Paleoenvironments and Paleoecologies of Cenozoic Mammals from Western China Based on Stable Carbon and Oxygen Isotopes Dana Michelle Biasatti

Home , Acerorhinus, Amphicyon, Amynodontidae, Cervavitus, Chilotherium, Dicrocerus

Florida State University Libraries

Electronic Theses, Treatises and Dissertations The Graduate School

2009 Paleoenvironments and Paleoecologies of Cenozoic Mammals from Western China Based on Stable Carbon and Oxygen Isotopes Dana Michelle Biasatti

Follow this and additional works at the FSU Digital Library. For more information, please contact [email protected] FLORIDA STATE UNIVERSITY

COLLEGE OF ARTS AND SCIENCES

PALEOENVIRONMENTS AND PALEOECOLOGIES OF CENOZOIC MAMMALS FROM

WESTERN CHINA BASED ON STABLE CARBON AND OXYGEN ISOTOPES

DANA MICHELLE BIASATTI

A Dissertation submitted to the Department of Geological Sciences in partial fulfillment of the requirements for the degree of Doctor of Philosophy

Degree Awarded: Spring Semester, 2009

The members of the Committee approve the Dissertation of Dana Michelle Biasatti defended on February 16, 2009.

______Yang Wang Professor Directing Dissertation

______Gregory Erickson Outside Committee Member

______Leroy Odom Committee Member

______Vincent Salters Committee Member

Approved:

______Leroy Odom, Chair, Department of Geological Sciences

The Graduate School has verified and approved the above named committee members.

To my family.

iii

ACKNOWLEDGEMENTS

I would like to extend special thanks to my supervisor, Dr. Yang Wang, for her advice, encouragement, and financial support throughout this project. I am extremely grateful to Dr. Wang for the research opportunities I have been granted throughout my time at Florida State University. I also thank Dr. Wang for her constructive reviews of this work. This research was funded by the U.S. National Science Foundation (INT-0204923 and EAR-0716235 to Yang Wang). I would also like to thank the Florida State University Department of Geological Sciences and the National High Magnetic Field Laboratory Geochemistry Division for supporting this research. In particular, I would like to thank Dr. Leroy Odom and Dr. Neil Lundberg of Florida State University for finding continuous financial support for me. Furthermore, I thank the FSU Foundation and the Congress of Graduate Students for providing financial support. In addition, I am grateful to Dr. Gregory Erickson, Dr. Leroy Odom, and Dr. Vincent Salters for serving on my advisory committee. I thank Dr. Yingfeng Xu of Florida State University and Dr. Jason Curtis of University of Florida for assistance with sample analyses and Ted Zateslo and Dr. Michael Bizimis of Florida State University for technical assistance in the lab. I would like to thank Dr. Tao Deng of the Institute of Vertebrate Paleontology and Paleoanthropology in Beijing, China, and the Hezheng Museum of Natural History in Hezheng County, Gansu Province, China, and Dr. Feng Gao of the Yunnan Institute of Cultural Relics and Archaeology in Kunming, Yunnan Province, China, for their generous contributions of fossil specimens and literature for this research and for invaluable assistance in the field. I am grateful Dr. Bruce MacFadden of University of Florida and Dr. Lawrence Flynn of Harvard University for assistance in the field and helpful discussions on the specimens. Furthermore, I thank Mabry Gaboardi of Florida State University for helpful discussions and assistance in the lab. Finally, I thank Dr. Louis Jacobs and Dr. Crayton Yapp of Southern Methodist University, who, as my Master’s Thesis advisors, encouraged and inspired me to step beyond the traditional field of paleontology into the world of paleobiogeochemistry.

TABLE OF CONTENTS

List of Tables ...... vii List of Figures ...... viii Abstract ...... xi

1. INTRODUCTION ...... 1

2. MEASUREMENT OF STABLE ISOTOPES OF CARBON AND OXYGEN AS A TOOL FOR PALEOECOLOGICAL AND PALEOCLIMATOLOGICAL RECONSTRUCTION ...... 4

2.1. Preservation of Stable Isotopic Composition in Tooth Enamel ...... 4 2.2. Carbon Isotopes in Paleodiet and Paleoecological Studies ...... 4 2.3. Oxygen Isotopes in Paleoclimate Studies ...... 9

3. STRENGTHENING OF THE EAST ASIAN SUMMER MONSOON REVEALED BY A MARKED SHIFT IN SEASONAL PATTERNS IN DIET AND CLIMATE AFTER 2-3 MA IN NORTHWEST CHINA ...... 13

3.1. Introduction ...... 13 3.2. Study Site ...... 15 3.3. Materials and Methods ...... 15 3.3.1. Sample Materials ...... 15 3.3.2. Experimental Methods ...... 18 3.4. Results and Discussion ...... 21 3.4.1. Assessment of Fossil Tooth Preservation ...... 21 3.4.2. 25 Million Years of Climate Variability in the Linxia Basin ...... 25 3.4.3. Changes in Seasonality and Monsoon Strength ...... 34 3.5. Conclusions ...... 48

4. PALEOECOLOGY OF CENOZOIC RHINOS FROM NORTHWEST CHINA: A STABLE ISOTOPE PERSPECTIVE ...... 52

4.1. Introduction ...... 52 4.2. Study Site ...... 54 4.3. Materials and Methods ...... 56 4.3.1. Sample Materials ...... 56 4.3.2. Laboratory Methods ...... 61 4.4. Results and Discussion ...... 63 4.4.1. Late Oligocene Rhinoceroses ...... 63

4.4.2. Middle Miocene Rhinoceroses ...... 73 4.4.3. Late Miocene Rhinoceroses ...... 76 4.4.4. Pliocene Rhinoceroses ...... 86 4.5. Conclusions ...... 87

5. PALEOECOLOGIES AND PALEOCLIMATES OF CENOZOIC MAMMALS FROM YUNNAN PROVINCE, CHINA, BASED ON STABLE CARBON AND OXYGEN ISOTOPES ...... 89

5.1. Introduction ...... 89 5.2. Study Site ...... 90 5.3. Materials and Methods ...... 97 5.3.1. Sample Materials ...... 97 5.3.2. Laboratory Methods ...... 98 5.4. Results and Discussion ...... 100 5.4.1. Assessment of Fossil Tooth Preservation ...... 100 5.4.2. Carbon Isotope Compositions of Fossil Mammals from Six Localities in Yunnan ...... 100 5.4.3. Oxygen Isotope Compositions of Fossil Mammals from Six Localities in Yunnan ...... 109 5.5. Conclusions ...... 114

6. CONCLUSION ...... 117

APPENDIX A: Data from Analyses of Bulk Carbonate Samples ...... 122

APPENDIX B: Data from Analyses of Serial Carbonate Samples ...... 130

APPENDIX C: Data from Analyses of Organic Samples ...... 143

APPENDIX D: Data from Analyses of Phosphate Samples ...... 145

APPENDIX E: Details of Experimental Methods ...... 146

REFERENCES ...... 150

BIOGRAPHICAL SKETCH ...... 167

LIST OF TABLES

Table 3.1: Statistical significance of differences in mean oxygen isotopic compositions of various taxa at different ages as determined by two-tailed t-tests ...... 27

Table 3.2: Carbon isotope compositions of plants from the Linxia Basin ...... 49

Table 4.1: Lithology and paleontology of the Linxia Basin stratigraphic sequence ...... 58

Table 4.2: Statistical significance of differences in mean carbon and oxygen isotopic compositions of various rhino taxa at different ages as determined by two-tailed t-tests ...... 67

Table 4.3: Statistical significance of differences in mean carbon and oxygen isotopic compositions of various rhino taxa at different ages as determined by two-tailed t-tests ...... 72

Table 5.1: Carbon isotope compositions of soil carbonates and organic matter from Yunnan Province ...... 106

Table 5.2: Carbon isotope compositions of grasses from Lufeng and Yuanmou localities in Yunnan province ...... 106

vii

LIST OF FIGURES

Figure 2.1: Carbon isotope fractionation in mammalian herbivores ...... 6

Figure 2.2: C3 and C4 photosynthetic pathways ...... 8

Figure 2.3: Oxygen isotope fractionation in the hydrologic cycle and its variation with latitude ...... 11

Figure 2.4: The measurement of oxygen isotopes in tooth enamel as a tool for paleoclimate reconstruction ...... 12

Figure 3.1: Location of the Linxia Basin ...... 16

Figure 3.2: Late Cenozoic sedimentary sequence of the Linxia Basin ...... 17

18 Figure 3.3: Δ Oc-p values of fossil rhino enamel from this study versus 18 calculated linear regressions from Δ Oc-p values of tooth enamel and bone from extant mammals and the shells of extant marine organisms Late Cenozoic sedimentary sequence of the Linxia Basin ...... 23

Figure 3.4: Differences in δ13C and δ18O values between fossil mammalian enamel samples and their coexisting matrix carbonates 13 18 (Δ Ce-m and Δ Oe-m) versus age ...... 24

Figure 3.5: Carbon and oxygen isotopic compositions of horses and rhinos that lived concurrently in the Linxia Basin throughout the last 25 million years ...... 26

Figure 3.6: Individual δ18O values of fossil tooth enamel from different taxonomic groups from the Linxia Basin versus age ...... 30

Figure 3.7: Individual δ13C values of fossil tooth enamel from different taxonomic groups from the Linxia Basin versus age ...... 33

Figure 3.8: Location of selected International Atomic Energy Agency (IAEA) stations within and outside of the summer monsoon region in China ...... 36

Figure 3.9: Weighted monthly mean δ18O values of precipitation at selected International Atomic Energy Agency (IAEA) stations within the

viii

summer monsoon region in China ...... 37

Figure 3.10: Weighted monthly mean δ18O values of precipitation at selected International Atomic Energy Agency (IAEA) stations outside of the summer monsoon region in China ...... 38

Figure 3.11: Expected patterns in serial δ18O and δ13C records for tooth enamel samples from localities within and outside of the summer monsoon region in China ...... 40

Figure 3.12a: Serial carbon and oxygen isotope compositions of horse tooth enamel from 11.5 to 6 Ma ...... 41

Figure 3.12b: Serial carbon and oxygen isotope compositions of horse tooth enamel from 4 to 0.05 Ma ...... 42

Figure 3.13: Serial carbon and oxygen isotope compositions of elephant (Gomphotherium and Platybelodon) and giraffe (Palaeotragus) tooth enamel from 17 to 4 Ma ...... 43

Figure 3.14: Serial carbon and oxygen isotope compositions of bovid tooth enamel from 2.5 Ma to the present ...... 44

Figure 3.15: Horse and rhino mean bulk δ18O values versus δ18O ranges of all serial-sampled mammals of various ages ...... 45

Figure 4.1: Map of China showing location of the Linxia Basin ...... 55

Figure 4.2: Sedimentary sequence of the Linxia Basin ...... 57

Figure 4.3: Bulk carbon isotope compositions of tooth enamel from rhinos from the Linxia Basin versus age ...... 64

Figure 4.4: Bulk oxygen isotope compositions of tooth enamel from rhinos from the Linxia Basin versus age ...... 65

Figure 4.5: δ13C range versus δ18O range for rhino individuals of particular genera at given geologic ages ...... 66

Figure 4.6: Serial carbon and oxygen isotope compositions of rhino tooth enamel from the Late Oligocene and Middle Miocene ...... 71

Figure 4.7: Serial carbon and oxygen isotope compositions of rhino tooth enamel from the Late Miocene ...... 79

Figure 5.1: Map of Yunnan Province ...... 91

Figure 5.2: Carbon and oxygen isotope compositions of fossil mammalian tooth enamel and coexisting matrix ...... 101

Figure 5.3: Individual δ13C values of fossil tooth enamel from six localities in Yunnan Province versus age ...... 102

Figure 5.4: Variations in carbon isotopic compositions of various mammalian taxa from Yuanmou with time ...... 103

Figure 5.5: Individual δ18O values of fossil tooth enamel from six localities in Yunnan Province versus age ...... 110 Figure 5.6: Variations in oxygen isotopic compositions of various mammalian taxa from Yuanmou over time ...... 111

Figure E.1: Sampling methods ...... 147

ABSTRACT

Three main objectives in this study were: 1) to examine climate variability throughout the Late Cenozoic and test hypotheses regarding the development of C4 ecosystems and the dynamics of the Asian monsoons in NW China; 2) to reconstruct the diets, habitats, and paleoclimates of fossil rhinocerotoids from the Linxia Basin, Gansu, China; and 3) to examine paleodiets, paleoecologies, and paleoclimates of extinct taxa and test previous hypotheses regarding expansion of C4 grasses in SW China. To examine climate variability in NW China throughout the Late Cenozoic and to test hypotheses regarding the development of C4 ecosystems and the dynamics of the Asian monsoons, the carbon and oxygen isotopic compositions of bulk tooth enamel samples from 158 fossil mammals from the Linxia Basin, ranging in age from 25 Ma to the present, were determined and serial carbon and oxygen isotopic analyses of 368 samples from 23 fossil and modern herbivore teeth were performed. The results indicated significant changes in the climates and diets of mammalian taxa from the Linxia basin, as well as in the seasonal patterns of diet and climate, over the last 25 million years. The bulk oxygen isotope data indicated an unstable climate in the Linxia Basin from 25 to 0.05 Ma and fluctuations in the oxygen isotope data throughout the entire sedimentary sequence were consistent with previous studies that indicated a general global warming trend from ~26 Ma to 15 Ma and two major cooling phases during the Neogene. In addition, a positive δ18O shift in the data was similar in timing to a positive δ18O shift observed in fossils and paleosols from Pakistan and Nepal, suggesting a shift toward a drier and/or warmer climate on both the north and south sides of the Tibetan Plateau during the Late Miocene. Bulk δ18O values of fossil tooth enamel from bovids, deer, giraffes, pigs, and elephants were consistent with the positive and negative trends in horse and rhino mean δ18O values. The bulk carbon isotope results indicated that horses and rhinos from the Linxia Basin had pure C3 diets throughout most of the Late Cenozoic. The horse bulk δ13C values indicated a change to a mixed C3/C4 diet after 2.5 Ma, suggesting that C4 grasses may have not spread into the basin until after 2.5 Ma. This is much later than the proposed global C4 expansion during the Late Miocene and indicates a strengthening of the Asian summer monsoon after ~2.5 Ma, as C4

xi plants require sufficient summer precipitation. The data also indicated an open environment, such as a savannah or mixed woodland/steppe biome in the Linxia Basin from ~25 to 0.05 Ma. The carbon isotope compositions of enamel from bovids, deer, giraffes, pigs, and elephants were similar to those of horses and rhinos at any given age. Serial oxygen isotopic analyses showed that positive shifts to either drier and/or warmer conditions after 14, 9.5, 7.5, and 2.5 Ma were accompanied by increased seasonality and negative shifts in the bulk data at 11.5, 6.0, 4.0, and 1.2 Ma were associated with decreased seasonality. A marked increase in the serial δ18O ranges of both horses and bovids after 2.5 Ma is consistent with a strengthening of the summer monsoon in the region after ~2-3 Ma. The serial carbon isotope results showed that prior to 1.2 Ma, all sampled mammalian taxa had pure C3 diets. The δ13C ranges of all horses from or prior to the age of 2.5 Ma were smaller than those of horses from 1.2 and 0.05 Ma, which further supports changes in the composition of plant biomass in the Linxia Basin after ~2.5 Ma, as taxa with mixed C3/C4 diets would have increased δ13C ranges in their enamel compared to those with pure C3 diets. A negative correlation between the δ18O and δ13C values of horses from ~1.2 and ~0.05 Ma is consistent with that expected in summer monsoon regions within China and strongly supports a strengthening of the summer monsoon after 2-3 Ma. Serial analyses of five bovid individuals from ~2.5 Ma and later also showed an anti-correlation between δ13C and δ18O values for all individuals, providing further support for an enhanced monsoon climate since about 2-3 Ma. The diets, habitats, and paleoclimates of fossil rhinocerotoids from the Linxia Basin, Gansu, China, ranging in age from 25 to 2.5 Ma, were reconstructed based on bulk and serial carbon and oxygen isotope analyses of tooth enamel. Bulk isotope analyses of 47 rhino individuals representing 11 genera and serial analyses of 15 of those individuals were performed. In addition, the bulk C and O isotopic compositions of teeth from 5 9-Ma Chilotherium individuals from the nearby Tianshui Basin were determined. The results support many previous hypotheses inferred from taxonomy and cranial and limb morphology and offer new insight on the paleoecologies of some genera. The isotopic results support the following previous hypotheses: the Late Oligocene rhino Paraceratherium inhabited a forested environment, and the coexisting rhino Allacerops lived in a relatively open habitat and had a less specialized diet; the Early Miocene Hispanotherium grazed in open steppe territory, whereas the contemporaneous Alicornops had a more generalized diet in a forested environment; and the Late Miocene rhino

xii

Parelasmotherium grazed in an open steppe habitat. The data are inconsistent with previous inferences that the rhinos Acerorhinus and Dicerorhinus dwelled in forested environments. Instead, the results indicate that these two rhinos inhabited open steppe environments. The isotopic results are not conclusive concerning the habitat of Iranotherium, but support previous hypotheses that this rhino was a specialized C3 grazer. The results also suggest that Chilotherium was a forest-dweller throughout much of the Late Miocene, but occupied a more open environment by the end of the Late Miocene. Additionally, the results are consistent with previous hypotheses that the Pliocene rhinos Shansirhinus and Coelodonta were grazers in open habitats. In general, the oxygen isotope data suggest a warming and/or drying trend in the Linxia Basin from the Late Oligocene to Late Miocene with cooling phase throughout much of the Pliocene and indicate that the regional climate was not strongly influenced by the Asian monsoon prior to 6 Ma. Finally, the carbon isotope data support that all rhinos in this study were pure C3 feeders, which suggests that C4 grasses were not an important component of the plant biomass in the Linxia Basin prior to 2.5 Ma. To examine paleodiets, paleoecologies, and paleoclimates of extinct taxa and to test previous hypotheses regarding a global expansion of C4 grasses, the carbon and oxygen isotopic compositions of 164 fossil herbivore teeth and 10 soil samples from six localities in Yunnan Province, China, ranging in age from ~10 Ma to the present, were determined. The results reveal significant changes in the environments and diets of mammalian taxa from various regions of Yunnan Province over the last 10 million years. Prior to 2-3 Ma, while most mammals had pure or nearly pure C3 diets, some individuals may have consumed a small amount of C4 grasses (<30% C4). Since then, C4 grasses became a significant dietary component of most herbivores as indicated by higher enamel-δ13C values at Yuanmou and Shangri La, most likely reflecting an increased C4 biomass in local ecosystems. The carbon isotope results showed that the diets of mammals aged ~2.5 to 1.75 Ma from Shangri La ranged from pure C3 to pure C4, while 1.7 Ma horses from Yuanmou had 0-50% C4 in their diets. The paleoenvironment of Zhaotong was more open and/or water-stressed than that of Kaiyuan at ~4 Ma. Mammals living at both ~8 and 7.5 Ma in the Lufeng region had very similar diets, habitats, and experienced similar climatic conditions. Increased C4 biomass after ~2-3 Ma suggests a significant change in certain aspects of the regional climate, such as increased seasonality of rainfall or an increase in seasonal drought and fires, as these factors are important to modern grasslands. The data also indicated a change from a largely forested environment at ~8 Ma to an increasingly drier and more open

xiii environment with a mosaic of forests and grasslands after ~2-3 Ma in the Yuanmou region. Niche partitioning of various taxa from the Yuanmou Basin was also evident from the carbon isotope results. The data suggested that horses and rhinos fed in more closed or forested environments than did pigs, tragulids, and chalicotheres and that elephants may have fed in both closed and open environments. The oxygen isotope compositions of ~7.5 and 8 Ma mammals from Lufeng suggest very similar climatic conditions at both ages in that region. The oxygen isotope results show a positive shift after ~8-8.5 Ma in the Yuanmou region, which is similar in timing to shifts observed in horses, rhinos, and deer from the Linxia Basin and in fossils and paleosols from Pakistan and Nepal, suggesting a shift toward a drier and/or warmer climate at the northeast, southeast, and southern borders of the Tibetan Plateau during the Late Miocene. A negative shift in the δ18O values of rhinos from the Yuanmou Basin after ~5 Ma likely indicates a change to a wetter environment at that time interval, which is similar in timing to a negative shift observed in the Linxia Basin and Gyirong Basin. Overall, the oxygen isotope compositions of mammals from the Yuanmou Basin indicate a general drying of the local climate over time, which is consistent with carbon isotope results from that region.

xiv

CHAPTER 1

INTRODUCTION

Measurements of natural abundances of stable carbon and oxygen isotopes in fossil mammalian tooth enamel and terrestrial sediments are important tools in the study of past continental environments and ecologies. Carbon and oxygen isotopes in fossil tooth enamel and paleosols provide proxy records for paleoclimates and paleoecologies when direct evidence is not available. Because instrumental measurements cover less than 10-7 of the Earth’s climate record, proxy records such as these are necessary to gain a more complete perspective on the Earth’s climate history (Bradley, 1999). By observing long-term trends in climate, we may be able to identify mechanisms of climate change which can then be modeled to predict future variations in climate. Furthermore, because carbon and oxygen isotope compositions of mammalian tooth enamel also reflect paleodiets, behaviors, and physiological processes, the use of such proxies may allow insight into to the biology and ecology of ancient taxa and increase our understanding of the effects of a changing environment on the evolution of vertebrates. The objectives of this research were to reconstruct the paleoenvironments and paleoecologies of fossil mammalian herbivores from Late Cenozoic deposits in Gansu and Yunnan provinces, located along the northeastern and southeastern margins of the Tibetan Plateau in western China, respectively, and to establish temporal and spatial records of vegetation and climate change in western China, based on stable carbon and oxygen isotopic analyses of fossil tooth enamel and ancient sediments. Direct paleoenvironmental records from western China are rare, but carbon and oxygen isotopic compositions of fossil tooth enamel and sediments that were collected from deposits in the Linxia Basin in Gansu Province and six localities in Yunnan Province were used as proxies for paleoclimatic and paleoecological conditions in western China over the last 25 m.y. Because it has been suggested that during the Late Miocene, C4 grasses became a significant component of terrestrial ecosystems (e.g., Quade et al., 1989; Cerling, 1992; Cerling et al., 1993, 1997a,b; MacFadden et al., 1994; MacFadden

1 and Cerling, 1994; Wang et al., 1994), and because those two regions of western China have sequences of fossils and sediments that occur prior to and after the proposed global expansion of C4 grasses, stable isotope analyses of the fossils and paleosols from those sequences provided a means of testing the global C4 expansion hypothesis. Those analyses also provided a way to explore the dietary evolution of extinct taxa, as well as a means to test hypotheses where physiological and behavioral characteristics were based on skeletal morphology, and allowed a greater understanding of ancient ecologies. The records of vegetation and climate change based on those isotopic analyses, and comparison of those records with well-established records on the south side of the Himalayan Tibetan Plateau and in other regions of the world, allowed an investigation into the effects of the uplift of the Tibetan Plateau on regional and global climate and on the evolution of mammalian species. It is well-known that the Tibetan Plateau strongly influences the Asian Monsoon circulation (Kutzbach et al., 1993; Yanai and Li, 1994; Webster et al., 1998; An et al., 2000; Wang and Deng, 2005; Wang et al., 2008a,b), although the timing of the uplift of the plateau, and thus, the timing of the development of the Asian Monsoon system, is strongly debated (e.g., Harrison et al., 1992; Coleman and Hodges, 1995; Yin and Harrison, 2000; Zheng et al., 2000; Dettman et al., 2001; Garzione et al., 2000; Spicer et al., 2003; Wang et al., 2006). By reconstructing ancient environments along the margins of the Tibetan Plateau, we may increase our understanding of how tectonic processes influence climate change and evolutionary processes. In this study, the carbon and oxygen isotopic compositions of bulk enamel samples from 158 individual fossil herbivore teeth from the Linxia Basin, ranging in age from 25 Ma to the present, were determined and serial carbon and oxygen isotopic analyses of 368 samples from 23 fossil and modern herbivore teeth were performed. In addition, the carbon and oxygen isotopic compositions of tooth enamel samples from 166 individual fossil herbivore teeth from Yunnan Province, ranging in age from 10 Ma to the present, were determined and carbon and oxygen isotopic analyses of carbonates and organic matter from 28 soil samples from the Yunnan region were performed. The carbon and oxygen isotope data were used to examine the climate variability in NW and SW China during the Late Cenozoic and to test the hypotheses regarding the development of C4 ecosystems and the dynamics of the Asian monsoons. The carbon isotope compositions of 54 modern plants from the Linxia Basin and 9 modern plants from

Yunnan Province were determined for comparison with plant compositions of ancient ecosystems. In order to examine the dietary evolution of the Rhinocerotoidea in NW China, to understand how these ancient animals lived and interacted with their environment, and to test hypotheses based on morphological characters of teeth and bones, the stable carbon and oxygen isotopic ratios of bulk tooth enamel samples from 47 rhino individuals, ranging in age from 25 to 2.5 Ma and representing 11 genera from the Linxia Basin, and 5 individuals representing 1 genus from the nearby Tianshui Basin were determined. In addition, the carbon and oxygen isotopic compositions were determined for 243 serial samples from 15 rhino individuals representing 6 genera that existed in the Linxia Basin from 25 to 6 Ma. The bulk and serial isotopic compositions of common genera were compared, both temporally and spatially, in order to examine behavioral, physiological, and climatic effects on the evolution of those genera. The results of this study are presented as follows: Chapter 2, Measurement of Stable Isotopes of Carbon and Oxygen as a Tool for Paleoecological and Paleoclimatological Reconstruction, provides background information about the use of stable isotopes in reconstructing paleoclimate and paleoecology in terrestrial ecosystems; Chapter 3, Strengthening of the East Asian Summer Monsoon Revealed by a Marked Shift in Seasonal Patterns in Diet and Climate after 2-3 Ma in Northwest China, presents an isotopic record of Late Cenozoic climate and ecosystem changes in the Linxia Basin located on the NE margin of the Tibetan Plateau; Chapter 4, Paleoecology of Cenozoic Rhinos from Northwest China: A Stable Isotope Perspective, discusses the paleoecologies of rhinos in western China based on stable carbon and oxygen isotopic analyses of fossils; and Chapter 5, Paleoecologies and Paleoclimates of Cenozoic Mammals from Yunnan Province, China, Based on Stable Carbon and Oxygen Isotopes, reports the results of stable carbon and oxygen isotopic analyses of fossil mammals and interprets the isotope data in terms of paleodiet, paleoecology and paleoenvironment.

CHAPTER 2

MEASUREMENT OF STABLE ISOTOPES OF CARBON AND OXYGEN AS A TOOL FOR PALEOECOLOGICAL AND PALEOCLIMATOLOGICAL RECONSTRUCTION

2.1. Preservation of Stable Isotopic Composition in Tooth Enamel

Tooth enamel consists of an inorganic component (>96% by weight) and an organic component (<1% by weight). The inorganic component is mainly in the form of hydroxyapatite - (Ca10(PO4)6(OH)2), which contains a small amount (1-3%) of structural carbonate (CO3 ) substituting for phosphate and hydroxyl ions. The organic component is mainly in the form of a protein (collagen). Dentine and bone consist of 65-75 % inorganic component and 17-26 % organic component. Porosity increases from about 1 % to about 40 % from enamel to dentine to bone (Brudevold and Soremark, 1967; Rowles, 1967; Trautz, 1967; Wang and Cerling, 1994). Because tooth enamel has a low organic content, low porosity, resulting in less influx of diagenetic fluids, and a large crystal size (Ayliffe et al., 1994; Wang and Cerling, 1994) tooth enamel is better suited, relative to other forms of biogenic apatite such as that in bones and dentine, to preserve the original isotopic composition of an animal (Wang and Cerling, 1994). Therefore, tooth enamel is the most suitable form of biogenic apatite to analyze in paleoenvironmental studies that involve stable isotope measurements.

2.2. Carbon Isotopes in Paleodiet and Paleoecological Studies

There are two stable isotopes of carbon, 13C and 12C, which differ by the mass of one neutron in the nucleus of the carbon atom. Because 12C is utilized preferentially over 13C in photosynthesis, organic matter derived from photosynthesis is 13C-depleted relative to atmospheric CO2 (O’Leary, 1988; Farquhar et al., 1989; Boutton, 1991). As animals ingest organic matter, an additional internal fractionation occurs as carbon becomes incorporated into

their tissues. In a single animal, muscle is generally 13C-enriched by ~1 ‰ relative to diet (Koch et al., 1994), whereas tooth enamel carbonate is generally 13C-enriched by ~14 ‰ relative to diet (Figure 2.1; Lee-Thorp and van der Merwe, 1987; Cerling and Harris, 1999). Further 13C- enrichment occurs progressively at each trophic level (Koch et al., 1994). The ratio of 13C to 12C is expressed in δ13C values, where δ is the deviation in per mil (‰) of the 13C/12C ratio of the sample of interest from the carbonate standard of Vienna Pee-Dee belemnite, or V-PDB (Gonfiantini et al., 1995). δ13C is defined as follows:

13 13 12 13 12 δ C = [ ( C/ C) SAMPLE – ( C/ C) V-PDB ] x 1000 13 12 ( C/ C)V-PDB

The measurement of natural abundances of stable carbon isotopes in tooth enamel structural carbonate, soil carbonate, and organic matter is an important tool for the study of paleovegetation and paleodiet and can provide valuable information regarding entire ecosystems (DeNiro and Epstein, 1978; Peterson and Fry, 1987; Cerling et al., 1989; Boutton, 1991; Gearing, 1991; Koch et al., 1994; Wang and Cerling, 1994; Koch, 1998; Cerling and Harris, 1999; MacFadden, 2000). Most of the carbon isotopic variation in mammalian herbivore tooth enamel is a result of isotope fractionation that occurred during photosynthesis of plants that were subsequently ingested by the herbivores (Farquhar et al., 1989; Wang and Deng, 2005). In the case of soil carbonate and soil organic matter, the carbon isotopic variation is a result of fractionation that occurred in plants that grew in the soil (Cerling et al., 1989; Koch, 1998; Wang and Deng, 2005). An additional source of isotopic variation is water availability. As plants become water-stressed or grow in a more open environment where more evaporation occurs, δ13C values increase (Wang and Deng, 2005). In closed-canopies or forests, δ13C values decrease, due to the influence of soil respiration (Sternberg et al., 1989; van der Merwe and Medina, 1989). There are three major photosynthetic types of terrestrial plants: C3, C4, and CAM. Each type has a unique carbon isotopic signature as a result of isotopic discrimination by CO2-fixing enzymes within plant tissues and limitations to the diffusion of CO2 into the leaf (O’Leary, 1988;

Boutton, 1991). C3 plants use the C3 or Calvin Cycle photosynthetic pathway and reduce CO2

+ 14 ‰ as plant becomes incorporated into tooth enamel

C3 PLANTS TOOTH ENAMEL all trees, cool season grasses, most shrubs and forbs: CARBONATE

13 is 13C-enriched relative to the plant δ CRange = -34 to -22 ‰ ingested 13 δ CAve = -27 ‰ 13 δ C < -8 ‰ ► Pure C3 Diet 13 δ C > -3 ‰ ► Pure C4 Diet C4 PLANTS warm season grasses, a few shrubs, corn, sugar cane, sorghum, bamboo:

13 δ CRange = -17 to -9 ‰ 13 δ CAve = -13 ‰

Figure 2.1. Carbon isotope fractionation in mammalian herbivores. There is an approximately 14 ‰ 13C-enrichment as ingested plant material becomes incorporated into mammalian tooth enamel (Lee-Thorp and van der Merwe, 1987; Cerling and Harris, 1999). 13 Therefore, an herbivore with a pure C3 diet (δ CDIET range of -34 to -22 ‰) will have tooth 13 13 enamel carbonate δ C values that are < -8 ‰, and an herbivore with a pure C4 diet (δ CDIET range of -17 to -9 ‰) will have enamel carbonate δ13C values that are > -3 ‰. A tooth enamel δ13C between -3 and –8 ‰ indicates a mixed C3/C4 diet. In water-stressed conditions or more open environments, δ13C values increase. In closed canopies, δ13C values decrease. δ13C values represent the deviation, in per mil, of 13C/12C of the sample from that of the standard V-PDB.

to a 3-C compound, phosphoglycerate or PGA, via the enzyme ribulose bisphosphate 13 carboxylase/oxygenase, or Rubisco (Figure 2.2). Rubisco discriminates against CO2, resulting in a relatively low carbon isotopic signature for C3 plants, which have a δ13C range of -34 to -22 ‰, with an average of -27 ‰ (Deines, 1980; O’Leary, 1988; Farquhar et al., 1989; Boutton, 1991). Because tooth enamel carbonate is 13C-enriched by ~14 ‰ relative to diet (Lee-Thorp and Van der Merwe, 1987; Cerling and Harris, 1999), tooth enamel δ13C values that are <-8 ‰ usually indicate a pure C3 diet (Figure 2.1; Wang and Deng, 2005). C3 species are currently the most abundant plant types on the Earth, as most temperate zone and all forest communities are dominated by C3 plants (Boutton, 1991). These plants include all trees, most shrubs, cool season grasses, and forbs (Wang and Deng, 2005). Most aquatic plants are C3 species and use bicarbonate as a carbon source rather than atmospheric CO2 (Smith and Walker, 1980). C3 plants are dominant in cooler regions, such as those at high latitudes (Cerling et al., 1993; Wang and Deng 2005).

C4 plants use the C4 or Hatch-Slack photosynthetic pathway and reduce CO2 to a 4-C acid via the enzyme phosphoenol pyruvate (PEP) carboxylase (Figure 2.2). PEP carboxylase 13 does not discriminate against CO2 to the extent that Rubisco does. This results in C4 plants having higher δ13C values relative to C3 plants. Extant C4 plants have a δ13C range of -17 to -9 ‰, with an average of -13 ‰ (Boutton, 1991; Wang and Deng, 2005). Thus, tooth enamel δ13C values that are >-3 ‰ indicate a pure C4 diet (Figure 2.1; Wang and Deng, 2005). Consequently, C3 and C4 plants each have distinct δ13C values that do not overlap and, in general, differ from each other by about 14 ‰ (Smith and Epstein, 1971). C4 plants include warm season grasses, a few shrubs, corn, sugar cane, sorghum, and bamboo. Today, these plants are adapted to low atmospheric CO2 levels, high temperatures, and water-stressed conditions and are abundant in warm regions with summer precipitation (Wang and Deng, 2005), such as those at low latitudes and low altitudes (e.g., Teeri and Stowe, 1976; Boutton, 1991; Cerling et al., 1993). CAM plants use the Crassulacean acid metabolism pathway and can minimize water loss 13 by fixing CO2 at night via the enzyme PEP carboxylase. Therefore, many CAM plants have δ C values typical of C4 plants. Some CAM plants are also able to fix CO2 during the day via Rubisco and, therefore, the δ13C values are dependent on the relative proportions of carbon that were fixed by Rubisco and PEP carboxylase. Most CAM plants have δ13C values that range from -20 to -10 ‰, but some have δ13C values as low as -28 ‰ (Boutton, 1991). Therefore, animals 7

Figure 2.2. C3 and C4 photosynthetic pathways. A) The C3 or Calvin Cycle photosynthetic pathway reduces CO2 to a 3-C compound, PGA, via the enzyme Rubisco. Rubisco discriminates 13 against CO2, resulting in a relatively low carbon isotopic signature for C3 plants. B) The C4 or Hatch-Slack photosynthetic pathway reduces CO2 to a 4-C acid via the enzyme PEP carboxylase, 13 which does not discriminate against CO2 to the extent that Rubisco does. This results in C4 plants having higher δ13C values than do C3 plants (Deines, 1980; O’Leary, 1988; Farquhar et al., 1989; Boutton, 1991).

that have strict CAM plant diets would be expected to have tooth enamel δ13C values that are between -14 and +4 ‰, with most having δ13C values between -6 and +4 ‰. CAM plants include the succulents and are currently commonly found in desert regions but are rare in other ecosystems (Boutton, 1991; Wang and Deng, 2005). The isotopic composition of soil organic matter reflects the photosynthetic pathway of the dominant plant species in a given area. In well-drained mineral soils, organic matter becomes 13C-enriched by 1 to 3 ‰ relative to the source material with increasing depth (Stout et al., 1981). Soil carbonate may be formed in alkaline, arid, or semi-arid soils and its carbon isotopic 13 composition is dependent on the δ C values of the soil CO2, which is generally about 5 ‰ more 13C-enriched than the associated organic matter due to diffusion, and an equilibrium isotope fractionation of about 10 ‰ (depending on temperature) during the formation of the calcium carbonate from soil CO2 (Cerling, 1984). Carbonates that are formed at depths greater than 30 cm in a soil with moderate to high respiration rates have δ13C values that are 13C-enriched by 14-17 ‰ relative to coexisting soil organic matter (Cerling et al., 1989).

2.3. Oxygen Isotopes in Paleoclimate Studies

Oxygen isotope ratios (18O/16O) of meteoric water vary from region to region as a result of evaporative and condensative processes. In these processes, evaporation favors water 16 molecules containing the lighter isotope of oxygen, H2 O, and condensation favors water 18 molecules with the heavier oxygen isotope, H2 O (Epstein and Mayeda, 1953; Craig, 1961). The ratio of 18O to 16O is expressed in δ18O values, where δ is the deviation in parts per mil (‰) of the 18O/16O ratio of the sample from that of a standard. The most commonly used standard for oxygen isotopes is Vienna Standard Mean Ocean Water, or V-SMOW (Gonfiantini, 1978). δ18O is defined by the following equation:

18 18 16 18 16 δ O = [ ( O/ O) SAMPLE – ( O/ O) V-SMOW ] x 1000 18 16 ( O/ O) V-SMOW

18 Typical modern ocean δ O values are 0±1 ‰ (Craig and Gordon, 1965). Values outside that range occur with evaporation, formation of sea ice, or addition of meteoric precipitation

(Epstein and Mayeda, 1953; Criss, 1999). Water vapor formed above the ocean is generally 11 to

14 ‰ more negative than the ocean water below. As water vapor moves inland, it becomes more depleted in 18O. The precipitation that occurs as a result of condensation of the vapor has a more positive δ18O value than the vapor, but becomes more depleted in 18O as the vapor moves inland (Figure 2.3; Epstein and Mayeda, 1953). δ18O values also decrease with high elevations and high latitudes (Craig, 1961). The measurement of oxygen isotopes in tooth enamel is an important tool for paleoclimate reconstruction. Because tooth enamel hydroxyapatite is precipitated in equilibrium with body water, the oxygen isotopic composition of phosphate and, most likely, structural carbonate in tooth enamel reflects the isotopic composition of body water, which, in turn, is determined by the isotopic composition of local meteoric water (i.e., drinking water and water in plants that are consumed). Physiological processes and behavioral characteristics also affect the oxygen isotopic composition of body water (Longinelli, 1984; Luz et al., 1984; Koch et al., 1989; Nagy, 1989; D’Angela and Longinelli, 1990; Kohn, 1996). Because the oxygen isotopic composition of meteoric water is affected by climatic variables such as ambient temperature, amount of precipitation, and seasonality of precipitation (Dansgaard, 1964; Rozanski et al., 1992), the δ18O values of phosphate or structural carbonate in tooth enamel serve as a proxies for climatic conditions that were present during the time of tooth growth (Longinelli, 1984; D’Angela and Longinelli, 1993; Koch et al., 1989; Wang and Deng, 2005). A significant shift in the δ18O values of tooth enamel from a given species over time would suggest a change in climate, such as a change in temperature or humidity (Figure 2.4; Wang and Deng, 2005). Because mammalian tooth growth is incremental, with the crown being the oldest part of the tooth and the base being the youngest, the tooth enamel preserves a time-series of δ18O values that reflect seasonal variations in climate within a given individual (Figure 2.4; Fricke and O’Neil, 1996; Kohn et al., 1998; Sharp and Cerling, 1998).

Figure 2.3. Oxygen isotope fractionation in the hydrologic cycle and its variation with latitude (based on data from Epstein and Mayeda, 1953; Craig, 1961; and Craig and Gordon, 1965). δ values represent the per mil deviation from the standard SMOW.

Figure 2.4. The measurement of oxygen isotopes in tooth enamel as a tool for paleoclimate reconstruction. A) The oxygen isotopic composition of tooth enamel carbonate reflects that of the local meteoric water that is consumed by animals. Because the δ18O of meteoric water is controlled by climate, a shift in δ18O indicates a change in regional climate (Dansgaard, 1964; Longinelli, 1984; Rozanski et al., 1992; D’Angela and Longinelli, 1993; Koch et al., 1989). B) Seasonality is ideally represented by a sinusoidal curve when plotting δ18O values against distance from the crown of the tooth of interest, where the peaks generally represent the summer season and the troughs represent winter months (Fricke and O’Neil, 1996; Kohn et al., 1998; Sharp and Cerling, 1998). In the summer monsoon region, however, the peaks would represent winter months and the troughs would represent summer months as a result of the summer precipitation having considerably lower δ18O values than the winter precipitation (Araguas- Araguas et al., 1998; Johnson and Ingram, 2004).

CHAPTER 3

STRENGTHENING OF THE EAST ASIAN SUMMER MONSOON REVEALED BY A MARKED SHIFT IN SEASONAL PATTERNS IN DIET AND CLIMATE AFTER 2-3 MA IN NORTHWEST CHINA

3.1. Introduction

Paleoclimatic and paleoecological records in western China are very important to the understanding of the effects of the uplift of the Tibetan Plateau on regional climate and on the evolution of mammalian species. Stable carbon and oxygen isotopic analyses of fossil teeth and ancient soils collected along the margin of the Tibetan Plateau provide proxies for those records, as direct paleoenvironmental records are rare for this region of China. Studies of carbon isotopic compositions of fossil mammalian tooth enamel and paleosols from around the world suggest that prior to the Late Miocene, terrestrial ecosystems consisted predominantly of plants that used the C3 photosynthetic pathway. After the Late Miocene (~7 to 5 Ma), C4 grasses became a significant component of low- to mid-latitude and low-elevation regions (Quade et al., 1989; Cerling, 1992; Cerling et al., 1993, 1997a,b; MacFadden et al., 1994; MacFadden and Cerling, 1994; Wang et al., 1994). Studies of oxygen isotopic compositions of paleosols from Pakistan have indicated that a dramatic increase in oxygen isotope ratios occurred with the Miocene expansion of C4 grasses, reflecting a change to a warmer and/or drier climate. It has been suggested that the Miocene global expansion of C4 plants occurred in response to declining atmospheric CO2 levels (Cerling et al., 1993, 1997a,b; Wang et al., 1994), and/or a strengthening of the Asian summer monsoon as a result of the uplift of the Tibetan Plateau (Quade et al., 1989). However, it has also been suggested that there was no global expansion of C4 plants

(Morgan et al., 1994) and that no evidence exists for decreasing CO2 levels during the Late Miocene (Pagani et al., 1999; Retallack, 2001; Royer et al., 2001).

Because direct paleoenvironmental records, such as paleobotanical specimens, are rare from western China, the isotopic analyses of fossil mammalian herbivore tooth enamel and soil carbonates from the Linxia Basin, Gansu Province, that have sequences of fossils and sediments that both pre- and post-date the Late Miocene global expansion of C4 grasses, are an important means of testing the global C4 expansion hypothesis. These analyses also provide a way to explore the dietary evolution of extinct taxa and allow a greater understanding of ancient ecologies. Comparison of isotopic records of vegetation and climate changes based on these analyses with well-established records on the south side of the Himalayan Tibetan Plateau and in other regions of the world will aid in understanding the effects of the Tibetan Plateau uplift on regional and global climate. Wang and Deng (2005) presented a late Cenozoic record of vegetation and environmental change from the Linxia Basin, based on the stable C and O isotope analyses of bulk enamel samples and paleosols. Their data showed that C4 grasses did not expand into the Linxia Basin during the period of “global C4 expansion” until the Quaternary. They hypothesized that the East Asian Summer monsoon (EASM), currently controlling the climatic conditions in the area, was probably not strong enough to affect this part of China throughout much of the Neogene and delayed C4 expansion into the area until the EASM was further strengthened after 2-3 Ma. Based on fossil enamel and paleosol carbonate isotope data from several localities on Chinese Loess Plateau, Passey et al. (2009) recently offered a contrasting hypothesis suggesting a stronger EASM during Late Miocene and Pliocene than today. In this study, the C and O isotopic compositions of bulk enamel samples from 158 individual fossil herbivore teeth from the Linxia Basin, ranging in age from 25 Ma to the present, were determined (Appendix A). In addition, serial C and O isotopic analyses of 368 samples from 23 fossil and modern herbivore teeth were performed (Appendix B). The new C and O isotope data and previously published data were utilized to examine the climate variability in NW China during the Late Cenozoic and to test the hypotheses regarding the development of C4 ecosystems and the dynamics of the Asian monsoons. The carbon isotope compositions of 54 modern plants from the Linxia Basin were determined (Appendix C) for comparison with estimated C3/C4 compositions of ancient ecosystems.

3.2. Study Site

The Linxia Basin (103°E, 35°N) is located in Gansu Province, China, about 100 km south of the provincial capital, Lanzhou, on the northeastern margin of the Tibetan Plateau (Figure 3.1). It is currently considered a temperate steppe biome, with an elevation of ~1917 m (Wang and Deng, 2005), a mean annual temperature of 7°C, and an annual rainfall of 515 mm. The present-day climate is strongly influenced by the East Asian monsoon system (An et al., 2000). The Linxia Basin is a flexural basin bounded by the Laijishan fault to the west, the North Qinling fault to the south, and the Maxian mountains to the north; but to the east, the boundary is poorly defined. It is believed that deposition in the Linxia Basin began at ~29 Ma and continued nearly uninterrupted until ~1.7 Ma (Fang et al., 1997, 2003) and that the Tibetan Plateau grew in a step-wise fashion towards the northeast through time (Fang et al., 1997, 2003; Tapponnier et al., 2001). Evidence suggests that the deformation front of the Tibetan Plateau had propagated into this region by 6 Ma (Fang et al., 1997, 2003). The Cenozoic sedimentary sequence in the Linxia Basin is well-exposed, with an overall thickness of more than 500 meters, and spans almost continuously from the Late Oligocene to the Holocene (Figure 3.2). Additionally, the strata contain abundant, diverse, and well-preserved mammalian fossils, making the Linxia Basin an ideal place to study paleoecology and paleoclimatology using stable carbon and oxygen isotope analyses of mammalian tooth enamel (Deng et al., 2004a, b; Wang and Deng, 2005). The chronology of the Cenozoic deposits has been determined by magnetostratigraphic and biostratigraphic techniques (Fang et al., 1997, 2003; Deng et al., 2004a, b). The deposits consist primarily of fluvial and lacustrine sediments, with Pleistocene eolian deposits covering most of the region (Fang et al., 2003; Deng et al., 2004a, b; Wang and Deng, 2005). The stratigraphic sequence is comprised of eleven formations based on lithofacies and paleontology (Deng et al., 2004b).

3.3. Materials and Methods

3.3.1. Sample Materials 158 well-preserved fossil tooth samples were selected for carbon and oxygen isotopic analyses from the Hezheng Museum of Natural History in Hezheng County, Gansu Province, or

Figure 3.1. Location of the Linxia Basin. The study site is located on the northeastern margin of the Tibetan Plateau, about 100 km south of Lanzhou, the capital city of Gansu Province, China (Modified from Wang and Deng, 2005).

Figure 3.2. Late Cenozoic sedimentary sequence of the Linxia Basin. This well-exposed sequence has an overall thickness of more than 500 meters and spans almost continuously from the L. Oligocene to the Holocene. The deposits are mainly fluvial and lacustrine sediments and Pleistocene eolian loess deposits cover most of the area. The chronology has been determined by magnetostratigrahy and biostratigraphy (Modified from Fang et al., 1997, 2003; Deng et al., 2004b).

from the Institute of Vertebrate Paleontology and Paleoanthropology (IVPP) in Beijing, China. All fossils were originally collected from the Linxia Basin by IVPP. The teeth were chosen from several groups of Late Cenozoic mammals and included the following genera: Bos, Capra, Gazella, Leptobus, Protoryx, Sinotragus, Hezhengia, Equus, Hipparion, Anchitherium, Coelodonta, Shansirhinus, Chilotherium, Dicerorhinus, Acerorhinus, Iranotherium, Parelasmotherium, Alicornops, Hispanotherium, Paraceratherium, Allacerops, Cervavitus, Turcocerus, Palaeotragus, Honanotherium, Samotherium, Microstonyx, Chleuastochoerus, Listriodon, Kubanochoerus, Tetralophodon, Platebelodon, and Zygolophodon. Most samples collected were third molars and premolars to reduce the effects of pre-weaning on the δ18O of the tooth enamel (e.g., Fricke and O’Neil, 1996; Bryant et al., 1996). Multiple specimens from each stratum were collected when possible so that variations in δ13C and δ18O values within populations, between different taxa of the same age, and between similar taxa of different ages could be determined. Matrix sediments were collected from the surface of fossil tooth enamel in order to compare their carbonate carbon and oxygen isotope compositions to the isotopic compositions of the fossil tooth enamel. In addition, 54 plants were collected throughout the Linxia basin for carbon isotope analyses in order to estimate the C3/C4 composition of the modern ecosystem and then compare it with ancient ecosystems inferred from the δ13C of fossil teeth.

3.3.2. Experimental Methods Bulk isotopic analyses of tooth enamel yield average delta values for the growth period of 13 2- a tooth. The bulk carbon and oxygen isotopic compositions of the enamel carbonate (δ C (CO3 18 2- ) and δ O (CO3 )) were determined for each of the 158 tooth samples. Many studies of oxygen isotopic compositions of mammalian tooth enamel have focused primarily on oxygen in tooth enamel phosphate, as carbonate is more susceptible to diagenesis than is phosphate. However, it 18 2- has been shown that δ O (CO3 ) values can be preserved in fossil enamel (Bocherens et al., 1996; Cerling et al., 1997c). The oxygen isotopic composition of tooth enamel phosphate (δ18O 3- (PO4 )) was also determined for 10 rhino samples (Appendix D) from various strata in order to test for the preservation of the primary isotopic compositions of the tooth carbonate samples by comparison with the phosphate isotopic data (e.g., Iacumin et al., 1996). Additionally, bulk carbon and oxygen isotopic analyses were performed on 36 sediment (matrix) samples

(Appendix A) that were collected from the surfaces of individual teeth in order to determine whether the fossil samples were diagenetically altered, as the isotopic compositions of the fossils and corresponding matrix should be similar or identical if alteration has occurred. Based on the isotopic results of the bulk enamel samples from Linxia Basin, tooth samples were selected that spanned the time intervals where significant changes in vegetation and climate occurred in order to determine δ13C and δ18O variation within the life-spans of 13 2- 18 2- individuals by serial sampling. δ C (CO3 ) and δ O (CO3 ) values were determined for 368 serial samples from 23 fossil herbivore teeth. Isotopic analyses of serial samples allow one to reconstruct seasonal patterns in diet and climate or changes in behaviors of individuals (e.g., Koch et al., 1995; Fricke and O’Neil, 1996; Sharp and Cerling, 1998). The results of this study are reported in standard notation as δ13C and δ18O in reference to V-PDB and V-SMOW, respectively (Gonfiantini, 1978; Gonfiantini et al., 1995). δ13C is defined as follows: δ13C = 13 12 13 12 13 12 18 18 18 [(( C / C)SAMPLE - ( C / C)VPDB)/ ( C / C)VPDB] x 1000 ‰. δ O is defined as: δ O = [(( O 16 18 16 18 16 / O)SAMPLE - ( O / O)VSMOW)/ ( O / O)VSMOW] x 1000 ‰. Bulk enamel samples were obtained by either drilling along the entire length of a tooth using a slow-speed rotary drill or by cutting off a section of each tooth, lengthwise from the crown to the root, and manually separating the enamel from the dentine using a rotary tool. Then, the samples were ground into a fine powder using a mortar and pestle. Serial samples were drilled, using a slow-speed rotary tool, perpendicular to the growth axis of each tooth from crown to root, with the youngest samples being near the root and the oldest samples being near the crown. Tooth enamel carbonate samples were pretreated in 5% reagent grade sodium hypochlorite for approximately 20 to 24 hours at room temperature to remove organic material from the tooth enamel. The samples were then centrifuged, decanted, and rinsed with deionized water to remove the sodium hypochlorite. Next, the samples were treated in 1M acetic acid for 4 to 15 hours at room temperature to remove non-structural carbonate from the enamel. The samples were then rinsed with deionized water. After the final rinse, the samples were freeze- dried for 3 to 5 days. To prepare the tooth enamel phosphate samples, tri-silver phosphate was precipitated from the enamel samples following the Dettman et al. (2001) procedure, which is a modification of the approach by O’Neil et al. (1994). First, samples were dissolved in 2M HF in an ultrasonic

19 bath overnight, which simultaneously precipitated CaF2. The solutions were then decanted and were brought to a nearly neutral pH with the addition of 20% NH3OH. Next, 2M AgNO3 was added to each sample, causing rapid precipitation of Ag3PO4. Finally, the samples were centrifuged, decanted, rinsed three times, and were freeze-dried for approximately 3 days. Carbonates from sediment matrix and plant samples were not pretreated prior to isotopic analyses. To analyze the carbonate samples, ~200 to 500 μg of carbonate standards and 3 to 6 mg of enamel or matrix carbonates were measured and placed into reaction vials capped with rubber septa. The sample vials were then loaded into a Thermo-Finnigan Gasbench II interfaced with a Delta Plus XP continuous flow isotope ratio mass spectrometer and were flushed for 5 minutes by injection of pure-grade helium. The samples were then converted to CO2 by injection of 100% phosphoric acid, which was left to react for 3 to 18 hours at 72ºC (bulk samples) or for approximately 72 hours at 25 ºC (serial samples; modified from McCrea, 1950). Then, the carbon and oxygen isotope ratios were measured by mass spectrometry. Ten aliquots of CO2 from each sample vial were measured and run against three aliquots of a CO2 reference gas. The δ13C and δ18O values reported are the average values for the ten sample aliquots. External errors were <0.06 ‰ for δ13C and <0.08 ‰ for δ18O. The expected internal error was <0.05 ‰ for δ18O 13 18 (CO2 reference gas; Thermo-Finnigan, 2002). The δ C and δ O values were calibrated by concurrent carbon and oxygen isotopic measurements of least two sets of three or more of the following carbonate standards: PDA, NBS-19, YW-CC-ST-1, ROY-CC, and MERK. Samples were analyzed on a Delta Plus XP IRMS at Florida State University or on a VG Prism in the Stable Isotope Lab at the University of Florida. To analyze phosphate samples, ~ 200 to 300 μg of phosphate standards and tooth enamel phosphate samples in the form of Ag3PO4 were measured into silver cups and then loaded into a Finnigan Thermal Conversion Elemental Analyzer (TC/EA) connected to a Delta Plus XP continuous flow isotope ratio mass spectrometer by a Conflo II open-split interface. The TC/EA uses high temperature conversion to convert oxygen that is present in a compound to CO. One aliquot of CO was analyzed for each sample and 2 aliquots of a CO reference gas were analyzed immediately before and after each sample. All samples and standards were run in triplicate to ensure that the TC/EA produced no memory effect. The expected internal error was <0.1 ‰ for δ18O (CO reference gas; Thermo-Finnigan, 2001). The expected external precision for the

analyses was <0.4 ‰ for δ18O (200 μg benzoic acid). The δ18O values were calibrated by concurrent oxygen isotopic measurements of least three of the following phosphate standards:

UMS-1, NIST-120c, NBS-120a, and KH2PO4. All phosphate samples were analyzed at Florida State University. To analyze the plant samples, ~ 2 to 3 mg of plant tissues were measured into tin cups and loaded into a Carlo Erba elemental analyzer (EA) connected to a Delta Plus XP continuous flow isotope ratio mass spectrometer by a Conflo II open-split interface. The carbon isotope compositions of the organic samples were determined by mass spectrometric measurement of

CO2 produced from combustion of the sample in the EA. Two aliquots of CO2 reference gas were analyzed immediately before and after each sample. The expected external precision for the analyses was <0.15 ‰ for δ13C (20 μg carbon). The δ13C values were calibrated by concurrent carbon isotopic measurements of at least two sets of three or more of the following organic standards: YWOMST-1 (sugar), YWOMST-2 (phenylalanine), YWOMST-3 (L-phenylalanine), YWOMST-4 (Costech acetamilide), and YWOMST-5 (urea).

