The artiodactyls are and for a very long time have been a very much larger and more variegated group than the perissodactyls, and the Old World has been and still is their headquarters and area of special development, where they are represented in far greater number and variety than in the New; the perissodactyls, on the other hand, flourished especially in North America, as was shown in the preceding chapter. At the present time the artiodactyls are the dominant ungulate order, far outnumbering all the others combined, and include an assemblage of varied types, which, when superficially examined, appear to be an arbitrary and unnatural group. What could seem more unlike than a dainty little mouse-deer, no larger than a hare, a stag, a camel, a giraffe, a bison and a hippopotamus? Yet, in spite of this wonderful diversity of size, proportions, appearance and habits, there is a genuine unity of structure throughout the order, which makes their association in a single group altogether natural and proper, especially as these structural characters are not found united in any other group.

It would be superfluous to enumerate all of the diagnostic characters which, on the one hand, unite all the living and extinct artiodactyls and, on the other, distinguish them from all other hoofed animals, and it will suffice to mention a few of the more significant of these features.

As the name implies, the artiodactyls typically have an even number of toes in each foot, four or two; though this rule may be departed from and we find members of the order with five digits or three, just as the tapirs and nearly all the Eocene genera of perissodactyls had four toes in the manus. Much more important is the fact that the plane of symmetry, which in the perissodactyls bisects the third digit and is therefore said to be mesaxonic, passes between the third and fourth digit and is paraxonic. The third and fourth digits always form an equal and symmetrical pair and are the “irreducible minimum,” beyond which the number of toes cannot be diminished. A single-toed artiodactyl would seem to be an anatomical impossibility; at all events, such a monstrosity was never known. Hence the term “cloven” or “divided” hoof, which seems to take the solid hoof of the horse as the norm; but “cloven or divided,” while expressing the appearance of the foot with sufficient accuracy, is erroneous, if taken to mean the splitting of what was once continuous.

Especially characteristic of the order is the structure of the ankle, or “hock-joint” of the hind limb. The ankle-bone, or astragalus, has a double pulley, the upper and lower ends being of quite similar shape; its lower end is almost equally divided between the cuboid and navicular, which are made concave to receive it. This type of astragalus is altogether peculiar to the artiodactyls, all of which possess it; it is unlike that of any other mammal whatever and may be recognized at a glance. The calcaneum, or heel-bone, has a large convex facet, by means of which it articulates with the fibula, or external leg-bone; there is no such articulation in the perissodactyls. The lower end of the calcaneum is narrow and fits into a step cut in the cuboid, which is thus every whit as peculiar and characteristic as the calcaneum and astragalus. The femur never has the third trochanter, which is always present in the perissodactyls. Another respect in which the artiodactyls differ from all perissodactyls except the horses is in the much more complex mode of articulation between the vertebræ of the lumbar and posterior dorsal regions, which the former display, and even the horses have no such elaborate arrangement. Finally, another very marked difference from the perissodactyls is in the teeth, for the premolars and molars are never alike, and only in very rare instances does the last premolar assume the molar-pattern. Of this pattern, there are two principal kinds, one exemplified by the peccaries, in which the crown supports a series, fundamentally two pairs, of conical cusps, and called bunodont, and the other, to be seen in all the ruminating animals, in which the crown is composed of two pairs of crescents and is therefore said to be selenodont. The bunodont was the primitive type, whence the other was derived, and many transitional forms are known.

The classification of the immense horde of living and extinct genera and species which are referable to the artiodactyls is an extremely difficult problem, which has found no thoroughly satisfactory solution and will not until much more is learned concerning the history of the order and conflicting opinions can be reconciled. The most important American families and genera are given below, though the arrangement is but tentative.

This list of families and genera, portentous as it is, would be greatly increased by the addition of the Old World forms, which outnumber those of the western hemisphere.

Suborder Suina. Swine-like Animals

The history of the American types of pig-like forms is, in one sense, very full and complete in that the successive genera may be traced back to the Eocene, but, in another sense, the story is exasperatingly imperfect, because so much of the material is fragmentary. Of most of the genera, nothing is known but teeth and jaws, and these, though sufficient for identification, tell but little of the structural changes which it is desirable to know. It is merely a question of time, when more adequate material will be obtained.

