A Comparison of Feeding Adaptations in Two Primitive Ruminants, Hypertragulus and Leptomeryx, from the Oligocene Deposits of Badlands National Park

William P. Wall And Christopher M. Collins
Department of Biology, Georgia College & State University, Milledgeville, GA 31061

Abstract—The craniodental morphology of two primitive Oligocene ruminants, Hypertragulus calcaratus and Leptomeryx evansi, was analyzed to determine the likely dietary habits of these sympatric species. Their craniodental morphologies were compared to each other and to a variety of extant ruminants whose dietary habits are known. A variety of quantitative measures, including hypsodonty index and relative premolar row length, were statistically analyzed. Ten qualitative characteristics were also used in comparison.

Hypertragulus showed overlap between browsers and mixed feeders for a variety of characters, but was most likely a mixed feeder. Leptomeryx grouped with browsers in almost all cases. It seems likely that these two closely related sympatric taxa reduced competition by consuming different food resources.


Hypertragulus and Leptomeryx are small, primitive, hornless ruminants belonging to the infraorder Tragulina that contains the families, Hypertragulidae, Tragulidae, and Leptomerycidae. The only extant members of this infraorder are the tragulids or chevrotains of southeastern Asia. Hypertragulids are the most primitive of the three families. Leptomerycids actually share more derived traits with horned ruminants than do the modern tragulids; therefore they are regarded as the most specialized of the tragulines (Webb and Taylor, 1980).

Hypertragulus calcaratus and Leptomeryx evansi are two of the most common taxa found in paleosols characteristic of savanna woodlands in the middle Oligocene beds exposed in Badlands National Park, South Dakota. Hypertragulus is not found in younger beds, but other species of Leptomeryx survive until the middle Miocene (Emry et al., 1987).

This study was conducted with two purposes in mind. First, to determine how much dietary overlap these two closely related sympatric taxa experienced. And second, to increase our knowledge of leptomerycids, a group that could have given rise to the horned ruminants, the most successful large terrestrial herbivores today.

Paleoecology and Geology of the Brule Formation

The middle Tertiary was a period of major climatic change. At the beginning of the Tertiary, during the Early Paleocene, the world was warm and equable with little latitudinal gradation. Broadleafed deciduous forests adapted to regimes of seasonal darkness were found in the Polar Regions and subtropical woodland extended to within polar confines. Paleocene forests were denser and more closedcanopied than that of the Cretaceous (Prothero, 1994). The thermal high point and the maximum diversity of mammals are represented by the transition from the Early to Middle Eocene. Following this period, a cooling and drying trend occurred (Janis, 1993; Prothero, 1994). Annual precipitation in the region of Badlands National Park was 1,000 mm in the early part of the late Eocene, 500900 mm in the early Oligocene, and only 350450 mm in the late Oligocene (Retallack, 1983).

The fauna of this period was transitional between the jungle-adapted fauna of the early Paleocene and the grassland fauna of today. Carnivores were more heavily proportioned and there is no evidence of pursuit predators. Ungulates were becoming more cursorial, but were still shorter-legged and less cursorial than modern forms (Janis, 1993).

Retallack (1983) studied the paleosols in Badlands National Park. From this he identified three major habitats found during deposition of the Scenic Member of the Brule (Orellan Land Mammal Age). These were stream-side swales with herbaceous vegetation that was probably mainly grasses, gallery woodlands lining the streams, and savanna woodland with scattered trees and intervening areas of bunch grasses, forbs, and small shrubs. Clark et al. (1967) and Retallack (1983) surveyed the fauna from each habitat. They found that in the gallery woodlands the predominant genera were Mesohippus and Merycoidodon while the fauna of the savanna woodlands was predominantly Paleolagus, Leptomeryx, and Hypertragulus. Little data were collected on the near-stream swales fauna, but according to Retallack, (1983), the aquatic rhinoceros Metamynodon was probably common (possibly restricted to this habitat).

The climate during deposition of the Scenic Member was changing from humid and subtropical to subhumid and warm temperate (Retallack, 1983). The dry season was becoming more severe and there were few flowing streams during the dry season. The greater abundance of Celtis seeds and increasing rarity of alligators also indicates that the climate was becoming cooler (Retallack, 1983).