3.4. Results and Discussion

3.4.1. Assessment of Fossil Tooth Preservation Tooth enamel is resistant to diagenetic alteration and, therefore, is well-suited to preserve the original isotopic composition of an animal. This is a result of a large crystal size relative to other biogenic materials, such as bone or dentine, and a low (~1 %) porosity that allows very little influx of diagenetic fluids (Ayliffe et al., 1994; Wang and Cerling, 1994). Tooth enamel has an inorganic component (>96% by weight) mainly in the form of hydroxyapatite crystals 2- (Ca10(PO4)6(OH)2), which contain ~1-3 % structural carbonate (CO3 ) substituting for phosphate 3- - (PO4 ) and hydroxyl (OH ) ions (Wang and Cerling, 1994). It has been demonstrated that an ~8- 9 ‰ fractionation occurs between coexisting phosphate and structural carbonate in enamel hydroxyapatite of extant mammals (Bryant et al., 1996; Iacumin et al., 1996). Because carbonates in fossil tooth enamel are more susceptible to dissolution and recrystallization processes during diagenesis than are phosphates, it is generally believed that phosphates are more likely to retain original oxygen isotope compositions (Kolodny and Luz, 1991; Ayliffe et 18 3- al., 1994). Even so, evidence suggests that microbial activity can modify δ O(PO4 ) values of

biogenic apatite during early diagenesis (Ayliffe et al., 1994; Kolodny et al., 1996; Sharp et al., 18 2- 18 3- 2000). Therefore, comparison of δ O(CO3 ) with δ O(PO4 ) values of fossil tooth enamel can be useful in evaluating the preservation of the original oxygen isotopic compositions of the structural carbonate and phosphate components of enamel hydroxyapatite (Iacumin et al., 1996; Fricke et al., 1998; Wang et al., 2008a). Ten fossil rhino enamel samples from the Linxia Basin showed a difference between 18 2- 18 3- 18 δ O(CO3 ) and δ O(PO4 ) values (Δ Oc-p) that ranged from 6.9 to 9.6 ‰. The average difference was 8.3 + 0.8 ‰, which is consistent with predicted values for the formation of coexisting structural carbonates and phosphates from the same body water in modern biogenic 18 apatite (Longinelli and Nuti, 1973; Bryant et al., 1996; Iacumin et al., 1996). The Δ Oc-p values of the fossil rhino enamel samples from this study are plotted against calculated linear 18 regressions from Δ Oc-p values of tooth enamel and bone from extant mammals (Bryant et al., 1996; Iacumin et al., 1996) and the shells of extant marine organisms (Longinelli and Nuti, 18 1973) in Figure 3.3. The Δ Oc-p values of the fossil rhino samples plotted on or near the equilibrium lines from previous studies, suggesting that there was little or no alteration of the oxygen isotopic compositions of either the structural carbonate or phosphate components of the fossil rhino samples. Comparison of the δ13C and δ18O values of fossil tooth enamel with those of carbonates within coexisting matrix provides another means of determining if significant alteration of the original carbon and oxygen isotopic compositions of the enamel has occurred, at least qualitatively. A significant difference between the isotopic compositions of the enamel and matrix carbonates would argue against considerable diagenetic alteration of the sample (e.g., Wang and Deng, 2005). Differences in δ13C and δ18O values between 32 fossil mammalian 13 18 enamel samples and their coexisting matrix carbonates (Δ Ce-m and Δ Oe-m) are plotted versus age in Figure 3.4. It has been suggested that the δ13C values of structural carbonate are much more resistant to diagenetic modification, even at high temperatures (120°C), than are δ18O 13 values (Wang and Cerling, 1994). Because the Δ Ce-m values at 1.2, 6, 7.5, and 11.5 Ma have 18 corresponding Δ Oe-m values that are large (>2 ‰), it is not likely that the small differences in δ13C values between the enamel and matrix at those ages are due to diagenesis. On the other hand, a few samples show small differences between enamel and matrix δ18O values at 0.05, 2.5, 7.5, and 9.5 Ma, which may suggest that some amount of diagenetic alteration has occurred at

25 SMOW

) 3

(CO O

18 20 δ

Longinelli & Nuti, 1973 15 Bryant et al., 1996 Iacumin et al., 1996 This study

0 5 10 15 20 25

18 δ O(PO4) SMOW

18 Figure 3.3. Δ Oc-p values of fossil rhino enamel from this study versus calculated linear 18 regressions from Δ Oc-p values of tooth enamel and bone from extant mammals (Bryant et al., 1996; Iacumin et al., 1996) and the shells of extant marine organisms (Longinelli and 18 Nuti, 1973). The Δ Oc-p values of the fossil rhino samples plotted on or near the equilibrium lines from previous studies, suggesting that there was little or no alteration of the oxygen isotopic compositions of either the structural carbonate or phosphate components of the fossil rhino samples.

0 1 2 3

5 6

(Ma) 7 8

Age ∆13-C (enamel-matrix) 9

∆18-O (enamel-matrix) 10 11 12 13

14 012345678910

Delta Values (‰, V-PDB)

Figure 3.4. Differences in δ13C and δ18O values between fossil mammalian enamel samples 13 18 13 and their coexisting matrix carbonates (Δ Ce-m and Δ Oe-m) versus age. Most Δ Ce-m and 18 Δ Oe-m values indicate a large difference (>2 ‰) between delta values of enamel and corresponding matrix carbonates, which suggests that minimal or no diagenetic alteration of the tooth enamel occurred for most fossil age groups.

13 18 those ages. Most Δ Ce-m and Δ Oe-m values indicate a large difference (>2 ‰) between delta values of enamel and corresponding matrix carbonates, suggesting that minimal or no diagenetic alteration of the tooth enamel occurred for most fossil age groups.

3.4.2. 25 Million Years of Climate Variability in the Linxia Basin The oxygen isotopic composition of tooth enamel from large-bodied mammalian herbivores largely reflects the isotopic composition of local meteoric water that is ingested by the herbivores, through drinking and plant consumption (Longinelli, 1984; Luz et al., 1984; Ayliffe and Chivas, 1990). In turn, the oxygen isotopic composition of meteoric water is controlled by climate (Dansgaard, 1964; Rozanski et al., 1992). As a result, the oxygen isotopic composition of mammalian tooth enamel can be a valuable tool for reconstruction of regional paleoclimates. A significant shift over time in the oxygen isotope compositions of tooth enamel of a given taxon, from a given region, indicates a change in regional climate (e.g., Ayliffe and Chivas 1990; Quade et al., 1992; Ayliffe et al., 1994; Bryant et al., 1994; Kohn and Cerling, 2002). At any given time interval, a large range of δ18O values for individuals of a given taxon may reflect differences in dietary behavior or local and seasonal variability in precipitation (MacFadden, 1998; Kohn and Cerling, 2002). Therefore, mean δ18O values can be used rather than individual data to reconstruct regional paleoclimates, as mean values attenuate δ18O variability caused by behavioral differences and/or local and seasonal weather changes (Kohn and Cerling, 2002; Wang and Deng, 2005). If, in fact, the mean δ18O values of tooth enamel reflect regional climate, the δ18O values of tooth enamel from different large-bodied mammalian taxa living in the same region should display similar shifts in their respective δ18O records. Comparing the oxygen isotopic compositions of horses and rhinos that lived concurrently in the Linxia Basin throughout the last 25 million years (Figure 3.5), it is demonstrated that the δ18O fluctuations throughout the Late Cenozoic are nearly identical for both taxa. The curves in Figure 3.5 represent calculated mean δ values of multiple horse and rhino tooth enamel samples at specific time intervals and new data were combined with that from Wang and Deng (2005) to provide improved time resolution as well as a more robust data set compared to the previous study. Two-tailed t-tests were performed in order to determine significant differences in mean oxygen isotopic compositions of various taxa at different ages (Table 3.1). Repeated fluctuations

Cooler Warmer 2 and/or and/or Wetter Drier Climate Climate 4

(Ma) 14

Age

20 (δ18O value of rhino from Lanzhou Basin, Horses d13C 22 adjacent to the Linxia Basin) Horses d18O Rhinos d13C 24 Rhinos d18O

-14-12-10-8-6-4-2 0

Delta Values (‰, V-PDB)

Figure 3.5. Carbon and oxygen isotopic compositions of horses and rhinos that lived concurrently in the Linxia Basin throughout the last 25 million years. Negative shifts in δ18O values indicate changes to cooler and/or wetter climates and positive shifts indicate changes to warmer and/or drier climates. δ13C values indicate diet composition (i.e., C3 or C4 plants) and shift to more positive values with water stress and more negative values in forested or closed- canopy environments. Error bars on δ18O curves represent 1 standard deviation from the mean.

Table 3.1. Statistical significance of differences in mean oxygen isotopic compositions of various taxa at different ages as determined by two-tailed t-tests.

Mean Significant Difference Difference Sample (‰) df t p at 95% CI? 18 18 Horse δ O (11.5 Ma) vs Horse δ O (9.5 Ma) 4.9 13 4.97 0.0003 Yes Horse δ18O (9.5 Ma) vs Horse δ18O (9 Ma) 1.7 7 -1.63 0.1468 No 18 18 Horse δ O (9.5 Ma) vs Horse δ O (7.5 Ma) 3.5 14 -5.69 0.0001 Yes Horse δ18O (9 Ma) vs Horse δ18O (7.5 Ma) 1.8 11 -1.90 0.0839 No 18 18 Horse δ O (6 Ma) vs Horse δ O (4 Ma) 1.6 14 3.16 0.0069 Yes 18 18 Horse δ O (6 Ma) vs Horse δ O (2.5 Ma) 3.3 19 4.38 0.0003 Yes 18 18 Horse δ O (4 Ma) vs Horse δ O (2.5 Ma) 1.7 15 1.70 0.1105 No 18 18 Horse δ O (2.5 Ma) vs Horse δ O (<2.5 Ma) 1.6 11 -1.01 0.3356 No 18 18 Rhino δ O (25 Ma) vs Rhino δ O (17 Ma) 1.6 10 -1.29 0.2261 No Rhino δ18O (17 Ma) vs Rhino δ18O (13 Ma) 0.7 6 1.21 0.2705 No Rhino δ18O (13 Ma) vs Rhino δ18O (11.5 Ma) 4.4 8 -10.51 <0.0001 Yes 18 18 Rhino δ O (11.5 Ma) vs Rhino δ O (9.5 Ma) 5.2 8 9.97 <0.0001 Yes 18 18 Rhino δ O (9.5 Ma) vs Rhino δ O (9 Ma) 2.8 6 -2.90 0.0274 Yes Rhino δ18O (9 Ma) vs Rhino δ18O (7.5 Ma) 1.9 7 2.74 0.0291 Yes Rhino δ18O (7.5 Ma) vs Rhino δ18O (7 Ma) 1.0 6 1.40 0.2118 No 18 18 Rhino δ O (7 Ma) vs Rhino δ O (6 Ma) 4.3 6 -3.96 0.0075 Yes Bovid δ18O (7 Ma) vs Bovid δ18O (4 Ma) 1.1 5 -0.59 0.5824 No Bovid δ18O (4 Ma) vs Bovid δ18O (2.5 Ma) 1.0 7 -1.20 0.2677 No Bovid δ18O (2.5 Ma) vs Bovid δ18O (0 Ma) 1.3 4 1.85 0.1373 No Deer δ18O (7 Ma) vs Deer δ18O (4 Ma) 6.4 2 -5.01 0.0376 Yes

in horse and rhino tooth enamel δ18O values occur throughout the sedimentary sequence in the Linxia Basin, which suggests that the climate was unstable from 25 to 0.05 Ma. Because the fluctuations in both horse and rhino δ18O values closely follow one another, this data strongly supports that the oxygen isotopic composition of mammalian herbivore tooth enamel largely reflects regional climate. As shown in Figure 3.5 and Table 3.1, the δ18O values of horse tooth enamel show significant shifts to more negative values after ~11.5 and ~6 Ma (p = 0.0003), indicating changes in regional climate toward either cooler temperatures or less arid conditions, or both. A large negative shift in the δ18O values of horse enamel also occurs after ~7.5 Ma, but more data is needed to determine statistical significance. Likewise, significant shifts to more negative values occur in the δ18O values of rhino enamel after ~11.5 (p = <0.0001) and ~9 Ma (p = 0.0291). Based on deep-sea δ18O and Mg/Ca records, a general cooling trend with two main cooling phases have occurred over the last 25 million years (Shackleton and Kennett, 1975; Shackleton et al., 1995; Lear et al., 2000; Zachos et al., 2001). The first significant cooling event related to the rapid expansion of the east Antarctic ice sheet occurred during the Middle Miocene from ~15 to 10 Ma (Shackleton & Kennett, 1975; Zachos et al., 2001). The negative shift in horse and rhino δ18O after 11.5 Ma indicates a shift to cooler temperatures and is roughly consistent with cooling due to the establishment of a major ice sheet by 10 Ma. The second cooling phase occurred during the Plio-Pleistocene with a small-scale expansion of the west Antarctic ice sheet followed by the onset of the Northern Hemisphere Glaciation (Shackleton and Kennett, 1975; Lear et al., 2000; Zachos et al., 2001). The horse and rhino δ18O values show a negative shift from 6 to 2.5 Ma, which is, again, in general agreement with the deep-sea records. Significant positive shifts in δ18O values after ~13 Ma in rhinos (p = <0.0001), ~9.5 Ma in horses (p = 0.0001) and rhinos (p = 0.0274), and ~7 Ma in rhinos (p = 0.0075) (Figure 3.5), indicate shifts to drier and/or warmer conditions. δ18O values of deer tooth enamel from the Linxia Basin also suggest a significant positive shift in δ18O values after ~7 Ma. The deep-sea δ18O and Mg/Ca records show a general warming trend from ~26 to 15 Ma (interrupted by an abrupt return to a cooler climatic state at ~23 Ma – the so-called “Early Miocene Mi-1 glaciation”), with a climatic optimum occurring at 17 to 15 Ma (Zachos et al., 2001). Rhino δ18O values show a warming trend from the Late Oligocene to the mid-Miocene that broadly agrees with deep-sea records. Dettman et al. (2003) suggested, based on δ18O values of fluvial and

lacustrine carbonates, that the Linxia Basin shifted to more arid and/or warmer conditions at 12 Ma. The positive shift observed in both horse and rhino δ18O values from 13 to 11.5 Ma also indicates drier conditions and/or increased temperatures. The positive shift in δ18O values of horses, rhinos and deer from the Linxia Basin after 7 Ma is similar in timing to a positive δ18O shift observed in fossils and paleosols from Pakistan and Nepal (Quade et al., 1989; Quade et al., 1992; Quade et al., 1995). This suggests a shift toward a drier and/or warmer climate on both the north and south sides of the Tibetan Plateau during the Late Miocene. Bovid samples from the Linxia Basin do not show significant shifts in mean δ18O values after 4 or 2.5 Ma, which is consistent with shifts in δ18O values of horses at those ages (Table 3.1). Individual δ18O values of fossil tooth enamel from different taxonomic groups from the Linxia Basin, including bovids, horses, rhinos, deer, giraffes, pigs, and elephants, are plotted versus age in Figure 3.6. All bovid, deer, giraffe, pig, and elephant data points fall within or close to (<2 ‰ difference) the ranges of horse and rhino δ18O values at any given age, with the exception of a single deer sample at 4 Ma, which is >4 ‰ more positive than any horse or rhino individual at that age. This difference could be the result of either diagenetic alteration or differences in dietary or drinking behavior between the deer at that age and all other taxa. A difference in dietary behavior is likely, as the deer δ18O values are again more positive than those of all other taxa at 14 and 7 Ma. Deer ingest a large amount of water from leaves and leaf water is strongly affected by relative humidity (Cormie et al., 1994; Koch, 1998). As a result, leaf δ18O values will increase with increases in aridity. Overall, the oxygen isotopic compositions of all mammalian taxa collected from the Linxia Basin for this study are consistent with the positive and negative trends in horse and rhino mean δ18O values shown in Figure 3.5, reflecting changes in regional climate. No fossil mammal samples between the ages of 25 and 17 Ma have been recovered from the Linxia Basin and, therefore, it was not possible to determine whether a negative shift in δ18O values occurred as a result of the Early Miocene Mi-1 glaciation based on the oxygen isotope compositions of fauna from the Linxia Basin. Interestingly, a single 20.4 Ma rhino sample from the Zhangjiaping locality (103.6°E, 36.3°N; elevation = 1673 m) within the Lanzhou Basin had a δ18O value of -12.5 ‰, which is 2.8 ‰ more negative than the mean δ18O value of 25 Ma rhinos

(Ma) 14 Age Bovids 16 Horses

18 Rhinos

Deer 20 Giraffes 22 Pigs Elephants 24

26 -14-13-12-11-10-9-8-7-6-5-4-3-2-10 δ18O (‰, V-PDB)

Figure 3.6. Individual δ18O values of fossil tooth enamel from different taxonomic groups from the Linxia Basin versus age. All bovid, deer, giraffe, pig, and elephant data points fall within or close to (<2 ‰ difference) the ranges of horse and rhino δ18O values at any given age, with the exception of a single deer sample at 4 Ma, which is >4 ‰ more positive than any horse or rhino individual at that age.

from the Linxia Basin (103°E, 35°N; elevation = 1917 m). The Zhangjiaping locality is situated approximately 30 miles northwest of the provincial capital, Lanzhou, and is ~130 km north of the Linxia Basin (Figure 3.1). There are two possible explanations for this apparent negative shift in δ18O: (1) it may truly represent a temporal shift in water δ18O in the region if the geographic difference in water δ18O between these two basins is small (<1-2 ‰); and (2) the apparent shift reflects spatial variations in water δ18O rather than temporal variation in climate. Here, it is argued that it may be possible to approximate paleoclimate fluctuations in the region of the Linxia Basin based on isotopic compositions of fossil mammals from the Lanzhou Basin as a result of their close proximity, as the difference in water δ18O between these two basins is likely to be small. For example, 9 Ma rhino (Chilotherium wimani) samples from the Tianshui locality (105.8°E, 34.6°N; elevation = ~1100 m), which is located approximately 300 km southeast of the Linxia Basin (Figure 3.1), have a mean δ18O value of -6.7 + SD 0.3 ‰, which is very similar to the mean δ18O value of 9 Ma Chilotherium wimani from the Linxia Basin, -7.3 + SD 2.1 ‰. This data supports that the oxygen isotopic compositions of meteoric water do not vary by large amounts within the geographic region. Therefore, the apparent negative shift in δ18O values at 20.4 Ma is consistent with cooling due to the Mi-1 glaciation during the Late Miocene. Carbon isotope compositions of mammalian herbivore tooth enamel serve as proxies for the diets of animals, as most carbon isotopic variation in enamel is a result of isotope fractionation that occurred during photosynthesis of plants that were subsequently ingested. C3 plants, including trees, shrubs, forbs, and cool season grasses, use the Calvin Cycle photosynthetic pathway and have an average δ13C value of -27 ‰ (δ13C range = -34 to -22 ‰). C4 plants, including warm season grasses, a few shrubs, corn, sugar cane, sorghum, and bamboo, use the Hatch-Slack photosynthetic pathway and have an average δ13C value of -13 ‰ (δ13C range = -17 to -9 ‰). Because tooth enamel carbonate is 13C-enriched by ~14 ‰ relative to diet, tooth enamel δ13C values that are < -8 ‰ usually indicate a pure C3 diet and δ13C values that are > -2 ‰ indicate a pure C4 diet (Lee-Thorp and van der Merwe, 1987; O’Leary, 1988; Farquhar et al., 1989; Koch, 1998; Cerling et al., 1997a; Cerling and Harris, 1999; Kohn and Cerling, 2002; Wang and Deng, 2005). If plants become water-stressed, δ13C values increase (Farquhar et al., 1989), and thus the conservative “cut-off” enamel-δ13C value for a pure C3 diet in water- stressed environments should be -8 ‰ for modern herbivores and -7 ‰ for fossil herbivores due

13 to changes in the δ C value of atmospheric CO2 (Cerling et al., 1997a; Wang et al., 2008b). In closed environments, such as forests, δ13C values decrease, due to the influence of soil respiration (Schleser and Jayasekera, 1985; Sternberg et al., 1989; van der Merwe and Medina, 1989). Therefore, temporal shifts in the δ13C values for horses or rhinos from the Linxia Basin may indicate either a shift in diet, habitat, or regional climatic conditions. Figure 3.5 shows that the mean δ13C values for horses and rhinos at nearly all time intervals are very similar and range from -8.2 to -12.4 ‰, which suggests that both taxa had pure C3 diets throughout most of the Late Cenozoic. Notably, at 1.2 Ma, the horse δ13C values increase to ~-5 ‰, indicating a change to a mixed C3/C4 diet after 2.5 Ma. Because all δ13C values are greater than -13 ‰, a relatively open environment, such as a savannah or mixed woodland/steppe biome is indicated at all time intervals. Most positive and negative shifts in mean δ13C values occur at the same time and in the same direction as the shifts in δ18O values, indicating that the variation in δ13C values is probably a result of fluctuations in water availability. As a result, this suggests that the variation in δ18O values most likely reflect changes in humidity or aridity, as opposed to changes in temperature. There is an exception at 11.5 Ma for both horses and rhinos, where δ13C values decrease as δ18O values increase. This suggests that the plants were not water-stressed at this time interval and, therefore, the positive shift in δ18O values may be due to warmer temperatures rather than increased aridity. Individual δ13C values of fossil tooth enamel from different taxonomic groups from the Linxia Basin are plotted versus age in Figure 3.7. All bovid, deer, giraffe, pig, and elephant data points fall within or near the ranges of horse and rhino δ13C values at any given age. The δ13C values of most individuals are lower than -8 ‰, indicating pure C3 diets for those animals. A single horse at 1.2 Ma, as mentioned above, had a δ13C value of -5.04 ‰, indicating a mixed C3/C4 diet. Additionally, a single bovid sample at 7.5 Ma had a δ13C value of -7.12 ‰, which suggests that this individual may have consumed a small amount of C4 plants. This would be consistent with a Late Miocene C4 expansion, but because there is no indication of mixed C3/C4 diets for any other individual (including horses) at 7.5 Ma, it is likely that very little, if any, C4 vegetation was present in the Linxia Basin at that time. Alternatively, this δ13C value could indicate that this individual was feeding on plants experiencing water stress (Cerling et al., 1997a; Wang et al., 2008b). In general, the carbon isotopic compositions of all mammalian taxa

12 (Ma)

Age Bovids 16 Horses 18 Rhinos Deer 20 Giraffes

22 Pigs Elephants 24

26 -14-13-12-11-10-9-8-7-6-5-4-3-2-10 δ13C (‰, V-PDB)

Figure 3.7. Individual δ13C values of fossil tooth enamel from different taxonomic groups from the Linxia Basin versus age. All bovid, deer, giraffe, pig, and elephant data points fall within or near the ranges of horse and rhino δ13C values at any given age. The δ13C values of most individuals are lower than -8 ‰, indicating pure C3 diets for those animals. A single horse at 1.2 Ma had a δ13C value of -5.04 ‰, indicating a mixed C3/C4 diet. A single bovid sample at 7.5 Ma had a δ13C value of -7.12 ‰, which suggests that this individual may have also had a mixed C3/C4 diet.

collected from the Linxia Basin for this study are consistent with the horse and rhino δ13C trends shown in Figure 3.5.

3.4.3. Changes in Seasonality and Monsoon Strength Seasonal patterns in diet and climate can be reconstructed from serial carbon and oxygen isotopic analyses of tooth enamel, respectively (e.g., Koch et al., 1995; Fricke and O’Neil, 1996; Sharp and Cerling, 1998). More specifically, the seasonality of precipitation and/or ambient temperature can be inferred from the stable oxygen isotope compositions of serial samples from mammalian herbivore tooth enamel and changes in the composition of an animal’s diet or in its feeding pattern can be determined from serial stable carbon isotope analyses. It has been shown that intra-tooth variation in the carbon and oxygen isotopic compositions of carbonate in modern and fossil tooth enamel represents seasonal fluctuations in diet and in regional climate experienced by an animal (Koch et al., 1989; Bryant et al., 1994, 1996; Fricke and O’Neil, 1996; Sharp and Cerling, 1998; Feranec and MacFadden, 2000). The intra-tooth variation in the carbon and oxygen isotopic compositions of fossil mammals from the Linxia Basin may offer clues to the behaviors of various fossil taxa and may provide evidence for the timing of the development of the Asian monsoon in this region. The uplift of the Tibetan Plateau during the Cenozoic Era is considered to be the driving force in the development of the Asian monsoons (Hahn and Manabe, 1975; Ye, 1981; Murakami, 1987; Kutzbach et al., 1993; Yanai and Li, 1994; An et al., 2001; Kitoh, 2004), and the timing history of this development is important to the understanding of mammalian evolution in China. The Asian monsoon is caused by the strongly contrasting heat capacities of the ocean and continent. The high elevation difference between the Tibetan Plateau and the ocean results in strong temperature contrasts of air masses, creating a pressure gradient that drives warm, moist winds from the ocean to over the land in the boreal summer (Summer Monsoon). As a result, large amounts of precipitation are deposited over the Asian continent during the summer monsoon season. In the winter, cold and dry winds move from Siberia southward across East Asia toward the ocean (Winter Monsoon), resulting in relatively dry winters. This leads to distinct seasonal differences in precipitation on the Asian continent (Webster, 1987; Clark et al., 2000). The oxygen isotope ratios of meteoric water vary from region to region as a result of evaporative and condensative processes. In these processes, evaporation favors water molecules

containing the lighter isotope of oxygen, 16O, and condensation favors water molecules with the heavier oxygen isotope, 18O (Epstein and Mayeda, 1953; Craig, 1961). In temperate zones that lie outside of monsoon regions and have relatively small amounts of annual precipitation, as well as large fluctuations in monthly temperatures, temporal variation in δ18O values largely reflects changes in ambient temperature as a result of evaporative processes. In this type of environment, δ18O values increase during summer months and decrease during winter months. In monsoonal regions, where there is seasonal fluctuation in temperature and in precipitation, temporal variation in δ18O values mainly reflects changes in rainfall amounts as a result of condensative processes, so called the “Amount Effect” (Dansgaard, 1964). Because precipitation favors 18O, the more it rains, the more 18O-depleted the atmospheric water vapor becomes. The precipitation that forms from the 18O-depleted water vapor then, likewise, becomes 18O-depleted. In a monsoonal environment, δ18O values decrease during summer months, even with higher ambient temperatures, and increase during winter months (Dansgaard, 1964; Gonfiantini, 1985; Rozanski et al., 1993; Araguas-Araguas et al., 1998). Weighted monthly mean δ18O values of precipitation at selected International Atomic Energy Agency (IAEA) stations (Figure 3.8) both within and outside of the summer monsoon region in China are shown in Figures 3.9 and 3.10. At the six localities outside the summer monsoon region, the mean δ18O values of precipitation are highest during the summer months (June-August) and are lowest during the winter months (December-February), with up to an 18 ‰ difference between summer and winter δ18O values. At the six localities within the summer monsoon region, the mean δ18O values of precipitation are lowest during the summer months (June-August), with up to a 16 ‰ difference between highest and lowest monthly δ18O values. These data are consistent with expected oxygen isotope compositions of precipitation in temperate regions outside monsoonal regions and those of precipitation within monsoonal regions. Because the δ18O of tooth enamel is a proxy for regional climate, serial isotopic analyses of fossil tooth enamel should show the same patterns observed in the IAEA precipitation data. That is, in regions that are not strongly influenced by monsoons, low amounts of precipitation and warmer temperatures should result in increased δ18O values during the summer months and within monsoon regions, high amounts of summer precipitation should result in decreased δ18O

Present-day monsoon limit ???

Figure 3.8. Location of selected International Atomic Energy Agency (IAEA) stations within and outside of the summer monsoon region in China. The dotted line is a rough estimation of the modern monsoon limit based on the δ18O records of precipitation from the IAEA stations.

Summer Monsoon Region 0

-2

-4

-6 O (‰) -8 18 δ -10 Changsha -12 Chengdu -14 -16 Kunming -18 Lhasa -20 Nanjing

Mean Weighted -22 Yantai -24 -26 123456789101112

Month

Figure 3.9. Weighted monthly mean δ18O values of precipitation at selected International Atomic Energy Agency (IAEA) stations within the summer monsoon region in China. At the six localities within the summer monsoon region, the mean δ18O values of precipitation are lowest during the summer months (June-August), with up to a 16 ‰ difference between highest and lowest monthly δ18O values. Shaded area represents summer months.

Outside Summer Monsoon Region 0 -2 -4 -6

O (‰) -8 18

δ -10 Baotou -12

-14 Hetian

-16 Lanzhou

-18 Qiqihar -20 Wulumuqi

Weighted Mean Mean Weighted -22 Zhangye -24 -26 123456789101112 Month

Figure 3.10. Weighted monthly mean δ18O values of precipitation at selected International Atomic Energy Agency (IAEA) stations outside of the summer monsoon region in China. At the six localities outside the summer monsoon region, the mean δ18O values of precipitation are highest during the summer months (June-August) and are lowest during the winter months (December-February), with up to an 18 ‰ difference between summer and winter δ18O values. Shaded area represents summer months.

values during the warmer months. Therefore, depending on the type of climate experienced by a fossil animal, either peaks or troughs in the serial data could reflect the summer months. As a result, it is necessary to determine whether the highest or the lowest δ18O values within the serial isotope records of given individuals occurred during the summer months in order to differentiate between mammals that lived within and those that dwelled outside of monsoon regions. It may be possible to distinguish between summer and winter δ18O values by looking at concurrent carbon isotope records (Figure 3.11). Because C3 plants are dominant year-round in regions with little summer precipitation, little or no change in δ13C values occur throughout the year. In this case, we may assume that the peaks, or increased δ18O values occur during the summer due to little precipitation and warmer temperatures and that this isotopic pattern reflects regions that are not strongly affected by the summer monsoon. On the other hand, because C4 grasses are able to grow in regions with summer precipitation and warm temperatures, an increase in δ13C values occurs in enamel formed during summer months when C4 grasses are present. Here, we may assume that the troughs or decreased δ18O values occur during the summer months as a result of high amounts of precipitation and that this pattern is reflective of a summer monsoon region. Thus, an inverse relationship exists between δ18O and δ13C values in a summer monsoon region. In order to investigate changes in seasonality in the Linxia Basin from the Miocene to the present and to determine the timing of the strengthening of the summer monsoon in northwestern China, serial tooth enamel samples from 12 fossil horses dated 11.5, 9.5, 7.5, 6.0, 4.0, 2.5, 1.2, and 0.05 Ma were analyzed, as were those from two 17 Ma and two 13 Ma elephants, two 4 Ma giraffes, and three 2.5 Ma and two modern bovids. The serial carbon and oxygen isotopic data for all fossil mammals in this study are presented in Figures 3.12a, 3.12b, 3.13 and 3.14. The results showed that, in general, positive δ18O shifts in the horse and rhino bulk data, indicating shifts to either drier and/or warmer conditions after 14, 9.5, 7.5, and 2.5 Ma, were accompanied by increased seasonality, as inferred from the relatively greater δ18O ranges in the serial data (Figure 3.15). Likewise, negative δ18O shifts in the bulk data at 11.5, 6.0, 4.0, and 1.2 Ma were associated with decreases in seasonality, or relatively smaller δ18O ranges in the serial data. The oxygen isotopic compositions of water sources have a great influence on seasonal δ18O signals in mammalian tooth enamel. It has been shown that precipitation displays much larger seasonal δ18O variability than groundwater, as the oxygen isotopic composition of

SUMMERTIME SUMMERTIME Outside Summer Within Summer Monsoon Region Monsoon Region

18 δ18O δ O

High amount of precipitation results in decrease Little precipitation and warmer temperatures of δ18O values during summer result in increase of δ18O values during summer

δ13C

C3 plants are dominant year-round in regions C4 grasses are able to grow in regions with with little summer precipitation; results in little or summer precipitation; results in increase of δ13C no change in δ13C values throughout year values during summer

Figure 3.11. Expected patterns in serial δ18O and δ13C records for tooth enamel samples from localities within and outside of the summer monsoon region in China. Because tooth enamel is a proxy for regional climate, serial isotopic analyses of fossil tooth enamel should show the same patterns observed in the IAEA precipitation data. That is, in regions that are not strongly influenced by monsoons, low amounts of precipitation and warmer temperatures should result in increased δ18O values during the summer months and within monsoon regions, high amounts of summer precipitation should result in decreased δ18O values during the warmer months.

Hipparion dermatorhinum (Ds-05) 6 Ma Hipparion dermatorhinum (Ds‐06) 6 Ma 0 0 Δ18O = 3.7 -2 -2 Δ13C = 0.5 Δ18O = 6.9 -4 -4 Δ13C = 0.6 -6 -6 -8 -8

-10 -10 -12 -12 -14 -14 0 5 10 15 20 25 30 35 40 45 50 55 0 5 10 15 20 25 30 35 40 45 50 55

Hipparion sp. (Hl-03) 7.5 Ma Hipparion sp. (Hl-10) 7.5 Ma 0 0 -2 Δ18O = 3.5 -2 Δ18O = 1.0 -4 Δ13C = 0.2 -4 Δ13C = 0.6

-6 -6 -8 -8 -10 -10 -12 -12 -14 -14 0 5 10 15 20 25 30 35 40 45 50 55 0 5 10 15 20 25 30 35 40 45 50 55

Hipparion dermatorhinum (Qj-06) 9.5 Ma 0

-2 18 Δ O = 3.5 -4 Δ13C = 1.1 18 V-PDB) Delta Values (‰, -6 δ O 13 -8 δ C -10 -12 -14 0 5 10 15 20 25 30 35 40 45 50 55

Hipparion dongxiangense (Gn-07) 11.5 Ma Hipparion dongxiangense (Gn-10) 11.5 Ma 0 0 -2 -2 -4 -4 -6 Δ18O = 5.0 -6 Δ18O = 4.2 -8 Δ13C = 1.7 -8 Δ13C = 0.9 -10 -10 -12 -12 -14 -14 0 5 10 15 20 25 30 35 40 45 50 55 0 5 10 15 20 25 30 35 40 45 50 55 Distance from Crown (mm)

Figure 3.12a. Serial carbon and oxygen isotope compositions of horse tooth enamel from 11.5 to 6 Ma. Open diamonds represent oxygen isotope compositions and closed diamonds represent carbon isotope compositions. The data labels in parentheses are the sample numbers of the horse individuals. Serial samples decrease in biological age with distance from the crown. Δ18O and Δ13C = the ranges of the δ18O and δ13C seasonal cycles.

Equus hemonius (By-05) 0.05 Ma Equus qingyangensis (Tz-02) 1.2 Ma 0 0 -2 Δ18O = 1.2 -2 Δ18O = 7.6 -4 Δ13C = 2.1 -4 Δ13C = 4.7 -6 -6 -8 -8 -10 -10

-12 -12 -14 -14 0 5 10 15 20 25 30 35 40 45 50 55 0 5 10 15 20 25 30 35 40 45 50 55

Equus sp. (Ld-10) 2.5 Ma Equus sp. (Ld-11) 2.5 Ma 0 0 -2 Δ18O = 2.6 -2 Δ18O = 2.1 13 13 -4 Δ C = 1.1 -4 Δ C = 1.1 -6 -6 -8 -8 -10 -10 -12 -12 -14 -14 0 5 10 15 20 25 30 35 40 45 50 55 0 5 10 15 20 25 30 35 40 45 50 55 Delta Values (‰, V-PDB) Hipparion sp. (Shl-1) 4 Ma 0 Δ18O = 3.3 -2 Δ13C = 0.7 -4 δ18O -6 13 -8 δ C -10 -12 -14 0 5 10 15 20 25 30 35 40 45 50 55

Distance from Crown (mm)

Figure 3.12b. Serial carbon and oxygen isotope compositions of horse tooth enamel from 4 to 0.05 Ma. Open diamonds represent oxygen isotope compositions and closed diamonds represent carbon isotope compositions. The data labels in parentheses are the sample numbers of the horse individuals. Serial samples decrease in biological age with distance from the crown. Δ18O and Δ13C = the ranges of the δ18O and δ13C seasonal cycles. Shaded areas represent inferred summer months.

Palaeotragus microdon (Sl-09) 4 Ma Palaeotragus microdon (Sl-10) 4 Ma 0 0 -2 -2 Δ18O = 4.5 Δ18O = 5.2 -4 Δ13C = 2.3 -4 Δ13C = 2.3 -6 -6 -8 -8 -10 -10 -12 -12 -14 -14 0 5 10 15 20 25 30 35 40 45 50 55 0 5 10 15 20 25 30 35 40 45 50 55

Platybelodon grangeri (Lg-08) 13 Ma Platybelodon grangeri (Lg-10)13 Ma 0 0

-2 Δ18O = 2.4 -2 Δ18O = 2.9 -4 Δ13C = 1.2 -4 Δ13C = 0.4 -6 -6 -8 -8 -10 -10 -12 -12 -14 -14 0 5 10 15 20 25 30 35 40 45 50 55 0 5 10 15 20 25 30 35 40 45 50 55 Delta Values (‰, V-PDB) Values Delta

Gomphotherium (Dl-04) 17 Ma Gomphotherium (Dl-07) 17 Ma 0 0 -2 Δ18O = 2.5 -2 Δ18O = 2.8 -4 Δ13C = 1.1 -4 Δ13C = 0.7 -6 -6 -8 -8 -10 -10 -12 -12 -14 -14

0 5 10 15 20 25 30 35 40 45 50 55 0 5 10 15 20 25 30 35 40 45 50 55

Distance from Crown (mm) δ18O δ13C

Figure 3.13. Serial carbon and oxygen isotope compositions of elephant (Gomphotherium and Platybelodon) and giraffe (Palaeotragus) tooth enamel from 17 to 4 Ma. Open diamonds represent oxygen isotope compositions and closed diamonds represent carbon isotope compositions. The data labels in parentheses are the sample numbers of the individuals. Serial samples decrease in biological age with distance from the crown. Δ18O and Δ13C = the ranges of the δ18O and δ13C seasonal cycles.

Modern cow (Lb-01) Modern goat (Lb-02) 0 0 -2 Δ18O = 2.8 -2 Δ18O = 6.8 13 -4 Δ C = 1.6 -4 Δ13C = 2.0 -6 -6

-8 -8 -10 -10 -12 -12 -14 -14 0 5 10 15 20 25 30 35 40 45 50 55 0 5 10 15 20 25 30 35 40 45 50 55

Gazella blacki (Ld-07) 2.5 Ma 0

-2 ? Δ18O = 1.6 -4 Δ13C = 1.5 18 -6 δ O 13 -8 δ C -10 -12 -14 0 5 10 15 20 25 30 35 40 45 50 55

Delta Values (‰, V-PDB) Leptobus amplifrontalis (Ld-08) 2.5 Ma Leptobus amplifrontalis (Ld-09) 2.5 Ma 0 0

-2 Δ18O = 3.7 -2 Δ18O = 2.3 -4 Δ13C = 2.2 -4 Δ13C = 0.6 -6 -6 -8 -8 -10 -10 -12 -12 -14 -14 0 5 10 15 20 25 30 35 40 45 50 55 0 5 10 15 20 25 30 35 40 45 50 55 Distance from Crown (mm)

Figure 3.14. Serial carbon and oxygen isotope compositions of bovid tooth enamel from 2.5 Ma to the present. Open diamonds represent oxygen isotope compositions and closed diamonds represent carbon isotope compositions. The data labels in parentheses are the sample numbers of the bovid individuals. Serial samples decrease in biological age with distance from the crown. Δ18O and Δ13C = the ranges of the δ18O and δ13C seasonal cycles. Shaded areas represent inferred summer months.

(Ma) 8

Age 10 Horse mean bulk d18O 12 values Rhino mean bulk d18O 14 values

Serial sample d18O ranges 16

18 -12 -10 -8 -6 -4 -2 0 2 4 6 8 10

Delta Values (‰, V-PDB)

Figure 3.15. Horse and rhino mean bulk δ18O values versus δ18O ranges of all serial- sampled mammals of various ages. In general, positive δ18O shifts in the horse and rhino bulk data, indicating shifts to either drier and/or warmer conditions after 14, 9.5, 7.5, and 2.5 Ma, were accompanied by increased seasonality, as inferred from the relatively greater δ18O ranges in the serial data. Likewise, negative δ18O shifts in the bulk data at 11.5, 6.0, 4.0, and 1.2 Ma were associated with decreases in seasonality, or relatively smaller δ18O ranges in the serial data.

45 groundwater reflects a weighted mean annual δ18O of precipitation in the catchment area modified by evaporation. Therefore, the δ18O values of lakes and longstanding rivers that consist predominantly of groundwater reflect average annual isotope compositions, and ephemeral streams and small ponds that consist mainly of local rainwater and have short water residence times have δ18O values that more accurately reflect the oxygen isotope composition of seasonal precipitation (Koch et al., 1989; Clark and Fritz, 1997). Thus, tooth enamel from animals that drank from large bodies of water would be expected to show dampened seasonal signals compared to that from animals that drank from smaller and more temporary ponds or pools. Furthermore, consumption of plant material affects the oxygen isotope compositions of animals. Plants consist mainly of water (> 80%) and leaf water is generally 18O-enriched relative to local meteoric water. This is due to the preferential loss of isotopically light water molecules during evapotranspiration. The extent of 18O-enrichment is dependent on relative humidity (Dongmann et al., 1974; Epstein et al., 1977; Yakir, 1992). Consequently, leaf water tends to have more enriched δ18O values and displays a larger range of δ18O variation in open habitats compared to more closed forested habitats in a given geographic area. Thus, it would be expected that individuals living in more open habitats or those migrating between habitats would display greater δ18O variability in their enamel than would non-migratory forest-dwellers. Therefore, mammalian individuals with relatively small δ18O ranges may have received most of their water intake from longstanding lakes, springs, or rivers, which tend to dampen seasonal δ18O signals, or they may have lived in or migrated between more forested habitats. This is consistent with a cooler and/or wetter environment as inferred from the bulk oxygen isotope data. Conversely, individuals with relatively larger δ18O ranges may have drank from more transient water sources or lived in a more open environment, which would be consistent with higher aridity and/or warmer temperatures. In addition, a marked increase in seasonality (~5 ‰) after 2.5 Ma, as indicated by intra-tooth δ13C and δ18O variations in both horses and bovids, (Figure 3.15) is consistent with a strengthening of the summer monsoon in the region after ~2-3 Ma. The δ18O range of Equus hemonius at 0.05 Ma was not as large as that of Equus qingyangensis at 1.2 Ma (Figure 3.12b), which could be due to having an incomplete seasonal record for E. hemonius, as the tooth was very worn. This could also be the result of a temporary decrease in seasonality during the years recorded by the 0.05 Ma individual. It has been shown

that the Asian monsoon has distinct interannual variations and that a year of heavy rainfall is often followed by a year of diminished rainfall (Clark et al., 2000). For example, extensive snow cover over Eurasia following a strong monsoon season can slow down the summer heating of the landmass, which is necessary for the formation of large-scale monsoon flow in the following year (Shukla, 1987; Barnett et al., 1989). The δ18O range of the modern cow was, likewise, not as large as that of the modern goat (Figure 3.14). Again, the records of these two individuals may represent two different time spans and the difference in seasonality between the two animals may be due to interannual variation in monsoon strength. It is also possible that the O isotopic difference between the modern cow and goat is due to different drinking and dietary behaviors of these two animals. Most horses showed weak or no cyclicity in the serial oxygen isotope data (that is, no clear sinusoidal pattern), which suggests that the horses may have been migratory, and therefore, had mixed seasonal signals from drinking at multiple locations. The serial carbon isotope results showed that prior to 1.2 Ma, all mammals had pure or nearly pure C3 diets (all δ13C values < -8 ‰). A positive shift in δ13C values occurred in horses after 2.5 Ma, indicating a change to a mixed C3/C4 diet. This suggests that C4 grasses may have not spread into the basin until after 2.5 Ma, which is much later than the proposed global C4 expansion during the Late Miocene. This also indicates a strengthening of the Asian summer monsoon after 2.5 Ma, as C4 plants require sufficient summer precipitation. The δ13C ranges of all horses from the ages 11.5, 9.5, 7.5, 6.0, 4.0, and 2.5 Ma were < 1.1 ‰, except for one individual at 11.5 Ma that had a δ13C range of 1.7 ‰ (Figures 3.12a and 3.12b ). At 1.2 and 0.05 Ma, the δ13C ranges of two horses increased to 4.7 and 2.1 ‰, respectively (Figure 3.12b), further supporting seasonal changes in the composition of plant biomass in the Linxia Basin after 2.5 Ma. Most importantly, decreases in the δ18O values within individual teeth of the horses from 1.2 and 0.05 Ma occurred simultaneously with increases in the δ13C values. As mentioned previously, C4 grasses require summer precipitation and an animal that consumes C4 vegetation will have increased δ13C values in its enamel formed during summer months. Increased summer precipitation (i.e., summer monsoon) will result in a decrease in δ18O values during summer months. That is, in the Asian monsoon region, one would expect that higher δ13C values correlate with lower δ18O values (representing summer months) within an individual tooth (Figure 3.11). Thus, the anti-correlation between δ13C and δ18O values observed in teeth of the

1.2 Ma and 0.05 Ma horses (Figure 3.12b) strongly supports a monsoonal environment in the Linxia Basin after, but not prior to ~2.5 Ma. The δ13C ranges of the elephants Gomphotherium sp. at 17 Ma and Platybelodon grangeri at 13 Ma are < 1.2 ‰, with all δ13C values < -8 ‰ (Figure 3.13), indicating no seasonal variation in feeding behavior or composition of diet. Palaeotragus microdon at 4 Ma has a relatively larger δ13C range of 2.3 ‰ (Figure 3.13), but all δ13C values are < -8 ‰ and there is no inverse relationship between δ18O and δ13C values. This suggests that the greater variation in the serial δ13C values of P. microdon is likely due to seasonal changes in relative humidity rather than seasonal variation in the proportion of C3 and C4 plants consumed by the animal. Three bovid individuals from 2.5 Ma and two modern bovids have δ13C ranges of 0.6, 2.2, 1.5, 2.0, and 1.6 ‰, with almost all δ13C values < -10 ‰ (Figure 3.14). Although all five individuals appear to have been pure C3 feeders based on their bulk δ13C values, there is an anti-correlation between δ13C and δ18O values for all individuals. This suggests that these animals may have incorporated a small amount of C4 grass into their diets (with up to 14% C4, assuming that the end-member enamel-δ13C values for pure C3 and C4 diet are -12‰ and +2‰, respectively) during the summer monsoon season (Figure 3.11). This is quite possible, as the modern flora in the Linxia Basin includes C4 grasses, which account for ~13% of all species collected from the basin (Table 3.2). Carbon isotope analysis of modern soil carbonates has shown that C4 plants are unevenly distributed over the modern landscape and make up about 10 to 40% of the local biomass in the Linxia Basin (Wang and Deng, 2005). Thus, both the modern cow and goat could have ingested some C4 vegetation during summer months. Also, because the modern-day environment of the Linxia basin is strongly influenced by the East Asian Summer Monsoon, increased serial δ13C values that occur in tandem with decreased serial δ18O values of tooth enamel from modern animals would require the incorporation of some C4 vegetation during the summer. These data are also consistent with a strengthening of the summer monsoon between 2-3 Ma.

3.5. Conclusions

Carbon and oxygen isotopic analyses of tooth enamel indicate that significant changes occurred in the climates and diets of mammalian taxa from the Linxia basin, as well as in the

Table 3.2. Carbon isotope compositions of plants from the Linxia Basin.

13 13 13 δ C δ C δ C Lab No. Value C3 or C4 Lab No. Value C3 or C4 Lab No. Value C3 or C4

Lx-01 -28.1 C3 Lx-19 -12.4 C4 Lx-37 -27.4 C3 Lx-02 -26.5 C3 Lx-20 -27.6 C3 Lx-38 -27.3 C3 Lx-03 -25.4 C3 Lx-21 -25.8 C3 Lx-39 -24.4 C3 Lx-04 -25.9 C3 Lx-22 -26.0 C3 Lx-40 -27.2 C3 Lx-05 -27.1 C3 Lx-23 -25.0 C3 Lx-41 -26.4 C3 Lx-06 -28.6 C3 Lx-24 -28.7 C3 Lx-42 -26.7 C3 Lx-07 -28.8 C3 Lx-25 -28.5 C3 Lx-43 -24.6 C3 Lx-08 -26.8 C3 Lx-26 -25.1 C3 Lx-44 -27.7 C3 Lx-09 -25.0 C3 Lx-27 -13.3 C4 Lx-45 -27.2 C3 Lx-10 -13.0 C4 Lx-28 -25.8 C3 Lx-46 -26.7 C3 Lx-11 -28.4 C3 Lx-29 -26.5 C3 Lx-47 -28.5 C3 Lx-12 -27.5 C3 Lx-30 -26.5 C3 Lx-48 -27.3 C3 Lx-13 -27.5 C3 Lx-31 -25.5 C3 Lx-49 -27.2 C3 Lx-14 -25.9 C3 Lx-32 -25.4 C3 Lx-50 -26.8 C3 Lx-15 -13.5 C4 Lx-33 -25.1 C3 Lx-51 -26.7 C3 Lx-16 -27.5 C3 Lx-34 -25.5 C3 Lx-52 -25.2 C3

Lx-17 -27.1 C3 Lx-35 -13.9 C4 Lx-53 -12.7 C4

Lx-18 -24.6 C3 Lx-36 -27.6 C3 Lx-54 -13.0 C4

seasonal patterns of diet and climate, throughout the last 25 million years. Fluctuations in mean bulk δ18O values were nearly identical for horses and rhinos collected throughout the entire sedimentary sequence in the Linxia Basin, suggesting an unstable regional climate from 25 to 0.05 Ma. Positive and negative shifts in the mean bulk δ18O values of enamel from both horses and rhinos were roughly consistent with deep-sea records that indicated a general warming trend from ~26 to 15 Ma and two major cooling phases during the Neogene, respectively. A positive shift in both horse and rhino mean bulk δ18O data was also consistent with a previous study that indicated the Linxia Basin shifted to more arid and/or warmer conditions at 12 Ma, based on oxygen isotope compositions of fluvial and lacustrine carbonates. A positive shift in the mean bulk δ18O values of horses, rhinos and deer from the Linxia Basin was similar in timing to a positive δ18O shift observed in fossils and paleosols from Pakistan and Nepal, suggesting a shift toward a drier and/or warmer climate on both the north and south sides of the Tibetan Plateau during the Late Miocene. Additionally, individual bulk δ18O values of fossil tooth enamel from bovids, deer, giraffes, pigs, and elephants were consistent with the positive and negative trends in horse and rhino mean δ18O values. The mean bulk δ13C values for horses and rhinos indicated that both taxa had pure C3 diets throughout most of the Late Cenozoic. At 1.2 Ma, the horse bulk δ13C values increased to ~-5 ‰, indicating a change to a mixed C3/C4 diet after 2.5 Ma. This suggests that C4 grasses may have not spread into the basin until after 2.5 Ma, which is much later than the proposed global C4 expansion during the Late Miocene. This also indicates a strengthening of the Asian summer monsoon after ~2.5 Ma, as C4 plants require summer precipitation. All horse and rhino bulk δ13C values were greater than -13 ‰, indicating an open environment, such as a savannah or mixed woodland/steppe biome in the Linxia Basin from ~25 to 0.05 Ma. Individual bulk δ13C values of fossil tooth enamel from bovids, deer, giraffes, pigs, and elephants fell within or near the ranges of horse and rhino bulk δ13C values at any given age. The bulk δ13C values of most of those individuals indicated pure C3 diets. In addition to the single horse at 1.2 Ma that had a bulk δ13C value indicating a mixed C3/C4 diet, the carbon isotope composition of tooth enamel from a single bovid at 7.5 Ma indicated that it may have also had a mixed C3/C4 diet or fed on C3 plants experiencing water stress. If this bovid indeed consumed a small amount of C4 plants, this would be consistent with a Late Miocene C4 expansion. But because there is no indication of

mixed C3/C4 diets for any other individual (including horses) at 7.5 Ma, it is likely that very little, if any, C4 vegetation was present in the Linxia Basin at that time. Serial oxygen isotopic analyses showed that, in general, positive δ18O shifts in the horse and rhino bulk data, indicating shifts to either drier and/or warmer conditions after 14, 9.5, 7.5, and 2.5 Ma, were accompanied by increased seasonality, as inferred from the relatively greater δ18O ranges in the serial data. Likewise, negative δ18O shifts in the bulk data at 11.5, 6.0, 4.0, and 1.2 Ma were associated with decreases in seasonality, or relatively smaller δ18O ranges in the serial data. A marked increase in the serial δ18O ranges of both horses and bovids after 2.5 Ma is consistent with a strengthening of the summer monsoon in the region after ~2-3 Ma. The serial carbon isotope results showed that prior to 1.2 Ma, all sampled mammalian taxa had pure C3 diets. The serial δ13C ranges of all horses from the ages 11.5, 9.5, 7.5, 6.0, 4.0, and 2.5 Ma were smaller than those of horses from 1.2 and 0.05 Ma. This increase in δ13C ranges further supports changes in the composition of plant biomass in the Linxia Basin after 2.5 Ma, as taxa with mixed C3/C4 diets would have increased δ13C ranges in their enamel compared to those with pure C3 diets. Interestingly, decreases in the δ18O values within individual teeth of horses from 1.2 and 0.05 Ma occurred simultaneously with increases in the δ13C values. This negative correlation between δ18O and δ13C values is consistent with that expected in summer monsoonal regions within China, but not outside of monsoonal regions and strongly supports a strengthening of the summer after ~2-3 Ma. Serial analyses of three bovid individuals from 2.5 Ma and two modern bovids also showed an anti-correlation between δ13C and δ18O values for all individuals, consistent with a strengthened monsoon circulation since about 2-3 Ma.