1. Tagassuidæ. Peccaries

The peccaries, or American swine, are now chiefly of Neotropical distribution, extending into the Sonoran region only as far as Arkansas; but this has been true only since the Pleistocene, for nearly the entire history of the family has been enacted in North America. In many points of structure the peccaries of the present day are more advanced and specialized than the far more varied and diversified true swine of the Old World, for it is a singular fact that such a long-lived and persistent stock as the peccaries should have given rise to so few variants and side-branches. Existing peccaries all belong to a single genus (Tagassu) and are relatively small animals, of unmistakably pig-like character and appearance, but far smaller than the Wild Boar (Sus scrofa) of Europe, or the Wart Hog (Phacochœrus æthiopicus) of Africa, to mention only two of the Old World swine.

One characteristic and thoroughgoing difference between the peccaries and the swine is the shape of the canine tusks. In the former, the tusks, though very effective weapons, are not very large and are straight and have a vertical direction, while in all the true swine the upper tusk is curved upward and outward, projecting strongly from the side of the jaw, and the great, curved lower tusk wears against its anterior side. The peccaries further have smaller and simpler molars, each with four principal, conical cusps (quadrituberculate pattern) arranged in two transverse pairs, with numerous very small cuspules around and between them, obscuring the plan. In the true swine the teeth are much larger and covered with innumerable wart-like cusps, large and small, seldom arranged according to any definite plan.

In the following particulars the modern peccaries show advance over the Old World swine: (1) the last lower premolar has taken on the molar-pattern, a very exceptional feature among the artiodactyls; (2) the ulna and radius are coössified; (3) there are but two functional digits in each foot; the fore foot has, in addition, two complete, but very reduced and slender, lateral digits and the hind foot only one, whereas in all the pigs of the eastern hemisphere there are four functional toes in each foot; (4) in the hind foot the two functional metatarsals, the third and fourth, have coalesced to form a “cannon-bone,” a structure which is not found in any other family of the suborder; (5) the stomach is complex, approximating that of a ruminant.

In the North American Pleistocene the predominating kind of peccary was a genus (†Platygonus) which was more advanced than the existing form (Tagassu), and, to all seeming, better fitted to survive, though for some inexplicable reason it failed to do so. It was a considerably larger animal, with proportionately longer and heavier legs. Its molar teeth are of special interest because they reproduced a type which has been so often repeated and independently acquired in so many different groups of mammals. In this molar the two conical cusps of each pair were fused into a high, transverse ridge or crest. Precisely the same modification took place among the true swine in the genus †Listriodon of the French middle Miocene. †Platygonus first appeared in the middle Pliocene, and its predecessor in the lower Pliocene and upper Miocene showed the crests of the molars in process of formation. In the latter stage it was accompanied by a true peccary with tuberculated teeth, which differed from the modern species in the simplicity of the hindmost premolar, which had not taken on the molar-pattern. If the feet and limbs of this upper Miocene peccary were known, they would doubtless prove to be much more primitive than those of Tagassu, but they still await discovery.

Little can be said of the peccaries of the middle and lower Miocene other than to record the fact of their presence in those formations, but those of the upper Oligocene (John Day) are, however, represented by well-preserved skulls, which show that more than one phylum of the family had arisen, though there was no great difference between them; they were considerably smaller animals than those of the Pliocene and Pleistocene. Still smaller was the White River genus (†Perchœrus) of which some fragmentary skeletons have been obtained. Although an undoubted peccary, this animal was not far from what the common progenitor of the peccaries and the true swine might be expected to resemble. The molars were quadrituberculate without the numerous accessory cuspules of the modern genus; the bones of the fore-arm were separate and the feet had four functional digits each, while there was no cannon-bone in the pes, the metatarsals remaining free.

