Multituberculata.—The name Multituberculata has been proposed for the group exhibiting the type of dentition last mentioned, and is generally adopted, although the term Allotheria has been also suggested. The essential characteristic of the dentition of this group is the presence of a single scalpriform incisor on each side of the lower jaw (Fig. 25) and of one larger incisor, and in some instances of one or two smaller ones in each premaxilla (Fig. 24). These incisors are separated by an interval or diastema from the first of the premolars. The true molars, and in some instances the premolars (Fig. 24), are characterised by having longitudinal rows of tubercles separated by one or more grooves; there being either two or three of these rows in the upper molars of those forms in which these teeth are known, while there are, at least usually, only two in those of the lower jaw. In other cases the premolars are of a secant type, with a highly convex cutting-edge, and usually either serrated or obliquely grooved (Figs. 25, 26). From a certain resemblance between these secant premolars and those of some of the smaller Macropodidæ it was at one time considered that we had in these mammals representatives of Diprotodont Marsupials. The great difference in the structure of the molar teeth of these forms, coupled with the circumstance that when the number of upper incisors is reduced below three it is the second in place of the first which becomes enlarged and opposed to the incisor of the lower jaw, seems to prevent the acceptation of this view. Moreover, in their peculiar structure the molars seem, on the whole, to make a nearer approximation to the teeth of Ornithorhynchus than to any other known mammal; and it has accordingly been suggested that the Multituberculata may really represent an order of Prototheria. Some support is afforded to this suggestion by certain fragmentary bones from the Cretaceous of the United States, which are regarded by Marsh as parts of a coracoid and interclavicle. The peculiar character of the whole dentition of these forms indicates that if they are really Prototherians they cannot be regarded as primitive and ancestral types.

It would be beyond the scope of the present work to describe in detail, or even to mention the names of all the members of this group, and it will therefore suffice to refer to a few of the principal types. Of the forms with tubercular premolars the best known is the genus Tritylodon (Fig. 24), which occurs typically in beds of Lower Mesozoic in South Africa, but is also known from the Trias of Stuttgart. In the Stonesfield Slate, near Oxford, which belongs to the lower part of the Jurassic system, and is separated from the Trias by the intervening Lias, a fragmentary jaw with three teeth (Fig. 27) appears to indicate an allied type, the teeth having three longitudinal ridges separated by grooves. In the Purbeck beds of Dorsetshire, forming the top of the Jurassic system, we find another member of this group, which has been described as Bolodon, closely allied to which is Allodon of the Upper Jurassic of the United States.

The first discovery of the remains of Mesozoic mammals was made in the Keuper or Upper Trias of the Rhætian Alps in Bavaria. In 1847 Professor Pleininger of Stuttgart, while sifting some sand from the Keuper of Diegerloch and Steinenbronn, found, among an immense mass of teeth, scales, and unrecognisable fragments of skeletons of fish and saurians, two minute teeth, each with well-defined, enamelled, tuberculated crowns and distinct roots, plainly showing their mammalian character. These were considered by their discoverer to indicate a predaceous and carnivorous animal of very small size, to which he gave the name of Microlestes antiquus. Subsequently Mr. C. Moore discovered in a bone bed of Rhætic (topmost Trias) age, filling a fissure in the Mountain Limestone at Holwell, near Frome in Somersetshire, various isolated teeth with their crowns much worn, but apparently including both upper and lower molars and a canine, which are assigned by Sir R. Owen to Pleininger’s genus Microlestes, and described specifically as M. moorei. Under the name of Hypsiprymnopsis rhæticus, Professor Boyd Dawkins described a single tooth with two roots discovered in the Rhætic Marlstone at Watchet in Somersetshire. Sir R. Owen referred the latter tooth to Microlestes, and if its describer is right in regarding it as a much worn premolar of the type of those of Plagiaulax (Fig. 25) there would be evidence that Microlestes was closely allied to the latter, from the molars of which those of Microlestes are scarcely distinguishable.

