William P. Wall and Walker Hickerson
Department of Biology
Georgia College
Milledgeville GA 31061


The Oligocene rhino Hyracodon is generally regarded as cursorial primarily because it is smaller than other rhinocerotoids. However, the cursorial abilities of this animal have never been tested biomechanically. The appendicular skeleton of Hyracodon is described and compared to Ceratotherium (white rhinoceros). The relative development of significant locomotor muscles is interpreted from muscle scars on specimens of Hyracodon. Relative lengths of each segment of the limb in Hyracodon are compared to a variety of modern ungulates ranging in locomotor habits from graviportal to cursorial. Hyracodon limb indices do not compare favorably with modern cursorial mammals. The following indices provide typical examples. The MCIII/radius index for Hyracodon (54.56) compared more favorably with Hippopotamus (56.30) than with the cursor Antilocapra (105.45). The tibia/femur index for Hyracodon (83.72) is very similar to the index for Sus (86.69) but falls far below that for Antilocapra (123.81). The data suggests a different lifestyle for Hyracodon than was earlier believed. Wild pigs or peccaries would be a more accurate modern locomotor analog for Hyracodon than cursorial ungulates.

The early rhinocerotoid Hyracodon is one of the most common fossils found in the Oligocene strata exposed in Badlands National Park, South Dakota. Hyracodon (family Hyracodontidae) is a primitive, long lived taxon that survived for nearly 10 million years in North America with very little change (Prothero et al, 1986). The superfamily Rhinocerotoidea is made up of three families (Amynodontidae, Hyracodontidae, Rhinocerotidae) united by derived characteristics of the skull, teeth, and post cranial skeleton as described by Prothero et al (1986). Rhinocerotoids are most likely derived from hyrachyid tapiroids (Wall and Manning, 1986).
Hyracodon lived contemporaneously with the true rhinoceroses, Trigonius and Subhyracodon (Scott, 1941) and the amynodontid Metamynodon (Wall, 1989). These animals are traditionally placed into three very different ecologic roles. Metamynodon probably filled a niche comparable to modern hippopotami. The rhinocerotids exhibit traits typical of large terrestrial herbivores. Hyracodon is invariably called the "running rhino" because of its smaller size and lighter build compared to other rhinocerotoids. The actual locomotor abilities of Hyracodon have never been examined biomechanically.

Limb structure and proportions are closely tied to an animals body size and life habits. Biomechanical analysis of limb/locomotor systems can be used to assign an animal to one of four traditional locomotor groups - graviportal, mediportal, subcursorial, and cursorial. Strict definition of these groups is not possible since they represent a continuum from one extreme (graviportal - limbs primarily designed for weight bearing) to another (cursorial - extreme development of cursorial adaptations as seen in most long distance runners; see Coombs, 1978 for a detailed discussion of these terms).

This paper presents the results of a comparative biomechanical investigation of Hyracodon and the modern white rhino, Ceratotherium, to a variety of modern ungulates.


This work is divided into two parts: first, a description of relevant portions of the limb anatomy of Hyracodon and Ceratotherium; and second, an interpretation of the locomotor abilities of Hyracodon based on a comparison of limb bone indices to a variety of modern ungulates.
Muscle reconstructions of Hyracodon were based on specimens housed in the American Museum of Natural History, New York City, NY (AMNH 1168, 1176) , Frick Collection (FAM 0-21-549, also housed at the American Museum), Georgia College Vertebrate Paleontology Collection, Milledgeville, GA (GCVP 4756, 4928), and South Dakota School of Mines and Technology, Rapid City, SD (SDSM 9789). Muscle origins and insertions were determined from identification of muscle scars and comparison to literature on recent mammalian anatomy (Beddard and Treves, 1889; Davis, 1964; Gregory, 1929; Howard 1973, 1975; Osborn, 1929; Sisson and Grossman, 1975; Windle and Parsons, 1901,1903).
All recent mammals cited, except Elephas, are housed in the Georgia College Mammal Collection (GCM). Modern mammals used for comparison included Dama (fallow deer, Artiodactyla: GCM 576), Tragelaphus (bush buck, Artiodactyla: GCM 577), Lama (llama, Artiodactyla: GCM 1587), Antilocapra (pronghorn, Artiodactyla: GCM 1229) as cursorial examples; Sus (pig, Artiodactyla: GCM 885) as a subcursorial example; Ceratotherium (white rhino, Perissodactyla: GCM 575) as a mediportal example; and Elephas (elephant, Proboscidea: from Osborn, 1929) as the graviportal example. Limb bone indices were calculated from measurements using Helios Vernier dial calipers for small elements, and a meter stick for larger elements.
The following standard limb ratios were used (each is multiplied by 100): Radius/Humerus; Olecranon/Propodium; metacarpal (MC) III/Humerus; MC III/Radius; Tibia/Femur; Calcaneum/ metatarsal (MT) III; MT III/Femur; and MT III/Tibia. Measurements and indices are given in Table 1. Lever arms for selected locomotor muscles were also calculated (see Hildebrand, 1982, for an introduction to locomotor biomechanics.