No peccaries have yet been found in the Uinta, but probably this is a mere accident of collecting. It is, however, possible that the White River genus was not of American derivation, but an immigrant from the Old World. In the middle Eocene, or Bridger stage, this series is known only from teeth and jaws and a very few scattered foot-bones, and these, though probably referable to the family, cannot be definitively assigned to it without more complete material. Several species, larger and smaller, of the genus †Helohyus occurred in the Bridger, where they were not uncommon, considering the general rarity of artiodactyls in that stage. Thus, the peccaries, though none of them were large, followed the usual law of mammalian development, and, beginning with very small forms, increased in size with each succeeding geological stage down to the Pleistocene.

2. †Entelodontidæ. †Giant Pigs

The †giant pigs, a most remarkable group of swine-like forms and of as yet unknown origin, appeared for the last time in North America in the lower Miocene, where the genus of that date (†Dinohyus) was the largest of known suilline animals, the hippopotamuses excepted. In nearly every part of the skeleton these great beasts displayed an unusual and aberrant kind of development. The incisors were long and pointed, and the canines formed stout and heavy, though not very long, tusks, which in shape were more like those of a bear than those of either peccaries or swine. The premolars were very simple, of compressed conical and trenchant shape, and occupied a very long space in the jaws, while the molars were relatively small and quadrituberculate, the crowns covered with very thick, coarsely wrinkled enamel. The skull was immensely elongate, especially the facial region in front of the eyes, while the brain-case was so absurdly small as to give the skull a reptilian aspect, when viewed from above. Evidently, these great pigs were profoundly stupid, in this respect rivalling the †titanotheres of the White River (p. 311). Beneath each eye-socket was a long, descending, bony flap, or process, and on the under side of the lower jaw were two pairs of prominent knobs, the function of which, as of the flaps beneath the eyes, is quite problematical. The eye-sockets themselves were completely encircled in bone, a rare character in the suborder.

The neck was short, as in the pigs generally, the body not very elongate and the tail of moderate length; at the shoulders, the spines of the dorsal vertebræ were very long, making a decided hump, and in the lumbar and posterior dorsal region the processes for articulation between the vertebræ were extremely elaborate. For one of the pigs, the limbs were very long and gave quite a stilted look to the animal. As in the modern peccaries, the fore-arm bones were indistinguishably fused together and the feet had only two toes each, the only members of the suborder in which digital reduction had proceeded so far, though the existing peccaries approximate this condition. There were, however, nodular vestiges of two other digits, which prove the derivation of this form from at least a four-toed type; no cannon-bone was formed. In view of the size of the animal, the hoofs were surprisingly small, which suggests that the weight was chiefly borne upon a pad. †Dinohyus was a very large animal, six feet or more in height at the shoulder.

In the upper Oligocene were very large species of another, but closely similar, genus (†Boöchœrus) though somewhat smaller than those of †Dinohyus, and the species of the upper White River beds (†Archæotherium) were little, if at all, smaller than those of the John Day. A number of specimens in the museum of Princeton University throw a welcome light upon the habits of these strange creatures. In one, the external, or third, upper incisor tooth has a deep, triangular notch worn in its postero-external face, and the lower canine has a well-defined groove worn on the posterior side at the base of the crown; other individuals show less distinct marks of similar kind. (See Fig. 194.) It is out of the question to suppose that these grooves and notches could have been produced by abrasion with other teeth, for no other teeth could reach the worn areas, and it is altogether probable that they were made in digging up roots. The root, held firmly in the ground at both ends and looped over the teeth which pulled until it broke, and being covered with abrasive grit, would wear just such marks as the teeth actually display.[7] While the †entelodonts were thus rooters, they were doubtless omnivorous, like other pigs, and did not disdain a meal of carrion when they could get it. It is likely that the heavy canine tusks were also used as weapons, both in defence against the attacks of carnivores and in fighting between the males of the same species. It must have been in some such encounter that the animal represented by a complete skeleton in the Princeton Museum received its broken rib; that the fracture was made during life is demonstrated by the large callus growths on the broken ends, but the pieces did not knit.

