moving succession does not require, or allow time for, their replacement by premolars. It must be noted, however, that, in the Mastodon in some respects the least specialised in tooth-structure, the M. americanus of North America, no vertical succession of the

molars has yet been observed, although vast numbers of specimens have been examined.

The Mastodons have fewer ridges on their molar teeth than the Elephants; the ridges are also less elevated, wider apart, have a thicker enamel-cover

ing, and scarcely any cement filling up the space between them. Sometimes (as in M. americanus) the ridges are simple transverse wedge-shaped elevations, with straight or concave edges. In

other species the sum- FIG. 187.-Oblique side and crown view of the last upper mits of the ridges are molar of Mastodon arvernensis. (From Owen.) more or less subdivided into conical cusps, and may have accessory cusps clustering around them (as in M. arvernensis, see Fig. 187). When the apices of these are worn by mastication, their surfaces present circles of dentine, surrounded by a border of enamel, and as the attrition proceeds different patterns are produced by the union of the bases of the cusps, a trilobed or trefoil form being characteristic of some species (Fig. 188).

As already mentioned, certain of the molariform teeth of the

middle of the series in Mastodons have the same number of principal ridges, those in front of them having fewer and those behind a greater number. These teeth were distinguished as "intermediate" molars by Dr. Falconer, and are three in number, namely the last milk-molar and the first and second true molars (or the The number of ridges

third, fourth, and fifth of the whole series). on these intermediate molars is nearly always three or four, and the tooth in front has usually one fewer and that behind one more, so that the ridge-formula of most Mastodons can be reduced either to 1, 2, 3, 3, 3, 4, or 2, 3, 4, 4, 4, 5. The former characterises the section called Trilophodon (of which an intermediate molar is shown in Fig. 188), and the latter that called Tetralophodon by Dr. Falconer. These divisions are very useful, as under one or the other all the present known species of Mastodon can be ranged, but observations. upon a larger number of individuals have shown that the number of ridges upon the teeth is not quite so constant as implied by the formulæ given above. Their exact enumeration is even difficult in many cases, as "talons" or small accessory ridges at the hinder end of the teeth occur in various stages of development, until they take on the character of true ridges. Transitional conditions have also been shown, at least in some of the teeth, between the trilophodont and the tetralophodont forms, and again between the latter and what has been called a "pentalophodont" type, which leads on towards the condition of dental structure characteristic of the true Elephants.

The range of the genus Mastodon in time was from the middle of the Miocene period to the end of the Pliocene in the Old World, when it became extinct; but in America several species-especially the one best known, owing to the abundance of its remains, which has been variously called M. americanus, M. ohioticus, and M. giganteus -survived to a late Pleistocene period.

The range in space will be best indicated by the following list of some of the better known species. (1) Trilophodont seriesM. angustidens, borsoni, pentelici, turicensis, from Europe; M. falconeri 1 Also found beyond the extreme north-western frontier of India.

and pandionis, from India; M. americanus, obscurus, and productus, North America; and M. cordillerum and humboldti, South America. (2) Tetralophodont series-M. arvernensis, M. longirostris, from Europe; M. latidens, sivalensis, and perimensis, from India; M. mirificus, from North America. Mastodon arvernensis and M. longirostris, together with a trilophodont species, occur in the crag-deposits of Norfolk and Suffolk.

Family DINOTHERIIDÆ.

An extinct family distinguished from the Elephantida by the whole series of permanent cheek-teeth being in use at the same time. Dinotherium.-Dentition of adult: i, c 8, p, m } = 22; all present at the

same time, there being no horizontal succession, but the premolars replacing milkteeth in the ordinary manner. The presence or absence of upper incisors has not yet been clearly ascertained.

Lower incisors

large, conical, descending, and slightly curved backwards, implanted in a greatly thickened and deflected beak or prolongation of the symphysis. In section they do not show the decussating striæ characteristic of Mastodons and Elephants. Crowns of molars carrying strong transverse, crenulated ridges, with deep valleys between, much resembling the lower ones of the Tapirs. Ridge-formula of the permanent molar series: 2, 2, 3, 2, 2. The three ridges of the first true molar are constant in both upper and 1, 2, 3, molars. lower jaws, although it is quite an anomalous character among Proboscideans for this molar to have more ridges than those which come behind it. The last milk-molar has also three ridges, the

FIG. 189.-Skull of Dinotherium

giganteum, from the Lower Pliocene (After Kaup.) p, 3, 4, premolars;

of Eppelsheim, Hessen-Darmstadt.

