A Craniodental Interpretation of the Dietary Habits of Poebrotherium wilsoni (Camelidae) from the Oligocene of Badlands National Park, South Dakota

WILLIAM P. WALL AND JACQUELINE M. HAUPTMAN

Dept. of Biology, Georgia College & State University, Milledgeville, GA 31061

____________________

Abstract— The craniodental morphology of the Oligocene camelid, Poebrotherium wilsoni was analyzed in an effort to determine the likely dietary habits of this species. Craniodental indices shown to correlate with dietary habits were applied to modern species of known dietary habits (browser, mixed feeder and grazer) and compared to Poebrotherium wilsoni. Hypsodonty index, relative muzzle width and relative length of the upper premolar row were utilized for all groups. Tooth wear was examined qualitatively and similar comparisons were made. Camelus dromedarius was separately subjected to the same analysis, and compared to Poebrotherium wilsoni to address differences in the two forms. The comparisons revealed that Poebrotherium wilsoni was not a grazer. It was more similar to the browsers and mixed feeders in hypsodonty, length of premolar row, and tooth topography and wear. Relative muzzle width was the only index in which Poebrotherium wilsoni was similar to the grazers. Locality data from the literature and our fieldwork suggests that Poebrotherium wilsoni ranged over a variety of habitats, further substantiating the probability of broad dietary preferences in this early camelid. Camelus dromedarius, a mixed feeder, was significantly different from Poebrotherium wilsoni in all indices.

____________________

INTRODUCTION

Camelid dentition is primitive among living seleno-dont artiodactyls in the retention of upper lateral incisors and three premolars. The three-chambered ruminating stomach found in camelids is also less efficient than that of the true ruminants with their four-chambered stomachs. The refugial distribution (Africa and South America) and specialized habitats (deserts and high mountains) of modern camelids is probably the result of their inability to compete effectively with advanced ruminants (Wheeler, 1995). The fossil record of camelids indicates that they had a much broader geographic range for most of their history. Modern camelids have proportionally long limbs, digitigrade feet, and a pacing gait that makes them efficient travelers in open terrain. The locomotor abilities of camelids therefore appear to be at least comparable to those of most ruminants. A possible cause for the inability of camelids to compete effectively with ruminants is lower efficiency at food processing. It is commonly believed that the development of the ruminant stomach was responsible for the decline of non-ruminant large herbivores. However, MacFadden (1992) does not believe this factor alone was responsible for the decline in diversity of horses. If that is true, then the history of camelids may be more complicated as well.

Poebrotherium wilsoni (Figure 1) from the Orellan of North America is the most well known early camelid, and a member of the central stock from which modern camels and llamas have evolved. Skeletal remains indicate P. wilsoni was a lightly built artiodactyl measuring approximately 75 cm in height. It was unique among its contemporaries in having a long slender neck and limbs that were well suited for efficient running. The dental characteristics of Poebrotherium are primitive for the family Camelidae.

The purpose of this paper is to describe the likely dietary habits of Poebrotherium wilsoni in order to better understand the selection pressures acting early in camelid evolution.
lateral view of the skull and jaws of Poebrotherium wilsoni
Figure 1. Lateral view of the skull and jaws of Poebrotherium wilsoni (GCVP 2119). Scale bar equals two centimeters.

CAMELID SYSTEMATICS

Camelids (suborder Tylopoda) underwent nearly forty million years of evolution in North America (McKenna, 1966; Schultz, 1959; Scott, 1929, 1940; Wilson, 1974; Wortman, 1898). During the Plio-Pleistocene (3 MYBP), some species of the llama-like genus Hemiauchenia migrated to South America via the Panamanian Land Bridge and gave rise to Llama and Vicugna about 2 MYBP (Webb; 1965, 1972, 1974). Contemporaneously, other species, possibly in the genus Paracamelus, migrated from North America to Eurasia via the Beringia Land Bridge ultimately giving rise to the two Old World camels, Camelus dromedarius and C.bactrianus (Gauthier-Pilthers and Dagg, 1981). Camels became extinct in North America as recently as 10,000-12,000 years ago with Camelops surviving well into the Pleistocene (Harrison, 1985; Schultz, 1959; Webb, 1965, 1972, 1974).