In the middle and lower substages of the White River the genus (†Archæotherium) was the same as in the upper substage of these beds, but the species were all smaller and some of them very much so, not exceeding an ordinary pig in size. Throughout the series, as we now have it, from the lower Oligocene into the lower Miocene, there is very little change except in size, all the essential features of structure remaining the same; the genera are therefore distinguished by modifications of very secondary importance, and it is a question whether all the species should not be included in a single genus. The European genus †Entelodon, which gives its name to the family, is so like the American forms that by most writers the White River species are referred to it. It is of interest to note that the †giant pigs have also been found in the marine Miocene of New Jersey, one of the few records of the Tertiary land mammals of the Atlantic seaboard.

At present, the †entelodonts proper cannot be traced back of the lower White River beds, nor are they found in any more ancient formations in Europe. It is, therefore, probable that they were immigrants in both of these continents, presumably from Asia.

The whole Eocene of North America had a series of pig-like animals, called the †achænodonts or †short-faced pigs, which seem to have been related to the †entelodonts. They ended their career in the Uinta just before the appearance of the †entelodonts, and it would be natural to suppose that the latter were descended from them. If, however, the principle that an organ or structure once lost can never be regained, is valid, then there can be no relation of ancestor and descendant between the two groups, for of the †achænodonts, even their most ancient representatives had lost the first premolar, giving the formula p 3/3, while in the †entelodonts it is constantly p 4/4. The †achænodonts, which are much less fully known than the †entelodonts, had teeth very similar in form to those of the latter; and their most conspicuous feature was the shortness of the face and jaws, as contrasted with the extreme elongation of these parts in the †entelodonts, nor did they have the bony flaps under the eyes or the knobs on the lower jaw which gave such a fantastic appearance to the †entelodont skull. Little is known of the skeleton except that there were four functional digits in the manus. The Uinta and Bridger genus (†Achænodon) was larger than the Wasatch form (†Parahyus), which was an immigrant, probably from the same region as afterwards sent out the †entelodonts to America and Europe; this would account for the similarity and probable relationship of the two subfamilies.

Suborder Artiodactyla †Primitiva. †Primitive Artiodactyls

No doubt, this suborder is an artificial assemblage of unrelated families, a sort of waste-basket, into which are thrown the groups of which no other disposition can be made in the present state of knowledge. As information becomes more complete, the various families will be redistributed among the groups with which they had a genuine relationship.

3. †Anthracotheriidæ. †Anthracotheres

This family was abundantly represented in Europe from the middle Eocene through the Oligocene, in Asia persisting even into the Pliocene, and were abundant in the Oligocene of Egypt. Migrants from the Old World reached America in White River times, but speedily died out, as they did not survive into the upper Oligocene. The most fully known of these animals is an American species of a European genus †Bothriodon. Almost complete skeletons of this genus have been obtained in the channel sandstones of the upper White River substage. In size and proportions, †Bothriodon was not unlike a domestic pig, but had a very long head with slender, pointed snout; it had also a short neck, long body, short limbs and feet. The primitive character of this genus is made clear by many features of its structure; the molar teeth were extremely low-crowned and their cusps were so imperfectly crescentic in form as to be called buno-selenodont, as indicating their transitional nature, and the upper molars had five cusps instead of four, a very primitive feature. Another very significant character was the five-toed manus; the first digit, or pollex, was much smaller than the others.

The second genus of the family which had American representatives was †Anthracotherium, which was much like †Bothriodon, but even more archaic in character; the molars could hardly be called selenodont at all.

4. †Oreodontidæ. †Oreodonts

This was one of the most characteristic of North American artiodactyl families, and its members were exceedingly abundant throughout the upper Eocene, the whole Oligocene and Miocene, ending their long career in the lower Pliocene. In distribution the family was exclusively North American, and no trace of it has been found in any other continent. In the course of their long history the †oreodonts underwent many transformations and branched out into several distinct phyla, yet through all these changes they remained singularly conservative, for the transformations, some of them sufficiently bizarre, affected chiefly the teeth and skull, the remainder of the skeleton changing but little. The †oreodonts were all small or of moderate size, none of them surpassing the Wild Boar in stature, nor was there any decided increase in size from stage to stage. One and all, they were strange beasts. Dr. Leidy, who first described and named most of the genera, spoke of them as combining the characters of camel, deer and pig, and called them “ruminating hogs,” a conception expressed in the names which he gave to some of them, such as †Merychyus and †Merycochœrus, both of which mean ruminant swine.

