William P. Wall and Michael J. Shikany
Department of Biology
Georgia College
Milledgeville, GA 31061


A comparison of feeding mechanics in Oligocene agriochoeres (Agriochoerus) and oreodonts (Merycoidodon and Promesoreodon) is presented. Measurements and observations obtained in this study resulted in the recognition of different dietary habits for agriochoeres and oreodonts. The most significant measurements were of tooth rows. Ratios of premolar to molar row length show a greater grinding area for food in the oreodonts. Examination of occlusal surfaces of dentition showed that oreodonts made tooth/tooth contact (characteristic of grazers), while agriochoeres made only tooth/food contact (more typical of browsers). Cranial morphology, jaw muscle reconstructions, and vector analyses showed significant differences between agriochoeres and oreodonts. The large sagittal crest and shallow angle of the jaw indicate agriochoeres were well adapted for shearing modes of plant mastication by means of a well developed temporalis. Oreodont skulls show adaptation for a grinding type of mastication emphasizing the masseter. Additionally, the oreodonts possess a complete orbit, which acted as a buttress to allow for increased jaw adductor musculature. Agriochoeres probably occupied a woodland habitat while oreodonts show specialization for a savanna grassland.


Agriochoerids are a little known group of mid-Tertiary
artiodactyls. Prior research on agriochoerids was primarily taxonomic or faunal in scope, leaving much speculation as to what niche and habitat this group occupied. Little is known about the feeding mechanics of agriochoerids, or their ecologic relationship to their more successful relatives, the oreodonts.
During the middle Eocene, primitive selenodonts gave

rise to the superfamily Merycoidodontoidea, which is subdivided into the families Agriochoeridae and Merycoidodontidae (Carroll,

1988). Agriochoerids range from the middle Eocene to the late Eocene. The more abundant oreodonts first appeared in the late Eocene and survived to the middle Pliocene.
The genus Agriochoerus , Scott (1929), a rare member of the White River Fauna of Badlands National Park, is the focus of this study. Like oreodonts, agriochoerids were endemic to North America, and are thought to have lived in open woodland or riverside areas (Osborn, 1910). Agriochoerids and oreodonts share various morphological characteristics in the skull and jaw region primarily due to common ancestry. These similarities imply that agriochoeres and oreodonts occupied similar niches. Like the oreodonts, agriochoerids have tusks formed from the upper canines and lower first premolars, also, in both, the lower canines are incisiform (Scott and Jepsen, 1940). Selenodont dentition is shared by both, however, there are subtle differences between occlusal patterns and general tooth structure. Both groups possess a small braincase, large temporal fossa, and prominent sagittal crest (Wortman, 1895). We undertook a comparative study of the jaw mechanics of agriochoerids and oreodonts to shed light on what niche the agriochoerids filled, and how this niche, if at all, overlapped with that of the oreodonts. The fact that agriochoerid fossils are rare and oreodont fossils are numerous could be important in determining habitat as well as indicating the relative success of these two families of artiodactyls.


The family, Agriochoeridae, contains three genera: Protoreodon, Diplobunops, and Agriochoerus. According to Golz (1976), this family is nothing more than a primitive form of oreodonts. Protoreodon represents the basal stock from which agriochoerids, as well as oreodonts arose (Carroll, 1988). This genus first appeared in the Uintan, and survived into the Chadronian (Savage & Russell, 1983). Protoreodon was the most abundant of all Uintan mammals, and inhabited areas now represented by California and the western interior of North America (Scott, 1889). The skull of Protoreodon is more primitive than Diplobunops and Agriochoerus in two significant ways. The skull is longer and narrower (more dolichocephalic), and its zygomatic arch is less robust than that of the other two agriochoerid genera (Golz, 1976).
Diplobunops exhibits a mosaic of characters intermediate between Protoreodon and Agriochoerus (Coombs, 1983). Like Protoreodon, Diplobunops retains the paraconule, a cusp between the paracone and protocone of the superior molars, lost in Agriochoerus (Golz, 1976). The molars tend to be relatively narrow, and their labial crescents are less extended across the crown than those of Agriochoerus. The dentary bone of Diplobunops shows a trend towards the condition in Agriochoerus, in that it becomes a little more robust and broadened than that of Protoreodon (Scott, 1945). Agriochoerus, the most advanced of the agriochoerids, appears to be a chimera of discordant body parts. Agriochoerus resembles oreodonts in the skull region, calicotheres in the foot region, and cats in the posterior limb region (Scott, 1929). Agriochoerus is reported from the upper Duchesnean, but the poor quality of these specimens makes this designation uncertain. The first positive identification of this genus is from the Chadronian beds of the White River Formation. According to Osborn (1910) agriochoerids inhabited forested woodland areas and were arboreal. Coombs (1983) concluded from an analysis of the manus and pes that agriochoerids show little adaptation for grasping and therefore were probably not arboreal.