Poebrotherium has long been recognized as an early member of the central stock from which modern camels and llamas evolved, however, not all authors agree on the ancestor of Poebrotherium (see for example, Gazin, 1955; Scott, 1940; Wilson, 1974). The poor fossil record of camelids prior to Poebrotherium complicates the situation. Wortman (1898) placed the Uintan artiodactyl, Protylopus petersoni, directly ancestral to Poebrotherium. Gazin (1955), however, believed Poebrodon kayi, also from the Uintan, was the direct ancestor to Poebrotherium. Gazin placed Protylopus in the family Oromerycidae. Wilson (1974) in his description of a new species of Poebrotherium, P. franki, from the Chadronian stated that P. franki was more closely related to Protylopus than to Poebrodon. The most recent systematic revision of camelids (Honey et al. 1998) questions the placement of P. franki in the genus Poebrotherium, regards Poebrodon to be the most primitive camelid, and places Poebrotherium as the sister taxon to all higher camelids.

Four species of Orellan Poebrotherium have been named (Scott, 1940): P. wilsoni, Leidy, 1847, lower Brule; P. labiatum, Cope, 1881, lower Brule; P. andersoni, Troxell, 1917, lower Brule; and P. eximium, Hay, 1902, base of the lower Brule. Loomis (1928) considered the size disparity and the difference in the anterior dentition between Poebrotherium labiatum and P. wilsoni as sexual differences. He synonymized P. labiatum and P. andersoni with P. wilsoni. Loomis accepted P. eximium as a distinct species because it lacks the diastema between the canine and premolar teeth characteristic of P. wilsoni.

Scott (1940) agreed with Loomis' assessment that P. andersoni was not a valid species. However, he questioned Loomis' (1928) failure to separate P. wilsoni and P. labiatum in light of the striking skeletal differences between the two species.

Prothero (1996) recognized P. wilsoni, P. eximium, P. franki, and he named a new species, P. chadronense. All of the specimens used in this study (from the Orellan of Badlands National Park) are P. wilsoni.

McKenna (1966) and Webb (1972) recognized four genera of late Oligocene camelids; Pseudolabis with no known descendants, Paratylopus the ancestral long-necked camelid, Dyseotylopus the basal representative of the modern camelids , and an unknown ancestor to the stenomyline camels. Prothero (1996) showed that Dyseotylopus is a junior synonym of Miotylopus.

MATERIALS AND METHODS

Data were collected on both fossil and modern specimens. The fossil specimens are housed in the Georgia College & State University Vertebrate Paleontology collection (GCVP); and the modern specimens are housed in the Georgia College & State University Mammal collection (GCM) or the University of Nebraska State Museum (UNSM). The specimens measured are listed in Appendix 1. All measurements were taken with Mitutoyo calipers with 0.05 mm accuracy. The mean values for each measurement were used to calculate each craniodental index. Significant differences in mean values were determined by implementing a t-test (p<0.05).

Dietary information was collected from the literature (Janis, 1988; Janis and Ehrhardt, 1988; Nowak, 1991; Solounias and Dawson-Saunders, 1988; Solounias et al., 1988). Specifically, Estes (1972) was used for the dietary information for Gazella granti. Grazers feed primarily on grasses (> 90% of diet is grass); browsers feed primarily on fruit, leaves and young shoots (<10% of diet is grass); mixed feeders feed on a variety of material (10-90% of diet is grass). Mixed feeders may vary their feeding seasonally or they may select different habitats throughout the year. The modern specimens were grouped into their respective feeding categories: browsers, mixed feeders and grazers based on dietary descriptions in the literature.

