Using Stable Carbon Isotope, Microwear, and Mesowear

USING STABLE CARBON ISOTOPE, MICROWEAR, AND MESOWEAR ANALYSES TO DETERMINE THE PALEODIETS OF NEOGENE UNGULATES AND THE PRESENCE 0F C4 OR C3 GRASSES IN NORTHERN AND CENTRAL FLORIDA

JONATHAN M. HOFFMAN

A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE

UNIVERSITY OF FLORIDA

2006

Jonathan M. Hoffman

ACKNOWLEDGMENTS

I would like to thank my advisor, Dr. Jonathan Bloch for his guidance on this project. I would also like to thank my committee members: Dr. David Hodell, Dr.

Richard Hulbert, and Dr. Andrew Zimmerman. Their suggestions and advice are greatly appreciated. In addition to their helpful advice, Dr. Bloch, Dr. Hulbert, and Dr.

Zimmerman all aided in collecting fossil samples. I would like to thank Art Poyer and

Jeremy Green for their help in collecting samples. I am especially appreciative of the

Englehard Corporation and Dave Mihalik for allowing me to collect fossils from their mines and being incredibly helpful at those sites.

I would like to thank Dr. Penny Higgins for her assistance in learning the sampling techniques for stable isotope analysis of fossil teeth and the chemical protocol for preparing those samples for analysis. I am greatly indebted to Dr. Jason Curtis for running my samples on the mass spectrometer and for all of his advice.

My fellow graduate students aided me with discussion and advice. I am very appreciative of P.J. Moore, Warren Grice, Jane Gustavson, and Derrick Newkirk for listening and providing support. Finally, I would like to thank my family, who has always supported me and encouraged me to pursue my passions.

iii TABLE OF CONTENTS

page

ACKNOWLEDGMENTS ...... iii

LIST OF TABLES...... vi

LIST OF FIGURES ...... vii

ABSTRACT...... ix

CHAPTER

1 INTRODUCTION ...... 1

2 FIELD AREA AND FAUNA...... 17

3 MESOWEAR METHOD ...... 29

Materials and Methods ...... 31 Results...... 32

4 MICROWEAR METHOD ...... 40

Materials and Methods ...... 41 Results...... 44

5 STABLE ISOTOPE ANALYSIS...... 48

Background...... 48 Materials and Methods ...... 54 Results...... 57

6 DISCUSSION...... 67

7 CONCLUSIONS ...... 80

APPENDIX

A MESOWEAR VALUES LISTED BY SAMPLE...... 82

B SERIAL STABLE ISOTOPE VALUES...... 86

iv LIST OF REFERENCES...... 88

BIOGRAPHICAL SKETCH ...... 102

LIST OF TABLES

Table page

2-1 Biochronological ranges of the Willacoochee Creek Fauna...... 23

3-1 Systematic and morphological information about the taxa used in the mesowear analysis...... 35

3-2 Observed percentages of mesowear attributes (% high and % low refer to the percentage of specimens with high or low occlusal relief and % sharp, % round, and % blunt refer to the percentage of specimens with sharp, round or blunt cusp shape)...... 36

3-3 Dietary classifications based on hierarchical cluster analyses of mesowear attributes from Table 3-2...... 36

4-1 Individual microwear counts for 3 specimens of Acritohippus isonesus and 7 specimens of Neohipparion trampasense...... 46

4-2 Calculated average microwear values and percentages for Acritohippus isonesus and Neohipparion trampasense...... 47

4-3 Isotopic data for equids from 4 central Florida sites...... 47

5-1 List of the 24 specimens analyzed for stable carbon and oxygen isotope analysis..60

5-2 Bulk stable carbon and oxygen isotope values for Willacoochee Creek Fauna...... 61

5-3 Descriptive statistics of the δ13C and δ18O values for 8 herbivores from the Willacoochee Creek Fauna...... 62

6-1 Isotopic data for equids from 4 central Florida sites...... 79

A-1 Abbreviations: UF ID = catalogue number in Florida Museum of Natural History collections,...... 82

B-1 Abbreviations: ELC = Englehard La Camelia Mine, MGF = Milwhite Gunn Farm Mine, and LC2 = La Camelia 2 Mine...... 86

LIST OF FIGURES

Figure page

2-1 Index map of middle Miocene localities in northern Florida...... 22

2-2 Composite section of the Torreya Formation and the location of the Willacoochee Creek Fauna within the Dogtown Member...... 24

2-3 Correlated stratigraphic sections of the Englehard La Camelia and Milwhite Gunn Farm Mines...... 25

2-4 Fresh cut at the newest site, the Crescent Lake Mine, in Decatur County, Georgia...... 26

2-5 Stratigraphic and temporal distribution of the 3 fossil sites studied in the mesowear analysis, as well as a fourth (Moss Acres) ...... 27

2-6 Tooth positions...... 28

3-1 Dendrogram illustrating the hierarchical cluster analysis of the 27 ‘typical’ grazers, browsers, and mixed feeders from Fortelius and Solounias (2000)...... 34

3-2 Examples of typical mesowear attributes...... 35

3-3 Dendrograms illustrating the hierarchical cluster analyses of 6 studied taxa amongst the 27 ‘typical’ grazers, browsers, and mixed feeders from Fortelius and Solounias (2000)...... 37

5-1 Schematic of isotopic fractionation between atmospheric carbon and C3 plants, as well as fractionation between C3 plants and ruminant herbivores...... 59

5-2 Plot of bulk δ13C vs. δ18O values for the 8 herbivores from the Willacoochee Creek Fauna...... 63

5-3 δ13C values for serial samples of two Aphelops specimens...... 64

5-4 δ18O values for serial samples of two Aphelops specimens...... 64

5-5 δ13C values for serial samples of two Merychippus primus specimens...... 65

5-6 δ18O values for serial samples of two Merychippus primus specimens...... 65

vii 5-7 δ13C values for serial samples of two Acriohippus isonesus specimens...... 66

5-8 δ18O values for serial sampling of two Acritohippus isonesus specimens...... 66

6-1 Plot of microwear index versus δ13C values...... 78

viii

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science

USING STABLE CARBON ISOTOPE, MICROWEAR, AND MESOWEAR ANALYSES TO DETERMINE THE PALEODIETS OF NEOGENE UNGULATES AND THE PRESENCE 0F C4 OR C3 GRASSES IN NORTHRN AND CENTRAL FLORIDA

Jonathan M. Hoffman

December 2006

Chair: Jonathan I. Bloch Major Department: Geology

Traditionally, hypsodont (high-crowned) teeth in North American ungulates

(hoofed mammals) were thought to have coevolved with grasses during the middle

Miocene. Isotopic evidence has demonstrated that tropical C4 grasses were not dominant

and therefore not abundant enough to be responsible for this adaptive radiation. It has

been proposed that high-altitude C3 grasses were extensive throughout the Great Plains

and were the dietary driving force behind the grazing adaptations. This study will test

this hypothesis to see if it applies to the middle Miocene of the Southeastern United

States. The δ13C values from 24 specimens of 8 ungulate taxa from the Willacoochee

Creek Fauna, an assemblage of middle Miocene mammals from northern Florida and

southern Georgia, are presented here to determine if there is a significant C4 grass

component in mammalian paleodiets. The δ13C values indicate that all 8 taxa were consuming C3 plant material, either browse or grass. Furthermore, microwear analyses

ix conducted on 3 specimens of the most hypsodont taxon indicate that the mammal was

13 eating grass. Combined with the δ C data, this study concludes that C3 grasses were

present in the middle Miocene of northern Florida and at least one hypsodont mammal

was consuming them. This evidence supports a C3 grass hypothesis of hypsodont

radiations. Also, this study combines the mesowear paleodietary analysis with previously

published isotopic data to 5 equid populations from 3 sites in central Florida, ranging in age from ~9.5 Ma to ~1.5 Ma, to trace the possible influence of C3 grasses on ungulate

diets. C3 grasses were the primary food source for these horses until approximately 7

Ma. After that, the horses fed on a mixed diet of C3 and C4 grasses until about 1.5 Ma. At

that point in Florida, the abundance of C3 grasses had diminished and the grazers

primarily fed on C4 grasses.

x CHAPTER 1 INTRODUCTION

Today, over 25% of North American natural biomes are nonforest (Webb, 1977).

The modern vegetation of North America is a sharp contrast to vegetation at the beginning of the Cenozoic, when nearly all of North America was covered with forests

(Webb, 1977). Traditionally, paleontologists have believed that, beginning in the

Paleocene, parts of North America underwent a stepwise progression from forest to

savanna to grassland biomes (Webb, 1977). For most of the Paleocene, North America was dominated by evergreen forests and cypress swamps (Wolfe, 1985; Wing and

Tiffney, 1987). It has been proposed that grasses originated amongst this vegetation in the early Paleocene, although no direct evidence of grasses has been found in sediments of that age (Linder, 1986; Crepet and Feldman, 1991). Nearly all the mammals in the

Paleocene belong to four orders: the Multiuberculata, Insectivora, Primates, and

‘Condylarthra’ (Webb and Opdyke, 1995). These mammals were small to medium in size and were arboreal or scansorial (Webb and Opdyke, 1995). The multituberculate

Ptilodus, for example, possesses morphological adaptations consistent with those of a

modern tree squirrel, suggestive of an arboreal habit (Jenkins and Krause, 1983).

The grassland transitional sequence began meekly, during the late Paleocene with

the appearance of possible protosavannas in scattered open-country areas that constitute

breaks in the forest coverage (Webb, 1977). In the late Paleocene Crazy Mountain Field

of Montana, Simpson (1937) found that 90% of the fossils collected from floodplain

sediments consisted of carnivores and archaic ungulates such as perptychids,

1 2 phenacodontids, and arctocyonids. In the same area, taxa collected from swampy woodland deposits were typically smaller, arboreal mammals (Simpson, 1937). A similar pattern is apparent in late Paleocene and early Eocene sediments of the Rocky Mountain intermontane basins. The late Paleocene Fort Union Formation from that region consists of gray sediments indicative of swampy woodlands bounded by alluvial plains (Van

Houten, 1945; Bown, 1980). Small arboreal mammals have been collected from these sediments, corroborating the environmental interpretation of woodlands (Van Houten,

1945). The early Eocene Willwood Formation, located in the same area, exhibits red- banded flood plain sediments and ungulates suited for open-country habitation (Van

Houten, 1945; Bown, 1980; Hickey, 1980; Wing, 1980). Van Houten (1945) concluded that the late Paleocene Rocky Mountain woodlands had, by the early Eocene, developed flood plain savannas that existed in breaks in the forest.

A number of subtle modifications are evident in probable open-country taxa during the late Paleocene and early Eocene. The condylarth Meniscotherium exhibits molarized premolars, molar crescents, and some cursorial limb elongation (Gazin, 1965), attributes that improved the mastication of coarser vegetation and open-country locomotion. In the

Torrejonian North American Land Mammal Age (NALMA) of the middle Paleocene, three orders of larger Asian immigrants arrived in North America: the Pantodonta,

Taeniodonta, and Dinocerata (Webb and Opdyke, 1995). Titanoides, a late Paleocene pantodont, bears digging forelimbs that would have excavated coarse roots (Webb, 1977).

Taeniodonts, large clawed opossum-like root grubbers, were the first mammals to develop hypsodont (high-crowned) teeth (Patterson, 1949). Crested molars and molarized premolars are also seen at this time in the uintatheres (Order Dinocerata), which bear

"hornlike protuberances" indicative of herding behavior seen in open-country ungulates

(Wheeler, 1961).

Global climate fluctuated during the Paleocene. Deciduous tree populations began

increasing at about 63 Ma, indicative of cooling climate (Rose, 1981). This cooling trend

began near the boundary of the Torrejonian and Tiffanian and continued through the

Tiffanian, accompanied by the disappearance of small mammals and an increased

presence of larger mammals (Webb and Opdyke, 1995). The Tiffanian cooling trend was

followed by the initial appearance of open-country habitats that coincides with a global

warming trend at the end of the Tiffanian (Koch, Zachos, and Gingerich, 1992). This

warming trend began in the middle Paleocene (about 59 Ma) and peaked and ended with

the early Eocene Climatic Optimum (EECO, 52–50 Ma) (Zachos, Pagani, Sloan, Thomas,

and Billups, 2001). At northern latitudes, the peak mean annual temperature in the early

Eocene was between 25°C and 30°C, 15°C to 20°C warmer than today (Novacek, 1999).

The warming trend and EECO are marked by a 1.5‰ decrease in δ18O values from

benthic foraminifera, the lowest such δ18O values in the Cenozoic (Zachos et al., 2001).

This global warmth had a major impact on both North American floral and faunal communities. Between 55 and 53 Ma, the number of macrofloral species doubled in the

Paleocene/Eocene of the Bighorn Basin (Wing, Alroy, and Hickey, 1995). Early Eocene fossil leaves exhibit “drip-tips” and smooth margins indicative of a tropical/subtropical climate (Wolfe, 1978; Prothero, 1994). The evergreen forests of the Paleocene had given way to early Eocene subtropical forests, while still maintaining some open-country enclaves (MacGinitie, 1974; Rose, 1981; Bown and Krause, 1981, 1987).

The beginning of the Clarkforkian (latest Paleocene), at about 56 Ma, is marked by

the first major immigration episode of Asian mammals capable of exploiting the new

North American open-country niches (Webb and Opdyke, 1995; Lofgren, Lillegraven,

Clemens, Gingerich, and Williamson, 2004). The newly arrived Asian orders included

the Tillodontia and Rodentia, as well as the pantodont family Coryphodontidae (Rose,

1981; Krause and Maas, 1990). Many of these mammals had adapted to grazing lifestyles in Asia. For example, Coryphodon, one of the Asian pantodonts that arrived in North

America in the late Paleocene, is believed to have already been a hippo-like amphibious grazer that, based on canine grooves, rooted for food (Simons, 1960). In all, nine genera immigrated from Asia to North America in the Clarkforkian (Stucky, 1990).

The Clarkforkian immigration episode continued into the Wasatchian, in the earliest Eocene (Webb and Opdyke, 1995). The Clarkforkian-Wasatchian boundary correlates with the Paleocene-Eocene boundary at about 55 Ma (Gingerich, 2001). The early Wasatchian was subject to the largest mammal immigration wave in the North

American fossil record. At this time, European mammals crossed the North Atlantic over the Thulean land bridge (Webb and Opdyke, 1995). This immigration episode included the first North American appearances of order Perissodactyla (Hyracotherium), order

Artiodactyla (Diacodexis), the creodont family Hyaenodontidae, and the primate families

Adapidae and Omomyidae (Rose, 1981). Direct evidence of grass during the late

Paleocene and early Eocene is rare, but there are some fossil grasses that indicate the presence of protosavannas: the oldest North American grass macrofossil, early Eocene in age, comes from the Paleocene/ Eocene Wilcox Formation of western Tennessee (Crepet and Feldman, 1991).

The morphological subtleties visible during the late Paleocene/early Eocene gave way, in the middle and late Eocene, to more pronounced adaptations and shifts in both floral and faunal communities due to the changing global climate. In the middle Eocene, the global climate began cooling again and becoming drier, as evident from a 3.0‰ increase in benthic foraminifera δ18O values over a 17 million year span, from 50 Ma to

33 Ma (Savin, 1977; Zachos et al., 2001). The increase in δ18O values in the middle

Eocene (50–48 Ma) resulted entirely from a decrease in deep-sea temperature from about

12°C to about 4.5°C (Zachos et al., 2001). Subsequent ice sheet growth began by the late

Eocene (34 Ma) and was responsible for further δ18O enrichment (Miller and Katz, 1987;

Zachos, Stott, and Lohmann, 1994; Zachos et al., 2001). It was also during the middle

Eocene that the first signs of seasonal aridity, such as evaporites and oxidized redbeds,

appear in the Rocky Mountain region (Webb and Opdyke, 1995).

As aridity and cooling increased in the middle Eocene, there was a major faunal

turnover in the subtropical forests. By the end of the Duchesnean, 80% of the terrestrial

mammal genera present in the Uintan (the previous age) had become extinct (Savage and

Russell, 1983; Stucky, 1990). The gradual disappearance of tropical forests in North

America prompted the decline of arboreal creatures such as primates (Webb, 1977). The

adapids and paromoyids, the last of the North American primates, disappeared at the end

of the Duchesnean (Prothero, 1994). The orders Condylarthra, Tillodontia, Dinocerata,

and Taeniodonta also disappeared in the late Duchesnean (~40 Ma), while many new

Asian immigrants arrived, most notably the eubrontotheres such as Duchesneodus

(Webb, 1977; Emry, 1981; Krishtalka et al., 1987; Prothero, 1994).

By the late Eocene, the cooler and drier climate allowed for the first true savannas to dominate the midcontinent (Webb, 1977; Wing and Tiffney, 1987). Savannas, as defined by Sears (1969), are any biomes that are subtropical open-country plains with

some trees, which includes areas such as thorn scrub and open deciduous forests, but do

not include open steppe or grasslands. The savannas that appeared in the late Eocene

were savanna woodlands, typified by the presence of the Leguminosae, Sapindaceae, and

Anacardiaceae floral families, similar to those of the modern Chihuahua region of

Mexico (Webb, 1977). As the global cooling continued, tropical flora and subtropical

forests retreated south of the Rocky Mountains (Leopold and MacGinitie, 1972).

Members of the grass family Poaceae are also present in the late Eocene.

Representatives of the subfamily Pooideae, from Tribes Stipeae (“e.g.”, Stipa florissanti)

and Phalarideae, appear in the Florissant floral assemblage of Colorado, at about 34 Ma

(MacGinitie, 1953; Stebbins, 1981). The presence of these two tribes, endmembers of

two separate evolutionary lineages, suggests that the Pooideae was well-differentiated by

the end of the Eocene (Stebbins, 1981).

The new savanna woodland habitats were exploited by a number of adapting North

American mammals, as well as by late Eocene Asian immigrants already adapted to

savanna habitats. The combined autochthounous and invasive taxa, which followed the

Duchesnean faunal turnover, are termed the “White River chronofauna” (Emry, 1981).

Chronofaunas are assemblages of species that remain compositionally stable over a

significant amount of time (Olson, 1952). The “White River chronofauna” consists of an

increase in herbivore genera and species from the late Eocene through the Oligocene,

beginning around 40 Ma (Krishtalka et al., 1987; Webb and Opdyke, 1995). The number

of identified browsing mammalian genera rose from 8 in the Duchesnean to about 40

after the Duchesnean (Stucky, 1990). Likewise, the overall number of mammalian

herbivore species rose from less than 40 to about 90 during the Eocene-Oligocene

transition (Savage and Russell, 1983). The higher species numbers were maintained

throughout the Oligocene (Webb, 1989). Numerous modern mammalian families make

their first appearance in North America during the late Eocene, including: soricid

insectivores; sciurid, castorid, cricetid, and heteromyid rodents; leporid lagomorphs;

canid and mustelid carnivores; camlids; tyassuids; and rhinocerotids (Webb and Opdyke,

1995).

By the middle Oligocene, the “White River chronofauna” had become the first

North American chronofauna to exhibit significant diversity of hypsodont herbivores,

from rodents to ungulates (Gregory, 1971; Webb and Opdyke, 1995). These hypsodont

taxa include: leporids, castorids, eomyids, rhinocerotids, hypertragulids, oromerycids,

and oreodonts (Webb, 1977). The oromerycid Montanatylopus, for example, has molars

significantly more hypsodont than its brachydont sister taxa (Prothero, 1986). The newest

Asian taxa, introduced through the connection of the North American and Asian continents (McKenna, 1972), included a suite of selenodont (crescent-toothed) artiodactyls already adapted for savanna feeding (Webb, 1977). Included in this group are the families Camelidae, Hypertragulidae, Leptomerycidae, and Agriochoeridae (Webb,

1977). The newly introduced taxa flourished in the fresh savanna environments. One

taxon native to North America, Hyopsodus, acquired more lophodont (crest-toothed)

molars to chew coarser food (Gazin, 1968). Native rodents, such as the protoptychids and cylindrodontids, developed open-country locomotive adaptations as well as dentitions

8 suited for coarser foods (Wood, 1962; Black and Dawson, 1966; Galbreath, 1969;

Wahlert, 1973). Larger herbivores existed in two general groups: semiamphibious streamdwellers such as amynodont rhinos and long-limbed, cursorial taxa that roamed the interfluves, such as equids (Mesohippus) and selenodont artiodactyls (Wall, 1982).

These adaptations, along with the diversification of many of these cursorial and hypsodont taxa, mark the dominance of the woodland savannas that had become widespread in the Oligocene (Webb, 1977). It is also likely that the “White River chronofauna” established a positive feedback loop with the savanna biomes:

Large herbivores preferred feeding in more open woodlands; in turn, expansion of open formations facilitated evolution of mixed-feeding herbivores. (p., 192, Webb and Opdyke, 1995)

Additionally, there was a negative correlation between browsers and the spread of open- country habitats (Webb and Opdyke, 1995). By the late Oligocene, browsers such as titanotheres had disappeared as woodland savannas continued to spread (Webb and

Opdyke, 1995).