Agriochoerus has an incomplete posterior ocular orbit, and is lacking a lachrymal pit (Wortman, 1895). The auditory bullae are inflated, but not filled with calcareous tissue distinguishing them from oreodont bullae (Wortman, 1895). The zygomatic arch is long and expanded, as in oreodonts, but its palate and jaws extend further laterally (Greaves, 1978). The nasals narrow proximally (Gregory, 1920), providing an easy visual distinction from oreodonts. The mandible of Agriochoerus is little different from Diplobunops.
The teeth of Agriochoerus are brachyodont and selenodont, with the same primitive dental pattern as that of all agriochoerids, 3/3, 1/1, 4/4, 3/3. Although there are some differences between the teeth of Agriochoerus and oreodonts, most researchers stressed the similarities between the two groups when interpreting possible dietary habits. Agriochoerus has a diastema between the superior canines and the first premolar in the maxilla, and between the caniniform first premolar, and second premolar of the mandible (Wortman, 1895). All fourth premolars are molariform in appearance, and there is a reduction in the hypocone of the fourth superior premolar (Greaves, 1972). The molars of Agriochoerus are tetraselenodont, lacking a protoconule, which was present in Protoreodon and Diplobunops (Carroll, 1988). The inferior molars are of the typical selenodont pattern, however, the superior molars differ, resembling those of the anthracothere possessing high cusps, and an anterior cingulum (Zittel, 1925), with deeply concave external crescents which are rounded (Wortman, 1895). The third inferior molar possesses a hypoconulid.


This study is divided into two sections: First, the teeth of Agriochoerus were compared to Oligocene anthracotheres and oreodonts. Observations of the teeth included a visual inspection of tooth cusp patterns, wear facets, relative size of tooth rows, and presence or absence of diastemas. Measurements of the teeth included length, width, surface area (length multiplied times width), and height (tooth/jaw interface to highest cusp). Various aspects of tooth rows were also measured, such as: size of diastema, if present, length of entire tooth row, and length of molar row. Wear facets on teeth were interpreted in accordance with Butler (1972) and Greaves (1973). If enamel is present over the majority of the occlusal surface, the teeth are considered to exhibit abrasion facets. However, if enamel dentine interfaces are present, the teeth are considered to exhibit attrition facets (Peyer, 1968). Specific aspects of the facets were also examined, such as the leading and trailing edge of the facets, to help determine where chewing cycles were initiated and terminated. Attrition facets can also define the manner in which food is broken down during mastication, i.e., puncturing, shearing, or grinding. These three methods of mastication greatly contribute to unique wear facets and may be used to define masticatory relationships between taxa (Butler, 1972).