CHAPTER 4

PALEOECOLOGY OF CENOZOIC RHINOS FROM NORTHWEST CHINA: A STABLE ISOTOPE PESPECTIVE

4.1. Introduction

The Superfamily Rhinocerotoidea, which includes the families Amynodontidae, Hyracodontidae, and Rhinocerotidae, was the largest and most ecologically diverse group of perissodactyls throughout the Cenozoic. The three groups of rhinocerotoids diverged in the Late Eocene of Asia and North America (Prothero et al., 1989) and the rhinocerotoids in China flourished until the Quaternary (Deng and Downs, 2002). The amynodonts reached their peak diversity in Asia during the Late Eocene and Early Oligocene (Prothero et al., 1989), becoming extinct by the Early Miocene (Wall, 1989). The hyracodonts first appeared in the Middle Eocene and became more advanced and diversified until the Late Oligocene (Prothero et al., 1989; Deng et al., 2004a). This group included both giant forms that browsed tree-tops and smaller dog- or goat-sized forms (Radinsky, 1967; Prothero et al., 1989). The hyracodonts vanished from Asia by the Middle Miocene. The Rhinocerotidae, or true rhinoceroses, are first known from the Late Eocene of Eurasia and became increasingly diversified during the Oligocene (Prothero et al., 1989). The Chinese Neogene Rhinocerotidae, represented by 25 species, had a temporal range from the Early Miocene to the Late Pliocene and were widespread, as well as taxonomically and ecologically diverse. The Middle Miocene and Late Miocene were stages of high diversity for the Chinese rhinocerotids, whereas the Early Miocene, early Late Miocene, and Pliocene were stages of low diversity (Deng and Downs, 2002). Very few species of rhinoceroses survived both the Late Miocene and Pleistocene extinction events; there are only five extant species of rhinos in Asia and Africa and all are currently in danger of extinction. Because rhinoceros diversity and morphology are closely related to environmental factors and are sensitive to fluctuations in ambient temperature and humidity (Deng and Downs, 2002),

52 many inferences regarding the paleoecologies of rhinos have been based on taxonomic diversity, as well as cranial and limb morphology (e.g., Radinsky, 1967; Heissig, 1989a, 1999; Prothero et al., 1989; Cerdeño and Nieto, 1995; Deng and Downs, 2002). Even so, most studies of Chinese Late Cenozoic rhinoceroses are primarily taxonomic and rarely involve paleoecological analyses (Deng and Downs, 2002). Periods of high taxonomic diversity of rhinos from Europe and Asia have been linked with warm and moist climates and stages of low diversity with cooler and drier climates (Cerdeño and Nieto, 1995; Deng and Downs, 2002). Cranial morphological characteristics, such as muzzle shape or the presence of horns with growth lines, have aided researchers in reconstructions of paleodiets and have helped to infer the types of climates that fossil rhinos likely inhabited (e.g., Fortelius, 1983; Deng and Downs, 2002). Dental morphological studies have allowed researchers to determine if particular rhino species had adaptations, such as hypsodonty or other specializations of dentition, for grazing or browsing (e.g., Radinsky, 1967; Heissig, 1989a, 1989b, 1999; Lucas and Sobus, 1989; Deng and Downs, 2002; Deng, 2003, 2005a, 2005b, 2006, 2007) and limb morphology has allowed insight regarding the cursorial behavior of ancient rhinos, or the lack thereof (e.g., Heissig, 1989a, 1989b; Prothero et al., 1989; Cerdeño, 1998; Deng and Downs, 2002; Deng, 2002, 2004, 2008). Stable carbon and oxygen isotope analysis of tooth enamel has been established as a valuable tool for reconstructing the diets and environments of ancient animals (e.g., Wang et al., 1994; Cerling et al., 1997a; MacFadden, 1998), and therefore, it is possible to test previous hypotheses concerning the diet, behavior and ecology of fossil rhinos using a stable isotope approach. Carbon isotope compositions of mammalian herbivore tooth enamel serve as proxies for the type of vegetation consumed by the animals, as most carbon isotopic variation in enamel is a result of isotope fractionation that occurred during photosynthesis of plants that were ingested. C3 plants (e.g., trees, shrubs, forbs, and high latitude and high elevation grasses) use the Calvin Cycle photosynthetic pathway and have a δ13C range of -34 to -22 ‰, with an average δ13C value of -27 ‰. Because tooth enamel carbonate is 13C-enriched by ~14 ‰ relative to diet, tooth enamel δ13C values that are < -8 ‰ usually indicate a pure C3 diet. C4 plants (e.g., warm season grasses, a few shrubs, corn, sugar cane, sorghum, and bamboo) use the Hatch-Slack photosynthetic pathway and have a δ13C range of -17 to -9 ‰, with an average of -13 ‰. Thus, tooth enamel δ13C values that are > -3 ‰ indicate a pure C4 diet (Lee-Thorp and Van der Merwe, 1987; O’Leary, 1988; Farquhar et al., 1989; Koch, 1998; Cerling et al., 1997a; Cerling

and Harris, 1999; Kohn and Cerling, 2002; Wang and Deng, 2005). Water availability also affects the carbon isotopic compositions of plants. As plants become water-stressed or grow in a relatively open environment where more evaporation occurs, δ13C values increase. In closed- canopies or forests, δ13C values decrease, due to the influence of soil respiration (Schleser and Jayasekera, 1985; Sternberg et al., 1989; van der Merwe and Medina, 1989). Oxygen isotopic compositions of tooth enamel from large-bodied mammalian herbivores largely reflect the isotopic composition of local meteoric water that is ingested by the herbivores, either by drinking or by plant consumption (Longinelli, 1984; Luz et al., 1984; Ayliffe and Chivas, 1990). In turn, the oxygen isotopic composition of meteoric water is controlled by climate (Dansgaard, 1964; Rozanski et al., 1992). As a result, the oxygen isotopic compositions of mammalian tooth enamel can be utilized for reconstruction of regional paleoclimates. Late Cenozoic deposits in the Linxia Basin have produced well-preserved rhinoceros fossils that range in age from 25 to 2.5 Ma and include individuals from 11 genera. In this study, the stable carbon (C) and oxygen (O) isotopic ratios of bulk tooth enamel samples from 47 rhino individuals representing all 11 genera from the Linxia Basin and 5 individuals representing 1 genus from the nearby Tianshui Basin were determined (Appendix A). In addition, the carbon and oxygen isotopic compositions were determined for 243 serial samples from 15 rhino individuals representing 6 genera that existed in the Linxia Basin from 25 to 6 Ma (Appendix B). The objective was to examine the dietary evolution of the Rhinocerotoidea in NW China, to understand how these ancient animals lived and interacted with their environment, and to test hypotheses based on morphological characters of teeth and bones. In addition, the bulk and serial isotopic compositions of common genera were compared, both temporally and spatially, in order to examine behavioral, physiological, and climatic effects on the evolution of those genera.

4.2. Study Site

The Linxia Basin is located in southeastern Gansu Province, China, about 100 km south of the provincial capital, Lanzhou, on the northeastern margin of the Tibetan Plateau (Figure 4.1). It is a temperate steppe biome with a mean annual temperature of 7°C and an annual rainfall of 515 mm. The present-day climate is strongly influenced by the East Asian monsoon system

Figure 4.1. Map of China showing location of the Linxia Basin. The study site is located on the northeastern margin of the Tibetan Plateau, about 100 km south of Lanzhou, the capital city of Gansu Province, China (Modified from Wang and Deng, 2005).

(An et al., 2000). The Late Cenozoic deposits in the Linxia Basin are thick and well-exposed and span almost continuously from the Late Oligocene to the Holocene (Figure 4.2). The strata contain abundant, diverse, and well-preserved mammalian fossils, and the chronology of the sequence has been determined by magnetostratigraphy and biostratigraphy (Fang et al., 1997, 2003; Deng et al., 2004a, b). The deposits consist primarily of fluvial and lacustrine sediments, with Pleistocene eolian deposits covering most of the region (Fang et al., 2003; Deng et al., 2004a, b; Wang and Deng, 2005). The stratigraphic sequence is comprised of eleven formations based on lithofacies and paleontology (Figure. 4.2 and Table 4.1). The Tianshui Basin is located approximately 300 km southeast of the Linxia Basin (Figure 4.1). Like the Linxia Basin, the Tianshui Basin is a Late Cenozoic sedimentary basin containing thick sequences of fluvial, lacustrine, and eolian deposits. The present-day climate in the area is semi-arid with a mean annual rainfall of 561 mm and an annual mean temperature of 10.5°C. The Late Cenozoic deposits in the basin have yielded abundant fossils, including rhinos (Li et al., 2007). As it is in close proximity to the Linxia Basin, its present-day climate is also temperate and greatly influenced by the East Asian monsoon.

4.3. Materials and Methods

4.3.1. Sample Materials Late Cenozoic sedimentary sequences of the Linxia Basin have yielded an abundance of well-preserved rhinoceros fossils that range in age from 25 to 2.5 Ma and include individuals from 11 genera within the families Hyracodontidae and Rhinocerotidae. Individuals from two genera within 2 subfamilies of the Hyracodontidae, Paraceratherium (giant rhinoceros) and Allacerops, have been recovered from the Linxia Basin. Supraspecific classification of the Linxia Basin hyracodonts follows that of Prothero and Schoch (1989).

Family Hyracodontidae Cope, 1879 Subfamily Indricotheriinae Borissiak, 1923 Genus Paraceratherium Forster Cooper, 1911 Subfamily Allaceropinae Wood, 1932 Genus Allacerops Wood, 1932

Figure 4.2. Sedimentary sequence of the Linxia Basin. This well-exposed sequence has an overall thickness of more than 500 meters and spans almost continuously from the L. Oligocene to the Holocene. The deposits are mainly fluvial and lacustrine sediments and Pleistocene eolian loess deposits cover most of the area. The chronology has been determined by magnetostratigrahy and biostratigraphy (Modified from Fang et al., 1997, 2003; Deng et al., 2004b).

Table 4.1. Lithology and Paleontology of the Linxia Basin Stratigraphic Sequence (from Deng et al., 2004a,b.)

Formation Age (epoch) Lithology Paleontology

Tala Early brownish red no known mammalian fossils Oligocene conglomerates, sandstones and mudstones

Jiaozigou Late Oligocene brownish yellow fossils include Tsaganomys sp., Dzungariotherium orgosense, sandstones and brownish Allacerops sp., Ronzotherium sp., Aprotodon sp., red mudstones Schizotherium sp., and Paraentelodon macrognathus

Shangzhuang Early Miocene brownish red silty fossils include Gomphotherium sp., Aprotodon sp., and mudstones and bluish gray Turcocerus sp. carbonate-cemented sandstones

Dongxiang Early Middle brownish red sandstones fossils include Hemicyon sp., Choerolophodon sp., Miocene and mudstones with Gomphotherium sp., Anchitherium sp., Alicornops sp., grayish white marlite beds Hispanotherium matritense, Chalicotherium sp., and Kubanochoerus sp.

Hujialiang Late Middle grayish yellow fine fossils include: Alloptox sp., Pliopithecus sp., Hemicyon sp., Miocene conglomerates and Amphicyon sp., Gomphotherium sp., Platybelodon grangeri, sandstones Zygolophodon sp., Anchitherium sp., Alicornops sp., Hispanotherium matritense, Kubanochoerus gigas, Listriodon sp., Palaeotragus sp., and Turcocerus sp.

Lower Liushu Late Miocene red clay with sandstones fossils include Dinocrocuta gigantea, Machairodus sp., and conglomerates Tetralophodon sp., Hipparion dongxiangense, interbedded at some Parelasmotherium simplym, Parelasmotherium linxiaense, localities, resulting in and Shaanxispira sp. variable thickness in different sections

Middle Late Miocene red clay with sandstones fossils include Pararhizomys hipparionum, Promephitis sp., Liushu and conglomerates Promephitis hootoni, Melodon majori, Ictitherium sp., interbedded at some Hyaenictitherium wongii, Hyaenictitherium hyaenoides, localities, resulting in Dinocrocuta gigantea, Machairodus palanderi, Felis sp., variable thickness in Tetralophodon exoletus, Hipparion sp., Hipparion chiai, different sections Hipparion weihoensis, Acerorhinus hezhengensis, Chilotherium wimani, Iranotherium morgani, Chleuastochoerus stehlini, Dicrocerus sp., Samotherium sp., Honanotherium schlosseri, Gazella sp., Hezhengia bohlini, and Miotragocerus sp.

Table 4.1 continued

Formation Age (epoch) Lithology Paleontology

Upper Liushu Late Miocene red clay with sandstones and fossils include Hystrix gansuensis, Pararhizomys conglomerates interbedded at hipparionum, Simocyon sp., Promephitis sp., Promephitis some localities, resulting in hootoni, Parataxidea sinensis, Pleisiogulo sp., Ictitherium variable thickness in sp., Hyaenictitherium wongii, Hyaenictitherium different sections hyaenoides, Adcrocuta variabilis, Machairodus sp., Machairodus palanderi, Metailurus sp., Metailurus minor, Felis sp., Hipparion sp., Hipparion coelophyes, Hipparion dermatorhinum, Acerorhinus hezhengensis, Chilotherium wimani, Dicerorhinus ringstromi, Ancylotherium sp., Chleuastochoerus stehlini, Microstonyx major, Metacervulus sp., Cervavitus novorossiae, Honanotherium schlosseri, Palaeotragus microdon, Miotragocerus sp., Sinotragus sp., Protoryx sp., and Gazella sp.

Hewangjia Early Pliocene red clay and a basal fossils include Hystrix gansuensis, Promephitis sp., conglomerate with variable Chasmaporthetes sp., Hyaenictitherium wongii, Hipparion thickness at different sp., Shansirhinus ringstromi, Cervavitus novorossiae, localities Palaeotragus sp., Sinotragus sp., and Gazella sp.

Jishi Late Pliocene gray and partially carbonate- no known mammalian fossils cemented coarse conglomerates

Wucheng Early yellowish brown massive fossils include Aepyosciurus orientalis, Marmota sp., Loess Pleistocene siltstones Cromeromys cf., Cromeromys gansunicus, Bahomys sp., Sericolagus brachypus, Macaca anderssoni, Paradolichopithecus sp., Vulpes chikushanensis, Canis sp., Meles sp., Chasmaporthetes progressus, Pachycrocuta licenti, Crocuta honanensis, Homotherium crenatidans, Megantereon nihowanense, Sivapanthera sp., Panthera palaeosinensis, Felis teilhardi, Lynx shansius, Hipparion sinense, Equus sp., Coelodonta nihowanensis, Hesperotherium sp., Nipponicervus sp., Gazella cf., Gazella blacki, Leptobos brevicornis, and Hemibos sp.

Lishi Loess Middle yellowish brown massive fossils include Equus qingyangensis Pleistocene siltstones

Malan Loess Late Pleistocene yellowish brown massive fossils include Equus hemionus siltstones

The rhinocerotid samples from the Linxia Basin represent 9 genera from 4 tribes within 2 subfamilies. Supraspecific classification of the Linxia Basin rhinocerotids follows that of Heissig (1999) except that an additional tribe, Chilotheriini, was established within subfamily Aceratheriinae by Qiu et al. (1987).

Family Rhinocerotidae Gill, 1872 Subfamily Aceratheriinae Dollo, 1885 Tribe Aceratheriini Dollo, 1885 Genus Alicornops Ginsburg and Guérin, 1979 Alicornops laogouense Deng, 2004 Tribe Chilotheriini Qiu et al., 1987 Genus Chilotherium Ringström, 1924 Chilotherium wimani Ringström, 1924 Genus Acerorhinus Kretzoi, 1942 Acerorhinus hezhengensis Qiu et al., 1987 Genus Shansirhinus Kretzoi, 1942 Shansirhinus ringstroemi Kretzoi, 1942 Subfamily Rhinocerotinae Dollo, 1885 Tribe Elasmotheriini Dollo, 1885 Genus Hispanotherium Crusafont and Villalta, 1947 Hispanotherium matritense Prado, 1864 Genus Iranotherium Ringström, 1924 Iranotherium morgani (Mecquenem, 1908) Genus Parelasmotherium Killgus, 1923 Parelasmotherium linxiaense Deng, 2001 Tribe Rhinocenotini Dollo, 1885 Genus Dicerorhinus Gloger, 1841 Dicerorhinus ringstroemi Arambourg, 1959 Genus Coelodonta Bronn, 1831 Coelodonta nihowanensis Kahlke, 1969

52 well-preserved fossil rhino teeth were selected for this study. All rhino teeth were collected from Linxia Basin deposits by the Institute of Vertebrate Paleontology and Paleoanthropology (IVPP) and the specimens were housed in the Hezheng Museum in Hezheng County, Gansu, China, or at IVPP. A total of 52 bulk and 243 serial enamel samples were obtained from these teeth for stable C and O isotope analyses. These samples represent all 11 genera within the families Hyracodontidae and Rhinocerotidae found in the Linxia Basin and 1 genus from the Tianshui Basin. Rhinoceroses were ideal for this study because this important group spans almost continuously from the late Oligocene to the Pliocene within the Linxia Basin, has relatively large teeth with thick enamel, and is abundant in the fossil record. Most samples collected were third molars and premolars, so that the effects of pre-weaning on the δ18O of the tooth enamel were reduced (e.g., Fricke and O’Neil, 1996; Bryant et al., 1996). Multiple specimens from each stratum were collected whenever possible so that variations in δ13C and δ18O values within populations, between different species of the same age, and between same species of different ages could be determined.

4.3.2. Laboratory Methods The bulk sampling of tooth enamel for isotopic analyses was accomplished by either drilling along the entire length of a tooth using a slow-speed rotary drill or by, first, cutting off a section of each tooth, lengthwise from the crown to the root, with a hammer and chisel or with a rotary tool. Then, the enamel was manually separated from the dentine using a rotary tool. Finally, the samples were ground into a fine powder using a mortar and pestle. Any matrix that was initially present was manually removed from each sample prior to cutting. It was necessary to sample along the whole length of a tooth in order to obtain results that represented average isotopic compositions over the life of an individual. Serial samples were drilled, using a slow-speed rotary tool with a diamond point, perpendicular to the growth axis of each tooth from crown to root, with the youngest samples being near the root and the oldest samples being near the crown. The outer surfaces of all teeth were manually cleaned prior to sampling. Tooth enamel carbonate samples were treated in 5% reagent grade sodium hypochlorite for approximately 20 to 24 hours at room temperature to remove organic material from the tooth enamel. The samples were then centrifuged, decanted, and rinsed 4 to 5 times with deionized

water to remove the sodium hypochlorite. Next, the samples were treated in 1M acetic acid for 4 to 15 hours at room temperature to remove non-structural carbonate from the enamel. After the final rinse, the samples were dried under vacuum in a freeze-dryer for 3 to 5 days. To analyze the carbonate samples, ~200 to 500 μg of carbonate standards and 3 to 6 mg of enamel carbonate samples were measured and placed into reaction vials capped with rubber septa. After loading the sample vials into a Thermo-Finnigan Gasbench II interfaced with a Delta Plus XP continuous flow isotope ratio mass spectrometer, all carbonate samples were flushed for 5 minutes by injection of pure-grade helium in order to force air and moisture from the vial. The

samples were then converted to CO2 by injection of 100% phosphoric acid, which was left to react for 3 to 18 hours at 72ºC or for approximately 72 hours at 25 ºC (modified from McCrea,

1950). After the carbonate samples were converted to CO2, the carbon and oxygen isotope ratios

were measured by mass spectrometry. Ten aliquots of CO2 from each sample vial were measured

and run against three aliquots of a CO2 reference gas. The results are reported in standard notation as δ13C and δ18O in reference to V-PDB and V-SMOW, respectively (Gonfiantini, 1978; 13 13 13 12 13 12 Gonfiantini et al., 1995). δ C is defined as follows: δ C = [(( C / C)SAMPLE - ( C / C)VPDB)/ 13 12 18 18 18 16 18 16 18 ( C / C)VPDB] x 1000 ‰. δ O is defined as: δ O = [(( O / O)SAMPLE - ( O / O)VSMOW)/ ( O 16 / O)VSMOW] x 1000 ‰. The δ13C and δ18O values reported are the average values for the ten sample aliquots. Acceptable standard deviations from the mean delta values of the samples, or external errors, were < 0.06 ‰ for δ13C and < 0.08 ‰ for δ18O. The expected internal error was < 0.05 ‰ for 18 13 18 δ O (CO2 reference gas; Thermo-Finnigan, 2002). The δ C and δ O values were calibrated by concurrent carbon and oxygen isotopic measurements of least two sets of three or more of the following carbonate standards: PDA, NBS-19, YW-CC-ST-1 (lab standard), ROY-CC (lab standard), and MERK. Although most samples were analyzed on a Delta Plus XP IRMS at Florida State University, some tooth enamel samples were analyzed on a VG Prism in the Stable Isotope Lab at the University of Florida.

4.4. Results and Discussion

4.4.1. Late Oligocene Rhinoceroses The hyracodontids Allacerops sp. and Paraceratherium sp. have been recovered from the sandstones of the Late Oligocene Jiaozigou Formation (Deng et al., 2004a). Allacerops and Paraceratherium lived concurrently in the Linxia Basin at 25 Ma. All hyracodonts had long, laterally compressed metapodials, indicating cursorial behavior or ancestry (Prothero et al., 1989). Allacerops was a large hyracodont that had large canine tusks and no particular specialization of the anterior dentition (Radinsky, 1967). Heissig (1989b) noted, based on limb structure, that the allaceropine hyracodonts were less cursorial than were the smaller hyracodonts. Paraceratherium belonged to a group of hyracodonts known as the indricotheres and was the largest land mammal that ever lived. Paraceratherium is characterized by its large size and derived or specialized anterior dentition, where the first pair of incisors are very large and conical, with the upper incisor vertical and the lower procumbent, and the posterior incisors and canines are lost (Radinsky, 1967; Lucas and Sobus, 1989). Its large body size and specialized anterior dentition support that Paraceratherium cropped vegetation from tree-tops (Lucas and Sobus, 1989). Although gigantic in size, Paraceratherium retained long metapodials, which is reflective of a cursorial ancestry (Prothero et al., 1989). Bulk carbon and oxygen isotope compositions of tooth enamel from four Allacerops individuals and five Paraceratherium individuals were determined in this study (Figures 4.3 and 4.4, respectively). The group of Allacerops individuals had a mean δ13C value of -9.5 ‰, a δ13C range of 1.1 ‰ (-9.0 ‰ max., -10.1 min.), a mean δ18O value of -7.9 ‰, and a δ18O range of 2.7 ‰ (-6.8 ‰ max., -9.5 min.). The Paraceratherium group had a mean δ13C value of –10.2 ‰, a δ13C range of 0.5 ‰ (-9.9 ‰ max., -10.4 ‰ min.), a mean δ18O value of -11.5 ‰, and a δ18O range of 1.1 ‰ (-10.9 ‰ max., -12.0 min.). The carbon isotope results indicate that both Allacerops and Paraceratherium had pure C3 diets (all δ13C values < -8 ‰), although Allacerops had a more 13C-enriched diet. Significant differences in mean δ13C (p = 0.0135) and δ18O values (p = 0.0004) between Allacerops and Paraceratherium (Table 4.2) suggest niche partitioning of the two genera. The larger δ13C range for Allacerops (Figure 4.5) suggests that this rhino was a less selective feeder than was Paraceratherium. This is consistent with Paraceratherium having a more specialized dentition. The larger δ18O range and higher δ18O values for Allacerops

0 2 4 Coelodonta 6 Shansirhinus

8 Chilotherium

Dicerorhinus 10 12 Acerorhinus (Ma) Iranotherium 14 Parelasmotherium

Age 16 Alicornops 18 Hispanotherium 20 Paraceratherium 22 Allacerops 24 Chilotherium (Tianshui)

-13 -12 -11 -10 -9 -8 -7 -6 -5 -4 -3 13 δ C (‰, V-PDB)

Figure 4.3. Bulk carbon isotope compositions of tooth enamel from rhinos from the Linxia Basin versus age. All rhino samples were collected in the Linxia Basin, with the exception of five 9 Ma Chilotherium individuals from the Tianshui Basin (closed circles).

0 2

4 Coelodonta 6 Shansirhinus 8 Chilotherium 10 Dicerorhinus

12 Acerorhinus

(Ma) Iranotherium 14 Parelasmotherium 16

Age Alicornops 18 Hispanotherium 20 Paraceratherium 22 Allacerops 24 Chilotherium (Tianshui) 26 -13 -12 -11 -10 -9 -8 -7 -6 -5 -4 -3 δ18O (‰, V-PDB)

Figure 4.4. Bulk oxygen isotope compositions of tooth enamel from rhinos from the Linxia Basin versus age. All rhino samples were collected in the Linxia Basin, with the exception of five 9 Ma Chilotherium individuals from the Tianshui Basin (closed circles).

3 9(2) (‰, V-PDB) 6(5) 25(4) Chilotherium 13(5) 7(2) Parelasmotherium 2 Alicornops O Range Range O Paraceratherium

δ 25(5) Allacerops 7.5(6) 1 Chilotherium (Tianshui) 9.5(4) 9(5) 17(3) 11.5(5) 0 012345 δ13C Range (‰, V-PDB) Explanation of data labels: age in Ma (n samples)

Figure 4.5. δ13C range versus δ18O range for rhino individuals of particular genera at given geologic ages. All rhino samples were collected in the Linxia Basin, with the exception of five Chilotherium individuals from the Tianshui Basin (open circle).

Table 4.2. Statistical significance of differences in mean carbon and oxygen isotopic compositions of various rhino taxa at different ages as determined by two-tailed t-tests.

Mean Significant difference difference at Sample (‰) df t p 95% CI? 13 13 Allacerops δ C (25 Ma) vs Paraceratherium δ C (25 Ma) 0.7 7 3.28 0.0135 Yes 18 18 Allacerops δ O (25 Ma) vs Paraceratherium δ O (25 Ma) 3.6 7 6.38 0.0004 Yes

Alicornops δ13C (13 Ma) vs Alicornops δ13C (17 Ma) 1.3 6 1.35 0.2256 No Alicornops δ18O (13 Ma) vs Alicornops δ18O (17 Ma) 0.7 6 1.21 0.2705 No

13 13 Chilotherium δ C (9.5 Ma) vs Chilotherium δ C (9 Ma) 0.2 4 0.67 0.5398 No Chilotherium δ18O (9.5 Ma) vs Chilotherium δ18O (9 Ma) 2 4 -2.11 0.1030 No

13 13 Chilotherium δ C (9 Ma) vs Chilotherium δ C (7.5 Ma) 1.3 6 -2.91 0.0270 Yes Chilotherium δ18O (9 Ma) vs Chilotherium δ18O (7.5 Ma) 1.6 6 1.91 0.1053 No

13 13 Chilotherium δ C (7.5 Ma) vs Chilotherium δ C (7 Ma) 1.4 6 3.28 0.0168 Yes Chilotherium δ18O (7.5 Ma) vs Chilotherium δ18O (7 Ma) 1.0 6 1.40 0.2118 No

13 13 Chilotherium δ C (7 Ma) vs Chilotherium δ C (6 Ma) 2.1 5 -6.22 0.0016 Yes 18 18 Chilotherium δ O (7 Ma) vs Chilotherium δ O (6 Ma) 4.1 5 -3.56 0.0162 Yes

(Figure 4.5) suggests that this genus occupied relatively open habitats or exhibited a more migratory behavior than did Paraceratherium. This is a result of enamel-δ18O being determined by the δ18O of body water, and for herbivores, body water is derived from drinking water and water in food plants (e.g., Longinelli, 1984; Luz et al., 1984; Bryant and Froelich, 1995; Kohn, 1996). Water bodies that provide drinking water for animals (e.g., puddles, ponds, lakes) tend to have higher δ18O values in open environments (due to higher evaporation rates) than in more closed forested environments. Also, plants (leaves & fruits) are > 80% water. Studies have shown that leaf water is generally enriched in 18O relative to local meteoric water due to preferential loss of isotopically light water molecules during evapotranspiration and that the extent of 18O-enrichment is dependent on the relative humidity (Dongmann et al., 1974; Epstein et al., 1977; Yakir, 1992). Consequently, leaf water tends to have more enriched δ18O values and displays a larger range of δ18O variation in open habitats than in more closed forested habitats in a given geographic area. Thus, it would be expected that individuals living in more open habitats or those migrating between habitats would display greater δ18O variability in their enamel than would non-migratory forest-dwellers. Additionally, if the Paraceratherium individuals received most of their water intake from longstanding lakes or rivers where groundwater predominated, and Allacerops drank from multiple alternative sources that consisted mainly of surface waters, including ephemeral streams and puddles that are more likely to reflect the oxygen isotope compositions of the local rainfall, then it would be expected that Paraceratherium would have less δ18O variation between individuals living in the same region concurrently, as more permanent river and lake waters display a much smaller seasonal δ18O variability compared to precipitation, reflecting the weighted mean annual δ18O of precipitation in the catchment area modified by evaporation (Clark and Fritz, 1997). The higher δ13C values of Allacerops’ tooth enamel may also suggest that this genus lived in a more open habitat than did Paraceratherium, which is consistent with the δ18O data. As stated previously, plants that grow in a relatively open environment where more evaporation occurs have increased δ13C values. Therefore, if Allacerops ingested plants that grew in a relatively open habitat compared to that of Paraceratherium, it would be expected that Allacerops would have higher δ13C values. Also, it has been demonstrated that in dense closed- canopy forests, the δ13C values of foliage collected near the forest floor is 13C-depleted relative

to foliage from the top of the canopy (Vogel, 1978; Medina & Minchin, 1980; Medina et al.,

1986, 1991; Sternberg et al., 1989) due to the incorporation of soil-respired CO2 (Marshall et al., 2007). Given that Paraceratherium was the largest land mammal to have ever lived and must have browsed tree-tops, then it would be expected that the diet of this genus consisted of large amounts of tree leaves and that this rhino required large numbers of trees to feed from. Therefore, it is reasonable to assume that Paraceratherium lived in a forested environment. If the relatively smaller Allacerops also inhabited this forested environment, but fed at a lower level in the canopy, it would be expected that Allacerops would be 13C-depleted relative to Paraceratherium. Because Allacerops was found to be 13C-enriched relative to Paraceratherium, it is unlikely, according to this line of reasoning, that Allacerops was as restricted to a closed woodland dwelling. Thus, both δ13C and δ18O data suggest that Paraceratherium lived in a forested habitat whereas Allacerops preferred more open habitats that experienced seasonal or periodic water stress. The higher δ18O values of Allacerops relative to those of Paraceratherium may also be due to Paraceratherium spending more time each day in an aquatic environment or in mud wallows. All modern species of rhinos are obligate drinkers and are dependent on waterholes for wallowing to cool their bodies (Owen-Smith and Berger, 2006). Therefore, it is reasonable to assume that the extinct rhinos were also obligate drinkers. It is also logical to assume that these extinct rhinos relied on waterholes for wallowing as do the extant rhinos, as both Allacerops and Paraceratherium were large-bodied mammals that must have required the cooling effects of wallowing. Because Paraceratherium was a gigantic rhino and was consequently much larger than Allacerops, Paraceratherium may have necessarily spent much more time in water or in mud wallows than did Allacerops. Bocherens et al. (1996) found that large mammals that spend most of their day within aquatic environments have more negative δ18O values than do large- bodied and more-terrestrial mammals within the same ecosystem. This is due to the intake of greater quantities of water from lakes, rivers, and aquatic plants, which is 18O-depleted relative to water in terrestrial plants, as well as ingestion of terrestrial plants during night, when there is a reduction in the 18O-enriching effects of evaporative transpiration with the cessation of photosynthetic activity. Hence, if Paraceratherium spent most of its day submerged in water and fed from tree-tops during night when photosynthesis did not occur and Allacerops, on the other hand, spent less time in water and fed during the daytime when plants were photosynthetically

active, then it would be expected that Paraceratherium would have more negative δ18O values than would Allacerops. In addition to the bulk analyses, serial carbon and oxygen isotope analyses were performed on two Paraceratherium individuals (Figure 4.6). The results revealed almost no variation in δ13C values throughout the entire tooth growth period for either individual (Δ13C < 1 ‰; Table 4.3), suggesting that there was little or no seasonal variation in diet. This, along with very negative (< -8 ‰) δ13C values, is consistent with Paraceratherium having a specialized and strictly C3 diet. A strict C3 diet is expected for all rhinos that lived prior to the Late Miocene global carbon shift at ~7 Ma (Cerling et al., 1993), when C4 grasses suddenly became a major component in many terrestrial ecosystems. Serial analysis of tooth enamel from a horse that lived in the Linxia Basin at 1.2 Ma, well after the global C4 expansion, showed greatly fluctuating δ13C values throughout the growth period of the tooth (Δ13C = ~5 ‰). It was suggested that the highest values reflected the horse’s diet during the summer months, when C4 grasses were available, and the relatively low values reflected the winter months, when C3 grasses were more dominant (Biasatti et al., manuscript in preparation). Distinct and uniform seasonal cycles in the δ18O curves suggest that these Paraceratherium individuals experienced regular seasonality within their habitats. The negative δ18O values (mean δ18O values = -11.8 ‰ and -11.1 ‰ for both individuals; Table 4.3) and uniform cyclicity of the δ18O curve are consistent with Paraceratherium inhabiting a closed forested habitat. It would be expected that in more open habitats, intra-tooth δ18O variations would be more irregular due to the ingestion of water from a greater variety of water bodies and plants. Both individuals display very similar intra-tooth δ18O and δ13C variations (Figure 4.6), suggesting equivalent diets and habitats for the two individuals. Consequently, the serial data are consistent with inferences made regarding the paleoecology of Paraceratherium based on the bulk C and O isotopic data and support that the climate in the Linxia Basin was relatively cooler and/or wetter during the Late Oligocene compared to the Miocene (Biasatti et al., manuscript in preparation).

0 -2 -4 -6 Alicornops -8 13 Ma -10 Δ18O = 2.8 (Lg‐04) 18 -12 (Lg-04) (Lg-05) Δ O = 2.9 (Lg‐05) -14 0 5 10 15 20 0 5 10 15 20 25 30 35 40 45 50 55 60 0 -2 -4 -6 Hispanotherium -8 14 Ma -10 Δ18O = 2.9 -12 (Lgo-05)

(‰, V-PDB) -14 0 5 10 15 20 25 30 35 40 45 50 55 0 -2 -4 -6 Alicornops -8 17 Ma Values Delta -10 Δ18O = 3.7 (Dl‐01) (Dl-01) (Dl-02) -12 Δ18O = 2.1 (Dl‐02) -14 0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 0 -2 -4 -6 Paraceratherium -8 25 Ma -10 Δ18O = 2.5 (Tl‐02) -12 Δ18O = 2.4 (Tl‐03) (Tl-02) (Tl-03) -14 0 5 10 15 20 25 30 0 5 10 15 20 25 30

Distance from Crown (mm) δ18O δ13C

Figure 4.6. Serial carbon and oxygen isotope compositions of rhino tooth enamel from the Late Oligocene and Middle Miocene. Open diamonds represent oxygen isotope compositions and closed diamonds represent carbon isotope compositions. The data labels in parentheses are the sample numbers of rhino individuals. Serial samples represent decreasing age with distance from the crown. Δ18O = the amplitude of the δ18O seasonal cycles.

Table 4.3. Δ18O and Δ13C values of serial rhino tooth enamel samples.

Δ13C Δ18O Bulk Bulk Mean serial Mean serial (serial (serial 13 18 13 18 Sample No. Species Locality Age δ C value δ O value δ C value δ O value samples) samples) Ds-11 Chilotherium wimani Upper Liushu Fm. 6 -8.9 -4.9 -9.2 -4.6 0.5 2.2 Hl-14 Chilotherium wimani Upper Liushu Fm. 7.5 -9.7 -8.4 -9.5 -8.8 0.5 2.1 Hl-15 Chilotherium wimani Upper Liushu Fm. 7.5 -9.6 -8.4 -9.5 -8.2 0.7 2.0 Ls-06 Iranotherium morgani Middle Liushu Fm. 9.2 -11.0 -3.1 -11.6 -2.3 0.9 1.4 Qj-07 Chilotherium wimani Middle Liushu Fm. 9.5 -12.0 -11.8 -11.0 -10.5 0.7 2.7 Qj-09 Chilotherium wimani Middle Liushu Fm. 9.5 -11.0 -9.6 -11.1 -9.8 0.4 2.1 Gn-13 Parelasmotherium linxiaense Lower Liushu Fm, 11.5 -9.9 -4.7 -10.1 -4.8 0.7 3.6 Gn-14 Parelasmotherium linxiaense Lower Liushu Fm, 11.5 -9.8 -4.5 -10.1 -4.1 0.9 2.4 Lg-04 Alicornops laogouense Laogou Fm. 13 -10.8 -10.4 -10.8 -10.3 0.1 2.8 Lg-05 Alicornops laogouense Laogou Fm. 13 -8.9 -8.5 -9.3 -7.4 0.6 2.9 Lgo-05 Hispanotherium matritense Laogou Fm. 14 -8.7 -7.1 -9.2 -7.0 0.8 2.9 Dl-01 Alicornops laogouense Dongxiang Fm. 17 -8.4 -8.0 -9.0 -7.8 0.7 3.7 Dl-02 Alicornops laogouense Dongxiang Fm. 17 -9.2 -8.1 -9.0 -8.2 0.8 2.1 Tl-02 Paraceratherium sp. Jiaozigou Fm. 25 -9.3 -10.7 -10.2 -11.1 0.8 2.5 Tl-03 Paraceratherium sp. Jiaozigou Fm. 25 -10.2 -11.5 -10.7 -11.8 0.8 2.4

4.4.2. Middle Miocene Rhinoceroses The rhinocerotids Alicornops laogouense and Hispanotherium matritense have been recovered from the sandstones and conglomerates of both the Early Middle Miocene Dongxiang Formation and the Late Middle Miocene Hujialiang Formation (Deng, 2004; Deng et al., 2004a). Alicornops and Hispanotherium lived concurrently in the Linxia Basin throughout the Middle Miocene. It has been demonstrated that a warm and moist environment supports a maximum diversity of rhinoceroses. While the diversity of the family Rhinocerotidae was very low during the Early Miocene, indicating a cold climate, the diversity of the Rhinocerotidae was high during the Middle Miocene, which indicates a warmer climate (Deng and Downs, 2002). In the Linxia Basin, Alicornops and Hispanotherium have been found alongside a large number of Amebelodontidae fossils, which favored habitats near water. This suggests an abundance of lakes and rivers in the Linxia Basin during the Middle Miocene (Deng, 2003; Deng, 2004). Alicornops, an aceratheriine, was a small rhinoceros (Prothero et al., 1989) with short limbs and robust metapodials which were adapted for life on soft soils (Cerdeño, 1998; Deng, 2004). Alicornops laogouense is the largest known species of Alicornops and is considered to have inhabited open woodlands with many lakes and rivers (Deng, 2004). In western Europe, Alicornops is known to have lived in open woodlands with associated lakes and swamps (Guérin, 1980). The elasmothere Hispanotherium had well-developed hypsodont dentition with a thick cementum cover and strong enamel plications, reduced incisors, and a lowered head, which is indicative of a typical steppe grazer (Heissig, 1989a; Deng and Downs, 2002; Deng, 2003). Hispanotherium also had slender limb bones, suggesting cursorial behavior in an open steppe habitat (Deng and Downs, 2002). Although Hispanotherium matritense was considered to live in warm and dry conditions in Europe (Cerdeño and Nieto, 1995), H. matritense was recovered from fluvial sandstones with gravel along with the amebelodontid Platybelodon grangeri in the Linxia Basin, and therefore, it is evident that the climate was not very dry in the Linxia Basin during the Middle Miocene (Deng, 2003). Considering that Alicornops was likely a forest-dweller and that Hispanotherium preferred an open-steppe environment, the Linxia Basin was probably a mixed open steppe/woodland biome during the Middle Miocene. Bulk carbon and oxygen isotope compositions of tooth enamel from three 17 Ma Alicornops individuals, five 13 Ma Alicornops individuals, and one 14 Ma Hispanotherium individual were determined in this study (Figures 4.3 and 4.4, respectively). The group of 17 Ma

Alicornops individuals had a mean δ13C value of -9.4 ‰, a δ13C range of 2.3 ‰ (-8.4 ‰ max., - 10.7 min.), a mean δ18O value of -8.2 ‰, and a δ18O range of 0.7 ‰ (-8.0 ‰ max., -8.7 min.). The group of 13 Ma Alicornops individuals had a mean δ13C value of -10.7 ‰, a δ13C range of 3.5 ‰ (-8.9 ‰ max., -12.4 min.), a mean δ18O value of -9.0 ‰, and a δ18O range of 2.3 ‰ (-8.1 ‰ max., -10.4min.). The 14 Ma Hispanotherium individual had a δ13C value of -8.7 and a δ18O value of -7.1. The carbon isotope results indicate that Alicornops and Hispanotherium had pure C3 diets, although Hispanotherium had a more 13C-enriched diet. The oxygen isotope composition of Hispanotherium was also 18O-enriched relative to Alicornops at both 17 and 13 Ma. Therefore, the carbon and oxygen isotope results indicate that Hispanotherium was ingesting both plants and water in a more open and water-stressed habitat than that of Alicornops. This supports that Hispanotherium grazed in a steppe biome whereas Alicornops preferred to browse within a relatively humid woodland habitat. There were no significant differences in mean δ13C (p = 0.2256) and δ18O values (p = 0.2705) between Alicornops at 17 Ma and Alicornops at 13 Ma (Table 4.2) suggesting similar diets and habitats for that genus at the two different ages. The large δ13C range for Alicornops (Figure 4.5) suggests that this rhino was a less selective feeder than all other rhinos in this study. This is in agreement with a previous suggestion by Heissig (1989a) that the tribe Aceratheriini was less specialized than the other tribes of the subfamily Aceratheriinae. The δ13C and δ18O ranges for Alicornops at 13 Ma were larger than those for Alicornops at 17 Ma (Figure 4.5). Because the mean δ13C and δ18O values for Alicornops are more negative at 13 Ma than at 17 Ma, it is not likely that the larger ranges for Alicornops at 13 Ma are due to a more open habitat at 13 Ma. Instead, the difference in the δ18O range may be due to either (1) error, as a result of small sample size or (2) to Alicornops getting most of its water intake from lakes at 17 Ma and from multiple sources at 13 Ma. The difference in the δ13C range may be a result of (1) again, small sample size or (2) Alicornops being a less selective feeder at 13 Ma than at 17 Ma due to a warming climate, which would allow more species of plants to be available, or as an adaptation to survive among an increasing diversity of large mammals. Serial carbon and oxygen isotope analyses were performed on two 17 Ma Alicornops individuals, two 13 Ma Alicornops individuals, and one 14 Ma Hispanotherium individual (Figure 4.6). The carbon isotope results showed almost no variation in δ13C values for all five

individuals (Δ13C < 1 ‰ for all individuals; Table 4.3), which suggests there was little or no seasonal variation in their diets. This, as well as the rhinos having very negative δ13C values (δ13C < -8 ‰), is consistent with all of these rhinos having pure C3 diets. The mean serial δ13C values are very similar for all individuals (~9 ‰), with the exception of one Alicornops individual at 13 Ma, that had a more negative mean δ13C value of -10.8 (Table 4.3). This is consistent with the inference made based on the bulk C-isotope data that Alicornops was a generalized feeder, as it would be expected that individuals within a population of generalized feeders would not have the exact same diets. Alternative explanations for the differing δ13C values of the Alicornops individuals at 13 Ma are that the two individuals may not have lived at precisely the same time and the more negative values occurred during a time of less water-stress in the region, or that the two individuals may have had dissimilar habitats and/or diets due to different biological ages or sex. It is not possible at this time to determine which explanation is more likely, as the ages and sexes of the two individuals are not currently known and there is no detailed climate record from the Hujialiang Formation in the Linxia Basin to compare results. Well-defined seasonal cycles in the δ18O curves for all Alicornops and Hispanotherium individuals suggest that all individuals experienced regular seasonality within their habitats. The δ18O curves for the two Alicornops individuals at 17 Ma are similar in frequency and in mean δ18O values (mean δ18O = -7.8 and -8.2 ‰; Figure 4.6), but vary in amplitude (Δ18O = 3.7 and 2.1 ‰). This difference in Δ18O values could be due the individuals having different water sources, or if these animals did not live at exactly the same time, it could be due to small-scale or local climate change. Again, if Alicornops was a food generalist, the two individuals would not necessarily be feeding or drinking in the exact same location, so the more 18O-enriched individual may have drank or ingested plant water in a more open environment than did the more 18O-depleted individual. In any case, this serial data illustrates how differing behaviors within populations or between various taxa or how local climatic events may not be evident from bulk C and O isotopic analyses. The serial oxygen isotope data showed that Hispanotherium was 18O- enriched compared to Alicornops at 17 and 13 Ma, which is consistent with the bulk δ18O data. Again, this suggests a more open habitat for Hispanotherium. The seasonal cycles in the Hispanotherium δ18O data have amplitudes that are equivalent to those of Alicornops at 13 Ma and are similar to those of Alicornops at 17 Ma. This suggests that Hispanotherium may have

water sources that were similar to those of Alicornops in its relatively open habitat. The two Alicornops individuals had seasonal δ18O cycles that were very similar in frequency and amplitude (Δ18O = 2.8 and 2.9 ‰), but one individual was 18O-enriched by ~3 ‰, compared to the other. This suggests that the more 18O-depleted individual inhabited a more closed environment or possibly spent more time wallowing than did the more 18O-enriched rhino. It is unlikely that this difference is due to experimental error, as all serial samples were collected and pretreated using the same methods and were reacted and run on the same mass spectrometer at 25ºC. All in all, these serial data indicate that the climate was relatively stable from 17 to 13 Ma and are consistent with previous suggestions that the Linxia Basin was a mixed woodland/steppe biome with an abundance of lakes and rivers in the region during the Middle Miocene. Although the local climate appears to have been unchanged throughout the Middle Miocene, these data also indicate that the local climate during the Middle Miocene was relatively warmer and/or drier than that of the Late Oligocene.

4.4.3. Late Miocene Rhinoceroses The rhinocerotids Parelasmotherium linxiaense, Iranotherium morgani, Acerorhinus hezhengensis, Chilotherium wimani, and Dicerorhinus ringstroemi have been recovered from the red clay of the Late Miocene Liushu Formation (Deng et al., 2004a). At the start of the Late Miocene, there was a decrease in rhinocerotid diversity in China. Although it has been suggested that a cooling event caused this decrease in diversity, the presence of the browser Acerorhinus and the grazer Parelasmotherium implies that the humidity in the early Late Miocene was comparable to that of the Middle Miocene (Deng and Downs, 2002). The giant elasmotherine rhinocerotid Parelasmotherium lived in the Linxia Basin during the early Late Miocene at 11.5 Ma. Parelasmotherium had a huge nasal horn and hypsodont dentition with massive cement filling, well-developed secondary folds, and wrinkled enamel (Deng, 2007). The specialized dentition of Parelasmotherium was thought to be an adaptation to an abrasive high-fiber diet (Heissig, 1989a; Deng, 2007), suggesting that this rhino grazed on tough grasses (Deng, 2007). After the short cooling event at the beginning of the Late Miocene, there was a rapid recovery of rhinocerotid abundance, indicating a warming climate. During this stage, the tribes Rhinocerotini and Elasmotheriini declined in numbers and the Chilotheriini, including Chilotherium and Acerorhinus, became dominant (Deng and Downs, 2002). Chilotherium has

been recovered from Linxia Basin deposits aged at 9.5, 9, 7.5, 7, and 6 Ma and from 9 Ma Tianshui Basin deposits. Chilotherium wimani had a wide mandibular symphysis, huge lower incisors with upturned medial flanges forming tusks, hypsodont cheek teeth, and relatively small, well-worn premolars (Deng and Downs, 2002; Deng, 2006). These characters suggest that Chilotherium was a grazer (Deng and Downs, 2002). Heissig (1999) indicated, as well, that the increased height of the molars and relatively small size of the premolars in Chilotherium were an indication of a progressive specialization for grazing. The aceratheres and chilotheres had short and robust limb bones that were not adapted for cursorial locomotion and Ch. wimani had limb bones that were shorter and less slender than most genera within the subfamily Aceratheriinae (Deng, 2002; Deng and Downs, 2002). Therefore, Ch. wimani was most likely a woodland inhabitant. Heissig (1989a) noted that although Chilotherium had hypsodont dentition, there was no sign of neck bending as seen in other grazing rhinoceroses. He also noted that because Chilotherium had enlarged incisors, as opposed to horns, it must have required a horizontal positioning of the head, and therefore, could only graze by shortening the limbs to lower the head to the ground. It was concluded that Chilotherium may not have been a true grazer. The elasmothere Iranotherium lived in the Linxia Basin at 9.2 Ma. Iranotherium was a large rhinocerotid with a huge nasal horn and hypsodont teeth with a cement cover and filling and slightly wrinkled enamel (Deng, 2005a). Like all other elasmotheres, the specialized dentition of Iranotherium was thought to be an adaptation to an abrasive diet. Deng (2005a) suggested that, according to the faunal composition of the Liushu Formation, Iranotherium lived in an open steppe habitat. In addition, a pollen analysis of the red clay of Liushu Formation indicated a significant increase in grasses during the Late Miocene and showed that the floral composition of the Liushu Formation is consistent with that expected in a more open habitat (Ma et al., 1998). The chilothere Acerorhinus lived concurrently with Chilotherium in the Linxia Basin at 9 Ma. Acerorhinus had shortened limb bones that were more massive than those of Chilotherium (Heissig, 1989a; Deng, 2002), indicating a woodland habitat for that genus (Deng and Downs, 2002). It has been suggested that Chilotherium had a diet rich in grasses, while Acerorhinus remained a brush feeder with an increasing adaptation toward tough, dry vegetation (Heissig, 1999). Deng and Downs (2002) considered Acerorhinus a browser, while Heissig (1989a) classified Acerorhinus as a grazer.

Dicerorhinus ringstroemi, belonging to the tribe Rhinocenotini, lived concurrently with Chilotherium in the Linxia Basin at 6 Ma. Dicerorhinus had a cursorial limb structure and hypsodont dentition, but because of its gigantic size, it is considered to have been a woodland dweller (Guérin, 1980; Deng and Downs, 2002). Heissig (1989a) classified Dicerorhinus as a browser, although it had a lowered head and lost incisors. The great diversity of rhincerotids during the Late Miocene indicates a warm and moist environment for that time period. Even so, an extinction event occurred at the end of the Late Miocene causing a decline in rhinocerotid diversity and Dicerorhinus ringstroemi was the only known survivor in China (Deng and Downs, 2002). Bulk carbon and oxygen isotope compositions of tooth enamel from five 11.5 Ma Parelasmotherium individuals, four 9.5 Ma Chilotherium individuals, one 9.2 Ma Iranotherium individual, one 9 Ma Acerorhinus individual, seven 9 Ma Chilotherium individuals, six 7.5 Ma Chilotherium individuals, two 7 Ma Chilotherium individuals, one 6 Ma Dicerorhinus individual, and five 6 Ma Chilotherium individuals were determined (Figures 4.3 and 4.4, respectively). The group of 11.5 Ma Parelasmotherium individuals had a mean bulk δ13C value of -9.9 ‰, a δ13C range of 0.3 ‰ (-9.8 ‰ max., -10.1 min.), a mean bulk δ18O value of -4.6 ‰, and a δ18O range of 0.2 ‰ (-4.5 ‰ max., -4.7 min.). The carbon isotope results indicate that Parelasmotherium had a pure C3 diet. The small δ13C range (Figure 4.5) suggests that this rhino was a very selective feeder, which is consistent with its specialized dentition. The small δ18O range (Figure 4.5) for Parelasmotherium may indicate that this rhino either received most of its water intake from lakes or lived in a forested environment without much migration between habitats. Because of its large size, Parelasmotherium may have spent a lot of time wallowing in lakes, which is consistent with a small δ18O range. The δ18O values of Parelasmotherium were higher than those of most rhinos in this study, which may indicate that Parelasmotherium lived in a relatively open C3 grassland environment, such as a savannah. On the other hand, because there are no other contemporary rhinos to compare with this genus and because co-existing horses from the Linxia Basin had even higher δ18O values (Wang and Deng, 2005), it is not possible to rule out a forested habitat for Parelasmotherium at 11.5 Ma. Serial carbon and oxygen isotope analyses were performed on two Parelasmotherium individuals (Figure 4.7). The serial carbon isotope results showed very negative δ13C values (δ13C < -8 ‰) with little variation throughout the growth periods of the teeth for both individuals 78

0 -2 -4 -6 Chilotherium -8 6 Ma -10 Δ18O = 2.2 -12 (Ds-11) -14 0 5 10 15 20 25 30 35 40 45 50 55 0 -2 -4 -6 Chilotherium -8 7.5 Ma -10 Δ18O = 2.1 (Hl‐14) -12 (Hl-14) (Hl-15) Δ18O = 2.0 (Hl‐15) -14 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35 0 -2 (‰, V-PDB) -4 -6 Iranotherium -8 9.2 Ma -10 Δ18O = 1.4 -12 (Ls-06) -14 0 5 10 15 20 25 30 35 40 45 50 55 0 Values Delta -2 -4 -6 Chilotherium -8 9.5 Ma -10 Δ18O = 2.7 (Qj‐07) -12 (Qj-07) (Qj-09) Δ18O = 2.1 (Qj‐09) -14 0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 -2 -4 -6 Parelasmotherium -8 11.5 Ma -10 Δ18O = 3.6 (Gn‐13) -12 (Gn-13) (Gn-14) Δ18O = 2.5 (Gn‐14) -14 0 5 10 15 20 25 30 35 40 45 0 5 10 15 20 25 30 35 40 45 18 Distance from Crown (mm) δ O δ13C

Figure 4.7. Serial carbon and oxygen isotope compositions of rhino tooth enamel from the Late Miocene. Open diamonds represent oxygen isotope compositions and closed diamonds represent carbon isotope compositions. The data labels in parentheses are the sample numbers of rhino individuals. Serial samples represent decreasing age with distance from the crown. Δ18O = the amplitude of the δ18O seasonal cycles.