The expansion of woodland savannas is supported not only by hypsodont radiations but other lines of both floral and faunal evidence. Aquatic reptiles in the Rocky Mountain area underwent severe population decreases as a result of aridity and seasonality during the late Eocene and Oligocene (Hutchinson, 1982). The late Eocene Florissant Flora of

Colorado records a decrease in the percentage of entire-margined leaves, indicating a drop in mean annual temperature from 10°C to ~12.5°C (MacGinitie, 1962; Wolfe,

1985). Additionally, pedological studies on the Brule Formation of South Dakota (~33

Ma, early Orellan) and the Upper John Day Formation in central Oregon (~30 Ma,

9 earliest Arikareenan) suggest the presence of desert bunch grasslands during the early/middle Oligocene (Retallack, 1997; Retallack, 2001; Retallack, 2004).

Grass, however, was still relatively rare in the fossil record of the Eocene and

Oligocene (Frederickson, 1981; Webb and Opdyke, 1995; Jacobs, Kingston, and Jacobs,

1999). There are two plausible explanations for this rarity. The first is that woody shrubs dominated the landscape prior to the profusion of grasses (Huber, 1982). The second possibility is that grasses suffer from a taphonomic bias, preventing the preservation of grasses despite their actual abundance (Webb and Opdyke, 1995). Grasses are known to be abundant elsewhere during the middle Eocene, including Australia (Truswell and

Harris, 1982) and Europe (Litke, 1968).

Immediately following the Eocene-Oligocene transition, the Drake and Tasmania

Passages opened (at ~31 and 32 Ma, respectively), increasing ocean circulation and proliferating the global cooling trend from the Eocene (Lawver and Gahagan, 2003). This also allowed the establishment and preservation of permanent Antarctic ice sheets

(Hambrey, Ehrmann, and Larsen, 1991). The decrease in global temperature continued until the late Oligocene (26 to 27 Ma), when another warming trend began (Miller,

Wright, and Fairbanks, 1991; Wright, Miller, and Fairbanks, 1992). This warming trend reduced Antarctic ice sheet volume until the middle Miocene, ~15 Ma (Miller et al.,

1991; Wright et al., 1992). There were some short glaciation events interspersed throughout this approximately 12 million year interval (Wright and Miller, 1993). This warming reached its zenith during the middle Miocene Climatic Optimum (MCO), from

17-15 Ma, which is evident from decreased δ18O values (Vincent, Killingley, and Berger,

1985; Flower and Kennett, 1995). Oceanic and atmospheric cooling and ice sheet growth

followed the MCO, marked by about a 1‰ increase in foraminifera δ18O values from

14.0 to 13.8 Ma (Flower and Kennett, 1995).

There are two hypotheses for the mechanism that drove the middle Miocene

climate variability: greenhouse gases and oceanic circulation (Zachos et al., 1994). From

16.5 to 13.5 Ma, overlapping the MCO, benthic foraminifera δ13C values were elevated

as high as 2.2‰ (Vincent and Berger, 1985). This event, termed the Monterey Excursion,

possibly resulted from the draw down of atmospheric pCO2 through organic carbon

burial in marginal marine sediments (Vincent and Berger, 1985). It has been proposed

that the Monterey Excursion drove middle Miocene climate variability, although there is

a 2.5 million year lag between the onset of the proposed pCO2 draw down (16.5 Ma) and

the atmospheric cooling indicated by the δ18O increase at 14 Ma (Vincent and Berger,

1985; Hodell and Woodruff, 1994). Atmospheric CO2 levels and global climate may have

remained high during this lag due to the outgassing of CO2 from the Columbia River

Flood Basalt from 17 to 14.5 Ma (Hodell and Woodruff, 1994). However, other

researchers have suggested that atmospheric CO2 levels were low from 17 Ma to 14 Ma

(Pagani, Freeman, and Arthur, 1999; Flower, 1999; Royer et al., 2001). The middle

Miocene climate variability, therefore, may have resulted from the opening and closing of

tectonic gateways that altered oceanic circulation (Woodruff and Savin, 1989, 1991;

Raymo, 1994).

The first immigration episode of the Miocene occurred at approximately the same

time as the beginning of the MCO, from 18-17 Ma in the middle Hemingfordian (Webb

and Opdyke, 1995). These Asian immigrants include eomyid rodents, the

biostratigraphically important cricetid rodent Copemys, and the first true cats in the New

World such as Pseudaelurus (Webb and Opdyke, 1995). The immigration also included

four “megaherbivores”: the rhinocerotids Teleoceras and Aphelops and the proboscideans

Miomastodon and Gomphotherium (Webb and Opdyke, 1995). It is assumed that these

“megaherbivores” modified the savanna landscape much like modern elephants,

however, there is no direct evidence to support this hypothesis (Owen-Smith, 1988).

These immigrants comprise part of the “Sheep Creek chronofauna” of the early and

middle Miocene. The “Sheep Creek chronofauna” is also important because it chronicles

the grazing advancement of horses, with the transition of the browsing/mixed feeding

Parahippus to the grazing Merychippus (Hulbert and MacFadden, 1991).

Treeless grassland prairies were initially thought to have become widespread in the

Great Plains at the beginning of the Barstovian Land Mammal Age (middle Miocene,

~15.8 Ma), replacing the steppe savannas (Kowalevsky, 1872; Webb, 1983). This timing

would correlate with the end of the middle Miocene climatic optimum. This

interpretation for grassland expansion was based largely on the prevalence of mammals

with grazing adaptations, especially horses, and has been viewed as a classic example of coevolution. Mesodont horses such as Parahippus gave way to hypsodont horses, like

Merychippus and Hipparion, that also exhibited increased enamel folding and cement deposition on the cheek teeth (Webb, 1977). These adaptations, which improved the grinding surface of the tooth, were also featured in camels, pronghorns, oreodonts, rhinoceroses (diceratherine and teleoceratine), and at least four genera of gomphotheriid

proboscideans (Webb, 1977). The diversification of these taxa is directly correlated to the

degree of hypsodonty; the higher-crowned lineages experienced greater radiations

(Webb, 1977). The succession of grazers, starting at the beginning of the Barstovian

(15.8 Ma) and lasting through the end of the Clarendonian (8.8 Ma), has been termed the

'Clarendonian chronofauna' (Webb, 1983). It has also been suggested that this faunal succession (and therefore the appearance of grasses) began during the beginning of the late Hemingfordian at 17.5 Ma (Theodor, Janis, and Broekhuizen, 1998).

In addition to dental adaptations for grazing, many ungulates acquired elongated limb modifications for open-country locomotion (Webb, 1977). Hypsodont horses, for example, developed digital springing ligaments at about the same time that they developed grazing dentitions (Camp and Smith, 1942). Rodents also adapted to savanna habitats in the Miocene. Their teeth became higher-crowned (Rensberger, 1973) and their limbs adapted for burrowing habits (Webb, 1977). Mylagaulids (Fagan, 1960), heteromyids (Lindsay, 1972), geomyoids (Rensberger, 1971), and ochotonid rabbits

(Green, 1972) all diversified into an array of hypsodont burrowers. The abundance of needlegrass taxa, specifically Stipidium and Berriochloa, in the High Plains region (Elias,

1942) seemingly confirmed the dominance of grasslands in the Great Plains during the

middle Miocene.

The traditional view of grassland evolution was modified by a study of the late

Miocene (12–13 Ma) Kilgore Flora of Nebraska. This floral assemblage depicted an environment consisting of savannas with mesic and open grassy forests, but lacking any

open prairies (MacGinitie, 1962). Faunal evidence corroborates this claim. The presence

of arboreal rodents, primates, insectivores, brachydont browsers and mixed feeders

suggests that the Great Plains was still a woodland savanna with riparian forests through

the late Miocene (Gregory, 1971). Grasses, therefore, were abundant through the late

Miocene, but grasslands had yet to dominate the landscape. It was not until the Late

Hemphillian (Early Pliocene, ~5 Ma), that treeless steppe grasslands swept across the

Great Plains. The vertebrate evidence for this consists of the absence of arboreal and

browsing taxa, a limited diversity of grazing taxa, and an overall lower diversity of all vertebrate taxa (Gregory, 1971).

The traditional grassland story has been further modified, in more recent years, from multiple types of studies. Paleosol studies indicate the presence of short sod grasslands in the Great Plains region in the early Miocene, ~19 Ma, and tall sod grasslands by the late Miocene, ~7 Ma (Retallack, 1997; Retallack, 2001). Paleosol studies also indicate the presence of sod grasslands in the Hemingfordian (early Miocene,

~ 19 Ma) in central Oregon (Retallack, 2004). Phytoliths, the silica granules found in many plants, have also been used to identify savanna environments. Morphological studies on phytolith assemblages from northwestern Nebraska indicate that open-habitat

grasses were present in savanna and woodland environments by the early Miocene

(Strömberg, 2002; Strömberg, 2004).

The most extreme revisions have come from stable isotope studies. Stable carbon isotopes from paleosols and fossil tooth enamel from various regions reveal a global increase in the C4 biomass during the late Miocene and early Pliocene (7–5 Ma) (Quade

et al., 1992; Cerling, Wang, and Quade, 1993; Wang, Cerling, and MacFadden, 1994;

MacFadden and Cerling, 1996; Cerling et al., 1997b). This observation was based

primarily on the bioapatite of the fossil teeth, which reflect the C3/C4 plant constituents of

13 the animal’s diet through δ C ratios. It was proposed that the increase in C4 biomass was

due to lowered concentrations of CO2 in the atmosphere, which would have caused the less efficient C3 plants to diminish and the more efficient C4 plants (mainly tropical

grasses) to flourish (Ehleringer, Sage, Flanagan, and Pearcy, 1991; Cerling et al., 1993;

Cerling et al., 1997b). This theory has been countered by Morgan, Kingston and Marino

(1994), who claim that there was no expansion of the C4 biomass and no connection

between the atmosphere and any changes in the C3/ C4 biomass. Rather, Morgan and

colleagues suggest that the apparent increase in C4 consumption by mammalian herbivores is the result of faunal immigration or in situ speciation, not a change in the floral populations. Also, the proposed decrease in atmospheric CO2 concentrations has been challenged due to a lack of direct evidence of a change in the partial pressure of

CO2 (Pagani et al., 1999).

The isotopic evidence modifies the grassland story in two ways: (1) It pushes the

dominance of C4 grasslands back as early as 7 Ma and, more importantly, (2) Denies a

significant C4 presence before 7 Ma, casting doubt on the idea that grassy savannas

preceded the steppe environment in the Great Plains. Initially, the isotopic and faunal

evidence provided an interesting dilemma. The faunal morphological changes suggested

a strong presence of grass, presumably C4, in the Great Plains at 15.8 Ma (Webb, 1983) or 17.5 Ma (Theodor et al., 1998). However, a major global increase in C4 biomass is not

documented in the paleodiets until 7 Ma (Cerling et al., 1993). This indicates an apparent

change in faunal morphology that began 8.8 to 10.5 million years before the ungulates

began eating C4 grasses, and therefore long before the dominance of the grasses believed

to have caused the adaptations (Fox and Koch, 2003).

More recent isotopic evidence has addressed this discrepancy by pushing the

appearance of C4 grasses back prior to the original 15.8 Ma age based on faunal

morphology. Fox and Koch (2003) looked at paleosol stable isotopes from the Great

Plains region of North America and suggested that the C4 biomass first appeared no later

than the early Miocene (about 23 Ma), but did not become dominant until the late

Miocene (as evident from the biomass shift seen at 7 Ma). Based on the paleosol and

faunal data, Fox and Koch (2003) suggested that typical Great Plains habitats in the late

Miocene consisted of C3 trees and shrubs over a light carpet of grass.

The apparent temporal gap between grazer morphology and C4 dominance could

then be the result of the expansion of cool climate C3 grasses (Wang et al., 1994; Fox and

Koch, 2003). Expansive C3 grasses in the middle Miocene could have been responsible

for the ungulate grazing adaptations, since C4 grasses were still too relatively low in

abundance to cause such expansive adaptive radiations (Fox and Koch, 2003). The idea

of C3 grasses driving hypsodonty radiations denotes a paradigm shift in how paleodiets

13 are assessed through stable isotopes. Initially, δ C values that reflected C3 diets were

typically ascribed to browse material, such as shrubbery and leaves. C3-consumers were

then categorized as “browsers” while C4 consumers were categorized as “grazers.” This

was due, in large part, to the assumption that paleoenvironments are analogous to modern

environments. This assumption has recently been reevaluated. Wang et al (1994) first suggested that a unique grassland environment composed of low-latitude C3 grasses

could have existed in the middle Miocene due to lower concentrations of atmospheric

CO2. Janis, Damuth, and Theodor (2002) concluded that early Miocene ungulates are

unlike any ungulates from modern grassland and forest environments and, therefore, represent a paleoenvironment unlike any seen today. This interpretation has been supported by Fox and Koch (2003), who proposed the hypothesis of expansive C3 grasses

in woodland environments. Woodland savannas with C3 grasses would then constitute a

unique environment with no modern analogues.

In order to further assess this claim, and to evaluate the spread of C4 grasses across

North America, this study focuses on the middle Miocene to early Pleistocene (15–5 Ma)

of northern/central Florida and southern Georgia and seeks the presence of grasses in the

diets of ungulates. Of particular interest is the Willacoochee Creek Fauna, an early

Barstovian (~15 Ma) faunal assemblage that is mostly composed of possible early grazers

or mixed feeders (ungulates bearing mesodont to hypsodont teeth). This faunal

community existed at a pivotal point in the evolutionary history of mammals; when open-

habitat morphologies rapidly expanded. This study utilizes three methods of paleodiet

analysis: stable carbon isotope analysis, mesowear analysis, and microwear analysis. The

combination of stable isotopes and mesowear or microwear makes it possible to either

substantiate or refute the presence of C4 grasses in the paleodiets of middle Miocene

ungulates, as well as address alternative driving mechanisms for the adaptation of

hypsodonty. The objectives of this study are (1) Characterize the dietary habits of the

community of herbivores from the Willacoochee Creek Fauna, (2) Determine the

presence of any grasses, C3 or C4, in northern Florida during the middle Miocene, and (3)

Determine the presence of any grasses, C3 or C4, in Florida leading up to the carbon biomass shift at ~7 Ma.

CHAPTER 2 FIELD AREA AND FAUNA

The taxa in this study were excavated from sediments from seven sites in Florida and Georgia. Three of the sites are located in northern Gadsen County, Florida: the

Englehard La Camelia Mine, the Milwhite Gunn Farm Mine, and the La Camelia 2 Mine

(Figure 2-1). Sediment samples were collected from the Englehard La Camelia and

Milwhite Gunn Mines in the late 1980's by field crews from the Florida Museum of

Natural History (FLMNH) and the University of Florida Department of Geological

Sciences (Bryant, 1991). Vertebrate fossils were also collected from these sites. Most of the fossils were recovered from spoil piles, but these piles were positively associated with the Dogtown Member of the Torreya Formation (Bryant, 1991). These fossils comprise the Willacoochee Creek Fauna (WCF) assemblage, an early Barstovian (middle Miocene) assemblage of mammals, birds, amphibians, and reptiles found in the Dogtown Member

(Bryant, 1991). The early Barstovian designation of the assemblage is based on the presence of Copemys, Perognathus, Rakomeryx, and Ticholeptus, as well as the overlapping age ranges of several other mammals (Bryant, 1991; Table 2-1). The beginning of the Barstovian is defined by the appearance of Copemys and the early

Barstovian is characterized by the appearance of Perognathus, Rakomeryx, and

Ticholeptus (Tedford et al., 1987). The WCF bears mammals with known age ranges that begin in the early Barstovian, restricting the age of the WCF to an upper limit of early

Barstovian (Bryant 1991). The overlapping age ranges of other mammals, such as those

17 18 of the Merychippine horses, support this age correlation. The absolute age of the early

Barstovian is between about 16.6 to 14.4 Ma (Tedford et al., 1987).

The Torreya Formation is part of the Hawthorn Group and it is the only part of that group that is present in the eastern Florida panhandle (Scott, 1988; Huddlestun, 1988).

The formation is siliciclastic with scattered carbonate and phosphate deposits and can be found throughout the eastern Florida panhandle and into southern Georgia (Bryant,

1991). Bryant (1991) described the Dogtown Member of the Torreya Formation as

"largely clay, with varying amounts of sand and dolomite, and is primarily present in

Gadsen County, Florida, and adjacent Decatur County, Georgia" (Figure 2-2). The

Englehard La Camelia Mine was designated as the type section of the Dogtown Member

(Bryant, 1991). Vertebrate fossils are found in the sand and sand/clay layers (Figure 2-3).

As mentioned earlier, the Milwhite Gunn Farm Mine also belongs to the Dogtown

Member, but it represents a unique lithology. It consists of well-indurated, weathering- resistant, carbonate-cemented sandstone (Bryant, 1991). Bryant (1991) suggested that the outcrop at the Milwhite Gunn Farm Mine represented a "subaerial exposure surface, but no pedogenic horizonation is preserved." The Milwhite Gunn Farm Mine and Englehard

La Camelia Mine layers are contemporaneous, based on the in situ presence of the rodents Copemys and Perognathus at both sites (Bryant, 1991).

There are two new sites, the La Camelia 2 and Crescent Lake Mines. The La

Camelia 2 Mine is located in northern Gadsen County, Florida, near the original

Englehard La Camelia Mine site. Over 2,000 pounds of sediment, as well as some vertebrate fossils, were collected from La Camelia Mine 2 by an FLMNH/UF Geology

Department field crew in May, 2004. Like the previous sites, the sediment samples and

fossils were collected from spoil piles at this site. These piles, however, were positively

associated with a single unit at the site. This unit consists of interfingering sand and clay

layers and has been correlated with the Dogtown Member and aged as early Barstovian.

The unit correlation is based on lithology and the Barstovian age, which is derived from the overlapping presences of Aphelops sp., which first appears in the Late

Hemingfordian, and Acritohippus isonesus (Figure 2-2).

The second new site, the Crescent Lake Mine, is located 15 miles north of the

Florida-Georgia Border in Decatur County. Vertebrate fossils and sediment were collected from this site by FLMNH field crews in May, 2005. Unlike the other sites, these samples were collected in situ along a fresh cut at the mine (Figure 2-4). The sediments at the Crescent Lake Mine constitute a clayey sand layer representing terrestrial and marine environments. Marine depositional environments are evident by the presence of abundant invertebrate fossils, such as gastropods, bivalves, and echinoids (sea urchins), as well as garfish scales. Based on the lithology of the unit and mammal assemblage, this site correlates with the Dogtown Member of the Torreya Formation and an early Barstovian age. The age is based on the presence of the brachydont Anchitherium clarencei and the mesodont Merychippus primus, two horses with biochronological ranges in the WCF that extended, and terminated, in the early Barstovian (Bryant, 1991; Figure 2-2).

Additionally, the hypsodont horse Acritohippus isonesus and the rhino Aphelops have overlapping ranges in the Barstovian (Bryant, 1991).

The two sites discovered in the 1980's (Englehard La Camelia and Milwhite Gunn

Mines) have both been analyzed for strontium-isotopic and paleomagnetic dating. Four reliable 87Sr/86Sr age estimates place the Dogtown Member between 16.6 ± 1.0 and 14.7

± 1.5 Ma (Bryan, MacFadden, and Mueller, 1992). The Dogtown Member is entirely of

reversed polarity, and, with biochronologic data, the unit correlates with Chron C5B-R

(Bryant et al., 1992). This correlation narrows the age of the Dogtown Member, and therefore the WCF, to between 16.2 and 15.3 Ma (Bryant et al., 1992). This range has

been restricted further, to between 15.9 and 15.3 Ma, based on the Hemingfordian-

Barstovian boundary (Woodburne, Tedford, and Swisher, 1990). Bryant et al. (1992)

also determined that the Dogtown Member is time-transgressive, with sediments getting

younger to the north.

In addition to the middle Miocene sites, this study looks at younger sediments to

trace the presence of grasses in Florida. The three younger sites are all located in central

Florida. Ages for two of these sites are determined by biochronology. The Love Bone

Bed Site is late Miocene and designated as late Clarendonian in age (~9.5 Ma) and the

Upper Bone Valley Formation is early Pliocene and late Hemphillian (~4.5 Ma) (Figure

2-5; Hulbert, 1992; Morgan, 1994). The accuracy of these ages is within about ± 0.5 Ma

(MacFadden and Cerling, 1996). The third site is the Leisey Shell Pit, which has been

dated at approximately 1.5 Ma, based on biochronological, 87Sr/86Sr, and paleomagnetic

data (Webb et al., 1989). The accuracy of this age is about ± 0.1 Ma and places the

Liesey Shell Pit in the early Pleistocene with an early Irvingtonian age (Figure 2-5).