The Poleslide Member of the Brule (Whitneyan Land Mammal Age) contains numerous calcareous nodules indicating a more arid environment than during deposition of the Scenic Member (Retallack, 1983). Few streams flowed permanently and the water table was generally below two meters. The dry climate and sparse vegetation contributed to accelerated sediment accumulation and ash preservation.

Systematic and Anatomical Review

Members of the suborder Ruminantia are typically divided into two infraorders, the Tragulina and the Pecora. Webb and Taylor (1980) consider the hypertragulids to be the most primitive ruminants based on features found in Hypertragulus. Primitive cranial features cited by them include an incomplete postorbital bar, an extensively exposed mastoid, the marginal position of the tympanohyal, and the primitive structure of the petrosal. The premolars are also simple in that they have no lingual elaboration, but it is not clear whether this is a primitive condition or if they were secondarily simplified (Janis, 1987).

The family Tragulidae includes two extant genera with four species (Nowak, 1991). Living tragulids occur exclusively in the Old World and are most common in the subtropical latitudes. The fossil record of tragulids only extends back to the Miocene. Webb and Taylor (1980) place them below leptomerycids due to the concave articulation of the malleolar and calcaneum. The derived state of this articulation in leptomerycids, gelocids and all higher ruminants is concavo-convex. Tragulids also share with all other ruminants except hypertragulids fusion of the trapezoid and magnum, absence of the trapezium, loss of metacarpal I, an incomplete fibula and a distinct malleolar. These features indicate that tragulids diverged from the main lineage of ruminants after the Hypertragulidae.

The most advanced traguline family is the Leptomerycidae. Derived characteristics that separate them from all other tragulines and ally them with higher ruminants are: the posterior position and narrow exposure of the mastoid, fusion of the magnum and trapezoid, reduction of metacarpals II and V, loss of metacarpal I, a well developed malleolar, parallel ginglymi on each end of the astragulus, fused metatarsals III and IV, and the reduction of metatarsals II and V to fused proximal rudiments (Webb and Taylor, 1980).

The dentition of Hypertragulus is simpler than Leptomeryx (Figure 1) and is considerably different except for the molars, which are of the basic selenodont pattern in both animals. The dental formula of Hypertragulus is 0/3 1/1 4/4 3/3 and the dental formula of Leptomeryx is 0/3 1/1 3/4 3/3 (Scott, 1940).

The lower incisors of Hypertragulus are styliform with the crowns spatulate and outcurved. The canine is incisiform, and the first premolar takes the canines place both functionally and in shape. The first incisor in Leptomeryx is much larger than the others. It is procumbent and points almost directly forward. The second and third incisors are much smaller. The canine is incisiform and is closely appressed to the third incisor. P1 is isolated from the other teeth by a diastema in front and behind. It is caniniform in shape but is very small and has only one root (Scott, 1940).

The large upper canines of Hypertragulus were possibly used as defensive fangs. However, the canines do show a significant separation into two size classes, which could indicate this character is sexually dimorphic. The upper canines of Leptomeryx were in contrast small and vestigial, but they probably protruded below the gum line (Scott, 1940; Webb and Taylor, 1980).

Figure 1—Lateral view of the skull and jaws of (A) Hypertragulus and (B), Leptomeryx (modified from Scott, 1940).

In Hypertragulus, P1 is a small, sharp pointed tooth with a simple crown and has two widely divergent roots. It is separated by a diastema both anteriorly and posteriorly. P1 in Leptomeryx is lost (Scott, 1940).

The premolars in the cheek teeth of Hypertragulus are simple compared to those of Leptomeryx. Both P2 and P2 are simple conical teeth. P3 has a weak protocone and lingual cingulum, making it triangular, and P3 has a weak metaconid and hypoconid, giving it a wedge-shaped appearance. P4 and P4 are more elaborate than the other premolars and a small paraconid is added to P4 (Webb and Taylor, 1980).