(Δ13C < 1 ‰; Table 4.3), which suggests there was little or no seasonal variation in their diets and is consistent with both rhinos having pure C3 diets. The mean serial δ13C values are identical for both individuals (-10.1 ‰), indicating very similar specialized diets. The seasonal cycles in the δ18O curves for the two individuals have irregular frequencies and there is a 1.1 ‰ difference in amplitude between the two rhinos. The irregular cycles suggest that Parelasmotherium lived in an open environment and had multiple water sources. The δ18O values are higher than those of most rhinos in this study, which is also consistent with an open habitat. The difference in cycle amplitude between the two individuals is likely a result of each individual having a variety of, and not necessarily the same water sources. This data is consistent with inferences made, based on bulk C and O isotope data, regarding the diet and habitat of Parelasmotherium. This data also supports that the Linxia Basin became relatively warmer and/or drier in the early Late Miocene. The 9.5 Ma Chilotherium individuals had a mean bulk δ13C value of -11.1 ‰, a δ13C range of 0.7 ‰ (-10.8 ‰ max., -11.5 min.), a mean bulk δ18O value of -9.4 ‰, and a δ18O range of 0.8 ‰ (-9.0 ‰ max., -9.8 min.). Again, the carbon isotope results indicate a pure C3 diet for this rhino. The small δ13C range (Figure 4.5) suggests that this rhino was a selective feeder, which is consistent with a specialized dentition. The small δ18O range (Figure 4.5), along with the very negative δ13C and δ18O values, implies that Chilotherium was a forest-dweller that may have spent most of its time wallowing in waterholes. If Chilotherium at 9.5 Ma was a grazer, it must have grazed on C3 grasses located within a relatively closed environment. Serial carbon and oxygen isotope analyses were performed on two 9.5 Ma Chilotherium individuals (Figure 4.7). The serial carbon isotope results revealed δ13C values < -8 ‰ with almost no variation throughout the growth periods of the teeth for both individuals (Table 4.3), which implies little or no seasonal variation in their diets and is consistent with both rhinos having pure C3 diets. The mean serial δ13C values are almost identical for both individuals indicating very similar specialized diets. The seasonal cycles in the δ18O curves for the two individuals have regular frequencies and the amplitude is very similar for both rhinos. The regular cycles, along with very negative mean δ18O values (~-10 ‰) for both individuals, suggest that Chilotherium at 9.5 Ma lived in forested habitats and both rhinos had similar water sources. This data is consistent with conclusions, based on bulk C and O isotope data, regarding the paleoecology of Chilotherium at 9.5 Ma.

The 9.2 Ma Iranotherium individual had a bulk δ13C value of -11.0 ‰ and a bulk δ18O value of -3.1 ‰. Based on the carbon isotope results, this rhino was a pure C3 feeder. Because Iranotherium had a dentition that was specialized for an abrasive diet, it can be speculated that this rhino fed on C3 grasses. The high δ18O value may suggest that Iranotherium drank in an open environment, where δ18O values were relatively high due to evaporative processes. Alternatively, the high δ18O values could indicate a change in regional climate towards warmer or drier conditions. These isotopic results are consistent with faunal and pollen analyses that suggest the Liushu Formation was an open steppe habitat during the Late Miocene. Serial carbon and oxygen isotope analyses were performed on one 9.2 Ma Iranotherium individual (Figure 4.7). The serial carbon isotope results showed very negative δ13C values (δ13C < -8 ‰) with little variation throughout the growth period of the tooth for this individual (Δ13C < 1 ‰; Table 4.3), suggesting little or no seasonal variation in this rhino’s diet. This is consistent with a specialized and strictly C3 diet. Seasonal cycles are not evident in the δ18O curve for this individual and, thus, there is very little variation in δ18O values (Δ18O = 1.4). The absence of seasonal cycles, combined with high δ18O values, may suggest that Iranotherium drank from a large body of water, such as a lake, in an open environment with relatively high amounts of evaporation. The δ18O values are higher than those of all other rhinos in this study, which is consistent with an open habitat, although a shift in meteoric water δ18O due to climate change (e.g., a change in moisture sources, higher temperatures, or increased aridity) could be responsible for the higher δ18O values at 9.2 Ma. The mean bulk and serial δ13C values for this individual are very negative (-11.0 and -11.6 ‰, respectively), which, conversely, suggests a forested habitat for Iranotherium. Because it was inferred from previous morphological studies that Iranotherium likely inhabited an open steppe environment, it is not expected that the negative δ13C values are due to Iranotherium living in a closed habitat. Instead, the carbon isotope data may suggest that the region was less water-stressed at 9.2 Ma than at 9.5 Ma. However, it is not possible at this time to determine whether or not this C and O isotope data support previous hypotheses regarding the habitat of Iranotherium. Analyses of additional Iranotherium individuals and other contemporary genera, if they become available, could help to elucidate the paleoecology of this genus.

The 9 Ma Acerorhinus individual had a bulk δ13C value of -9.6 ‰ and a bulk δ18O value of -6.3 ‰. The 9 Ma Chilotherium individuals from Linxia Basin had a mean bulk δ13C value of -11.3 ‰, a δ13C range of 0.2 ‰ (-11.2 ‰ max., -11.4 min.), a mean bulk δ18O value of -7.3 ‰, and a δ18O range of 3.0 ‰ (-5.8 ‰ max., -8.8 min.). The 9 Ma Chilotherium individuals from Tianshui Basin had a mean bulk δ13C value of -8.5 ‰, a δ13C range of 0.8 ‰ (-8.1 ‰ max., -8.9 min.), a mean bulk δ18O value of -6.7 ‰, and a δ18O range of 0.8 ‰ (-6.4 ‰ max., -7.2 min.). According to the carbon isotope results, both genera apparently had a pure C3 diet, although Acerorhinus had a more 13C-enriched diet than that of Chilotherium from Linxia Basin. This suggests that, in the Linxia Basin, Acerorhinus was feeding in a more open environment than that of Chilotherium. The small δ13C range of Chilotherium from Linxia Basin (Figure 4.5) indicates a specialized diet, which, again, is supported by a specialized dentition. The larger δ13C range of Chilotherium from Tianshui Basin (Figure 4.5) indicates a more generalized diet for those individuals compared to Chilotherium from Linxia Basin, which could be due to differences in the environments of those two populations or a result of Chilotherium from Tianshui Basin facing greater competition from other taxa. The higher δ18O value of the Acerorhinus individual (compared to contemporary Chilotherium) implies that it drank in an open habitat, which is consistent with the δ13C data. The large δ18O range of the Chilotherium individuals from Linxia Basin (Figure 4.5) suggests that Chilotherium migrated between forested habitats or had water sources in both forested and relatively open habitats. The smaller δ18O range of the Chilotherium individuals from Tianshui Basin, along with the relatively high δ13C and δ18O values, suggest that these rhinos inhabited a more open environment than did Chilotherium from Linxia Basin. Overall, this data suggests that Acerorhinus and Chilotherium from Tianshui Basin were C3 grazers in relatively open habitats compared to that of Chilotherium from Linxia Basin, and again, if Chilotherium from Linxia Basin was a grazer, it must have grazed on C3 grasses within a more forested environment. Additionally, there were no significant differences in δ13C and δ18O values between Chilotherium from the Linxia Basin at 9.5 and 9 Ma (Table 4.2), indicating no significant change in habitat for that genus between those time intervals. The group of 7.5 Ma Chilotherium individuals had a mean bulk δ13C value of -10.0 ‰, a δ13C range of 1.2 ‰ (-9.6 ‰ max., -10.8 min.), a mean bulk δ18O value of -8.9 ‰, and a δ18O

range of 1.2 ‰ (-8.3 ‰ max., -9.5 min.). The 7 Ma Chilotherium individuals had a mean bulk δ13C value of -11.5 ‰, a δ13C range of 0.2 ‰ (-11.4 ‰ max., -11.6 min.), a mean bulk δ18O value of -9.9 ‰, and a δ18O range of 2.2 ‰ (-8.8 ‰ max., -11.0 min.). The carbon isotope results show that the diets of the Chilotherium individuals at 7.5 Ma were 13C-enriched relative to Chilotherium at 9 and 7 Ma. These significant differences in δ13C values (p=0.0270 and 0.0168, for Chilotherium at 7.5 Ma compared to Chilotherium at 9 Ma and at 7 Ma, respectively; Table 4.2) are probably due to increased water-stress in plants at 7.5 Ma, which is consistent with the previously published horse oxygen isotope record from the Linxia Basin, which suggests the Linxia Basin was relatively warmer and/or drier at 7.5 Ma compared to 9 and 7 Ma (Wang and Deng, 2005). Because there were no significant differences in δ18O values between Chilotherium at 9, 7.5, and 7 Ma (Table 4.2), the O-isotope data does not support that Chilotherium inhabited a more open environment at 7.5 Ma compared to that at 9 and 7 Ma. Nevertheless, if the higher δ13C values of Chilotherium at 7.5 Ma are due to water-stress, a more open environment is implied for that age. The δ18O values of Chilotherium decrease from 9 to 7.5 Ma and then further decrease from 7.5 to 7 Ma. (Figure 4.4), which is not entirely consistent with the shift observed in the δ18O values of horses from the Linxia Basin, as the horse δ18O values increase from 9 to 7.5 Ma and then decrease from 7.5 to 7 Ma (Wang and Deng, 2005). This difference in direction of the δ18O shift between horses and rhinos from 9 to 7.5 Ma may be due to changes in the behavior of rhinos during that time interval. For example, Chilotherium at 7.5 Ma may have spent more time wallowing than did Chilotherium at 9 Ma, which could have caused the O-isotope composition of this rhino to be more negative at 7.5 Ma even though the climate was warmer and/or drier at that age compared to that at 9 Ma. The large δ13C range of Chilotherium individuals at 7.5 Ma (Figure 4.5), compared to those at 7 and 9 Ma, suggest that this rhino was a less selective feeder at 7.5 Ma, which, again, could be due to a warmer environment that would allow more species of plants to be available or a means to survive amongst competition from other large mammals. The small δ13C range of Chilotherium individuals at 7 Ma (Figure 4.5) is consistent with this rhino having a very specialized dentition. The relatively large δ18O range of Chilotherium at 9 and 7 Ma compared to that of Chilotherium at 7.5 Ma (Figure 4.5) suggests that Chilotherium at 7.5 Ma may have migrated less between forested habitats than did Chilotherium at 9 and 7 Ma, or had fewer water sources at 7.5 Ma,

which supports the climate being warmer and/or drier during that time interval. By and large, the carbon and oxygen isotope results for Chilotherium at 9.5, 9, 7.5, and 7 Ma, along with previous morphological data, suggest that this genus was a forest-dweller that grazed on C3 grasses, but had the ability to adapt to a changing environment. Serial carbon and oxygen isotope analyses were performed on two Chilotherium individuals at 7.5 Ma (Figure 4.7). The serial carbon isotope results showed δ13C values < -8 ‰ with little variation throughout the growth periods of the teeth for both individuals (Δ13C < 1 ‰; Table 4.3), suggesting that there was little or no seasonal variation in their diets and that both rhinos had pure C3 diets. The mean serial δ13C values are identical for both individuals (-9.5 ‰), indicating very similar specialized diets. The δ13C values of these 7.5 Ma rhinos are higher than those of Chilotherium at 9.5, which indicates that Chilotherium adjusted to a more water-stressed environment at 7.5 Ma. The seasonal cycles in the δ18O curves for the two individuals are almost identical in amplitude and frequency, suggesting similar habitats and water sources for the two individuals. The regular cycles in the δ18O curves of the two individuals suggest that these rhinos inhabited a closed or forested environment, but because the δ18O values of the 7.5 Ma rhinos are higher than those of Chilotherium at 9.5, it can be inferred that the climate became warmer and/or more arid by 7.5 Ma. This data is consistent with inferences made, based on bulk C and O isotope data, regarding the diet and habitat of Chilotherium and supports that the Linxia Basin became relatively warmer and/or drier at 7.5 Ma. The 6 Ma Dicerorhinus individual had a bulk δ13C value of -9.2 ‰ and a bulk δ18O value of -4.6 ‰. The group of 6 Ma Chilotherium individuals had a mean bulk δ13C value of -9.4 ‰, a δ13C range of 1.0 ‰ (-8.9 ‰ max., -9.9 min.), a mean bulk δ18O value of -5.8 ‰, and a δ18O range of 2.7 ‰ (-4.6 ‰ max., -7.3 min.). As with all other rhinos analyzed in this study, the carbon isotope results indicate that Dicerorhinus and Chilotherium at 6 Ma had pure C3 diets. The δ13C and δ18O values of Chilotherium at 6 Ma are significantly higher than those of Chilotherium at 7 Ma (Table 4.2), suggesting increased water-stress in plants or more open habitats at 6 Ma. This is consistent with the warming/drying trend observed in horse and rhino mean δ18O values (Wang and Deng, 2005; Biasatti et al., manuscript in preparation) and supports that the Linxia Basin became more open and arid through time. The δ13C and δ18O values of Dicerorhinus were very similar to those of Chilotherium at 6 Ma, which suggests that these two

84 genera had similar diets and habitats. The oxygen isotope composition of Dicerorhinus was slightly 18O-enriched compared to that of Chilotherium at 6 Ma, indicating that Dicerorhinus was also drinking water in an open environment. In fact, the δ18O values of these two genera at 6 Ma are similar to the δ18O values of rhinos such as Hispanotherium, Parelasmotherium, and Iranotherium, which are considered grazers that lived in an open steppe habitats. Along with increased δ18O values, Chilotherium at 6 Ma also had a large δ18O range (Figure 4.5), indicating multiple water sources or migration between open habitats. Chilotherium at 6 Ma had a larger δ13C range than did Chilotherium at 7 Ma (Figure 4.5), indicating a more generalized diet for this genus at 6 Ma. The δ13C range of Chilotherium at 6 Ma is very similar to that of Chilotherium at 7.5 Ma (Figure 4.5), where warming temperatures could have increased the availability or diversity of plants or where increased temperatures could have forced this rhino to become a more generalized feeder due to increased competition for resources. On the whole, Chilotherium was able to adapt to increasingly arid conditions in the Linxia Basin and apparently moved from a forested habitat at 9 Ma to an open steppe dwelling at 6 Ma. The carbon and oxygen isotope data do not support that Dicerorhinus was a woodland dweller that browsed, but supports that this rhino was likely grazing on C3 grasses in an open steppe environment. Serial carbon and oxygen isotope analyses were performed on one 6 Ma Chilotherium individual (Figure 4.7). The serial carbon isotope results showed very negative δ13C values (δ13C < -8 ‰) with little variation throughout the entire tooth growth period (Δ13C < 1 ‰; Table 4.3). This suggests that there was little or no seasonal variation in its diet and the data are also consistent with a specialized pure C3 diet. The mean serial δ13C value of the 6 Ma individual (- 9.2 ‰) is very similar to the mean δ13C values of the two 7.5 Ma Chilotherium individuals (-9.5 ‰), indicating very similar diets. The seasonal cycles in the δ18O curve for the 6 Ma individual are irregular, indicating a more open habitat with multiple water sources for Chilotherium at 6 Ma. The amplitude of the δ18O curve for the 6 Ma individual is almost identical to those of the 7.5 Ma rhinos, indicating that all three rhinos may have experienced similar seasonality. The mean δ18O value of the 6 Ma rhino is ~4 ‰ higher than that of either Chilotherium individual at 7.5, suggesting that the climate became relatively warmer and/or drier by 6 Ma. Again, Chilotherium apparently adapted to a changing climate. This data is consistent with inferences made, based on bulk C and O isotope data, regarding the diet and habitat of Chilotherium.

4.4.4. Pliocene Rhinoceroses The aceratherine rhinocerotid Shansirhinus ringstroemi was recovered from the red clay of the Early Pliocene Hewangjia Formation (Deng, 2005b), and the rhinocerotine Coelodonta nihowanensis was collected from the Late Pliocene Wucheng Loess deposits (Deng, 2008). S. ringstroemi is the only known rhinoceros species from the Early Pliocene deposits of the Linxia Basin. Shansirhinus had high-crowned teeth with strong wear and well-developed secondary folds and enamel plications, as well as a premolar morphology that is unique to horses and certain rhinos with grazing adaptations. These traits would allow the teeth of Shansirhinus to resist the abrasion of a high-fiber diet and, therefore, indicate that Shansirhinus probably grazed on tough grasses (Deng, 2005b). Even so, Qiu and Yan (1982) considered S. ringstroemi (= Ch. cornutum) to be a browser, feeding on soft twigs and leaves. Based on the faunal composition of the Hewangjia Formation, which includes rodents, perissodactyls, and artiodactyls, it has been suggested that the Linxia Basin had an open and more arid ecological environment during the Early Pliocene compared to the Miocene (Deng, 2005b). The woolly rhino, Coelodonta nihowanensis, from the Linxia Basin is the earliest known member of this genus. Because the earliest known occurrence of the woolly rhino is in northwest China at 2.5 Ma, it is apparent that the presence of Coelodonta in Europe and North Asia is a result of dispersal from northwestern China (Deng, 2008). The woolly rhino had a flattened nasal horn that was transversely banded. It has been suggested that the bands represent annual growth zones, implying strong seasonality in the Linxia Basin during the Late Pliocene (Fortelius, 1983). The presence of Coelodonta, a typical glacial mammal, in the Linxia Basin indicates a cold climate during the Late Pliocene, which would be expected with the onset of the Northern Hemisphere Glaciation. Coelodonta nihowanensis was smaller and had a more cursorial limb structure than did the more derived species of Coelodonta (Deng, 2008). In addition, Coelodonta had moderately hypsodont dentition and probably used its wide muzzle and flat nasal horn to forage on grasses through the snow (Deng and Downs, 2002). Bulk carbon and oxygen isotope compositions of tooth enamel from one Shansirhinus individual and one Coelodonta individual were determined (Figures 4.3 and 4.4, respectively). The 4 Ma Shansirhinus individual had a bulk δ13C value of -9.8 ‰ and a bulk δ18O value of -5.7 ‰. The 2.5 Ma Coelodonta individual had a bulk δ13C value of -9.9 ‰ and a bulk δ18O value of

-6.5 ‰. The carbon isotope results indicate that both rhinos had pure C3 diets that were similar to those of apparent grazing rhinos that lived during the Miocene in the Linxia Basin and were 13C-enriched compared to rhinos that are thought to have been forest-dwellers. The δ18O values of Shansirhinus and Coelodonta are high, suggesting that these rhinos drank water in an open environment. The oxygen isotope composition of Shansirhinus is slightly 18O-enriched compared to that of Coelodonta, which would be expected, as temperatures were cooler in the Linxia basin during the Late Pliocene due to the Northern Hemisphere Glaciation. The carbon and oxygen isotope data is very limited for rhinos during the Pliocene, but overall, the data supports that both Coelodonta and Shansirhinus were C3 grazers that lived in an open steppe environment that became colder throughout the Pliocene.

4.5. Conclusions

Carbon and oxygen isotope analysis of tooth enamel is a useful and important tool in the study of the paleoecologies of ancient mammalian herbivores and can be utilized to test previous investigations of various taxa based on morphological or taxonomical studies. The bulk and serial C and O isotope results for rhinocerotoids that lived in the Linxia Basin from 25 to 2.5 Ma supported most previous hypotheses that were formed from analyses of taxonomic and cranial and limb morphological characteristics and allowed new insight on some genera. The isotope data supported that Paraceratherium at 25 Ma lived in a closed forested environment and spent a great amount of time wallowing in water or mud, which is consistent with the gigantic size of this rhino and that its dentition was apparently specialized for consuming tree leaves. It also supported that Allacerops, who coexisted with Paraceratherium at 25 Ma, lived in a relatively open habitat and had a less specialized diet. The data suggested that the Early Miocene rhinos Alicornops and Hispanotherium had dissimilar diets and habitats, and were consistent with previous hypotheses that Hispanotherium was a grazer in an open steppe environment, whereas Alicornops was a more generalized feeder in a forested environment. The isotope data were consistent with previous inferences that the Late Miocene rhino Parelasmotherium grazed and dwelled in an open steppe habitat, but were inconsistent with previous hypotheses that the rhinos Acerorhinus and Dicerorhinus inhabited closed forested environments. Rather, the results indicated that these two rhinos inhabited open-steppe regimes.

The isotopic results were not conclusive in regard to the habitat of Iranotherium, but supported previous hypotheses of a specialized C3 grass diet for Iranotherium. The results also suggested that the rhino Chilotherium dwelled in a forested environment throughout most of the Late Miocene, but inhabited a more open environment by the end of the Late Miocene, indicating that Chilotherium was able to adapt to a changing environment. It was not possible to determine from the isotopic results whether Chilotherium was grazing or browsing within the forest. Finally, the results were consistent with previous hypotheses that the Pliocene rhinos Shansirhinus and Coelodonta were grazers that lived in an open habitat and support that the climate was cooling during much of the Pliocene. Overall, the oxygen isotope data imply a general warming and/or drying trend in the Linxia Basin from the Late Oligocene to Late Miocene. The serial oxygen isotope data show almost no changes in seasonality in the Linxia basin from 25 to 6 Ma, indicating that climate in the region did not become strongly influenced by the Asian monsoon prior to 6 Ma. Furthermore, the carbon isotope data support that all rhinos in this study were pure C3 feeders and indicate that C4 grasses were not an important component of the plant biomass in the Linxia Basin prior to 2.5 Ma, which is contrary to the global C4 grass expansion at ~7 Ma.

CHAPTER 5

PALEOECOLOGIES AND PALEOCLIMATES OF CENOZOIC MAMMALS FROM YUNNAN PROVINCE, CHINA, BASED ON STABLE CARBON AND OXYGEN ISOTOPES

5.1. Introduction

Stable carbon and oxygen isotopic compositions of fossil mammalian tooth enamel and ancient soils collected throughout Yunnan Province, China, provide a means of understanding of regional climate change and the evolution of mammalian species along the southeastern margin of the Tibetan Plateau. Carbon and oxygen isotopes in fossil tooth enamel and paleosols serve as records of paleoecology and paleoclimate in terrestrial ecosystems. Carbon isotopes in tooth enamel or soil reflect the types of plants (C3, C4, or CAM) ingested by an herbivore or that grew in the local soil, respectively (e.g., Cerling et al., 1989; Koch, 1998; Kohn and Cerling, 2002), and also reflect changes in relative humidity, as plant carbon isotope compositions are affected by water availability (Farquhar et al., 1989). Carbon isotope compositions of tooth enamel and soil are also influenced by soil respiration rates, which vary with ecosystem type (i.e., closed forest versus open grassland; Schleser and Jayasekera, 1985; Sternberg et al., 1989; van der Merwe and Medina, 1989). As a result, carbon isotope analyses allow insight into the feeding behaviors and habitat preference of particular fossil taxa. Oxygen isotopes in mammalian tooth enamel or soil carbonates reflect the oxygen isotopic composition of local meteoric water ingested by an animal or that was present in the region, respectively (Longinelli, 1984; Luz et al., 1984; Ayliffe and Chivas, 1990). Because the oxygen isotopic composition of meteoric water is controlled by climate (Dansgaard, 1964; Rozanski et al., 1992), oxygen isotopes in fossil tooth enamel or paleosols are proxies for paleoclimate (e.g., Ayliffe and Chivas 1990; Quade et al., 1992; Ayliffe et al., 1994; Bryant et al., 1994; Kohn and Cerling, 2002). Previous studies, based on carbon isotope compositions of fossil mammalian tooth enamel and paleosols, suggest that terrestrial ecosystems consisted predominantly of C3 plants

prior to the Late Miocene. Following the Late Miocene, between ~7 and 5 Ma, C4 grasses became a significant component of low- to mid-latitude and low-elevation regions around the world (Quade et al., 1989; Cerling, 1992; Cerling et al., 1993, 1997a,b; MacFadden et al., 1994; MacFadden and Cerling, 1994; Wang et al., 1994). It was hypothesized that this global spread of

C4 plants occurred in response to declining atmospheric CO2 levels (Cerling et al., 1993, 1997a,b; Wang et al., 1994) and/or a strengthening of the Asian summer monsoon due to the uplift of the Tibetan Plateau (Quade et al., 1989). Others have suggested that there was no global expansion of C4 plants (Morgan et al., 1994) and that there is a lack of evidence supporting

decreasing CO2 levels during the Late Miocene (Pagani et al., 1999; Retallack, 2001; Royer et al., 2001). In this study, the carbon and oxygen isotopic compositions of tooth enamel samples from 164 individual fossil herbivore teeth and 8 paleosol carbonates from Yunnan Province, ranging in age from 10 Ma to the present, were determined in order to reconstruct the paleodiets, paleoecologies, and paleoenvironments of extinct taxa from the region and to test previous hypotheses regarding the development of C4 ecosystems and the dynamics of the Asian monsoons. By comparing the results of this study to records from other localities along the margin of the Tibetan Plateau, we may gain more understanding of the effects of the uplift of the Tibetan Plateau on regional and global climate. In addition, the carbon isotope compositions of 9 modern plants from Yunnan Province were determined for comparison with plant compositions of ancient ecosystems.

5.2. Study Site

Yunnan Province is located in southwestern China, along the southeastern margin of the Tibetan Plateau (Figure 5.1). It is bounded by Myanmar (Burma) to the west, by Vietnam and Laos to the south, and merges with the Tibetan Plateau to the northwest. Yunnan is a mountainous region and the elevations of many of its mountain ranges toward the northwest exceed 4000 m. Toward the east of the province, a broad plateau with a more gentle relief is formed. The Yunnan Plateau lies just north of the Tropic of Cancer at an elevation of ~2000 m (Hodell et al., 1999; Leloup et al., 1995). Yunnan spans three tectonic domains: the Yangzi paraplatform to the northeast, the Sichuan-Eastern Tibet fold belts to the northwest, and the

CHINA

Figure 5.1. Map of Yunnan Province. Yunnan Province is located on the southeastern margin of the Tibetan Plateau. Samples were collected from six localities within the province: Baoshan, Kaiyuan, Lufeng, Shangri La, Yuanmou, and Zhaotong.

Indochina block to the south, with the Red River Fault zone separating the latter from the first two domains (Leloup et al., 1995). Seasonal rainfall in Yunnan is strongly influenced by the Indian monsoon. Most precipitation occurs between the months of May and October under the influence of a strong southwesterly summer monsoon that originates from the Bay of Bengal. From November to March, during the winter monsoon, winds shift to the N-NW, bringing dry air and very little rainfall by eastward-moving frontal disturbances. Evidence has shown that monsoon circulation strongly affects the amount and the oxygen isotopic composition of precipitation in Kunming, the provincial capital, and the δ18O values of summer precipitation can be ~15 ‰ more negative than those of winter rainfall (Hodell et al., 1999). Six fossil localities in Yunnan Province, ranging in age from 10 Ma to 9 Ka, were of interest for this study: (1) Baoshan: Baoshan Prefecture (25°N, 99°E) is located ~350 km west of Kunming (Figure 5.1) and has an elevation of ~1655 m. Mean annual temperature and rainfall are 15°C and 97.1 cm, respectively. For this study, fossils were collected from two sites within the Baoshan Prefecture, the Yangyi coal mine and the Tanzigou Neolithic site. The Yangyi locality is situated on the eastern slopes of the Gaoligongshan Mountains, between the Nu and Lancang Rivers, about 30 km south of the city Baoshan. The age of the Yangyi Formation has not been reliably established, but the fauna of the Yangyi coal mine correspond well with other Pliocene faunas from China, such as that from the Shagou Formation in the Yuanmou Basin, which has an estimated age of ~3-4 Ma based on magnetostratigraphy. The fauna also correlates with the Yushean-aged fauna from northern China, with an age of 5.2 to 2.6 Ma (Yun, 1975; Harrison et al., 2002). In addition, stratigraphic (Bureau of Geology and Mineral Resources of Yunnan Province, 1990; Ge and Li, 1999) and palynological evidence (Tang and Hu, 1993) suggests that the Yangyi deposits are Late Pliocene in age. Therefore, the Yangyi deposits were tentatively assigned an age of ~3-5 Ma. The coal-bearing sediments of the Yangyi Formation consist of lignite, mudstone, sandstone, and conglomerate (Bureau of Geology and Mineral Resources of Yunnan Province, 1990). Fossils associated with the Yangyi site include Lufengpithecus sp., Sinoadapis sp., Zygolophodon sp., Chilotherium yunnanensis, Stegolophodon yanyiensis, Stegodon elephantoides, and Axis sp. (Jablonski et al., 2003). The Tanzigou Neolithic site is located ~30 km southwest of the city Baoshan and has been assigned an age of ~8-10 Ka, based

on the discovery of a partial human skull and a few roughly crafted stone and bone tools (Gen and Zhang, 1992). Associated fossil remains include fresh water mussel shells, water fowl, fish, otters, several species of deer and muntjac, sheep, rhinoceroses, monkeys, ferret-badgers, rodents, and bats (Greenman, 2003; Jablonski et al., 2003). (2) Kaiyuan: The city of Kaiyuan (23.7°N, 103.2°E), is located ~210 km south of the provincial capital, Kunming (Figure 5.1), and has an elevation of 1030-1110 m. The site of interest, the Xiaolongtan coal mine, is an approximately 21-km2 northeast-trending ellipse that is situated northwest of the city of Kaiyuan (Dong, 1987; Harrison et al., 2002). Exposed sediments at the Xiaolongtan coal mine consist of (from top to bottom): an approximately 100 m-thick top unit of gray-white moderately massive to massive marls containing partial calcarenites bearing plant fragments and invertebrates; an approximately 250 m-thick upper unit of black and black- brown thin to moderately stratified lignites grading to gray-white and tan mudstones with fossil mammals and reptiles found in the top of this unit; and an approximately 100 m-thick lower unit of light yellow, gray-yellow, yellow-brown, and light gray mudstones, clays, and sandy gravels (Dong, 1987; Zhang et al., 1999). Fauna associated with the Xiaolongtan coal system are currently considered to be slightly older than that from the Lufeng locality and are correlated to the upper part of the Lower Siwaliks, or the Chinji Formation (~10 Ma; Zhang et al., 1999; Harrison et al., 2002). Paleomagnetic studies indicate an age of ~8.3 Ma (Yin, 1994), which is younger than that inferred from the mammalian fauna. Vertebrate fossils from the Xiaolongtan coal system include Dryopithecus kaiyuanensis, Chleuastochoerus cf. stehlini, Zygolophodon chinjiensis, Gomphotherium xiaolongtanensis, Gomphotherium cf. macrognathus, Potamochoerus sp., Hippopotamus sp., Rhynchotherium sp., and Listriodon sp. (Zhang et al., 1999). (3) Lufeng: The site of interest, Shihuiba (25°N, 102°E), is located ~60 km northwest of the provincial capital, Kunming (Figure 5.1), and ~9 km north of the town Lufeng on the outskirts of the small village Shihuiba and has an elevation of ~1570 m (Badgley et al., 1988; Harrison et al., 2002). Mean annual temperature and rainfall are 15.5°C and 84 cm, respectively. The fossil-bearing sediments at Shihuiba consist of massive lignites, alternating with carbonaceous clays, silts, and sands with interbedded lignites (Qi, 1979, 1985, 1993; Harrison et al., 2002). The deposits are ~8 m thick and the outcrop is limited to a few hundred square meters (Badgley et al., 1988). The sequence consists of eight units and the sedimentology is described in

detail by Qi (1979, 1985, 1993), Chen (1986), and Badgley et al. (1988). More than 80 species of mammals have been recovered from the site at Shihuiba, as well as the fossil remains of birds, reptiles, fishes, mollusks, and pollen. Based on geological, floral, and faunal analyses, previous studies indicate that Lufeng was an upland forest community during the late Miocene and was probably more similar in climate to the Indian subcontinent than to northern China (Badgley et al., 1988). Based on faunal assemblages, Qui et al. (2000) suggest that Lufeng was not a grassland but was clearly a tropical forest. The Lufeng fauna is inferred to be younger than the Xiaolongtan fauna, as indicated by the more derived proboscideans Zygolophodon lufengensis and Tetralophodon cf. exoletus found at Shihuiba (Qiu and Qiu, 1995; Harrison et al., 2002). Additionally, the mammalian fauna indicate that the Lufeng fauna is equivalent to Turolian assemblages in Europe that are late Miocene in Age and a correlation to the biostratigraphic sequence of rhizomyid rodents from the Siwalik faunas of Pakistan suggest an age of ~8 Ma for the Lufeng fauna (Flynn and Qi, 1982; Li et al., 1984; Badgley et al., 1988). Paleomagnetic studies indicate that the hominoid-bearing lignites at Lufeng correlate to ~4 Ma (Yin, 1994). Harrison et al. (2002) imply that the date derived from magnetostratigraphy is incorrect, noting that the estimate is clearly too young to be concordant with the fauna. Vertebrate fossils recovered from the Shihuiba site include: Laccopithecus robustus, Sinoadapis shihuibaensis, Tetralophodon cf. exoletus, Zygolophodon lufengensis, Hipparion cf. nagriensis, Chilotherium sp., Aceratherium sp., Macrotherium salinum, Tapirus sp., Hyotherium sp., Hyotherium cf. palaeochoerus, Lophochoerus lufengensis, Potamochoerus, Dorcabune progressus, Yunnanotherium simplex, Moschus sp., Dicroceros sp., Metacervulus cf. simplex, Muntiacus cf. nanus, and Selenoportax sp. (Badgley et al., 1988). (4) Shangri La: The city of Shangri La (27.8°N, 99.6°E), also named Zhongdian, is located in northwest Yunnan Province on the southeastern margin of the Qinghai-Tibetan Plateau within the Deqin Tibetan Autonomous Prefecture (Figure 5.1). The Deqin Tibetan Autonomous Prefecture lies at the center of the Hengduan mountain range and has an elevation range of 1530 to 5545 m and the average altitude is ~3380 m. Mean annual temperature and rainfall are 5.5°C and 70 cm, respectively (Xie et al., 2001). The Deqin region consists of high mountains and deep valleys and the Lancang and Jinsha rivers run southward along the Hengduan Mountains, forming the famous gorges in the area. The range of altitudes in the region produces a great diversity of plants. The vegetation consists of subtropical dry and semi-dry in the lower river

94 valleys, open bush lands, evergreen broad-leafed forests and evergreen conifer forests in the temperate and sub-alpine zones, and alpine scrub and meadows, alpine screes, and periglacial vegetation in the frigid zone (Xie et al., 2001). The sites of interest are the Yie Ka Village (27°36’26.6”N, 99°42’20.6”E; elev. ~2903-2953 m) and Wuzhuang localities (27°45’91.3”N, 99°42’35.8”E; elev. ~3294 m), which are reported to be 2.5 and 1.75 Ma in age, respectively (Zong 1987; Zong et al., 1996). Fossil mammals collected at the Yie Ka Village and Wuzhuang localities include horses, deer, rhinos, bovids, and pigs (present study). (5) Yuanmou: The Yuanmou Basin (25.58°N, 102°E) is located ~110 km northwest of the provincial capital, Kunming (Figure 5.1), at an elevation of 1050 to1380 m. The Yuanmou Basin has a subtropical climate that is controlled by the Indian monsoon during the rainy season and by the continental tropical air mass during the dry season. Mean annual temperature and rainfall are 21.9°C and 61.1 cm, respectively (Qian, 1993). This north-south elongated basin is bordered by elevated regions of metamorphic rocks and granites of the Precambrian basement and by Jurassic to Cretaceous sediments. The basin is mainly filled by a thick series of Tertiary and Quaternary fluvial and lacustrine sediments that contain abundant mammalian fossils (Pan and Zong, 1993; Qian, 1993; Harrison et al., 2002; Zhu et al., 2002). Within the Yuanmou Basin, the Xiaohe and Yuanmou Formations are of interest for this study. The Late Miocene hominoid- bearing sediments of the Xiaohe Formation lie mainly in the northwestern section of the basin and consist of ~80 m of clays, silts, partially cemented fine sands, and gravels (Qian, 1993; Harrison et al., 2002). Within the Xiaohe Formation, fossil hominoids and many other fossil mammalians have been recovered from four localities: Baozidongqing (locality # 8601, 8602, 8603, 8604, 8605, 8606), Hudieliangzi (locality # 8701, 8702, 8704), and Fangbeiliangzi, or Gaipailiangzi, (locality # 8703, 8801, 8802, 9001, 9002) are located near the villages of Zhupeng and Xiaohe, while Dashuqingliangzi is located near Leilao village about 8 km southwest of Zhupeng (Harrison et al., 2002). The age and characters of the Xiaohe fauna are similar to those of the South Asian Siwaliks, the Lufeng fauna, and the Baode fauna (Johnson and Vondra, 1972; Flynn et al., 1990; Pan and Zong, 1993; Qian, 1993; Quade et al., 1995). Earlier magnetostratigraphic evidence suggested that the age of the Xiaohe Formation was ~4 to 5.5 Ma (Qian, 1993). The Xiaohe mammalian fauna are similar to the Lufeng fauna, but based on faunal assemblages, Ni and Qiu (2002) concluded that the Xiaohe fauna was slightly older than the Lufeng fauna, with an estimated age of ~9 Ma (Harrison et al., 2002). Recent paleomagnetic

evidence indicates that the age Xiaohe Formation should be ~7-9 Ma (Yin and Liang, 1998; Ni and Qiu, 2002). More than 40 species of mammals have been recovered from the Xiaohe Formation. Some were adapted to tropical habitats and many express adaptation to grasslands or dense forest biomes. Therefore, it has been suggested that the Yuanmou Basin must have had a warm climate and abundant rainfall, with thick forests as well as grasslands (Pan and Zong, 1993). Vertebrate fossils recovered from the Xiaohe Formation include: Sinoadapis sp., Laccopithecus sp., Stegolophodon banguoensis, Stegotetrabelodon primitivum, Hipparion platyodus, Hipparion sp., Macrotherium salinum, Chilotherium yunnanensis, Chilotherium sp., Potamochoerus sp., Muntiacus sp., Yunnanotherium sp., Cervavitus sp., and Cervus sp. (Pan and Zong, 1993). The Longchuan and Yuanmou Formations overly the Xiaohe Formation (Qian, 1993; Harrison et al., 2002). The Plio-Pleistocene sediments of the Yuanmou Formation consist of alternating sand, silt, clay, and gravel layers, and this formation has been divided into four members. Members 1 and 2 contain the Shagou fauna and have respective ages of 3.4 - 3.0 Ma and 3.0 - 2.5 Ma, and members 3 and 4 contain the Yuanmou fauna and have respective ages of 2.5 - 1.8 Ma and 1.8 -1.3 Ma, based on magnetostratigraphy (Pan and Zong, 1993; Qian, 1993). Within the Shagou fauna, several taxa were adapted to live in dense forests or riparian forest valleys, while artiodactyls typical of grasslands were rare. Therefore, it has been suggested that the Yuanmou Basin was covered with dense forests, while grasslands were insignificant during this time (Pan and Zong, 1993). Vertebrate fossils recovered from members 1 and 2 (the Shagou fauna) include: Stegolophodon stegodontoides progressus, Stegodon zhaotongensis, Stegodon yuanmouensis, Stegodon elephantoides, Stegodon sp., Rhinoceros sp., and muntiacus nanus (Pan and Zong, 1993). Within the Yuanmou fauna, member 3 is dominated by carnivores that are typical of dense forests, with Equus yunnanensis as a principle component, and member 4 has relatively few carnivores that are adapted to dense forests and an abundance of cervids and perissodactyls. Therefore, it has been suggested that member 3 represents a dense forest interspersed with sparsely-treed grassy planes and member 4 reflects a reduction in the dense forests and expansion of the grassy planes (Pan and Zong, 1993). Vertebrate fossils recovered from members 3 and 4 (the Yuanmou fauna) include: Stegodon elephantoides, Stegodon sp., Equus yunnanensis, Nestoritherium sp., Rhinoceros sp., Rhinoceros sinensis, Sus scrofa, Sus sp., Eostylocerus longchuanensis, Metacervulus capreolinus, Paracervulus attenuatus, Muntiacus lacustris, Cervoceros ultimus, Axis shansius, Axis sp., Bibos sp., and Gazella sp. (Pan and Zong,

1993). Homo erectus yuanmouensis (Yuanmou Man) was also recovered from member 4 and the sedimentary layer from which Yuanmou Man was collected has an age of ~1.7 Ma, based on paleomagnetic data (Pan and Zong, 1993; Qian, 1993). (6) Zhaotong: Zhaotong Basin (27°N, 103°E) is located in the northeastern region of Yunnan Province, about 300 km northeast of the provincial capital, Kunming (Figure 5.1). The average elevation of the basin is ~1959 m and mean annual temperature and rainfall are 12°C and 82 cm, respectively. Based on faunal assemblages, the fossil-bearing deposits of the Zhaotong Basin are believed to be Late Pliocene to Pleistocene in age (Chow and Zhai, 1962; Zhang et al., 1999), but additional records, such as faunal and magnetostratigraphic data, are needed in order to establish a more conclusive age. These “Plio-Pleistocene” deposits make up the Tuobuka Formation, which consists of loosely consolidated conglomerates grading to silts, clays, carbonates, and sandstones. The lower section of the formation contains ten lignite units, and the upper section contains numerous limonite zones (Zhang et al., 1999). Zhang et al. (1999) suggested that based on the conclusions of former workers (Colbert, 1943; Pei, 1961; Chow and Zhai, 1962), the age of the Tuobuka Formation at Zhaotong should be Early Pleistocene, or equivalent to the Yuanmou Formation, and that Pliocene sediments underlie the Tuobuka Formation, based on a discovery of the proboscidean Rynchotherium in the Zhaotong region (Chow and Zhang, 1974), which is believed to be Pliocene in age. Fossil mammalians recovered from this locality include: Felis sp., Equus cf. yunnanensis, Sus sp., Tapirus sp., Muntiacus sp., Zygolophodon sp., Elephas sp., Stegodon zhaotongensis, and Rynchotherium huananensis (Chow and Zhai, 1962; Chow and Zhang, 1974).

5.3. Materials and Methods

5.3.1. Sample Materials 164 well-preserved fossil teeth were collected for carbon and oxygen isotopic analyses from the Yunnan Institute of Cultural Relics and Archaeology in Kunming, Yunnan province, or were collected in the field (Appendix A). The teeth were chosen from several groups of Late Cenozoic herbivores, including bovids, horses, rhinos, elephants, cervids, pigs, tapirs, chalicotheres, and tragulids. Third molars and premolars were collected whenever possible to reduce the effects of pre-weaning on the δ18O of the tooth enamel (e.g., Fricke and O’Neil, 1996;

Bryant et al., 1996), although many specimens collected in the field were tooth fragments that

could only be identified as herbivores. Multiple specimens from each stratum were collected when possible so that variations in δ13C and δ18O values within populations, between different taxa of the same age, and between similar taxa of different ages could be determined. 13 matrix sediment samples were collected (Appendix A) from the surface of fossil teeth in order to compare their carbonate carbon and oxygen isotope compositions to the isotopic compositions of the fossil tooth enamel. In addition, 8 carbonate nodule samples were collected (Appendix A) from Yuanmou sediments for carbon and oxygen isotope analyses and 9 grass samples from Lufeng and Yuanmou and 2 organic sediment samples from Shangri La were collected (Appendix C) for carbon isotope analyses in order to estimate the C3/C4 compositions of modern and ancient ecosystems and to compare with ecosystems inferred from the δ13C of fossil teeth.

5.3.2. Laboratory Methods Bulk isotopic analyses were performed on all fossil and modern teeth, plants, and sediments. Bulk isotopic analyses of tooth enamel yield average delta values for the growth period of a tooth. The bulk carbon and oxygen isotopic compositions of the enamel carbonate were determined for each of the 164 tooth samples. Carbon and oxygen isotopic analyses were also performed on 13 sediment (matrix) samples that were collected from the surfaces of individual teeth in order to determine whether the fossil samples were diagenetically altered, as the isotopic compositions of the fossils and corresponding matrix should be similar or identical if alteration has occurred. Carbon isotopic compositions were determined for all plant samples and organic-rich sediments and δ13C and δ18O values were determined for all carbonate-rich sediment samples. The results of this study are reported in standard notation as δ13C and δ18O in reference to V-PDB and V-SMOW, respectively (Gonfiantini, 1978; Gonfiantini et al., 1995). 13 13 13 12 13 12 13 12 δ C is defined as follows: δ C = [(( C / C)SAMPLE - ( C / C)VPDB)/ ( C / C)VPDB] x 1000 ‰. 18 18 18 16 18 16 18 16 δ O is defined as: δ O = [(( O / O)SAMPLE - ( O / O)VSMOW)/ ( O / O)VSMOW] x 1000 ‰. Bulk enamel samples were obtained by drilling along the entire length of a tooth using a slow-speed rotary drill. Tooth enamel carbonate samples were pretreated in 5% reagent grade sodium hypochlorite for approximately 20 to 24 hours at room temperature to remove organic material from the tooth enamel. The samples were then centrifuged, decanted, and rinsed with deionized water to remove the sodium hypochlorite. Next, the samples were treated in 0.1M

acetic acid for 4 to 6 hours at room temperature to remove non-structural carbonate from the enamel. The samples were then rinsed with deionized water. After the final rinse, the samples were freeze-dried for 3 to 5 days. Plant samples, organic soils, and soil carbonates were dried at room temperature and then ground into powder using a mortar and pestle, but were not pretreated prior to isotopic analyses. To analyze the carbonate samples, ~200 to 500 μg of carbonate standards and 3 to 6 mg of enamel or matrix carbonates were measured and placed into reaction vials capped with rubber septa. The sample vials were then loaded into a Thermo-Finnigan Gasbench II interfaced with a Delta Plus XP continuous flow isotope ratio mass spectrometer and were flushed for 5 minutes

by injection of pure-grade helium. The samples were then converted to CO2 by injection of 100% phosphoric acid, and were left to react for approximately 72 hours at 25 ºC (modified from McCrea, 1950). Then, the carbon and oxygen isotope ratios were measured by mass

spectrometry. Ten aliquots of CO2 from each sample vial were measured and run against three 13 18 aliquots of a CO2 reference gas. The δ C and δ O values reported are the average values for the ten sample aliquots. External errors were <0.06 ‰ for δ13C and <0.08 ‰ for δ18O. The expected 18 13 internal error was <0.05 ‰ for δ O (CO2 reference gas; Thermo-Finnigan, 2002). The δ C and δ18O values were calibrated by concurrent carbon and oxygen isotopic measurements of at least two sets of three or more of the following carbonate standards: PDA, NBS-19, YW-CC-ST-1, ROY-CC, and MERK. Samples were analyzed on a Delta Plus XP IRMS at Florida State University. To analyze the organic samples, ~ 2 to 3 mg of plant tissues or organic soils were measured into tin cups and loaded into a Carlo Erba elemental analyzer (EA) connected to a Delta Plus XP continuous flow isotope ratio mass spectrometer by a Conflo II interface. The carbon isotope compositions of the organic samples were determined by mass spectrometric

measurement of CO2 produced from combustion of the sample in the EA. Two aliquots of CO2 reference gas were analyzed immediately before and after each sample. The expected external precision for the analyses was <0.15 ‰ for δ13C (20 μg carbon). The δ13C values were calibrated by concurrent carbon isotopic measurements of at least two sets of three or more of the following organic standards: YWOMST-1 (sugar), YWOMST-2 (phenylalanine), YWOMST-3 (L- phenylalanine), YWOMST-4 (Costech acetamilide), and YWOMST-5 (urea).

5.4. Results and Discussion

5.4.1. Assessment of Fossil Tooth Preservation Tooth enamel is more resistant to diagenetic alteration (as a result of large crystal size and low (~1%) porosity that allows very little influx of diagenetic fluids) than other calcified tissues and tends to preserve the original isotopic composition of an animal (Ayliffe et al., 1994; Wang and Cerling, 1994). Comparison of the δ13C and δ18O values of fossil tooth enamel with those of carbonates within coexisting matrix provides a means of determining if significant alteration of the original carbon and oxygen isotopic compositions of the enamel has occurred. A significant difference between the isotopic compositions of the enamel and matrix carbonates would argue against considerable diagenetic alteration of the enamel (e.g., Wang and Deng, 2005). δ13C and δ18O values of 13 fossil mammalian enamel samples from Baoshan, Shangri La, and Yuanmou localities, and their coexisting matrix carbonates are plotted in Figure 5.2. The 13 differences between the δ C values of almost all tooth enamel samples and their corresponding matrix carbonates are >2 ‰, so it is not likely that the carbon isotope compositions of the enamel 18 samples were altered by diagenetic processes. Differences in the δ O values between tooth enamel samples and their corresponding matrix carbonates range from ~1 to ~8 ‰, suggesting minimal or no diagenetic alteration of the oxygen isotope compositions of the enamel samples.

5.4.2. Carbon Isotope Compositions of Fossil Mammals from Six Localities in Yunnan Bulk carbon isotopic results for mammals from various localities in Yunnan Province are presented in Figures 5.3 and 5.4. Most carbon isotopic variation in mammalian herbivore tooth enamel is a result of isotope fractionation that occurred during photosynthesis of plants that were ingested by the animal. C3 plants, including trees, most shrubs, forbs, and cool season grasses, use the Calvin Cycle photosynthetic pathway and have an average δ13C value of -27 ‰ (δ13C range = -35 to -20 ‰). Under water-stressed conditions, C3 plants tend to become enriched in 13C and have δ13C values >-27‰ (Farquhar et al., 1989). C3 plants growing under closed forest canopy typically have δ13C values <-27‰, due to the influence of soil respiration (Schleser and Jayasekera, 1985; Sternberg et al., 1989; van der Merwe and Medina, 1989). C4 plants, including mainly warm season grasses, use the Hatch-Slack photosynthetic pathway and have an average δ13C value of -13 ‰ (δ13C range = -17 to -9 ‰). Because tooth enamel carbonate is 13C-enriched

100

-8

-9

-10

-11 Yg-28 Yg-28 matrix -12 SGL-01 SGL-01 matrix SGL-03 SGL-03 matrix -13 SGL-20 SGL-20 matrix

V-PDB) (‰, GP-1 GP-1 matrix O -14 GP-02 GP-02 matrix 18 δ GP-03 GP-03 matrix -15 GP-04 GP-04 matrix GP-05 GP-05 matrix XH-03 XH-03 matrix -16 Yg-34 Yg-34 matrix Yg-38 Yg-38 matrix -17 YMH-1 YMH-1 matrix

-18 -16 -14 -12 -10 -8 -6 -4 -2 0

δ13C (‰, V-PDB)

Figure 5.2. Carbon and oxygen isotope compositions of fossil mammalian tooth enamel and coexisting matrix. δ13C and δ18O values of 13 fossil enamel samples (solid shapes) from Baoshan, Shangri La, and Yuanmou localities, and their coexisting matrix carbonates (open 13 shapes) are shown. The differences between the δ C values of almost all tooth enamel samples and their corresponding matrix carbonates are >2 ‰, so it is not likely that the carbon isotope compositions of the enamel samples were altered by diagenetic processes. Differences in the 18 δ O values between tooth enamel samples and their corresponding matrix carbonates range from ~1 to ~8 ‰, suggesting minimal or no diagenetic alteration of the oxygen isotope compositions of the enamel samples.

101

1 2

(Ma) 5 Baoshan 6 Age Age Kaiyuan 7 Lufeng

8 Shangrila

9 Yuanmou

10 Zhaotong

11 -20 -18 -16 -14 -12 -10 -8 -6 -4 -2 0

δ13C (‰, V-PDB)

Figure 5.3. Individual δ13C values of fossil tooth enamel from six localities in Yunnan Province versus age. δ13C values indicate diet composition (i.e., C3 or C4 plants) and shift to more positive values with water stress and more negative values in forested or closed-canopy environments.

102

1 Mammals from Yuanmou

2 3 horses 4 (Ma) rhinos 5 chalicotheres tragulids Age 6 pigs 7 elephants 8 tapir 9 -16-15-14-13-12-11-10-9-8-7-6-5-4 δ13C (‰, V-PDB)

Figure 5.4. Variations in carbon isotopic compositions of various mammalian taxa from Yuanmou with time. The data suggest that horses and rhinos fed in more closed or forested environments than did pigs, tragulids, and chalicotheres. Elephants may have fed in both closed and open environments. Temporally increasing δ13C values for rhinos, pigs, and horses suggests that the Yuanmou Basin became more arid and/or warmer through time. At 1.75 Ma, horses had significant C4 plants in their diets as indicated by a δ13C value greater than -5 ‰.

103

by ~14 ‰ relative to diet, a tooth enamel δ13C value of ~-13 ‰ would indicate a pure C3 diet and a δ13C value of +1 ‰ would indicate a pure C4 diet (Lee-Thorp and van der Merwe, 1987; O’Leary, 1988; Farquhar et al., 1989; Koch, 1998; Cerling et al., 1997a; Cerling and Harris, 1999; Kohn and Cerling, 2002; Wang and Deng, 2005). In arid and semi-arid environments (or water-stressed environments), a “cut-off” enamel-δ13C value for a pure C3 diet could be -8 ‰ for modern herbivores and -7 ‰ for fossil herbivores, due to changes in the δ13C value of atmospheric CO2 (Cerling et al., 1997a; Wang et al., 2008b). In a dense forested environment, the end-member enamel C-13 value for a pure C3 diet could be as low as -15‰ or -16‰ (Cerling et al., 1997). Therefore, temporal shifts in the δ13C values of mammals from given localities may indicate shifts in diet, habitat, or regional climatic conditions. Bulk δ13C values of mammalian tooth enamel from six localities in Yunnan Province are shown in Figure 5.3. The carbon isotope data indicate that all mammals from Baoshan, Kaiyuan, Lufeng, and Zhaotong may have had pure C3 diets at all ages, as all enamel δ13C values are less than -10 ‰ (δ13C range = -14.8 to -10.4 ‰). On the other hand, abundant coal and lignite deposits found in the Late Cenozoic strata at these fossil localities suggest that the region had plenty of water and was probably not under severe water stress through out much of the Neogene and the very negative enamel δ13C values (<-13‰) suggest a densely forested environment. In such an environment, an enamel δ13C value of -10.4 ‰ could indicate a mixed C3/C4 diet, consisting of ~29% C4, assuming that the end-member enamel δ13C values for pure C3 and C4 diets were -15‰ and +1‰, respectively. In that case, the data suggest that forests dominated the landscape in these areas, perhaps with patches of open habitats. The enamel δ13C of mammals (rhinos) aged ~9 Ka from Baoshan ranged from -14.8 to -14.1 ‰, with a mean δ13C value of - 14.4 ‰. The enamel δ13C of mammals (elephants) aged ~5 Ma from Baoshan ranged from -12.3 to -11.5 ‰, with a mean δ13C value of -11.9 ‰. The higher δ13C values of mammals at 5 Ma compared to those at 9 Ka suggest that the paleoenvironment of the Baoshan region was more open and/or water-stressed at 5 Ma than at 9 Ka. Alternatively, the differences in δ13C values may be due to differences in diets and/or habitat preferences of elephants and rhinos. Even so, the enamel δ13C values of the 9 Ka mammals from Baoshan were <-13 ‰, suggesting that the Baoshan region was a closed forest at ~9 Ka, whereas the >-13 ‰ δ13C values at ~5 Ma suggest a mixed woodland/grassland biome at this earlier age.

