humerus, radius and complete manus, the left innominate bone, the femur, tibia and fibula and complete pes; of the axial skeleton is included the atlas, axis and several scattered vertebral centra; also a series of peculiarly coalesced post-sacral or caudal vertebræ.

It is possible that the condition of the caudals found in this specimen was pathological, but they bear a perfectly normal appearance, and point to a unique caudal appendage. Behind the sacrum (which is wanting) we find remains of sub-cylindrical caudal centra, followed by a series of flattened centra which rapidly diminish in size posteriorly; these flattened centra become united together by the adjoining faces, and send out wide flattened transverse processes; but the most peculiar feature is the coalescence of the neural spines and laminæ into a long solid ridge which tapers off posteriorly and anteriorly, and renders this portion of the tail absolutely rigid. The only interpretation of this structure seems to be that the proximal portion of the tail was flexible, while the distal half formed a broad solid plate. The humerous suggestion has been made that this appendage supplied Coryphodon with a steering apparatus while swimming in the Big Horn Lake; it is impossible to make any serious conjecture as to the purpose which such a tail subserved.

THE FOOT STRUCTURE OF CORYPHODON.

The fore and hind feet of Coryphodon have been figured and described by both Marsh and Cope, yet neither of these authors has given an accurate idea of their real structure. Marsh' has figured both the manus and pes in the digitigrade position like the feet of the Elephant. In Cope's latest paper upon the Amblypoda' he says of Coryphodon and other members of the order, "The feet are always short and plantigrade." Elsewhere," however, he speaks of the movements of the Coryphodons as resembling those of the Elephant (i. e., digitigrade); Cope's figures correspond with those here published, but fail to represent the actual position of the feet.

1 Dinocerata, p. 184, figs. 150, 151.

American Naturalist, Nov., 1884, p. 1110.

3 Tertiary Vertebrata, p. 524.

The fact is that the positions of the fore and hind feet of Coryphodon were absolutely different, the fore foot was digitigrade

like that of the Elephant, the hind foot was plantigrade like that of the Bear. In other words the carpus was entirely raised from the ground, and the manus rested upon the distal ends of the metacarpals and upon the spreading phalanges, while the calcaneum and tarsus rested directly upon the ground together with the entire plantar surface of the foot. This substantial difference between the advanced state of evolution of the fore foot, and retarded evolution of the hind foot, is of great interest; it is clearly shown in the

We may therefore restate the characters of the feet of the Coryphodon as follows:

The absence of the fibular facet upon the calcaneum has been observed in a specimen in the Smithsonian collection; also the cuneiform resting upon the fifth metacarpal. These differences,

as well as the variations in the tibiale facet, require fuller investigation.

It will be noted that the manus has many points of functional parallelism with that of the Elephant, especially the enlargement of the lunar. The pes, on the other hand, is much more primitive than that of the Elephant.

Fig. 15. Coryphodon anax. Left hind foot, external view of the foot in the natural position. One-fourth natural size.

The figure of the pes of Coryphodon given by Marsh is somewhat similar to that of Uintatherium (Dinoceras), and is wholly different from those belonging to Coryphodon in the collections we have examined, since the astragalus is represented as covering the entire upper surface of the cuboid, and, as above noted, the foot is represented as digitigrade instead of plantigrade; the figure of the manus agrees with those we have examined, except that it is of a higher, narrower type. This author rarely errs in his figures, but in this case it would appear either that the astragalus is wrongly figured, or that these feet belong to some of the Dinocerata.

THE HOMOLOGIES OF THE MOLAR CUSPS IN CORYPHODON.

While leaving to Mr. Earle the full discussion of the evolution of the molars of the Coryphodons, as indicated by the numerous/ variations in the molar pattern, the homologies of the molar elements with those of Pantolambda and of the Perissodactyla may be briefly pointed out. Cope is correct in the interpretation of the lower molar elements, but his interpretation of the upper molar homologies is much more uncertain. We have three means of determining the latter.

1. The comparison with the molars of Pantolambda.

2. The actual typical structure of the molars.

3. The vestiges of ancestral structure seen in specific variations.

The molar teeth of Pantolambda are tritubercular.

There are

two external subequal selenoid cusps, the paracone and metacone, separated by a slender median buttress, or mesostyle, with a strong anterior buttress, or parastyle; there are two faintly-marked intermediate tubercles, protoconule and metaconule, and a strong selenoid protocone; there are also anterior and posterior cingula.

The question is how has this selenodont molar been transformed into the lopho-selenodont molar of Coryphodon? Cope considers that the anterior crest, or protoloph, of the Coryphodon molar represents the union of the protocone and parastyle; that the median external cusp is the greatly reduced paracone, while the postero-external crescent certainly is the paracone. This theory is rendered clear by a study of the Anchitherium molar, and it is supported by this comparison, because it is shown in the Equida that where there are two external crescents the protoloph is formed by the union of the protocone with the protoconule and parastyle.

Another interpretation is the following: that the protoloph of Coryphodon represents the union of the protocone with the paracone; that the median external buttress represents the mesostyle; this is supported by the fact that the antero-external cusp in Coryphodon is often sub-crescentic as in C. radians.

Upon the whole, however, the evidence seems to favor Cope's theory. It is interesting to observe that in examining a large number of Coryphodon molars we find traces of the protoconule and metaconule; also faint traces of the mesostyle.

The anterior crest in the Coryphodons is therefore probably homologous with the anterior crest in the Perissodactyla, especially in the form exhibited in the Equida; the posterior crest (or metaloph) of the Perissodactyla is wholly wanting in Coryphodon; the external crest (ectoloph) of the Coryphodontia is homologous with the ectoloph or the posterior crest in the Dinocerata.

Order PERISSODACTYLA.

Family TAPIRIDE.

Subfamily SYSTEMODONTINÆ.

Genus Systemodon Cope.

Dentition, 1, 1, . Superior dental series continuous.

First lower pre

molar contiguous to canine, followed by narrow diastema. Third and fourth superior premolars with two external cusps and a single internal lobe. Paracone and metacone subequal, conic. Protoloph and metaloph complete. Large third lobe upon third lower molar.

-ml

hy a

This Eocene Tapir ranks next to Hyracotherium in abundance during the Wahsatch period. In this collection there are nearly fifty specimens which may be referred to it, including numerous lower jaws and several fragmentary skulls. The only skeleton preserved embraces portions of a hind foot which unfortunately is of somewhat doubtful reference, as the associated teeth are only partially preserved. As Cope has shown, there are strong grounds for considering Systemodon an ancestor of the Tapirs, and certainly the Tapir stamp in the molar teeth is most striking, as shown in the accompanying figures (Figs. 16, 17) yet our opinion must be reserved until we learn the foot structure with certainty. There are two species, S. tapirinus Cope, and S. semihians Cope. Our material enables us to distinguish them somewhat more clearly than Cope has done as follows:

A

med
B

end

Fig. 16. Systemodon and Tapirus, first Superior and Inferior Molars. A. Systemodon semihians, showing Primary Cusps and Parastyle. B, Tapirus indicus, showing complete

crests.