These three sites were chosen for this study because stable carbon isotope analyses have

been previously conducted on large populations of fossil horses from these sites

(MacFadden and Cerling, 1996). These large populations are also ideal for mesowear

analyses.

The specimens analyzed in this study span approximately 14 million years.

Specimens from the Englehard La Camelia, Milwhite Gunn Farm, and La Camelia 2

Mines are representatives of the WCF from the Dogtown Member of the Torreya

Formation and are of early Barstovian (about 15.9 to 15.3 Ma) age (Bryant, 1991).

Before any analyses could be conducted, it was necessary to identify all of the

specimens. Despite the previous description of the WCF by Bryant (1991), many of the

collected fossils were still unidentified and not catalogued. Additionally, all of the fossils

from the new site needed to be identified and catalogued. These fossils, which are all

teeth, were identified by referencing published literature as well as identified specimens

from the FLMNH collections. The vast majority of unidentified specimens were horse

teeth. In addition to identifying the taxon, the tooth position was determined for each

tooth by comparing the widths of the parastyle and mesostyle and assessing the angle of

the mesostyle and plane of occlusal surface (Figure 2-6). Typically, the parastyle is wider

than the mesostyle in horse molars and vice versa for horse premolars (Bode, 1931).

More posterior-leaning mesostyle angles and shallower angles of the occlusal plane are

indicative of molars, while more anteriorly-leaning mesostyles and more pronounced

angles of the occlusal plane are apparent in premolars (Bode, 1931). Identification of the

tooth position is important in this study since paleodiet analyses often use specific teeth

as a standard for paleodiet assessment.

Figure 2-1. Index map of middle Miocene localities in northern Florida. 1 = Milwhite Gunn Farm Mine, 2 = Englehard La Camelia Mine and La Camelia 2 Mine. All of the Willacoochee Creek Fauna taxa were collected from the Dogtown Member of the Torreya Formation. (Modified with permission from Bryant, 1991)

Table 2-1. Biochronological ranges of the Willacoochee Creek Fauna. Solid lines indicate known ranges and asterisks indicate range extensions. Note that all of the ranges overlap in the early Barstovian. Also, Ticholeptus hypsodus, Bouromeryx cf. parvus, and Rakomeryx sp. all occur only in the early Barstovian. These ranges denote an early Barstovian age for the Dogtown Member of the Torreya Formation. (Modified with permission from Bryant, 1991) TAXON HEMINGFORDIAN BARSTOVIAN CLARENDONIAN EARLY LATE EARLY LATE EARLY LATE Lanthanotherium sp. ********* Mylagaulus sp. cf. Protospermophilus sp. Perognathus cf. minutus Proheteromys sp. Copemys sp. "Cynorca " cf. proterva Ticholeptus hypsodus Bouromeryx cf. parvus Rakomeryx sp. Anchitherium clarencei ********* Merychippus gunteri Merychippus primus ********* Acritohippus isonesus Aphelops sp. KNOWN BIOCHRONOLOGICAL RANGE ********* BIOCHRONOLOGICAL RANGE EXTENSION

Figure 2-2. Composite section of the Torreya Formation and the location of the Willacoochee Creek Fauna within the Dogtown Member. Note that the Dogtown member is composed mostly of clayey sand and some clay. (Modified with permission from Bryant, 1991)

Figure 2-3. Correlated stratigraphic sections of the Englehard La Camelia and Milwhite Gunn Farm Mines. Note that the Willacoochee Creek Fauna are found in layers composed of clayey sand, sand, and sand with carbonate cement. (Modified with permission from Bryant, 1991)

Figure 2-4. Fresh cut at the newest site, the Crescent Lake Mine, in Decatur County, Georgia. A) Cross-section view of sand and clay layers above the fossiliferous layer. B) Bottom of the cut, where the fossiliferous clayey sand is exposed. In the foreground is where terrestrial vertebrate fossils were found, while marine invertebrates were found elsewhere on the cut.

Figure 2-5. Stratigraphic and temporal distribution of the 3 fossil sites studied in the mesowear analysis, as well as a fourth (Moss Acres) discussed later. Neohipparion trampasense, Cormohipparion plicatile, and Cormohipparion ingenuum are from the Love Bone Bed, Nannippus aztecus is from the Upper Bone Valley Formation, and Equus “leidyi” is from the Leisey Shell Pits. (Modified with permission from MacFadden and Cerling, 1996)

Figure 2-6. Tooth positions. A) The modern horse jaw with the angles of the mesostyles and occlusal surfaces for each tooth position. Abbreviations: pm2 = upper 2nd premolar, pm3 = upper 3rd premolar, pm4 = upper 4th premolar, m1 = upper 1st 2 nd 3 rd st molar, m = upper 2 molar, m = upper 3 molar, pm1 = lower 1 premolar, nd rd st pm2 = lower 2 premolar, pm3 = lower 3 premolar, m1 = lower 1 molar, m2 nd rd th = lower 2 molar, m3 = lower 3 molar. B) Buccal view of upper 4 premolar (a) and upper 1st molar (b). Note the differences in parastyle and mesostyle widths and the differences in angles of the plane of occlusal surface. (Modified with permission from Bode, 1931)

CHAPTER 3 MESOWEAR METHOD

The mesowear method was devised by Fortelius and Solounias (2000) as a quick and inexpensive process of determining the lifelong diet of a taxon. They determined that, for extant mammals, broad conclusions on the diets, such as a grazer or browser classification, can be deduced from the shape of the buccal cusps (either the paracone or metacone) and the relative difference in height between the tip of the cusps and the intercusp valley. The robustness of this method was confirmed by blind test studies that revealed that there are no statistical differences in the scoring of attributes between individual researchers (Kaiser et al., 2000).

The attributes that are evaluated, termed cusp shape and occlusal relief, were originally only applied to the upper second molar (M2) and require at least 20 specimens to obtain a reliable classification. Fortelius and Solounias (2000) used cusp shape and occlusal relief to establish a "typical" set of 27 extant browsers, grazers, and mixed feeders. This set was illustrated in a hierarchical cluster analysis (Figure 3-1), which shows clusters of mammals that had similar wear (attrition- or abrasion-dominated). The classification of each cluster (“e.g.”, “grazer”) was confirmed by direct observations made on the diets of the animals. The mesowear technique can be extended to an extinct taxon, which can then be included in the hierarchical cluster analysis to determine its dietary classification. Mesowear analyses require large sample populations (>20), which can be problematic for some localities, but the method yields an accurate depiction of an animal’s average lifelong diet.

29 30

In a study on hipparionine and extant equids, the mesowear method was “extended”

to broaden its application beyond the upper M2, to specific combinations of the upper

third premolar (P3), upper fourth premolar (P4), upper first molar (M1), and the upper

third molar (M3) (Kaiser and Solounias, 2003). This study was conducted with the goal

of expanding the application of the mesowear method beyond ungulate populations with

an abundance of M2s. The “extended” mesowear method, the combination of tooth

positions that yielded results most consistent with the M2 results, was that of the

P4+M1+M2+M3 teeth.

However, the “extended” method still requires positive identification of the tooth

positions. While this is not problematic when studying associated teeth, it can be difficult

to identify isolated teeth. This is especially true for equids. Bode (1931) showed that it

was possible to identify tooth position of unassociated teeth through the anterior-posterior tilt of each tooth, but it has been noted that this method is time-consuming and does not always result in a positive identification (Hulbert, 1987). This problem is often encountered when trying to distinguish between a P3 and P4. Kaiser and Solounias

(2003) showed that the P3 in hipparionine equids was an unreliable indicator of paleodiet,

causing a shift towards a grazer classification. Additionally, they showed that the P4 is a

reliable indicator, particularly when grouped with molars (as in the “extended”

combination of P4-M3). To ensure correct dietary classifications based on unassociated

teeth, this study will apply the mesowear method only to molars, avoiding dubious

premolars. Previous studies have shown that analyses based on M1s and M2s yield the

same classifications as analyses based on only the M2s (Kaiser and Solounias, 2003;

Hoffman, unpublished).

Materials and Methods

Mesowear analyses were conducted on five equids from the younger central Florida sites:

Neohipparion trampasense, Cormohipparion plicatile, and Cormohipparion ingenuum

from the Love Bone Bed Site (~9.5 Ma); Nannippus aztecus from the Upper Bone Valley

Formation (~4.5 Ma), and Equus “leidyi” from the Leisey Shell Pit (~1.5 Ma) (Table 3-

1). Based on dental and post-cranial morphology, Neohipparion trampasense is a

probable grazer to mixed feeder while the Cormohipparion species are both expected to be mixed feeders (MacFadden and Cerling, 1996). The taxon analyzed from the Upper

Bone Valley Formation is Nannippus aztecus, an expected grazer. Finally, Equus

“leidyi,” an extremely hypsodont grazer from the Leisey Shell pit, was also analyzed.

Following the techniques described in Fortelius and Solounias (2000), the cheek teeth of six different taxa were assessed for mesowear analysis, in terms of cusp shape and occlusal relief (Figure 3-2). The cusp shape rating (sharp, round, or blunt) describes the shape of the apex of the sharper cusp. The occlusal relief describes the height of the cusps (high or low) relative to the valley between them. This rating can be quantified by drawing a line that connects the apices of the two cusps, then measuring the vertical distance between that line and the center of cusp valley. This value is then divided by the length of the whole tooth. For equids, values greater than 0.1 signify high occlusal relief and values below 0.1 signify low occlusal relief (Fortelius and Solounias, 2000). To avoid teeth that were in the early or late stages of wear at time of deposition, only teeth that are 25% to 75% of the maximum crown height were analyzed. Waterworn teeth were also excluded from the sample set. When necessary, a magnifying lens was used in rating the attributes.

As stated in Fortelius and Solounias (2000), the mesowear signal becomes stable

when there are more than 20 teeth, although a reasonable approximation is attained at

about 10 samples. In this study, at least 20 samples were measured for each taxon to

ensure correct classifications. The teeth that were analyzed include positively identified

M1s and M2s, as well as molars that could be either an M1 or an M2. Once the

measurements were recorded, the percentages for each attribute were calculated. Next,

these percentages were examined by hierarchical cluster analyses, using SPSS v.11.5

software, to determine the dietary classification of each taxon. The variables entered into

the hierarchical cluster analyses are percentage high occlusal relief, percentage sharp

cusp shape, and percentage blunt cusp shape. Complete linkage and normalized

Euclidean distance were used in the analysis, following Fortelius and Solounias (2000).

Each taxon was entered into a separate cluster analysis, to avoid altering the morphology

of the dendrogram and possibly yielding biased clusters. The cluster placement of each

fossil taxon was used to determine its dietary classification. For example, a taxon that is

located within the cluster of ‘typical’ extant grazers would be classified as a grazer.

Results

The cusp shape and occlusal relief ratings for each sample are listed in the

Appendix A. The calculated percentages of the attribute ratings for each taxon are listed

in Table 3-2. In the hierarchical cluster analyses, all five horses (Neohipparion

trampasense, Cormohipparion plicatile, Cormohipparion ingenuum, Nannippus aztecus,

and Equus “leidyi”) are placed within the cluster of confirmed extant grazers (Table 3-3,

Figure 3-3). N. trampasense clusters closest to Ceratotherium simum (White rhinoceros),

while C. plicatile is closest to Damaliscus lunatus (topi) and C. ingennum is placed

closest to the subcluster of Alcelaphus buselaphus (hartebeest) and Connochaetes

33 taurinus (wildebeest). Conversely, N. aztecus forms its own subcluster with Alcelaphus buselaphus, placing closer than Connochaetes taurinus. Finally, E. “lediyi” is placed closest to Bison bison (American plains bison). All of these extant herbivores are typical hypsodont grazers.

Figure 3-1. Dendrogram illustrating the hierarchical cluster analysis of the 27 ‘typical’ grazers, browsers, and mixed feeders from Fortelius and Solounias (2000). Browser abbreviations are in upper case, grazer abbreviations are in lower case, and mixed feeder abbreviations are upper and lower case. Abbreviations: AA= Alces alces, RS= Rhinoceros sondaicus, DB= Diceros bicornis, OV= Odocoileus virginianus, OJ= Okapia johnstoni, GC= Giraffa camelopardalis, OH= Odocoileus hemionus, DS= Dicerorhinus sumatrensis, Gg= Gazella granti, Gt= Gazella thomsoni, Om= Ovibos moschatus, To= Taurotragus oryx, Ts= Tragelaphus scriptus, Cc= Cervus canadensis, Cs= Capricornis sumatraensis, Me= Aepyceros melampus, ab= Alcelaphus buselaphus, ct= Connochaetes taurinus, he= Hippotragus equinus, rr= Redunca redunca, ke= Kobus ellipsipyrmnus, hn= Hippotragus niger, eb= Equus burchelli, eg= Equus grevyi, dl= Damaliscus lunatus, cs= Ceratotherium simum, bb= Bison bison.

Table 3-1. Systematic and morphological information about the taxa used in the mesowear analysis. Abbreviations: MSCH = Maximum crown height measured from the occlusal surface to the base of the crown along the mesostyle. (N. trampasense tribe, age, and MSCH data from Hulbert, 1987; C. plicatile and C. ingenuum tribe, age, an MSCH data from Hulbert, 1988; N. aztecus tribe, age and MSCH data from Hulbert, 1988 and Hulbert, 1990; Equus “leidyi” tribe, age, and MSCH data from Hulbert, 1995).

Molar Tooth Taxon Tribe Site and Age Level MSCH Morphology Neohipparion trampasense Hipparionini Love Bone Bed, ~9.5 Ma 60 mm Hypsodont Cormohipparion plicatile Hipparionini Love Bone Bed, ~9.5 Ma 58 mm Hypsodont Moderately Cormohipparion ingenuum Hipparionini Love Bone Bed, ~9.5 Ma 49 mm hypsodont Moderately Nannippus aztecus Hipparionini Upper Bone Valley Fm., ~4.5 Ma 51 mm hypsodont Equus "leidyi" Equinni Liesey Shell Pit, ~1.5 Ma 91.3 mm Extremely hypsodont

Figure 3-2. Examples of typical mesowear attributes. The three types of cusp shape are sharp, round, and blunt. The types of occlusal relief are high and low. The dotted lines indicate the height of the occlusal relief, from the bottom of intercusp valley to the apex of the highest cusp. (Modiefied with permission from Kaiser and Fortelius, 2003)

Table 3-2. Observed percentages of mesowear attributes (% high and % low refer to the percentage of specimens with high or low occlusal relief and % sharp, % round, and % blunt refer to the percentage of specimens with sharp, round or blunt cusp shape).

% % % % % Taxon N high low sharp round blunt

Neohipparion trampasense 49 24.5 75.5 2 57.2 40.8

Cormohipparion plicatile 45 35.6 64.4 8.9 60 31.1

Cormohipparion ingenuum 26 50 50 0 84.6 15.4

Nannippus aztecus 35 54.3 45.7 0 82.9 17.1

Equus "leidyi" 30 0 100 6.7 20 73.3

Table 3-3. Dietary classifications based on hierarchical cluster analyses of mesowear attributes from Table 3-2.

Taxon Site Age Classification

Neohipparion trampasense Love Site 9.5 ± 0.5 Ma Grazer

Cormohipparion plicatile Love Site 9.5 ± 0.5 Ma Grazer

Cormohipparion ingenuum Love Site 9.5 ± 0.5 Ma Grazer

Nannippus aztecus Upper Bone Valley Fm. 4.5 ± 0.5 Ma Grazer

Equus "leidyi" Leisey Shell Pit 1.5 ± 0.1 Ma Grazer

Figure 3-3. Dendrograms illustrating the hierarchical cluster analyses of 6 studied taxa amongst the 27 ‘typical’ grazers, browsers, and mixed feeders from Fortelius and Solounias (2000). A) Neohipparion trampasense (Nt). B) Cormohipparion plicatile (Cp). C) Cormohipparion ingenuum (Ci). D) Nannippus aztecus (Na). E) Equus “leidyi” (El). Note that all five equids place within the “Grazers” cluster. Browser abbreviations are in upper case, grazer abbreviations are in lower case, and mixed feeder abbreviations are in both upper and lower case. Abbreviations: AA= Alces alces, RS= Rhinoceros sondaicus, DB= Diceros bicornis, OV= Odocoileus virginianus, OJ= Okapia johnstoni, GC= Giraffa camelopardalis, OH= Odocoileus hemionus, DS= Dicerorhinus sumatrensis, Gg= Gazella granti, Gt= Gazella thomsoni, Om= Ovibos moschatus, To= Taurotragus oryx, Ts= Tragelaphus scriptus, Cc= Cervus canadensis, Cs= Capricornis sumatraensis, Me= Aepyceros melampus, ab= Alcelaphus buselaphus, ct= Connochaetes taurinus, he= Hippotragus equinus, rr= Redunca redunca, ke= Kobus ellipsipyrmnus, hn= Hippotragus niger, eb= Equus burchelli, eg= Equus grevyi, dl= Damaliscus lunatus, cs= Ceratotherium simum, bb= Bison bison.

CHAPTER 4 MICROWEAR METHOD

The microwear method is a dietary analysis that quantifies the microscopic wear on herbivore teeth. The consumption of herbivorous diets leaves microscopic scratches and

pits on the tooth enamel known as microwear. Studies on microwear of extant herbivores

with known diets have revealed a correlation between microwear and diet (Teaford and

Walker, 1984; Grine, 1986; Teaford, 1988; Solounias and Hayek, 1993) One noticeable

trend is that grazers typically have more scratches than browsers while browsers bear

more pits than grazers (Solounias and Semprebon, 2002). These observations have led to

the establishment of a Microwear Index (MI) (MacFadden, Solounias, and Cerling,

1999). For a stated area of enamel (“e.g.”, 0.5mm x 0.5mm), the MI is calculated as the

total number of scratches divided by the total number of pits. As a standard, an MI below

1.5 indicates a browsing diet and an MI above 1.5 indicates a grazing diet (MacFadden et

al., 1999). Additional conclusions can be drawn by comparing the analyzed tooth to a

database of the microwear of extant animals with observed known diets. Plots of the

number of pits versus the number of scratches for these extant herbivores reveal distinct

morphospaces indicative of browsing, grazing, and mixed diets (Solounias and

Semprebon, 2002).

Microwear studies were initially conducted at high magnification using Scanning

Electron Microscopy (SEM) imagery (Grine, 1986). Recently, however, a low-

magnification technique was developed to reduce the time and financial costs of the

microwear method. Analyzing microwear at only 35x magnification has been shown to

40 41

reveal the same results as high magnification (Solounias and Semprebon, 2002).

Solounias and Semprebon (2002) also created four additional quantifying characters to

attain a more detailed understanding of dietary habits. These characters can be used in

hierarchical cluster analyses to determine which kind of extant herbivores most closely

resemble the fossil taxa in terms of microwear, and therefore diet. However, microwear is

subject to the “Last Supper Effect,” the limitation of reflecting an individual’s last few meals, rather than the life-history of diet (Grine, 1986). Therefore, microwear analyses are based on the assumption that the animal’s typical diet is reflected by these final meals.

Materials and Methods

Microwear analyses were conducted on three samples of the most hypsodont taxon from the WCF assemblage, Acritohippus isonesus. Microwear analyses were conducted instead of mesowear because of the small sample size. The standard tooth position for microwear analysis is the second molar (upper or lower) (Solounias and Semprebon,

2002; Rivals and Deniaux, 2003). No other horses, such as Merychippus primus or

Merychippus gunteri, from the WCF were analyzed for microwear because there are not enough M2s for a statistical analysis. Due to the very limited sample size of the WCF, only one tooth of A. isonesus (part of an associated dentition) was positively identified as a second molar (M2). The other two samples were identified as first or second molars

(M1/2s). Based on the curve of the mesostyle, it appears likely that both of these teeth are second molars, but it cannot be stated with certainty. Statistical analyses, such as the unpaired Students t-test and nonparametric Mann-Whitney test, were conducted on the collected results to determine whether there exist significant differences between the

M1/2s and the one positively identified M2. Microwear analyses were also conducted on

specimens of the hypsodont horse, Neohipparion trampasense, from the Love Bone Bed

site (~9.5 Ma) of north central Florida. These microwear results were used for comparison to A. isonesus.

This study followed the procedure for microwear analysis outlined in Solounias and

Semprebon (2002). The first step is to clean the occlusal surface of each tooth using 95% alcohol and cotton swabs. A mold is then made of the tooth using high-precision polyvinylsiloxane dental impression material. The mold is removed and discarded in order to remove any remaining debris from the enamel. A second mold is then made.