The following craniodental indices were employed: relative muzzle width (RMW), relative length of the upper premolar row (RLUPR) and hypsodonty index (HI). Relative muzzle width was calculated by dividing the mean palatal width by the mean muzzle width (Figure 2). Palatal width (PAW) was measured as the distance between the upper second molars on the lingual side at the level of the protocone. Muzzle width (MZW) was measured as the maximum outer distance of the junction of the premaxillary and maxillary bones. Relative length of the upper premolar row was calculated by dividing the mean premolar row length by the mean molar row length. Premolar row length (PRL) and molar row length (MRL) were measured as the length along the base of the crown on the labial side. Hypsodonty index was calculated by dividing the mean of the unworn height of the second molar by the mean of the width of the second molar. The height was measured on the labial side from the base of the crown to the apex of the protoconid. The width was measured as the distance between the protoconid and the metaconid.

The source for the hypsodonty data for all groups except Poebrotherium was Janis (1988). Relative muzzle width and relative length of the upper premolar row data were collected from multiple specimens of the same species. Familiarization of head musculature was made possible by the dissection of a fresh specimen of Antilocapra americana, housed in the Georgia College & State University Mammal collection.

Dental wear facets were examined to reveal clues about the types of food comminuted and about jaw movement during mastication (Butler, 1972; Costa and Greaves, 1981; Fortelius, 1985; Greaves, 1973, 1991; Janis, 1990, 1995; Rensberger et al, 1984).
ventral view of skull illustrating cranial measurements taken

Figure 2. Ventral view of skull (GCVP 2679) illustrating cranial measurements taken. Abbreviations: MZW, muzzle width; PAW, palatal width.

RESULTS

Table 1.Data summary for cranial and dental indices. The numbers given are the mean, the number of observations, and the standard deviation. Abbreviations: HI, hypsodonty index; RMW, relative muzzle width; RLUPR, relative length of upper premolar row.
Taxon
HI
RMW
RLUPR
Poebrotherium
1.61(5) 0.03
1.30(3) 0.08
1.18(5) 0.17
Browsers
1.82(3) 0.73
1.96(5) 0.54
0.85(4) 0.15
Mixed Feeders
3.02(4) 1.42
1.67(7) 0.23
0.71(7) 0.18
Grazers
4.34(6) 0.78
1.27(8) 0.12
0.58(5) 0.05
Camelus
2.52(4) 0.30
1.83(3) 0.14
0.55(3) 0.05

A summary of the data is presented in Table 1. The results obtained from comparing the modern taxa are presented first. The following results were obtained after test ing for significant differences in the mean values of each index (HI, RMW, and RLUPR) for each dietary group (browser, mixed feeder and grazer.)

The hypsodonty results showed that browsers (1.82) were significantly (p<0.001) less hypsodont than grazers (4.34). Mixed feeders (3.02) were intermediate in hypsodonty, but not significantly less hypsodont than grazers (4.34) or significantly more hypsodont than browsers (1.82). Therefore, browsers were distinguished from grazers, but mixed feeders were not distinguished from browsers or grazers.

Study of relative muzzle width produced the following results. Grazers (1.27) had a significantly broader muzzle than either browsers (1.96) or mixed feeders (1.70) p<0.001 and p<0. 01, respectively. Browsers had a narrower muzzle (1.96) than mixed feeders (1.70), but not significantly nar
rower. Therefore, grazers were distinguished from mixed feeders and browsers, but browsers were not distinguished from mixed feeders.

Analysis of the relative length of the upper premolar row produced the following results. Browsers (0.85) had a significantly (p<0.01) longer premolar row than grazers (0.58); mixed feeders (0.71) were intermediate in premolar row length, but not significantly longer than grazers (0.58) nor significantly shorter than browsers (0.85). Therefore, browsers were distinguished from grazers, but mixed feeders were not distinct from browsers or grazers.

The following results were obtained after testing for significant differences between Poebrotherium and each dietary group. Poebrotherium was the least hypsodont (HI=1.61) of all taxa. It was significantly (p<0.001) less hypsodont than the grazers (mean HI=4.34). It was not significantly less than the mixed feeders (but this may be an artifact of the small sample size since the mixed feeder mean is HI=3.02), nor was it significantly less than the browsers (mean HI=1.82). Poebrotherium had a relatively broad muzzle (RMW=1.30), which is significantly (p<0.05) broader than the mixed feeders (mean RMW=1.67). However, Poebrotherium did not test significantly broader than the narrowest of all groups, the browsers (mean RMW=1.96), but this is probably due to the small sample size. Relative length of the upper premolar row was the greatest in Poebrotherium (1.18). It tested significantly greater than the browsers (mean RLUPR=0.85, p<0.02), the mixed feeders (mean RLUPR=0.71, p<0.001) and the grazers mean (mean RLUPR=0.58, p>0.01). In comparative terms, Poebrotherium can be described as having a broad muzzle, a long upper premolar row and low crowned molars.