The second major section of this study involves description, measurements, and biomechanical analysis of cranial structures. Agriochoerus cranial proportions were compared to oreodonts and anthracotheres using distortion grids (Hildebrand, 1988). Cranial measurements were chosen based on their significance in showing similarities or differences among the taxa. A vector analysis of the jaw regions of these artiodactyls was performed. The vectors drawn in this section represent the lines of action of the major adductor muscles in these herbivorous mammals, the masseter, temporalis, and pterygoideous (for comparison with modern ungulates see Herring, 1976; and Janis, 1983). The muscular anatomy of the head/jaw region of the selected artiodactyls was reconstructed using a method similar to Greaves (1972). After the origins and insertions for the adductor muscles were determined, individual muscle masses were estimated by the size of grooves and ridges located in the areas of tendon attachment. The exact size or mass of a muscle is impossible to determine from these attachments, however, relative sizes can be estimated, which are sufficient for comparative purposes. Once the relative size of each muscle was determined, vector diagrams were drawn; a separate vector was drawn for each muscle analyzed. The direction of each vector was determined by locating the center of mass for each muscle, and magnitude was estimated based on the relative size of the muscle. The total of all adductor muscles was set at one hundred percent. The contribution of each muscle towards the total was then determined, through observation of the muscle reconstructions. A scale of ten percent equals one centimeter of vector length was used to obtain the magnitude of each vector in the original drawings. After all the muscle vectors were properly placed, their lever arms were determined. To simplify the study we selected the jaw joint as the fulcrum for our analysis (see Gans, 1974; Alexander, 1983; Greaves, 1991; and Hildebrand, 1988, for the relevance of lever arm studies in biomechanics).
All measurements were taken with a Helios calipers accurate to 0.01mm. All of the specimens used in this study were collected in Badlands National Park, South Dakota and are housed in the Georgia College Vertebrate Paleontology Collection (GCVP), Milledgeville, Georgia.


There was a gradual increase in the upper premolar to molar ratio from the anthracothere Aepinacodon, the most primitive genus, to Merycoidodon and Promesoreodon (oreodonts), the most advanced organisms in this study. The upper premolars of Aepinacodon are approximately forty-four percent of the length of the upper molars, in Agriochoerus this ratio was approximately fifty-one percent, and in the two oreodont genera the upper premolars were fifty-seven percent of the length of the molars. The ratio of lower premolars to molars in Agriochoerus was seventy-four percent, and the average ratio for the oreodonts was sixty-five percent.
Facet patterns on the cheek teeth showed a clear transition from primitive to advanced forms. The occlusal surfaces on the upper and lower dentition of Agriochoerus and Aepinacodon primarily exhibited abrasion facets. Even though abrasion marks, indicative of jaw motion during mastication, were observed, enamel covered the entire occlusal surface (Figure 1B). In addition, leading and trailing edges of the wear facets could be identified but were not greatly pronounced. Contrasting this, the oreodonts had attrition facets on their occlusal surfaces which show a well developed dentine/enamel interface. The leading edges of this interface were sharp and highly developed, in comparison to the developed, but rounded trailing edges (Figure 1A).

Significant cranial differences between the three families are revealed by the distortion grids (Figure 2) The anthracothere skull shows a greater resemblance to the agriochoere skull than it does to the oreodont. A lachrymal pit, just anterior to the orbit, is present in the oreodont, but absent in the anthracothere and agriochoere. The orbit is open posteriorly in the anthracothere and agriochoere, while the more advanced oreodont orbit is complete. Several changes in the oreodont jaw are also evident. First, there is no diastema in oreodont dentition, but a pronounced diastema exists in agriochoeres and anthracotheres. Second, the angle of the jaw, as well as the entire posterior jaw of the oreodonts is expanded compared to that of agriochoeres and anthracotheres. A final difference in skull structure is the sagittal crest, which is less prominent and less convex in the oreodont than in the more primitive forms.
Muscle reconstructions involving the three major adductors of the jaw indicate muscle mass increased from anthracotheres to oreodonts. The relative importance of each muscle also differs in these taxa. The anthracothere adductor mass was 50% masseter, 40% temporalis, 10% pterygoideus, for the agriochoere 55% masseter, 35% temporalis, 10% pterygoideus, and for oreodonts 60% masseter, 30% temporalis, 10% pterygoideus. These results show there is a gradual increase in the size of the masseter and a gradual decrease in size of the temporalis from anthracotheres to oreodonts.
Vector analyses drawn from these reconstructions are shown in Figure 3. The lever arm of the temporalis exhibited a gradual decrease in size from anthracotheres to oreodonts. Contrasting this, the lever arm of the masseter shows an increase in the more advanced forms. Changes in the lever arm are due to changes in muscle mass and from variation of skull morphologies centered around the craniomandibular joint.