After the second mold has hardened, it is used to make a clear, high-quality epoxy cast

(using a resin to hardener ratio of 5:1). After 1–2 days, the cast has hardened and is ready for analysis.

The standard for microwear analysis consists of looking at a 0.4 mm x 0.4 mm area of the cast under 35x magnification (Solounias and Semprebon, 2002). Those dimensions

(0.4mm x 0.4mm) were used because the objective crosshairs had 0.4 mm increments at that magnification. It should be noted that the only stereomicroscope with crosshairs available for this study had 50x magnification, not 35x. Additionally, the crosshair increments had larger spacing, causing the search area to be larger, about 0.5mm x 0.5 mm. The increased magnification and search area result in higher numbers of scratches and pits than would be found using the standard magnification and search field. While this has no effect on the calculated ratio of scratches to pits or on the percentages of other quantitative categories, the increase in average pit and scratch numbers prevents the use of hierarchical cluster analyses to assess dietary subcategories. However, these

subcategories are not essential for this study, which is only concerned with whether the

taxa were eating grass or browse.

To observe the micowear features, a light source is shone through the cast at a

shallow angle to the occlusal surface. The casts are analyzed in a dark room and

adjustments to the intensity and angle of the light source are often made to best observe

the microwear features. Two 0.5 mm x 0.5 mm areas were assessed for each tooth.

Following Solounias and Semprebon (2002), these areas were located along the shear

facet of the second enamel band on the paracone of each upper tooth. For lower teeth, the

search areas were restricted to the shear facet on the protocone. Pits are defined as

microscopic defects in the enamel that have a length:width ratio < 4, while scratches are

defects that have a ratio ≥ 4 (Grine, 1986; Teaford and Robinson, 1987). These categories

are obvious under low magnification. Pits usually appear as fairly round dots with a

length:width ratio of about 1 or 2, while scratches are always much longer and typically

have length:width ratios much higher than 4. For each tooth, the pit and scratch numbers

from the two search areas are averaged. Pits are further classified as large or small. Small

pits are the most numerous type and are fairly rounded. There are usually a few pits that

are at least twice the diameter of the small pits. These are categorized as large pits and

appear deeper and dark due to less refraction. Scratches are subdivided into fine or

coarse. Fine scratches are the narrowest scratches. They are shallow and often are white

from light reflection. Coarse scratches are often dark and longer than most fine scratches.

It is also noted when a sample bears more than four cross scratches. Cross scratches are

any scratches that run roughly perpendicular to the majority of scratches. Finally, the average microwear index for each taxon is calculated.

Results

The observed counts of microwear features for each specimen are recorded in Table

4-1 and the calculated MI and percentages are presented in Table 4-2. Table 4-1 follows the format and uses some of the quantitative parameters established by Solounias and

Semprebon (2002). The average number of pits and scratches for each taxon is listed, followed by the microwear index (the ratio of scratches to pits), and two of the newer quantitative variables. Since the microscope field used in this study is not the commonly used size, these variables cannot be utilized in cluster analyses to determine more detailed similarities to the diets of extant mammals. However, these variables are helpful in making general inferences. The listed percentages in Table 4-2 for each of these attributes refer to the percentages of specimens that bear that corresponding attribute.

Unpaired Students t-tests and nonparametric Mann-Whitney tests were used to examine the concordance of the three A. isonesus specimens. The MI’s of the two dubiously identified teeth (UF 223080 and UF 223081) were compared to the MI of the positively identified M2 (UF 223063). The Students t-test and Mann-Whitney p-values for comparing UF 223080 UF 223063 are 0.6826 and 0.4386, respectively. The Students t-test and Mann-Whitney p-values for comparing UF 223081 to UF 223063 are 0.2195 and 0.1213, respectively. For both of the M1/2 specimens, the Mann-Whitney test and

Students t-test yield p-values greater than 0.05 and reveal no significant differences between the MI’s. For each sample of both A. isonesus and N. trampasense, the average number of scratches is almost double the average number of pits. The average MI’s for A.

isonesus and N. trampasense are 1.97 and 1.98, respectively. The MI’s of A isonesus and

N trampasense were compared for statistical differences using an unpaired Students t-test

and nonparametric Mann-Whitney test. The p-values for the Students t-test and Mann-

Whitney test are 0.5707 and 0.8690, respectively. These p-values are greater than 0.05

and show that the MI’s of A. isonesus and N. trampasense are not statistically distinct.

Additionally, all of the samples for both taxa had scratches that were predominantly fine.

Each sample also had at least four cross scratches.

Table 4-1. Individual microwear counts for 3 specimens of Acritohippus isonesus and 7 specimens of Neohipparion trampasense. Each specimen has two areas (A and B) assessed for microwear. Abbreviations: UF ID = catalogue number in the Florida Museum of Natural History collections, # x-scratches = number of cross-scratches.

# fine # coarse Taxon UF ID # pits # scratches # x-scratches scratches scratches Acritohippus 223063A 22 86 39 83 3 isonesus 223063B 43 96 20 81 15 223080A 17 58 16 53 5 223080B 37 57 10 53 4 223081A 83 118 27 110 8 223081B 65 111 40 106 5

Neohipparion 27991A 29 70 15 62 8 trampasense 27991B 42 48 13 45 3 27992A 45 59 26 55 6 27992B 22 57 17 52 5 27993A 11 34 18 28 6 27993B 23 39 5 35 4 32280A 68 103 31 98 5 32280B 42 76 16 67 9 32291A 25 49 9 48 1 32291B 24 85 25 79 6 36287A 29 58 26 53 5 36287B 30 69 25 58 11 32114A 37 103 17 99 4 32114B 32 60 18 54 6

Table 4-2. Calculated average microwear values and percentages for Acritohippus isonesus and Neohipparion trampasense. Abbreviations: N = sample size, Avg. # pits = the average number of pits for the sample population, Avg. # scratches = the average number of scratches for the sample population, MI = microwear index, % Fine scratches = the percentage of the sample population that possess scratches that are predominantly fine, % coarse scratches = the percentage of the sample population that possess scratches that are predominantly coarse, % cross scratches = the percentage of the sample population that exhibits at least 4 cross scratches.

Avg. Avg. % Fine % Coarse % Cross Taxon N # pits # scratches MI scratches scratches scratches A. isonesus 3 44.5 87.67 1.97 100 0 100 N. trampasense 7 32.79 65 1.98 100 0 100

Table 4-3. Isotopic data for equids from 4 central Florida sites. Abbreviations: UF ID = catalogue number in Florida Museum of Natural History collections. (Modified from MacFadden and Cerling, 1996)

Taxon Site UF ID Material δ13C(‰)

9.5 Ma Level, late Clarendonian, late Miocene

Neohipparion trampasense Love Bone Bed 32230 R. P3-M3 -11.6 Cormohipparion plicatile Love Bone Bed 32265 R. P2 -12.6 Cormohipparion plicatile Love Bone Bed 32265 R. P3 -12.5 Cormohipparion plicatile Love Bone Bed 32265 R. P4 -11.0 Cormohipparion plicatile Love Bone Bed 32265 R. M1 -12.0 Cormohipparion ingenuum Love Bone Bed 60396 R. P2 -10.8 Cormohipparion ingenuum Love Bone Bed 35979 R. P2 -10.8

7.0 Ma level, "middle" Hemphillian, late Miocene

Nannippus aztecus Moss Acres 69933 R. M3 -7.9 cf. Nannippus Moss Acres None R. p2 -5.3

4.5 Ma level, latest Hemphillian, early Pliocene Nannippus Upper Bone Valley None L. M -6.0 Nannippus aztecus Upper Bone Valley 63633 L. M2 -9.1 Nannippus aztecus Upper Bone Valley 65796 R. M1 -2.4

1.5 Ma level, early Irvingtonian, early Pleistocene Equus "leidyi" Leisey 1A 80047 R. M3 -1.5

Equus "leidyi" Leisey 1A None R. P2 -3.5

CHAPTER 5 STABLE ISOTOPE ANALYSIS

Background

Plants utilize three different types of metabolic pathways to process and use carbon.

These pathways are: the C4 pathway (or Hatch-Slack cycle), the C3 photosynthetic pathway (the Calvin cycle), and the CAM (Crassulacean acid metabolism) pathway

(Bender, 1971). The C4-dicarboxylic acid pathway utilizes CO2 through carboxylation of phosphoenolpyruvate (O’Leary, 1988). The Calvin Cycle, used by C3 plants, uses the enzyme ribulose biphosphate carboxylase to fix CO2 (O’Leary, 1988). The CAM pathway also uses ribulose biphosphate carboxylase to take in CO2, but the process is more similar to what happens in the bundle sheath cells of C4 plants (O’Leary, 1988).

Plants that use the C4 pathway are primarily tropical grasses, but also include some fruits and vegetables (O’Leary, 1988). C3 plants include trees, shrubs, and cool-climate grasses, and dicots (O’Leary, 1988). The CAM plants consist primarily of desert succulents (O’Leary, 1988).

Each plant metabolic pathway results in the isotopic fractionation of carbon (Figure

5-1) as it is taken into the plant cells from CO2 (Bender, 1971). The heavier carbon isotope, 13C, is discriminated against to varying degrees dependent on the pathway used, and the 13C/12C isotopic ratios of plants decreases relative to atmospheric isotopic ratios

(Bender, 1971). The lighter isotope is preferentially used due to the physical and chemical properties associated with its mass (O’Leary, 1988). The carbon isotopic ratio

48 49

(δ13C) is represented as the parts per thousand difference between the sample and a standard, the Peedee Belemnite from South Carolina (Craig, 1957).

13 C3 plants have an average δ C value of -27.1‰ ± 2.0‰ and a range from about

13 -35‰ to -22‰, while C4 plants have an average δ C value of -13.1‰ ± 1.2‰ and have a

more limited range from -14‰ to -10‰ (O’Leary, 1988). CAM plants have a range of

-10‰ to -20‰ that distinguishes them from C3 plants, but not C4 plants (O’ Leary, 1988).

Since CAM plants are known to be dominant only in xeric habitats (Ehleringer et al.,

1991), they are typically not considered in paleodiet analyses of temperate or subtropical paleoenvironments.

When mammalian herbivores consume plants, the carbon isotopes are fractionated once again. By determining the fractionation factor for the mammal in question, the carbon isotopic signature can be used to determine the diet, in terms of C3 and C4 consumption, of the herbivore. Stable carbon isotope analyses were first used in archaeology to determine the paleodiets of ancient human populations (MacFadden and

Cerling, 1996). These studies analyzed the inorganic apatite in bone, which exists primarily as hydroxyapatite, or Ca10(PO4)6(OH)2 (Hillson, 1986; Wang and Cerling,

1994; MacFadden and Cerling, 1996). Trace amounts of carbon are present in this

-2 mineral when it is commonly altered during skeletal formation, with carbonate (CO3) replacing phosphate to yield Ca10(PO4,CO3)6(OH)2 (Hillson, 1986; Newesely, 1989;

McClellan and Kauwenbergh, 1990). Due to the porous nature of bone collagen, the

hydroxyapatite with structural carbonate is extremely susceptible to diagenetic alteration

(Quade et al., 1992). Dental enamel, however, is much more resistant to diagenesis

because it is has very low porosity and is more than 96% inorganic by weight (Quade et

al., 1992; Wang and Cerling, 1994). Additionally, enamel is more than 95%

hydroxyapatite (Hillson, 1986). Multiple studies have found that the biogenic apatite of enamel does not undergo diagenesis in most depositional situations and therefore retains primary isotopic values (Quade et al., 1992; Wang and Cerling, 1994).

Cerling and Harris (1999) determined an enrichment factor of +14.1 ± 0.5‰ between the δ13C of the plants consumed and the bioapatite in the enamel of extant

ruminant mammals. This consistent enrichment establishes distinct ranges of isotope

ratios for ruminants that consume C4 or C3 plants. Grazers (animals that feed primarily

on C4 grasses) fall in a range of 0‰ to +4‰ while browsers (animals that feed off of

leaves from C3 shrubs and trees) fall within a range of -21‰ to -8‰ (Cerling and Harris,

1999). Stable carbon isotope analyses can therefore be conducted on ruminant teeth to determine diet. This method has been extended to extinct ruminants to determine a general classification of feeding habit. To do this, paleodietary analyses have adopted the categories used to describe extant herbivores, established by Hofmann and Stewart

(1972). There are three general categories: concentrate selectors (browsers), bulk and roughage feeders (grazers), and intermediate feeders (mixed feeders that consume both

browse and grass) (Hofmann and Stewart, 1972). These categories have been subdivided

further to detail the complexity of the feeding strategy. These subdivisions, however, are

rarely used in paleodiet analyses because it is difficult to calculate percentages of plant

type for an extinct animal. It is also important to recognize that paleodiet studies that

utilize stable isotope analyses rely on two assumptions: 1) The metabolic pathways of the

plants behaved similarly in the past as they do today (“i.e.”, fractionated to the same

degree), and 2) the metabolic pathways and enrichment factor of ruminant mammals have

been consistent throughout time.

There is another factor that must be corrected when analyzing stable carbon

isotopes in fossils. As stated above, the δ13C values of modern ruminants are derived

from the fractionation of carbon from plant intake, which is fractionated from the fixation

13 from atmospheric CO2. Atmospheric δ C values have decreased from approximately

-6.5‰ to approximately -8.0‰ over the course of the last 200 years, due to the burning

of fossil fuels (Friedli, Lötscher, Oeschger, Siegenthaler, and Stauffer, B, 1986; Marino,

McElroy, Salawitch, and Spaulding, 1992). Due to this change in atmospheric δ13C values, fossil bioapatite values are about 0.5‰ to 1.3‰ more positive than the values of modern mammals (Koch, Hoppe, and Webb, 1998). This adjustment shifts the endmembers on the scale for paleodiet interpretation: a diet strictly consisting of C3

plants would be no more positive than -8.7‰ and a pure C4 diet would be no more

negative than -0.5‰ (Feranec, 2003). Using these endmembers, a browser (less then 10%

13 C4 intake) would have a δ C value less than -7.9‰, a grazer (less than 10% C3 intake)

would have a δ13C value greater than -1.3‰, and a mixed feeder would have a δ13C value between -7.9‰ and -1.3‰ (Feranec, 2003).

In addition to the utility of δ13C values, valuable dietary and environmental

information can also be attained from the δ18O values of structural carbonate and

phosphate in enamel. Several studies have determined that the δ18O of a mammal’s body

water is directly correlated to the δ18O of ingested water from drinking water and food,

such as plants (Luz, Kolodny, and Horowitz, 1984; Luz and Kolodny, 1985; Bryant and

Froelich, 1995; Bryant, Froelich, Showers, and Genna, 1996a; Kohn, 1996). Furthermore,

these δ18O values are recorded in the structural carbonate and phosphate of mammalian

tooth enamel, which mineralizes in isotopic equilibrium with body water (Longinelli,

1984; Luz et al., 1984). This conclusion has been corroborated by the calculation of a

consistent fractionation factor for oxygen between body water and structural carbonate in

modern equids (Bryant, Koch, Froelich, Showers, and Genna, 1996). Therefore, the δ18O values of mammals can be used to determine the δ18O values of their water source.

However, it must first be determined whether that particular mammal obtains most of its

body water from plants or drinking water.

For obligate drinkers, tooth enamel can reflect the δ18O values of meteoric water.

The δ18O values of precipitation are controlled by temperature (Dansgaard, 1964).

Typically, this link causes the δ18O values of meteoric water to be enriched during

periods of warm weather and depleted during periods of cool weather (McCrea, 1950;

Bryant et al., 1996a). The δ18O values of meteoric water can therefore be used to interpret

seasonality. As mentioned previously, the δ18O values of water sources are reflected in

the carbonate and phosphate of mammalian tooth enamel (Longinelli, 1984; Luz et al.,

1984). However, the metabolic rate of the mammal can influence the δ18O values

recorded in the enamel (Bryant and Froelich, 1995; Kohn, 1996; Kohn, Schoeninger, and

Valley, 1998; Zhow and Zheng, 2002). This poses a problem for using small mammals

as environmental indicators, but Bryant and Froelich (1995) note that “because the

proportion of oxygen taken up as liquid water increases while the food requirement

decreases, the proportion of surficial drinking water reflected in [δ18O of body water] will

increase with increasing body size.” The δ18O of surface water will then be reflected in

the enamel phosphate and carbonate of large mammals that weigh more than 1 kg (Bryant

and Froelich, 1995). The use of δ18O values from tooth enamel as environmental indicators has been extended to fossils, where the variation of δ18O values of enamel

phosphate along serially sampled teeth from Miocene horses (Bryant et al., 1996a) and

Holocene bison and sheep (Gadbury, Todd, Jahren, and Amundson, 2000) has been

determined to reflect seasonality. Similar studies have been conducted using the

structural carbonate in tooth enamel (Cerling and Sharp, 1996; Higgins and MacFadden,

2004; MacFadden and Higgins, 2004). Therefore, along the serially sampled tooth of an

obligate drinker, more positive (or enriched) δ18O values will indicate summer and more

negative (or depleted) values will indicate winter (Fricke and O’Neil, 1996; Feranec and

MacFadden, 2000).

For herbivores that are not obligate drinkers, the δ18O of structural carbonate in

fossil teeth reflects the δ18O of the water in consumed leaves (Longinelli, 1984; Bryant,

Froehlich, Showers, and Genna, 1996; Koch, 1998). The δ18O of water in plants is

influenced by temperature, and humidity (Dongman, Nürnberg, Förstel, and Wagener,

1974; Epstein, Thompson, and Yapp, 1977; Sternberg, Mulkey, and Wright, 1989). Due

to this relationship, evapotranspiration causes leaves in forest canopies and dry, open

habitats to have higher δ18O values than ground-level leaves in cool, humid forests

(Förstel, 1978; Kohn, Schoeninger, and Valley, 1996; Cerling, Harris, Ambrose, Leakey,

and Solounias, 1997a). These values can be enriched by more than 20‰ compared to the

δ18O of local precipitation (Förstel, 1978; Kohn et al., 1996). For non-obligate drinkers,

the δ18O of structural carbonate can yield some general information about the mammal’s

habitat (“e.g.”, arid, open areas versus cool forests), but it does not accurately reflect seasonality.

Materials and Methods

The assemblage of eight ungulates analyzed ranges from expected browsers

(brachydont dentitions found in the unidentified artiodactyl, oreodont, one rhinocerotid, and one equid) to expected mixed feeders/grazers (mesodont to hypsodont dentitions found in three equids and one rhinocerotid). Due to the eruption order of teeth and the weaning stage of mammals, the tooth positions most appropriate for a post-nursing diet analysis are the third premolar (P3), fourth premolar (P4), and third molar (M3) (Bryant et al., 1996a; Fricke and O’Neil, 1996; Hoppe, Stover, Pascoe, and Amundson, 2004).

Other teeth were analyzed for taxa that lacked any positively identified P3s, P4s, or M3s.

These taxa include Anchitherium clarencei, Teleoceras, and Aphelops. In all, twenty-four teeth were selected for bulk stable carbon and oxygen isotope analysis of enamel carbonate (Table 5-1). Bulk samples, taken along the entire height of the tooth, yield an average stable isotopic ratio for the time during which the enamel mineralized. Six specimens, two from three different taxa, were also chosen for serial sampling to obtain isotopic ratios with finer resolution. The taxa chosen were: Aphelops sp., the brachydont, long-limbed rhinocerotid expected to be an open-country cursorial browser; Merychippus primus, a hypsodont horse presumed to be a mixed feeder; and Acritohippus isonesus, the most hypsodont taxon from the assemblage that is expected to be a grazing horse.

Serial samples of bioapatite were taken from 6 teeth at 5-millimeter intervals along the length of each tooth. Serial stable isotope analyses were conducted in order to analyze the variation of diet for an individual. Since a tooth mineralizes from the crown to the base over the course of a few months to two years (Hillson, 1986), serial samples reveal a dietary interpretation from a smaller time interval than a bulk sample. However, the isotopic signals are masked due to the slightly non-perpendicular mineralization of tooth

55 enamel (Passey and Cerling, 2002). Despite this effect, the serial samples still yield an isotopic ratio that is averaged over shorter time intervals than a bulk sample and allows some variation to be discerned (MacFadden and Higgins, 2004). For the purposes of this study, it is important to show any C4 dietary influences throughout the mineralization stage of the animal.