Camelus was analyzed separately from the mixed feeders. Camelus was not significantly more or less hypsodont (HI=2.52) than the browsers (mean HI=l.82) or the mixed feeders (mean HI=3.02), respectively. Camelus was significantly (p<0.01) less hypsodont than the grazers (mean HI=4.34). Camelus was not significantly more or less narrow in muzzle width (RMW=1.83) than the mixed feeders (mean RMW=1.67) or the browsers (mean RMW=1.96), respectively. Camelus was significantly (p<0.001) narrower in muzzle width than the grazers (mean RMW=1.27). Camelus had a significantly (p<0.05) shorter premolar row (RLUPR=0.55) than the browsers (mean RLUPR=0.85). Camelus did not have a significantly shorter premolar row than either the mixed feeders (mean RLUPR=0.71) or the grazers (mean RLUPR=0.58). For all indices, Camelus was not significantly different from the mixed feeders.
A, upper and B, lower dentition of Poebrotherium wilsoni
Figure 3. A, upper and B, lower dentition of Poebrotherium wilsoni (GCVP 2679 and 3819). Scale bar equals one centimeter.

Poebrotherium was significantly different from Camelus for all three indices. Poebrotherium was less hypsodont (1.61 vs. 2.52 p<0.02), in possession of a broader muzzle (1.3 vs. 1.83 p<0.01) and had a longer premolar row (1.18 vs. 0.55 p<0.001).

Dental wear of all the groups was compared, and the following observations summarize the comparisons. The mixed feeders, and to a lesser degree, the browsers, exhibited similar patterns of wear. On the lingual slopes of the endolophs, there were small ovoid facets with fine, parallel striations. Tragulus sp. deviated slightly from the other browsers, as it exhibited wear more similar to the mixed feeders. The browsers and the mixed feeders were more similar in having greater occlusal relief, and in abrasion wear.

Examination of the wear facets of Poebrotherium (Figure 3) produced the following results. The direction of the jaw movement was inferred by inspecting the enamel-dentine interface. The ectoloph of the upper molar exhibited wear frequently observed in living selenodont artiodactyls. Namely, there was a smooth, uninterrupted enamel-dentine interface at the leading edge and an uneven, abrupt enamel-dentine interface at the trailing edge. The striations etched on the wear facets were directed perpendicular to the enamel-dentine ridges. Wear facets on the slopes of the endoloph were ovoid in shape, and they were positioned more vertically than horizontally. Abrasion wear was characteristic of these teeth.

The wear pattern of Poebrotherium molars was similar to the wear of the mixed feeders, specifically, Hydropotes inermis and Muntiacus muntjak. It should be noted that the browsers and mixed feeders used in this study are all comparable in size to Poebrotherium.

DISCUSSION

Poebrotherium wilsoni is a relatively rare member of the ungulate fauna from the Scenic Member of the Brule formation. The relative abundance of this taxon increases significantly to the south and west of Badlands National Park (Honey et al. 1998). The paleoecology of the Scenic Member has been described from fossil soil studies (Retallack, 1983). Paleosol evidence depicts the change from a warm, humid Eocene (55-35 MYIBP) to an increasingly arid Oligocene (35-25 MYBP) marked with episodes of severe drought. Fossil soils also indicate that woodlands were widespread during the early Oligocene with interspersed streams or marsh areas. The vegetation during the deposition of the Scenic Member can be divided into three predominant habitats: swampy streamside swales, with early successional vegetation, an extensive woodland bordering the streams, and a savanna habitat. The savanna habitat was populated with bunch grass and small forbs.