104

The 4 Ma mammal (elephant) from Kaiyuan had an enamel δ13C value of -14.1 ‰ and the 10 Ma mammals (pig and tapir) from Kaiyuan had an enamel δ13C range of -11.3 to -9.9 ‰, with a mean δ13C value of -10.6 ‰. The higher δ13C values of mammals at 10 Ma compared to those at 4 Ma suggest that the Kaiyuan region was more open and/or water-stressed at 10 Ma than at 4 Ma. These data also suggest that the Kaiyuan region was a closed forest at 4 Ma. Because pigs and tapirs are known to be more restricted to forested environments and elephants are known to inhabit more open environments, such as savannahs, the more negative δ13C value of the elephant at ~4 Ma compared to that of the pig and tapir at 10 Ma, most likely does not reflect differences in dietary behavior or habitat preference, but instead, reflects a general change in relative humidity or ecosystem type. The 4 Ma mammal (elephant) from Zhaotong had an enamel δ13C value of -11.7 ‰, which is ~2.5 ‰ enriched in 13C compared to that of the 4 Ma elephant individual from Kaiyuan, but is very similar to that of the two ~5 Ma elephant individuals from Baoshan. These data suggest that the paleoenvironment of Zhaotong was more open and/or water-stressed than that of Kaiyuan at ~4 Ma, but was very similar to that of Baoshan at ~5 Ma. The 7.5 Ma mammalian individuals (elephant and horses) from Lufeng had an enamel δ13C range of -13.8 to -10.9 ‰, with a mean δ13C value of -12.2 ‰ and the enamel δ13C of 8 Ma mammals (bovids, horses, rhinos, a chalicothere, deer, and a bear) from Lufeng ranged from - 14.2 to -10.4 ‰, with a mean δ13C value of -12 ‰. The similarity in enamel δ13C values of the 7.5 and 8 Ma mammals from Lufeng suggest very similar diets, habitats, and climatic conditions for mammals living at both ages in that region. The low δ13C values of mammals from that age indicate that the animals were not ingesting significant amounts of C4 vegetation at ~8 Ma (no more than 29% C4 in diet, as mentioned previously). Much of the province is presently within the “Subtropical Broadleaved Evergreen Forest” vegetation zone that contains significant amounts of C4 grasses (Liu, 1988; Yin and Li, 1997). All the grasses we collected from Lufeng are C4, based on carbon isotope compositions (Table 5.2). The enamel δ13C data suggest that Lufeng was largely a forested environment at ~8 Ma, but a small amount of C4 plants may have been present in local ecosystems at that time. Serial analysis of individual teeth would help determine if the local ecosystems indeed contained a small amount of C4. Mammals from Shangri La had diets that ranged from pure C3 to pure C4, as indicated by enamel δ13C values. Mammals from Yuanmou had mostly C3-based diets (some individuals 105

Table 5.1. Carbon isotope compositions of soil carbonates and organic matter from Yunnan Province.

Estimated Age Δ13C Estimated Sample No. Sample Type Locality (Ma) (PDB) % C4 YPS-8 Soi l carbonate nodule N iuj iangbao, Yuanmou 1.7 -3.2 63 YPS-9 Soi l carbonate nodule N iujiangbao, Yuanmou 1.7 -4.9 51 YPS-9 Soi l carbonate nodule N iujiangbao, Yuanmou 1.7 -5.1 49 YPS-10 Soi l carbonate nodule N iujiangbao, Yuanmou 1.7 -8.6 25 YPS-11 Soi l carbonate nodule N iujiangbao, Yuanmou 1.7 -1.4 76 YPS-1 Soi l carbonate nodule Baozidongqian, Yuanmou 7.1 -4.8 51 YPS-1 Soi l carbonate nodule Baozidongqian, Yuanmou 7.1 -5.1 49 YPS-2 Soi l carbonate nodule Baozidongqian, Yuanmou 7.1 -7.2 35 YPS-2 Soi l carbonate nodule Baozidongqian, Yuanmou 7.1 -7.3 33 YPS-3 Soi l carbonate nodule Baozidongqian, Yuanmou 7.1 -7.4 33 YPS-5 Soi l carbonate nodule Baozidongqian, Yuanmou 7.1 -7.1 35 SGL-45 Buried recent prairie soil Shangrila 0.0 -24.8 15 SGL-45 Buried recent prairie soil Shangrila 0.0 -24.4 18

Table 5.2. Carbon isotope compositions of grasses from Lufeng and Yuanmou localities in Yunnan province.

13 Sample δ C No. Sample Type Locality (PDB) Lf-01 Grass Lufeng -12.85 Lf-01 Grass Lufeng -13.06 Lf-02 Grass Lufeng -12.60 Lf-02 Grass Lufeng -12.36 Lf-03 Grass Lufeng -14.54 Lf-03 Grass Lufeng -14.61 Lf-04 Grass Lufeng -14.01 Lf-04 Grass Lufeng -13.86 Lf-08 Grass Lufeng -13.80 Lf-08 Grass Lufeng -13.77 Lf-05 Grass Baozidongqian, Yuanmou -13.62 Lf-05 Grass Baozidongqian, Yuanmou -13.69 Lf-06 Grass Baozidongqian, Yuanmou -14.24 Lf-07 Grass Baozidongqian, Yuanmou -13.17 Lf-09 Grass Baozidongqian, Yuanmou -13.27

106

may have incorporated a small amount of C4 (< 20-30%) in their diets) until ~3.5 Ma and then appear to have incorporated significant amounts of C4 plants (up to 60%) in their diets at ~1.75 Ma and later. Modern mammals (a deer and a pig) from Shangri La had an enamel δ13C range of -10.5 to -8.1 ‰ and a mean enamel δ13C value of -9.3 ‰. 1.75 Ma mammals (a horse and misc. unidentified herbivores) from Shangri La had an enamel δ13C range of -10.6 to -0.2 ‰ and a mean enamel δ13C value of -6.1 ‰. 2.5 Ma mammals (bovids, a horse, rhinos, deer, a pig, and misc. unidentified herbivores) from Shangri La had an enamel δ13C range of -11.6 to -1.1 ‰ and a mean enamel δ13C value of -8.1 ‰. These results indicate that the diets of modern and fossil mammals from Shangri La were 13C-enriched compared to mammals of all ages from all other sites in Yunnan Province, with the exception of two 1.75 Ma horses and one modern pig from Yuanmou. Because no samples of equivalent ages from other localities are available for comparison with the Shangri La specimens and because there are no samples older than 2.5 Ma from Shangri La to compare with those from other localities, it is not yet possible to determine if significant amounts of C4 grasses had spread only into the Shangri La and Yuanmou regions or also into other regions of Yunnan Province by ~2-3 Ma, or if the expansion of C4 grasses into Shangri La occurred at an earlier time. The present-day ecosystem in Shangri La consists of a mosaic of forests and prairies. Two modern prairie soil samples yielded an average δ13C value of -24.6‰ (Table 5.1), suggesting C3 dominance on the prairie with C4 grasses accounting for no more than ~18% of prairie plant biomass. The fossil enamel δ13C values indicate that the vegetation in the Shangri La region was not composed solely of C4 vegetation at ~1.75 Ma, as approximately 1/3 of the mammals apparently had pure C3 diets. Overall, the carbon isotope data show (1) that fossil mammals from Shangri La had variable diets that ranged from pure C3 to pure C4 from ~2.5 to 1.75 Ma, which indicates the presence of significant C4 biomass in local ecosystems at those ages; and (2) limited modern data indicates that diets were dominated by C3 plants, with C4 accounting for no more than 35% of the diet, which is consistent with the present C3 dominance in the area. The enamel δ13C value of a modern mammal (pig) from Yuanmou was -8.1 ‰, and the enamel δ13C values of mammals (horses) aged ~1.75 Ma from Yuanmou ranged from -9.4 to -4.7 ‰, with a mean δ13C value of -7 ‰. The enamel δ13C value of a 3.5 Ma mammal (rhino) from Yuanmou was -12.9 ‰, and the enamel δ13C values of mammals (rhinos and an elephant) aged

107

~4 Ma from Yuanmou ranged from -13.2 to -10.2 ‰, with a mean δ13C value of -12.1 ‰. The enamel δ13C values of 5 Ma mammals (rhinos) from Yuanmou ranged from -13.2 to -10.8 ‰, with a mean δ13C value of -12.2 ‰, and the enamel δ13C values of mammals (horses, rhinos and elephants) aged ~8 Ma from Yuanmou ranged from -15.3 to -12.2 ‰, with a mean δ13C value of -13.6 ‰. The enamel δ13C values of 8.15 Ma mammals (horses, rhinos, chalicotheres, tragulids, pigs, and a tapir) from Yuanmou ranged from -15.4 to -10.5 ‰, with a mean δ13C value of -12.6 ‰, and the enamel δ13C values of mammals (rhinos and elephants) aged ~8.5 Ma from Yuanmou ranged from -14.7 to -12 ‰, with a mean δ13C value of -13.3 ‰. The very negative δ13C values (as low as -15.3 ‰) indicate that the area had dense forests at ~8 Ma. The large range of δ13C values at about 8 Ma suggest that patches of more open area may have existed where the forest canopies were broken or on flood plains. The presence of soil carbonate nodules in paleosols at ~7 Ma suggests the presence of more open habitats that may have contained up to ~50% C4 grasses, as indicated by the δ13C values of paleosol carbonate nodules (Table 5.1). These results indicate that although most mammals that lived from 8.5 to 3.5 Ma in the Yuanmou Basin had pure or nearly pure C3 diets, some individuals may have consumed small amounts of C4 grasses (no more than 30% C4 in diets assuming end-member d13C values for pure C3 and C4 diets were -15 and +1‰, respectively). Again, serial analysis of individual teeth would provide further insight into whether C4 grasses were indeed a dietary component for some of the animals. More positive δ13C values at ~1.75 Ma suggests that C4 biomass increased significantly in the Yuanmou region after ~2-4 Ma, indicating a shift from a more closed forested environment at ~8 Ma to an increasingly more open and/or drier environment around 1.75 Ma. Furthermore, because moderate rainfall, seasonal drought, and fires are all important factors in the development of modern grasslands (Walter, 1971), the increased C4 biomass may indicate an increased drought or a strengthening of the monsoon circulation in the region after ~2-4 Ma. Carbon isotope analyses of carbonate nodules (Table 5.1) indicate that the Yuanmou ecosystem at ~7.1 Ma consisted of ~33 to 51% C4 vegetation and at ~1.7 Ma the ecosystem consisted of ~25 to 76% C4. The increase in C4 vegetation between ~7.1 and 1.7 Ma is consistent with the increase in δ13C values of mammalian tooth enamel from similar ages. All modern grasses collected from the Yuanmou region have δ13C values which indicate they are C4 plants (Table 5.2). Only one modern mammalian sample (a pig) from Yuanmou was analyzed and it apparently

108 had a mixed C3-C4 diet (~35% C4 assuming -13 ‰ and +1 ‰ as the end-member enamel δ13C values for pure C3 and pure C4 diets, respectively). A more detailed look at bulk carbon isotopic compositions of mammals from Yuanmou versus age is presented in Figure 5.4. The results suggest niche partitioning of the various taxa. From ~8.5 to 8 Ma, all horses and rhinos plot between ~-12.5 and -15.5 ‰ and all pigs, tragulids, and chalicotheres plot between ~-10.5 and -13 ‰. Elephants plot from ~-12 to -13 ‰. The data suggest that horses and rhinos fed in more closed or forested environments than did pigs, tragulids, and chalicotheres. Elephants may have fed in both closed and open environments. Temporally increasing δ13C values for rhinos, pigs, and horses suggests that the Yuanmou Basin became more arid and/or warmer through time. At 1.75 Ma, horses began to incorporate significant amounts of C4 plants into their diets as indicated by a δ13C value greater than -5 ‰.

5.4.3. Oxygen Isotope Compositions of Fossil Mammals from Six Localities in Yunnan Bulk oxygen isotopic results for mammals from various localities in Yunnan Province are presented in Figures 5.5 and 5.6. The oxygen isotopic composition of mammalian herbivore tooth enamel largely reflects the isotopic composition of local meteoric water that is ingested by the herbivores (Longinelli, 1984; Luz et al., 1984; Ayliffe and Chivas, 1990). In turn, the oxygen isotopic composition of meteoric water is controlled by climate (Dansgaard, 1964; Rozanski et al., 1992). As a result, the oxygen isotopic compositions of mammalian tooth enamel can be utilized to reconstruct regional paleoclimates. Significant shifts over time in the oxygen isotope compositions of tooth enamel of a given taxon, from a given region, indicates a change in regional climate (e.g., Ayliffe and Chivas 1990; Quade et al., 1992; Ayliffe et al., 1994; Bryant et al., 1994; Kohn and Cerling, 2002). At any given time interval, a large range of δ18O values for individuals of a given taxon may reflect differences in dietary behavior or local and seasonal variability in precipitation (MacFadden, 1998; Kohn and Cerling, 2002). As shown in Figure 5.5, the enamel δ18O of two rhinos aged ~9 Ka from Baoshan ranged from -8.5 to -8.3 ‰, with a mean δ18O value of -8.4 ‰. The enamel δ18O of two elephants aged ~5 Ma from Baoshan ranged from -11.1 to -8.4 ‰, with a mean δ18O value of -9.7 ‰. The 4 Ma mammal (elephant) from Kaiyuan had an enamel δ18O value of -8.8 ‰ and the 10 Ma mammals from Kaiyuan had an enamel δ18O range of -10.3 (pig) to -6.9 (tapir) ‰, with a mean δ18O value of -8.6 ‰. The results show that the δ18O value of the tapir at 10 Ma was higher

109

1 2

4 5 (Ma) Baoshan 6 Age Kaiyuan 7 Lufeng 8 Shangrila

9 Yuanmou 10 Zhaotong

11 -20 -18 -16 -14 -12 -10 -8 -6 -4 -2 0

δ18O (‰, V-PDB)

Figure 5.5. Individual δ18O values of fossil tooth enamel from six localities in Yunnan Province versus age. Negative shifts in δ18O values for a given taxon from a given locality indicate changes to cooler and/or wetter climates and positive shifts indicate changes to warmer and/or drier climates.

110

0 Mammals from Yuanmou 1 2 3 horses 4 rhinos (Ma) 5 chalicotheres tragulids

Age 6 pigs 7 elephants 8 tapir 9 -16-15-14-13-12-11-10-9-8-7-6-5-4 δ18O (‰, V-PDB)

Figure 5.6. Variations in oxygen isotopic compositions of various mammalian taxa from Yuanmou over time. The data indicate positive shifts in δ18O values from ~8.5 to 8 Ma in elephants, from ~8 Ma to the present in horses, from ~ 8 to 5 Ma in rhinos, and from ~8 to 2 Ma in pigs, which suggests that the local climate became warmer and/or drier over time. Negative shifts in δ18O values of rhinos and an elephant after ~5 Ma may indicate a shift to a wetter climate during the time interval from ~5 to 3.5 Ma in the Yuanmou Basin.

111

than that of the elephant at 4 Ma. Again, because elephants are known to favor more open habitats, whereas tapirs prefer to remain in forested environments, the results suggest a negative shift in δ18O values after 10 Ma. The implied negative shift in mean δ18O values occurs in the same direction as the shift in δ13C values for the Kaiyuan region, indicating that the variation in δ13C values may be a result of variation in water availability. As a result, this suggests that the variation in δ18O values most likely reflect changes in humidity or aridity, as opposed to changes in temperature. More samples are needed from Kaiyuan in order to establish a more conclusive climatic record of this region. The 4 Ma mammal (elephant) from Zhaotong had an enamel δ18O value of -12.7 ‰, which is ~4 ‰ enriched in 18O compared to that of the 4 Ma elephant individual from Kaiyuan, and is ~3 ‰ 18O-enriched compared to that of the two 5 Ma elephant individuals from Baoshan. These data suggest that the paleoenvironment of Zhaotong was colder and/or wetter than that of Kaiyuan at ~4 Ma and Baoshan at 5 Ma. Again, the carbon isotope data suggested that the paleoenvironment of Zhaotong was more open and/or water-stressed than that of Kaiyuan at ~4 Ma, but was very similar to that of Baoshan at ~5 Ma. From all available data, it is not possible to discern whether the climate of Zhaotong differed from the Baoshan and Kaiyuan localities due to differences in temperature, relative humidity, or ecosystem type (closed forests vs. open grasslands, for example). More samples of various ages are needed from this region to establish a better climate record. The 7.5 Ma mammalian individuals (elephant and horses) from Lufeng had an enamel δ18O range of -13.6 to -8.7 ‰, with a mean δ18O value of -10.7 ‰ and the enamel δ18O of 8 Ma mammals (bovids, horses, rhinos, a chalicothere, deer, and a bear) from Lufeng ranged from - 13.8 to -5.9 ‰, with a mean δ18O value of -10.8 ‰. The similarity in enamel δ18O values of the 7.5 and 8 Ma mammals from Lufeng suggest very similar climatic conditions for mammals living at both ages in that region. The enamel δ18O value of a modern mammal (pig) from Yuanmou was -8.3 ‰, and the enamel δ18O values of mammals (horses) aged ~1.75 Ma from Yuanmou ranged from -8.1 to - 6.1 ‰, with a mean δ18O value of -7.1 ‰. The enamel δ13C value of a 3.5 Ma mammal (rhino) from Yuanmou was -10.9 ‰, and the enamel δ18O values of mammals (rhinos and an elephant) aged ~4 Ma from Yuanmou ranged from -14.3 to -10.7 ‰, with a mean δ18O value of -12.3 ‰.

112

The enamel δ18O values of 5 Ma mammals (rhinos) from Yuanmou ranged from -10.1 to -5.7 ‰, with a mean δ18O value of -8.9 ‰, and the enamel δ18O values of mammals (horses, rhinos and elephants) aged ~8 Ma from Yuanmou ranged from -12.6 to -6.9 ‰, with a mean δ18O value of - 10.4 ‰. The enamel δ18O values of 8.15 Ma mammals (horses, rhinos, chalicotheres, tragulids, pigs, and a tapir) from Yuanmou ranged from -15.8 to -5 ‰, with a mean δ18O value of -10.7 ‰, and the enamel δ18O values of mammals (rhinos and elephants) aged ~8.5 Ma from Yuanmou ranged from -14.2 to -10.4 ‰, with a mean δ18O value of -12.1 ‰. Positive shifts in the enamel δ18O values occur after ~8-8.5 Ma and ~4 Ma, indicating shifts to warmer and/or more arid climatic conditions. The positive shift in the δ18O values of mammals from the Yuanmou Basin after ~8-8.5 Ma is similar in timing to positive δ18O shifts observed in horses, rhinos, and deer from the Linxia Basin and in fossils and paleosols from Pakistan and Nepal (Quade et al., 1989; Quade et al., 1992; Quade et al., 1995; Wang and Deng, 2005; Biasatti, manuscript in preparation). This suggests a shift toward a drier and/or warmer climate northeast, southeast, and south of the Tibetan Plateau during the Late Miocene. A positive shift in the δ18O values of mammals from the Yuanmou Basin after ~4 Ma is similar in timing to the so-called “Early Pliocene warm period,” which occurred from ~5 to 3.5 Ma. Based on deep-sea δ18O and Mg/Ca records, a general cooling trend with two main cooling phases have occurred over the last 25 million years (Shackleton and Kennett, 1975; Shackleton et al., 1995; Lear et al., 2000; Zachos et al., 2001). The first significant cooling event occurred during the Middle Miocene from ~15 to 10 Ma (Shackleton & Kennett, 1975; Zachos et al., 2001). The second cooling phase occurred during the Plio-Pleistocene with a small-scale expansion of the west Antarctic ice sheet followed by the onset of the Northern Hemisphere Glaciation (Shackleton and Kennett, 1975; Lear et al., 2000; Zachos et al., 2001). A possible negative shift in the δ18O values of mammals from the Yuanmou Basin after ~2 Ma is in general agreement with the deep-sea records, but more data is needed to determine if, in fact, a negative shift occurs after ~2 Ma, and the negative shift in δ18O values of mammals from the Yuanmou Basin after ~5 Ma is similar in timing to a negative shift observed in the Linxia Basin. In addition, negative shifts in the δ18O values of rhinos and an elephant after ~5 Ma, roughly coinciding with a negative shift in the δ18O of lacustrine and fluvial carbonates in the Gyirong Basin in southern Tibet and a period of inferred enhancement

113

of the summer monsoons (Li et al., 2008; Wang et al., submitted), may indicate a wetter environment during the time interval from ~5 to 3.5 Ma in the Yuanmou Basin. Bulk oxygen isotopic compositions of the various mammals from Yuanmou versus age are presented in Figure 5.6. The data indicate positive shifts in δ18O values from ~8.5 to 8 Ma in elephants, from ~8 Ma to 1.75 Ma in horses, from ~ 8 to 5 Ma in rhinos, and from ~8 to 2 Ma in pigs, which suggests that the local climate became warmer and/or drier over time. This is in agreement with the carbon isotope data.

5.5. Conclusions

Carbon and oxygen isotopic analyses of mammalian tooth enamel indicate that significant changes occurred in the climates and diets of mammalian taxa from Yunnan Province over the last 10 million years. The carbon isotope results indicated that all mammals collected from Baoshan, Kaiyuan, Lufeng and Zhaotong had pure C3 or mixed C3/C4 diets (with up to 29% C4) at all ages. The carbon isotope compositions of mammals from Baoshan suggest that the paleoenvironment of the Baoshan region was more open and/or water-stressed at 5 Ma than at 9 Ka and that the region was a closed forest at ~9 Ka and a mixed woodland/grassland biome at ~5 Ma. The carbon isotope compositions of mammals from Kaiyuan suggest that the region was more open and/or water-stressed at 10 Ma than at 4 Ma. The data also suggest that the Kaiyuan region was a closed forest at 4 Ma. The carbon isotope results suggest that the paleoenvironment of Zhaotong was more open and/or water-stressed than that of Kaiyuan at ~4 Ma, but was very similar to that of Baoshan at ~5 Ma. A similarity in enamel δ13C values of 7.5 and 8 Ma mammals from Lufeng suggest very similar diets, habitats, and climatic conditions for mammals living at both ages in that region. Carbon isotope analyses of 2.5 and 1.75 Ma mammals from Shangri La indicated diets that ranged from pure C3 to pure C4 and showed that the diets were 13C-enriched compared to mammals from all other sites in Yunnan Province, with the exception of 1.75 Ma mammals from Yuanmou. The carbon isotope results indicated that all mammals that lived from 8.5 to 3.5 Ma in the Yuanmou Basin had diets that varied from pure C3 to mixed C3/C4 diets (with up to 30% C4) and the carbon isotope compositions of ~7 Ma paleosol carbonates from Yuanmou suggest the presence of significant C4 biomass in more open and relatively dry areas, where the soil carbonates were formed. On the other hand, a shift to

114 more positive enamel δ13C values at ~1.75 Ma suggests that C4 plants may have not been present in the Yuanmou region in significant quantities until after ~3.5 Ma. In addition, the carbon isotope results indicate a change from a largely forested environment at ~8 Ma to an increasingly drier and more open environment at ~1.75 Ma in the Yuanmou region. Furthermore, because C4 grasses require sufficient summer precipitation and seasonal drought, the increased C4 biomass may indicate a strengthening of the monsoon circulation in the region after ~2-4 Ma. Carbon isotope analyses of carbonate nodules from the Yuanmou Basin are consistent with the δ13C values of mammalian tooth enamel from similar ages. The carbon isotope results of mammals from Yuanmou also suggest niche partitioning of the various taxa. The data suggest that horses and rhinos fed in more closed or forested environments than did pigs, tragulids, and chalicotheres and that elephants may have fed in both closed and open environments. Temporally increasing δ13C values for rhinos, pigs, and horses suggests that the Yuanmou Basin became more arid and/or warmer through time. The oxygen isotope results suggest a negative shift occurred in the δ18O values of mammals from Kaiyuan after ~10 Ma, indicating a change to colder temperatures or a wetter environment between 10 and 4 Ma. The oxygen isotope results also suggest that the paleoenvironment of Zhaotong at 4 Ma was colder and/or wetter than that of Kaiyuan at ~4 Ma and Baoshan at 5 Ma. A similarity in enamel δ18O values of 7.5 and 8 Ma mammals from Lufeng suggest very similar climatic conditions at both ages in that region. Positive shifts in enamel δ18O values occur after ~8-8.5 Ma and ~4 Ma in the Yuanmou region, indicating shifts to warmer and/or more arid climatic conditions. The positive shift after ~8-8.5 Ma is similar in timing to positive δ18O shifts observed in horses, rhinos, and deer from the Linxia Basin and in fossils and paleosols from Pakistan and Nepal, suggesting a shift toward a drier and/or warmer climate at the northeast, southeast, and southern borders of the Tibetan Plateau during the Late Miocene. A positive shift in the δ18O values of mammals from the Yuanmou Basin after ~4 Ma is similar in timing to the so-called “Early Pliocene warm period,” which occurred from ~5 to 3.5 Ma. Negative shifts in the δ18O values of rhinos and an elephant after ~5 Ma, roughly coinciding with a negative shift in the δ18O of lacustrine and fluvial carbonates in the Gyirong Basin in southern Tibet and a period of inferred enhancement of the summer monsoons (Li et al., 2008; Wang et al., submitted), may indicate a wetter environment during the time interval from ~5 to 3.5 Ma in

115 the Yuanmou Basin. Overall, the oxygen isotope compositions of mammals from the Yuanmou Basin indicate a general warming and/or drying of the local climate over time, which is in agreement with the carbon isotope results from that region.

116

CHAPTER 6

CONCLUSION

This study examined the paleoenvironments and paleoecologies of fossil mammalian herbivores from Late Cenozoic deposits in the Linxia Basin in Gansu Province and six localities in Yunnan Province, located along the northeastern and southeastern margins of the Tibetan Plateau in western China, respectively, using stable carbon and oxygen isotopic records of fossil tooth enamel and ancient sediments. The isotopic results served as proxies for paleoclimatic and paleoecological conditions in western China over the last 25 m.y. In Chapter 3, carbon and oxygen isotopic analyses of tooth enamel indicated that significant changes occurred in the climates and diets of mammalian taxa from the Linxia basin, as well as in the seasonal patterns of diet and climate, throughout the last 25 million years. Positive and negative shifts in the mean bulk δ18O values of enamel from both horses and rhinos were roughly consistent with deep-sea records that indicated a general warming trend from ~26 to 15 Ma and two major cooling phases during the Neogene, respectively. A positive shift in the mean bulk δ18O values of horses, rhinos and deer from the Linxia Basin was similar in timing to a positive δ18O shift observed in fossils and paleosols from Pakistan and Nepal, suggesting a shift toward a drier and/or warmer climate on both the north and south sides of the Tibetan Plateau during the Late Miocene. The mean bulk δ13C values for horses and rhinos indicated that both taxa had pure C3 diets throughout most of the Late Cenozoic. At 1.2 Ma, the horse bulk δ13C values increased to ~-5 ‰, indicating a change to a mixed C3/C4 diet after 2.5 Ma. This suggests that C4 grasses may have not spread into the basin until after 2.5 Ma, which is much later than the proposed global C4 expansion during the Late Miocene. This also indicates a strengthening of the Asian summer monsoon after ~2.5 Ma, as C4 plants require sufficient summer precipitation. All horse and rhino bulk δ13C values were greater than -13 ‰, indicating an open environment, such as a savannah or mixed woodland/steppe biome in the Linxia Basin from ~25 to 0.05 Ma.

117

Serial oxygen isotopic analyses showed that, in general, positive δ18O shifts in the horse and rhino bulk data, indicating shifts to either drier and/or warmer conditions after 14, 9.5, 7.5, and 2.5 Ma, were accompanied by increased seasonality, as inferred from the relatively greater δ18O ranges in the serial data. Likewise, negative δ18O shifts in the bulk data at 11.5, 6.0, 4.0, and 1.2 Ma were associated with decreases in seasonality, or relatively smaller δ18O ranges in the serial data. A marked increase in the serial δ18O ranges of both horses and bovids after 2.5 Ma is consistent with a strengthening of the summer monsoon in the region after ~2-3 Ma. The serial δ13C ranges of all horses from the ages 11.5, 9.5, 7.5, 6.0, 4.0, and 2.5 Ma were smaller than those of horses from 1.2 and 0.05 Ma. This increase in δ13C ranges further supports changes in the composition of plant biomass in the Linxia Basin after 2.5 Ma, as taxa with mixed C3/C4 diets would have increased δ13C ranges in their enamel compared to those with pure C3 diets. Interestingly, decreases in the δ18O values within individual teeth of horses from 1.2 and 0.05 Ma occurred simultaneously with increases in the δ13C values. This negative correlation between δ18O and δ13C values is consistent with that expected in summer monsoonal regions within China, but not outside of monsoonal regions and strongly supports a strengthening of the summer after ~2-3 Ma. Serial analyses of three bovid individuals from 2.5 Ma and two modern bovids also showed an anti-correlation between δ13C and δ18O values for all individuals, consistent with a strengthened monsoon circulation since about 2-3 Ma. In Chapter 4, bulk and serial C and O isotope results for rhinocerotoids that lived in the Linxia Basin from 25 to 2.5 Ma supported most previous hypotheses that were formed from analyses of taxonomic and cranial and limb morphological characteristics and allowed new insight on some genera. The isotope data supported that Paraceratherium at 25 Ma lived in a closed forested environment and spent a great amount of time wallowing in water or mud, which is consistent with the gigantic size of this rhino and that its dentition was apparently specialized for consuming tree leaves. It also supported that Allacerops, who coexisted with Paraceratherium at 25 Ma, lived in a relatively open habitat and had a less specialized diet. The data suggested that the Early Miocene rhinos Alicornops and Hispanotherium had dissimilar diets and habitats, and were consistent with previous hypotheses that Hispanotherium was a grazer in an open steppe environment, whereas Alicornops was a more generalized feeder in a forested environment. The isotope data were consistent with previous inferences that the Late

118

Miocene rhino Parelasmotherium grazed and dwelled in an open steppe habitat, but were inconsistent with previous hypotheses that the rhinos Acerorhinus and Dicerorhinus inhabited closed forested environments. Rather, the results indicated that these two rhinos inhabited open- steppe regimes. The isotopic results were not conclusive in regard to the habitat of Iranotherium, but supported previous hypotheses of a specialized C3 grass diet for Iranotherium. The results also suggested that the rhino Chilotherium dwelled in a forested environment throughout most of the Late Miocene, but inhabited a more open environment by the end of the Late Miocene, indicating that Chilotherium was able to adapt to a changing environment. It was not possible to determine from the isotopic results whether Chilotherium was grazing or browsing within the forest. Finally, the results were consistent with previous hypotheses that the Pliocene rhinos Shansirhinus and Coelodonta were grazers that lived in an open habitat and support that the climate was cooling during much of the Pliocene. Overall, the oxygen isotope data imply a general warming and/or drying trend in the Linxia Basin from the Late Oligocene to Late Miocene. The serial oxygen isotope data show almost no changes in seasonality in the Linxia basin from 25 to 6 Ma, indicating that climate in the region did not become strongly influenced by the Asian monsoon prior to 6 Ma. Furthermore, the carbon isotope data support that all rhinos in this study were pure C3 feeders and indicate that C4 grasses were not an important component of the plant biomass in the Linxia Basin prior to 2.5 Ma, which is contrary to the global C4 grass expansion at ~7 Ma. In Chapter 5, carbon and oxygen isotopic analyses of mammalian tooth enamel indicated that significant changes occurred in the climates and diets of mammalian taxa from Yunnan Province over the last 10 million years. The carbon isotope results indicated that all mammals collected from Baoshan, Kaiyuan, Lufeng and Zhaotong had pure C3 or mixed C3/C4 diets (with up to 29% C4) at all ages. The carbon isotope compositions of mammals from Baoshan suggest that the paleoenvironment of the Baoshan region was more open and/or water-stressed at 5 Ma than at 9 Ka and that the region was a closed forest at ~9 Ka and a mixed woodland/grassland biome at ~5 Ma. The carbon isotope compositions of mammals from Kaiyuan suggest that the region was more open and/or water-stressed at 10 Ma than at 4 Ma. The data also suggest that the Kaiyuan region was a closed forest at 4 Ma. The carbon isotope results suggest that the paleoenvironment of Zhaotong was more open and/or water-stressed than that of Kaiyuan at ~4 Ma, but was very similar to that of Baoshan at ~5 Ma. A similarity in enamel δ13C values of 7.5

119

and 8 Ma mammals from Lufeng suggest very similar diets, habitats, and climatic conditions for mammals living at both ages in that region. Carbon isotope analyses of 2.5 and 1.75 Ma mammals from Shangri La indicated diets that ranged from pure C3 to pure C4. The carbon isotope results indicated that all mammals that lived from 8.5 to 3.5 Ma in the Yuanmou Basin had diets that varied from pure C3 to mixed C3/C4 diets (with up to 30% C4) and the carbon isotope compositions of ~7 Ma paleosol carbonates from Yuanmou suggest the presence of significant C4 biomass in more open and relatively dry areas, where the soil carbonates were formed. On the other hand, a shift to more positive enamel δ13C values at ~1.75 Ma suggests that C4 plants may have not been present in the Yuanmou region in significant quantities until after ~3.5 Ma. In addition, the carbon isotope results indicate a change from a largely forested environment at ~8 Ma to an increasingly drier and more open environment at ~1.75 Ma in the Yuanmou region. Furthermore, because C4 grasses require sufficient summer precipitation and seasonal drought, the increased C4 biomass may indicate a strengthening of the monsoon circulation in the region after ~2-4 Ma. The carbon isotope results of mammals from Yuanmou also suggest niche partitioning of the various taxa. The data suggest that horses and rhinos fed in more closed or forested environments than did pigs, tragulids, and chalicotheres and that elephants may have fed in both closed and open environments. Temporally increasing δ13C values for rhinos, pigs, and horses suggests that the Yuanmou Basin became more arid and/or warmer through time. The oxygen isotope results suggest a negative shift occurred in the δ18O values of mammals from Kaiyuan after ~10 Ma, indicating a change to colder temperatures or a wetter environment between 10 and 4 Ma. The oxygen isotope results also suggest that the paleoenvironment of Zhaotong at 4 Ma was colder and/or wetter than that of Kaiyuan at ~4 Ma and Baoshan at 5 Ma. A similarity in enamel δ18O values of 7.5 and 8 Ma mammals from Lufeng suggest very similar climatic conditions at both ages in that region. Positive shifts in enamel δ18O values occur after ~8-8.5 Ma and ~4 Ma in the Yuanmou region, indicating shifts to warmer and/or more arid climatic conditions. The positive shift after ~8-8.5 Ma is similar in timing to positive δ18O shifts observed in horses, rhinos, and deer from the Linxia Basin and in fossils and paleosols from Pakistan and Nepal, suggesting a shift toward a drier and/or warmer climate at the northeast, southeast, and southern borders of the Tibetan Plateau during the Late Miocene. A positive shift in the δ18O values of mammals from the Yuanmou Basin after ~4 Ma is similar in

120

timing to the so-called “Early Pliocene warm period,” which occurred from ~5 to 3.5 Ma. Negative shifts in the δ18O values of rhinos and an elephant after ~5 Ma, roughly coinciding with a negative shift in the δ18O of lacustrine and fluvial carbonates in the Gyirong Basin in southern Tibet and a period of inferred enhancement of the summer monsoons (Li et al., 2008; Wang et al., submitted), may indicate a wetter environment during the time interval from ~5 to 3.5 Ma in the Yuanmou Basin. Overall, the oxygen isotope compositions of mammals from the Yuanmou Basin indicate a general warming and/or drying of the local climate over time, which is in agreement with the carbon isotope results from that region.

121

APPENDIX A

DATA FROM ANALYSES OF BULK CARBONATE SAMPLES

Lab No. Sample Taxon Collection Locality δ13C δ18O Age Type (PDB) (PDB) (Ma) Lb-01 Enamel Modern Cow Linxia Basin, Gansu Province -11.0 -6.4 0.00 Lb-02 Enamel Modern Goat Linxia Basin, Gansu Province -12.0 -5.6 0.00 Ld-06 Enamel Gazella blacki Linxia Basin, Gansu Province -11.6 -4.4 2.50 Ld-07 Enamel Gazella blacki Linxia Basin, Gansu Province -10.5 -5.6 2.50 Ld-08 Enamel Leptobus amplifrontalis Linxia Basin, Gansu Province -10.7 -3.6 2.50 Ld-09 Enamel Leptobus amplifrontalis Linxia Basin, Gansu Province -10.5 -5.2 2.50 Sl-01 Enamel Gazella sp. Linxia Basin, Gansu Province -12.3 -5.9 4.00 Sl-02 Enamel Gazella sp. Linxia Basin, Gansu Province -9.7 -5.6 4.00 Sl-04 Enamel Gazella sp. Linxia Basin, Gansu Province -9.8 -3.8 4.00 Sl-05 Enamel Gazella sp. Linxia Basin, Gansu Province -10.5 -5.4 4.00 Sl-06 Enamel Gazella sp. Linxia Basin, Gansu Province -9.6 -8.0 4.00 E-12 Enamel Bovidae Linxia Basin, Gansu Province -8.8 -9.6 7.00 Ls-04 Enamel Gazella sp. Linxia Basin, Gansu Province -9.8 -4.0 7.00 Ls-11 Enamel Protoryx sp. Linxia Basin, Gansu Province -7.1 -4.9 7.50 Ls-13 Enamel Sinotragus sp. Linxia Basin, Gansu Province -10.1 -6.0 8.00 E-19 Enamel Gazella sp. Linxia Basin, Gansu Province -9.7 -7.3 9.00 E-28 Enamel Gazella sp. Linxia Basin, Gansu Province -9.1 -6.6 9.00 Ls-12 Enamel Hezhengia bohlini Linxia Basin, Gansu Province -9.1 -3.5 9.20 By-05 Enamel Equus hemionus Linxia Basin, Gansu Province -8.2 -5.8 0.05 Tz-02 Enamel Equus qingyangensis Linxia Basin, Gansu Province -5.0 -5.6 1.20 Ld-01 Enamel Equus sp. Linxia Basin, Gansu Province -10.5 -7.7 2.50 Ld-02 Enamel Equus sp. Linxia Basin, Gansu Province -10.6 -8.3 2.50 Ld-03 Enamel Equus sp. Linxia Basin, Gansu Province -9.9 -7.9 2.50 Ld-04 Enamel Equus sp. Linxia Basin, Gansu Province -11.9 -3.1 2.50 Ld-05 Enamel Equus sp. Linxia Basin, Gansu Province -11.9 -3.4 2.50 Ld-10 Enamel Equus sp. Linxia Basin, Gansu Province -9.5 -7.1 2.50 Ld-11 Enamel Equus sp. Linxia Basin, Gansu Province -8.5 -7.1 2.50 Ld-12 Enamel Equus sp. Linxia Basin, Gansu Province -8.8 -8.2 2.50 Ld-14 Enamel Equus sp. Linxia Basin, Gansu Province -9.6 -10.3 2.50 Ld-15 Enamel Equus sp. Linxia Basin, Gansu Province -9.3 -9.1 2.50 Ld-17 Enamel Equus sp. Linxia Basin, Gansu Province -9.2 -8.1 2.50 Sl-11 Enamel Hipparion sp. Linxia Basin, Gansu Province -10.3 -4.3 4.00 S1-12 Enamel Hipparion sp. Linxia Basin, Gansu Province -10.7 -4.3 4.00 S1-13 Enamel Hipparion sp. Linxia Basin, Gansu Province -10.1 -7.4 4.00 Sl-14 Enamel Hipparion sp. Linxia Basin, Gansu Province -10.2 -6.3 4.00 Sl-15 Enamel Hipparion sp. Linxia Basin, Gansu Province -10.0 -5.7 4.00 Sl-16 Enamel Hipparion sp. Linxia Basin, Gansu Province -10.0 -5.8 4.00 Ds-01 Enamel Hipparion dermatorhinum Linxia Basin, Gansu Province -10.6 -2.9 6.00

122

Lab No. Sample Taxon Collection Locality δ13C δ18O Age Type (PDB) (PDB) (Ma) Ds-02 Enamel Hipparion dermatorhinum Linxia Basin, Gansu Province -10.3 -2.6 6.00 Ds-03 Enamel Hipparion dermatorhinum Linxia Basin, Gansu Province -10.0 -5.4 6.00 Ds-04 Enamel Hipparion dermatorhinum Linxia Basin, Gansu Province -9.8 -4.4 6.00 Ds-05 Enamel Hipparion dermatorhinum Linxia Basin, Gansu Province -10.0 -3.6 6.00 Ds-06 Enamel Hipparion dermatorhinum Linxia Basin, Gansu Province -10.1 -4.6 6.00 Ds-07 Enamel Hipparion dermatorhinum Linxia Basin, Gansu Province -9.6 -3.9 6.00 Ds-08 Enamel Hipparion dermatorhinum Linxia Basin, Gansu Province -9.3 -4.5 6.00 Ds-09 Enamel Hipparion dermatorhinum Linxia Basin, Gansu Province -10.2 -3.9 6.00 Ds-10 Enamel Hipparion dermatorhinum Linxia Basin, Gansu Province -9.2 -4.6 6.00 E-11 Enamel Hipparion sp. Linxia Basin, Gansu Province -10.0 -8.4 7.00 Hl-01 Enamel Hipparion sp. Linxia Basin, Gansu Province -9.7 -5.2 7.50 Hl-02 Enamel Hipparion sp. Linxia Basin, Gansu Province -9.5 -6.3 7.50 Hl-03 Enamel Hipparion sp. Linxia Basin, Gansu Province -9.3 -3.2 7.50 Hl-04 Enamel Hipparion sp. Linxia Basin, Gansu Province -9.4 -5.2 7.50 Hl-05 Enamel Hipparion sp. Linxia Basin, Gansu Province -9.5 -4.0 7.50 Hl-06 Enamel Hipparion sp. Linxia Basin, Gansu Province -9.2 -6.1 7.50 Hl-07 Enamel Hipparion sp. Linxia Basin, Gansu Province -10.2 -6.8 7.50 Hl-08 Enamel Hipparion sp. Linxia Basin, Gansu Province -9.6 -3.3 7.50 Hl-09 Enamel Hipparion sp. Linxia Basin, Gansu Province -8.4 -4.3 7.50 Hl-10 Enamel Hipparion sp. Linxia Basin, Gansu Province -9.9 -5.1 7.50 E-21 Enamel Hipparion sp. Linxia Basin, Gansu Province -9.8 -5.8 9.00 E-22 Enamel Hipparion sp. Linxia Basin, Gansu Province -11.5 -9.2 9.00 E-26 Enamel Hipparion sp. Linxia Basin, Gansu Province -9.5 -5.2 9.00 Qj-01 Enamel Hipparion dermatorhinum Linxia Basin, Gansu Province -11.0 -8.4 9.50 Qj-02 Enamel Hipparion dermatorhinum Linxia Basin, Gansu Province -11.6 -9.5 9.50 Qj-03 Enamel Hipparion dermatorhinum Linxia Basin, Gansu Province -9.9 -8.6 9.50 Qj-04 Enamel Hipparion dermatorhinum Linxia Basin, Gansu Province -9.1 -7.2 9.50 Qj-05 Enamel Hipparion dermatorhinum Linxia Basin, Gansu Province -10.5 -7.2 9.50 Qj-06 Enamel Hipparion dermatorhinum Linxia Basin, Gansu Province -10.9 -9.9 9.50 Gn-01 Enamel Hipparion dongxiangense Linxia Basin, Gansu Province -9.1 -3.4 11.50 Gn-02 Enamel Hipparion dongxiangense Linxia Basin, Gansu Province -8.7 -3.0 11.50 Gn-03 Enamel Hipparion dongxiangense Linxia Basin, Gansu Province -9.3 -4.3 11.50 Gn-04 Enamel Hipparion dongxiangense Linxia Basin, Gansu Province -9.4 -2.3 11.50 Gn-05 Enamel Hipparion dongxiangense Linxia Basin, Gansu Province -10.5 -9.0 11.50 Gn-06 Enamel Hipparion dongxiangense Linxia Basin, Gansu Province -10.0 -3.0 11.50 Gn-08 Enamel Hipparion dongxiangense Linxia Basin, Gansu Province -9.0 -3.4 11.50 Gn-09 Enamel Hipparion dongxiangense Linxia Basin, Gansu Province -9.9 -1.8 11.50 Gn-10 Enamel Hipparion dongxiangense Linxia Basin, Gansu Province -9.6 -1.8 11.50 Lgo-04 Enamel Anchitherium sp. Linxia Basin, Gansu Province -8.8 -6.9 14.00 Wcg-01 Enamel Coelodonta nihowanensis Linxia Basin, Gansu Province -9.9 -6.5 2.55 Hw-01 Enamel Shansirhinus ringstroemi Linxia Basin, Gansu Province -9.8 -5.7 4.00 Ds-11 Enamel Chilotherium wimani Linxia Basin, Gansu Province -8.9 -4.9 6.00 Ds-12 Enamel Chilotherium wimani Linxia Basin, Gansu Province -9.4 -4.9 6.00 Ds-13 Enamel Chilotherium wimani Linxia Basin, Gansu Province -9.1 -4.6 6.00 Ds-14 Enamel Chilotherium wimani Linxia Basin, Gansu Province -9.8 -7.3 6.00 Ds-15 Enamel Chilotherium wimani Linxia Basin, Gansu Province -9.9 -7.1 6.00 Ls-15 Enamel Dicerorhinus ringstroemi Linxia Basin, Gansu Province -9.2 -4.6 6.00 E-15 Enamel Chilotherium wimani Linxia Basin, Gansu Province -11.6 -8.8 7.00 E-16 Enamel Chilotherium wimani Linxia Basin, Gansu Province -11.4 -11.0 7.00 Hl-11 Enamel Chilotherium wimani Linxia Basin, Gansu Province -10.8 -9.4 7.50 Hl-12 Enamel Chilotherium wimani Linxia Basin, Gansu Province -10.8 -9.5 7.50 Hl-13 Enamel Chilotherium wimani Linxia Basin, Gansu Province -9.6 -9.5 7.50 123

Lab No. Sample Taxon Collection Locality δ13C δ18O Age Type (PDB) (PDB) (Ma) Hl-14 Enamel Chilotherium wimani Linxia Basin, Gansu Province -9.7 -8.4 7.50 Hl-15 Enamel Chilotherium wimani Linxia Basin, Gansu Province -9.6 -8.4 7.50 Hl-16 Enamel Chilotherium wimani Linxia Basin, Gansu Province -9.7 -8.3 7.50 E-17 Enamel Chilotherium wimani Linxia Basin, Gansu Province -11.4 -8.8 9.00 E-25 Enamel Chilotherium wimani Linxia Basin, Gansu Province -11.2 -5.8 9.00 Ls-10 Enamel Acerorhinus hezhengensis Linxia Basin, Gansu Province -9.6 -6.3 9.00 Ls-06 Enamel Iranotherium morgani Linxia Basin, Gansu Province -11.0 -3.1 9.20 Qj-07 Enamel Chilotherium wimani Linxia Basin, Gansu Province -10.8 -9.8 9.50 Qj-08 Enamel Chilotherium wimani Linxia Basin, Gansu Province -11.4 -9.6 9.50 Qj-09 Enamel Chilotherium wimani Linxia Basin, Gansu Province -10.9 -9.0 9.50 Qj-10 Enamel Chilotherium wimani Linxia Basin, Gansu Province -11.5 -9.0 9.50 Gn-11 Enamel Parelasmotherium linxiaense Linxia Basin, Gansu Province -9.8 -4.6 11.50 Gn-12 Enamel Parelasmotherium linxiaense Linxia Basin, Gansu Province -10.0 -4.6 11.50 Gn-13 Enamel Parelasmotherium linxiaense Linxia Basin, Gansu Province -9.9 -4.7 11.50 Gn-14 Enamel Parelasmotherium linxiaense Linxia Basin, Gansu Province -9.8 -4.5 11.50 Gn-15 Enamel Parelasmotherium linxiaense Linxia Basin, Gansu Province -9.9 -4.5 11.50 Lg-01 Enamel Alicornops laogouense Linxia Basin, Gansu Province -11.7 -8.4 13.00 Lg-02 Enamel Alicornops laogouense Linxia Basin, Gansu Province -12.4 -9.4 13.00 Lg-03 Enamel Alicornops laogouense Linxia Basin, Gansu Province -9.8 -8.1 13.00 Lg-04 Enamel Alicornops laogouense Linxia Basin, Gansu Province -10.8 -10.4 13.00 Lg-05 Enamel Alicornops laogouense Linxia Basin, Gansu Province -8.9 -8.5 13.00 Lgo-05 Enamel Hispanotherium matritense Linxia Basin, Gansu Province -8.7 -7.1 14.00 Dl-01 Enamel Alicornops sp. Linxia Basin, Gansu Province -8.4 -8.0 17.00 Dl-02 Enamel Alicornops sp. Linxia Basin, Gansu Province -9.2 -8.1 17.00 Dl-03 Enamel Alicornops sp. Linxia Basin, Gansu Province -10.7 -8.7 17.00 Tl-01 Enamel Paraceratherium sp. Linxia Basin, Gansu Province -10.3 -12.0 25.00 Tl-02 Enamel Paraceratherium sp. Linxia Basin, Gansu Province -10.0 -10.9 25.00 Tl-03 Enamel Paraceratherium sp. Linxia Basin, Gansu Province -10.2 -11.5 25.00 Tl-04 Enamel Paraceratherium sp. Linxia Basin, Gansu Province -10.4 -11.6 25.00 Tl-05 Enamel Paraceratherium sp. Linxia Basin, Gansu Province -10.1 -11.4 25.00 Tl-06 Enamel Allacerops sp. Linxia Basin, Gansu Province -9.3 -7.1 25.00 Tl-07 Enamel Allacerops sp. Linxia Basin, Gansu Province -9.5 -8.2 25.00 Tl-08 Enamel Allacerops sp. Linxia Basin, Gansu Province -10.1 -9.5 25.00 Tl-09 Enamel Allacerops sp. Linxia Basin, Gansu Province -9.0 -6.8 25.00 Sl-07 Enamel Cervavitus novorassiae Linxia Basin, Gansu Province -9.0 -2.8 4.00 Sl-08 Enamel Cervavitus novorassiae Linxia Basin, Gansu Province -10.8 -0.3 4.00 E-13 Enamel Cervavitus novorassiae Linxia Basin, Gansu Province -10.9 -8.1 7.00 E-14 Enamel Cervavitus novorassiae Linxia Basin, Gansu Province -10.9 -7.8 7.00 E-20 Enamel Cervavitus novorassiae Linxia Basin, Gansu Province -9.9 -7.6 9.00 E-27 Enamel Cervavitus novorassiae Linxia Basin, Gansu Province -10.0 -6.9 9.00 Lgo-03 Enamel Turcocerus sp. Linxia Basin, Gansu Province -10.0 -6.5 14.00 Sl-09 Enamel Palaeotragus microdon Linxia Basin, Gansu Province -10.1 -7.8 4.00 Sl-10 Enamel Palaeotragus microdon Linxia Basin, Gansu Province -10.4 -7.7 4.00 Ls-01 Enamel Palaeotragus microdon Linxia Basin, Gansu Province -10.3 -3.6 7.50 Ls-02 Enamel Honanotherium schlosseri Linxia Basin, Gansu Province -9.2 -4.5 7.50 E-18 Enamel Samotherium sp. Linxia Basin, Gansu Province -10.0 -7.3 9.00 E-23 Enamel Samotherium sp. Linxia Basin, Gansu Province -9.7 -6.2 9.00 E-24 Enamel Samotherium sp. Linxia Basin, Gansu Province -9.6 -5.4 9.00 Ls-05 Enamel Samotherium sp. Linxia Basin, Gansu Province -9.4 -8.4 9.10 Ls-09 Enamel Microstonyx major Linxia Basin, Gansu Province -10.3 -9.7 7.50 Ls-14 Enamel Chlenastochoerus stehlini Linxia Basin, Gansu Province -10.7 -7.7 9.00 Lgo-01 Enamel Listriodon sp. Linxia Basin, Gansu Province -10.5 -7.6 14.00 124

Lab No. Sample Taxon Collection Locality δ13C δ18O Age Type (PDB) (PDB) (Ma) Lgo-02 Enamel Kubanochoerus gigas Linxia Basin, Gansu Province -9.7 -8.9 14.00 Ls-07 Enamel Tetralophodon exoletus Linxia Basin, Gansu Province -11.4 -8.7 9.10 Ls-03 Enamel Tetralophodon sp. Linxia Basin, Gansu Province -10.4 -6.4 10.00 Lg-06 Enamel Platebelodon grangeri Linxia Basin, Gansu Province -10.1 -8.8 13.00 Lg-07 Enamel Platebelodon grangeri Linxia Basin, Gansu Province -9.6 -8.7 13.00 Lg-08 Enamel Platebelodon grangeri Linxia Basin, Gansu Province -8.9 -7.9 13.00 Lg-09 Enamel Platebelodon grangeri Linxia Basin, Gansu Province -9.2 -7.3 13.00 Lg-10 Enamel Platebelodon grangeri Linxia Basin, Gansu Province -8.8 -7.0 13.00 Lgo-06 Enamel Zygolophodon sp. Linxia Basin, Gansu Province -9.6 -8.6 14.00 Dl-04 Enamel Gomphotherium sp. Linxia Basin, Gansu Province -9.2 -7.2 17.00 Dl-05 Enamel Gomphotherium sp. Linxia Basin, Gansu Province -9.5 -7.3 17.00 Dl-06 Enamel Gomphotherium sp. Linxia Basin, Gansu Province -9.0 -6.5 17.00 Dl-07 Enamel Gomphotherium sp. Linxia Basin, Gansu Province -8.7 -7.4 17.00 Dl-08 Enamel Gomphotherium sp. Linxia Basin, Gansu Province -8.7 -7.4 17.00 Dl-09 Enamel Gomphotherium sp. Linxia Basin, Gansu Province -9.3 -6.7 17.00 Dl-10 Enamel Gomphotherium sp. Linxia Basin, Gansu Province -9.2 -6.7 17.00 TS-1 Enamel Chilotherium wimani Tianshui, Gansu Province -8.9 -6.8 9.00 TS-2 Enamel Chilotherium wimani Tianshui, Gansu Province -8.7 -6.5 9.00 TS-3 Enamel Chilotherium wimani Tianshui, Gansu Province -8.5 -6.4 9.00 TS-4 Enamel Chilotherium wimani Tianshui, Gansu Province -8.4 -7.2 9.00 TS-5 Enamel Chilotherium wimani Tianshui, Gansu Province -8.1 -6.7 9.00 E-10 Enamel Aprotodon lanzhouensis Lanzhou Basin, Gansu Province -10.6 -12.5 20.40 YG-1 Enamel Rhino Baoshan, Yunnan Province -14.8 -8.5 0.009 YG-2 Enamel Rhino Baoshan, Yunnan Province -14.1 -8.3 0.009 YG-12 Enamel Elephant Baoshan, Yunnan Province -11.5 -8.4 5.00 Yg-28 Enamel Stegolophodon baoshan Baoshan, Yunnan Province -11.9 -11.1 5.30 YG-23 Enamel Pig Kaiyuan, Yunnan Province -9.9 -10.3 10 Yg-30 Enamel Stegodon yuxiensis Kaiyuan, Yunnan Province -14.1 -8.8 4.00 YI-24 Enamel Tapir Kaiyuan, Yunnan Province -11.3 -6.9 10.00 LF-4 Enamel Bovid Lufeng, Yunnan Province -10.4 -8.7 8.00 LF-25 Enamel Bovid Lufeng, Yunnan Province -12.1 -12.1 8.00 LF-27 Enamel Bovid Lufeng, Yunnan Province -11.1 -9.7 8.00 LF-33 Enamel Bovid Lufeng, Yunnan Province -11.0 -9.3 8.00 LF-36 Enamel Cow Lufeng, Yunnan Province -12.5 -11.9 8.00 LF-37 Enamel Cow Lufeng, Yunnan Province -12.1 -12.8 8.00 LF-48 Enamel Bovid Lufeng, Yunnan Province -13.2 -11.6 8.00 LF-49 Enamel Bovid Lufeng, Yunnan Province -11.6 -13.8 8.00 LF-52 Enamel Bovid Lufeng, Yunnan Province -11.5 -11.8 8.00 LF-55 Enamel Bovid Lufeng, Yunnan Province -11.5 -11.0 8.00 YN-1 Enamel Cormohipparion sp. Lufeng, Yunnan Province -13.8 -9.8 7.50 YN-1* Enamel Cormohipparion sp. Lufeng, Yunnan Province -11.9 -13.6 7.50 LF-21 Enamel Hipparion sp. Lufeng, Yunnan Province -13.0 -10.6 8.00 LF-22 Enamel Hipparion sp. Lufeng, Yunnan Province -13.9 -13.0 8.00 LF-45 Enamel Hipparion sp. Lufeng, Yunnan Province -14.1 -12.6 8.00 LF-10 Enamel Rhino Lufeng, Yunnan Province -11.1 -11.5 8.00 LF-10 Enamel Rhino Lufeng, Yunnan Province -11.4 -11.5 8.00 LF-35 Enamel Rhino Lufeng, Yunnan Province -10.9 -12.9 8.00 LF-51 Enamel Rhino Lufeng, Yunnan Province -14.2 -11.1 8.00 YI-1 Enamel Rhino Lufeng, Yunnan Province -13.4 -11.4 8.00 YI-35 Enamel Rhino Lufeng, Yunnan Province -10.5 -11.7 8.00 YI-36 Enamel Rhino Lufeng, Yunnan Province -13.1 -12.1 8.00 YI-32 Enamel Chalicothere Lufeng, Yunnan Province -10.4 -6.8 8.00 125