1 Kaup, Isis, vol. xxii. p. 401 (1829).

penultimate but two. The cranium is much depressed, with comparatively little development of air-cells. The remainder of the skeleton is imperfectly known, but apparently agrees in its general characters with that of the other Proboscideans.

Remains of Dinotherium giganteum, an animal of elephantine proportions, strikingly characterised by the pair of huge tusks descending nearly vertically from the front of the lower jaw, were first discovered at Eppelsheim, near Darmstadt, and described by Kaup. They have since been met with in various Lower Pliocene and higher Miocene formations in the south of Germany, France, Greece, and Asia Minor. Closely allied forms also occur in the Lower Pliocene and Upper Miocene of India, but none are known from America.

Suborder AMBLYPODA.

Uintatherium.1-Among the most remarkable of the comparatively recent discoveries in the higher Eocene formations of the western states of North America has been one of a group of animals of huge size, approaching that of the largest existing Elephants, presenting a combination of characters quite unlike those known among other recent or extinct creatures, and of which there were evidently many species living contemporaneously, but all of which became extinct before the close of the Eocene period. To form some idea of their appearance, we must imagine animals very elephantine in general proportions and in the structure of their limbs. The feet had five short toes. The tail, as in the Elephants, was long and slender, but the neck, though still short, was not so much abbreviated as in the Proboscideans, and there is no evidence that these animals possessed a trunk. The head differed greatly from that of the Elephants, being long and narrow, more like that of a Rhinoceros, and, as in that animal, was elevated behind into a great occipital crest, and it had developed upon its upper surface three pairs of conspicuous, laterally diverging protuberances—one pair in the parietal region, one on the maxillaries in front of the orbits, and one (much smaller) near the fore part of the elongated nasal bones. Whether these were merely covered by bosses of callous skin, as the rounded form and ruggedness of their extremities would indicate, or whether they formed the bases of attachment for horns of still greater extent, like those of the Rhinoceros or of the Cavicorn Ruminants, can only be a matter of conjecture. There were no upper incisors, but usually three on each side below, of comparatively small size, as was also the lower canine. A huge, compressed, curved, sharp-pointed canine tusk, very similar in form

1 Leidy, Proc. Ac. Nat. Sci. Philad. 1872, p. 169.

and position to that of the Musk-Deer, descended from each side of the upper jaw. These were present in both sexes, but very much smaller in the female, as was also the flange-like process of the lower jaw by which they were guarded. Behind these, and at some distance from them, were on each side above and below six cheek-teeth, of comparatively small size, placed in continuous series, each with a pair of oblique ridges conjoined internally and diverging externally in a V-like manner, and provided with a stout basal cingulum. The normal dental formula was therefore if, c, ps, m = 34; and the dentition had thus already attained a remarkable degree of specialisation, although the brain was smaller and more rudimentary in characters than in almost any other

FIG. 190.-Skeleton of Uintatherium mirabile.natural size. (From Marsh,
Am. Journ. Sci. vol. xii. pl. 2.)

known mammal. In its comparative length and the absence of a third trochanter the femur of these animals resembles that of the Proboscidea. The first discovered evidences of the existence of animals of this group were described by Leidy in 1872, under the name of Uintatherium (from the Uinta mountains, near which they were found). Subsequently the names Dinoceras, Tinoceras, Loxolophodon, etc., have been applied to various members of the group, but the characters by which they are distinguished do not seem of sufficient importance to allow of their separation from the type genus Uintatherium.1

Coryphodon.2-Another interesting form referred to this suborder is Coryphodon, which appears to connect the Uintatheriida with the most primitive Perissodactyla. It was first described by Owen in

1 For detailed descriptions and figures of this group, see Marsh, "Monograph of the Dinocerata," Rep. U.S. Geol. Surv. vol. x. (1884).

2 Owen, Brit. Foss. Mamm. and Birds, p. 299 (1846).