size than Anchitherium, to be considered immediately. The odontoid process of the axis is just beginning to assume the characteristic spout-like shape of that of the existing Horse and many modern Ungulates. The median digit of both fore- and hind-limbs has become greatly enlarged as compared with the corresponding digit of earlier forms.
It is held, however, that Anchitherium is not on the direct line of descent either in America or in Europe, in both of which it occurs. Its teeth are in some respects less Horse-like than in some of the more ancient genera, to which the converse would be expected on the descent theory. Its hoofs are much elongated and flattened, a mark of specialisation and not appropriate to a creature holding an intermediate position in the equine series. Both the American (A. equinum) and the European species (A. aureliense) are of very large size, larger than its successors, and such "alternations in bulk are unlikely."
The genus Desmatippus of Professor Scott[[155]] fills in the gap between Miohippus and Protohippus. The molars and premolars are brachyodont, but there is a thin deposit of cement in the tooth valleys, leading towards the more complete filling of these valleys with cement, which is found in Protohippus. This genus of Horses, of which there is at present but one species, D. crenidens, was three-toed, and "the lateral digits, so far as can be judged by fragmentary remains, were still fairly developed, and though much more reduced than in Miohippus, appear to be somewhat less so than in Protohippus."
To recapitulate, the following is the probable series of equines in America—Mesohippus, Miohippus, Desmatippus, Protohippus.
The development of the limbs of the Horse shows a most interesting series, of stages, which correspond in part to the ancestral forms which palaeontology seems to prove to be the line of the descent of our existing Equidae. This matter has recently been elucidated by Professor Ewart, who details the following facts and comparisons:—
In the youngest embryo (about 20 mm. in length) the humerus is somewhat curved, and considerably longer than the radius and carpus taken together. The first-named bone is shorter in the adult, and the proportions of that bone in the young as well as its curvature are suggestive of that ancient
Ungulate Phenacodus (see p. [202]). In the next stage (an embryo of 25 mm.) the humerus has slightly decreased in proportionate length, and has come to be more like that of Hipparion. In both of these embryos it should be noted that the ulna is complete and separate from the radius. In the second of the two it has more distinctly acquired the form which it will possess in the adult. The second metacarpal—one of the splint bones of the adult—is tipped with a small nodule of cartilage, which is clearly the representative of one or more of the phalanges belonging to that digit.
Fam. 2. Tapiridae.—The Tapirs may be distinguished from the Horse and from the Rhinoceros tribe by a few characters, which are as follows:—
The dentition is generally the full one of forty-four teeth. The premolars in the more ancient forms are unlike the molars, but like them in more recent forms. The molars of the upper jaw have two crests parallel and united by an outer crest. The fore-feet have four, the hind-feet three toes.
The family is fully as ancient as that of the Equidae, but the specialisation of the toes never advances so far. The modern representatives of the order are, so far as the feet are concerned, in the condition of very early representatives of the equine stock. Nor do the teeth of the Tapirs ever reach the complicated pattern of that presented by at least the modern Horses, or indeed of the Palaeotheres. Apart from this it is not an easy matter to distinguish accurately between these several families, including the Lophiodontidae, which, as already mentioned, is placed nearer to the Tapiridae than to the Palaeotheriidae. Indeed the differentiation of these two families, the Tapiridae and the Lophiodontidae, seems to be a matter of the greatest difficulty. The difficulty is well emphasised by the fact that naturalists disagree most profoundly as to the relations of various genera of extinct Tapir-like animals. For Mr. Lydekker the genus Lophiodon includes also the American genera Isectolophus and Systemodon, which are placed by Zittel in the sub-family Tapirinae as opposed to Lophiodontinae, which contains Lophiodon and Helaletes. The existing Tapirs can be differentiated from the existing Horses with great ease, as the following account of the existing genera will show.