The premolars in the cheek teeth of Leptomeryx are more complex. P2 has three labial cusps, a large paracone with smaller basal cusps anteriorly and posteriorly. P3 and P4 have strong protocones in addition to the three labial cusps, with the protocone forming a crescent on P4. P2 and P3 have a paraconid, a tall protoconid, and a small hypoconid connected by three longitudinal crests. P4 has a strong metaconid lingual to the protoconid in addition to the other three crests found on P3. Leptomeryx premolars are more submolariform than are those of Hypertragulus. The premolars of Hypertragulus show little wear, while those of Leptomeryx often show considerable wear (Webb and Taylor, 1980).

The molars of both Hypertragulus and Leptomeryx are of the typical selenodont pattern, having four elongated cusps. The molars of Hypertragulus are somewhat higher crowned than those of Leptomeryx, but they are not hypsodont or even mesodont (Scott, 1940). The upper molars of Leptomeryx differ from those of Hypertragulus by the presence of a strong mesostyle, which is absent in Hypertragulus (Janis, 1987, Matthew, 1908). The lower molars of Leptomeryx differ from those of Hypertragulus by the presence of the Palaeomeryx fold. In Leptomeryx, a small fissure separates the entoconulid and hypoconulid on M3 posteriorly, while in Hypertragulus, they are united (Heaton and Emry, 1996). Hypertragulus first appears in the White River group in the Orellan and persists through to the end of the Whitneyan (Emry, 1978).

Retallack (1983) studied the paleosols found in Badlands National Park, and found that Hypertragulus fossils were most abundant in deposits that were formed in savanna woodlands that probably developed in broad floodplains adjacent to streams. Root traces and the nature of the paleosols indicated that they supported abundant herbaceous ground cover with well-spaced and only weakly clumped trees. Celtis seeds are also found in these paleosols, giving further evidence of trees.

The genus Leptomeryx includes seven species (Emry and Heaton, 1996). Some of the more primitive forms are placed in the genus Hendryomeryx by some workers (e.g., Black, 1978; Storer, 1981). The type species for this genus is L. evansi, which is common in Orellan deposits in the Great Plains, and is apparently the only species of Leptomeryx found during the Orellan. There was greater diversity in Chadronian forms, and they are found in deposits ranging from Saskatchewan to Texas (Emry and Heaton, 1996). Leptomerycids survived until the middle part of the Miocene. According to Retallack (1983), Leptomeryx fossils are most common in the same type of deposits as those where Hypertragulus are the most common, but in greater abundance. In addition to these near stream savanna woodland deposits, Leptomeryx remains are also common in savanna woodland deposits that were not associated with streams.

Materials And Methods

Craniodental measurements were taken on both fossil and recent specimens. With the exception of specimen number SDSM 3083 on loan from the South Dakota School of Mines, all fossil specimens used in this study are housed in the Georgia College & State University Vertebrate Paleontology collection (GCVP). Modern forms used for comparison are housed in the Georgia College & State University Mammal collection (GCM). Measurements were taken with Mitutoyo calipers accurate to within 0.05 mm.

The craniodental indices used were hypsodonty index, relative premolar row length, and relative muzzle width. Hypsodonty index was determined by dividing the height of M3 by its width. Height was measured from the base of the crown to the tip of the protoconid on teeth that were fully erupted and showed only slight wear. Width was measured between the outer surfaces of the protoconid and entoconid. Relative premolar row length was determined by dividing the premolar row length by the molar row length. Measurements were taken on the labial side of each series at the base of the crown. Relative muzzle width was determined by dividing the palatal width by the muzzle width. Palatal width was measured as the distance between the second molars at the level of the protocone, and muzzle width was measured at the junction of the premaxillary and maxillary bones.

Statistical analysis of the data was calculated by Instat 2.03. The test used to analyze the data was dependent on sample size and differences in standard deviations between the groups being compared. Because the unpaired t-test assumes equal standard deviations, this test was only used for sets of data with equal standard deviations. When comparing a trait from a single species with the same trait of a group of species within a certain feeding type, the standard deviations of each group would not be expected to be equal. When this occurred, the Mann Whitney test was used. This test checks for significant differences between the medians of the groups being analyzed and makes no assumptions about their standard deviations. This test was used to identify significant differences between Hypertragulus and Leptomeryx and the modern forms.