Results obtained from the tooth measurements indicated a higher premolar to molar ratio in the upper dentition, and a lower ratio in the lower dentition, for Oligocene oreodonts compared to Agriochoerus. Even though the upper premolars of oreodonts occupied a greater portion of the total tooth row, the ratio of M3 to M2 was significantly higher in oreodonts indicating an increase in the total length of their molars. The large value for the molar region shows that oreodonts were better adapted for grinding foodstuffs than agriochoeres.
The occlusal surfaces of the dentition in both oreodonts and agriochoeres exhibited facet wear patterns that are typical of ungulate mastication. Subtle differences in these facet patterns, however, indicate emphasis on different parts of this cycle. The abrasion facets of agriochoere teeth indicate there was no actual contact between teeth during occlusion. Mastication was achieved through tooth/food contact, which is typical of organisms whose chewing cycles are dominated by shear chewing methods. Shearing mastication would allow agriochoeres to break down larger, tougher plant sources, including shrubs and roots. Striations found on the occlusal surfaces of agriochoere teeth are oriented labial/lingual, showing that most of the work done by the teeth was performed in a transverse motion. Therefore, it may be assumed that tooth/food contact was made during the late orthal retraction phase through the early buccal phase of the chewing cycle, when transverse motion of the jaw is initiated. The attrition facet patterns found on the occlusal surfaces of oreodont dentition show that tooth/tooth contact was made during mastication. The dentine/enamel interface on occlusal surfaces causes a rasping action during chewing cycles and allows a greater processing of foods which require much grinding, such as grasses. The attrition interfaces of the oreodont are arranged in a transverse orientation, like the agriochoere's, however, the dentine/enamel interface indicates a more powerful grinding action during mastication. This means tooth/tooth contact occurred primarily during the buccal phase, which is the most powerful stroke of the chewing cycle. Differences in occlusal facet patterns between oreodonts and agriochoeres are evident in the relative development of adductor jaw musculature in the two groups.