The sampling and preparation procedures for stable isotope analysis are comprised of two main components: 1) physical sampling and 2) chemical processing. The physical sampling consists of the actual removal of enamel for analysis and the chemical processing is the procedural treatment of the enamel to remove organic compounds and all other contaminants that will affect the stable isotope signature. The physical sampling begins by identifying any enamel flakes on the tooth that appear pliable. The surface of the flake is cleaned using brushes, and then removed. A drill is then used to remove any dentine remaining on the inside surface of the flake. It is then ground in a clean mortar and pestle, and the enamel powder is placed in a labeled microcentrifuge vial. If there are no flakes apparent, then a part of the tooth must be drilled, preferably where the enamel is thick. The surface is prepared by cleaning it with carbide dental drill bits and brushes.

All the cementum and soil must be removed from the tooth to avoid contaminating the sample. Once the surface is cleaned, a Foredom drill is used at low RPM’s to remove approximately 5 milligrams of pristine enamel. The drill operator must be careful not drill into the dentine, as dentine has a different δ13C value from enamel and will contaminate the sample. Bulk samples were taken along the entire height of the tooth, while serial samples were taken at 5-millimeter increments starting at the base of the

tooth. All the enamel powder is collected on weighing paper and placed into a labeled

microcentrifuge vial.

The first step in the chemical treatment of the enamel samples is to remove all

organic material that might affect the carbon isotope ratio. To do this, 1 mL of 30%

H2O2 is added to each sample. Then, the vials are closed and shaken on a Thermolyne

Type 16700 Mixer in order to thoroughly mix the enamel powder with the H2O2. The samples are then placed, with the lids off, in a reaction cabinet overnight. After approximately 24 hours, some of the samples may still be reacting with the H2O2, indicating the presence of abundant organic residues. The samples are all treated a second time with H2O2 to remove these organic materials.

After the second H2O2 treatment, the samples are centrifuged and the H2O2 is removed. One milliliter of distilled water is added to each vial and then mixed. The samples are centrifuged, and the water is removed using a pipetter. This rinsing step is repeated two more times. After the third rinse, 1 mL of 0.1 M acetic acid is added to the vials to remove carbonates from the samples. The samples are shaken and left in the reaction cabinet overnight. The samples are not allowed to react with the acetic acid for more than 24 hours, since it has been shown that prolonged exposure to acetic acid can affect the carbon isotope signatures (Lee-Thorpe, Sealy, and van der Merwe, 1989;

Vennemann, Hegner, Cliff, and Benz, 2001).

The next day, the samples are centrifuged and the acetic acid is removed. Distilled water is used to rinse the samples three more times. After the rinses, 95% methanol is added to each sample to remove water. Again, the samples are shaken to mix the

methanol with the enamel, and then the methanol is removed. Finally, the samples are

placed in the reaction cabinet to dry overnight.

After the samples have dried, they are analyzed using a VG Prism mass

spectrometer at the Department of Geological Sciences at the University of Florida.

Approximately 1 mg of each sample is placed in a small sample boat and loaded into the

mass spectrometer, accompanied by standards. The bulk samples are standardized to

either the MEme (MacFadden Elephantus maximus enamel, δ13C = -10.43‰) or the

NBS-19 (δ13C = +1.95‰; Coplen, 1996) standard to ensure the precision of the results

and the calibration of the mass spectrometer. All of the serial samples are calibrated to

the NBS-19 standard. All of the measured isotopic values are then calibrated to the

universal V-PDB (Vienna Pee Dee Belemnite). The values are presented in standard

13 18 δ-notation: δ C or δ O = [(Rsample/RV-PDB) – 1] × 1000, where Rsample is the measured

13C/12C or 18O/16O ratio of the sample. The analytical precision of samples run with the

MEme standard is ±0.05 and ±0.07 for δ13C and δ18O, respectively. The analytical

precision of samples run with the NBS-19 standard is ±0.09 and ±0.18 for δ13C and δ18O, respectively. The analytical precision is measured by calculating one standard deviation of the corrected δ13C and δ18O values of the standards.

Results

The bulk δ13C values for all the sampled specimens range from -11.39‰ to -8.30‰

(Table 5-2, Table 5-3, Figure 5-2). The most positive mean bulk δ13C value belongs to

Aphelops sp. and the most negative belongs to Teleoceras sp., although Teleoceras has a

sample size of only one. The bulk δ18O values for all the sampled specimens range from

-2.15‰ to 2.71‰ (Table 5-3, Figure 5-2).

Three taxa were selected for serial sampling: one expected browser with

brachydont dentition (Aphelops sp.) and two expected grazing/mixed feeding horses with

hypsodont dentitions (Merychippus primus and Acritohippus isonesus). Individual serial stable isotope values are presented in Appendix B. The serial stable isotope samples for the first specimen of Aphelops (UF 116827) range from -8.37‰ to -11.72‰ for carbon and from -1.99‰ to -0.31‰ for oxygen. (Table 5-3, Figures 5-3 and 5-4). The second

specimen (UF 104227) has a narrower range of -9.76‰ to -10.73‰ for carbon and a much wider and more positive range of 0.71‰ to 3.46‰ for oxygen (Table 5-3, Figures

5-3 and 5-4). For Merychippus primus, the serial δ13C values range from -9.26‰ to

-10.54‰ (UF 221419) and from -9.39‰ to -9.72‰ (UF 221427) and the serial δ18O

values range from 0.15‰ to 1.99‰ (UF 221419) and from 1.12‰ to 2.83‰ (UF

221427) (Table 5-3, Figures 5-5 and 5-6). Acritohippus isonesus exhibits serial δ13C ranges of -9.92‰ to -10.66‰ (UF 221407) and -10.33‰ to -11.96‰ (UF 217590) as well as serial δ18O ranges of 0.25‰ to 2.56‰ (UF 221407) and 0.62‰ to 2.37‰ (UF

217590) (Table 5-3, Figures 5-7 and 5-8).

Figure 5-1. Schematic of isotopic fractionation between atmospheric carbon and C3 plants, as well as fractionation between C3 plants and ruminant herbivores. 13 13 δ C values are listed in parentheses. The average δ C value for C3 plants is 13 approximately -27.1‰ and the average δ C value for C3-consuming ruminants is approximately -13‰. The average isotopic fractionation between atmospheric carbon and C3 plants is a depletion of about -19.5‰. The isotopic fractionation between plants and ruminants is an enrichment of approximately 14.1‰. (Modified with permission from Koch et al., 1992)

Table 5-1. List of the 24 specimens analyzed for stable carbon and oxygen isotope analysis. Specimens are representatives of the Willacoochee Creek Fauna from the early Barstovian (middle Miocene) Dogtown Member of the Torreya Formation. Abbreviations: UF ID = catalogue number in the Florida Museum of Natural History collections, ELC = Englehard La Camelia Mine, LC2 = La Camelia 2, and MGF = Milwhite Gunn Farm Mine.

UF ID Taxon Crown Height Tooth Site 221429 Teleoceras sp. Brachydont R. I ELC 104227 Aphelops sp. Brachydont Tooth fragment MGF 116827 Aphelops sp. Brachydont R. MX fragment ELC 217565 cf. Aphelops Brachydont L. PX LC2

114723 Anchitherium clarencei Brachydont L. P2 ELC

221402 Anchitherium clarencei Brachydont R. M1 ELC 217590 Acritohippus isonesus Hypsodont L. P4 LC2 217562 Acritohippus isonesus Hypsodont R. M3 LC2

221405 Acritohippus isonesus Hypsodont L. M3 ELC 221407 cf. Acritohippus isonesus Hypsodont R. P3/4 ELC 114721 Merychippus gunteri Mesodont R. M3 ELC

116829 Merychippus gunteri Mesodont L. P3/4 ELC 221416 Merychippus gunteri Mesodont L. P3/4 ELC 221408 Merychippus gunteri Mesodont L. M3 ELC 114976 Merychippus primus Mesodont R. P3/4 ELC 104208 Merychippus primus Mesodont L. P3/4 MGF 221415 Merychippus primus Mesodont R. P3 ELC 221419 Merychippus primus Mesodont L. P4 ELC

221426 Merychippus primus Mesodont R. M3 ELC

221427 Merychippus primus Mesodont R. M3 ELC

221428 Merychippus primus Mesodont R. M3 ELC

221418 Merychippus primus Mesodont R. M3 ELC 116823 Ticholeptus hypsodus Brachydont R. M3 ELC

221434 Artiodactyla Brachydont R. M3 ELC

Table 5-2. Bulk stable carbon and oxygen isotope values for Willacoochee Creek Fauna. Abbreviations: UF ID = catalogue number in Florida Museum of Natural History collections, Unident. = unidentified tooth position, ELC = Englehard La Camelia Mine, MGF = Milwhite Gunn Farm Mine, and LC2 = La Camelia 2 Mine.

UF ID Taxon δ13C (‰) δ18O (‰) Tooth Site 221429 Teleoceras sp. -11.39 -0.41 R. I ELC 104227 Aphelops sp. -9.48 1.01 Unident. MGF 116827 Aphelops sp. -8.72 -2.15 R. MX ELC 217565 cf. Aphelops -10.23 0.38 L. PX LC2

114723 Anchitherium clarencei -8.90 2.32 L. P2 ELC

221402 Anchitherium clarencei -10.23 1.45 R. M1 ELC 217590 Acritohippus isonesus -10.87 2.03 L. P4 LC2 217562 Acritohippus isonesus -11.00 -0.35 R. M3 LC2

221405 Acritohippus isonesus -9.67 1.95 L. M3 ELC 221407 cf. Acritohippus isonesus -10.16 1.02 R. P3/4 ELC 114721 Merychippus gunteri -11.09 -0.65 R. M3 ELC

116829 Merychippus gunteri -8.30 -0.16 L. P3/4 ELC 221416 Merychippus gunteri -9.72 1.86 L. P3/4 ELC 221408 Merychippus gunteri -11.19 0.67 L. M3 ELC 114976 Merychippus primus -8.78 1.36 R. P3/4 ELC 104208 Merychippus primus -9.54 -0.82 L. P3/4 MGF 221415 Merychippus primus -8.79 2.10 R. P3 ELC 221419 Merychippus primus -10.00 0.85 L. P4 ELC

221426 Merychippus primus -9.76 1.14 R. M3 ELC

221427 Merychippus primus -8.70 1.03 R. M3 ELC

221428 Merychippus primus -9.73 0.45 R. M3 ELC

221418 Merychippus primus -9.50 0.97 R. M3 ELC 116823 Ticholeptus hypsodus -9.83 0.33 R. M3 ELC

221434 Artiodactyla -10.87 0.66 R. M3 ELC

Table 5-3. Descriptive statistics of the δ13C and δ18O values for 8 herbivores from the Willacoochee Creek Fauna. Abbreviations: Unident. = unidentified, N = sample size, SD = standard deviation.

Bulk Sampling Taxon N δ13C (‰) δ18O (‰) Mean SD Range Mean SD Range Anchitherium clarencei 2 -9.57 0.94 -10.23 to -8.90 1.89 0.94 1.45 to 2.32 Merychippus primus 8 -9.35 0.51 -10.39 to -8.70 0.89 0.84 -0.82 to 2.10 Merychippus gunteri 4 -10.08 1.36 -11.19 to -8.30 0.43 1.10 -0.65 to 1.86 Acritohippus isonesus 5 -10.43 0.62 -11.00 to -9.67 1.16 1.11 -0.35 to 2.03 Teleoceras sp. 1 -11.40 - - -0.41 - - Aphelops sp. 3 -9.48 0.76 -10.23 to -8.72 -0.25 1.67 -2.15 to 1.01 Ticholeptus hypsodus 1 -9.83 - - 0.33 - - Unident. artiodactyl 1 -10.87 - - 0.66 - -

Serial Sampling Taxon N δ13C (‰) δ18O (‰) Mean SD Range Mean Range Aphelops sp. 19 -10.52 0.82 -10.23 to -8.72 0.06 -1.99 to 3.46 Merychippus primus 9 -9.67 0.43 -10.39 to -8.70 1.66 0.15 to 2.90 Acritohippus isonesus 12 -10.66 0.65 -11.19 to -8.30 1.51 0.25 to 2.56

Figure 5-2. Plot of bulk δ13C vs. δ18O values for the 8 herbivores from the Willacoochee Creek Fauna. Note that all the herbivores bear bulk δ13C values below -8.30‰ and most specimens cluster between -9.50‰ and -10.50‰.

Figure 5-3. δ13C values for serial samples of two Aphelops specimens. Note that all the values are lower than -8.37 ‰. Position along tooth refers to the distance from the base of the tooth.

Figure 5-4. δ18O values for serial samples of two Aphelops specimens. Note that both specimens exhibit variation outside of the range of error. Position along tooth refers to the distance from the base of the tooth.

Figure 5-5. δ13C values for serial samples of two Merychippus primus specimens. Note that all the values are lower than -9.26‰. Position along tooth refers to the distance from the base of the tooth.

Figure 5-6. δ18O values for serial samples of two Merychippus primus specimens. Note that both specimens exhibit variation outside of the range of error. Position along tooth refers to the distance from the base of the tooth.

Figure 5-7. δ13C values for serial samples of two Acriohippus isonesus specimens. Note that all the values are lower than -9.92‰. Position along tooth refers to the distance from the base of the tooth.

Figure 5-8. δ18O values for serial sampling of two Acritohippus isonesus specimens. Note that both specimens exhibit variation outside of the range of error. Position along tooth refers to the distance from the base of the tooth.

CHAPTER 6 DISCUSSION

Of the eight taxa analyzed from the Willacoochee Creek Fauna (WCF), three are

expected to yield significant proportions of browse in their diets: Anchitherium clarencei,

Ticholeptus hypsodus, and Aphelops. A. clarencei is a three-toed horse with low-crowned

cheek teeth and stocky limbs suggestive of a browse diet in a woodland habitat (Janis et

al., 2002). The only positively identified artiodactyl from the WCF, T. hypsodus, has a

mesodont dentition and is expected to be a mixed feeder (Lander, 1998). Aphelops is a

hornless, brachydont aceratherine rhino with long limbs suggestive of a cursorial habit

(Matthew, 1932; Janis, 1982). Aphelops is traditionally described as an open-country

browser, with the extant black rhino (Diceros bicornis) cited as a modern analog

(Matthew, 1932). Furthermore, Janis (1982) described all aceratherine rhinos as browsers

within woodland savannas. There has been some variance on the dietary classification of

Aphelops. Webb (1983) recognized Aphelops and Teleoceras as grazers from the

‘Clarendonian Chronofauna,’ the succession of Barstovian-Clarendonian grazers.

However, stable isotopic analyses of Aphelops specimens from multiple Florida localities

(ranging from 9.5 to 4.5 Ma), combined with crown-height data, suggest that Aphelops

was eating C3 browse material before and after the ~7 Ma spread of C4 grasses

(MacFadden, 1998). The WCF Aphelops specimens are approximately 6 million years older than the previously analyzed specimens, but the dental and postcranial morphologies still suggest a diet of C3 browse.

67 68

Teleoceras is a moderately hypsodont teleoceratine rhino with short limbs. It has

been traditionally interpreted as semi-aquatic grazer, comparable to the extant

hippopotamus, Hippopotamus amphibious (Scott, 1937; Voorhies, 1981; Webb, 1983;

Prothero, 1992). Matthew (1932) rejected the hypothesis of a semi-aquatic habitat for

Teleoceras, and instead suggested that the rhino grazed on grassy plains. Voorhies and

Thomasson (1979) determined that Teleoceras was a grazer, based on the presence of grass anthoecia in the oral and body cavities of specimens from the Ashfall Fossil Beds in

Nebraska (10 Ma). They were, however, unable to determine whether that grass

originated from mesic or lacustrine environments. Recent isotopic evidence, however, has revealed that: 1) Teleoceras was, based on δ18O values, not primarily aquatic; and 2)

Teleoceras was a mixed feeder, consuming significant portions of C3 grass prior to the 7

Ma spread of C4 grasses and then shifting to a diet of C4 grasses after 7 Ma (MacFadden,

1998). This interpretation likens Teleoceras to the extant white rhino (Ceratotherium

simum), a grazer (MacFadden, 1998). Based on this interpretation of Teleoceras, the

specimens from the WCF would be expected to have δ13C signatures indicative of a

mixed diet, possibly composed of C3 grasses.

The remaining equids from the WCF are all hypsodont three-toed horses. The

hypsodont teeth of Merychippus primus and Merychippus gunteri suggest a

grazing/mixed diet for these equids. Janis (1988) points out that, despite fairly hypsodont

cheek teeth, the Merychippus dentition is more similar to that of extant mixed feeders than to grazers. Whether Merychippus was a grazer or mixed feeder, the tooth morphology suggests that there is at least some grazing component to the animal’s diet.

Acritohippus isonesus, the most hypsodont taxon of the WCF, would be expected to be a

grazer.

Based on average bulk stable carbon isotope values, all eight taxa from the WCF

fell within the range of a C3–dominated diet, between -21‰ and -8‰ (Cerling and

Harris, 1999). Additionally, only one sample had a δ13C value more positive than -8.7‰

(a Merychippus gunteri specimen had a value of -8.30‰), indicating that nearly the entire

WCF community was consuming strictly C3 plants, according to the pre-industrial ranges of Feranec (2003). Even the most positive value was still more negative than -7.9‰, indicating that the animal was eating less than 10% C4 grass.

However, atmospheric δ13C values during the middle Miocene were considerably

higher than during the pre-industrial Holocene (Vincent and Berger, 1985). During the

Monterey Excursion, from 16.5 Ma to 13.5 Ma, δ13C values of benthic foraminifera were as high as +2.2‰, indicative of higher atmospheric δ13C (Zachos et al., 2001). The higher

atmospheric δ13C would raise the δ13C of plants, slightly raising the δ13C of the enamel of

mammals that consume the plants. This increase would be no greater than about 1‰.

Taking this enrichment into account, the δ13C signatures of all the samples are well

within a pure C3 diet. This suggests that each ungulate was either grazing on C3 grasses or

browsing, and not consuming any significant amounts of C4 grass. The serial sampling of

Aphelops sp., Merychippus primus, and Acritohippus isonesus, also reveal no significant

C4 fluctuations in diet at any point during the mineralization of the enamel. For both

Aphelops sp. and Merychippus primus, the most enriched serial sample was not greater

13 than -8.7‰, indicating a pure C3 diet. Once again, when the higher atmospheric δ C is considered, the serial samples fall well within the range of a pure C3 diet.

The stable carbon isotope analyses support the browser interpretations based on

dental morphologies for A. clarencei, T. hypsodus, and Aphelops. Conversely, the dentitions and lack of a C4 signal for Teleoceras and the three hypsodont equids suggest

that they were consuming C3 grasses. Unfortunately, the sample sizes of M. primus, M.

gunteri, and Teleoceras are inadequate for further investigation of grass consumption

(“e.g.”, microwear or mesowear analyses). Likewise, the sample sizes of the suspected

browsers are also insufficient for direct confirmation of a browse-dominated diet. A.

isonesus, however, can be assessed for evidence of a grass diet through microwear

analysis.

While conducting the microwear analyses, there was some doubt as to whether

two of the A. isonesus molars were M2s. However, since the parametric and

nonparametric statistical tests found no significant differences between the microwear

indices (MI) of each A. isonesus specimen, they are accepted here as all being M2 teeth.

The average microwear index (MI) for A. isonesus is 1.97, well above the 1.5 threshold

for grazers (MacFadden et al., 1999). The MI for N. trampasense is 1.98, which is also

significantly higher than 1.5. This supports the mesowear analysis, which classifies N.

trampasense as a grazer. The MI of A. isonesus is not statistically different from that of

N. trampasense. The relatively high MI of A. isonesus, coupled with the lack of a

significant difference to the MI from the grazing N. trampasense, suggests that A.

isonesus was also dominantly a grazer, with no significant browse component. The δ13C

data and microwear index indicate that A. isonesus was a C3 grazer (Figure 6-1). This

conclusion is supported by the scratch morphology. The quantified scratches in extant

grazers have indicated that a prevalence of fine scratches are observed in consumers of C3

grasses, while coarse scratches dominate the enamel wear of C4 grazers (Solounias and

Semprebon, 2002). For each sample of A. isonesus, fine scratches composed at least 84% of the total number of scratches.

Further paleoecological interpretations can be made from the bulk and serial δ13C

values of the WCF. In closed canopy forests, CO2 becomes trapped near the forest floor,

13 elevating the CO2 concentration near the forest floor and in turn depleting the δ C of that air which, in turn, results in the depletion of 13C in plant tissues (Medina and Minchin,

1980; Medina, Montes, Cuevas, and Rokzandic, 1986; van der Merwe and Medina, 1989;

Cerling et al., 1997a). Further depletion of 13C in plant tissues results from the lack of light reaching the forest floor, caused by the dense cover of the canopy (Medina and

Minchin, 1980; Medina et al., 1986; van der Merwe and Medina, 1989). These factors

cause the understory plants of a closed canopy forest to bear δ13C values as low as -37‰, a phenomenon known as the “Canopy Effect” (Medina and Minchin, 1980; Medina et al.,

1986; van der Merwe and Medina, 1989). The average δ13C value for these plants is -

30.9‰ (Ehleringer, Field, Lin, and Kuo, 1986). Modern C3 plants that grow in open

areas and do not suffer from water stress exhibit a δ13C range from -26 to -27‰ (Cerling et al., 1997a).