Lab No. Sample Taxon Collection Locality δ13C δ18O Age Type (PDB) (PDB) (Ma) LF-3 Enamel Deer Lufeng, Yunnan Province -12.0 -6.0 8.00 LF-5 Enamel Deer Lufeng, Yunnan Province -10.5 -8.4 8.00 LF-11 Enamel Deer Lufeng, Yunnan Province -11.7 -11.4 8.00 LF-23 Enamel Deer Lufeng, Yunnan Province -13.6 -9.4 8.00 LF-24 Enamel Deer Lufeng, Yunnan Province -13.2 -12.5 8.00 LF-26 Enamel Deer Lufeng, Yunnan Province -13.5 -14.2 8.00 LF-28 Enamel Deer Lufeng, Yunnan Province -11.7 -9.3 8.00 LF-29 Enamel Deer Lufeng, Yunnan Province -11.9 -10.3 8.00 LF-30 Enamel Deer Lufeng, Yunnan Province -11.6 -12.2 8.00 LF-31 Enamel Deer Lufeng, Yunnan Province -12.4 -8.9 8.00 LF-32 Enamel Deer Lufeng, Yunnan Province -11.4 -15.6 8.00 LF-34 Enamel Deer Lufeng, Yunnan Province -13.0 -12.1 8.00 LF-38 Enamel Deer Lufeng, Yunnan Province -12.0 -11.5 8.00 LF-39 Enamel Deer Lufeng, Yunnan Province -12.2 -12.5 8.00 LF-40 Enamel Deer Lufeng, Yunnan Province -10.7 -11.6 8.00 LF-41 Enamel Deer Lufeng, Yunnan Province -10.6 -13.2 8.00 LF-42 Enamel Deer Lufeng, Yunnan Province -11.0 -13.1 8.00 LF-43 Enamel Deer Lufeng, Yunnan Province -12.5 -7.1 8.00 LF-44 Enamel Deer Lufeng, Yunnan Province -11.0 -9.5 8.00 LF-46 Enamel Deer Lufeng, Yunnan Province -11.0 -12.8 8.00 LF-47 Enamel Deer Lufeng, Yunnan Province -13.0 -13.7 8.00 LF-50 Enamel Deer Lufeng, Yunnan Province -12.3 -10.1 8.00 LF-53 Enamel Deer Lufeng, Yunnan Province -12.7 -12.4 8.00 LF-54 Enamel Deer Lufeng, Yunnan Province -11.5 -12.8 8.00 YI-33 Enamel Deer Lufeng, Yunnan Province -13.6 -11.2 8.00 YG-13 Enamel Elephant Lufeng, Yunnan Province -10.9 -8.7 7.50 YI-34 Enamel Bear Lufeng, Yunnan Province -12.6 -9.4 8.00 LF-1 Enamel Herbivore Lufeng, Yunnan Province -11.5 -5.9 8.00 LF-2 Enamel Herbivore Lufeng, Yunnan Province -13.9 -7.2 8.00 LF-12 Enamel Herbivore Lufeng, Yunnan Province -11.5 -11.5 8.00 LF-13 Enamel Herbivore Lufeng, Yunnan Province -11.7 -8.1 8.00 LF-14 Enamel Herbivore Lufeng, Yunnan Province -11.6 -7.3 8.00 LF-15 Enamel Herbivore Lufeng, Yunnan Province -11.4 -7.7 8.00 LF-16 Enamel Herbivore Lufeng, Yunnan Province -10.8 -8.4 8.00 SGL-21 Enamel Bovid Shangri-La, Yunnan Province -9.7 -16.4 2.50 SGL-27 Enamel Bovid Shangri-La, Yunnan Province -7.7 -15.3 2.50 SGL-28 Enamel Bovid Shangri-La, Yunnan Province -4.9 -10.8 2.50 SGL-35 Enamel Bovid Shangri-La, Yunnan Province -9.5 -14.0 2.50 SGL-20 Enamel Horse Shangri-La, Yunnan Province -9.7 -15.3 1.75 SGL-1 Enamel Horse Shangri-La, Yunnan Province -8.5 -16.3 2.50 1 Enamel Rhino Shangri-La, Yunnan Province -11.6 -13.2 2.50 2 Enamel Rhino Shangri-La, Yunnan Province -11.7 -13.2 2.50 SGL-31 Enamel Rhino Shangri-La, Yunnan Province -6.3 -17.2 2.50 SGL-44 Enamel Deer Shangri-La, Yunnan Province -10.5 -13.0 0.00 SGL-22 Enamel Deer Shangri-La, Yunnan Province -1.1 -16.4 2.50 SGL-25 Enamel Deer Shangri-La, Yunnan Province -5.1 -17.1 2.50 SGL-43 Enamel Pig Shangri-La, Yunnan Province -8.1 -13.3 0.00 SGL-3 Enamel Pig Shangri-La, Yunnan Province -10.2 -8.8 2.50 SGL-4 Enamel Herbivore Shangri-La, Yunnan Province -10.6 -12.5 1.75 SGL-5 Enamel Herbivore Shangri-La, Yunnan Province -8.2 -7.4 1.75 SGL-6 Enamel Herbivore Shangri-La, Yunnan Province -4.7 -14.6 1.75 SGL-7 Enamel Herbivore Shangri-La, Yunnan Province -4.3 -16.7 1.75 126

Lab No. Sample Taxon Collection Locality δ13C δ18O Age Type (PDB) (PDB) (Ma) SGL-8 Enamel Herbivore Shangri-La, Yunnan Province -2.7 -9.3 1.75 SGL-9 Enamel Herbivore Shangri-La, Yunnan Province -10.4 -11.4 1.75 SGL-10 Enamel Herbivore Shangri-La, Yunnan Province -4.3 -16.6 1.75 SGL-11 Enamel Herbivore Shangri-La, Yunnan Province -3.3 -15.3 1.75 SGL-12 Enamel Herbivore Shangri-La, Yunnan Province -3.7 -12.3 1.75 SGL-13 Enamel Herbivore Shangri-La, Yunnan Province -6.6 -13.4 1.75 SGL-14 Enamel Herbivore Shangri-La, Yunnan Province -6.5 -16.6 1.75 SGL-15 Enamel Herbivore Shangri-La, Yunnan Province -10.4 -13.5 1.75 SGL-16 Enamel Herbivore Shangri-La, Yunnan Province -0.2 -16.6 1.75 SGL-18 Enamel Herbivore Shangri-La, Yunnan Province -5.6 -14.4 1.75 SGL-23 Enamel Herbivore Shangri-La, Yunnan Province -6.2 -14.2 2.50 SGL-24 Enamel Herbivore Shangri-La, Yunnan Province -10.2 -14.8 2.50 SGL-26 Enamel Herbivore Shangri-La, Yunnan Province -10.6 -12.0 2.50 SGL-30 Enamel Herbivore Shangri-La, Yunnan Province -5.9 -14.5 2.50 SGL-34 Enamel Herbivore Shangri-La, Yunnan Province -5.9 -12.0 2.50 SGL-36 Enamel Herbivore Shangri-La, Yunnan Province -11.4 -12.1 2.50 SGL-39 Enamel Herbivore Shangri-La, Yunnan Province -10.2 -9.3 2.50 YG-3 Enamel Equus yunnanensis Yuanmou Basin, Yunnan Province -9.4 -6.1 1.75 YG-4 Enamel Equus yunnanensis Yuanmou Basin, Yunnan Province -4.7 -8.1 1.75 YG-21 Enamel Hipparion sp. Yuanmou Basin, Yunnan Province -13.2 -11.2 8.00 YN-2 Enamel Cormohipparion sp. Yuanmou Basin, Yunnan Province -13.2 -10.8 8.00 YN-3 Enamel Cormohipparion sp. Yuanmou Basin, Yunnan Province -13.3 -11.0 8.00 YN-4 Enamel Cormohipparion sp. Yuanmou Basin, Yunnan Province -15.3 -12.6 8.00 YN-5 Enamel Cormohipparion sp. Yuanmou Basin, Yunnan Province -14.4 -12.1 8.00 YN-6 Enamel Cormohipparion sp. Yuanmou Basin, Yunnan Province -14.4 -11.9 8.00 GP-02 Enamel Hipparion sp. Yuanmou Basin, Yunnan Province -14.4 -13.0 8.15 XH-03 Enamel Hipparion sp. Yuanmou Basin, Yunnan Province -14.4 -11.6 8.15 YG-5 Enamel Rhino Yuanmou Basin, Yunnan Province -12.9 -10.9 3.50 YM-SG-1 Enamel Siwarhino Yuanmou Basin, Yunnan Province -10.2 -12.0 4.00 YM-SG-2 Enamel Siwarhino Yuanmou Basin, Yunnan Province -12.8 -10.7 4.00 YG-6 Enamel Rhino Yuanmou Basin, Yunnan Province -12.2 -9.8 5.00 YG-7 Enamel Rhino Yuanmou Basin, Yunnan Province -12.3 -8.4 5.00 YG-8 Enamel Rhino Yuanmou Basin, Yunnan Province -12.5 -9.7 5.00 YG-9 Enamel Rhino Yuanmou Basin, Yunnan Province -12.3 -10.1 5.00 YG-10 Enamel Rhino Yuanmou Basin, Yunnan Province -10.8 -10.0 5.00 YG-11 Enamel Rhino Yuanmou Basin, Yunnan Province -13.2 -5.7 5.00 YG-17 Enamel Rhino Yuanmou Basin, Yunnan Province -13.7 -12.3 8.00 YG-18 Enamel Rhino Yuanmou Basin, Yunnan Province -12.6 -6.9 8.00 YG-19 Enamel Rhino Yuanmou Basin, Yunnan Province -14.6 -11.8 8.00 YG-20 Enamel Rhino Yuanmou Basin, Yunnan Province -13.9 -8.2 8.00 XH-02 Enamel Acerorhinus sp. Yuanmou Basin, Yunnan Province -14.5 -12.0 8.15 YI-2 Enamel Rhino Yuanmou Basin, Yunnan Province -13.8 -9.8 8.15 YI-3 Enamel Rhino Yuanmou Basin, Yunnan Province -13.5 -11.3 8.15 YI-4 Enamel Rhino Yuanmou Basin, Yunnan Province -14.3 -11.4 8.15 YI-5 Enamel Rhino Yuanmou Basin, Yunnan Province -12.7 -9.3 8.15 YI-9 Enamel Rhino Yuanmou Basin, Yunnan Province -14.9 -9.0 8.15 YI-10 Enamel Rhino Yuanmou Basin, Yunnan Province -14.1 -5.9 8.15 YI-11 Enamel Rhino Yuanmou Basin, Yunnan Province -12.3 -8.8 8.15 YI-12 Enamel Rhino Yuanmou Basin, Yunnan Province -13.4 -10.9 8.15 YI-13 Enamel Rhino Yuanmou Basin, Yunnan Province -15.2 -11.1 8.15 YI-14 Enamel Rhino Yuanmou Basin, Yunnan Province -13.1 -12.1 8.15 YI-15 Enamel Rhino Yuanmou Basin, Yunnan Province -15.4 -9.0 8.15 127

Lab No. Sample Taxon Collection Locality δ13C δ18O Age Type (PDB) (PDB) (Ma) YI-16 Enamel Rhino Yuanmou Basin, Yunnan Province -13.6 -8.2 8.15 YI-18 Enamel Rhino Yuanmou Basin, Yunnan Province -11.5 -8.5 8.15 YI-21 Enamel Rhino Yuanmou Basin, Yunnan Province -13.8 -7.0 8.15 Yg-34 Enamel Rhino Yuanmou Basin, Yunnan Province -14.0 -10.8 8.50 YMH-2 Enamel Rhino Yuanmou Basin, Yunnan Province -14.7 -11.6 8.50 YMH-3 Enamel Rhino Yuanmou Basin, Yunnan Province -13.7 -10.4 8.50 YI-6 Enamel Chalicothere Yuanmou Basin, Yunnan Province -11.4 -8.6 8.15 YI-19 Enamel Chalicothere Yuanmou Basin, Yunnan Province -10.7 -7.9 8.15 YI-20 Enamel Chalicothere Yuanmou Basin, Yunnan Province -13.1 -8.9 8.15 GP-01 Enamel Yunnanotherium sp. Yuanmou Basin, Yunnan Province -12.0 -12.6 8.15 GP-03 Enamel Yunnanotherium sp. Yuanmou Basin, Yunnan Province -11.6 -8.2 8.15 GP-04 Enamel Yunnanotherium sp. Yuanmou Basin, Yunnan Province -12.1 -15.8 8.15 YI-26 Enamel Tragulidae Yuanmou Basin, Yunnan Province -11.2 -11.0 8.15 YI-27 Enamel Tragulidae Yuanmou Basin, Yunnan Province -12.4 -15.8 8.15 YI-28 Enamel Tragulidae Yuanmou Basin, Yunnan Province -11.9 -13.1 8.15 YMH-4 Enamel Pig Yuanmou Basin, Yunnan Province -8.1 -8.3 0.00 GP-05 Enamel Molarocherus Yunnanensis Yuanmou Basin, Yunnan Province -11.3 -11.1 8.15 YI-8 Enamel Hippopotamodon sp. Yuanmou Basin, Yunnan Province -10.9 -12.5 8.15 YI-17 Enamel Hippopotamodon sp. Yuanmou Basin, Yunnan Province -12.6 -13.7 8.15 YI-22 Enamel Chleustochoerus sp. Yuanmou Basin, Yunnan Province -11.6 -12.7 8.15 YI-23 Enamel Chleustochoerus sp. Yuanmou Basin, Yunnan Province -11.9 -10.7 8.15 YI-25 Enamel Molarochoerus sp. Yuanmou Basin, Yunnan Province -11.0 -13.4 8.15 YI-30 Enamel Hippopotamodon sp. Yuanmou Basin, Yunnan Province -10.9 -11.4 8.15 YI-31 Enamel Hippopotamodon sp. Yuanmou Basin, Yunnan Province -10.5 -9.3 8.15 Yg-29 Enamel Stegolophodon sp. Yuanmou Basin, Yunnan Province -13.2 -14.3 4.00 YG-14 Enamel Elephant Yuanmou Basin, Yunnan Province -12.2 -8.6 8.00 YG-15 Enamel Elephant Yuanmou Basin, Yunnan Province -13.0 -9.8 8.00 YG-16 Enamel Elephant Yuanmou Basin, Yunnan Province -13.1 -8.3 8.00 Yg-38 Enamel Elephant Yuanmou Basin, Yunnan Province -11.9 -14.3 8.50 YMH-1 Enamel Elephant Yuanmou Basin, Yunnan Province -12.1 -14.2 8.50 YI-7 Enamel Tapirus sp. Yuanmou Basin, Yunnan Province -10.5 -5.0 8.15 Yg-27 Enamel Stegodon zhaotongensis Zhaotong, Yunnan Province -11.5 -14.2 4.00 Yg-27-1 Enamel Stegodon zhaotongensis Zhaotong, Yunnan Province -11.8 -12.6 4.00 Yg-27-2 Enamel Stegodon zhaotongensis Zhaotong, Yunnan Province -11.8 -12.5 4.00 Yg-27-3 Enamel Stegodon zhaotongensis Zhaotong, Yunnan Province -11.6 -11.6 4.00

Lab No. Sample Type Sample Description Collection Locality δ13C δ18O Age (PDB) (PDB) (Ma) YPS-1 Soil Carbonate Soil Carbonate Nodule Yuanmou Basin, Yunnan -4.8 -12.0 7.10 YPS-1 Soil Carbonate Soil Carbonate Nodule Yuanmou Basin, Yunnan -5.1 -12.0 7.10 YPS-2 Soil Carbonate Soil Carbonate Nodule Yuanmou Basin, Yunnan -7.2 -12.2 7.10 YPS-2 Soil Carbonate Soil Carbonate Nodule Yuanmou Basin, Yunnan -7.3 -12.3 7.10 YPS-3 Soil Carbonate Soil Carbonate Nodule Yuanmou Basin, Yunnan -7.4 -13.5 7.10 YPS-5 Soil Carbonate Soil Carbonate Nodule Yuanmou Basin, Yunnan -7.1 -12.6 7.10 YPS-8 Soil Carbonate Soil Carbonate Nodule Yuanmou Basin, Yunnan -3.2 -11.9 1.70 YPS-9 Soil Carbonate Soil Carbonate Nodule Yuanmou Basin, Yunnan -4.9 -13.0 1.70 YPS-9 Soil Carbonate Soil Carbonate Nodule Yuanmou Basin, Yunnan -5.1 -12.9 1.70 YPS-10 Soil Carbonate Soil Carbonate Nodule Yuanmou Basin, Yunnan -8.6 -13.4 1.70 YPS-11 Soil Carbonate Soil Carbonate Nodule Yuanmou Basin, Yunnan -1.4 -12.3 1.70

128

Lab No. Sample Type Sample Description Collection Locality δ13C δ18O Age (PDB) (PDB) (Ma) By-05* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -4.1 -5.7 0.05 Tz-02* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -4.1 -8.6 1.20 Ld-11* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -7.0 -8.0 2.50 Ld-13* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -7.3 -9.0 2.50 Ld-14* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -7.0 -7.8 2.50 Ld-15* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -7.1 -8.0 2.50 Ld-16* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -7.5 -7.7 2.50 Ld-17* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -7.1 -7.8 2.50 SL-11*cc Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -6.9 -9.5 4.00 SL-12cc Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -7.0 -9.6 4.00 Ds-01* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -9.5 -9.0 6.00 Ds-02* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -7.3 -9.8 6.00 Ds-05* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -10.4 -9.2 6.00 Ds-06* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -8.4 -8.8 6.00 Ds-07* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -8.3 -9.1 6.00 Ds-11* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -8.8 -8.0 6.00 Ds-12* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -7.8 -8.3 6.00 Hl-03* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -8.5 -9.5 7.50 Hl-06* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -8.6 -9.6 7.50 Hl-09* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -8.9 -9.6 7.50 Hl-10* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -8.3 -9.6 7.50 Hl-11* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -8.0 -9.2 7.50 Hl-12* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -7.7 -8.7 7.50 Hl-14* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -7.0 -8.0 7.50 Hl-15* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -7.0 -8.5 7.50 Hl-15-cc Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -6.1 -9.1 7.50 Ls-01 cc Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -6.7 -10.1 7.50 Qj-06* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -7.4 -9.9 9.50 Gn-06* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -7.7 -8.7 11.50 Gn-07* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -7.6 -9.5 11.50 Gn-08* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -8.2 -8.4 11.50 Gn-09* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -7.7 -10.5 11.50 Gn-10* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -7.5 -9.6 11.50 Gn-12* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -7.0 -7.9 11.50 Gn-14* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -6.1 -7.8 11.50 Lgo-03* Soil Carbonate Matrix from Tooth Surface Linxia Basin, Gansu -5.6 -9.6 14.00 Yg-28* Soil Carbonate Matrix from Tooth Surface Baoshan, Yunnan -1.5 -10.2 5.30 SGL-20* Soil Carbonate Matrix from Tooth Surface Shangri-La, Yunnan -1.2 -16.0 1.75 SGL-1* Soil Carbonate Matrix from Tooth Surface Shangri-La, Yunnan -7.1 -17.4 2.50 SGL-3* Soil Carbonate Matrix from Tooth Surface Shangri-La, Yunnan -7.1 -16.4 2.50 GP-01* Soil Carbonate Matrix from Tooth Surface Yuanmou Basin, Yunnan -8.6 -11.8 8.15 GP-02* Soil Carbonate Matrix from Tooth Surface Yuanmou Basin, Yunnan -8.5 -11.5 8.15 GP-03* Soil Carbonate Matrix from Tooth Surface Yuanmou Basin, Yunnan -8.8 -13.6 8.15 GP-04* Soil Carbonate Matrix from Tooth Surface Yuanmou Basin, Yunnan -9.1 -14.7 8.15 GP-05* Soil Carbonate Matrix from Tooth Surface Yuanmou Basin, Yunnan -8.9 -13.5 8.15 XH-03* Soil Carbonate Matrix from Tooth Surface Yuanmou Basin, Yunnan -8.9 -10.1 8.15 Yg-34* Soil Carbonate Matrix from Tooth Surface Yuanmou Basin, Yunnan -11.0 -12.7 8.50 Yg-38* Soil Carbonate Matrix from Tooth Surface Yuanmou Basin, Yunnan -12.8 -14.6 8.50 YMH-1* Soil Carbonate Matrix from Tooth Surface Yuanmou Basin, Yunnan -8.9 -11.7 8.50

129

APPENDIX B

DATA FROM ANALYSES OF SERIAL CARBONATE SAMPLES

Sample No. Taxon Collection Locality Distance from δ13C δ18O Age Crown (mm) (Ma) By-05-s0 Equus hemionus Beiyuan Loc., Malan loess 1.00 -6.9 -6.3 0.05 By-05-S1a Equus hemionus Beiyuan Loc., Malan loess 4.00 -7.1 -6.0 0.05 By-05-S2a Equus hemionus Beiyuan Loc., Malan loess 7.00 -7.5 -6.0 0.05 By-05-s3a Equus hemionus Beiyuan Loc., Malan loess 10.00 -7.5 -5.6 0.05 By-05-s4a Equus hemionus Beiyuan Loc., Malan loess 13.00 -7.4 -5.3 0.05 By-05-s5a Equus hemionus Beiyuan Loc., Malan loess 16.00 -7.4 -5.3 0.05 By-05-s6a Equus hemionus Beiyuan Loc., Malan loess 19.00 -7.3 -5.6 0.05 By-05-s7a Equus hemionus Beiyuan Loc., Malan loess 22.00 -6.8 -6.4 0.05 By-05-s8a Equus hemionus Beiyuan Loc., Malan loess 25.00 -7.7 -6.5 0.05 By-05-s9a Equus hemionus Beiyuan Loc., Malan loess 28.00 -8.9 -5.5 0.05 By-05-s11 Equus hemionus Beiyuan Loc., Malan loess 31.00 -8.3 -5.6 0.05 Tz-02-2-s2 Equus qingyangensis Taizicun Loc., Lishi loess 2.80 -4.6 -6.6 1.2 Tz-02-2-s3 Equus qingyangensis Taizicun Loc., Lishi loess 4.60 -3.9 -5.6 1.2 Tz-02-2-s4 Equus qingyangensis Taizicun Loc., Lishi loess 6.40 -3.5 -5.7 1.2 Tz-02-2-s5 Equus qingyangensis Taizicun Loc., Lishi loess 8.20 -3.5 -5.7 1.2 Tz-02-2-s6 Equus qingyangensis Taizicun Loc., Lishi loess 10.00 -3.4 -8.5 1.2 Tz-02-2-s8 Equus qingyangensis Taizicun Loc., Lishi loess 13.60 -2.8 -11.3 1.2 Tz-02-2-s9 Equus qingyangensis Taizicun Loc., Lishi loess 15.40 -2.5 -11.4 1.2 Tz-02-2-s10 Equus qingyangensis Taizicun Loc., Lishi loess 17.20 -2.7 -10.5 1.2 Tz-02-2-s11 Equus qingyangensis Taizicun Loc., Lishi loess 19.00 -2.7 -13.0 1.2 Tz-02-2-s12 Equus qingyangensis Taizicun Loc., Lishi loess 20.80 -3.1 -10.0 1.2 Tz-02-2-s13 Equus qingyangensis Taizicun Loc., Lishi loess 22.60 -3.2 -12.1 1.2 Tz-02-2-s14 Equus qingyangensis Taizicun Loc., Lishi loess 24.40 -3.6 -7.6 1.2 Tz-02-2-s15 Equus qingyangensis Taizicun Loc., Lishi loess 26.20 -3.9 -6.5 1.2 Tz-02-2-s16 Equus qingyangensis Taizicun Loc., Lishi loess 28.00 -4.2 -6.5 1.2 Tz-02-2-s17 Equus qingyangensis Taizicun Loc., Lishi loess 29.80 -4.8 -6.0 1.2 Tz-02-2-s18 Equus qingyangensis Taizicun Loc., Lishi loess 31.60 -4.9 -6.1 1.2 Tz-02-2-s19 Equus qingyangensis Taizicun Loc., Lishi loess 33.40 -5.8 -5.3 1.2 Tz-02-2-s20 Equus qingyangensis Taizicun Loc., Lishi loess 35.20 -5.6 -5.4 1.2 Tz-02-2-s21 Equus qingyangensis Taizicun Loc., Lishi loess 37.00 -5.9 -5.4 1.2 Tz-02-2-s22 Equus qingyangensis Taizicun Loc., Lishi loess 38.80 -6.2 -8.6 1.2 Tz-02-2-s23 Equus qingyangensis Taizicun Loc., Lishi loess 40.60 -6.0 -9.2 1.2 Tz-02-2-s24 Equus qingyangensis Taizicun Loc., Lishi loess 42.40 -6.4 -10.2 1.2 Tz-02-2-s25 Equus qingyangensis Taizicun Loc., Lishi loess 44.20 -7.0 -9.2 1.2 Tz-02-2-s26 Equus qingyangensis Taizicun Loc., Lishi loess 46.00 -7.2 -10.6 1.2 Ld-10-2-s4 Equus sp. Longdan, Linxia 6.10 -9.2 -7.4 2.5 Ld-10-2-s5 Equus sp. Longdan, Linxia 7.80 -9.3 -7.7 2.5 Ld-10-2-s6 Equus sp. Longdan, Linxia 9.50 -9.4 -7.7 2.5

130

Sample No. Taxon Collection Locality Distance from δ13C δ18O Age Crown (mm) (Ma) Ld-10-2-s7 Equus sp. Longdan, Linxia 11.20 -9.3 -7.2 2.5 Ld-10-2-s8 Equus sp. Longdan, Linxia 12.90 -9.3 -7.1 2.5 Ld-10-2-s9 Equus sp. Longdan, Linxia 14.60 -9.3 -6.8 2.5 Ld-10-2-s10 Equus sp. Longdan, Linxia 16.30 -9.3 -6.9 2.5 Ld-10-2-s11 Equus sp. Longdan, Linxia 18.00 -9.4 -6.2 2.5 Ld-10-2-s12 Equus sp. Longdan, Linxia 19.70 -9.3 -6.1 2.5 Ld-10-2-s13 Equus sp. Longdan, Linxia 21.40 -9.4 -6.3 2.5 Ld-10-2-s14 Equus sp. Longdan, Linxia 23.10 -9.4 -6.4 2.5 Ld-10-2-s15 Equus sp. Longdan, Linxia 24.80 -9.4 -6.6 2.5 Ld-10-2-s16 Equus sp. Longdan, Linxia 26.50 -9.5 -7.0 2.5 Ld-10-2-s17 Equus sp. Longdan, Linxia 28.20 -9.6 -7.1 2.5 Ld-10-2-s18 Equus sp. Longdan, Linxia 29.90 -9.7 -7.3 2.5 Ld-10-2-s19 Equus sp. Longdan, Linxia 31.60 -9.7 -7.5 2.5 Ld-10-2-s20 Equus sp. Longdan, Linxia 33.30 -9.9 -8.1 2.5 Ld-10-2-s21 Equus sp. Longdan, Linxia 35.00 -10.0 -8.3 2.5 Ld-10-2-s22 Equus sp. Longdan, Linxia 36.70 -10.1 -8.5 2.5 Ld-10-2-s23 Equus sp. Longdan, Linxia 38.40 -10.1 -8.6 2.5 Ld-10-2-s24 Equus sp. Longdan, Linxia 40.10 -10.3 -8.5 2.5 Ld-10-2-s25 Equus sp. Longdan, Linxia 41.80 -10.3 -8.5 2.5 Ld-11-s1 Equus sp. Longdan, Linxia 1.00 -9.9 -6.1 2.5 Ld-11-s2 Equus sp. Longdan, Linxia 3.00 -9.6 -6.1 2.5 Ld-11-s3 Equus sp. Longdan, Linxia 5.00 -9.4 -6.2 2.5 Ld-11-s4 Equus sp. Longdan, Linxia 7.00 -9.5 -6.1 2.5 Ld-11-s5 Equus sp. Longdan, Linxia 9.00 -9.2 -5.5 2.5 Ld-11-s6 Equus sp. Longdan, Linxia 11.00 -9.2 -5.2 2.5 Ld-11-s7 Equus sp. Longdan, Linxia 13.00 -9.2 -5.0 2.5 Ld-11-s8 Equus sp. Longdan, Linxia 15.00 -9.2 -5.1 2.5 Ld-11-s9 Equus sp. Longdan, Linxia 17.00 -9.3 -5.2 2.5 Ld-11-s10 Equus sp. Longdan, Linxia 19.00 -9.4 -5.1 2.5 Ld-11-s11 Equus sp. Longdan, Linxia 21.00 -9.4 -5.4 2.5 Ld-11-s12 Equus sp. Longdan, Linxia 23.00 -9.1 -5.3 2.5 Ld-11-s13 Equus sp. Longdan, Linxia 25.00 -9.1 -5.4 2.5 Ld-11-s14 Equus sp. Longdan, Linxia 27.00 -9.0 -5.4 2.5 Ld-11-s15 Equus sp. Longdan, Linxia 29.00 -8.9 -5.8 2.5 Ld-11-s16 Equus sp. Longdan, Linxia 31.00 -9.3 -6.1 2.5 Ld-11-s17 Equus sp. Longdan, Linxia 33.00 -9.1 -6.5 2.5 Ld-11-s18 Equus sp. Longdan, Linxia 35.00 -8.8 -6.6 2.5 Ld-11-s19 Equus sp. Longdan, Linxia 37.00 -8.8 -7.0 2.5 Shl-1-s2a Hipparion sp. Shilidong, Linxia 6.20 -10.6 -5.1 4 Shl-1-s3a Hipparion sp. Shilidong, Linxia 8.80 -10.7 -3.8 4 Shl-1-s4a Hipparion sp. Shilidong, Linxia 11.40 -10.7 -4.0 4 Shl-1-s5a Hipparion sp. Shilidong, Linxia 14.00 -10.6 -4.8 4 Shl-1-s6a Hipparion sp. Shilidong, Linxia 16.60 -10.2 -4.1 4 Shl-1-s7a Hipparion sp. Shilidong, Linxia 19.20 -10.2 -4.6 4 Shl-1-s8a Hipparion sp. Shilidong, Linxia 21.80 -10.3 -5.7 4 Shl-1-s9a Hipparion sp. Shilidong, Linxia 24.40 -10.6 -4.3 4 Shl-1-s11 Hipparion sp. Shilidong, Linxia 28.30 -10.6 -6.3 4 Shl-1-s11a Hipparion sp. Shilidong, Linxia 29.60 -10.5 -5.5 4 Shl-1-s12 Hipparion sp. Shilidong, Linxia 30.90 -10.4 -6.0 4 Shl-1-s12a Hipparion sp. Shilidong, Linxia 32.20 -10.2 -6.3 4 Shl-1-s13 Hipparion sp. Shilidong, Linxia 33.50 -10.4 -6.2 4 Shl-1-s13a Hipparion sp. Shilidong, Linxia 34.80 -10.4 -5.4 4

131

Sample No. Taxon Collection Locality Distance from δ13C δ18O Age Crown (mm) (Ma) Shl-1-s14 Hipparion sp. Shilidong, Linxia 36.10 -10.5 -6.3 4 Shl-1-s14a Hipparion sp. Shilidong, Linxia 37.40 -10.3 -6.1 4 Shl-1-s15 Hipparion sp. Shilidong, Linxia 38.70 -10.6 -4.7 4 Shl-1-s16 Hipparion sp. Shilidong, Linxia 40.00 -10.6 -4.8 4 Shl-1-s17 Hipparion sp. Shilidong, Linxia 41.30 -10.4 -5.0 4 Shl-1-s18 Hipparion sp. Shilidong, Linxia 42.60 -10.5 -4.7 4 Shl-1-s19 Hipparion sp. Shilidong, Linxia 43.90 -10.7 -5.0 4 Shl-1-s20 Hipparion sp. Shilidong, Linxia 45.20 -10.5 -4.9 4 Shl-1-s21 Hipparion sp. Shilidong, Linxia 46.50 -10.7 -5.1 4 Shl-1-s22 Hipparion sp. Shilidong, Linxia 47.80 -10.8 -6.2 4 Shl-1-s23 Hipparion sp. Shilidong, Linxia 49.10 -10.6 -7.1 4 Ds-05-s2 Hipparion dermatorhinum Dashanzhuang, Linxia 3.00 -10.2 -2.9 6 Ds-05-s3 Hipparion dermatorhinum Dashanzhuang, Linxia 5.00 -10.2 -3.9 6 Ds-05-s4 Hipparion dermatorhinum Dashanzhuang, Linxia 7.00 -10.1 -4.1 6 Ds-05-s5 Hipparion dermatorhinum Dashanzhuang, Linxia 9.00 -10.1 -5.1 6 Ds-05-s6 Hipparion dermatorhinum Dashanzhuang, Linxia 11.00 -10.2 -4.9 6 Ds-05-s7 Hipparion dermatorhinum Dashanzhuang, Linxia 13.00 -10.1 -5.7 6 Ds-05-s8 Hipparion dermatorhinum Dashanzhuang, Linxia 15.00 -10.2 -6.4 6 Ds-05-s10 Hipparion dermatorhinum Dashanzhuang, Linxia 19.00 -10.2 -6.4 6 Ds-05-s11 Hipparion dermatorhinum Dashanzhuang, Linxia 21.00 -10.4 -6.5 6 Ds-05-s12 Hipparion dermatorhinum Dashanzhuang, Linxia 23.00 -10.2 -6.6 6 Ds-05-s13 Hipparion dermatorhinum Dashanzhuang, Linxia 25.00 -10.2 -6.2 6 Ds-05-s14 Hipparion dermatorhinum Dashanzhuang, Linxia 27.00 -10.2 -6.6 6 Ds-05-s15 Hipparion dermatorhinum Dashanzhuang, Linxia 29.00 -10.3 -6.3 6 Ds-05-s16 Hipparion dermatorhinum Dashanzhuang, Linxia 31.00 -10.3 -6.4 6 Ds-05-s17 Hipparion dermatorhinum Dashanzhuang, Linxia 33.00 -10.3 -6.4 6 Ds-05-s18 Hipparion dermatorhinum Dashanzhuang, Linxia 35.00 -10.3 -5.9 6 Ds-05-s19 Hipparion dermatorhinum Dashanzhuang, Linxia 37.00 -10.4 -6.0 6 Ds-05-s20 Hipparion dermatorhinum Dashanzhuang, Linxia 39.00 -10.5 -5.0 6 Ds-05-s21 Hipparion dermatorhinum Dashanzhuang, Linxia 41.00 -10.3 -4.8 6 Ds-05-s22 Hipparion dermatorhinum Dashanzhuang, Linxia 43.00 -9.9 -6.1 6 Ds-06-s1 Hipparion dermatorhinum Dashanzhuang, Linxia 1.00 -10.3 -4.7 6 Ds-06-s2 Hipparion dermatorhinum Dashanzhuang, Linxia 2.50 -10.4 -4.7 6 Ds-06-s3 Hipparion dermatorhinum Dashanzhuang, Linxia 4.00 -10.3 -5.6 6 Ds-06-s4 Hipparion dermatorhinum Dashanzhuang, Linxia 5.50 -10.5 -6.2 6 Ds-06-s5 Hipparion dermatorhinum Dashanzhuang, Linxia 7.00 -10.6 -6.6 6 Ds-06-s6 Hipparion dermatorhinum Dashanzhuang, Linxia 8.50 -10.7 -7.0 6 Ds-06-s7 Hipparion dermatorhinum Dashanzhuang, Linxia 10.00 -10.6 -6.5 6 Ds-06-s8 Hipparion dermatorhinum Dashanzhuang, Linxia 11.50 -10.6 -6.8 6 Ds-06-s9 Hipparion dermatorhinum Dashanzhuang, Linxia 13.00 -10.7 -9.2 6 Ds-06-s10 Hipparion dermatorhinum Dashanzhuang, Linxia 14.50 -10.4 -10.6 6 Ds-06-s11 Hipparion dermatorhinum Dashanzhuang, Linxia 16.00 -10.5 -9.7 6 Ds-06-s12 Hipparion dermatorhinum Dashanzhuang, Linxia 17.50 -10.4 -9.1 6 Ds-06-s13 Hipparion dermatorhinum Dashanzhuang, Linxia 19.00 -10.3 -9.5 6 Ds-06-s14 Hipparion dermatorhinum Dashanzhuang, Linxia 20.50 -10.2 -9.8 6 Ds-06-s15 Hipparion dermatorhinum Dashanzhuang, Linxia 22.00 -10.2 -10.1 6 Ds-06-s16 Hipparion dermatorhinum Dashanzhuang, Linxia 23.50 -10.3 -9.0 6 Ds-06-s17 Hipparion dermatorhinum Dashanzhuang, Linxia 25.00 -10.1 -5.8 6 Ds-06-s19 Hipparion dermatorhinum Dashanzhuang, Linxia 28.00 -10.3 -4.7 6 Ds-06-s20 Hipparion dermatorhinum Dashanzhuang, Linxia 29.50 -10.1 -4.3 6 Ds-06-s21 Hipparion dermatorhinum Dashanzhuang, Linxia 31.00 -10.2 -3.8 6 Ds-06-s22 Hipparion dermatorhinum Dashanzhuang, Linxia 32.50 -10.1 -4.1 6

132

Sample No. Taxon Collection Locality Distance from δ13C δ18O Age Crown (mm) (Ma) Hl-03-s1 Hipparion sp. Heilinding, Linxia 1.00 -9.6 -4.2 7.5 Hl-03-s2 Hipparion sp. Heilinding, Linxia 2.50 -9.4 -5.5 7.5 Hl-03-s3 Hipparion sp. Heilinding, Linxia 4.00 -9.5 -4.7 7.5 Hl-03-s4 Hipparion sp. Heilinding, Linxia 5.50 -9.5 -5.0 7.5 Hl-03-s5 Hipparion sp. Heilinding, Linxia 7.00 -9.6 -5.2 7.5 Hl-03-s6 Hipparion sp. Heilinding, Linxia 8.50 -9.5 -5.0 7.5 Hl-03-s7 Hipparion sp. Heilinding, Linxia 10.00 -9.5 -5.6 7.5 Hl-03-s8 Hipparion sp. Heilinding, Linxia 11.50 -9.5 -5.9 7.5 Hl-03-s9 Hipparion sp. Heilinding, Linxia 13.00 -9.6 -6.5 7.5 Hl-03-s10 Hipparion sp. Heilinding, Linxia 14.50 -9.5 -6.4 7.5 Hl-03-s11 Hipparion sp. Heilinding, Linxia 16.00 -9.6 -7.3 7.5 Hl-03-s12 Hipparion sp. Heilinding, Linxia 17.50 -9.5 -7.2 7.5 Hl-03-s13 Hipparion sp. Heilinding, Linxia 19.00 -9.6 -7.0 7.5 Hl-03-s14 Hipparion sp. Heilinding, Linxia 20.50 -9.4 -7.7 7.5 Hl-10-s0 Hipparion sp. Heilinding, Linxia 1.00 -10.0 -4.4 7.5 Hl-10-s1a Hipparion sp. Heilinding, Linxia 3.60 -10.3 -3.8 7.5 Hl-10-s2a Hipparion sp. Heilinding, Linxia 6.20 -10.3 -4.0 7.5 Hl-10-s3a Hipparion sp. Heilinding, Linxia 8.80 -10.1 -3.8 7.5 Hl-10-s4a Hipparion sp. Heilinding, Linxia 11.40 -10.0 -3.7 7.5 Hl-10-s5a Hipparion sp. Heilinding, Linxia 14.00 -9.8 -3.5 7.5 Hl-10-s6a Hipparion sp. Heilinding, Linxia 16.60 -9.8 -3.8 7.5 Hl-10-s7a Hipparion sp. Heilinding, Linxia 19.20 -10.1 -4.1 7.5 Hl-10-s8a Hipparion sp. Heilinding, Linxia 21.80 -10.3 -3.8 7.5 Hl-10-s9a Hipparion sp. Heilinding, Linxia 24.40 -10.3 -3.8 7.5 Hl-10-s10a Hipparion sp. Heilinding, Linxia 27.00 -10.2 -4.1 7.5 Hl-10-s11a Hipparion sp. Heilinding, Linxia 29.60 -10.0 -3.7 7.5 Hl-10-s12a Hipparion sp. Heilinding, Linxia 32.20 -9.7 -3.7 7.5 Hl-10-s13a Hipparion sp. Heilinding, Linxia 34.80 -9.7 -3.8 7.5 Qj-06-s0 Hipparion dermatorhinum Qiaojia, Linxia 1.00 -11.5 -7.7 9.5 Qj-06-s1a Hipparion dermatorhinum Qiaojia, Linxia 4.20 -11.5 -8.0 9.5 Qj-06-s2a Hipparion dermatorhinum Qiaojia, Linxia 7.40 -11.5 -8.6 9.5 Qj-06-s3a Hipparion dermatorhinum Qiaojia, Linxia 10.60 -11.5 -9.4 9.5 Qj-06-s4a Hipparion dermatorhinum Qiaojia, Linxia 13.80 -11.6 -9.2 9.5 Qj-06-s5a Hipparion dermatorhinum Qiaojia, Linxia 17.00 -11.5 -8.9 9.5 Qj-06-s6a Hipparion dermatorhinum Qiaojia, Linxia 20.20 -11.5 -9.4 9.5 Qj-06-s7a Hipparion dermatorhinum Qiaojia, Linxia 23.40 -11.1 -9.7 9.5 Qj-06-s8a Hipparion dermatorhinum Qiaojia, Linxia 26.60 -11.0 -9.6 9.5 Qj-06-s9a Hipparion dermatorhinum Qiaojia, Linxia 29.80 -10.9 -10.5 9.5 Qj-06-s11 Hipparion dermatorhinum Qiaojia, Linxia 33.00 -10.9 -11.2 9.5 Qj-06-s12 Hipparion dermatorhinum Qiaojia, Linxia 34.60 -10.9 -10.9 9.5 Qj-06-s13 Hipparion dermatorhinum Qiaojia, Linxia 36.20 -11.2 -10.5 9.5 Qj-06-s14 Hipparion dermatorhinum Qiaojia, Linxia 37.80 -11.0 -9.6 9.5 Qj-06-s15 Hipparion dermatorhinum Qiaojia, Linxia 39.40 -11.0 -9.1 9.5 Qj-06-s16 Hipparion dermatorhinum Qiaojia, Linxia 41.00 -11.2 -9.1 9.5 Qj-06-s17 Hipparion dermatorhinum Qiaojia, Linxia 42.60 -11.2 -8.8 9.5 Qj-06-s18 Hipparion dermatorhinum Qiaojia, Linxia 44.20 -10.9 -8.4 9.5 Qj-06-s19 Hipparion dermatorhinum Qiaojia, Linxia 45.80 -10.5 -8.5 9.5 Gn-07-S1 Hipparion dongxiangense Guonigou, Linxia 1.00 -10.3 -1.6 11.5 Gn-07-S2 Hipparion dongxiangense Guonigou, Linxia 3.00 -10.8 -1.9 11.5 Gn-07-S3 Hipparion dongxiangense Guonigou, Linxia 5.00 -10.5 -2.4 11.5 Gn-07-S4 Hipparion dongxiangense Guonigou, Linxia 7.00 -10.7 -3.7 11.5 Gn-07-S5 Hipparion dongxiangense Guonigou, Linxia 9.00 -10.5 -3.5 11.5

133

Sample No. Taxon Collection Locality Distance from δ13C δ18O Age Crown (mm) (Ma) Gn-07-S6 Hipparion dongxiangense Guonigou, Linxia 11.00 -10.6 -1.7 11.5 Gn-07-S7 Hipparion dongxiangense Guonigou, Linxia 13.00 -10.6 -2.2 11.5 Gn-07-S8 Hipparion dongxiangense Guonigou, Linxia 15.00 -10.6 -2.5 11.5 Gn-07-S9 Hipparion dongxiangense Guonigou, Linxia 17.00 -10.7 -2.5 11.5 Gn-07-S10 Hipparion dongxiangense Guonigou, Linxia 19.00 -10.7 -3.1 11.5 Gn-07-S11 Hipparion dongxiangense Guonigou, Linxia 21.00 -10.7 -3.0 11.5 Gn-07-S12 Hipparion dongxiangense Guonigou, Linxia 23.00 -10.6 -3.4 11.5 Gn-07-S13 Hipparion dongxiangense Guonigou, Linxia 25.00 -10.5 -3.2 11.5 Gn-07-S14 Hipparion dongxiangense Guonigou, Linxia 27.00 -10.8 -6.0 11.5 Gn-07-S15 Hipparion dongxiangense Guonigou, Linxia 29.00 -10.7 -4.9 11.5 Gn-07-S16 Hipparion dongxiangense Guonigou, Linxia 31.00 -10.6 -4.8 11.5 Gn-07-S17 Hipparion dongxiangense Guonigou, Linxia 33.00 -10.6 -4.8 11.5 Gn-07-S18 Hipparion dongxiangense Guonigou, Linxia 35.00 -10.6 -4.2 11.5 Gn-07-S19 Hipparion dongxiangense Guonigou, Linxia 37.00 -10.7 -3.8 11.5 Gn-07-S20 Hipparion dongxiangense Guonigou, Linxia 39.00 -10.4 -4.4 11.5 Gn-07-S21 Hipparion dongxiangense Guonigou, Linxia 41.00 -10.3 -3.9 11.5 Gn-07-S22 Hipparion dongxiangense Guonigou, Linxia 43.00 -10.0 -3.2 11.5 Gn-07-S23 Hipparion dongxiangense Guonigou, Linxia 45.00 -9.8 -2.3 11.5 Gn-07-S24 Hipparion dongxiangense Guonigou, Linxia 47.00 -9.7 -1.0 11.5 Gn-07-S25 Hipparion dongxiangense Guonigou, Linxia 49.00 -9.4 -1.5 11.5 Gn-07-S26 Hipparion dongxiangense Guonigou, Linxia 51.00 -9.3 -1.8 11.5 Gn-07-S27 Hipparion dongxiangense Guonigou, Linxia 53.00 -9.1 -2.3 11.5 Gn-07-S28 Hipparion dongxiangense Guonigou, Linxia 55.00 -9.2 -2.2 11.5 Gn-10-s1 Hipparion dongxiangense Guonigou, Linxia 1.00 -10.9 -3.0 11.5 Gn-10-s2 Hipparion dongxiangense Guonigou, Linxia 2.90 -10.8 -3.2 11.5 Gn-10-s3 Hipparion dongxiangense Guonigou, Linxia 4.80 -10.9 -3.6 11.5 Gn-10-s4 Hipparion dongxiangense Guonigou, Linxia 6.70 -11.0 -2.6 11.5 Gn-10-s5 Hipparion dongxiangense Guonigou, Linxia 8.60 -10.9 -3.3 11.5 Gn-10-s6 Hipparion dongxiangense Guonigou, Linxia 10.50 -10.8 -3.1 11.5 Gn-10-s7 Hipparion dongxiangense Guonigou, Linxia 12.40 -10.9 -3.3 11.5 Gn-10-s8 Hipparion dongxiangense Guonigou, Linxia 14.30 -10.9 -3.6 11.5 Gn-10-s9 Hipparion dongxiangense Guonigou, Linxia 16.20 -10.7 -3.7 11.5 Gn-10-s10 Hipparion dongxiangense Guonigou, Linxia 18.10 -10.6 -4.1 11.5 Gn-10-s11 Hipparion dongxiangense Guonigou, Linxia 20.00 -10.5 -4.1 11.5 Gn-10-s12 Hipparion dongxiangense Guonigou, Linxia 21.90 -10.4 -5.4 11.5 Gn-10-s13 Hipparion dongxiangense Guonigou, Linxia 23.80 -10.3 -5.4 11.5 Gn-10-s14 Hipparion dongxiangense Guonigou, Linxia 25.70 -10.3 -5.8 11.5 Gn-10-s15 Hipparion dongxiangense Guonigou, Linxia 27.60 -10.3 -6.7 11.5 Gn-10-s16 Hipparion dongxiangense Guonigou, Linxia 29.50 -10.2 -5.7 11.5 Gn-10-s17 Hipparion dongxiangense Guonigou, Linxia 31.40 -10.3 -5.9 11.5 Gn-10-s18 Hipparion dongxiangense Guonigou, Linxia 33.30 -10.2 -5.7 11.5 Gn-10-s19 Hipparion dongxiangense Guonigou, Linxia 35.20 -10.1 -6.3 11.5 Gn-10-s20 Hipparion dongxiangense Guonigou, Linxia 37.10 -10.2 -4.1 11.5 Gn-10-s21 Hipparion dongxiangense Guonigou, Linxia 39.00 -10.3 -3.8 11.5 Gn-10-s22 Hipparion dongxiangense Guonigou, Linxia 40.90 -10.4 -3.9 11.5 Gn-10-s23 Hipparion dongxiangense Guonigou, Linxia 42.80 -10.3 -3.8 11.5 Gn-10-s24 Hipparion dongxiangense Guonigou, Linxia 44.70 -10.3 -3.6 11.5 Ds-11-s1 Chilotherium wimani Dashanzhuang, Linxia 1.00 -9.3 -4.2 6 Ds-11-s2 Chilotherium wimani Dashanzhuang, Linxia 2.70 -9.4 -5.7 6 Ds-11-s3 Chilotherium wimani Dashanzhuang, Linxia 4.40 -9.3 -5.9 6 Ds-11-s4 Chilotherium wimani Dashanzhuang, Linxia 6.10 -9.3 -5.4 6 Ds-11-s5 Chilotherium wimani Dashanzhuang, Linxia 7.80 -9.2 -4.5 6