Plagiaulax, of the Dorsetshire Purbeck (Figs. 24, 25), is at once distinguished from Tritylodon by its secant premolars, which, as already mentioned, recall those of some of the Macropodidæ, although readily distinguished by the convexity of the cutting edge and their oblique grooving. This remarkable and highly specialised type has been the occasion of one of the most interesting discussions on the inferences which may be drawn as to the affinities and habits of an otherwise unknown animal from the structure of a small portion of its organisation which occurs in the annals of natural history—a discussion carried on with great ability, ingenuity, and wealth of illustration on both sides. Dr. Falconer maintained that it was more nearly allied to the Rat-Kangaroo (Potorous or Hypsiprymnus) than to any other existing form, and that, as it is known that these animals feed upon grass and roots, “it may be inferred of Plagiaulax that the species were herbivorous or frugivorous. I can see nothing in the character of their teeth,” he adds, “to indicate that they were either insectivorous or omnivorous.” Sir R. Owen, on the other hand, from the same materials came to the conclusion that “the physiological deductions from the above-described characteristics of the lower jaw and teeth of Plagiaulax are that it was a carnivorous Marsupial. It probably found its prey in the contemporary small insectivorous mammals and Lizards, supposing no herbivorous form like Stereognathus to have co-existed during the Upper Oolitic period.”

It is impossible here to give at any length the arguments by which these opposing views are respectively supported, but it may be indicated that the first-mentioned is strongly countenanced by the consideration of the following facts: (1) all existing Marsupials may be divided, so far as their dentition is concerned, into two groups—(a) those which have a pair of large more or less procumbent incisors close to the symphysis of the lower jaw, and rudimentary or no canines (diprotodont dentition), and (b) those which have numerous small incisors and large pointed canines (polyprotodont dentition); (2) the vast majority of the former group are purely vegetable feeders, and almost all of the latter are carnivorous or insectivorous; and (3) Plagiaulax, so far as its structure is known, shows an analogy with the former group; and, as we have no sure basis for inferences as to the habits of an unknown animal, but the knowledge of the habits of such as are known, we have no grounds for supposing that its habits differed from those forms having an analogous type of dental structure.[31]

Allied types, such as Ctenacodon, are also met with in the Upper Jurassic of North America; and the Plagiaulacidæ also persisted into the lower part of the Eocene division of the Tertiary period; Neoplagiaulax being a Tertiary form common to Europe and the United States, while Liotomus and Ptilodus are at present known only from the latter country.

The present group is also represented in the upper Cretaceous of the United States by Selenacodon (Meniscoëssus in part), Cimoliomys, etc. Polymastodon, of the Lowest or Puerco Eocene of New Mexico is the largest known form, and is characterised by the presence of only one premolar and the elongated molars. The angle of the mandible is inflected after the Marsupial fashion.

Polyprotodont Types.—The second type of mammalian dentition found in the Mesozoic period resembles that occurring among recent Polyprotodont Marsupials—that is to say there are at least three lower incisors, the canines are well developed, and the premolars and molars are cuspidate, the number of the latter reaching in some cases to seven or eight. There has been much discussion as to the taxonomic position of these forms, and while the majority of writers admit the Marsupial affinities of at least a moiety, it has been contended that others indicate distinct ordinal groups more or less closely allied to the Insectivora. At present, however, there is no decisive evidence to support such a view. Important proof of the Marsupial affinity of one of these forms is afforded by the replacement of the teeth, which appears to be of the same nature as in the existing Marsupials, that is to say, the last premolar alone is preceded by a milk-tooth.

The most generalised forms appear to be Dromatherium and Microconodon, from Lower Mesozoic beds in the United States, of which enlarged views of the teeth are given in Fig. 4 (1, 2), p. 31. Professor Osborn points out the extremely simple character of these teeth, and it is quite possible that these forms may prove to be Prototheria. There are three premolars and seven molars in the lower jaw of Dromatherium.

A common form in the Purbeck of Dorsetshire is Triconodon (Triacanthodon), in which the formula of the lower teeth is i 3, c 1, p 4, m 3-4. A lower jaw is shown in Fig. 28, and an enlarged view of a molar tooth in Fig. 4 (5). The molar teeth consist of three flattened cones placed in the same antero-posterior line, those of the upper and lower jaw being alike. Priacodon, of the Jurassic of the United States, is probably inseparable from Triconodon. In the genus Phascolotherium (Fig. 29) of the Lower Jurassic Stonesfield Slate, the lower teeth may be classified as i 4, c 1, p 3, m 4, the premolars and molars being much alike. The molars approximate to the type of those of Triconodon, but the anterior and posterior cones are relatively smaller. Like that of the last-named genus, the mandible of Phascolotherium is remarkable for the extremely low position of its articular condyle. In Amphilestes (Fig. 30) of the Stonesfield Slate the molars appear to be of the same general type as those of Phascolotherium, but are more numerous, although their exact number cannot be determined. A somewhat different type of lower molar is displayed by the genus Amblotherium, of the Dorsetshire Purbeck, to which Amphitherium of the Stonesfield Slate was probably allied. This type of tooth is shown in Fig. 4 (8, 9, 12) p. 31, and, as there stated, represents that modification of the tritubercular type known as the tubercular sectorial. The three primitive tritubercular cusps form what is known as the blade of the tooth, behind which there is the talon or hypocone. A similar form of molar occurs in the existing Opossums and Bandicoots. The number of lower teeth in Amblotherium is i 4, c 1, p 4, m 7-8. Numerous allied types, such as Achyrodon and Dryolestes occur in the Upper Jurassic of Europe or the United States, while from only one side of the jaw being exposed in each case so-called genera like Stylodon and Stylacodon have been formed upon specimens showing the opposite side to that which is exposed in the types of Amblotherium and Amphitherium. The only parallel among existing forms to the excessive number of molar teeth found in these Mesozoic genera occurs in the Marsupial genus Myrmecobius, of which a description is given in a succeeding chapter. Jaws more or less closely resembling those described under the names mentioned above are also found in the uppermost Cretaceous of the United States. A feature common to these Mesozoic mammals and Myrmecobius and some other existing forms is the presence of a narrow channel on the inner side of the mandibular ramus known as the mylohyoid groove (Fig. 29).