The range of bulk δ13C values for all the sampled WCF is -11.39‰ to -8.3‰,

which, considering analytical and fractionation error, indicates a diet that ranges in

isotopic composition from -25.49‰ ± 0.59‰ to -22.40‰ ± 0.59‰. An adjustment for

the Monterey Excursion of -1‰ changes the range to approximately -26.49‰ ± 0.59‰ to

-23.40‰ ± 0.59‰. With this adjustment, all of the bulk samples have δ13C values above

the maximum limit (-27‰) for the “Canopy Effect,” although the error range of the

Teleoceras sample extends down to -27.08‰. Even the two brachdyont taxa,

Anchitherium clarencei and Aphelops, have dietary bulk δ13C ranges that preclude them

from eating ground-level browse in a closed canopy forest. The serial sampling also does

not support a closed-canopy dietary component. All the serial samples for M. primus,

Aphelops, and A. isonesus indicate diets that have δ13C values above -27‰, except one A.

isonesus sample that yields a dietary value of -27.06‰. The error range of this sample,

along with the higher serial δ13C values for this specimen, suggests that this specimen

was still not eating in a closed-canopy forest. The range of bulk and serial δ13C values for the WCF reveals that the mammals lived and ate in an open habitat, such as a woodland savanna or grassland.

Paleoecological interpretations can also be made from the bulk δ13C values of the

13 WCF. When C3 plants are suffering water stress, they become enriched in C (Ehleringer et al., 1986; Ehleringer and Cooper, 1988; Ehleringer, 1991). During episodes of water stress, plants close their stomata to conserve water, consequently reducing CO2 intake

and enriching the δ13C of C3 plant tissues as high as -22‰ (Ehleringer et al., 1986;

Ehleringer and Cooper, 1988; Ehleringer, 1991). As mentioned above, the calibrated

isotopic range of the WCF diet is -25.49‰ ± 0.59‰ to -22.40‰ ± 0.59‰. An adjustment

for the Monterey Excursion of -1‰ changes the range to approximately -26.49‰ ±

0.59‰ to -23.40‰ ± 0.59‰. This range suggests that the WCF were feeding off of C3

plants in open areas that periodically experienced water stress. This interpretation is supported by the serial sampling, which reveals dietary ranges of -24.36‰ ± 0.59‰ to

-25.64‰ ± 0.59‰ for two serially sampled M. primus specimens and -25.02‰ ± 0.59‰ to -27.06‰ ± 0.59‰ for the two serially sampled A. isonesus specimens. The serially

sampled specimens of Aphelops also support an interpreted diet of water-stressed C3 plants, with significantly variant dietary ranges of -23.47‰ ± 0.59‰ to -26.82‰ ±

0.59‰. The stable carbon isotopic evidence suggests that the WCF were consuming water-stressed plants in open arid and seasonal environment, such as open-country plains.

Interpretations about local seasonality can be made from the serial δ18O values

from the structural carbonate in the enamel of the WCF. Assuming that equids from the

middle Miocene were obligate drinkers like modern equids, the serial sampling should

reveal a noticeable curve in δ18O values. For M. primus, specimens UF221419 and

UF221427 display trends that are nearly mirror images of one another. Enamel

mineralization, which begins at the crown, began during the winter for UF 221419, which is evident from the low δ18O values at 15 and 10 mm from the base of the tooth. Enamel mineralization, which takes about 1.5 to 2.8 years for cheek teeth in modern equids

(Bryant et al., 1996a), ended in the summer. Contrastingly, enamel mineralization for

UF221427 began in spring or early summer and was mostly completed by winter, according to the high δ18O values from 20 to 5 mm from the tooth base that are within

error of one another. Although modern equid cheek teeth take a minimum of about 1.5

years to mineralize, it is possible for the teeth of Merychippus primus to take less than a

year since the teeth are less high-crowned than modern equids. It should be noted that

each sample along the tooth bears a δ18O value that is time-averaged over a few months.

The other equid that was serially sampled, A. isonesus, reveals two distinct curves that

are also mirror images. UF221407 appears to have begun enamel mineralization towards the end of summer or fall and continued through the entire winter and into the next

74 summer. UF217590 began enamel mineralization in early summer or spring, in the winter.

Modern rhinocerotids are obligate drinkers (Clauss et al., 2005). Assuming that

Aphelops is also an obligate drinker, the δ18O values of the enamel carbonate should reflect any seasonal variation. UF 104227 displays considerable variation. Tooth formation began during the δ18O lows of winter. There are two summer peaks and three winter troughs, indicating that the Aphelops cheek tooth formed over roughly two years.

The δ18O variation of UF 116827 is not as pronounced as UF 104227, but a curve outside the error ranges is discernible. Tooth mineralization for UF 104227 began in a summer peak. The curve features two summer peaks and two winter lows, suggesting that the

Aphelops molar took slightly more than a year to mineralize. The lower δ18O values of

UF 116827 indicates that the tooth mineralized during a period of cooler temperatures than UF 104227. Overall, the δ18O variation in both Aphelops specimens supports the interpretation seasonality in northern Florida during the middle Miocene.

The variation of δ18O values for the serially sampled specimens of M. primus, A. isonesus, and Aphelops suggests that the WCF experienced significant seasonality. This stable oxygen isotopic data supports the interpretations of periodic water-stress made from the stable carbon isotopic data. The seasonality experienced by the WCF is typical of the warming that took place in the middle Miocene.

The middle Miocene (early Barstovian) WCF represent an interesting transition in the paleodiets of Floridian mammals. The mesodont equid Parahippus leonensis can be found at the early Miocene (Hemingfordian) Thomas Farm locality, from about 19 to 18

Ma (Hulbert and MacFadden, 1991). The rate of wear for P. leonensis cheek teeth is

about half of wear rate of grazing equids, indicating a mixed diet of grass and browse

(Hulbert, 1984). This interpretation has been supported by mesowear analyses (Hoffman,

unpublished). This suggests that grasses, either C3 or C4, existed in central Florida as far

back as 19 Ma, but the mesodont equids at were still consuming significant amounts of browse. Only hypsodont equids, like Acritohippus isonesus, were capable of exploiting

the more abrasive grasses. Thomas Farm also yields the brachydont equids Anchitherium

and Archaeohippus. A diminuitive presence of grasses would explain the lack of

hypsodont taxa at the Thomas Farm site. It was only after C3 grasses became more

abundant, by 15.8 Ma in northern Florida, that equids radiated into hypsodont lineages.

Mesowear analyses revealed grazing diets for all five of the hypsodont equids from

the last 10 million years. Coupled with previous isotopic data (Table 6-1), it is possible to

trace the existence of C3 grasses over the last 10 million years in Florida. All three horses from the Love Site (Neohipparion trampasense, Cormohipparion plicatile, and

13 Cormohipparion ingenuum) have δ C values indicative of a C3-dominated diet.

Furthermore, the most positive δ13C value of these samples is -10.8‰, well below -8.7‰.

This indicates that the diets of these horses consisted purely of C3 plants. Since the

mesowear analyses indicate that these taxa were primarily grazers, they must have been

13 feeding on C3 grasses. The δ C values from enamel indicate a range of -26.7‰ to

-24.9‰ for the consumed plants. This suggests that the equids were eating water-stressed

C3 grasses in seasonal, open habitat. This interpretation compares favorably with other

studies. The population dynamics of N. trampasense from the Love Site reveal a high rate

of tooth wear, comparable to the modern zebra Equus burchelli, which indicates a highly

abrasive grass-dominated diet (Hulbert, 1982). The presence of discrete age classes at the

site suggests that the area was a wooded grassland savanna with seasonal rains that

controlled the migratory and birthing patterns (Hulbert, 1982). N. trampasense migrated

away from the Love site during the wet season to give birth, then returned during the dry

season (Hulbert, 1982). This interpretation is supported by the range of δ13C values,

which suggest that the equids at the Love Site were consuming water-stressed grasses.

The next site, the Upper Bone Valley Formation, features specimens of Nannippus

aztecus that represent the 4.5 Ma level. The three δ13C values for N. aztecus indicate a

diet composed of both C3 and C4 plant material (Table 6-1). Combined with the

mesowear grazer classification, N. aztecus can be interpreted as an opportunistic grazer, feeding on both C3 and C4 grasses and possibly some browse. This level is dated after the

global shift in carbon biomass, when C4 grasses became dominant. However, the N.

aztecus specimens consist of an M1, an M2, and an M1/2. First molars are completely mineralized during the weaning process, suggesting that the isotopic compositions of

M1s are influenced by the mother’s milk (Bryant et al., 1996a; Fricke and O’Neil, 1996;

Hoppe et al., 2004). Second molars are not completely mineralized until after the weaning process ends, but they can still reflect the isotopic composition of the nursing diet (Bryant et al., 1996a; Fricke and O’Neil, 1996; Hoppe et al., 2004). The M2 and

M1/2 δ13C values, therefore, might be skewed.

A study conducted on 6 sympatric horses from Upper Bone Valley Formation

concluded that, as recently as 5 million years ago, C4 tropical grasses in Florida coexisted

with C3 grasses. Using microwear, this study showed that Nannippus aztecus, along with

the sympatric late Hemphillian horse Pseudhipparion simpsoni, consumed both C3 and C4 grasses (MacFadden et al., 1999). This study substantiates the interpretation of

Nannippus aztecus as a C3/C4 grazer. Furthermore, older N. aztecus specimens from the 7

Ma Moss Acres Racetrack site (Table 6-1) have δ13C values of -7.9‰ and -5.3‰. These

values indicate a mixed diet of C3 and C4 plants. Mesowear analyses could not be conducted on this population because the population size was too small (<20). However, if this population behaved similarly to the younger Upper Bone Valley Nannippus populations (“i.e.”, they were primarily grazing), then this population was likely composed of C3/C4 grazers.

Finally, the Leisey Shell Pit represents the 1.5 Ma level. The specimens of Equus

13 “leidyi” have δ C values that are more negative than -1.3‰, indicating a C4-dominated

diet with a minor C3 component. Modern Equus is predominantly a grazer, but is known

to eat other locally available plants (Berger, 1986). The isotopic and mesowear data suggests that, like modern Equus, Equus “leidyi” was a dominant C4 grazer at 1.5 Ma.

Previous work (MacFadden et al., 1999) as well as the mesowear analyses of

various Florida equid populations have shown that C3 grasses were present in Florida from ~10 to ~5 Ma. In light of the microwear and isotopic data retrieved from the WCF, it appears that the record of C3 grasses in northern Florida and southern Georgia extends

back to at least 15 million years ago. As evidenced by the consumption C3 grasses by a

hypsodont taxon, it is possible that C3 grasses forced the adaptation of hypsodonty in

ungulates.

2.5

2.3 Pure C3 Mixed Diet Pure C4 2.1

1.9

1.7 Grazers 1.5

1.3

1.1

0.9 Browsers Acritohippus isonesus 0.7 Neohipparion trampasense Microwear Indices (Scratches/Pits)

0.5 -12-11-10-9-8-7-6-5-4-3-2-10 1 δ13C (V-PDB)

Figure 6-1. Plot of microwear index versus δ13C values. The dashed line represents a microwear index of 1.5, the boundary between a browser and a grazer classification. Solid lines mark the boundaries between ‘Pure C3 diet’, ‘Mixed 13 diet’, ‘Pure C4 diet” and transitional zones. δ C values below -8.7‰ indicate a pure C3 diet. Note that both Acritohippus isonesus and Neohipparion trampasense have microwear indices above 1.5 and δ13C values below -8.7‰, suggesting diets composed purely of C3 grasses.

Table 6-1. Isotopic data for equids from 4 central Florida sites. Abbreviations: UF ID = catalogue number in Florida Museum of Natural History collections. (Modified woth permission from MacFadden and Cerling, 1996)

Taxon Site UF ID Material δ13C (‰)

9.5 Ma Level, late Clarendonian, late Miocene

7.0 Ma level, "middle" Hemphillian, late Miocene

Nannippus aztecus Moss Acres 69933 R. M3 -7.9 cf. Nannippus Moss Acres None R. P2 -5.3

4.5 Ma level, latest Hemphillian, early Pliocene Nannippus Upper Bone Valley None L. M -6.0 Nannippus aztecus Upper Bone Valley 63633 L. M2 -9.1 Nannippus aztecus Upper Bone Valley 65796 R. M1 -2.4

1.5 Ma level, early Irvingtonian, early Pleistocene Equus "leidyi" Leisey 1A 80047 R. M3 -1.5

Equus "leidyi" Leisey 1A None R. P2 -3.5

CHAPTER 7 CONCLUSIONS

Stable carbon isotope analyses of the Willacoochee Creek Fauna, the oldest

isotopically sampled taxa in Florida, provide no evidence for the presence of C4 grasses

in the diets of ungulates in northern Florida and southern Georgia at from 15.3 to 15.9

Ma. This suggests that C4 grasses were not present in the area at important time in mammal evolution and, since they are not a significant dietary component, it is unlikely

that they were responsible for hypsodonty adaptations. Furthermore, the microwear of the

most hypsodont taxon from the WCF, Acritohippus isonesus, indicates a grazing diet.

13 Combined with the C3-consuming classification, based on δ C values from bioapatite, the microwear suggests that Acritohippus isonesus ate C3 grasses. This substantiates the

presence of C3 grasses in the middle Miocene of northern Florida and supports Fox and

Koch’s (2003) claim that C3 grasses were responsible for the hypsodonty adaptations.

Stable carbon and oxygen isotope values also indicate that the Willacoochee Creek Fauna

ate many water-stressed plants and lived in an open, arid, and seasonal environment, such

as open-country plains

Mesowear analyses and previously published isotope data reveal that the C3 grasses were the primary dietary component in some Floridian horses at ~9.5 Ma. Even after the global carbon biomass shift at ~7 Ma, C3 grasses persisted in Florida and were significant dietary components until at least ~4.5 Ma. By ~1.5 Ma, it appears that abundance of C3 grasses finally subsided in Florida, where the dominant grazer (Equus

“leidyi”) was consuming mostly C4 grass. This long reliance off of C3 grasses suggests

80 81

that C3 grasses were a driving mechanism for the appearance of hypsodonty in ungulates.

The combinations of mesowear and stable isotope analyses and microwear and stable isotope analyses have proven to be valuable tools in assessing the origins of hypsodonty.

Future work will involve applying these combined methods to the northern Great Plains, where there is a rich record of ungulate fossils.

APPENDIX A MESOWEAR VALUES LISTED BY SAMPLE

Table A-1. Abbreviations: UF ID = catalogue number in Florida Museum of Natural History collections, CS = Cusp Shape, OR = Occlusal Relief, BVF = Bone Valley Formation, Leisey = Leisey Shell Pit, B = Blunt, R = Round, S = Sharp, H = High, L = Low. UF ID Taxon Locality Material CS OR 62251 Neohipparion trampasense Love Site L. M1 R L 25637 Neohipparion trampasense Love Site L. M1 R L 27992 Neohipparion trampasense Love Site L. M2 R L 32273 Neohipparion trampasense Love Site L. M2 B L 32258 Neohipparion trampasense Love Site R. M2 R L 27991 Neohipparion trampasense Love Site L. M2 R H 32253 Neohipparion trampasense Love Site L. M2 R L 53428 Neohipparion trampasense Love Site L. M2 B L 53427 Neohipparion trampasense Love Site L. M1 B L 32256 Neohipparion trampasense Love Site L. M2 R H 32272 Neohipparion trampasense Love Site L. M2 R H 32252 Neohipparion trampasense Love Site L. M2 R L 53429 Neohipparion trampasense Love Site L. M1/2 B L 53230 Neohipparion trampasense Love Site L. M1/2 R L 53243 Neohipparion trampasense Love Site L. M1/2 R L 53256 Neohipparion trampasense Love Site L. M1/2 R L 53269 Neohipparion trampasense Love Site R. M1/2 B L 53244 Neohipparion trampasense Love Site L. M1/2 B L 53257 Neohipparion trampasense Love Site L. M1/2 B L 53232 Neohipparion trampasense Love Site L. M1/2 R H 53271 Neohipparion trampasense Love Site R. M1/2 B L 53246 Neohipparion trampasense Love Site R. M1/2 R H 53272 Neohipparion trampasense Love Site R. M1/2 B L 53234 Neohipparion trampasense Love Site L. M1/2 B L 53247 Neohipparion trampasense Love Site L. M1/2 B L 53273 Neohipparion trampasense Love Site R. M1/2 R H 53235 Neohipparion trampasense Love Site L. M1/2 B L 53248 Neohipparion trampasense Love Site L. M1/2 B L 53236 Neohipparion trampasense Love Site L. M1/2 R H 53249 Neohipparion trampasense Love Site L. M1/2 R H 53275 Neohipparion trampasense Love Site R. M1/2 B L 53237 Neohipparion trampasense Love Site L. M1/2 B L 53250 Neohipparion trampasense Love Site L. M1/2 R H 53276 Neohipparion trampasense Love Site R. M1/2 S H 53225 Neohipparion trampasense Love Site L. M1/2 B L 53251 Neohipparion trampasense Love Site L. M1/2 R L

82 83

Table A-1. Continued UF ID Taxon Locality Material CS OR 53277 Neohipparion trampasense Love Site R. M1/2 R H 53265 Neohipparion trampasense Love Site R. M1/2 B L 53278 Neohipparion trampasense Love Site R. M1/2 R H 53282 Neohipparion trampasense Love Site R. M1/2 R L 96934 Cormohipparion plicatile Love Site R. M2 B L 53338 Cormohipparion plicatile Love Site R. M2 R H 36289 Cormohipparion plicatile Love Site R. M2 B L 27993 Cormohipparion plicatile Love Site R. M2 S H 96933 Cormohipparion plicatile Love Site L. M2 R L 32270 Cormohipparion plicatile Love Site R. M1 B L 32260 Cormohipparion plicatile Love Site L. M2 R L 32250 Cormohipparion plicatile Love Site R. M2 R L 35891 Cormohipparion plicatile Love Site L. M1 R L 32262 Cormohipparion plicatile Love Site L. M2 R L 27316 Cormohipparion plicatile Love Site R. M2 R H 32264 Cormohipparion plicatile Love Site R. M2 R H 32283 Cormohipparion plicatile Love Site L. M2 B L 32266 Cormohipparion plicatile Love Site L. M2 B L 69811 Cormohipparion plicatile Love Site R. M1 R L 53346 Cormohipparion plicatile Love Site R. M1/2 R L 53377 Cormohipparion plicatile Love Site L. M1/2 R H 53347 Cormohipparion plicatile Love Site R. M1/2 R H 53348 Cormohipparion plicatile Love Site R. M1/2 B L 53363 Cormohipparion plicatile Love Site L. M1/2 R L 53379 Cormohipparion plicatile Love Site L. M1/2 B L 53331 Cormohipparion plicatile Love Site R. M1/2 R H 53349 Cormohipparion plicatile Love Site R. M1/2 B L 53364 Cormohipparion plicatile Love Site L. M1/2 R H 53351 Cormohipparion plicatile Love Site R. M1/2 R H 53365 Cormohipparion plicatile Love Site L. M1/2 S L 53352 Cormohipparion plicatile Love Site R. M1/2 R L 53366 Cormohipparion plicatile Love Site L. M1/2 B L 53336 Cormohipparion plicatile Love Site R. M1/2 R H 53353 Cormohipparion plicatile Love Site R. M1/2 S H 53367 Cormohipparion plicatile Love Site L. M1/2 R L 53370 Cormohipparion plicatile Love Site L. M1/2 R L 53339 Cormohipparion plicatile Love Site R. M1/2 R H 53371 Cormohipparion plicatile Love Site L. M1/2 S H 32300 Cormohipparion ingenuum Love Site L. M2 R H 32254 Cormohipparion ingenuum Love Site R. M2 R L 32296 Cormohipparion ingenuum Love Site L. M2 R L 53426 Cormohipparion ingenuum Love Site R. M1 R L 53396 Cormohipparion ingenuum Love Site R. M2 R L 53375 Cormohipparion ingenuum Love Site L. M1/2 R H 53391 Cormohipparion ingenuum Love Site R. M1/2 B L 53395 Cormohipparion ingenuum Love Site R. M1/2 B L 53394 Cormohipparion ingenuum Love Site R. M1/2 R H