This study was undertaken to determine the likely dietary habits of Poebrotherium wilsoni compared to modern selenodont artiodactyls of known dietary habits. For all three indices examined there were significant differences between the browsers and the grazers. The grazers and the mixed feeders were significantly different in only one index, relative muzzle width. The mixed feeders and the browsers did not differ significantly for any index. Qualitatively, the tooth wear differed for each group, but there were greater similarities between the browsers and mixed feeders.

The hypsodonty index for Poebrotherium wilsoni (1.61) was least like that of the grazers (4.34). It was not significantly lower than the mixed feeders (3.02), but this is probably an artifact of the sample size. The hypsodonty index for Poebrotherium was most similar to the browsers (1.82) as a group. Comparisons with individual taxa are possibly more meaningful. Poebrotherium has a hypsodonty index particularly close to Tragulus sp. (HI=1.40), a browser, and two mixed feeders Hydropotes (HI=1.84) and Muntiacus (HI=1.81). Tragulus, the mouse deer lives in wooded areas and feeds on the leaves of bushes, fallen fruit and some grass. Hydropotes, the Chinese water deer, feeds on a variety of reeds, grasses, and vegetables near streams. Muntiacus, the barking deer, lives in dense forests where it feeds on shrubs, grasses, and ripe fruits.

Poebrotherium had the longest premolar row (1.18) of any group measured. Tragulus sp. (1.04) was the only modern species that was nearly as long. Poebrotherium exhibited abrasion wear indicative of a diet consisting of at least in part, pulpy food items. Dental wear in Poebrotherium was most similar to the mixed feeders Hydropotes and Muntiacus and to a lesser degree to the browser Tragulus.

In summary, Poebrotherium wilsoni shared more similarities with the browsers and the mixed feeders. From this analysis, it can be reasonably suggested that Poebrotherium wilsoni was a browser or a mixed feeder. The difficulty in further clarification rests in the inability to distinguish mixed feeders from browsers with these methods.

Locality data from the literature suggested that Poebrotherium frequented a variety of habitats. Clark et al. (1967) collected specimens of Poebrotherium from all three principal habitats represented in the deposits (open plain, near-stream woodland, and stream-side swales). Because of its locomotor adaptations, Clark et al. expected Poebrotherium to be common in the deposits interpreted as open plains, but they were not expecting to find this taxon in the other habitats. Our extensive fieldwork in the Orellan of the Southern Unit of Badlands National Park confirms the widespread nature of this taxon.

Poebrotherium wilsoni is commonly referred to in paleontological literature for its cursorial adaptations (Webb, 1972). The limbs were elongated, and the front limb was nearly as long as the hind limb. Additionally, the lateral toes were reduced to tiny splints (Scott, 1940; Webb, 1972). These limb modifications are reminiscent of modern cursorial ungulates living in open plains. However, the occurrence of Poebrotherium wilsoni in a variety of habitats does not support a strong preference for any one habitat. The cursorial adaptations of this early Oligocene camelid and the later development of digitigrade feet and a pacing gait exhibited in the two Miocene camelids, Michenia and Protolabis, has been attributed to the success of the group during the transition from a woodland, savanna habitat to an open, grassland habitat. Later forms have been described as becoming more hypsodont with the coinciding spread of grasslands in the Miocene (Webb, 1972).

The modern camels, such as Camelus dromedarius, are more hypsodont (2.52) than Poebrotherium wilsoni (1.61), and modern camels are mixed feeders. However, the habitat of Camelus dromedarius, the Arabian camel or dromedary, is vastly different from the habitats occupied by Poebrotherium wilsoni, or the habitats of the other mixed feeders in this study. The dentition of the modern camel is moderately hypsodont, with molarization of premolars three and four, and a leveling of the occlusal surface. The combination of these dental features suggests a coarse diet that is verified through observation (Gauthier-Pilters and Dagg, 1981). In contrast, a low level of hypsodonty, a long premolar row and an occlusal surface of high relief characterize the dentition of Poebrotherium wilsoni. This combination of features suggests a diet consisting of pulpy food items.