Measurements for relative muzzle width for all modern forms were taken from Janis and Ehrhardt, (1988). According to them, only grazers could be distinguished from other feeding types with a high degree of confidence, but mixed feeders in open habitats usually have the narrowest muzzles.

Several qualitative characteristics were also used to distinguish feeding types. These included 1) massiveness of the muzzle, 2) length of M3 versus M2, 3) height of the molar basal pillars, 4) morphology of the central cavities of the molars, 5) size of the maxilla in lateral view, 6) presence or absence of a prominent protuberance above M1, 7) the position of the orbit, 8) relative size of the ridge below the orbit, 9) relative size of the coronoid process, and 10) the shape of the ventral and posterior rims of the dentary.

Individual molar and premolar lengths were used to determine if a size difference existed between Hypertragulus and Leptomeryx. An unpaired t-test was used to determine if the size difference was significant for M1 and M2, and the Mann Whitney test was used to test for significant differences in M3. Janis (1995) provides a thorough review of the reliability of comparing fossil and modern taxa.


Table 1 summarizes the data obtained for hypsodonty index for each group measured. Hypertragulus could not be distinguished from high level browsers (p=.090), unspecialized browsers (p=.178), and mixed feeders in closed habitats (p=.467). Hypertragulus is significantly more hypsodont than selective browsers are (p=.026), and significantly less hypsodont than mixed feeders in open habitats (p=.005), and grazers (p=.036). The Mann Whitney test was used to calculate p values for all groups except high level browsers and unspecialized browsers, whose p value was found using the Table 1—Median, mean, and range of hypsodonty index for each group measured.

Table 2—Qualitative characteristics observed in Hypertragulus and Leptomeryx.

Characteristic Hypertragulus Leptomeryx

Muzzle massiveness browser or browser or

mixed feeder mixed feeder

Length of M2 versus M3 browser browser

Molar basal pillars browser browser

Molar central cavities browser browser

Size of maxilla browser browser

Prominence above M1 —— browser

Position of orbit browser browser

Size of ridge below orbit grazer browser

Size of the Coronoid Process mixed feeder browser

Shape of the dentary mixed feeder browser?

n Median Mean Range

Hypertragulus 2 2.15 2.15 2.12-2.17

Leptomeryx 7 1.67 1.67 1.62-1.72

High level browsers 5 1.32 1.45 1.18-2.23

Selective browsers 3 1.47 1.50 1.30-1.72

Unspecialized browsers 8 1.60 1.64 1.23-2.29

Mixed closed habitats 19 1.97 2.07 1.12-3.03

Mixed open habitats 37 3.90 3.90 2.12-5.30

Fresh grass grazers 7 3.59 3.39 2.35-4.05

Grazers 9 4.87 4.87 3.77-6.12

unpaired t test. There was not a large enough sample size to calculate a p value using the Mann Whitney test and the standard deviations were not equal so an unpaired t test could not be used to test for a significant difference between Hypertragulus and fresh grass grazers.

Leptomeryx could not be distinguished from high level browsers (p=1.06), selective browsers (p=.383), and unspecialized browsers (p=.121). Leptomeryx is significantly less hypsodont than mixed feeders in closed habitats (p=.010), mixed feeders in open habitats (p=<.0001), fresh grass grazers (p=.0006), and grazers (p=.0002). The Mann Whitney test was used to calculate all p values.

Hypertragulus is significantly more hypsodont than Leptomeryx with an unpaired t test p value of <.0001. There was not a significant difference between the relative premolar row length of the browsers and mixed feeders that were measured (p=0.1075).

The mean relative premolar row length for Leptomeryx was .917 and the median was .915. This was not significantly different (p=.257) from browsers whose mean relative premolar row length was .859 and median was .822. Leptomeryx had a significantly longer relative premolar row length than mixed feeders (p=.002) whose mean was .734 and median was .762.