The following description of Hyracodon skeletal elements supplements and reinterprets the detailed description by Scott (1941).


The glenoid fossa of AMNH 1176 is more oval in shape than previously described and what Scott calls the coracoid appears to actually be the supraglenoid tubercle. Therefore the coracobrachialis either had a different origin or was absent. In either case it appears that the biceps brachii was the better developed of the two muscles.


The humerus of Hyracodon is lightly built and relatively short compared to the ulna. Proximally the humerus is medio-laterally flattened. The greater tubercle is prominent and forms a deep bicipital groove with the lesser tubercle. A narrow deltoid tuberosity extends approximately half the length of the humeral shaft. Distinct muscle scars on the tubercles and tuberosity are not visible. Inferences about muscle development can be made based on the relative size of their sites of attachment. The supraspinatous and infraspinatous that insert onto the tubercles; and the deltoid that inserts onto the deltoid tuberosity were all well developed in Hyracodon. The lateral head of the triceps and the brachialis anticus both had large areas for attachment due to the flattening of the proximal humerus.
The teres major tuberosity is very small but does provide a point of reference and site of attachment for the teres major and lattisimus dorsi. Behind this tuberosity the medial head of the triceps had only a small area for its origin while the lateral head had much more area behind the large deltoid tuberosity. No evidence of an anconeus can be determined on the specimens studied.
Distally, Scott (1941) describes an intercondylar ridge that is not present in either AMNH 1176 or AMNH 9789. The ridge on the lateral condyle where the extensor carpi radialis originates is well developed. Just distal to this is a triangular area for the origin of the extensor digitorum lateralis and the extensor carpi ulnaris. There is also a prominence for the origin of the extensor communis digitorum. On the median condyle are origin sites for the flexor carpi ulnaris, flexor digitorum profundus and the flexor carpi radialis.

The radius of Hyracodon is about equal in length to the humerus in AMNH 1176 (right radius). The shaft is compressed in an anterior-posterior fashion and is curved anteriorly. A roughened area for the biceps brachii is present. On the anterior-medial surface another roughened area is visible for the attachment of the brachialis anticus. This second area is about equal in size to the scar left by the biceps. No evidence of an extensor metacarpi obliquus is present, but this might be an artifact of preservation in the specimens studied.


The ulna is closely associated with the radius its entire length. It has a laterally compressed, curved shaft, and a well developed, rugose olecranon process. The semilunar notch has a curve of about 120 degrees (Scott, 1941). The area for insertion of the triceps is prominent and the olecranon is posteriorly deflected. The surface from which the olecranon portion of the flexor digitorum profundus and flexor carpi ulnaris arise is concave in shape to increase surface for attachment. The distal end does not articulate with the lunar, and the cuneiform facet lacks a saddle.


The proximal phalanx of metacarpal III is long but compressed palmo-dorsally. The second phalanx is shorter than the proximal one by about one-half, but similarly shaped. The ungal is long and broad.

Innominate Bone
The pelvis of Hyracodon is long and lightly built with a long compressed iliac body that expands into the wing. The acetabulum is large, deeply concave and with an almost circular outline. The ischium is short and has a prominent tuberosity posteriorly on the ischial plate. The pubis is short, slender, rod-like and somewhat broad at the symphysis. The obturator foramen is a large elongate oval.