The general proportions of most of the species were quite as in the peccaries, though, for the most part, with much longer tails; they had a short neck, elongate body, short limbs and feet. In one genus (†Mesoreodon) of the lower Miocene a rudimentary collar-bone has been found, and probably all of the more ancient genera possessed it, but only by an unusually lucky chance would so small and loosely attached a bone be preserved in place. As the collar-bone is superfluous in hoofed animals, in which the limbs are used only for locomotion and move in planes parallel with that of the backbone, it is almost universally absent in them, and in only one other group of ungulates, the extinct †Typotheria of South America, has its presence been demonstrated. In all of the †oreodonts the bones of the fore-arm and lower leg remained separate. The teeth were in continuous series, and there was a peculiar feature in the dentition common to nearly every one of the genera. On casual examination, one would say that the animals had four lower incisors on each side and that the lower canine closed behind the upper one, a most exceptional arrangement. More careful study shows that the apparent fourth incisor was the canine, a transformation which has also taken place in all of the ruminants except the camels, and the tooth which had assumed the form and function of the lower canine was really the first lower premolar; this latter change is not found among the ruminants, but was repeated in a few other extinct families.

Only two genera of †oreodonts (†Merychyus and †Merycochœrus) survived into the lower Pliocene. Both had the proportions common throughout the family, but †Merychyus was much more slender and lightly built, its lateral digits were reduced in size and very thin and it had hypsodont grinding teeth; while †Merycochœrus was of larger size (about that of a large domestic pig) and stouter build and had low-crowned teeth; its head, however, had a very different appearance, given by the possession of a short proboscis, the presence of which is indicated by the greatly reduced nasal bones; the jaws and face were also much shortened. The eye-sockets presented obliquely forward and upward, instead of laterally, as is usual among mammals, and were placed high in the head. This position of the eyes and of the entrance to the ear renders it probable that †Merycochœrus was largely aquatic in its habits. Both genera had short, four-toed feet, as was general throughout the family and in no genus did the reduction of digits proceed beyond the loss of the first of the original five, the pollex and hallux.

The two genera above described, representatives of two distinct phyla within the family, held over, as it were, from the upper Miocene without essential change. The phylum of the hypsodont and slender †Merychyus went back, with only minor modifications, into the upper substage of the lower Miocene, but cannot as yet be traced to an Oligocene ancestry; it is therefore still impossible to say just where and when it branched off from the main stem of the family. Future discoveries in the Oligocene will no doubt clear up this problem. The real terminal and most highly specialized member of the †Merycochœrus phylum and the most extraordinary member of the entire family was confined to the upper Miocene. The extreme peculiarity of this genus (†Pronomotherium) was displayed only in the head, which was an exaggeration of the †Merycochœrus type, the face being excessively shortened and the nasals so reduced as to show that the proboscis was much better developed than in the parent genus. The shortening of the face and the great vertical height of the skull and lower jaw gave a decided likeness to the skull of a great ape, though the proboscis would mask any such resemblance in the living head. †Merycochœrus itself went back to the upper division of the lower Miocene, but in the lower division it was replaced by an ancestral genus, †Promerycochœrus, which had an elongate face and jaws and no proboscis; but in other characteristic features, such as the extreme thickness and roughness of the zygomatic arches, it was like its descendant. †Promerycochœrus contained the largest known species of †oreodonts, some of them equalling a Wild Boar in stature, and its remains are found so abundantly in the middle and lower Miocene and upper Oligocene, that there must have been great herds of these animals over the plains. Probably it was itself derived from some of the larger species of †Eporeodon of the upper White River beds, but there is a gap in the history, due to the fact that the lower part of the John Day is almost barren of fossils and the connecting link has not been recovered.