Table 2 lists the qualitative characteristics observed in Hypertragulus and Leptomeryx and whether they are more like browsers or grazers in these features.

The lengths of M2 and M3 were both browser-like in that they were approximately equal in both Hypertragulus and Leptomeryx. The molar basal pillars are also browser-like in that they are small and do not reach the occlusal surface. The molar central cavities are also browser-like, being simple crests with no complex folding.

In lateral view (Figure 1), the size of the maxilla in Hypertragulus and Leptomeryx is small like modern browsers. There is no maxillary protrusion like that often found in grazers in either Hypertragulus or Leptomeryx.

The position of the orbit in both Hypertragulus and Leptomeryx is browser-like in that it starts above M2 in both. In Hypertragulus, the zygomandibularis leaves a pronounced ridge on the zygomatic arch like that found in grazers. In Leptomeryx, this ridge is small and browser-like.

Leptomeryx has a relatively more massive coronoid pro

cess than Hypertragulus. The large coronoid process in Leptomeryx would indicate that the temporalis muscle and therefore the orthal retraction phase of the chewing cycle were more important in this animal than in Hypertragulus. The ventral rim of the dentary of Hypertragulus is like that of mixed feeders in that it is concave and thin lateromedially. The ventral rim of the dentary of Leptomeryx is fuller and more like that of browsers. The posterior rims of the dentaries of both Hypertragulus and Leptomeryx are concave like those of browsers.

There was a significant size difference between Hypertragulus and Leptomeryx, with Hypertragulus being much smaller. The mean length of M1 of Hypertragulus is 5.09 mm and the mean length of M1 of Leptomeryx is 6.53mm. The unpaired t test p value is <0.0001. The mean length of M2 of Hypertragulus is 5.35 mm and the mean length of M2 of Leptomeryx is 6.78mm. The unpaired t test p value is <0.0001. The difference in standard deviations between M3 was too large to calculate a p value using the unpaired t test. The Mann Whitney p value was significant at <0.0001.


The molars in Hypertragulus are significantly higher crowned than those in Leptomeryx (Table 1), however, neither falls into the range required for hypsodonty (Janis, 1988). The crown height for both taxa falls into the mesodont range (Leptomeryx at the low end and Hypertragulus in the middle). The crown height in Hypertragulus groups it with high level browsers, unspecialized browsers, and mixed feeders in closed habitats. The crown height exhibited by Leptomeryx groups it clearly with the browsers (no overlap with either group of mixed feeders).

Hypertragulus groups with mixed feeders such as Antilocapra, the North American pronghorn, and Boselaphus, the Indian nilgais, with regard to relative premolar row length. Leptomeryx has a premolar row length typical of modern browsers.

A comparison of the qualitative characters examined (Table 2) reveals that Leptomeryx is browser-like in almost all traits, while Hypertragulus shares some traits with browsers and others with mixed feeders. Relative muzzle width is not significantly different in Hypertragulus and Leptomeryx. The narrow muzzles of both taxa clearly indicate they are not like modern grazers in this feature. The relatively larger coronoid process on the lower jaw of Leptomeryx indicates orthal retraction, the food acquisition phase of mastication, took more effort in this animal than in Hypertragulus. This distinction could indicate Hypertragulus was more selective in its dietary habits than Leptomeryx. The angular shape of the lower jaw in Hypertragulus and Leptomeryx is clearly different (Figure 1). Hypertragulus is more like that of primitive artiodactyls, possibly indicating a more generalized jaw muscular pattern. Leptomeryx has a relatively full and thick mandibular angle indicating a more specialized jaw muscle arrangement. This trait would therefore imply a more specialized feeding pattern for Leptomeryx.

Based on their similar body size and anatomical traits, the closest modern analog for Hypertragulus appears to be Moschus the Asiatic musk deer. Moschus has a varied diet consisting of grass, moss, twigs, and other leafy material (Nowak, 1991). Leptomeryx, on the other hand compares most favorably to the modern Pudu, the South American pudu, and Tragulus, the Asiatic mouse deer. The pudu is a generalist browser and the mouse deer is a selective browser (Nowak, 1991).