The femur of Hyracodon has a flattened posterior surface. It has a well developed, rugose greater and third trochanter. The free border of the third trochanter is rugose and anteriorly deflected. The lesser trochanter is marked by a low, rugose ridge on the posterior-medial surface. The trochanteric fossa is appreciable in size due to the size of the greater trochanter. The shaft is broad proximally, narrows at the third trochanter and widens again distally. Just above the medial and lateral condyles are large rugosities for the gastrocnemius. The large greater trochanter provides insertion for a well developed gluteus medius and gluteus minimus. Likewise the long third trochanter provides a large attachment site for the gluteus maximus. The large concave trochanteric fossa allows for insertion of a well developed obturator internus. The origin sites for the gastrocnemius are prominent above the condyles indicating the muscle was large. In the middle 1/3 of the posterior surface is a long, thin rugose area for the attachment of the adductors. No distinction between the brevis, magnus or longus can be made. Just above the internal condyle is a prominence for the insertion of the semimembranosus. Based on the size of this condyle the semimembranosus would have been a well developed muscle. No rugosities for the vastus internus and vastus externus are visible, however, there was ample space on the femur for their attachment.


The tibia of Hyracodon (AMNH 1480) is shorter than the femur of the same specimen. Scott (1941) states that the tibia is almost equal in length to the radius. The proximal end is large and triangular shaped, with a deep pit anterior for the insertion of the patellar ligament. Medial to the tibial crest is a roughened area for the insertion of the gracilis and sartorius. Lateral to the tibial crest is a fossa for the tibial origin of the tibialis anticus.


Scott (1941) described the fibula as being reduced and slender in size, and forming a narrow external malleolus. There is no tendency for the fibula to co-ossify with the tibia. There is, however, a long, close connection between the two.

The tarsus is high and lightly built like the carpus. The astragalus is short and has a relatively wide trochlea, and an open median groove. The two condyles do not rise sharply from the groove. The navicular facet has the typical Perrisodactyl saddle shape. The calcaneum tuber is laterally compressed but well developed. The cuboid is high proximo-distally. It articulates with the astragalus and calcaneum proximally , and with metatarsal IV distally. The entocuneiform is almost equal in size to the navicular and articulates proximally with it. Distally the entocuneiform articulates with metatarsals II and III. Metatarsal III is the longest of the three found in
Hyracodon (II, III, and IV). Metatarsal III is not quite as long as metacarpal III. Scott (1941) states that the phalanges of the pes are somewhat longer than those of the manus, but otherwise very similar.