134

Sample No. Taxon Collection Locality Distance from δ13C δ18O Age Crown (mm) (Ma) Ds-11-s6 Chilotherium wimani Dashanzhuang, Linxia 9.50 -9.1 -4.8 6 Ds-11-s7 Chilotherium wimani Dashanzhuang, Linxia 11.20 -9.0 -3.8 6 Ds-11-s8 Chilotherium wimani Dashanzhuang, Linxia 12.90 -9.2 -4.0 6 Ds-11-s9 Chilotherium wimani Dashanzhuang, Linxia 14.60 -9.1 -3.7 6 Ds-11-s10 Chilotherium wimani Dashanzhuang, Linxia 16.30 -9.1 -5.0 6 Ds-11-s11 Chilotherium wimani Dashanzhuang, Linxia 18.00 -8.9 -3.9 6 Ds-11-s12 Chilotherium wimani Dashanzhuang, Linxia 19.70 -9.1 -3.8 6 Hl-14-s1 Chilotherium wimani Heilinding, Linxia 1.00 -9.6 -9.6 7.5 Hl-14-s2 Chilotherium wimani Heilinding, Linxia 3.50 -9.4 -9.8 7.5 Hl-14-s3 Chilotherium wimani Heilinding, Linxia 6.00 -9.2 -9.8 7.5 Hl-14-s4 Chilotherium wimani Heilinding, Linxia 8.50 -9.2 -9.3 7.5 Hl-14-s5 Chilotherium wimani Heilinding, Linxia 11.00 -9.5 -8.7 7.5 Hl-14-s6 Chilotherium wimani Heilinding, Linxia 13.50 -9.6 -7.7 7.5 Hl-14-s7 Chilotherium wimani Heilinding, Linxia 16.00 -9.5 -7.7 7.5 Hl-14-s8 Chilotherium wimani Heilinding, Linxia 18.50 -9.6 -7.9 7.5 Hl-14-s9 Chilotherium wimani Heilinding, Linxia 21.00 -9.7 -8.3 7.5 Hl-14-s10 Chilotherium wimani Heilinding, Linxia 23.50 -9.7 -8.4 7.5 Hl-14-s11 Chilotherium wimani Heilinding, Linxia 26.00 -9.6 -8.8 7.5 Hl-14-s12 Chilotherium wimani Heilinding, Linxia 28.50 -9.5 -9.0 7.5 Hl-14-s13 Chilotherium wimani Heilinding, Linxia 31.00 -9.3 -9.1 7.5 Hl-14-s14 Chilotherium wimani Heilinding, Linxia 33.50 -9.4 -9.2 7.5 Hl-14-s15 Chilotherium wimani Heilinding, Linxia 36.00 -9.4 -9.1 7.5 Hl-15-s0 Chilotherium wimani Heilinding, Linxia 1.00 -9.9 -8.3 7.5 Hl-15-s1 Chilotherium wimani Heilinding, Linxia 3.00 -9.7 -9.3 7.5 Hl-15-s2 Chilotherium wimani Heilinding, Linxia 5.00 -9.5 -9.4 7.5 Hl-15-s3 Chilotherium wimani Heilinding, Linxia 7.00 -9.5 -8.8 7.5 Hl-15-s4 Chilotherium wimani Heilinding, Linxia 9.00 -9.6 -8.0 7.5 Hl-15-s5 Chilotherium wimani Heilinding, Linxia 11.00 -9.6 -7.6 7.5 Hl-15-s6 Chilotherium wimani Heilinding, Linxia 13.00 -9.6 -7.5 7.5 Hl-15-s7 Chilotherium wimani Heilinding, Linxia 15.00 -9.5 -7.7 7.5 Hl-15-s8 Chilotherium wimani Heilinding, Linxia 17.00 -9.5 -7.8 7.5 Hl-15-s9 Chilotherium wimani Heilinding, Linxia 19.00 -9.6 -8.1 7.5 Hl-15-s10 Chilotherium wimani Heilinding, Linxia 21.00 -9.4 -8.5 7.5 Hl-15-s11 Chilotherium wimani Heilinding, Linxia 23.00 -9.2 -8.6 7.5 Hl-15-s12 Chilotherium wimani Heilinding, Linxia 25.00 -9.2 -8.6 7.5 Hl-15-s13 Chilotherium wimani Heilinding, Linxia 27.00 -9.3 -7.8 7.5 Hl-15-s14 Chilotherium wimani Heilinding, Linxia 29.00 -9.3 -8.0 7.5 Hl-15-s15 Chilotherium wimani Heilinding, Linxia 31.00 -9.4 -7.7 7.5 Ls-06-s1 Iranotherium morgani Liushu Fm., Linxia 1.00 -11.5 -3.1 9.2 Ls-06-s2 Iranotherium morgani Liushu Fm., Linxia 4.20 -11.4 -2.5 9.2 Ls-06-s3 Iranotherium morgani Liushu Fm., Linxia 7.40 -11.6 -2.7 9.2 Ls-06-s4 Iranotherium morgani Liushu Fm., Linxia 10.60 -11.6 -2.6 9.2 Ls-06-s5 Iranotherium morgani Liushu Fm., Linxia 13.80 -11.6 -2.4 9.2 Ls-06-s6 Iranotherium morgani Liushu Fm., Linxia 17.00 -11.8 -2.4 9.2 Ls-06-s6 Iranotherium morgani Liushu Fm., Linxia 17.00 -11.8 -2.3 9.2 Ls-06-s7 Iranotherium morgani Liushu Fm., Linxia 23.40 -12.0 -2.3 9.2 Ls-06-s8 Iranotherium morgani Liushu Fm., Linxia 26.60 -11.8 -2.6 9.2 Ls-06-s9 Iranotherium morgani Liushu Fm., Linxia 29.80 -11.7 -2.5 9.2 Ls-06-s10 Iranotherium morgani Liushu Fm., Linxia 33.00 -11.9 -2.0 9.2 Ls-06-s11 Iranotherium morgani Liushu Fm., Linxia 36.20 -11.8 -1.7 9.2 Ls-06-s11 Iranotherium morgani Liushu Fm., Linxia 36.20 -11.8 -1.7 9.2 Ls-06-s12 Iranotherium morgani Liushu Fm., Linxia 39.40 -11.6 -2.4 9.2

135

Sample No. Taxon Collection Locality Distance from δ13C δ18O Age Crown (mm) (Ma) Ls-06-s12 Iranotherium morgani Liushu Fm., Linxia 39.40 -11.6 -2.4 9.2 Ls-06-s13 Iranotherium morgani Liushu Fm., Linxia 42.60 -11.4 -2.0 9.2 Ls-06-s14 Iranotherium morgani Liushu Fm., Linxia 45.80 -11.4 -1.7 9.2 Ls-06-s15 Iranotherium morgani Liushu Fm., Linxia 49.00 -11.3 -2.3 9.2 Ls-06-s16 Iranotherium morgani Liushu Fm., Linxia 52.20 -11.1 -2.5 9.2 Qj-07-s1 Chilotherium wimani Qiaojia, Linxia 1.00 -10.7 -11.6 9.5 Qj-07-s2 Chilotherium wimani Qiaojia, Linxia 3.00 -10.6 -11.5 9.5 Qj-07-s3 Chilotherium wimani Qiaojia, Linxia 5.00 -10.7 -11.2 9.5 Qj-07-s4 Chilotherium wimani Qiaojia, Linxia 7.00 -10.9 -10.3 9.5 Qj-07-s5 Chilotherium wimani Qiaojia, Linxia 9.00 -11.1 -10.0 9.5 Qj-07-s6 Chilotherium wimani Qiaojia, Linxia 11.00 -11.2 -8.9 9.5 Qj-07-s7 Chilotherium wimani Qiaojia, Linxia 13.00 -11.2 -9.3 9.5 Qj-07-s8 Chilotherium wimani Qiaojia, Linxia 15.00 -11.1 -10.2 9.5 Qj-07-s9 Chilotherium wimani Qiaojia, Linxia 17.00 -11.1 -10.6 9.5 Qj-07-s10 Chilotherium wimani Qiaojia, Linxia 19.00 -11.0 -10.3 9.5 Qj-07-s11 Chilotherium wimani Qiaojia, Linxia 21.00 -11.2 -9.5 9.5 Qj-07-s12 Chilotherium wimani Qiaojia, Linxia 23.00 -11.2 -9.1 9.5 Qj-07-s13 Chilotherium wimani Qiaojia, Linxia 25.00 -11.0 -9.5 9.5 Qj-07-s14 Chilotherium wimani Qiaojia, Linxia 27.00 -10.8 -11.6 9.5 Qj-07-s15 Chilotherium wimani Qiaojia, Linxia 29.00 -10.9 -11.5 9.5 Qj-07-s16 Chilotherium wimani Qiaojia, Linxia 31.00 -11.0 -11.5 9.5 Qj-07-s17 Chilotherium wimani Qiaojia, Linxia 33.00 -11.1 -11.6 9.5 Qj-09-s1 Chilotherium wimani Qiaojia, Linxia 1.00 -11.3 -10.0 9.5 Qj-09-s2 Chilotherium wimani Qiaojia, Linxia 2.70 -11.0 -10.1 9.5 Qj-09-s3 Chilotherium wimani Qiaojia, Linxia 4.40 -10.9 -10.0 9.5 Qj-09-s4 Chilotherium wimani Qiaojia, Linxia 6.10 -11.1 -10.0 9.5 Qj-09-s5 Chilotherium wimani Qiaojia, Linxia 7.80 -11.0 -10.0 9.5 Qj-09-s6 Chilotherium wimani Qiaojia, Linxia 9.50 -11.0 -9.4 9.5 Qj-09-s7 Chilotherium wimani Qiaojia, Linxia 11.20 -11.1 -9.1 9.5 Qj-09-s8 Chilotherium wimani Qiaojia, Linxia 12.90 -11.1 -8.5 9.5 Qj-09-s9 Chilotherium wimani Qiaojia, Linxia 14.60 -11.1 -8.7 9.5 Qj-09-s10 Chilotherium wimani Qiaojia, Linxia 16.30 -11.1 -9.1 9.5 Qj-09-s11 Chilotherium wimani Qiaojia, Linxia 18.00 -11.2 -9.8 9.5 Qj-09-s12 Chilotherium wimani Qiaojia, Linxia 19.70 -11.3 -10.6 9.5 Qj-09-s13 Chilotherium wimani Qiaojia, Linxia 21.40 -11.2 -10.4 9.5 Qj-09-s14 Chilotherium wimani Qiaojia, Linxia 23.10 -11.1 -10.4 9.5 Qj-09-s15 Chilotherium wimani Qiaojia, Linxia 24.80 -11.1 -10.4 9.5 Qj-09-s16 Chilotherium wimani Qiaojia, Linxia 26.50 -10.9 -10.1 9.5 Gn-13-s1 Parelasmotherium linxiaense Guonigou, Linxia 1.00 -9.9 -6.1 11.5 Gn-13-s2 Parelasmotherium linxiaense Guonigou, Linxia 3.30 -10.3 -4.7 11.5 Gn-13-s3 Parelasmotherium linxiaense Guonigou, Linxia 5.60 -10.3 -5.6 11.5 Gn-13-s4 Parelasmotherium linxiaense Guonigou, Linxia 7.90 -10.3 -5.1 11.5 Gn-13-s5 Parelasmotherium linxiaense Guonigou, Linxia 10.20 -10.3 -4.1 11.5 Gn-13-s6 Parelasmotherium linxiaense Guonigou, Linxia 12.50 -9.9 -4.6 11.5 Gn-13-s7 Parelasmotherium linxiaense Guonigou, Linxia 14.80 -10.1 -4.7 11.5 Gn-13-s8 Parelasmotherium linxiaense Guonigou, Linxia 17.10 -10.0 -5.0 11.5 Gn-13-s9 Parelasmotherium linxiaense Guonigou, Linxia 19.40 -10.0 -4.0 11.5 Gn-13-s10 Parelasmotherium linxiaense Guonigou, Linxia 21.70 -10.1 -3.7 11.5 Gn-13-s11 Parelasmotherium linxiaense Guonigou, Linxia 24.00 -10.0 -3.6 11.5 Gn-13-s12 Parelasmotherium linxiaense Guonigou, Linxia 26.30 -10.0 -3.5 11.5 Gn-13-s14 Parelasmotherium linxiaense Guonigou, Linxia 30.90 -9.8 -4.8 11.5 Gn-13-s15 Parelasmotherium linxiaense Guonigou, Linxia 33.20 -10.1 -7.1 11.5

136

Sample No. Taxon Collection Locality Distance from δ13C δ18O Age Crown (mm) (Ma) Gn-13-s16 Parelasmotherium linxiaense Guonigou, Linxia 35.50 -9.9 -5.1 11.5 Gn-13-s17 Parelasmotherium linxiaense Guonigou, Linxia 37.80 -10.5 -4.1 11.5 Gn-13-s18 Parelasmotherium linxiaense Guonigou, Linxia 40.10 -10.2 -4.5 11.5 Gn-13-s19 Parelasmotherium linxiaense Guonigou, Linxia 42.40 -9.9 -4.9 11.5 Gn-13-s20 Parelasmotherium linxiaense Guonigou, Linxia 44.70 -10.1 -5.5 11.5 Gn-14-s1 Parelasmotherium linxiaense Guonigou, Linxia 1.00 -10.5 -4.6 11.5 Gn-14-s2 Parelasmotherium linxiaense Guonigou, Linxia 3.00 -10.3 -4.5 11.5 Gn-14-s3 Parelasmotherium linxiaense Guonigou, Linxia 5.00 -10.1 -4.3 11.5 Gn-14-s4 Parelasmotherium linxiaense Guonigou, Linxia 7.00 -10.1 -4.0 11.5 Gn-14-s5 Parelasmotherium linxiaense Guonigou, Linxia 9.00 -10.1 -3.1 11.5 Gn-14-s6 Parelasmotherium linxiaense Guonigou, Linxia 11.00 -10.1 -3.5 11.5 Gn-14-s7 Parelasmotherium linxiaense Guonigou, Linxia 13.00 -10.0 -3.6 11.5 Gn-14-s8 Parelasmotherium linxiaense Guonigou, Linxia 15.00 -9.8 -4.1 11.5 Gn-14-s9 Parelasmotherium linxiaense Guonigou, Linxia 17.00 -9.7 -4.4 11.5 Gn-14-s10 Parelasmotherium linxiaense Guonigou, Linxia 19.00 -9.6 -4.5 11.5 Gn-14-s11 Parelasmotherium linxiaense Guonigou, Linxia 21.00 -9.7 -5.0 11.5 Gn-14-s12 Parelasmotherium linxiaense Guonigou, Linxia 23.00 -9.7 -4.7 11.5 Gn-14-s13 Parelasmotherium linxiaense Guonigou, Linxia 25.00 -9.9 -4.0 11.5 Gn-14-s14 Parelasmotherium linxiaense Guonigou, Linxia 27.00 -10.1 -3.3 11.5 Gn-14-s15 Parelasmotherium linxiaense Guonigou, Linxia 29.00 -10.3 -3.4 11.5 Gn-14-s16 Parelasmotherium linxiaense Guonigou, Linxia 31.00 -10.2 -4.2 11.5 Gn-14-s17 Parelasmotherium linxiaense Guonigou, Linxia 33.00 -10.1 -5.0 11.5 Gn-14-s18 Parelasmotherium linxiaense Guonigou, Linxia 35.00 -10.1 -5.3 11.5 Gn-14-s19 Parelasmotherium linxiaense Guonigou, Linxia 37.00 -10.1 -3.8 11.5 Gn-14-s20 Parelasmotherium linxiaense Guonigou, Linxia 39.00 -10.3 -2.9 11.5 Gn-14-s21 Parelasmotherium linxiaense Guonigou, Linxia 41.00 -10.6 -3.9 11.5 Lg-04-s1 Alicornops laogouense Laogou, Linxia 1.00 -10.8 -11.5 13 Lg-04-s2 Alicornops laogouense Laogou, Linxia 2.90 -10.8 -11.2 13 Lg-04-s3 Alicornops laogouense Laogou, Linxia 4.80 -10.7 -10.6 13 Lg-04-s4 Alicornops laogouense Laogou, Linxia 6.70 -10.8 -10.0 13 Lg-04-s5 Alicornops laogouense Laogou, Linxia 8.60 -10.8 -9.6 13 Lg-04-s6 Alicornops laogouense Laogou, Linxia 10.50 -10.7 -8.7 13 Lg-04-s7 Alicornops laogouense Laogou, Linxia 12.40 -10.8 -9.2 13 Lg-04-s8 Alicornops laogouense Laogou, Linxia 14.30 -10.8 -10.2 13 Lg-04-s9 Alicornops laogouense Laogou, Linxia 16.20 -10.8 -11.1 13 Lg-04-s10 Alicornops laogouense Laogou, Linxia 18.10 -10.8 -10.7 13 Lg-04-s11 Alicornops laogouense Laogou, Linxia 20.00 -10.7 -10.3 13 Lg-05-s1 Alicornops laogouense Laogou, Linxia 1.00 -9.5 -6.9 13 Lg-05-s2 Alicornops laogouense Laogou, Linxia 3.00 -9.5 -6.1 13 Lg-05-s3 Alicornops laogouense Laogou, Linxia 5.00 -9.2 -6.1 13 Lg-05-s4 Alicornops laogouense Laogou, Linxia 7.00 -9.4 -6.5 13 Lg-05-s5 Alicornops laogouense Laogou, Linxia 9.00 -9.4 -7.3 13 Lg-05-s6 Alicornops laogouense Laogou, Linxia 11.00 -9.3 -8.0 13 Lg-05-s8 Alicornops laogouense Laogou, Linxia 15.00 -9.4 -8.1 13 Lg-05-s9 Alicornops laogouense Laogou, Linxia 17.00 -9.4 -7.4 13 Lg-05-s10 Alicornops laogouense Laogou, Linxia 19.00 -9.4 -6.8 13 Lg-05-s11 Alicornops laogouense Laogou, Linxia 21.00 -9.5 -6.5 13 Lg-05-s12 Alicornops laogouense Laogou, Linxia 23.00 -9.5 -6.3 13 Lg-05-s13 Alicornops laogouense Laogou, Linxia 25.00 -9.4 -7.0 13 Lg-05-s14 Alicornops laogouense Laogou, Linxia 27.00 -9.3 -7.7 13 Lg-05-s15 Alicornops laogouense Laogou, Linxia 29.00 -9.2 -8.4 13 Lg-05-s16 Alicornops laogouense Laogou, Linxia 31.00 -9.3 -8.6 13

137

Sample No. Taxon Collection Locality Distance from δ13C δ18O Age Crown (mm) (Ma) Lg-05-s17 Alicornops laogouense Laogou, Linxia 33.00 -9.4 -8.6 13 Lg-05-s18 Alicornops laogouense Laogou, Linxia 35.00 -9.5 -7.9 13 Lg-05-s19 Alicornops laogouense Laogou, Linxia 37.00 -9.4 -7.7 13 Lg-05-s20 Alicornops laogouense Laogou, Linxia 39.00 -9.3 -7.1 13 Lg-05-s21 Alicornops laogouense Laogou, Linxia 41.00 -9.2 -6.9 13 Lg-05-s22 Alicornops laogouense Laogou, Linxia 43.00 -9.2 -6.9 13 Lg-05-s23 Alicornops laogouense Laogou, Linxia 45.00 -9.2 -6.7 13 Lg-05-s24 Alicornops laogouense Laogou, Linxia 47.00 -9.2 -6.7 13 Lg-05-s25 Alicornops laogouense Laogou, Linxia 49.00 -9.1 -7.7 13 Lg-05-s26 Alicornops laogouense Laogou, Linxia 51.00 -9.0 -8.3 13 Lg-05-s27 Alicornops laogouense Laogou, Linxia 53.00 -8.9 -9.0 13 Lg-05-s28 Alicornops laogouense Laogou, Linxia 55.00 -8.9 -8.9 13 Lgo-05-s1 Hispanotherium matritense Laogou, Linxia 1.00 -8.7 -7.7 14 Lgo-05-s2 Hispanotherium matritense Laogou, Linxia 3.75 -9.1 -8.1 14 Lgo-05-s3 Hispanotherium matritense Laogou, Linxia 6.50 -9.0 -7.7 14 Lgo-05-s4 Hispanotherium matritense Laogou, Linxia 9.25 -9.1 -7.3 14 Lgo-05-s5 Hispanotherium matritense Laogou, Linxia 12.00 -9.2 -6.2 14 Lgo-05-s6 Hispanotherium matritense Laogou, Linxia 14.75 -9.0 -6.3 14 Lgo-05-s7 Hispanotherium matritense Laogou, Linxia 17.50 -9.2 -6.4 14 Lgo-05-s8 Hispanotherium matritense Laogou, Linxia 20.25 -9.5 -7.7 14 Lgo-05-s9 Hispanotherium matritense Laogou, Linxia 23.00 -9.5 -8.2 14 Lgo-05-s10 Hispanotherium matritense Laogou, Linxia 25.75 -9.4 -7.7 14 Lgo-05-s11 Hispanotherium matritense Laogou, Linxia 28.50 -9.4 -6.6 14 Lgo-05-s12 Hispanotherium matritense Laogou, Linxia 31.25 -9.4 -6.2 14 Lgo-05-s13 Hispanotherium matritense Laogou, Linxia 34.00 -9.1 -5.3 14 Dl-01-s1 Alicornops sp. Dalanggou, Linxia 1.00 -9.4 -8.2 17 Dl-01-s2 Alicornops sp. Dalanggou, Linxia 3.70 -9.3 -6.3 17 Dl-01-s3 Alicornops sp. Dalanggou, Linxia 6.40 -9.1 -6.0 17 Dl-01-s4 Alicornops sp. Dalanggou, Linxia 9.10 -9.0 -6.0 17 Dl-01-s5 Alicornops sp. Dalanggou, Linxia 11.80 -9.1 -6.2 17 Dl-01-s6 Alicornops sp. Dalanggou, Linxia 14.50 -9.1 -7.2 17 Dl-01-s7 Alicornops sp. Dalanggou, Linxia 17.20 -9.2 -8.1 17 Dl-01-s8 Alicornops sp. Dalanggou, Linxia 19.90 -9.1 -8.8 17 Dl-01-s9 Alicornops sp. Dalanggou, Linxia 22.60 -8.9 -8.9 17 Dl-01-s10 Alicornops sp. Dalanggou, Linxia 25.30 -8.7 -9.3 17 Dl-01-s10 Alicornops sp. Dalanggou, Linxia 25.30 -9.0 -9.7 17 Dl-01-s11 Alicornops sp. Dalanggou, Linxia 28.00 -8.8 -9.1 17 Dl-01-s12 Alicornops sp. Dalanggou, Linxia 30.70 -8.7 -8.1 17 Dl-01-s13 Alicornops sp. Dalanggou, Linxia 33.40 -8.9 -7.3 17 Dl-01-s13 Alicornops sp. Dalanggou, Linxia 33.40 -9.0 -7.5 17 Dl-01-s14 Alicornops sp. Dalanggou, Linxia 36.10 -8.7 -7.2 17 Dl-01-s15 Alicornops sp. Dalanggou, Linxia 38.80 -8.8 -7.6 17 Dl-01-s16 Alicornops sp. Dalanggou, Linxia 41.50 -9.0 -8.6 17 Dl-01-s17 Alicornops sp. Dalanggou, Linxia 44.20 -8.8 -8.7 17 Dl-02-s1 Alicornops sp. Dalanggou, Linxia 1.00 -9.4 -8.0 17 Dl-02-s2 Alicornops sp. Dalanggou, Linxia 3.10 -9.2 -8.5 17 Dl-02-s3 Alicornops sp. Dalanggou, Linxia 5.20 -9.3 -8.8 17 Dl-02-s4 Alicornops sp. Dalanggou, Linxia 7.30 -9.3 -9.1 17 Dl-02-s5 Alicornops sp. Dalanggou, Linxia 9.40 -9.2 -8.9 17 Dl-02-s6 Alicornops sp. Dalanggou, Linxia 11.50 -9.1 -9.0 17 Dl-02-s7 Alicornops sp. Dalanggou, Linxia 13.60 -9.0 -8.8 17 Dl-02-s8 Alicornops sp. Dalanggou, Linxia 15.70 -8.7 -8.5 17

138

Sample No. Taxon Collection Locality Distance from δ13C δ18O Age Crown (mm) (Ma) Dl-02-s9 Alicornops sp. Dalanggou, Linxia 17.80 -9.0 -8.3 17 Dl-02-s10 Alicornops sp. Dalanggou, Linxia 19.90 -8.9 -7.6 17 Dl-02-s11 Alicornops sp. Dalanggou, Linxia 22.00 -8.8 -7.3 17 Dl-02-s12 Alicornops sp. Dalanggou, Linxia 24.10 -8.8 -7.0 17 Dl-02-s13 Alicornops sp. Dalanggou, Linxia 26.20 -8.7 -7.0 17 Dl-02-s14 Alicornops sp. Dalanggou, Linxia 28.30 -8.6 -7.3 17 Tl-02-s1 Paraceratherium sp. Tala, Linxia 1.00 -10.7 -12.5 25 Tl-02-s2 Paraceratherium sp. Tala, Linxia 3.20 -10.6 -12.1 25 Tl-02-s3 Paraceratherium sp. Tala, Linxia 5.40 -10.5 -11.5 25 Tl-02-s4 Paraceratherium sp. Tala, Linxia 7.60 -10.2 -10.9 25 Tl-02-s5 Paraceratherium sp. Tala, Linxia 9.80 -10.1 -10.2 25 Tl-02-s6 Paraceratherium sp. Tala, Linxia 12.00 -10.0 -9.9 25 Tl-02-s7 Paraceratherium sp. Tala, Linxia 14.20 -10.1 -10.7 25 Tl-02-s8 Paraceratherium sp. Tala, Linxia 16.40 -10.1 -11.2 25 Tl-02-s9 Paraceratherium sp. Tala, Linxia 18.60 -10.2 -11.3 25 Tl-02-s10 Paraceratherium sp. Tala, Linxia 20.80 -10.1 -11.5 25 Tl-02-s11 Paraceratherium sp. Tala, Linxia 23.00 -10.1 -11.5 25 Tl-02-s12 Paraceratherium sp. Tala, Linxia 25.20 -9.9 -10.8 25 Tl-02-s13 Paraceratherium sp. Tala, Linxia 27.40 -9.9 -10.2 25 Tl-02-s14 Paraceratherium sp. Tala, Linxia 29.60 -10.0 -10.6 25 Tl-02-s15 Paraceratherium sp. Tala, Linxia 31.80 -10.0 -11.3 25 Tl-03-s1 Paraceratherium sp. Tala, Linxia 1.00 -10.8 -10.4 25 Tl-03-s2 Paraceratherium sp. Tala, Linxia 3.50 -10.7 -11.0 25 Tl-03-s3 Paraceratherium sp. Tala, Linxia 6.00 -10.7 -11.9 25 Tl-03-s4 Paraceratherium sp. Tala, Linxia 8.50 -10.7 -12.0 25 Tl-03-s5 Paraceratherium sp. Tala, Linxia 11.00 -10.7 -12.2 25 Tl-03-s6 Paraceratherium sp. Tala, Linxia 13.50 -10.6 -12.1 25 Tl-03-s7 Paraceratherium sp. Tala, Linxia 16.00 -10.4 -11.9 25 Tl-03-s8 Paraceratherium sp. Tala, Linxia 18.50 -10.5 -11.0 25 Tl-03-s9 Paraceratherium sp. Tala, Linxia 21.00 -10.6 -11.3 25 Tl-03-s10 Paraceratherium sp. Tala, Linxia 23.50 -11.0 -12.2 25 Tl-03-s11 Paraceratherium sp. Tala, Linxia 26.00 -11.2 -12.9 25 Tl-03-s12 Paraceratherium sp. Tala, Linxia 28.50 -10.9 -12.5 25 Lb-01-s1 Modern Cow Linxia Basin 1.00 -9.9 -7.9 0 Lb-01-s2 Modern Cow Linxia Basin 2.90 -10.1 -7.9 0 Lb-01-s3 Modern Cow Linxia Basin 4.80 -10.2 -7.7 0 Lb-01-s4 Modern Cow Linxia Basin 6.70 -10.5 -7.7 0 Lb-01-s5 Modern Cow Linxia Basin 8.60 -10.8 -7.5 0 Lb-01-s6 Modern Cow Linxia Basin 10.50 -10.8 -7.7 0 Lb-01-s7 Modern Cow Linxia Basin 12.40 -11.0 -7.4 0 Lb-01-s8 Modern Cow Linxia Basin 14.30 -11.0 -6.6 0 Lb-01-s9 Modern Cow Linxia Basin 16.20 -11.0 -6.3 0 Lb-01-s10 Modern Cow Linxia Basin 18.10 -10.9 -5.6 0 Lb-01-s11 Modern Cow Linxia Basin 20.00 -10.7 -5.3 0 Lb-01-s12 Modern Cow Linxia Basin 21.90 -10.4 -5.3 0 Lb-01-s13 Modern Cow Linxia Basin 23.80 -10.0 -5.1 0 Lb-01-s14 Modern Cow Linxia Basin 25.70 -9.6 -5.2 0 Lb-01-s15 Modern Cow Linxia Basin 27.60 -9.5 -5.9 0 Lb-01-s16 Modern Cow Linxia Basin 29.50 -9.6 -6.1 0 Lb-01-s17 Modern Cow Linxia Basin 31.40 -9.8 -6.2 0 Lb-02-s1 Modern Goat Linxia Basin 1.00 -12.4 -2.4 0 Lb-02-s2 Modern Goat Linxia Basin 3.20 -12.2 -1.8 0

139

Sample No. Taxon Collection Locality Distance from δ13C δ18O Age Crown (mm) (Ma) Lb-02-s3 Modern Goat Linxia Basin 5.40 -12.1 -1.2 0 Lb-02-s4 Modern Goat Linxia Basin 7.60 -12.0 -1.4 0 Lb-02-s5 Modern Goat Linxia Basin 9.80 -11.7 -2.0 0 Lb-02-s6 Modern Goat Linxia Basin 12.00 -11.6 -2.1 0 Lb-02-s7 Modern Goat Linxia Basin 14.20 -11.7 -2.6 0 Lb-02-s8 Modern Goat Linxia Basin 16.40 -11.3 -3.7 0 Lb-02-s9 Modern Goat Linxia Basin 18.60 -11.2 -5.1 0 Lb-02-s10 Modern Goat Linxia Basin 20.80 -11.4 -5.9 0 Lb-02-s11 Modern Goat Linxia Basin 23.00 -11.1 -6.8 0 Lb-02-s12 Modern Goat Linxia Basin 25.20 -11.2 -7.7 0 Lb-02-s13 Modern Goat Linxia Basin 27.40 -11.4 -8.0 0 Lb-02-s14 Modern Goat Linxia Basin 29.60 -11.9 -7.4 0 Lb-02-s15 Modern Goat Linxia Basin 31.80 -12.2 -6.3 0 Lb-02-s16 Modern Goat Linxia Basin 34.00 -12.5 -5.0 0 Lb-02-s17 Modern Goat Linxia Basin 36.20 -13.1 -4.0 0 Ld-07-s0 Gazella blacki Longdan, Linxia 1.00 -10.9 -7.0 2.5 Ld-07-s1 Gazella blacki Longdan, Linxia 2.90 -11.4 -6.6 2.5 Ld-07-s2 Gazella blacki Longdan, Linxia 4.80 -11.5 -6.5 2.5 Ld-07-s3 Gazella blacki Longdan, Linxia 6.70 -11.0 -7.0 2.5 Ld-07-s4 Gazella blacki Longdan, Linxia 8.60 -10.7 -7.9 2.5 Ld-07-s5 Gazella blacki Longdan, Linxia 10.50 -10.7 -7.6 2.5 Ld-07-s6 Gazella blacki Longdan, Linxia 12.40 -10.6 -7.8 2.5 Ld-07-s7 Gazella blacki Longdan, Linxia 14.30 -10.5 -7.6 2.5 Ld-07-s8 Gazella blacki Longdan, Linxia 16.20 -10.4 -6.8 2.5 Ld-07-s9 Gazella blacki Longdan, Linxia 18.10 -10.0 -6.3 2.5 Ld-08-s1 Leptobus amplifrontalis Longdan, Linxia 1.00 -10.3 -5.6 2.5 Ld-08-s2 Leptobus amplifrontalis Longdan, Linxia 2.75 -10.8 -4.4 2.5 Ld-08-s3 Leptobus amplifrontalis Longdan, Linxia 4.50 -11.0 -4.0 2.5 Ld-08-s4 Leptobus amplifrontalis Longdan, Linxia 6.25 -11.1 -3.9 2.5 Ld-08-s5 Leptobus amplifrontalis Longdan, Linxia 8.00 -11.5 -2.7 2.5 Ld-08-s6 Leptobus amplifrontalis Longdan, Linxia 9.75 -11.8 -2.3 2.5 Ld-08-s7 Leptobus amplifrontalis Longdan, Linxia 11.50 -12.0 -2.0 2.5 Ld-08-s8 Leptobus amplifrontalis Longdan, Linxia 13.25 -11.9 -2.7 2.5 Ld-08-s9 Leptobus amplifrontalis Longdan, Linxia 15.00 -12.4 -1.9 2.5 Ld-09-s1 Leptobus amplifrontalis Longdan, Linxia 1.00 -11.2 -4.3 2.5 Ld-09-s2 Leptobus amplifrontalis Longdan, Linxia 3.10 -11.1 -4.7 2.5 Ld-09-s3 Leptobus amplifrontalis Longdan, Linxia 5.20 -10.9 -5.1 2.5 Ld-09-s4 Leptobus amplifrontalis Longdan, Linxia 7.30 -10.8 -5.5 2.5 Ld-09-s5 Leptobus amplifrontalis Longdan, Linxia 9.40 -10.9 -6.0 2.5 Ld-09-s6 Leptobus amplifrontalis Longdan, Linxia 11.50 -10.9 -6.5 2.5 Ld-09-s7 Leptobus amplifrontalis Longdan, Linxia 13.60 -10.6 -6.6 2.5 Ld-09-s8 Leptobus amplifrontalis Longdan, Linxia 15.70 -10.8 -6.1 2.5 Sl-09-s1 Palaeotragus microdon Shilidong, Linxia 1.00 -12.3 -9.3 4 Sl-09-s2 Palaeotragus microdon Shilidong, Linxia 3.60 -11.7 -8.5 4 Sl-09-s3 Palaeotragus microdon Shilidong, Linxia 6.20 -11.0 -7.8 4 Sl-09-s4 Palaeotragus microdon Shilidong, Linxia 8.80 -10.5 -6.5 4 Sl-09-s5 Palaeotragus microdon Shilidong, Linxia 11.40 -10.3 -6.3 4 Sl-09-s6 Palaeotragus microdon Shilidong, Linxia 14.00 -10.1 -6.0 4 Sl-09-s7 Palaeotragus microdon Shilidong, Linxia 16.60 -10.2 -6.4 4 Sl-09-s8 Palaeotragus microdon Shilidong, Linxia 19.20 -10.3 -6.9 4 Sl-09-s9 Palaeotragus microdon Shilidong, Linxia 21.80 -10.5 -8.5 4 Sl-09-s10 Palaeotragus microdon Shilidong, Linxia 24.40 -10.4 -9.2 4

140

Sample No. Taxon Collection Locality Distance from δ13C δ18O Age Crown (mm) (Ma) Sl-09-s11 Palaeotragus microdon Shilidong, Linxia 27.00 -10.3 -10.1 4 Sl-09-s12 Palaeotragus microdon Shilidong, Linxia 29.60 -10.0 -10.4 4 Sl-10-s1 Palaeotragus microdon Shilidong, Linxia 1.00 -11.8 -9.5 4 Sl-10-s2 Palaeotragus microdon Shilidong, Linxia 3.40 -11.7 -8.3 4 Sl-10-s3 Palaeotragus microdon Shilidong, Linxia 5.80 -10.9 -6.9 4 Sl-10-s4 Palaeotragus microdon Shilidong, Linxia 8.20 -10.1 -5.5 4 Sl-10-s5 Palaeotragus microdon Shilidong, Linxia 10.60 -10.0 -5.6 4 Sl-10-s6 Palaeotragus microdon Shilidong, Linxia 13.00 -10.0 -6.1 4 Sl-10-s7 Palaeotragus microdon Shilidong, Linxia 15.40 -10.0 -6.7 4 Sl-10-s8 Palaeotragus microdon Shilidong, Linxia 17.80 -10.0 -7.6 4 Sl-10-s9 Palaeotragus microdon Shilidong, Linxia 20.20 -10.1 -8.7 4 Sl-10-s10 Palaeotragus microdon Shilidong, Linxia 22.60 -9.9 -9.6 4 Sl-10-s11 Palaeotragus microdon Shilidong, Linxia 25.00 -9.8 -10.1 4 Sl-10-s12 Palaeotragus microdon Shilidong, Linxia 27.40 -9.4 -10.7 4 Lg-08-s1 Platybelodon grangeri Laogou, Linxia 1.00 -10.1 -8.8 13 Lg-08-s1 Platybelodon grangeri Laogou, Linxia 1.00 -10.1 -8.6 13 Lg-08-s2 Platybelodon grangeri Laogou, Linxia 3.70 -9.5 -8.0 13 Lg-08-s2 Platybelodon grangeri Laogou, Linxia 3.70 -9.4 -7.8 13 Lg-08-s3 Platybelodon grangeri Laogou, Linxia 6.40 -9.6 -7.5 13 Lg-08-s4 Platybelodon grangeri Laogou, Linxia 9.10 -9.5 -6.9 13 Lg-08-s5 Platybelodon grangeri Laogou, Linxia 11.80 -9.3 -6.4 13 Lg-08-s6 Platybelodon grangeri Laogou, Linxia 14.50 -9.3 -7.2 13 Lg-08-s7 Platybelodon grangeri Laogou, Linxia 17.20 -8.8 -7.4 13 Lg-10-s1 Platybelodon grangeri Laogou, Linxia 1.00 -9.2 -4.6 13 Lg-10-s2 Platybelodon grangeri Laogou, Linxia 3.70 -9.3 -5.6 13 Lg-10-s3 Platybelodon grangeri Laogou, Linxia 6.40 -9.4 -7.4 13 Lg-10-s4 Platybelodon grangeri Laogou, Linxia 9.10 -9.3 -7.5 13 Lg-10-s5 Platybelodon grangeri Laogou, Linxia 11.80 -9.1 -7.4 13 Lg-10-s6 Platybelodon grangeri Laogou, Linxia 14.50 -9.1 -6.4 13 Lg-10-s7 Platybelodon grangeri Laogou, Linxia 17.20 -9.3 -4.8 13 Lg-10-s8 Platybelodon grangeri Laogou, Linxia 19.90 -9.3 -4.7 13 Lg-10-s9 Platybelodon grangeri Laogou, Linxia 22.60 -9.4 -5.0 13 Lg-10-s10 Platybelodon grangeri Laogou, Linxia 25.30 -9.5 -5.0 13 Dl-04-s1 Gomphotherium sp. Dalanggou, Linxia 1.00 -9.9 -6.2 17 Dl-04-s2 Gomphotherium sp. Dalanggou, Linxia 3.20 -9.5 -5.3 17 Dl-04-s3 Gomphotherium sp. Dalanggou, Linxia 5.40 -9.7 -5.6 17 Dl-04-s4 Gomphotherium sp. Dalanggou, Linxia 7.60 -9.9 -6.3 17 Dl-04-s5 Gomphotherium sp. Dalanggou, Linxia 9.80 -10.1 -7.0 17 Dl-04-s6 Gomphotherium sp. Dalanggou, Linxia 12.00 -10.4 -7.4 17 Dl-04-s7 Gomphotherium sp. Dalanggou, Linxia 14.20 -10.5 -7.5 17 Dl-04-s8 Gomphotherium sp. Dalanggou, Linxia 16.40 -10.5 -7.7 17 Dl-04-s9 Gomphotherium sp. Dalanggou, Linxia 18.60 -10.6 -7.6 17 Dl-04-s10 Gomphotherium sp. Dalanggou, Linxia 20.80 -10.6 -7.4 17 Dl-04-s11 Gomphotherium sp. Dalanggou, Linxia 23.00 -10.6 -7.2 17 Dl-04-s12 Gomphotherium sp. Dalanggou, Linxia 25.20 -10.5 -7.3 17 Dl-07-s1 Gomphotherium sp. Dalanggou, Linxia 1.00 -8.8 -6.6 17 Dl-07-s1 Gomphotherium sp. Dalanggou, Linxia 1.00 -8.6 -6.2 17 Dl-07-s2 Gomphotherium sp. Dalanggou, Linxia 3.50 -8.5 -6.2 17 Dl-07-s3 Gomphotherium sp. Dalanggou, Linxia 6.00 -8.6 -6.2 17 Dl-07-s4 Gomphotherium sp. Dalanggou, Linxia 8.50 -8.6 -6.5 17 Dl-07-s5 Gomphotherium sp. Dalanggou, Linxia 11.00 -8.7 -6.9 17 Dl-07-s6 Gomphotherium sp. Dalanggou, Linxia 13.50 -8.8 -7.4 17

141

Sample No. Taxon Collection Locality Distance from δ13C δ18O Age Crown (mm) (Ma) Dl-07-s7 Gomphotherium sp. Dalanggou, Linxia 16.00 -8.8 -7.1 17 Dl-07-s8 Gomphotherium sp. Dalanggou, Linxia 18.50 -8.8 -8.0 17 Dl-07-s9 Gomphotherium sp. Dalanggou, Linxia 21.00 -9.2 -8.7 17 Dl-07-s10 Gomphotherium sp. Dalanggou, Linxia 23.50 -8.9 -5.9 17

142

APPENDIX C

DATA FROM ANALYSES OF ORGANIC SAMPLES

Lab No. Sample Type Age Collection Locality δ13C C3 (PDB) or C4 Lx-01 Plant/Grass Modern Linxia Basin, Gansu -28.1 C3 Lx-02 Plant/Grass Modern Linxia Basin, Gansu -26.5 C3 Lx-03 Plant/Grass Modern Linxia Basin, Gansu -25.4 C3 Lx-04 Plant/Grass Modern Linxia Basin, Gansu -25.9 C3 Lx-05 Plant/Grass Modern Linxia Basin, Gansu -27.1 C3 Lx-06 Plant/Grass Modern Linxia Basin, Gansu -28.6 C3 Lx-07 Plant/Grass Modern Linxia Basin, Gansu -28.8 C3 Lx-08 Plant/Grass Modern Linxia Basin, Gansu -26.8 C3 Lx-09 Plant/Grass Modern Linxia Basin, Gansu -25.0 C3 Lx-10 Plant/Grass Modern Linxia Basin, Gansu -13.0 C4 Lx-11 Plant/Grass Modern Linxia Basin, Gansu -28.4 C3 Lx-12 Plant/Grass Modern Linxia Basin, Gansu -27.5 C3 Lx-13 Plant/Grass Modern Linxia Basin, Gansu -27.5 C3 Lx-14 Plant/Grass Modern Linxia Basin, Gansu -25.9 C3 Lx-15 Plant/Grass Modern Linxia Basin, Gansu -13.5 C4 Lx-16 Plant/Grass Modern Linxia Basin, Gansu -27.5 C3 Lx-17 Plant/Grass Modern Linxia Basin, Gansu -27.1 C3 Lx-18 Plant/Grass Modern Linxia Basin, Gansu -24.6 C3 Lx-19 Plant/Grass Modern Linxia Basin, Gansu -12.4 C4 Lx-20 Plant/Grass Modern Linxia Basin, Gansu -27.6 C3 Lx-21 Plant/Grass Modern Linxia Basin, Gansu -25.8 C3 Lx-22 Plant/Grass Modern Linxia Basin, Gansu -26.0 C3 Lx-23 Plant/Grass Modern Linxia Basin, Gansu -25.0 C3 Lx-24 Plant/Grass Modern Linxia Basin, Gansu -28.7 C3 Lx-25 Plant/Grass Modern Linxia Basin, Gansu -28.5 C3 Lx-26 Plant/Grass Modern Linxia Basin, Gansu -25.1 C3 Lx-27 Plant/Grass Modern Linxia Basin, Gansu -13.3 C4 Lx-28 Plant/Grass Modern Linxia Basin, Gansu -25.8 C3 Lx-29 Plant/Grass Modern Linxia Basin, Gansu -26.5 C3 Lx-30 Plant/Grass Modern Linxia Basin, Gansu -26.5 C3 Lx-31 Plant/Grass Modern Linxia Basin, Gansu -25.5 C3

143

Lab No. Sample Type Age Collection Locality δ13C C3 (PDB) or C4 Lx-33 Plant/Grass Modern Linxia Basin, Gansu -25.1 C3 Lx-34 Plant/Grass Modern Linxia Basin, Gansu -25.5 C3 Lx-35 Plant/Grass Modern Linxia Basin, Gansu -13.9 C4 Lx-36 Plant/Grass Modern Linxia Basin, Gansu -27.6 C3 Lx-37 Plant/Grass Modern Linxia Basin, Gansu -27.4 C3 Lx-38 Plant/Grass Modern Linxia Basin, Gansu -27.3 C3 Lx-39 Plant/Grass Modern Linxia Basin, Gansu -24.4 C3 Lx-40 Plant/Grass Modern Linxia Basin, Gansu -27.2 C3 Lx-41 Plant/Grass Modern Linxia Basin, Gansu -26.4 C3 Lx-42 Plant/Grass Modern Linxia Basin, Gansu -26.7 C3 Lx-43 Plant/Grass Modern Linxia Basin, Gansu -24.6 C3 Lx-44 Plant/Grass Modern Linxia Basin, Gansu -27.7 C3 Lx-45 Plant/Grass Modern Linxia Basin, Gansu -27.2 C3 Lx-46 Plant/Grass Modern Linxia Basin, Gansu -26.7 C3 Lx-47 Plant/Grass Modern Linxia Basin, Gansu -28.5 C3 Lx-48 Plant/Grass Modern Linxia Basin, Gansu -27.3 C3 Lx-49 Plant/Grass Modern Linxia Basin, Gansu -27.2 C3 Lx-50 Plant/Grass Modern Linxia Basin, Gansu -26.8 C3 Lx-51 Plant/Grass Modern Linxia Basin, Gansu -26.7 C3 Lx-52 Plant/Grass Modern Linxia Basin, Gansu -25.2 C3 Lx-53 Plant/Grass Modern Linxia Basin, Gansu -12.7 C4 Lx-54 Plant/Grass Modern Linxia Basin, Gansu -13.0 C4 Lf-01 Plant/Grass Modern Lufeng, Yunnan -12.9 C4 Lf-01 Plant/Grass Modern Lufeng, Yunnan -13.1 C4 Lf-02 Plant/Grass Modern Lufeng, Yunnan -12.6 C4 Lf-02 Plant/Grass Modern Lufeng, Yunnan -12.4 C4 Lf-03 Plant/Grass Modern Lufeng, Yunnan -14.5 C4 Lf-03 Plant/Grass Modern Lufeng, Yunnan -14.6 C4 Lf-04 Plant/Grass Modern Lufeng, Yunnan -14.0 C4 Lf-04 Plant/Grass Modern Lufeng, Yunnan -13.9 C4 Lf-05 Plant/Grass Modern Yuanmou, Yunnan -13.6 C4 Lf-05 Plant/Grass Modern Yuanmou, Yunnan -13.7 C4 Lf-06 Plant/Grass Modern Yuanmou, Yunnan -14.2 C4 Lf-07 Plant/Grass Modern Yuanmou, Yunnan -13.2 C4 Lf-08 Plant/Grass Modern Lufeng, Yunnan -13.8 C4 Lf-08 Plant/Grass Modern Lufeng, Yunnan -13.8 C4 Lf-09 Plant/Grass Modern Yuanmou, Yunnan -13.3 C4 SGL-45 Buried Prairie Soil Recent Shangrila, Yunnan -24.8 15% C4 SGL-45 Buried Prairie Soil Recent Shangrila, Yunnan -24.4 18% C4

144

APPENDIX D

DATA FROM ANALYSES OF PHOSPHATE SAMPLES

18 18 18 Sample No. δ O (PO4) δ O (CO3) Average Average δ O 18 18 (‰, SMOW) (‰, SMOW) δ O (PO4) δ O (CO3) (CO3-PO4) (‰) (‰) (‰) DS-12-PO4 18.0 25.8 18.0 25.8 7.8 DS-12-PO4 18.1 25.8 DS-12-PO4 17.9 25.8 DS-13-PO4 18.4 26.1 18.5 26.1 7.6 DS-13-PO4 18.6 26.1 DS-13-PO4 18.6 26.1 DS-15-PO4 15.2 23.5 15.2 23.5 8.3 DS-15-PO4 15.6 23.5 DS-15-PO4 14.8 23.5 FG-01-PO4 13.3 22.2 13.6 22.2 8.6 FG-01-PO4 13.6 22.2 FG-01-PO4 13.9 22.2 FG-02-PO4 14.5 23.0 14.5 23.0 8.4 FG-02-PO4 14.4 23.0 FG-02-PO4 14.6 23.0 HL-11-PO4 11.6 21.2 11.6 21.2 9.6 HL-11-PO4 12.0 21.2 HL-11-PO4 11.2 21.2 HL-14-PO4 13.8 22.2 13.9 22.2 8.2 HL-14-PO4 13.9 22.2 HL-14-PO4 14.0 22.2 HL-15-PO4 13.2 22.3 13.2 22.3 9.0 HL-15-PO4 13.3 22.3 HL-15-PO4 13.3 22.3 WC-01-PO4 15.2 22.0 15.1 22.0 6.9 WC-01-PO4 15.2 22.0 WC-01-PO4 14.8 22.0 Wcg-01-PO4 15.1 24.1 15.1 24.1 9.0 Average 8.3 SD 0.8

145

APPENDIX E

DETAILS OF EXPERIMENTAL METHODS

Bulk Sampling The bulk sampling of enamel for isotopic analyses was approached in two different ways. Using the first method, the teeth from Linxia Basin were cut by hammer and chisel or with a Dremel tool along the growth axis. The enamel samples were then manually separated from dentine and any other matrix. Finally, the samples were ground into fine powder using a mortar and pestle. Using the second method, the teeth from Yunnan Province were drilled parallel to the growth axis from the crown to the root (Figure E.1) using a slow-speed Dremel tool with a hand attachment and a diamond-tipped drill bit, with care being taken to avoid contamination of the sample by the underlying dentine. The outer surfaces of all teeth were manually cleaned prior to sampling. Plants were air-dried and the dried plants and sediments were ground into fine powder using a mortar and pestle.

Serial Sampling Serial samples were drilled perpendicular to the growth axis from crown to root, with the youngest samples being near the root and the oldest samples being near the crown (Figure E.1). A slow-speed Dremel tool with a hand attachment and a diamond-tipped drill bit was used. The outer surfaces of all teeth were manually cleaned prior to sampling.

Pretreatment of Samples Tooth enamel carbonate samples were treated in 5% reagent grade sodium hypochlorite for approximately 20 to 24 hours at room temperature to remove organic material from the tooth enamel. The samples were then centrifuged, decanted, and rinsed 4 to 5 times with deionized water to remove the sodium hypochlorite. Next, the samples were treated in 1M acetic acid for 4 to 15 hours at room temperature to remove non-structural carbonate from the enamel. Originally, samples were treated in 1M acetic acid for 15 hours, but it was discovered that 4 to 6 hours of sample treatment in 1M acetic acid was sufficient for the removal of non-structural carbonate (Lee-Thorp and van der Merwe, 1991; Koch et al., 1997; Garvie-Lok et al., 2004). The samples were then centrifuged, decanted, and rinsed 2 to 3 times with deionized water to remove the acetic acid. After the final rinse, the samples were dried under vacuum in a freeze-dryer for 3 to 5 days. To prepare the tooth enamel phosphate samples, tri-silver phosphate was precipitated from the enamel samples following the Dettman et al. (2001) procedure, which is a modification of the approach by O’Neil et al. (1994). First, 5 to 10 mg-samples were dissolved in 1000 μl of 2M HF in an ultrasonic bath overnight, which simultaneously precipitated CaF2. The solutions were then decanted and were brought to a nearly neutral pH with the addition of 200 μl of 20%

146

Figure E.1. Sampling methods. A. Each bulk sample consisted of enamel that was removed parallel to the growth axis along the entire length of a tooth, in order to determine the average carbon and oxygen isotopic compositions over the entire growth period of the tooth. B. Serial samples were drilled perpendicular to the growth axis along the entire length of each tooth to look at changes in carbon and oxygen isotopic compositions throughout the growth period of each tooth, in order to reconstruct seasonal patterns in diet and/or climate or changes in the behaviors of individuals.

147

NH3OH. Next, 1000 μl of 2M AgNO3 was added to each sample, causing rapid precipitation of Ag3PO4. Although Dettman et al. (2001) suggested the addition of 800 μl of 20% NH3OH, it was found that the solution became very basic, rather than neutral, with the addition of 800 μl of 20% NH3OH and consequently, much higher quantities of 2M AgNO3 were necessary to precipitate Ag3PO4. Finally, samples were centrifuged, decanted, rinsed three times, and were freeze-dried for approximately 3 days. No plant samples, organic sediments, or soil carbonates were pretreated prior to isotopic analyses.

Isotopic Analyses To analyze the carbonate samples, ~200 to 500 μg of carbonate standards and 3 to 6 mg of enamel carbonate or soil carbonate samples were measured and placed into reaction vials capped with rubber septa. After loading the sample vials into a Thermo-Finnigan Gasbench II interfaced with a Delta Plus XP continuous flow isotope ratio mass spectrometer, all carbonate samples were flushed for 5 minutes by injection of pure-grade helium in order to force air and moisture from the vial. The samples were then converted to CO2 by injection of 100% phosphoric acid, which was left to react for 3 to 18 hours at 72ºC (bulk samples from Linxia Basin) or for approximately 72 hours at 25 ºC (serial samples from Linxia Basin and bulk samples from Yunnan Province; modified from McCrea, 1950). During this reaction (simplified as follows: CaCO3 +H3PO4 → CO2 + H2O + CaHPO4), most calcium in solution is present as the calcium phosphate ion pairs (CaPO4)- and CaHPO4 and water that is produced is taken up by excess P2O5 to form more H3PO4 (Sharp et al., 2000). Any water vapor that remains in the sample vial is removed as the sample is transferred from the sample vial to the gas bench interface by a gas-tight but hygroscopic Nafion tube that serves as a water trap. After the carbonate samples were converted to CO2, the carbon and oxygen isotope ratios were measured by mass spectrometry. Ten aliquots of CO2 from each sample vial were measured and run against 13 18 three aliquots of a CO2 reference gas. The δ C and δ O values reported are the average values for the ten sample aliquots. Acceptable standard deviations from the mean delta values of the samples, or external errors, were <0.06 ‰ for δ13C and <0.08 ‰ for δ18O. The expected internal 18 13 18 error was <0.05 ‰ for δ O (CO2 reference gas; Thermo-Finnigan, 2002). The δ C and δ O values were calibrated by concurrent carbon and oxygen isotopic measurements of least two sets of three or more of the following carbonate standards: PDA, NBS-19, YW-CC-ST-1 (lab standard), ROY-CC (lab standard), and MERK. To analyze organic samples, ~ 2 to 3 mg of plant tissues or organic soils were measured into small tin cups, which were then sealed by folding over the tops of the cups and the cups were shaped into spheres using forceps. Next, the samples were loaded into the autosampler of a Carlo Erba elemental analyzer connected to a Delta Plus XP continuous flow isotope ratio mass spectrometer by a Conflo II open-split interface. This instrumentation uses dynamic flash combustion, where the samples are dropped from the autosampler into a 1020°C quartz combustion furnace which is flooded with oxygen-enriched helium. The combination of high temperature and excess oxygen causes a great enthalpy reaction, resulting in the ignition of the tin cup and an increase in the local temperature to a value between 1700 and 1800°C. The resultant combustion gases flow through an inorganic oxygen source (chromium oxide and silvered cobaltous cobaltic oxide), which facilitates complete oxidation, and then the gases pass over pure copper, which removes any excess oxygen and reduces nitrogen oxides to N2. Water is then removed by a trap containing magnesium perchlorate and quartz chips and the remaining

148

mixture is directed onto a chromatographic column where the gases are separated and eluted as N2 and CO2. These gases then proceed to the Conflo II interface, where the isotope ratio mass spectrometer accesses them and performs the isotopic analyses. For this study, only the carbon isotope compositions of the organic samples were determined by mass spectrometric measurement of eluted CO2. Only one aliquot of CO2 was analyzed for each sample and 2 aliquots of CO2 reference gas were analyzed immediately before and after each sample. The expected external precision for the analyses was <0.15 ‰ for δ13C (20 μg carbon). The δ13C values were calibrated by concurrent carbon isotopic measurements of at least two sets of three or more of the following organic standards: YWOMST-1 (sugar), YWOMST-2 (phenylalanine), YWOMST-3 (L-phenylalanine), YWOMST-4 (Costech acetamilide), and YWOMST-5 (urea). To analyze phosphate samples, ~ 200 to 300 μg of phosphate standards and tooth enamel phosphate samples in the form of Ag3PO4 were measured into small silver cups, which were then sealed by folding over the tops of the cups and the cups were shaped into spheres using forceps. Next, the samples were loaded into the autosampler of a Finnigan Thermal Conversion Elemental Analyzer (TC/EA) connected to a Delta Plus XP continuous flow isotope ratio mass spectrometer by a Conflo II open-split interface. The TC/EA uses high temperature conversion, also known as pyrolysis, to convert oxygen that is present in a compound to CO and hydrogen present in a compound is converted to H2. In this process, the samples are dropped from the autosampler into a high temperature (1450°C) reactor, which consists of an outer ceramic mantle tube of aluminum oxide and a reducing inner glassy carbon tube packed with a glassy carbon filling. The space between the internal carbon tube and external aluminum oxide tube is continuously flushed with helium to avoid any undesired oxidation of the sample. The reaction gases, CO and H2, are then separated in an isothermal gas chromatographic column, and finally, the gases are transferred to the isotope ratio mass spectrometer via the Conflo II interface (Thermo-Finnigan, 2001). For this study, only the oxygen isotope compositions of the phosphate samples were determined by mass spectrometric measurement of eluted CO. One aliquot of CO was analyzed for each sample and 2 aliquots of a CO reference gas were analyzed immediately before and after each sample. All samples and standards were run in triplicate to ensure that the TC/EA produced no memory effect. The expected internal error was <0.1 ‰ for δ18O (CO reference gas; Thermo-Finnigan, 2001). The expected external precision for the analyses was <0.4 ‰ for δ18O (200 μg benzoic acid). The δ18O values were calibrated by concurrent oxygen isotopic measurements of least three of the following phosphate standards: UMS-1, NIST-120c, NBS-120a, and KH2PO4.