Skull and jaw morphologies of agriochoeres and oreodonts exhibit differences which can best be attributed to selection for optimal mastication in these organisms. The complete orbit found in oreodonts, but absent in agriochoeres, gave oreodonts a structural advantage by allowing for a greater power stroke and larger range of lateral motion of the jaw than in agriochoeres. The posterior portion of the orbit in oreodonts serves as a buttress which adds support to the orbit area. According to Scapino (1972) and Greaves (1984) this extra support permits increased muscle mass along with increased lateral motion of the entire jaw, which are both characteristic of oreodont mastication.
The lack of a diastema in the lower jaw of oreodonts, provides more room for teeth, however, other factors are of greater significance. The purpose of a diastema in herbivorous mammals is to allow manipulation of food, with the tongue, during mastication. The space provided by the diastema allows food to be moved from the cropping mechanism to the molars, via the tongue. Modern grazers, such as horses, show good examples of diastema function, however, the surface area ratio of premolars to molars, as well as the entire palate area, is much greater in these modern grazers than in oreodonts and agriochoeres. Organisms with such a limited palate area, like the oreodonts, would not have much trouble moving material, such as grass, from the cropping mechanism to the grinding mechanism during mastication, without the aid of a diastema. Also, lack of a diastema would shorten the length of the jaw, thereby decreasing the magnitude of the resistance arm of the adductors. This would give the oreodonts a greater capacity for powerful grinding motions of the teeth near the proximal end of the jaw. The presence of a diastema in agriochoeres is substantial evidence that food was manipulated by the tongue during the chewing cycle. Tongue manipulation would allow tougher and larger vegetation, such as shrubs and roots, to be consumed in an organism with a palate the size of an agriochoere's.
These two groups also differ in the relative development of the angle of the jaw and sagittal crest. The expanded angle of the jaw of oreodonts permits greater insertion of the masseter. Action of the masseter is greatest during the buccal phase of the chewing cycle, when grinding is taking place. Thus, a larger region for insertion of the masseter would reinforce the hypothesis that grinding predominated in oreodont mastication. Conversely the sagittal crest in agriochoeres provides a large area for the temporalis to originate. A large temporalis would have been used in three ways. First, a large temporalis, whose action causes upward and rearward adduction of the jaw, would allow agriochoeres to more easily acquire large vegetation and/or roots from the ground. Second, resistance from the root or shrub would generate a force at the craniomandibular joint as the animal pulled on the plant. This stress on the jaw joint can best be countered by a force in the opposite direction. The temporalis is the only adductor muscle in proper alignment to resist this force and thereby stabilize the jaw joint. And third, action of the temporalis is maximized during the orthal retraction and early buccal phases of mastication, when shearing predominates. Larger plant material must be sheared into manageable particles before it can be ground into a small bolus suitable for swallowing. Muscle reconstructions and vector analyses presented in the results also support the hypothesis that agriochoeres were browsers. The large lever arm of the masseter in oreodonts reinforces the idea that grinding predominated in oreodont mastication. Contrasting this, the large lever arm of the temporalis in agriochoeres would be advantageous in shearing modes of mastication. Even though the pterygoideus represents 10% of the jaw adductor muscle mass in agriochoeres and oreodonts, there is an increase in the overall mass in oreodonts. Increase in total muscle mass would represent increased development of the pterygoideus, which along with the masseter acts to pull the jaw transversely. Thus, oreodonts would exhibit greater efficiency at grinding food.

Post cranial characters corroborate our analysis of feeding in agriochoeres and oreodonts. Oreodonts have a reduced number of digits and digitigrade stance, locomotor traits suitable for the cursorial lifestyle typical of many grazers. The manus and pes of agriochoeres indicate a different lifestyle. Their plantigrade stance is suited for a subcursorial mode of locomotion. Additionally, the prominent claws on agriochoere digits, whose presence is still not adequately explained, could have been used for digging and probing for roots. Coombs (1983) disagrees with this idea, pointing out that the hind limb of agriochoeres is not as developed as most diggers. This, however, does not rule out the possibility that agriochoeres were occasional diggers, and dug for roots as a secondary food source.

Our study suggests that agriochoeres and oreodonts maintained low levels of competition during the Oligocene since they did not compete directly for food resources. Agriochoeres favored a diet of large vegetation and tubers. This choice of food, and the fact that agriochoeres retained a primitive digit formula supports the notion of Osborn (1910) that agriochoeres were woodland specialists. Oreodonts, on the other hand, probably fed on grass, and had a reduced number of digits, making them more agile runners. These two characteristics made them better adapted for survival in a savanna habitat.
The early extinction of agriochoeres probably resulted from changes in the ecology of Nearctica. During the late Oligocene to early Miocene, woodland areas, which predominated up to this point (Wolfe, 1971) were giving way to savannas (Savage & Russell, 1983). This left abundant resources for oreodonts, but restricted the habitat for agriochoeres. The early demise of agriochoeres was probably due to their inability to successfully compete with their specialized relatives in a purely savanna environment.


Alexander, R.M., 1983. Animal Mechanics: Blackwell Scientific Publications, Oxford, 301 p.

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

Carroll, R.L., 1988. Vertebrate Paleontology and Evolution: W.H. Freeman Co.,New York, 698 p.

Coombs, M.C., 1983. Large mammalian clawed herbivores: Transactions of the American Philosophical Society, v. 73, pt. 7, p. 1-96.