The scapula of Ceratotherium is broad and well developed while the scapula of Hyracodon is more lightly built. Both genera can rotate the scapula as part of the forelimb, however, this is more likely a result of their lineage rather than a specific cursorial adaptation. The humeri of Ceratotherium and Hyracodon both provide a large area for the origin of the lateral head of the triceps. The medial head of the triceps has a small origin site in Hyracodon but a relatively larger site in Ceratotherium. The lateral condyle of Ceratotherium provides a relatively larger site for the origin of the extensor carpi ulnaris than is found in Hyracodon. This important flexor seems to be better developed in Ceratotherium than in Hyracodon. The proximal radii of both genera are expanded somewhat over the ulna. This expansion in Ceratotherium is not as extensive as in Hyracodon and seems to be an adaptation for weight bearing. The proximal articular surfaces of the radius of Ceratotherium are shallow, while Hyracodon has more deeply notched surfaces. These would seat the humerus more firmly into place and help restrict movement to a single plane. The distal end has simpler articular surfaces than Ceratotherium. The facet that articulates with the lunar of the carpus is concave. The facet for the scaphoid has both concave and convex components but they are not as pronounced as in Ceratotherium.
The ulna of Ceratotherium is much better developed than the ulna of Hyracodon. Both genera have a well developed, posteriorly deflected olecranon process. This is the insertion point for the triceps, and the origin for the flexor digitorum profundus. The olecranon of Hyracodon has a concave area to increase surface area for the origin of these muscles. The posterior deflection of the olecranon allows for a more complete extension of the elbow for weight bearing by the radius. In Ceratotherium the distal ulna is tightly associated with the radius, in Hyracodon the ulna/radius association is over the entire length of the radius. This and the reduction of the distal articular surfaces of the Hyracodon ulna inhibit rotations of the forearm. This reduction in rotational abilities would reduce weight distally by reducing the muscles associated with this action.
The manus of Ceratotherium has large carpals and metacarpals while Hyracodon has a lightly built manus. The manus of Ceratotherium is large and fleshy to distribute its weight. Hyracodon had a relatively lighter manus due to its smaller size and weight. This smaller manus size, along with the muscle reductions mentioned above, results in a lighter distal portions of the front limb for Hyracodon.
The femora of Ceratotherium and Hyracodon have well developed rugosities for origin of the gastrocnemius. The tibia and fibula of Hyracodon have a long close association with the fibula being reduced and slender. The tibia and fibula of Ceratotherium are well developed. The distal articulations on Hyracodon tibia are more deeply grooved than in Ceratotherium. This difference indicates Hyracodon was more tightly limited to movement in a single plane than Ceratotherium.
Both genera have well developed calcanea for insertion of the gastrocnemius. The tarsals and metatarsals of Ceratotherium are large and well developed, much like the carpals and metacarpals of the forelimb. Hyracodon has a lightly built pes similar to its manus. These arrangements result in a higher angular momentum possible for the hind feet of Hyracodon as was described for the forefeet.


Measurements of limb bones were used to generate various indices useful in determining the actual locomotor capabilities of Hyracodon. Table 1 presents a comparison of these figures for Hyracodon and various modern ungulates. The indices range from the cursorial Antilocapra, to the graviportal Elephas, accurately reflecting the different lifestyles of these two animals. The radius/humerus and tibia/femur indices for Antilocapra were 123.17 and 123.81 respectively. The MCIII/radius index was 105.45 and the MTIII/tibia index was 83.85. The olecranon/radius index was 21.34 and the calcaneum/MTIII index was 37.42. The same indices for Elephas are as follows: radius/humerus=84.57; tibia/femur=60.59; MCIII/radius=26.72; MTIII/tibia=22.33; olecranon/radius=23.81 (no calcaneum measurements were available for Elephas). Hyracodon grouped closer to Ceratotherium and Sus for all indices than to any of the truly cursorial ungulates.


Muscle vectors for selected limb muscles of Hyracodon and Ceratotherium were used to determine lever arms, li and lo, for the in and out forces (Figures 1 and 2). Computed ratios for these lever arms are compared to Lama in Table 2. The ratios for the flexor group of Hyracodon and Ceratotherium are similar, 29.97 and 27.32 respectively. The flexor group ratio for Lama was 32.89. The triceps ratio was also similar for the two rhinos (Hyracodon 19.35 and Ceratotherium 22.80). The triceps ratio for Lama was 12.36. The only ratio showing a marked difference between Hyracodon and Ceratotherium was the gastrocnemius ratios of 30.07 and 40.85 respectively. Hyracodon was, however, still far from the value for the cursorial Lama which had a gastrocnemius ratio of 11.28.


The osteological characteristics of Ceratotherium and Hyracodon limb elements show a spectrum of similarities and differences. Similarities between the two are the product of their common ancestry and to comparable locomotor adaptations. Differences between the two appear to be due primarily to size.

Most of the front-limb ratios of Hyracodon and Ceratotherium do not reflect any significant cursorial adaptations. The olecranon/epipodium ratio of 28.46 for Hyracodon does compare favorably with Dama at 25.08. This indicates a higher gear muscle system than the figure of 33.31 for Ceratotherium. The radius/humerus indices for Ceratotherium (98.13) and Hyracodon (99.52) show no lengthening of epipodium relative to propodium. The cursorial Antilocapra has a radius/humerus index of 123.17, and the graviportal Elephas has a radius/humerus index of 84.57. The index for MCIII/radius for Ceratotherium (48.71) and Hyracodon (54.56) compare well to the mediportal Hippopotamus (56.30, Osborn, 1929) but falls far below that of Antilocapra (105.45).