CONCLUSIONS

Tooth height, muzzle width and premolar row length all have functional implications related to a particular dietary habit. Interpretation of the craniodental morphology and dental wear of Poebrotherium wilsoni has led to the conclusion that Poebrotherium was not a grazer. The results indicate a browsing or mixed feeding dietary habit. Modern browsers select a variety of succulent food items; fruit, tender shoots, shrubs, and some grasses. Mixed feeders also take these items, but more of their diet will consist of grasses, shrubs, and in the case of the Chinese water deer, rushes and reeds. Clark et al. (1967) collected Poebrotherium wilsoni in deposits that represent three principal habitats: open savanna; near steam woodland; and streamside swales. These habitats and plant communities parallel the habitats frequented by the modern species featured in this study. It is reasonable to suggest that Poebrotherium wilsoni may have made seasonal or daily movements selecting a variety of foods from any one of these habitats.

ACKNOWLEDGEMENTS

We thank Ms. Rachel Benton of Badlands National Park for her extensive support of our research efforts and Mr. Vince Santucci for his enthusiasm and support for paleontological research in the National Parks. Three anonymous reviewers provided useful comments on the manuscript. This research was partially funded by faculty research grants from Georgia College & State University.

REFERENCES

Butler, P.M., 1972. Some functional aspects of molar evolution. Evol. 26(3): 474-483.

Clark, J., J.R. Beerbower, and K.K. Kietzke, 1967. Oligocene sedimentation, stratigraphy, paleoecology, and paleoclimatology in the Big Badlands of South Dakota. Fieldiana Geology Memoirs 5:1-158.

Costa, R. and W.S. Greaves, 1981. Experimentally produced toothwear facets and the direction of jaw motion. Journal of Paleontology 55(3):635-638.

Estes, R.D., 1972. The comparative behavior of Grant's Gazelle and Thompson's Gazelle. Journal of Mammalogy 48:182-209.

Fortelius, M., 1985. Ungulate cheek teeth: developmental, functional, and evolutionary interrelations. Acta Zoologica Fennica 180:1-76.

Gauthier-Pilthers, H. and K.L. Dagg, 1981. The Camel: its evolution, ecology, behaviour, and relationship to man. University of Chicago Press: Chicago. 208 pp.

Gazin, C.L., 1955. A review of the upper Eocene artiodactyls of North America. Smithsonian Contributions to Paleontology 5:1-29.

Greaves, W. S., 1973. The inference of jaw motion from tooth wear facets. Journal of Paleontology 47:1000-1001.

_____, 1991. A relationship between premolar loss and jaw elongation in selenodont artiodactyls. Zoological Journal of the Linnean Society. 101:121-129.

Harrison, J.A., 1985. Giant camels from the Cenozoic of North America. Smithsonian Contributions to Paleontology. 5:1-29.

Honey, J.G., J.A. Harrison, D.R. Prothero, and M.S. Stevens, 1998. Camelidae. In: Evolution of Tertiary Mammals of North America Volume 1: Terrestrial Carnivores, Ungulates, and Ungulatelike Mammals (C.M. Janis, K.M. Scott, and L.L. Jacobs, eds.), p. 439-462.

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: Teeth Revisited (D.E. Russell, J.P. Santoro, and D. Sigogneau-Russell, eds.). Memoirs de Museum National d'Histoire Naturelle 53: 367-387.

_____, 1990. The correlation between diet and dental wear in herbivorous mammals, and its relationship to the determination of diets of extinct species. In: Paleobiological Evidence for Rates of Coevolution and Behavioral Evolution (A. J. Boucot, ed.). Elsevier: New York. Pp. 241-259

Janis, C.M., 1995. Correlation between craniodental morphology and feeding behavior in ungulates: Reciprocal illumination between living and fossil taxa. In: Functional Morphology in Vertebrate Paleontology (G. J. Thomason, ed.). Cambridge U. Press: Cambridge. Pp. 76-98.

Janis, C.M. and D. Ehrhardt, 1988. Correlation of relative muzzle width and relative incisor width with dietary preference in ungulates. Zoological Journal of the Linnean Society 92:267-284.

Loomis, F.B., 1928. Poebrotherium. American Journal of Science 16:137-142.