Analysis of the craniodental morphology of these primitive ruminants indicates that Hypertragulus was probably a mixed feeder and Leptomeryx was a browser. These small herbivores were clearly sympatric spatially and temporally during much of the Orellan. Based on the results of this study, however, it would appear that these two taxa did not overlap significantly in food requirements. Hypertragulus does not appear in the fossil record after the Whitneyan. Whether Leptomeryx played an indirect role in the termination of the Hypertragulus lineage could not be determined from this study. Based on their success through the middle Miocene it would appear that leptomerycids were well adapted to a browsing mode of life. Questions regarding the adaptive radiation of the horned ruminants should consider the browsing mode of life as a likely starting point for the evolution of modern pecorans.


We thank Dr. Philip Bjork for access to specimens in the South Dakota School of Mines collection. We also thank Ms. Rachel Benton of Badlands National Park for her extensive support of our research efforts. Dr. Bill McDaniel of Georgia College & State University provided useful insight on statistical techniques. Finally we thank Mr. Vince Santucci for his enthusiasm and support for paleontological research in the National Parks. This research was partially funded by faculty research grants from Georgia College & State University.


Black, C. C. 1978. Paleontology and Geology of the Badwater Creek Area, Central Wyoming Part 14. The Artiodactyls. Annals, Carnegie Museum of Natural History. 47(10): 223-

Clark, J., J. R. Beerbower, and K. Kietzke. 1967. Oligocene Sedimentation, Stratigraphy, Paleoecology and Paleoclimatology in the Big Badlands of South Dakota. Fieldiana: Geology Memoirs. 5:1-158.

Emry, R. J. 1978. A New Hypertragulid (Mammalia, Ruminantia) From the Early Chadronian of Wyoming and Texas. Journal of Paleontology. 52: 1004-1115.

———, L. S. Russell, and P. R. Bjork. 1987. The Chadronian, Orellan, and Whitneyan North American Land Mammal Ages. In: Cenozoic Mammals of North America (M. O. Woodburne, ed.). University of California Press. Berkeley. Pp. 118-152.

Heaton, T. and R. J. Emry. 1996. Leptomerycidae. In: The Terrestrial Eocene-Oligocene Transition in North America (D. R. Prothero and R. J. Emry, eds.). Cambridge University Press. Cambridge. Pp. 581-608.

Janis, C. M.. 1987. Grades and Clades in Hornless Ruminant Evolution: The Reality of the Gelocidae and the Systematic Position of Lophiomeryx and Bachitherium. Journal of Vertebrate Paleontology. 79(2): 200-216.

———. 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 dHistoire Naturelle. Paris. 53: 367-387.

———. 1993. Tertiary Mammal Evolution in the Context of Changing Climates, Vegetation, and Tectonic Events. Annual Review of Ecology and Systematics. 24: 467-500.

———. 1995. Correlations Between Craniodental Morphology and Feeding Behavior in Ungulates: Reciprocal Illumination Between Living and Fossil Taxa. IN: Functional Morphology in Vertebrate Paleontology. (J. J. Thomason, ed.). Cambridge University Press. Pp. 76-94.

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.

Matthew, W. D. 1908. Osteology of Blastomeryx and Phylogeny of the American Cervidae. Bulletin, American Museum of Natural History. 24: 535-562.

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

Prothero, D. R. 1994. The Eocene-Oligocene Transition. Columbia U. Press: New York.

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.

Scott, W. B. 1940. Atiodactyls Part IV. IN: Scott, W. B. and G. L. Jepsen, The Mammalian Fauna of the White River Oligocene. Transactions of the American Philosophical Society. 28: 363-746.

Storer, J. E. 1981. Leptomerycid Artiodactyla of the Calf Creek Local Fauna (Cypress Hills Formation, Oligocene, Chadronian), Saskatchewan. Saskatchewan Museum of Natural History Contributions. 3: 1-32.

Webb, D. S., and B. E. Taylor. 1980. The Phylogeny of Hornless Ruminants and a Description of the Cranium of Archaeomeryx. Bulletin of the American Museum of Natural History. 167(3): 121-154.