The hind limb of each genus is similarly lacking in cursorial adaptations. The tibia/femur index indicates a longer propodium than epipodium for both genera (68.13 for Ceratotherium and 83.72 for Hyracodon). These compare well to Hippopotamus (66.67, Osborn, 1929), Elephas (60.59), and Sus (86.69), but not at all with Antilocapra (123.81). The MTIII/femur index for Hyracodon (43.80) is somewhat higher than that of Ceratotherium (31.96) but far below that of Antilocapra (103.81).

The calcaneum/MTIII index for Ceratotherium (76.81) is almost two times that of Dama (38.86). The same index for Hyracodon is 65.49, somewhat lower than Ceratotherium but high compared to the cursorial animals. This figure and the olecranon/epipodium figure given above indicate relative importance of different gear ratio muscles for the different genera. The lower numbers indicate higher gear ratios for these two muscle systems which would be advantageous in cursors.

The triceps in force, out force and muscle vectors for Hyracodon and Ceratotherium (Figure 1, 2 and Table 2) produce li/lo ratios of 19.35 and 22.80 respectively. This indicates a slightly higher gear ratio for Hyracodon relative to Ceratotherium for the triceps, but is not comparable with the figure for Lama of 12.36. This represents a class I lever system and a short in force lever arm is needed to produce a high velocity at the out force contact point. The gastrocnemius represents a class II lever system. The gastrocnemius index for Hyracodon of 30.07 is much higher than that for the cursor Lama (11.28), but is well below that of Ceratotherium at 40.85.

The flexor group of the forearm represents a class III lever system. In class III levers lo is always longer than li therefore the best mechanical advantage comes from having li long relative to lo (i.e. a higher li/lo index). This is the situation cursors would benefit most from. Lama has a flexor index of 32.89 while Hyracodon has an index of 29.97 and Ceratotherium has an index of 27.32.


The muscle/bone systems, limb bone indices, and locomotor muscle in force/out force ratios all indicate that Hyracodon was not a cursorial animal. It would seem that Hyracodon was only slightly better at running than Ceratotherium. Although Ceratotherium can run fast for short distances, it has to use muscle power alone to attain such speeds, and cannot sustain them for long periods of time. Hyracodon does not have any significant adaptations to increase stride length necessary for a cursorial way of life. Its smaller size alone would make it easier for Hyracodon to run than Ceratotherium, but it was not a truly cursorial animal. Wild pigs or peccaries would be better locomotor analogs for Hyracodon than any of the modern cursorial ungulates. It would seem that Hyracodon was not subjected to the same selection pressures as the horses of its time. If Hyracodon had been a truly cursorial animal it should have flourished in the emerging savannas of the Miocene (Savage and Russell, 1983), instead of dying out.

Beddard, F.E. and Treves, F., 1889. On the Anatomy of Rhinoceros Sumatrensis: Proceedings of Zoological Society, p. 7-25.

Coombs, W.P., Jr., 1978. Theoretical Aspects of Cursorial Adaptations in Dinosaurs: The Quarterly Review of Biology, v. 53, p. 393-417.

Davis, D.D., 1964. The Giant Panda: A Morphological Study of Evolutionary Mechanisms: Fieldiana, Zoology, v. 3, 339 p.

Gregory, W.K., 1929. The Muscular Anatomy and the Restoration of the Titanotheres; in Osborne, H.F. The Titanotheres of Ancient Wyonimg, Dakota, and Nebraska: v. 2, p. 703-726, U.S. Geological Survey. Monograph 55.

Hildebrand, M., 1982. Analysis of Vertebrate Structure: John Wiley and Sons, New York, 654 p.

Osborn, H.F., 1929. The Titanotheres of Ancient Wyoming, Dakota, and Nebraska: United States Geological Survey, Monograph 55, 953 p.

Prothero, D.R., Manning, E.M., and Hanson, C.B., 1986. The Phylogeny of the Rhinocerotoidea (Mammalia, Perissodactyla): Zoological Journal of the Linnean Society, v. 87, p. 341-366.