The last type of molar dentition occurring among the Mesozoic Mammalia is that found in the lower jaws (Fig. 31), upon which the genus Spalacotherium was established, the upper jaws, described as Peralestes, being apparently referable to the same animal. Upper and lower teeth of this form are represented in Fig. 4 (6, 7), p. 31, where they are described as typical examples of the tritubercular type of molars, the upper teeth having one inner and two outer cusps, and the reverse condition obtaining in the lower ones. This type of molar presents a marked resemblance to that found in the existing Insectivorous genus Chrysochloris; the number of lower teeth in Spalacotherium is, however, i 3, c 1, p + m 10, by which it is widely distinguished from all the Insectivora. Menacodon, of the Upper Jurassic of the United States, appears to be allied to Spalacotherium.

Tertiary Mammals.—The more important types of Tertiary mammals will, as already mentioned, be noticed under the heads of the groups to which they are severally allied; but a few general remarks on this subject may be advantageously recorded in this chapter. In the first place, it may be observed that the comparatively scanty evidence of mammalian life hitherto yielded by the Cretaceous, coupled with the number and variety of forms approximating to the existing groups found even in the lowest Tertiary, indicates a great imperfection of the geological record. At present, indeed, we have no decisive evidence of the existence of any members of the Eutherian subclass previously to the Tertiary; but it can hardly be doubted that in some part of the world they had made their appearance before that epoch. The Eutherian mammals of the lowest Eocene, both in Europe and the United States, are of an extremely generalised type; and although many of them approximate to existing groups, they show such a combination of characters, now restricted to individual groups, as to indicate that several of the various orders into which the subclass is now divided were at that period very intimately connected. A marked feature of these early Eutherians is the prevalency of trituberculism in the dentition, not less noteworthy being the frequent occurrence of pentadactylism in the feet, while many of the individual bones were devoid of the grooves and ridges found in those of later types. By the time that we reach the upper division of the Eocene period, such as the horizon of the well-known gypsum of the Paris basin, nearly all the chief groups of mammals had become clearly differentiated from one another, although their representatives were usually more generalised than their existing allies. From this date to the later geological periods there is a gradual approximation to the types of mammalian life existing at the present day.

In addition to the features of trituberculism and pentadactylism so characteristic of the oldest known Eutherians, we may notice some other points in connection with the earlier types. Thus the older Tertiary mammals, as we have already stated, had relatively smaller and simpler brains than the later types, so that a gradual evolution in this respect may be traced from the Eocene to the Pleistocene. Again, there is a great tendency among the Eocene forms to a retention of the typical Eutherian dental formula noticed on page 25, and also to the absence of an interval, or diastema, in the dental series. Concomitantly with this feature we may notice the short crowns and simpler structure of the molar teeth of the earlier Ungulates as compared with those of to-day, of which details will be given in a later chapter. Another instance of the more generalised characters of the earlier mammals is afforded by the absence or slight development of horns, antlers, and tusks among the Ungulata. Thus the earlier Rhinoceroses were hornless, and the Deer either without antlers or with antlers of a very simple kind, while the male Swine were not furnished with the formidable tusks of the existing Wild Boars. Finally, all, or nearly all of the mammals, from the lowest Eocene of Rheims present the peculiarity of having a vertical perforation in the astragalus.

The intimate connection existing during the Middle Tertiary between many families of mammals now widely distinguished from one another may be more conveniently noted when we come to the consideration of the families in question.