MacFadden, B.J., 1992. Fossil horses: Systematics, Paleobiology and Evolution of the family Equidae. Cambridge University Press. 369 p.

McKenna, M.C., 1966. Synopsis of Whitneyan and Arikareean camelid phylogeny. American Museum Novitates 2253:1-11.

Nowak, R. M., 1991. Walker's Mammals of the world. Johns Hopkins University Press: Maryland. 1629 pp.

Prothero, D.R., 1994. The Eocene-Ohgocene Transition. Columbia University Press: New York. 283 pp.

_____, 1996. Camelidae. In. The terrestrial Eocene-Oligocene transition in North America (D. R. Prothero and R. J. Emry eds.) Cambridge University Press: New York. Pp. 609-651.

Radinsky, L.B., 1986. Patterns in the evolution of ungulate jaw shape. American Zoologist 25:315-330.

Rensberger, L.B., A. Forsten., and M. Fortelius, 1984. Functional evolution of the cheek tooth pattern and chewing direction in Tertiary horses. Paleobiology 10(4):439-452.

Retallack, G.J., 1983. A paleopedological approach to the interpretation of terrestrial sedimentary rocks: the mid-Tertiary fossil soils of Badlands National Park, South Dakota. Geological Society of America Bulletin 94:823-840.

Schultz, B., 1959. The camel story. University of Nebraska Museum Notes no. 8.

Scott, W. B., 1929. History of Land Mammals in the Western Hemisphere. Macmillan Company, New York. 693 pp.

_____, 1940. Artiodactyla part IV. In: The Mammalian Fauna of the White River Oligocene (W. B. Scott and G. L. Jepsen, Eds.). Transactions of the American Philosophical Society 28:363-746.

Solounias, N. and B. Dawson-Saunders, 1988. Dietary adaptations and paleoecology of the Miocene ruminants from Pikermi and Samos in Greece. Palaeogeography, Palaeoclimatology, Palaeoecology 65:149-172.

Solounias, N., M. Tedford, and A. Walker, 1988. Interpreting the diet of extinct ruminants: the case of a non-browsing giraffid. Paleobiology 14:287-300.

Webb, S.D., 1965. The osteology of Camelops. Bulletin of Los Angeles County Museum, no. 1.

_____, 1972. Locomotor evolution in camels. Forma et Functio 5:99-112.

_____, 1974. Pleistocene llamas of Florida, with a brief review of the lamini. In: Pleistocene Mammals of Florida (S. D. Webb, ed.). University Press Florida; Gainesville.

Wheeler, J.C., 1995. Evolution and present situation of the South American Camelidae. Biological Journal of the Linnean Society 54:271-295.

Wilson, J.A., 1974. Early Tertiary vertebrate faunas, Vieja group and Buck Hill group, Trans-Pecos Texas: Protoceratidae, Camelidae, Hypertragulidae. Bulletin Texas Memorial Museum no. 23. 34 pp.

Wortman, J.L., 1898. The extinct Camelidae of North America and some associated forms. Bulletin of the American Museum of Natural History 10:93-142.

APPENDIX 1

Poebrotherium wilsoni: GCVP 258, 263, 264, 2011, 2012, 2118, 2119, 2679, 2704, 3816, 3819, 3879, 3932.

Browsers: GCM 411 Madoqua sp.; GCM 17 Odocoileus virginiaus; GCM 918 Gazella granti; Tragulus sp.; Litocranius sp. (both in a private collection)

Mixed Feeders: GCM 44 Oreotragus oreotragus; GMC 405, 406 Hydropotes inermis; GCM 407, 408 Muntiacus muntiac; GCM 914 Llama glama; GCM 135 Antilocapra arnericana.

Grazers: UNSM-16018 Llama pacos; UNSM-16921 Vicugna vicugna; UNSM-5081 Damaliscus lunatus; UNSM-15514 D. dorcas; UNSM-5019 Kobus kob; UNSM-5023 K ellipsiprymnus; UNSM-15482 Hippotragus niger; UNSM-5070 AIcelaphus buselaphus. Modern camels: GCM 9741 975, 976 Camelus dromedarius.