Scott, W.B., 1941. The Perrisodactyla; in Scott, W. B. and Jepsen, G. L. The Mammalian Fauna of the White River Oligocene, Part 5: Transactions of the American Philosphical Society, p. 747-946.

Sisson, S., 1975. Equine Myology; in Sisson and Grossman's; The Anatomy of the Domesticated Animal 1: Fifth Edition, W B. Saunders Company. 1211p.

Wall, W.P., 1989. The phylogenetic history and adaptive radiation of the Amynodontidae (Perissodactyla, Rhinocerotoidea); in Evolution of the Perissodactyla: Prothero, D. and Schoch, R., eds., Oxford Univ. Press., p. 341-354.

Wall, W.P. and Manning, E., 1986. Rostriamynodon grangeri New Genus, New Species of Amynodontid (Perissodactyla, Rhinocerotoidea) With Comments on the Phylogenetic History of Eocene Amynodontidae: Journal of Paleontology, v. 60(4), p. 911-919.

Windle, B.C.A. and Parsons F.G., 1901. On the Muscles of the Ungulata: Proceedings of the Zoological Society, p. 656-703.

Windle, B.C.A. and Parsons F.G. 1903. On the Muscles of the Ungulata: Proceedings of the Zoological Society, p. 260-296.


Figure 1. A and B, medial views of right forelimb of Hyracodon. C and D, medial views of right forelimb of Ceratotherium. In A and C the in force (li) and out force (lo) of the resultant force (R) of the lateral and medial heads of the triceps (T1) and the long head of the triceps (T2) are shown. In B and D the muscle vector for the flexor group (F) is illustrated, its in force (li), and out force (lo).
Figure 2. A, medial view of left hindlimb of Hyracodon; B, medial view of left hindlimb of Ceratotherium; showing the muscle vector for the gastrocnemius (G), its in force (li), and out force (lo).

Table 1. Limb bone measurements (in mm) and indices.  Data for Elephas from Osborn (1929)MEASUREMENT		HYRACODON	CERATOTHERIUM	SUS	DAMA	TRAGELAPHUS	ANTILOCAPRA	ELEPHAS    	AMNH 1176	GMC 575	GMC 885	GMC 576	GMC 577	GMC 1229 Humerus	209.5		374.5	208.0	185.5	153.0	164.0	810.0Radius		208.5		367.5		168.0	210.5	171.0	202.0	685.0Epipodium		208.5		367.0		168.0	210.5	171.0	202.0	685.0 Olecranon		59.3		122.2		52.8	35.0MC III		113.7		179.0		77.0	199.5	169.5	213.0	183.0Femur		258.0		511.5		248.0	237.5	213.5	210.0	1020.0Tibia		216.0		348.5		215.0	277.5	242.5	260.0	618.0Calacneum		74.0		125.5		87.2	70.0	MT III		113.0		163.5		77.0	224.5	177.5	218.0	138.0Radius/Humerus	99.52		98.13		80.77	113.48	111.76	123.17	84.5Olecranon/Epipodium	28.46		33.31		25.08	20.48	MC III/Humerus	54.3		47.8		37.02	107.55	110.78	129.88	22.59MC III/Radius	54.56	48.71		45.83		94.77	99.12	105.45	26.72Tibia/Femur	83.72	68.13		86.69		116.84	113.58	123.81	60.59Calcaneum/MT III	65.49		76.81		38.86	39.44	MT III/Femur	43.8	31.96		31.05		94.53	83.14	103.81	13.53MT III/Tibia	52.31	46.92	35.81	80.9	73.2	83.85	22.33	Table 2.  Muscle lever arms and gear ratios (li/lo). Measurements in mm.   GENUS             TRICEPS    li/lo    FLEXOR    li/lo    GASTROCNEMIUS    li/loHyracodon  li	26.68	19.35	16.65	29.97	17.21	 30.07                lo	  137.89	55.56		57.23Ceratotherium  li	28.89	22.80	13.35	27.32	24.97	40.85                     lo	126.70		48.87		61.12Lama  li	64.00	12.36	7.40	32.89	65.00	11.28         lo	518.00		22.50		576.00
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