149

REFERENCES

An, Z., Porter, S., Kutzbach, J., Wu, X., Wang, S., Liu, X., Li, X., and Zhou, W. 2000. Asynchronous Holocene optimum of the East Asian monsoon. Quaternary Science Reviews, 19: 743-762.

An, Z.S., Kutzbach, J.E., Prell, W.L., and Porter, S.C. 2001. Evolution of Asian monsoons and phased uplift of the Himalaya–Tibetan Plateau since late Miocene times. Nature, 411: 62–66.

Araguas-Araguas, L., Froehlich, K., and Rozanski, K. 1998. Stable isotope composition of precipitation over southeast Asia. Journal of Geophysical Research, 103: 28721–28742.

Arambourg, C. 1959. Vertébrés continentaux du Miocène supérieur de l'Afrique du Nord. Publications du Service de la carte géologique de l'Algérie (nouvelle serie), Paléontologie, Alger, Mémoir. 4: 1-159.

Ayliffe, L. and Chivas, A. 1990. Oxygen isotope composition of the bone phosphate of Australian kangaroos: potential as a paleoenvironmental recorder. Geochimica et Cosmochimica Acta, 54: 2603–2609.

Ayliffe, L.K., Chivas, A.R., and Leakey, M.G. 1994. The retention of primary oxygen isotope compositions of fossil elephant skeletal phosphate. Geochimica et Cosmochimica Acta, 58: 5291-5298.

Badgley, C., Qi, G., Chen, W., and Han, D. 1988. Paleoecology of a Miocene, Tropical, Upland Fauna: Lufeng, China. National Geographic Research, 4(2):178-195.

Barnett, T. P., Dumenil, L., Schlese, U., Roekler, E., and Latif, M. 1989. The effect of Eurasian snow cover on regional and global climate variations. Journal of the Atmospheric Sciences, 46: 661–685.

Bocherens, H., Koch, P.L., Mariotti, A., Geraads, D., and Jaeger, J.J. 1996. Isotope biogeochemistry of mammalian enamel from African Pleistocene hominid sites. Palaios, 11: 306-318.

Borissiak, A.A. 1923. O rod Indricotherium n.g. (sem. Rhinocerotidae). Zapiski Rossiskaya Akademiya Nauk, 35: 1–128.

150

Boutton, T.W. 1991. Stable carbon isotope ratios of natural materials: II. Atmospheric, terrestrial, marine, and freshwater environments. In D.C. Coleman and B. Fry (eds.) Carbon Isotope Techniques. Academic Press, San Diego, 173-185.

Bradley, R.S. 1999. Paleoclimatology: Reconstructing Climates of the Quaternary, Second Edition. Academic Press, San Diego, 1.

Bronn, H.G. 1831. Über die fossilen Zähne eines neuen Geschlechtes der Dickhäuter-Ordnung, Coelodonta, Höhlenzahn. Neues Jahrbuch für Mineralogie, Geologie und Paläontologie, 1831: 51-61.

Brudevold, F. and Soremark, R. 1967. Chemistry of the mineral phase of enamel—Crystalline organization of dental mineral. In A.E.D. Miles (ed.) Structural and Chemical Organization of Teeth. Academic Press, London, 2: 247-277.

Bryant, D. and Froelich, P. 1995. A model of oxygen isotope fractionation in body water of large mammals. Geochimica et Cosmochimica Acta, 59: 4523-4537.

Bryant, J., Luz, B., and Froelich, P. 1994. Oxygen isotopic composition of fossil horse tooth phosphate as a record of continental paleoclimate. Palaeogeography, Palaeoclimatology, Palaeoecology, 107: 303–316.

Bryant, J.D., Froelich, P.N., Showers, W.J., and Genna, B.J. 1996. Biologic and climatic signals in the oxygen isotopic composition of Eocene-Oligocene equid enamel phosphate. Palaeogeography, Palaeoclimatology, Palaeoecology, 126: 75-89.

Bureau of Geology and Mineral Resources of Yunnan Province. 1990. Regional Geology of Yunnan Province. Geological Memoirs, Series 21. Geological Publishing House, Beijing.

Cerdeño, E. 1998. Diversity and evolutionary trends of the family Rhinocerotidae (Perissodactyla). Palaeogeography, Palaeoclimatology, Palaeoecology, 141: 13-34.

Cerdeño, E. and Nieto, M. 1995. Changes in western European Rhinocerotidae related to climatic variations. Palaeogeography, Palaeoclimatology, Palaeoecology, 114: 325-338.

Cerling, T.E. 1984. The stable isotopic composition of modern soil carbonate and its relationship to climate. Earth and Planetary Science Letters, 71: 229-240.

Cerling, T.E. 1992. Development of grasslands and savannahs in East Africa during the Neogene. Palaeogeography, Palaeoclimatology, Palaeoecology, 97: 241-247.

Cerling, T.E. and Harris, M. 1999. Carbon isotope fractionation between diet and bioapatite in ungulate mammals and implications for ecological and paleoecological studies. Oecologia, 120: 347-363.

151

Cerling, T.E., Quade, J., Wang, Y., and Bowman, J.R. 1989. Carbon isotopes in soils and paleosols as ecology and paleoecology indicators. Nature, 341: 138-139.

Cerling, T.E., Wang, Y., and Quade, J. 1993. Expansion of C4 ecosystems as an indicator of global ecological change in the late Miocene. Nature, 361: 344– 345.

Cerling, T.E., Harris, J.M., and MacFadden, B.J. 1997a. Carbon isotopes, diets of North American equids, and the evolution of North American C4 grasslands. In H. Griffiths, D. Robinson, and P. van Gardingen (eds.) Stable Isotopes and the Integration of Biological, Ecological, and Geochemical Processes. Bios Scientific Publishers, Oxford, 363-379.

Cerling, T.E., Harris, J.M., MacFadden, B.J., Leakey, M.G., Quade, J., Elsenmann, V., and Ehleringer, J.R. 1997b. Global vegetation change through the Miocene/Pliocene boundary. Nature, 389: 153-158.

Cerling, T.E., Harris, J.M., Ambrose, S.H., Leakey, M.G., and Solounias, N. 1997c. Dietary and environmental reconstruction with stable isotope analyses of herbivore tooth enamel from the Miocene locality of Fort Ternan, Kenya. Journal of Human Evolution, 33: 635-650.

Chen, W. 1986. Preliminary studies of sedimental environment and taphonomy in the hominoid fossil site of Lufeng. Acta Anthropologica Sinica, 5: 89-100.

Chow, M. and Zhai, R. 1962. A new species of Stegodon from Zhaotong, Yunnan and a discussion of Stegodon taxonomy and chronology. Vertebrata Palasiatica, 6(2): 138-149.

Chow, M.C. and Zhang, Y.P. 1974. Proboscidea of China. Science Press, Beijing.

Clark, I. and Fritz, P. 1997. Environmental Isotopes in Hydrogeology. Lewis Publishers, Boca Raton, 1-328.

Clark, C.O., Cole, J.E., and Webster, P.J. 2000. Indian Ocean SST and Indian Summer Rainfall: Predictive Relationships and Their Decadal Variability. Journal of Climate, 13: 2503- 2519.

Colbert, E.H. 1943. Pleistocene vertebrates collected in Burma by the American Southeast Asiatic Expedition. Transactions of the American Philosophical Society, 32: 395–429.

Coleman, M. and Hodges, K. 1995. Evidence for Tibetan Plateau uplift before 14 Myr ago from a new minimum age for east-west extension. Nature, 374: 49–52.

Cope, E.D. 1879. On the extinct species of Rhinoceridae of North America and their allies. Bulletin of the United States Geological and Geographical Survey of the Territories, 5: 227-237.

152

Cormie, A.B., Luz, B., and Schwarcz, H.P. 1994. Relationship between the hydrogen and oxygen isotopes of deer bone and their use in the estimation of relative humidity. Geochimica et Cosmochimica Acta, 58: 3439–3449.

Craig, H. 1961. Isotopic variations in meteoric waters. Science, 133: 1702-1703.

Craig, H. and Gordon, L.I. 1965. Deuterium and oxygen 18 variations in the ocean and the marine atmosphere. In E. Tongiorgi (ed.), 2nd Conference on Stable Isotopes in Oceanographic Studies and Paleotemperatures. Consiglio Nazionale delle Richerche, Rome, 9-130.

Criss, R.E. 1999. Principles of Stable Isotope Distribution. Oxford University Press, New York, 1-254.

Crusafont, M. and Villalta, J.F. 1947. Sobre un interesante rinoceronte (Hispanotherium nov. gen.) del Mioceno del Valle del Manzanares. Las Ciencias, 12: 869-883.

D’Angela, D. and Longinelli, A. 1990. Oxygen isotopes in living mammal’s bone phosphate: further results. Chemical Geology, 86: 75-82.

D’Angela, D. and Longinelli, A. 1993. Oxygen isotopic composition of fossil mammal bones of Holocene age: palaeoclimatological considerations. Chemical Geology, 103: 171-179.

Dansgaard, W. 1964. Stable Isotopes in precipitation. Tellus, 16: 436- 468.

Deines, P. 1980. The isotopic composition of reduced organic carbon. In P. Fritz and J. Fontes (eds.) Handbook of Environmental Isotope Geochemistry, Vol. I, The Terrestrial Environment, Part A. Elsevier Scientific Publishing Company, New York, 329-406.

Deng, T. 2001. New remains of Parelasmotherium (Perissodactyla, Rhinocerotidae) from the Late Miocene in Dongxiang, Gansu, China. Vertebrata Palasiatica, 39: 306-311.

Deng, T. 2002. Limb bones of Chilotherium wimani (Perissodactyla, Rhinocerotidae) from the Late Miocene of the Linxia Basin in Gansu, China. Vertebrata Palasiatica, 40: 305-316.

Deng, T. 2003. New material of Hispanotherium matritense (Rhinocerotidae, Perissodactyla) from Laogou of Hezheng County (Gansu, China), with special reference to the Chinese Middle Miocene elasmotheres. Geobios, 36: 141-150.

Deng, T. 2004. A new species of the rhinoceros Alicornops from the Middle Miocene of the Linxia Basin, Gansu, China. Palaeontology, 47: 1427-1439.

Deng, T. 2005a. New discovery of Iranotherium morgani (Perissodactyla, Rhinocerotidae) from the Late Miocene of the Linxia Basin in Gansu, China, and its sexual dimorphism. Journal of Vertebrate Paleontology, 25(2): 442-450.

153

Deng, T. 2005b. New cranial material of Shansirhinus (Rhinocerotidae, Perissodactyla) from the Lower Pliocene of the Linxia Basin in Gansu, China. Geobios, 38: 301-313.

Deng, T. 2006. A primitive species of Chilotherium (Perissodactyla, Rhinocerotidae) from the Late Miocene of the Linxia Basin (Gansu, China). Cainozoic Research, 5(1-2): 93-102.

Deng, T. 2007. Skull of Parelasmotherium (Perissodactyla, Rhinocerotidae) from the Upper Miocene in the Linxia Basin (Gansu, China). Journal of Vertebrate Paleontology, 27(2): 467-475.

Deng, T. 2008. Comparison between woolly rhino forelimbs from Longdan, Northwestern China and Tologoi, Transbaikalian region. Quaternary International, 179:196–207.

Deng, T. and Downs, W. 2002. Evolution of Chinese Neogene Rhinocerotidae and its response to climatic variations. Acta Geologica Sinica, 76: 139-145.

Deng, T., Wang, X., Ni, X., and Liu, L. 2004a. Sequence of the Cenozoic mammalian faunas of the Linxia Basin in Gansu, China. Acta Geologica Sinica, 78: 8-14.

Deng, T., Wang, X., Ni, X., Liu, L., and Liang, Z. 2004b. Cenozoic stratigraphic sequence of the Linxia Basin in Gansu, China and its evidence from mammal fossils. Vertebrata Palasiatica, 42: 45-66.

DeNiro, M.J. and Epstein, S. 1978. Influence of diet on the distribution of carbon isotopes in animals. Geochimica et Cosmochimica Acta, 42: 495-506.

Dettman, D.L., Kohn, M.J., Quade, J., Ryerson, F.J., Ojha, T.P., and Hamidullah, S. 2001. Seasonal stable isotope evidence for a strong Asian Monsoon throughout the past 10.7 m.y. Geology, 29: 31-34.

Dettman, D., Fang, X., Garzione, C., and Li, J. 2003. Uplift-driven climate change at 12 Ma: a long δ18O record from the NE margin of the Tibetan Plateau. Earth and Planetary Science Letters, 214: 267– 277.

Dollo, L. 1885. Rhinoceros vivants et fossils. Revue des questions scientifiques, 17: 293-299.

Dong, W. 1987. Further investigations upon the age and characteristics of the Xiaolongtan fauna, Kaiyuan Co., Yunnan Province. Vertebrata Palasiatica, 25(2): 116-123.

18 Dongmann, G., Nurnberg, H., Forstel, H., and Wagener, K., 1974. On the enrichment of H2 O in the leaves of transpiring plants. Radiation and Environmental Biophysics, 11: 41–52.

Epstein, S. and Mayeda, T. 1953. Variation of 18O content of waters from natural sources. Geochimica et Cosmochimica Acta, 4: 213-224.

154

Epstein, S., Thomas, P., and Yapp, C. 1977. Oxygen and hydrogen isotopic ratios in plant cellulose. Science, 198: 1209-1215.

Fang, X., Li, J., Zhu, J., Chen, H., and Cao, J. 1997. Determination and calibration of time scale of late Cenozoic sedimentary sequences in Linxia Basin, Gansu Province, China. Chinese Science Bulletin, 42: 1457-1471.

Fang, X., Garzione, C., Van der Voo, R., Li, J., and Fan, M. 2003. Flexural subsidence by 29 Ma on the NE edge of Tibet from the magnetostratigraphy of Linxia Basin, China. Earth and Planetary Science Letters, 210: 545-560.

Farquhar, G.D., Ehleringer, J.R., and Hubick, K.T. 1989. Carbon isotope discrimination and photosynthesis. Annual Review of Plant Physiology and Plant Molecular Biology, 40: 503-537.

Feranec, R.S. and MacFadden, B.J. 2000. Evolution of the grazing niche in Pleistocene mammals from Florida: evidence from stable isotopes. Palaeogeography, Palaeoclimatology, Palaeoecology, 162: 155–169.

Flynn, L.J. and Qi, G. 1982. Age of the Lufeng, China, hominoid locality. Nature, 298: 746-747.

Flynn, L., Pilbeam, D., Jacobs, L., Barry, J., Behrensmeyer, A., and Kappelman, J. 1990. The Siwaliks of Pakistan: Time and faunas in a Miocene terrestrial setting. Journal of Geology, 98: 589-604.

Forster Cooper, C. 1911. Paraceratherium bugtiense, a new genus of Rhinocerotidae from Bugti Hills of Baluchistan- preliminary notice. Annals and Magazine of Natural History, 8(8): 711-716.

Fortelius, M. 1983. The morphology and paleobiological significance of the horns of Coelodonta antiquitatis (Mammalia: Rhinocerotidae). Journal of Vertebrate Paleontology, 3: 125– 135.

Fricke, H.C. and O’Neil, J.R. 1996. Inter- and intra-tooth variation in the oxygen isotope composition of mammalian tooth enamel phosphate: implications for palaeoclimatological and palaeobiological research. Palaeogeography, Palaeoclimatology, Palaeoecology, 126: 91-99.

Fricke, H., Clyde, W., O’Neil, J., and Gingerich, P. 1998. Evidence for rapid climate change in North America during the latest Paleocene thermal maximum: oxygen isotope compositions of biogenic phosphate form the Bighorn Basin (Wyoming). Earth and Planetary Science Letters, 160: 193– 208.

Garvie-Lok, S., Varney, T., and Katzenberg, M. 2004. Preparation of bone carbonate for stable isotope analysis: The effects of treatment time and acid concentration. Journal of Archaeological Science, 31: 763-776.

155

Garzione, C., Dettman, D., Quade, J., DeCelles, P., and Butler, R. 2000. High times on the Tibetan Plateau: Paleoelevation of the Thakkhola graben, Nepal. Geology, 28: 339–342.

Ge, H.R. and Li, D.Y. 1999. Cenozoic Coal-Bearing Basins and Coal-Forming Regularity in West Yunnan. Yunnan Science and Technology Press, Kunming.

Gearing, J.N. 1991. The study of diet and trophic relationships through natural abundance 13C. In D.C. Coleman and B. Fry (eds.) Carbon Isotope Techniques. Academic Press, San Diego, 201-218.

Gen, D.M. and Zhang, X.Y. 1992. Tanzigou culture and its archeological significance. In X.Y. Zhang (ed.) Prehistoric Archeology of Baoshan. Yunnan Science and Technology Press, Kunming, 42–48.

Gill, T. 1872. Arrangement of the families of mammals and synoptical tables of characters of the subdivisions of mammals. Smithsonian Miscellaneous Collections, 11(1): 1-98.

Ginsburg, L. and Guérin, C. 1979. Sur l’origine et l’extension stratigraphique du petit Rhinocérotidé miocéne Aceratherium (Alicornops) simorrense (Lartet 1851) nov. subgen. Comptes Rendus Sommaires de la Société Géologique de France, 3: 114-116.

Gloger, C.W.L. 1841. Gemeinnütziges Hand- und Hilfsbuch der Naturgeschichte. Aug. Schultz and Company, Breslau, 1: 125.

Gonfiantini, R. 1978. Standards for stable isotope measurements in natural compounds. Nature, 271: 534-536.

Gonfiantini, R. 1985. On the isotopic composition of precipitation in tropical stations. Acta Amazonica, 15: 121– 139.

Gonfiantini, R., Stichler, W., and Rozanski, K. 1995. Standards and intercomparison materials distributed by the International Atomic Energy Agency for stable isotope measurements. In Reference and Intercomparison Materials for Stable Isotopes of Light Elements, IAEA TECDOC-825. IAEA, Vienna, 13-29.

Greenman, S. 2003. California Academy of Sciences Notes from the Field: Yunnan Province, China Expedition 2003. http://www.calacademy.org/notes_from_the_field/yunnan_2003.html

Guérin, C. 1980. Les rhinocéros (Mammalia, Perissodactyla) du Miocène terminal au Pléistocène supérieur en Europe occidentale: comparaison avec les espèces actuelles. Documents du Laboratoire de Géologie de la Faculté des Sciences de Lyon, 79: 1-1184.

Hahn, D. G. and Manabe, S. 1975. The role of mountains in the South Asian monsoon circulation. Journal of the Atmospheric Sciences, 32: 1515–1541.

156

Harrison, T.M., Copeland, P., Kidd, W.S.F., and Yin, A. 1992. Raising Tibet: Science, 255: 1663–1670.

Harrison, T., Xueping, J., and Su, D. 2002. On the systematic status of the late Neogene hominoids from Yunnan Province, China. Journal of Human Evolution, 43:207-227.

Heissig, K. 1989a. The Rhinocerotidae. In D.R. Prothero and R.M. Schoch (eds.) The Evolution of Perissodactyls. Oxford University Press, New York, 399-417.

Heissig, K. 1989b. The allaceropine hyracodonts. In D.R. Prothero and R.M. Schoch (eds.) The Evolution of Perissodactyls. Oxford University Press, New York, 355-357.

Heissig, K. 1999. Family Rhinocerotidae. In G.E. Rössner and K. Heissig (eds.) The Miocene Land Mammals of Europe. Verlag Dr. Friedrich Pfeil, München, 175-188.

Hodell, D.A., Brenner, M., Kanfoush, S.L., Curtis, J.H., Stoner, J.S., Xueliang, S., Yuan, W., and Whitmor, T.J. 1999. Paleoclimate of Southwestern China for the Past 50,000 yr Inferred from Lake Sediment Records. Quaternary Research, 52: 369-380.

Iacumin, P., Bocherens, H., Mariotti, A., and Longinelli, A. 1996. Oxygen isotope analyses of co-existing carbonate and phosphate in biogenic apatite: a way to monitor diagenetic alteration of bone phosphate? Earth and Planetary Science Letters, 142: 1-6.

Jablonski, N.G., Xueping, J., Chaplin, G., Lirui, W., Shengyi, Y., Guihua, L., and Zhicai, L. 2003. A preliminary report on new and previously known vertebrate paleontological sites in Baoshan Prefecture, Yunnan Province, China. Proceedings of the California Academy of Sciences, 54(11): 209-224.

Johnson, G.D. and Vondra, C.F. 1972. Siwalik sediments in a portion of the Punjab re-entrant: The sequence at Haritalyangar, District Bilashpur, H.P. Himalayan Geology, 2: 118-144.

Johnson, K.R. and Ingram, B.L. 2004. Spatial and temporal variability in the stable isotope systematics of modern precipitation in China: implications for paleoclimate reconstructions. Earth and Planetary Science Letters, 220: 365–377.

Kahlke, H.D. 1969. Die Rhinocerotiden-Reste aus den Kiesen von Süssenborn bei Weimar. Paläontologische Abhandlungen, A 3: 567–709.

Killgus, H. 1923. Unterpliozäne Säuger aus China. Paläontologische Zeitschrift, 5(3): 251-257.

Kitoh, A. 2004. Effects of mountain uplift on East Asian summer climate investigated by a coupled atmosphere-ocean GCM. Journal of Climate, 17: 783-802.

Koch, P.L. 1998. Isotopic reconstruction of past continental environments. Annual Review of Earth and Planetary Sciences, 26: 573-613.

157

Koch, P.L., Fisher, D.C., and Dettman, D. 1989. Oxygen isotope variation in the tusks of extinct proboscideans: a measure of season of death and seasonality. Geology, 17: 515-519.

Koch, P.L., Fogel, M.L., and Tuross, N. 1994. Tracing the diets of fossil animals using stable isotopes. In K. Lajtha and R.H. Michener (eds.) Stable Isotopes in Ecology and Environmental Science, Blackwell Scientific Publications, Oxford, 63-92.

Koch, P.L., Heisinger, J., Moss, C., Carlson, R.W., Fogel, M.L., and Behrensmeyer, A.K. 1995. Isotopic tracking of change in diet and habitat use in African elephants. Science, 267: 1340-1343.

Koch, P.L., Tuross, N., and Fogel, M.L. 1997. The effects of sample treatment and diagenesis on the isotopic integrity of carbonate in biogenic hydroxylapatite. Journal of Archaeological Science, 24: 417-429.

Kohn, M.J. 1996. Predicting animal δ18O: accounting for diet and physiological adaptation. Geochimica et Cosmochimica Acta, 60: 4811-4829.

Kohn, M. and Cerling, T.E. 2002. Stable isotope compositions of biological apatite. In M. Kohn, J. Rakovan, and J. Hughes (eds.) Phosphates—Geochemical, Geobiological, and Materials Importance. Reviews in Mineralogy and Geochemistry, v. 48, Mineralogical Society of America, Washington D.C., 455– 488.

Kohn, M.J., Schoeninger, M.J., and Valley, J.W. 1998. Variability in oxygen isotope compositions of herbivore teeth: reflections of seasonality or developmental physiology? Chemical Geology, 152: 97-112.

Kolodny, Y. and Luz, B. 1991. Oxygen isotopes in phosphates of fossil fish; Devonian to Recent. In H.P. Taylor, Jr., J.R. O’Neil, and I.R. Kaplan (eds) Stable Isotope Geochemistry: A Tribute to Samuel Epstein. The Geochemical Society Special Publication, San Antonio, Texas, 105–119.

Kolodny, Y., Luz, B., Sander, M., and Clemens, W. A. 1996. Dinosaur bones: fossils or pseudomorphs? The pitfalls of physiology reconstruction from apatitic fossils. Palaeogeography, Palaeoclimatology, Palaeoecology, 126: 161–171.

Kretzoi, M. 1942. Bemerkungen zum system der nachmiozänen Nashorn-Gattungen. Földtani Közlöny, 72: 4-12.

Kutzbach, J. E., Prell, W. L., and Ruddiman, W. F. 1993. Sensitivity of Eurasian climate to surface uplift of the Tibetan Plateau. The Journal of Geology, 101: 177–190.

Lear, C., Elderfield, H., and Wilson, P. 2000. Cenozoic deep-sea temperatures and global ice volumes from Mg/Ca in benthic foraminiferal calcite. Science, 287: 269– 272.

158

Lee-Thorp, J.A. and van der Merwe, N.J. 1987. Carbon isotope analysis of fossil bone apatite. South African Journal of Science, 83: 712-715.

Lee-Thorp, J.A. and van der Merwe, N.J. 1991. Aspects of the chemistry of modern and fossil biological apatites. Journal of Archaeological Science, 18: 343-354.

Leloup, P.H., Lacassin, R., Tapponnier, P., Scharer, U., Dalai, Z., Xiaohan, L., Liangshang, Z., Shaocheng, J., and Trinh, P.T. 1995. The Ailao Shan-Red River shear zone (Yunnan, China), Tertiary transform boundary of Indochina. Tectonophysics, 251: 3-84.

Li, C., Wu, W., and Qiu, Z. 1984. The Chinese Neogene: subdivision and correlation. Vertebrata Palasiatica, 22: 163-178.

Li, F., Rousseau, D., Wu, N., Hao, Q., and Pei, Q.Y. 2008. Late Neogene evolution of the East Asian monsoon revealed by terrestrial mollusk record in Western Chinese Loess Plateau: From winter to summer dominated sub-regime. Earth and Planetary Science Letters, 274: 439.

Li, J., Zhang, J., Song, C., Zhao, Z., Zhang, Y., Wang, X., Zhang, J., and Cui, Q. 2007. The significance of the Bawanghe Formation in Longzhong Basin. Science in China Press, 37(1): 52-60.

Longinelli, A. 1984. Oxygen isotopes in mammal bone phosphate: a new tool for paleohydrological and paleoclimatological research? Geochimica et Cosmochimica Acta, 48: 385-390.

Longinelli, A. and Nuti, S. 1973. Oxygen isotope measurements of phosphate from fish teeth and bones. Earth and Planetary Science Letters, 20: 337–340.

Lucas, S.G. and Sobus, J.C. 1989. The systematics of indricotheres. In D.R. Prothero and R.M. Schoch (eds.) The Evolution of Perissodactyls. Oxford University Press, New York, 358- 378.

Luz, B., Kolodny, Y., and Horowitz, M. 1984. Fractionation of oxygen isotopes between mammalian bone-phosphate and environmental drinking water. Geochimica et Cosmochimica Acta, 48: 1689-1693.

Ma, Y.Z., Li, J.J., and Fang, X.M. 1998. Records of the climatic variation and pollen flora from the red beds at 30.6- 5.0 Ma in Linxia district. Chinese Science Bulletin, 43: 301-304. (In Chinese).

MacFadden, B. 1998. Tale of two rhinos: isotopic ecology, paleodiet, and niche differentiation of Aphelops and Teleoceras from the Florida Neogene. Paleobiology, 24: 274-286.

159

MacFadden, B.J. 2000. Cenozoic mammalian herbivores from the Americas: reconstructing ancient diets and terrestrial communities. Annual Review of Ecology, Evolution, and Systematics, 31: 33-59.

MacFadden, B.J. and Cerling, T.E. 1994. Fossil horses, carbon isotopes, and global change. Trends in Ecology and Evolution, 9: 481-485.

MacFadden, B.J., Wang, Y., Cerling, T.E., and Anaya, F. 1994. South American fossil mammals and carbon isotopes: a 25 million-year sequence from the Bolivian Andes. Palaeogeography, Palaeoclimatology, Palaeoecology, 107: 257-268.

Marshall, J.D., Brooks, J.R., and Lajtha, K. 2007. Sources of variation in the stable isotopic composition of plants. In R. Mitchener and K. Lajtha (eds.) Stable Isotopes in Ecology and Environmental Science. Blackwell Scientific, Oxford, 22-60.

McCrea, J.M. 1950. On the isotopic chemistry of carbonates and a paleotemperature scale. Journal of Chemical Physics, 18: 849-857.

Mecquenem, R. de. 1908. Contribution à l’étude du gisement de vertébras de Maragha et de ses environs. Annales d’Histoire Naturelle, Paris, 1(1): 27-79.

Medina, E. and Minchin, P. 1980. Stratification of δ13C values of leaves in Amazonian rain forests. Oecologia, 45: 377–378.

Medina, E., Montes, G., Cuevas, E., and Roksandic, Z. 1986. Profiles of CO2 concentration and δ13C values in tropical rain forests of the Upper Rio Negro Basin, Venezuela. Journal of Tropical Ecology, 2: 207–217.

Medina, E., Sternberg, L. & Cuevas, E. 1991. Vertical stratification of δ13C values in closed natural and plantation forests in the Luquillo mountains, Puerto Rico. Oecologia, 87: 369–372.

Morgan, M.E., Kingston, J.D., and Marino, B.D. 1994. Carbon isotopic evidence for the emergence of C4 plants in the Neogene from Pakistan and Kenya. Nature, 367: 162-164.

Murakami, T. 1987. Effects of the Tibetan Plateau. In C.P. Chang and T.N. Krishnamurti (eds.) Monsoon Meteorology. Oxford University Press, 235-270.

Nagy, K.A. 1989. Double-labeled water studies of vertebrate physiological ecology. In P.W. Rundel, J.R. Ehleringer, and K.A. Nagy (eds.) Stable Isotopes in Ecological Research, Volume 68. Springer-Verlag, New York, 270-287.

Ni, X. and Qiu, Z. 2002. The micromammalian fauna from the Leilao, Yuanmou hominoid locality: Implications for biochronology and paleoecology. Journal of Human Evolution, 42: 535-546.

160

O’Leary, M.H. 1988. Carbon isotopes in photosynthesis. BioScience, 38: 328-336.

O’Neil, J.R., Roe, L.J., Reinhard, E., and Blake, R.E. 1994. A rapid and precise method of oxygen isotope analysis of biogenic phosphate. Israel Journal of Earth Sciences, 43: 203- 212.

Owen-Smith, N. and Berger, J. 2006. Rhinoceroses. In D. Macdonald (ed.) The Encyclopedia of Mammals. The Brown Reference Group plc., London, 476-481.

Pagani, M., Freeman, K., and Arthur, M. 1999. Late Miocene atmospheric CO2 concentrations and the expansion of C4 grasses. Science, 285: 876-879.

Pan, Y. and Zong, G. 1993. Stratigraphy and Paleontology of the Yuanmou Basin, Yunnan Province, China (Chapter 4: Paleontology, Section 1: Mammalian faunas). In F. Qian and G. Zhou (eds.) Quaternary Geology and Paleoanthropology of Yuanmou, Yunnan, China, 19-28.

Passey, B.H., Ayliffe, L.K., Kaakinen, A., Zhang, Z., Eronen, J.T., Zhu, Y., Zhou, L., Cerling, T.E., and Fortelius, M. 2009. Strengthened East Asian summer monsoons during a period of high-latitude warmth? Isotopic evidence from Mio-Pliocene fossil mammals and soil carbonates from northern China. Earth and Planetary Science Letters, 277: 443-452.

Pei, W. 1961. Early Pleistocene mammalian fossils from Yuanmou, Yunnan (an appendix to the research of the Gigantopithecus Dongma fossils from Liucheng, Guangxi). Vertebrata Palasiatica, 1: 16-31.

Peterson, B.J. and Fry, B. 1987. Stable isotopes in ecosystem studies. Annual Review of Ecology, Evolution, and Systematics, 18: 293-320.

Prado, C. de. 1864. Descripción fisica y geológica de la provincia de Madrid. Junta General de Estadistica Madrid, 1-219.

Prothero, D.R. and Schoch, R.M. 1989. Classification of the Perissodactyla. In D.R. Prothero and R.M. Schoch (eds.) The Evolution of Perissodactyls. Oxford University Press, New York, 530-537.

Prothero, D.R., Guérin, C., and Manning, E. 1989. The History of the Rhinocerotoidea. In D.R. Prothero and R.M. Schoch (eds.) The Evolution of Perissodactyls. Oxford University Press, New York, 321-340.

Qi, G. 1979. Pliocene mammalian fauna of Lufeng, Yunnan. Vertebrata Palasiatica, 17: 14-22.

Qi, G. 1985. Stratigraphic summarization of Ramapithecus fossil locality, Lufeng, Yunnan. Acta Anthropologica Sinica, 4: 55-69.

161

Qi, G. 1993. The environment and ecology of the Lufeng hominoids. Journal of Human Evolution, 24: 3-11.

Qian, F. 1993. Stratigraphy and Paleontology of the Yuanmou Basin, Yunnan Province, China (Chapter1: Physical Geography and Regional Geology and Chapter 2: Stratigraphy). In F. Qian and G. Zhou (eds.) Quaternary Geology and Paleoanthropology of Yuanmou, Yunnan, China, 1-19.

Qiu, Z.X. and Yan, D.F. 1982. A horned Chilotherium skull from Yushe, Shansi. Vertebrata Palasiatica, 20: 122-132.

Qiu, Z. and Qiu, Z. 1995. Chronological sequence and subdivision of Chinese Neogene mammalian faunas. Palaeogeography, Palaeoclimatology, Palaeoecology, 116: 41-70.

Qiu, Z.X., Xie, J.Y., and Yan, D.F. 1987. A new chilothere skull from Hezheng, Gansu, China, with special reference to the Chinese “Diceratherium”. Scientia Sinica, Series B, 5: 545- 552.

Quade, J., Cerling, T.E., and Bowman, J.R. 1989. Development of Asian monsoon revealed by marked ecological shift during the latest Miocene in northern Pakistan. Nature, 342: 163- 166.

Quade, J., Cerling, T.E., Barry, J., Morgan, M., Pilbeam, D., Chivas, A., Lee-Thorp, J., and van der Merwe, N. 1992. A 16-Ma record of paleodiet using carbon and oxygen isotopes in fossil teeth from Pakistan. Chemical Geology, 94: 183-192.

Quade, J., Cater, J. M. L., Ojha, T. P., Adam, J., and Harrison, T. M. 1995. Dramatic carbon and oxygen isotopic shift in paleosols from Nepal and late Miocene environmental change across the northern Indian sub-continent. Geological Society of America Bulletin, 107: 1381-1397.

Qui, Z., Han, D., Qi, G., and Yufen, L. 2000. A preliminary report on a micromammalian assemblage from the hominoid locality of Lufeng Co. Yunnan Province. Acta Anthropologica Sinica, 4(1): 13-32.

Radinsky, L.B. 1967. A review of the rhinocerotoid family Hyracodontidae (Perissodactyla). Bulletin of the American Museum of Natural History, 136(1): 1-45.

Retallack, G. 2001. A 300-million-year record of atmospheric carbon dioxide from fossil cuticles. Nature, 411: 287-290.

Ringström, T. 1924. Nashorner der Hipparion-fauna Nord-Chinas. Palaeontologia Sinica, Series C, 1(4): 1-159.

162

Rowles, S.L. 1967. Chemistry of the mineral phase of dentine. In A.E.D. Miles (ed.) Structural and Chemical Organization of Teeth. Academic Press, London, 2: 201-245.

Royer, D., Wing, S., Beerling, D., Jolley, D., Koch, P., Hickey, L., and Berner, R. 2001. Paleobotanical evidence for near present-day levels of atmospheric CO2 during part of the Tertiary. Science, 292: 2310-2313.

Rozanski, K., Araguas-Araguas, L., and Gonfiantini, R. 1992. Relation between long-term trends of oxygen-18 isotope composition of precipitation and climate. Science, 258: 981-985.

Rozanski, K., Arguas-Araguas, L., and Gonfiantini, R. 1993. Isotopic patterns in modern global precipitation. In P. Swart (ed.) Climate Change in Continental Isotope Records. American Geophysical Union, Washington, 1 –36.

Schleser, G.H. and Jayasekera, R. 1985. δ13C-variations of leaves in forests as an indication of reassimilated CO2 from the soil. Oecologia, 65: 536–542.

Shackleton, N. and Kennett, J. 1975. Paleotemperature history of the Cenozoic and the initiation of Antarctic glaciation: oxygen and carbon isotope analyses in DSDP sites 277, 279, and 281. Initial Reports of the Deep Sea Drilling Project Leg 29, 743– 955.

Shackleton, N., Hall, M., and Pate, D. 1995. Pliocene stable isotope stratigraphy of site 846. Proceedings of the Ocean Drilling Program, Scientific Results, 138: 337– 355.

Sharp, Z.D. and Cerling, T.E. 1998. Fossil isotope records of seasonal climate and ecology: straight from the horse’s mouth. Geology, 26: 219-222.

Sharp, Z.D., Atudorei, V., and Furrer, H. 2000. The effect of diagenesis on oxygen isotope ratios of biogenic phosphates. American Journal of Science, 300: 222–237.

Shukla, J. 1987. Interannual variability of monsoons. In J. S. Fein and P. L. Stephens (eds.) Monsoons. John Wiley and Sons, 399–464.

Smith, B.N. and Epstein, S. 1971. Two categories of 13C/12C ratios for higher plants. Plant Physiology, 47: 380-384.

Smith, F. and Walker, N. 1980. Photosynthesis by aquatic plants: effects of unstirred layers in relation to assimilation of CO2 and HCO3 and to carbon isotopic discrimination. New Phytologist, 86: 245-259.

Spicer, R., Harris, N., Widdowson, M., Herman, A., Guo, S., Valdes, P., Wolf, J., and Kelley, S. 2003. Constant elevation of southern Tibet over the past 15 million years. Nature, 421: 622–624.

Sternberg, L., Mulkey, S., and Wright, S. 1989. Ecological interpretation of leaf carbon isotope ratios: influence of respired carbon dioxide. Ecology, 70: 1317-1324. 163

Stout, J.D., Goh, K.M., and Rafter, T.A. 1981. Chemistry and turnover of naturally occurring resistant organic compounds in soil. In E.A. Paul and J.N. Ladd (eds.) Soil Biochemistry, Vol. 5. Marcel Dekker, Inc., New York, 1-73.

Tang, H.F. and Hu, Y.H. 1993. Late Pliocene palynolflora of bearing coal sediments in western Yunnan. Coal Geology China, 5: 25-29.

Tapponnier, P., Xu, Z., Roger, F., Meyer, B., Arnaud, N., Wittlinger, G., and Yang, J. 2001. Oblique stepwise rise and growth of the Tibet Plateau. Science, 294: 1671-1677.

Teeri, J.A. and Stowe, L.G. 1976. Climatic patterns and the distribution of C4 grasses in North America. Oecologia, 23: 1-12.

Thermo-Finnigan. 2001. High Temperature Conversion Elemental Analyzer (TC/EA) Operating Manual.

Thermo-Finnigan. 2002. GasBench II Operating Manual.

Trautz, O.R. 1967. Crystalline organization of dental mineral. In A.E.D. Miles (ed.) Structural and Chemical Organization of Teeth. Academic Press, London, 2: 165-200. van der Merwe, N. and Medina, E. 1989. Photosynthesis and 13C/12C ratios in Amazon rain forests. Geochimica et Cosmochimica Acta, 53: 1091-1094.

Vogel, J.C. 1978. Recycling of carbon in a forest environment. Oecologia Plantarum, 13: 89–94.

Wall, W.P. 1989. The phylogenetic history and adaptive radiation of the Amynodontidae. In D.R. Prothero and R.M. Schoch (eds.) The Evolution of Perissodactyls. Oxford University Press, New York, 341-354.

Walter, H. 1971. Ecology of Tropical and Subtropical Vegetation. J.H. Burnett (Ed.). Oliver and Boyd, Edinburgh.

Wang, Y. and Cerling, T.E. 1994. A model of fossil tooth and bone diagenesis: implications for paleodiet reconstruction from stable isotopes. Palaeogeography, Palaeoclimatology, Palaeoecology, 107: 281-289.

Wang, Y. and Deng, T. 2005. A 25 m.y. isotopic record of paleodiet and environmental change from fossil mammals and paleosols from the NE margin of the Tibetan Plateau. Earth and Planetary Science Letters, 236: 322-338.

Wang, Y., Cerling, T.E., and MacFadden, B.J. 1994. Fossil horses and carbon isotopes: new evidence for Cenozoic dietary, habitat, and ecosystem changes in North America. Palaeogeography, Palaeoclimatology, Palaeoecology, 107: 269-279.

164

Wang, Y., Deng, T., and Biasatti, D. 2006. Ancient diets indicate significant uplift of southern Tibet after ca. 7 Ma. Geology 34: 309–312.

Wang, Y., Wang, X., Xu, Y., Zhang, C., Li, Q., Tseng, Z.J., Takeuchi, G., and Deng, T. 2008a. Stable isotopes in fossil mammals, fish and shells from Kunlun Pass Basin, Tibetan Plateau: Paleo-climatic and paleo-elevation implications. Earth and Planetary Science Letters, 270: 73–85.

Wang, Y., Kromhout, E., Zhang, C., Xu, Y., Parker, W., Deng, T., and Qiu, Z., 2008b. Stable isotopic variations in modern herbivore tooth enamel, plants and water on the Tibetan Plateau: Implications for paleoclimate and paleoelevation reconstructions. Palaeogeography, Palaeoclimatology, Palaeoecology, 260: 359-374.

Wang, Y., Deng, T., Flynn, L., Wang, X., Xu, Y., Lochner, E., Parker, W., and An, Y. 2009. A late Neogene record of environmental and tectonic changes in southern Tibet (submitted).

Webster, P. 1987. The elementary monsoon. In J. Fein and P. Stephens (eds.) Monsoons. John Wiley and Sons, New York, 3–32.

Webster, P.J., Magana, V.O., Palmer, T.N., Shukla, J., Tomas, R.A., Yanai, M., and Yasunari, T. 1998. Monsoons: Processes, predictability, and the prospects for prediction. Journal of Geophysical Research, 103: 14451–14510.

Wood, H.E., II. 1932. Status of Epiaceratherium (Rhinocerotidae). Journal of Mammalogy, 13(2): 169-170.

Xie, H., Richard, C., Xu, J., and Wang, J. 2001. Collective management of improved forage in Zhongdian County, Deqin, Tibetan Autonomous Prefecture, Northwest Yunnan, P.R. China. Report of the Center for Biodiversity and Indigenous Knowledge, 1-7.

Yakir, D. 1992. Variations in the natural abundance of oxygen-18 and deuterium in plant carbohydrates. Plant, Cell and Environment, 15: 1005-1020.

Yanai, M. and Li, C. 1994. Mechanism of heating and the boundary layer over the Tibetan Plateau. Monthly Weather Review, 122: 305-323.

Ye, D. 1981. Some characteristics of the summer circulation over the Qinghai-Xizang (Tibet) Plateau and its neighborhood. Bulletin of the American Meteorological Society, 62: 14- 19.

Yin, J. 1994. The Late Cenozoic paleomagnetic chronostratigraphy of Yuanmou Basin and its paleontological significance. Yunnan Geology, 13: 306-311.

Yin, A. and Harrison, M. 2000. Geologic evolution of the Himalayan-Tibetan orogen: Annual Review of Earth and Planetary Sciences, 28: 211–280.

165

Yin, J. and Liang, Q. 1998. Paleomagnetic stratigraphy of hominoid-bearing strata in Zhupeng area, Yuanmou, China. In J. Cheng, Z. Jiang, and X. Ji (eds.) Collected works for the 30th Anniversary of Yuanmou Man Discovery and the International Conference on Paleoanthropological Studies. Yunnan Science and Technology Press, Kunming, 235.

Yin, L. and Li, M. 1997. A study of the geographic distribution and ecology of C4 plants in China: I. C4 plant distribution in China and their relation with regional climatic condition. Acta Ecologica Sinica, 17: 350-363.

Yun, B. 1975. New material of about elephant fossil from Yunnan Province. Vertebrata Palasiatica, 13: 229-233.

Zachos, J., Pagani, M., Sloan, L., Thomas, E., and Billups, K. 2001. Trends, rhythms, and aberrations in global climate 65 Ma to present. Science, 292: 686– 693.

Zhang, Y., Long, Y., Ji, H., and Ding, S. 1999. The Cenozoic Deposits of the Yunnan Region. Professional Papers on Stratigraphy and Paleontology, No.7, Geological Publishing House, Beijing, 1-21.

Zheng, H., Powell, C., An, Z., Zhou, J., and Dong, G. 2000. Pliocene uplift of the northern Tibetan Plateau: Geology, 28: 715–718.

Zhu, R., An, Z., Potts, R., Hoffman, K.A. 2002. Magnetostratigraphic dating of early humans in China. Earth Science Reviews, 1275: 1-19.

Zong, G. 1987. Note on some mammalian fossils from the Early Pleistocene of Di-Qing County, Yunnan. Vertebrata Palasiatica, 25: 69-76.

Zong, G., Chen, W., Huang, X., and Xu, Q. 1996. Cenozoic mammals and environment of Hengduan Mountains region.

166

BIOGRAPHICAL SKETCH

Dana M. Biasatti

EDUCATION Ph.D. (2009) Florida State University, Tallahassee, Florida Major: Geology Area of Specialization: Stable isotope geochemistry Dissertation title: Paleoenvironments and Paleoecologies of Cenozoic Mammals from Western China based on stable carbon and oxygen Isotopes. Chair, Dr. Yang Wang

M.S. (2002) Southern Methodist University, Dallas, Texas Major: Geology Area of Specialization: Vertebrate paleontology and stable isotope geochemistry Thesis title: Stable isotopic profiles of sea turtle humeri and epizoic barnacles: implications for ecology and migration. Chair, Dr. Louis Jacobs

B.A. (1996) University of Texas at Dallas, Richardson, Texas Major: Biology Area of concentration: Molecular and cellular biology

RESEARCH INTERESTS My primary interest is the application of stable isotopes and other geochemical and biological techniques to understand climatic, ecological, behavioral, and physiological processes of fossil and extant taxa. I employ principles and methods from several disciplines, including stable isotope geochemistry, vertebrate paleontology, anatomy and physiology, geology, and environmental science to address unique questions that require a multi-disciplinary approach.

RESEARCH EXPERIENCE Dissertation Research, Florida State University, Tallahassee, Florida (2002-Present) This research involved the stable isotopic analyses of carbonate and phosphate from fossil tooth enamel, carbonates and organic materials from soils, carbonate nodules, and plants, to study the paleoecologies and paleoclimates, as well as some behavioral and physiological aspects of a number of mammalian taxa from the Neogene of western China. I was responsible for collection of samples in the field and from museum collections and for preparing and analyzing samples on a Delta plus XP continuous flow mass spectrometer using the following peripherals: gas bench, elemental analyzer, and a high temperature elemental analyzer (TCEA).

167

Research Assistant, National High Magnetic Field Laboratory, Tallahassee, Florida (2002- 2005) I assisted in a collaborative research project involving the isotopic analyses of carbonate and phosphate samples from fossil mammal teeth, carbonates and organic materials from soils, carbonate nodules, plants, and water in order to study Cenozoic climate change in northwestern China and to examine the relationship between climate change and the uplift of the Tibetan Plateau. I was responsible for preparing and analyzing samples on a Delta plus XP continuous flow mass spectrometer using the following peripherals: gas bench, elemental analyzer, and a high temperature elemental analyzer (TCEA). I was also responsible for organization of the lab and the ordering of supplies and prepared training manuals for use of the mass spectrometer and other laboratory equipment.

Master’s Thesis Research, Southern Methodist University, Dallas, Texas (1998-2002) This research involved stable carbon and oxygen isotopic analyses of carbonate samples from extant marine turtle bones and epizoic barnacles, as well as organic samples from marine plants and sea turtle muscle tissue, in order to study the ecology, migratory behaviors, and physiology of the turtles. I was responsible for collection of the samples in the field and from museum collections. I prepared and cryogenically purified all CO2 samples on a vacuum line. The samples were analyzed on a Finnigan-MAT 251 mass spectrometer.

Chemist, Safety-Kleen, Denton, Texas (1996-1997) I assisted in a research project involving the development of new techniques for hazardous organic waste disposal. I was responsible for analyzing samples using the following techniques: gas chromatography, total organic carbon measurements, bomb calorimetry, centrifugation, UV spectroscopy, and specific gravity measurements.

TEACHING EXPERIENCE Teaching Assistant, FSU Department of Geological Sciences (2002- Present) Lab instructor for Dynamic Earth (intro to Earth science) course and Historical Geology.

Teaching Assistant, Southern Methodist University Department of Geology (1998-2002) Lab instructor for Intro to Earth Sciences and Intro to Environmental Science.

Teaching Assistant, University of Texas at Dallas Department of Geology (1996-1998) Lab instructor for Earth science and mineralogy courses.

FIELD EXPERIENCE Yunnan Province, southwestern China (2004) Two weeks of field work involving the collection of Cenozoic mammal teeth from numerous fossil localities and museum collections.

Gyirong Basin, Tibet (2004) Two weeks of field work involving the collection of fossil mammal teeth from Miocene sediments.

168

Linxia Basin, Gansu Province, northwestern China (2003) Two weeks of field work involving the collection of Cenozoic mammal teeth from numerous fossil localities and museum collections.

Northwestern Territory, Guyana (2001) Two weeks of fieldwork involving the collection of sea turtle specimens for my M.S. research.

Ostional and Tortuguero, Costa Rica (1998-2001) Several months of field work that involved the collection of sea turtle and barnacle specimens for my M.S. research and participation in the Caribbean Conservation Corporation turtle-tagging program.

Geology of Mexico, University of Texas at Dallas (1996) A three-week graduate field course that involved mapping of paleontological sites, excavation of invertebrate fossils, and the study of igneous formations throughout central and southern Mexico.

Big Bend National Park, Texas (1995-1999) I discovered two sauropod dinosaur sites and assisted in the excavation and mapping of those sites for five years.

PROFESSIONAL AFFILIATIONS Society of Vertebrate Paleontology Geological Society of America

PUBLICATIONS Peer-Reviewed Publications Xiaoming Wang, Zhuding Qiu, Qiang Li, Banyue Wang, Zhanxiang Qiu, William R. Downs, Guangpu Xie, Junyi Xie, Tao Deng, Gary T. Takeuchi, Zhijie J. Tseng, Meeman Chang, Juan Liu, Yang Wang, Dana Biasatti, Zhencheng Sun, Xiaomin Fang, and Qingquan Meng. Vertebrate paleontology, biostratigraphy, geochronology, and paleoenvironment of Qaidam Basin in northern Tibetan Plateau. Palaeogeography, Palaeoclimatology, Palaeoecology, 254, 363–385 (2007).

Wang, Y., Deng, T., and Biasatti, D., Ancient diets indicate significant uplift of southern Tibet after ca.7 Ma, Geology, 34, 309-312 (2006)

Biasatti, D.M., Stable carbon isotopic profiles of sea turtle humeri: implications for ecology and physiology, Palaeogeography, Palaeoclimatology, Palaeoecology, 206, 203-216 (2004)

Presentations, Posters & Abstracts Biasatti, D.; Wang, Y. and Deng, T., Reconstruction of paleoecologies and paleoclimates of Cenozoic mammals from Northwest China based on stable isotopes. Society of Vertebrate Paleontology, Austin, Texas, October 17-20 (2007); Published in Journal of Vertebrate Paleontology Abstracts of Papers, 27 (Supplement to No. 3), 48 (2007)

169

Biasatti, D.; Wang, Y. and Deng, T., Reconstruction of ancient diets and habitats of Cenozoic rhinos from Northwest China based on stable isotopes. Society of Vertebrate Paleontology, Ottawa, Canada, October 18-21 (2006); Published in Journal of Vertebrate Paleontology Abstracts of Papers, 26 (Supplement to No. 3), 42 (2006)

Biasatti, D.; Wang, Y. and Deng, T., Evidence for a Plio-Pleistocene Strengthening of the Asian Monsoon and its Importance to the Understanding of Mammalian Evolution in Northwest China, The Geological Society of America, Salt Lake City, UT, October 16-19 (2005); Published in GSA Abstracts with Programs, 37 (7) (2005)

Biasatti, D.; Wang, Y. and Deng, T., Evidence for a Plio-Pleistocene Strengthening of the Asian Monsoon and its Importance to the Understanding of Mammalian Evolution in Northwest China, Society of Vertebrate Paleontology, Mesa, AZ, October 19-22 (2005); Published in Journal of Vertebrate Paleontology Abstracts of Papers, 25 (Supplement to No. 3), 37 (2005)

Wang, Y.; Deng, T. and Biasatti, D., Ancient diets indicate significant uplift of Southern Tibet after 7 MYR ago, The Geological Society of America, Salt Lake City, UT, October 16- 19 (2005); Published in GSA Abstracts with Programs, 37 (7), 57 (2005) Biasatti, D. and Wang, Y., Paleoecology of the Cenozoic rhino fauna of Linxia Basin, Gansu, NW China, from carbon and oxygen isotopes in tooth enamel, Geological Society of America Annual Meeting, Denver, CO, November 7-9 (2004); Published in GSA Abstracts with Programs, 36 (5), 38 (2004)

Biasatti, D.; Wang, Y. and Deng, T., Paleoecology of the Cenozoic Rhino Fauna of Linxia Basin, Gansu, NW China, from carbon and oxygen isotopes in tooth enamel, 64th Annual Meeting of the Society of Vertebrate Paleontology, Denver, CO, November 3-5 (2004); Published in Journal of Vertebrate Paleontology Abstracts of Papers, 24, 39 (2004)

Biasatti, D., Analyses of modern sea turtle humeri suggest respiratory effects on carbon isotope compositions of biogenic structural carbonate, Society of Vertebrate Paleontology, St. Paul, MN, October 15-18 (2003); Published in Journal of Vertebrate Paleontology Abstracts of Papers, 23 (Supplement to No. 3), 1-128 (2003) Biasatti, D.M. Implications of carbon isotopic profiles of sea turtle humeri and epizoic barnacle communities, 22nd Annual Symposium on Sea Turtle Biology and Conservation, Miami, FL, April 4-7 (2002); Published in Proceedings of the 22nd Annual Symposium on Sea Turtle Biology and Conservation, NOAA technical Memorandum NMFS-SEFSC- 503, 277 (2002) Biasatti, D.M. Implications of isotopic profiles of sea turtle humeri and epizoic barnacle communities, Sixty-first annual meeting for the Society of Vertebrate Paleontology, Bozeman, MT, October 3-6 (2001), Published in Journal of Vertebrate Paleontology Abstracts of Papers, 21 (Supplement to No. 3), 34 (2001)

170