 Rodentia Frugivore Herbivore Dephomys_sp-ZMBEmin Rodentia Frugivore Insectivore Hybomys_univittatus-A584094 Rodentia Generalist Herbivore Hylomyscus_stella-A594301 Rodentia Frugivore Generalist Ichthyomys_hydrobates-A584738 Rodentia Insectivore Generalist Lophuromys_medicaudatus-ZMB83983 Rodentia Insectivore Granivore Lorentzimys_nouhuysii-ZMB60078 Rodentia Herbivore Insectivore Malacomys_sp-ZMB33402 Rodentia Insectivore Herbivore Micromys_minutus-A584052 Rodentia Generalist Herbivore Mus_musculus-HEL13.241 Rodentia Herbivore Insectivore Nectomys_squamipes-A582359 Rodentia Generalist Herbivore Niviventer_rapit-A591531 Rodentia Generalist Herbivore Notomys_mitchellii-A592841 Rodentia Generalist Insectivore Oenomys_hypoxanthus-A584066 Rodentia Herbivore Generalist Phloeomys_sp-ZMB44925 Rodentia Herbivore Insectivore Phyllotis_sp-A598232 Rodentia Herbivore Insectivore Praomys_jacksoni-A584294 Rodentia Generalist Herbivore Rattus_leucopus-ZMB91918 Rodentia Generalist Frugivore Reithrodontomys_mexicanus-ZMB92937 Rodentia Granivore Herbivore Rhabdomys_pumilio-A584571 Rodentia Generalist Herbivore Rhipidomys_sp-A583312 Rodentia Generalist Herbivore Sigmodon_hispidus-ZMB43184 Rodentia Insectivore Granivore Stochomys_longicaudatus-ZMB83974 Rodentia Frugivore Herbivore

220 Silvia Pineda-Munoz - April 2016

Chapter 11: Appendices

Sundamys_muelleri-ZMB39517 Rodentia Generalist Herbivore Uromys_caudimaculatus-ZMB39872 Rodentia Frugivore Herbivore Zygodontomys_sp-ZMB81389 Rodentia Granivore Insectivore

b) Discriminant Specimen Order Diet diet Ailurus_fulgens-1.537 Carnivora Herbivore Granivore Galerella_sanguinea-HEL678 Carnivora Insectivore Carnivore Otocyon_megalotis-ZMB65785 Carnivora Insectivore Frugivore Ursus_arctos-A585456 Carnivora Herbivore Generalist Ursus_maritimus-HELU201 Carnivora Carnivore Generalist Viverra_zibetha-HEL390 Carnivora Carnivore Insectivore Dactylopsila_trivirgata-NMVC4375_L Diprotodontia Insectivore Herbivore Dendrolagus_bennetianus_NMVC7116 Diprotodontia Frugivore Herbivore Dendrolagus_lumholtzi-NMVc7111 Diprotodontia Frugivore Gumivore Petauroides_volans-NMVC6816_L Diprotodontia Herbivore Gumivore Trichosurus_vulpecula_NMV_c3234 Diprotodontia Herbivore Frugivore Cebus_apella_MCZ30726R Primates Generalist Herbivore Chlorocebus_aethiops_MZC31976L Primates Generalist Herbivore Miopithecus_talapoin_MCZ34264L Primates Frugivore Herbivore

Silvia Pineda-Munoz - April 2016 221

Diet, ecology, and dental morphology in terrestrial mammals

Thirty white horses on a red hill,

First they champ,

Then they stamp,

Then they stand still.

[The riddle] was rather an old one, too, and Gollum knew the answer as

well as you do.“Chestnuts, chestnuts,” he hissed. “Teeth! teeth! my

preciousss; but we has only six!”

― J.R.R. Tolkien, The Hobbit

222 Silvia Pineda-Munoz - April 2016