It is an interesting and significant fact that ancestral and derivative genera may continue to live side by side in the same region. †Promerycochœrus, it is believed, gave rise to †Merycochœrus, but survived with it into the middle Miocene. †Merycochœrus, in its turn, produced †Pronomotherium, and, so far from being replaced by the latter, actually outlived it and persisted into the lower Pliocene.

A third phylum of the †oreodonts, which appeared for the last time in the middle Miocene (genus †Cyclopidius), was a series of small and very small species, of which the skull was almost as peculiar as that of †Pronomotherium, but in a different fashion. The face was very much shortened and on each side a great vacuity reduced the nasal bones to mere splints; the elevated position of the eye-sockets, which projected above the forehead, and of the tubular entrance to the ear is an evidence of an aquatic or amphibious mode of life, such as is illustrated by the hippopotamuses, which can float almost completely submerged, with only the ears, eyes and nostrils above the surface of the water. The tympanic bullæ (see p. 66) or bony chambers into which the ear-tubes opened, were of relatively enormous size and added much to the unusual appearance of the skull. The incisors were very small and the grinding teeth narrow and completely hypsodont, this and the †Merychyus series being the only two phyla of the family in which the hypsodont molar was fully acquired. The remainder of the skeleton differed but little from the type common to the whole family, except for a somewhat shorter tail.

The animals of this series were common in the middle and lower Miocene and in the upper substage of the White River, but have not been found in the intermediate John Day. This may have been a matter of geographical distribution, these creatures not extending west of the main ranges of the Rocky Mountains. In the upper White River the genus †Leptauchenia is extremely common, but below that level they suddenly and completely vanish and, as in the case of the †Merychyus phylum, it is not yet practicable to determine the point in time or space of their branching off from the main stem of the family. Were the †oreodonts not entirely confined to North America, we should, as a matter of course, explain the seemingly sudden appearance of †Leptauchenia as due to immigration, and it is entirely possible that the series did actually originate in some part of North America which has left no record of its Eocene or Oligocene terrestrial life. On the other hand, no one can imagine that everything that can be known of the mammals of the middle and lower White River has already been learned, and at any time the sought-for ancestor of †Leptauchenia may be found in those beds.

The fourth phylum may be regarded as the main or central stem of the family and was the one which underwent the least change, though it probably gave rise to all the other phyla, which branched off from it at various stages in its history. This series terminated in the middle Miocene and comprised several genera, all very much alike, in the lower stages of that epoch. One of these genera (†Mesoreodon) displayed a very remarkable peculiarity of structure in the ossification of the great cartilage of the larynx, which seems to point to the possession of uncommon vocal powers. It is impossible to say whether this feature was confined to the single genus, or was general in the family, for only in rare instances would so extremely delicate a structure be preserved. In the John Day the genus †Eporeodon, which was very abundant, was the representative of this phylum, and the same, or a closely similar, genus lived in the latter part of the White River stage.

In the middle and lower White River substages †oreodonts are the commonest of fossils, so that the collector soon wearies of them (see Fig. 136, p. 259); they must have lived in great herds in the forests and along the streams. There were several species, varying principally in size, the largest about as long as a wolf, but with shorter legs, and the smallest not so much as half of that size. All belonged to a single genus, for which the rigid law of priority compels us to use a most cumbrous name (†Merycoidodon), the widely used term †Oreodon being a synonym. This genus was the central stock of the family, from which most, if not all, the others were directly or indirectly derived, though, as previously pointed out, we cannot in all cases trace the connection. In these White River animals the grinding teeth were very low-crowned and had considerable resemblance to those of a deer; the molars were typically selenodont and made up of two pairs of crescentic cusps. The skull differed little from that of the succeeding genera of this phylum; the neck was short, body and tail long. An especially interesting fact is that the fore foot had five digits, the first, or pollex, very small and of no functional value, but complete in all its parts; the hind foot was four-toed. In all of the subsequent genera of the family the number of digits was uniformly four in both manus and pes.