Table A-1. Continued UF ID Taxon Locality Material CS OR 53390 Cormohipparion ingenuum Love Site R. M1/2 R H 53389 Cormohipparion ingenuum Love Site R. M1/2 R H 53383 Cormohipparion ingenuum Love Site L. M1/2 R H 53384 Cormohipparion ingenuum Love Site L. M1/2 R H 53385 Cormohipparion ingenuum Love Site L. M1/2 B L 62409 Cormohipparion ingenuum Love Site L. M1/2 R L 96377 Cormohipparion ingenuum Love Site L. M1/2 R H 62417 Cormohipparion ingenuum Love Site R. M1/2 R H 62391 Cormohipparion ingenuum Love Site L. M1/2 R H 62430 Cormohipparion ingenuum Love Site R. M1/2 R H 62398 Cormohipparion ingenuum Love Site L. M1/2 R L 62429 Cormohipparion ingenuum Love Site R. M1/2 R H 62395 Cormohipparion ingenuum Love Site L. M1/2 B L 62427 Cormohipparion ingenuum Love Site R. M1/2 R L 6700 Nannippus aztecus BVF R. M2 R H 57576 Nannippus aztecus BVF L. M1 R - 57576 Nannippus aztecus BVF L. M2 R H 212370 Nannippus aztecus BVF R. M1 R H 211769 Nannippus aztecus BVF R. M1/2 R H 208402 Nannippus aztecus BVF R. M1/2 R L 207957 Nannippus aztecus BVF R. M1 R H 203506 Nannippus aztecus BVF R. M1/2 R H 156921 Nannippus aztecus BVF R. M2 R H 156920 Nannippus aztecus BVF R. M1 R H 130079 Nannippus aztecus BVF R. M1/2 R H 124209 Nannippus aztecus BVF R. M1/2 R L 102613 Nannippus aztecus BVF R. M1/2 R H 102612 Nannippus aztecus BVF R. M1/2 R L 102611 Nannippus aztecus BVF R. M1/2 R L 100239 Nannippus aztecus BVF R. M1/2 B L 100241 Nannippus aztecus BVF R. M1/2 B L 93223 Nannippus aztecus BVF R. M1/2 R H 67981 Nannippus aztecus BVF R. M1/2 B L 63630 Nannippus aztecus BVF R. M1/2 R H 55933 Nannippus aztecus BVF R. M1/2 B L 55893 Nannippus aztecus BVF R. M1/2 R L 53953 Nannippus aztecus BVF R. M1 B L 47371 Nannippus aztecus BVF R. M1/2 R L 17252 Nannippus aztecus BVF R. M1 R H 208360 Nannippus aztecus BVF L. M1 R H 211787 Nannippus aztecus BVF L. M1 R H 211892 Nannippus aztecus BVF L. M1 R L 58296 Nannippus aztecus BVF L. M1 R H 63627 Nannippus aztecus BVF L. M1 R L 63633 Nannippus aztecus BVF L. M2 R H 63693 Nannippus aztecus BVF L. M1 R H 63964 Nannippus aztecus BVF L. M1 B L

Table A-1. Continued UF ID Taxon Locality Material CS OR 63967 Nannippus aztecus BVF L. M1 R L 17279 Nannippus aztecus BVF L. M1 R H 47362 Nannippus aztecus BVF L. M2 R L 80090 Equus “leidyi” Leisey L. M1 B L 85523 Equus “leidyi” Leisey L. M2 B L 85780 Equus “leidyi” Leisey R. M1 B L 83800 Equus “leidyi” Leisey R. M2 B L 85776 Equus “leidyi” Leisey L. M1 B L 85772 Equus “leidyi” Leisey R. M1 R L 85771 Equus “leidyi” Leisey L. M1 B L 82642 Equus “leidyi” Leisey L. M1 B L 84173 Equus “leidyi” Leisey L. M1/2 S L 85708 Equus “leidyi” Leisey L. M1/2 S L 85720 Equus “leidyi” Leisey L. M1/2 B L 85710 Equus “leidyi” Leisey L. M1/2 B L 85683 Equus “leidyi” Leisey L. M1/2 B L 84172 Equus “leidyi” Leisey L. M1/2 B L 85679 Equus “leidyi” Leisey L. M1/2 B L 81636 Equus “leidyi” Leisey L. M1/2 R L 85680 Equus “leidyi” Leisey L. M1/2 R L 81797 Equus “leidyi” Leisey L. M1/2 B L 85719 Equus “leidyi” Leisey L. M1/2 R L 85727 Equus “leidyi” Leisey L. M1/2 B L 85698 Equus “leidyi” Leisey L. M1/2 B L 85730 Equus “leidyi” Leisey L. M1/2 B L 85729 Equus “leidyi” Leisey L. M1/2 R L 82079 Equus “leidyi” Leisey L. M1/2 R L 85724 Equus “leidyi” Leisey L. M1/2 B L 85689 Equus “leidyi” Leisey L. M1/2 B L 85692 Equus “leidyi” Leisey L. M1/2 B L

APPENDIX B SERIAL STABLE ISOTOPE VALUES

Table B-1. Abbreviations: ELC = Englehard La Camelia Mine, MGF = Milwhite Gunn Farm Mine, and LC2 = La Camelia 2 Mine. Sample UF ID Taxon Height δ13C (‰) δ18O (‰) Tooth Site 116827 Aphelops sp. 0 mm -8.37 -1.89 R. MX ELC 116827 Aphelops sp. 5 mm -9.98 -1.95 R. MX ELC 116827 Aphelops sp. 10 mm -9.41 -1.65 R. MX ELC 116827 Aphelops sp. 15 mm -9.91 -1.26 R. MX ELC 116827 Aphelops sp. 20 mm -10.72 -1.42 R. MX ELC 116827 Aphelops sp. 25 mm -10.94 -1.55 R. MX ELC 116827 Aphelops sp. 30 mm -11.44 -1.99 R. MX ELC 116827 Aphelops sp. 35 mm -11.35 -1.19 R. MX ELC 116827 Aphelops sp. 40 mm -11.55 -0.80 R. MX ELC 116827 Aphelops sp. 45 mm -11.28 -0.90 R. MX ELC 116827 Aphelops sp. 50 mm -11.72 -0.31 R. MX ELC 104227 Aphelops sp. 0 mm -9.76 0.71 Unident. MGF 104227 Aphelops sp. 5 mm -10.55 2.34 Unident. MGF 104227 Aphelops sp. 10 mm -10.40 1.87 Unident. MGF 104227 Aphelops sp. 15 mm -10.53 1.22 Unident. MGF 104227 Aphelops sp. 20 mm -10.54 3.46 Unident. MGF 104227 Aphelops sp. 25 mm -10.29 3.15 Unident. MGF 104227 Aphelops sp. 30 mm -10.39 1.67 Unident. MGF 104227 Aphelops sp. 35 mm -10.73 1.60 Unident. MGF 221419 Merychippus primus 0 mm -9.26 1.99 L. P4 ELC 221419 Merychippus primus 5 mm -10.54 0.67 L. P4 ELC 221419 Merychippus primus 10 mm -10.19 0.15 L. P4 ELC 221419 Merychippus primus 15 mm -9.27 0.15 L. P4 ELC 221415 Merychippus primus 0 mm -9.39 1.15 R. P3 ELC 221415 Merychippus primus 5 mm -9.66 2.90 R. P3 ELC

86 87

Table B-1. Continued Sample UF ID Taxon Height δ13C (‰) δ18O (‰) Tooth Site 221415 Merychippus primus 10 mm -9.72 2.65 R. P3 ELC 221415 Merychippus primus 15 mm -9.49 2.48 R. P3 ELC 221415 Merychippus primus 20 mm -9.50 2.83 R. P3 ELC 221407 Acritohippus isonesus 0 mm -9.94 1.12 R. P3/4 ELC 221407 Acritohippus isonesus 5 mm -10.16 0.25 R. P3/4 ELC 221407 Acritohippus isonesus 10 mm -10.34 0.36 R. P3/4 ELC 221407 Acritohippus isonesus 15 mm -9.92 0.97 R. P3/4 ELC 221407 Acritohippus isonesus 20 mm -10.63 1.68 R. P3/4 ELC 221407 Acritohippus isonesus 25 mm -10.66 2.56 R. P3/4 ELC 217590 Acritohippus isonesus 0 mm -10.52 0.62 L. P4 LC2 217590 Acritohippus isonesus 5 mm -10.33 2.01 L. P4 LC2 217590 Acritohippus isonesus 10 mm -10.70 2.37 L. P4 LC2 217590 Acritohippus isonesus 15 mm -10.94 2.13 L. P4 LC2 217590 Acritohippus isonesus 20 mm -11.78 2.16 L. P4 LC2 217590 Acritohippus isonesus 25 mm -11.96 1.82 L. P4 LC2

LIST OF REFERENCES

Bender, M.M., 1971. Variations in the 13C/12C Ratios of Plants in Relation to the Pathway of Photosynthetic Caron Dioxide Fixation. Phytochemistry, 10: 1239-1245.

Berger, J., 1986. Wild horses of the Great Basin: Social competition and population size. Chicago University press, Chicago, 326 pp.

Black, C.C. and Dawson, M.R., 1966. A review of Late Eocene mammalian faunas from North America. Am. J. Sci., 264: 321-349.

Bode, F.D., 1931. Characters useful in determining the position of individual teeth in the permanent cheek-tooth series of merychippine horses. J. Mammal., 12: 118-129.

Bown, T.M., 1980. The Willwood Formation (lower Eocene) of the southern Bighorn Basin, Wyoming, and its mammalian faunas. P.D. Gingerich, P.D. (Editor), Early Cenozoic Paleontology and Stratigraphy of the Bighorn Basin. University of Michigan Papers in Paleontol., 24: 25-32.

Bown, T.M. and Krause, M.J., 1981. Lower Eocene alluvial Paleosols (Willwood Formation, northwest Wyoming, U.S.A.) and their significance for paleoecology, paleoclimatology, and basin analysis. Palaeogeogr., Palaeoclimatol., Palaeoecol., 34: 1-30.

Bown, T.M. and Kraus, M.J., 1987. Integration of channel and floodplain suites; I. Developmental sequence and lateral relationships of alluvial paleosols. J. Sedimentary Petrol., 57: 587-601.

Bryant, J.D., 1991. New Early Barstovian (Middle Miocene) Vertebrates from the Upper Torreya Formation, Eastern Florida Panhandle. J. Vertebrate Paleontol., 11: 472- 489.

Bryant, J.D. and Froelich, P.N., 1995. A model of oxygen isotope fractionation in body water of large mammals. Geochim. Cosmochim. Acta, 59: 4523-4537.

Bryant, J.D., Froelich, P.N., Showers, W.J., and Genna, B.J., 1996a. A tale of two quarries: Biologic and taphonomic signatures in the oxygen isotope composition of tooth enamel phosphate from modern and Miocene equids. Palaios, 11: 397-408.

Bryant, J.D., Froehlich, 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. Palaeogeogr., Palaeoclimatol., Palaeoecol., 126: 75-89.

88 89

Bryant, J.D., Koch, P.L., Froelich, P.N., Showers, W.J., and Genna, B.J., 1996., Oxygen isotope partitioning between phosphate and carbonate in mammalian apatite. Geochim. Cosmochim. Acta, 60: 5145-5148.

Bryant, J.D., MacFadden, B.J. and Mueller, P.A., 1992. Improved chronologic resolution of the Hawthorn and the Alum Bluff Groups in northern Florida: Implications for Miocene chronostratigraphy. Geological Society of America Bulletin, 104: 208- 218.

Camp, C.L. and Smith, N., 1942. Phylogeny and function of the digital ligaments of the horse. University of California Memoirs, 13: 69-124.

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

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

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

Cerling, T.E. and Sharp, Z.D., 1996. Stable carbon and oxygen isotope analysis of fossil tooth enamel using laser ablation. Palaeogeogr., Palaeoclimatol., Palaeoecol., 126: 173-186.

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.

Clauss, M., Polster, C., Kienzle, E., Wiesner, H., Baumgartner, K., von Houwald, F., Streich, W.J. and Dierenfeld, E., 2005. Energy and mineral nutrition and water intake in the captive Indian Rhinoceros (Rhinoceros unicornis). Zoo Biol., 24: 1- 14.

Coplen, T.B., 1996. More uncertainty than necessary. Paleoceanography, 11: 369-370.

Craig, H., 1957. Isotopic standards for carbon and oxygen and correction factors for mass spectrometric analysis of carbon dioxide. Geochim. Cosmochim. Acta, 12: 133- 149.

Crepet, W.L. and Feldman, G.D., 1991. The earliest remains of grasses in the fossil record. Am. J. Botany, 78: 1010-1014.

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

Dongman, G., Nürnberg, H.W., Förstel, H. and Wagener, K., 1974. On the enrichment of 18 the H2 O in the leaves of transpiring plants. Radiation and Environmental Biophysics, 11: 41-52.

Ehleringer, J.R., 1991. 13C/12C fractionation and its utility in terrestrial plant studies. In: D.C. Coleman and B. Fry (Editors), Carbon Isotope Techniques. Academic Press, New York: 187-200.

Ehleringer, J.R. and Cooper, T.A., 1988. Correlations between carbon isotope ratio and microhabitat in desert plants. Oecologia, 76: 562-566.

Ehleringer, J.R., Field, C.B., Lin, Z. and Kuo, C., 1986. Leaf carbon isotope and mineral composition in subtropical plants along an irradiance cline. Oecologia, 70: 520- 526.

Ehleringer, J.R., Sage, R.F., Flanagan, L.B., and Pearcy. R.W., 1991. Climate change and the evolution of C4 photosynthesis. Trends in Ecol. and Evol., 6: 95-99.

Elias, M.K., 1942. Tertiary prairie grasses and other herbs from the High Plains. Geological Society of America Special Papers, 41: 176 pp.

Emry, R. J., 1981. Additions to the mammalian fauna of the type Duchesnean, with comments on the status of the Duchesnean “age.” J. Paleontol., 55: 563-570.

Epstein, S., Thompson, P. and Yapp, C.J., 1977. Oxygen and hydrogen isotope relations in plants. Science, 198: 1209-1215.

Fagan, S. R., 1960. Osteology of Mylagaulus laevis, a fossorial rodent from the Upper Miocene of Colorado. University of Kansas Paleontological Contributions, 9: 1-32.

Feranec, R.S., 2003. Stable isotopes, hypsodonty, and the paleodiet of Hemiauchenia (Mammalia: Camelidae): a morphological specialization creating ecological generalization. Paleobiol., 29: 230-242.

Feranec, R.S. and MacFadden, B.J., 2000. Evolution of the grazing niche in Pleistocene mammals from Florida: evidence from stable isotopes. Palaeogeogr., Palaeoclimatol., Palaeoecol., 162: 155-169.

Flower, B.P., 1999. Warming without high CO2? Nature, 399: 313-314.

Flower, B.P. and Kennett, J.P., 1995. Middle Miocene deepwater paleoceanography in the Southwest Pacific; relations with East Antarctic ice sheet development. Paleoceanography, 10: 1095-1112.

Förstel, H., 1978. The enrichment of 18O in leaf water under natural conditions. Radiation and environmental biophysics, 15: 323-344.

Fortelius, M. and Solounias, N., 2000. Functional characterization of ungulate molars using the abrasion attrition wear gradient: A new method for reconstructing paleodiets. Am. Museum Novitates, 3301: 1-36.

Fox, D.L., and Koch, P.L., 2003. Tertiary History of C4 Biomass in the Great Plains, USA. Geol., 31: 809-812.

Frederickson, N.O., 1981. Middle Eocene to early Oligocene plant communities of the Gulf Coast. In: J. Gray, A.J. Boucot, and W.B.N. Berry (Editors), Communities of the past. Hutchinson Ross, Stroudsburg: 493-550.

Friedli, H., Lötscher, H., Oeschger, H., Siegenthaler, U., and Stauffer, B., 1986. Ice core 13 12 record of the C/ C ratio of atmospheric CO2 in the past two centuries. Nature, 324: 237-238.

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. Palaeogeogr., Palaeoclimatol., Palaeoecol., 126: 91-99.

Gadbury, C., Todd, L., Jahren, A.H., and Amundson, R., 2000. Spatial and temporal variations in the isotopic composition of bison tooth enamel from the early Holocene Hudson-Meng Bone Bed, Nebraska. Palaeogeogr., Palaeoclimatol., Palaeoecol., 157: 79-93.

Galbreath, E.C., 1969. Cylindrodont rodents from the lower Oligocene of northeastern Colorado. Transactions of the Illinois State Academy of Sci., 62: 94-97.

Gazin, C.L., 1965. A study of the Early Tertiary condylarthran mammal Meniscotherium. Smithsonian Miscellaneous Collections, 149: p. 98.

Gazin, C.L., 1968. A study of the Eocene condylarthran mammal, Hyopsodus. Smithsonian Miscellaneous Collections, 153: 1-90.

Gingerich, P.D., 2001. Biostratigraphy of the continental Paleocene-Eocene boundary interval on Polecat bench in the northern Bighorn Basin. In: P.D. Gingerich (Editor), Paleocene-Eocene stratigraphy and biotic change in the Bighorn and Clark’s Fork basins. Wyoming. University of Michigan Papers on Paleontol., 33: 37-71.

Green, M., 1972. Lagomorpha from the Rosebud Formation, South Dakota. J. Paleontol., 46: 377-385.

Gregory, J. T., 1971. Speculations on the significance of fossil vertebrates for the antiquity of the Great Plains of North America. Abhandlungen des Hessischen Landesamtes fuer Bodenforschung, 60: pp.64-72.

Grine, F.E., 1986. Dental evidence for dietary differences in Australopithecus and Paranthropus: A quantitative analysis of permanent molar microwear. J. Human Evol., 15: 783-822

Hambrey, M.J., Ehrmann, W.U., and Larsen, B., 1991. Cenozoic glacial record of the Prydz Bay continental shelf, east Antarctica. Proceedings of the Ocean Drilling Program, Scientific Results, 119: 77-132.

Hickey, L.J., 1980. Paleocene stratigraphy and flora of the Clark’s Fork Basin. In: Gingerich, P.D. (Ed.), Early Cenozoic Paleontology and Stratigraphy of the Bighorn Basin. University of Michigan Papers in Paleontol., 24: 33-49.

Hillson, S., 1986.Teeth. Uniof Cambridge, Cambridge. 376 pp.

Higgins, P. and MacFadden, B.J., 2004. “Amount Effect” recorded in oxygen isotopes of late Glacial horse (Equus) and bison (Bison) teeth from the Sonoran and Chihuahuan deserts, southwestern United States. Palaeogeogr., Palaeoclimatol., Palaeoecol., 206: 337-353.

Hodell, D.A. and Woodruff., 1994. Variations in the strontium isotopic ration of seawater during the Miocene: Stratigraphic and geochemical implications. Paleoceanography, 9: 405-426.

Hoppe, K.A., Stover, S.M., Pascoe, J.R., and Amundson, R., 2004. Tooth enamel biomineralization in extant horses: implications for isotopic microsampling. Palaeogeogr., Palaeoclimatol., Palaeoecol., 206: 355-365.

Huber, O., 1982. Significance of savanna vegetation in the Amazon Territory of Venezuela. In: Prance, G.T. (Ed.), Biological diversification in the tropics, Columbia University Press, New York: 221-224.

Huddlestun, P.F., 1988. A revision of the lithostratigraphic units of the coastal plain of Georgia: the Miocene through Holocene. Georgia Geologic Survey Bulletin, 104: 1-62

Hulbert, R.C., Jr., 1982. Population dynamics of the three-toed horse Neohipparion from the late Miocene of Florida. Paleobiol., 8: 159-167.

Hulbert, R. C. Jr., 1984. Paleoecology and population dynamics of the early Miocene (Hemingfordian) horse Parahippus leonensis from the Thomas Farm Site, Florida. J. Vertebrate Paleontol., 4: 547-558.

Hulbert, R.C., Jr., 1987. Late Neogene Neohipparion (Mammalia, Equidae) from the Gulf Coastal Plain of Florida and Texas. J. Paleontol., 61: 809-830.

Hulbert, R.C., Jr., 1988. Cormohipparion and Hipparion (Mammalia, Perissodactyla, Equidae) from the late Neogene of Florida. Bulletin of the Florida State Museum, Biological Sci., 33: 229-338.

Hulbert, R.C., Jr., 1992. A checklist of the fossil vertebrates of Florida. Papers in Florida Paleontol., 6: 1-35.

Hulbert, R.C., Jr., 1995. Equus from Leisey Shell Pit 1A and other Irvingtonian localities from Florida. Bulletin of the Florida Museum of Natural History, 37, Pt. II (17): 553-602.

Hulbert, R.C., Jr. and MacFadden, B.J., 1991. Morphological transformation and cladogenesis at the base of the adaptive radiation of Miocene hypsodont horses. Am. Museum Novitates, 3000: 1-61.