Gans, C., 1974. Biomechanics: an approach to vertebrate biology: J.B. Lippincott Company, Philadelphia, 261 p.

Golz, D.J., 1976. Eocene Artiodactyla of southern California: Natural History Museum of Los Angeles County, Science Bulletin v. 26, p. 29-39.

Greaves, W.S. 1972. Evolution of the Merycoidodont masticatory apparatus: Evolution, v. 26, p. 659-657.

Greaves, W.S., 1973. The inference of jaw motion from tooth wear facets: Journal of Paleontology, v. 47(5), p. 1000-1001.

Greaves, W.S., 1978. The posterior zygomatic root in oreodonts: Journal of Paleontology, v. 52, p. 740-743.

Greaves, W.S., 1984. The mammalian post orbital bar as a torsion resisting helical strut: Journal of Zoology, v. 207, p. 125-136.

Greaves, W.S., 1991. The orientation of the force of the jaw muscles and the length of the mandible in mammals: Zoological Journal of the Linnean Society, v. 102, p. 367-374
Gregory, W.K., 1920. Studies in comparative myology and osteology: Bulletin of the American Museum of Natural History, v. 42, p.187-193.

Herring, S.W., 1976. The dynamics of mastication in pigs: Archives Oral Biology, v. 27, p. 473-480.

Hildebrand, M., 1988. Analysis of Vertebrate Structure: John Wiley and Sons, New York, 654 p.

Janis, C.M., 1983. Muscles of the masticatory apparatus in two genera of Hyraxes (Procavia and Heterohyrax): Journal of Morphology, v. 176, p. 61-87.

Osborn, H.F., 1910. The Age Of Mammals In Europe, Asia, and North America: Macmillan Co., New York, p. 177-241.

Peyer, B., 1968. Comparative Odontology: Univ. of Chicago Press, Chicago, 347p.

Savage, D.E. and Russell, D.E., 1983. Mammalian Paleofaunas Of The World:Addison Wesley Co., London, 432 p.

Scapino, R.P., 1972. Adaptive radiation of mammalian jaws; in G.H. Schmuacher, ed., Morphology of the Maxillo-Mandibular Apparatus: IXth International Congress of Anatomists, Leipzig, p. 33-39.

Scott, W.B., 1889. The mammalia of the Uintan formation: Transactions of the American Philosophical Society, v. 16, pt. 3, p. 487-503.

Scott, W.B., 1929. A History Of Land Mammals In The Western Hemisphere: Macmillan Co., New York, p. 383-386.

Scott, W.B., 1945. The mammalia of the Duchesne river Oligocene: Transactions of the American Philosophical Society, v. 34, pt. 3, p. 230-238.

Scott, W.B. and Jepsen, G.L., 1940. The mammalian fauna of the White River Oligocene: Transactions of the American Philosophical Society, v. 28, part 1, p. 687-709.

Wolfe, J.A., 1971. Tertiary climatic fluctuations and methods of analysis of Tertiary floras: Paleogeography, Paleoclimatology, Paleoecology, v. 9, p. 27-57.

Wortman, J.L., 1895. On the osteology of Agriochoerus: Bulletin of the American Museum of Natural History: v. 7, p. 145-179.

Zittel, K.A., 1925. Text-Book of Paleontology: MacMillan Co., New York, v. 3, pp. 183-184.


Figure 1: A. Cross section of an oreodont tooth (modified from Greaves, 1973).B. Cross section of an agriochoere tooth . Arrows indicate direction of occluding tooth. Figure abbreviations: le-leading edge, li-leading interface, te-trailing edge, ti-trailing interface, d-dentine, and e-enamel.

Figure 2: Distortion grid comparison using an Aepinacodon skull (A) a standard for Agriochoerus (B) and Merycoidodon (C) skulls.

Figure 3: Vector analysis diagrams of A, Merycoidodon; B, Agriochoerus; and C, Aepinacodon.

Back to Table of Contents

United States Department of the Interior, National Park Service

Go to Nature Net Home Page Go to Park Net Home Page