Hutchinson, J.H., 1982. Turtle, crocodilian, and champosaur diversity changes in the Cenozoic of the north-central region of western United States. Palaeogeogr., Palaeoclimatol., Palaeoecol., 37: 149-164.

Jacobs, B.F., Kingston, J.D., and Jacobs, L.L., 1999. The origin of grass-dominated ecosystems. Annals of the Missouri Botanical Garden, 86: 590-643.

Janis, C.M., 1982. Evolution of horns in ungulates: ecology and paleoecology. Biological Review, 57: 261-326.

Janis, C.M., 1988. An estimation of tooth volume and hypsodonty indices in ungulate mammals, and the correlation of these factors with dietary preference. In: D.E. Russell, J.-P. Santoro, D. Sigogneau-Russell (Editors), Teeth Revisited: Proceedings of the VIIth International Symposium on Dental Morphology, Paris, 1986. Memoirs de Musee d’Histoire naturelle, Paris, Series C: 397-387.

Janis, C.M., Damuth, J., and Theodor, J.M., 2002. The origins and evolution of the North American grassland biome: the story from the hoofed mammals. Palaeogeogr., Palaeoclimatol., Palaeoecol., 177: 183-198.

Jenkins, F.A. and Krause, D.W., 1983. Adaptations for climbing in North American multituberculates. Science, 220: 712-715.

Kaiser, T.M. and Solounias, N., 2003. Extending the tooth mesowear method to extinct and extant equids. Geodiversitas, 25: 321-345.

Kaiser, T.M.; Solounias, N., Fortelius, M., Bernor, R.L.; and Schrenk, F., 2000. Tooth mesowear analysis on Hippotherium primigenium from the Vallesian Dinotheriensande (Germany): a blind test study. Carolinea, 58: 103-114.

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

Koch, P.L., Hoppe, K.A., and Webb, S.D., 1998. The isotopic ecology of late Pleistocene mammals in North America, Part 1, Florida. Chemical Geol., 152: 119-138.

Koch, P.L., Zachos, J.C., and Gingerich, P.D., 1992. Correlation between isotope records in marine and continental carbon reservoirs near the Palaeocene/Eocene boundary. Nature, 358: 319-322.

Kohn, M.J., 1996. Predicting animal δ18O: Accounting for diet and physiological adaptation. Geochim. Cosmochim. Acta, 60: 4811-4829.

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 Geol., 152: 97-112.

Kohn, M.J., Schoeninger, M.J. and Valley, J.W., 1996. Herbivore tooth oxygen isotope compositions: effects of diet and physiology. Geochim. Cosmochim. Acta, 60: 3889-3896.

Kovalevsky, W., 1873. Sur l’Anchitherium aurelianense Cuet sur l’histoire paleontologique des chevaux. Mem. Acad. Imp. Sci. St. Petersburg, Ser. 7, 20: 1- 73.

Krause, D.W. and Maas, M.C., 1990. The biogeographic origins of late Paleocene-early Eocene mammalian immigrants to the Western Interior of North America. In: T.M. Brown and K.D. Rose (Editors), Dawn of the Age of Mammals in the Northern Part of the Rocky Mountain Interior. Geological Society of America Special Paper, 243: 71-105.

Krishtalka, L., West, R.M., Black, C.C., Dawson, M.R., Flynn, J.J., Turnbull, W.D., Stucky, R.K., McKenna, M.C, Bown, T.M., Golz, D.J., and Lillegraven, J.A., 1987. Eocene (Wasatchian through Duchesnean) biochronology of North America. In: Woodburne, M.O. (Ed.), Cenozoic Mammals of North America: Geochronology and Biostratigraphy. University of California Press, Berkeley and Los Angeles.

Lander, E.B., 1998. Oreodontoidea In: Janis, C., Scot, K.M., and Jacobs, L.L. (Eds), Evolution of Tertiary Mammals of North America, 1: Terrestrial Carnivores, Ungulates, and Ungulatelike Mammals. University Press, New York, 402-420.

Lawver, L.A. and Gahagan, L.M., 2003. Evolution of Cenozoic seaways in the circum- Antarctic region. Palaeogeogr., Palaeoclimatol., Palaeoecol., 198: 11-37.

Lee-Thorpe, J.A., Sealy, J.C., and van der Merwe, N.J., 1989. Stable carbon isotope ratio differences between bone collagen and bone apatite, and their relationship to diet. J. Archaeological Sci., 16: 585-599.

Leopold, E. and MacGinitie, H.D., 1972. Development and affinities of Tertiary floras in the Rocky Mountains. In: Graham, A. (Ed.), Floristics and Paleofloristics of Asia and Eastern North America. Amsterdam, Elsevier: 147-200.

Linder, H.P., 1986. The evolutionary history of the Poales/Restionales – a hypothesis. Kew Bulletin, 42: 297-318.

Lindsay, E., 1972. Small mammal fossils from the Barstow Formation, California. University of California Publications in Geological Sci., 93, 1-104.

Litke, R., 1968. Uber den nachweis tertiarer Gramineen. Monatsber. Duet. Akad. Wiss. Berlin, 19: 462-471.

Lofgren, D.L., Lillegraven, J.A., Clemens, W.A., Gingerich, P.D. and Williamson, T.E., 2004. Paleocene biochronology: The Puercan through Clarkforkian land mammal ages. In: Woodburne, M.O. (Ed.), Late Cretaceous and Cenozoic mammals of North America: biostratigraphy and geochronology. Columbia University Press, New York: 43-105.

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

Luz, B. and Kolodny, Y., 1989. Oxygen isotope variation in bone phosphate. Applied Geochemistry, 4, 317-323.

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

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

MacFadden, B.J. and Cerling, T.E., 1996. Mammalian herbivore communities, ancient feeding ecology, and carbon isotopes: a 10 million-year sequence from the Neogene of Florida. J. Vertebrate Paleontol., 16: 103-115.

MacFadden, B.J. and Higgins, P., 2004. Ancient ecology of 15-million-year-old browsing mammals within C3 plant communities from Panama. Oecologia, 140: 169-182.

MacFadden, B.J., Solounias, N., and Cerling, T.E., 1999. Ancient diets, ecology, and extinction of 5-million-year-old horses from Florida. Science, 283: 824-827.

MacGinitie, H. D., 1953. Fossil plants of the Florissant beds, Colorado. Carnegie Institution of Washington Publication, 599: 198 pp.

MacGinitie, H.D., 1962. The Kilgore flora; A Late Miocene flora from northern Nebraska. University of California Publications in Geological Sci., 35: 67-158.

Marino, B.D, McElroy, M.B., Salawitch, R.J., and Spaulding, W.G., 1992. Glacial-to interglacial variations in the carbon isotopic composition of atmospheric CO2. Nature, 357: 461-466.

Matthew, W.D., 1932. A review of the rhinoceroses with a description of Aphelops [sic] material from the Pliocene of Texas. University of California Publications, Bulletin of the Department of Geological Sci., 20: 411-480.

McClellan, G.H. and Kauwenbergh, S.J., 1990. Mineralogy of sedimentary apatites. In: A.J.G. Notholt and I. Jarvis (Editors), Phosphorite Research and Development. Geological Society Special publication, London: 23-31.

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

McKenna, M.C., 1972. Possible biological consequences of plate tectonics. BioScience, 22: 519-525.

Miller, K.G. and Katz, M.E., 1987. Oligocene to Miocene benthic foraminiferal and abyssal circulation changes in the North Atlantic. Micropaleontol., 33: 97-149.

Miller, K.G., Wright, J.D., and Fairbanks, R.G., 1991. Unlocking the icehouse: Oligocene-Miocene oxygen isotopes, eustasy, and margin erosion. J. Geophysical Research, 96: 6829-6848.

Morgan, G.S., 1994. Miocene and Pliocene marine mammal faunas from the Bone Valley Formation of central Florida. In: A. Berta and T.A. Deméré (Editors), Contributions in Marine Mammal Paleontology Honoring Frank C. Whitmore, Jr. Proceedings of the San Diego Society of Natural History, 29: 239-268.

Morgan, M.E., Kingston, J.D., and Marino, B.D., 1994. Carbon isotope evidence for the emergence of C4 plants in the Neogene from Pakistan and Kenya. Nature, 367: 162- 165.

Newesely, H., 1989. Fossil bone apatite. Applied Geochemistry, 4: 233-245.

Novacek, M.J., 1999. 100 Million years of land vertebrate evolution: the Cretaceous– early Tertiary transition. Annals of the Missouri Botanical Garden, 86: 230-258.

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

Olson, E.C., 1952. The evolution of a Permian vertebrate chronofauna. Evol., 6: 181-196.

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

Passey, B.H. and Cerling, T.E. Tooth enamel mineralization in ungulates; implications for recovering a primary isotopic time-series., 2002. Geochim. Cosmochim. Acta, 66: 3225-3234.

Patterson, B., 1949. Rates of evolution in taeniodonts. In: G.L. Jepsen, E. Mayr, and G.G. Simpson (Editors), Genetics, Paleontology, and Evolution. Princeton University Press, Princeton: 243-278.

Prothero, D.R., 1986. A new Oromerycid (Mammalia, Artiodactyla) from the early Oligocene of Montana. J. Paleontol., 60: 458-465.

Prothero, D.R., 1992. Fifty million years or rhinoceros evolution. In: Ryder, O.A. (Ed.), Rhinoceros biology and conservation. Zoological Society, San Diego: 82-91.

Prothero, D.R., 1994. The Eocene-Oligocene Transition: Paradise Lost. Columbia University Press, New York: 291 pp.

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

Raymo, M.E., 1994. The Himalayas, organic carbon burial, and climate in the Miocene. Paleoceanography, 9: 399-404.

Rensberger, J.M., 1971. Entoptychine pocket gophers (Mammalia, Geomyoidea) of the Early Miocene John Day Formation, Oregon. University of California Publications in Geological Sci., 90: 1-163.

Rensberger, J.M., 1973. Pleurolicine rodents (Geomyoidae) of the John Day Formation, Oregon and their relationships to taxa from the Early and Middle Miocene, South Dakota. University of California Publications in Geological Sci., 102: 1-95.

Retallack, G.J., 1997. Neogene expansion of North American prairie. Palaios, v., 12: 380- 390.

Retallack, G.J., 2001. Cenozoic Expansion of Grasslands and Climatic Cooling. J. Geol., 109: 407-426.

Retallack, G.J., 2004. Late Oligocene bunch grassland and early Miocene sod grassland paleosols from central Oregon, USA. Palaeogeogr., Palaeoclimatol., Palaeoecol., 207: 203-237.

Rivals, F. and Deniaux, B., 2003. Dental microwear analysis for investigating the diet of an argali population (Ovis ammon antiqua) of mid-Pleistocene age, Caune de l’Arago Cave, eastern Pyrenees, France. Palaeogeogr., Palaeoclimatol., Palaeoecol., 193: 443-455.

Rose, K.D., 1981. The Clarkforkian land mammal age and mammalian faunal composition across the Paleocene-Eocene boundary. University of Michigan Papers in Paleontol., 26: 1-196.

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

Savage, D.E. and Russell., 1983. Mammalian paleofaunas of the worlds. Addison-Wesley Publishing Co., Reading, Mass.

Savin, S.M., 1977. The history of the earth’s surface temperature during the past 100 million years. Annual Review of Earth and Planetary Sci., 5: 319-355.

Scott, T.M., 1988. The lithostratigraphy of the Hawthorn Group (Miocene) of Florida. Florida Geological Survey Bulletin, 59: 1-48.

Scott, W.B., 1937. A history of land mammals in the western hemisphere. Macmillan, New York: 786 pp.

Sears, P.B., 1969. Lands Beyond the Forest. Prentice Hall, New York: 206 pp.

Simons, E.L., 1960. The Paleocene Pantodonta. Transactions of the Am. Philosophical Society, 50: 1-99.

Simpson, G.G., 1937. The Fort Union Crazy Mountain Field, Montana, and its mammalian faunas. Bulletin of the United States National Museum, 169: 1-287.

Solounias, N. and Hayek, L-A. C., 1993. New methods of tooth microwear analysis and application to dietary determination of two extinct antelopes. J. Zool., London. 229: 421-445.

Solounias, N. and Semprebon, G.M., 2002. Advances in the reconstruction of ungulate ecomorphology with applications to early fossil equids. Am. Museum Novitates, 3366: 1-49.

Stebbins, G.L., 1981. Coevolution of grasses and herbivores. Annals of the Missouri Botanical Garden, 68: 75-86.

Sternberg, L.S.L., Mulkey, S.S. and Wright, S.J., 1989. Oxygen isotope ratio stratification in a tropical moist forest. Oecologia, 81: 51-56.

Strömberg, C.A.E., 2002. The origin and spread of grass-dominated ecosystems in the late Tertiary of North America: preliminary results concerning the evolution of hypsodonty. Palaeogeogr., Palaeoclimatol., Palaeoecol., 177: 59-75.

Strömberg, C.A.E., 2004. Using phytolith assemblages to reconstruct the origin and spread of grass-dominated habitats in the great plains of North America during the late Eocene to early Miocene. Palaeogeogr., Palaeoclimatol., Palaeoecol., 207: 239- 275.

Stucky, R., 1990. Evolution of land mammal diversity in North America during the Cenozoic. Current Mammal., 2: 375-432.

Teaford, M.F., 1988. A review of dental microwear and diet in modern mammals. Scanning Micrscopy, 2: 1149-1166.

Teaford, M.F. and Robinson, J.G., 1987. Diet and dental microwear in Cebus nigrivittatus. Am. J. Physical Anthropol., 73: 261-262.

Teaford, M.F. and Walker, A., 1984. Quantitative differences in dental microwear between primate species with different diets and a comment on the presumed diet of Sivapithecus. Am. J. Physical Anthropol., 64: 191-200.

Tedford, R.H., Skinner, M.F., Fields, R.W., Rensberger, J.M., Whistler, D.P. Galusha, T., Taylor, B.E., Macdonald, J.R. and Webb, S.D., 1987. Faunal succession and biochronology of the Arikareean through Hemphillian interval (late Oligocene through earliest Pliocene epochs) in North America. In: M.O. Woodburne (Editor), Cenozoic Mammals of North America: Geochronology and biostratigraphy. University of California Press, Berkeley: 153-210.

Theodor, J.M., Janis, C.M., and Broekhuizen, L., 1998. Chronofaunas and ungulate turnover in the Cenozoic of North America. J. Vertebrate Paleontol., 18: 81A.

Truswell, E.M. and Harris, W.K., 1982. The Cainozoic palaeobotanical record in arid Australia: Fossil evidence for the origins of an arid-adapted flora. In: W.R. Barker and P.J.M. Greenslade (Editors), Evolution of the Flora and Fauna of arid Australia. Peacock Publications, Frewville, Australia: 57-76.

Van Houten, F.B., 1945. Review of latest Paleocene and Early Eocene mammalian faunas. J. Paleontol., 19: 421-461.

Vennemann, T.W., Hegner, E., Cliff, G., and Benz, G.W., 2001. Isotopic composition of recent shark teeth as a proxy for environmental conditions. Geochim. Cosmochim. Acta, 65: 1583-1599.

Voorhies, M.R., 1981. Dwarfing the St. Helens eruption: ancient ashfall creates a Pompeii of prehistoric animals. National Geographic, 159: 66-75.

Voorhies, M.R. and Thomasson, J.R., 1979. Fossil grass anthoecia within Miocene rhinoceros skeletons: diet in an extinct species. Science, 206: 331-333.

Vincent, E. and Berger, W.H., 1985. Carbon dioxide and cooling in the Miocene; the Monterey hypothesis. In: W.S. Broecker and E.T. Sundquist (Editors), The Carbon Cycle and Atmospheric CO2: Natural Variations Archean to Present. GeoPhysical Monograph, 32: 455-468.

Vincent, E., Killingley, J.S., and Berger, W.H., 1985. Geological Society of America Memoirs, 163: 103-130.

100

Wahlert, J.H., 1973. Protoptychus, a hystricomorphous rodent from the Late Eocene of North America. Breviora, 419: 1-14.

Wall, W.P., 1982. Evolution and biogeography of the Amynodontidae (Perissodactyla, Rhinocerotoidea). Third North American Paleontological Convention, Proceedings, 2: 563-567.

Wang, Y. and Cerling, T.E., 1994. A model of fossil tooth and bone diagenesis: implications of paleodiet reconstruction from stable isotopes. Palaeogeogr., Palaeoclimatol., Palaeoecol., 107: 281-289.

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. Palaeogeogr., Palaeoclimatol., Palaeoecol., 107: 269-279.

Webb, S.D., 1977. A History of Savanna Vertebrates in the New World: Part I. North America: Annual Reviews of Ecol. and Systematics, 8: 355-380.

Webb, S.D., 1983. The Rise and Fall of the Late Miocene Ungulate Fauna in North America. In: M.N. Nitecki (Editor), Coevolution. University of Chicago Press, Chicago: 267-306.

Webb, S.D., 1989. The fourth dimension in North American terrestrial mammal communities. In: D.W. Morris (Editor), Patterns in the structure of mammalian communities. Special Publications, Museum of Texas Technical University: 181- 203.

Webb, S.D., Morgan, G.S., Hulbert, R.C., Jr., Jones, D.S., MacFadden, B.J., and Mueller, P.A., 1989. Geochronology of rich early Pleistocene vertebrate fauna, Leisey Shell Pit, Tampa Bay, Florida. Quaternary Research, 32: 96-110.

Webb, S.D. and Opdyke, N.D., 1995. Global climatic influence on Cenozoic land mammal faunas. In: S.M. Stanley (Editor), Effects of past global change on life. National Academy Press, Washington, D.C.: 184-208.

Wheeler, W.H., 1961. Revision of the uintatheres. Bulletin of the Peabody Museum of Natural History, 14: 1-93.

Wing, S.L., 1980. Fossil floras and plant-bearing beds of the Central Bighorn Basin. In: P.D. Gingerich (Editor), Early Cenozoic Paleontology and Stratigraphy of the Bighorn Basin. University of Michigan Papers in Paleontol., 24: 119-125.

Wing, S.L., Alroy, J. and Hickey, L.J., 1995. Plant and mammal diversity in the Paleocene to early Eocene of the Bighorn Basin. Palaeogeogr., Palaeoclimatol., Palaeoecol., 115: 117-155.

Wing, S.L. and Tiffney, B.H., 1987. The reciprocal interaction of angiosperm evolution and tetrapod herbivory. Reviews of Palaeobotany and Palynol., 50: 179-210.

101

Wolfe, J.A., 1978. A paleobotanical interpretation of Tertiary climates in the Northern hemisphere. Am. Scientist, 66: 694-703.

Wolfe, J.A., 1985. Distribution of major vegetational types during the Tertiary. In: E.T. Sundquist and W.S. Broecker (Editors). Geophysical Monograph 32, Am. Geophyisical Union, Washington, D.C.: 357-374.

Wood, A.E., 1962. The Early Tertiary rodents of the family Paramyidae. Transactions of the Am. Philosophical Society, 52: 1-261.

Woodburne, M.O., Tedford, R.H., and Swisher, C.C., III., 1990. Lithostratigraphy, biostratigraphy, and geochronology of the Barstow formation, Mojave Desert, southern California. Geological Society of America Bulletin, 102: 456-477.

Woodruff, F. and Savin, S.M., 1989. Miocene deepwater oceanography. Paleoceanography, 4: 87-140.

Woodruff, F. and Savin, S.M., 1991. Mid-Miocene isotope stratigraphy in the deep sea: high-resolution correlations, paleoclimatic cycles, and sediment preservation. Paleoceanography, 6: 755-806.

Wright, J.D. and Miller, K.G., 1993. Southern ocean influences on late Eocene to Miocene deepwater circulation. In: J.P. Kennet and D.A. Warnke (Editors), The Antarctic paleoenvironment: A perspective on global change. Antarctic Research Series, 60: 1-25.

Wright, J.D., Miller, K.G., and Fairbanks, R.G., 1992. Early and middle Miocenes stable isotopes: Implications for deepwater circulation and climate. Paleoceanography, 7: 357-389.

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.

Zachos, J.C., Stott, L.D., and Lohmann, K.C., 1994. Evolution of early Cenozoic marine temperatures. Paleoceanography, 9: 353-387.

Zhow, G.-T. and Zheng, Y.-F., 2002. Kinetic mechanism of oxygen isotope disequilibrium in precipitated witherite and aragonite at low temperatures: an experimental study. Geochim. Cosmochim. Acta, 66: 63-71.

BIOGRAPHICAL SKETCH

Jonathan Hoffman was born on June 11, 1981 in Phoenix, Arizona. The oldest of three children, he grew up in Phoenix and graduated from North High School in 1999. He received his B.A. in geology from Occidental College in 2003. Jonathan received his

M.S. in the geological sciences from the University of Florida in 2006. He plans to pursue a doctorate degree and continue research in vertebrate paleontology.

102