The following are the principal characters distinguishing them from the Artiodactyla. Premolar and molar teeth in continuous series, with massive, quadrate, transversely ridged or complex crowns,—the posterior premolars often resembling the true molars in size and structure. Crown of the last lower molar commonly bilobed, and if a third lobe is present in this tooth it is wanting in the last lower milk-molar. Dorso-lumbar vertebræ never fewer than twenty-two, usually twenty-three in the existing species. Nasal bones expanded posteriorly. An alisphenoid canal. Femur with a third trochanter.[253] The middle or third digit on both fore and hind feet larger than any of the others, and symmetrical in itself, the free border of the ungual phalanx being evenly rounded (see Fig. 151). This may be the only functional toe, or the second and fourth may be subequally developed on each side of it. In the Tapirs and many extinct forms, the fifth toe also remains on the fore limb, but its presence does not interfere with the symmetrical arrangement of the remainder of the foot around the median line of the third or middle digit. Traces of a hallux have only been found in some extremely ancient and primitive forms. The astragalus has a pulley-like surface above for articulation with the tibia, but its distal surface is flattened and unites to a much greater extent with the navicular than with the cuboid, which bone is of comparatively less importance than in the Artiodactyla. The calcaneum does not articulate with the lower or distal extremity of the fibula. The stomach is always simple, the cæcum is large and capacious, the placenta diffused, and the mammæ are inguinal. The gall-bladder is invariably absent.

As regards the dentition, the whole of the premolar series may be preceded by milk-teeth; and it has been demonstrated in Rhinoceros that when there is no displacement of the first cheek-tooth that tooth is a persistent milk-molar; the same condition apparently holding good in Palæotherium. This feature indicates considerable dental specialisation, the milk-molars, according to the theory generally accepted by the leading English zoologists, being the acquired, and the premolars the original series. Another peculiar feature of the dentition of the Perissodactyla, very rarely met with among the Artiodactyla, is that the premolars tend to resemble the true molars; this feature occurring in all the existing genera, although not found in the earlier generalised types. The cheek-teeth of all the members of the suborder are primarily constructed on some modification of what is known as the lophodont plan. Thus the upper molars (Fig. 155, p. 375) have an outer antero-posterior wall from which proceed two transverse ridges, formed by the coalescence of the primitive inner and outer columns, towards the inner aspect of the crown; while in the lower molars there may be either two simple transverse ridges, or these ridges may be curved into crescents, coming into contact with one another at their extremities. Those forms having brachydont teeth show this plan of structure in its simplest modification; but in cases, as in the Horse, where the teeth assume an extremely hypsodont form, the original plan is so obscured by infoldings of the enamel that it can only be traced with difficulty.

At the present day the Perissodactyla are sharply differentiated into Horses, Tapirs, and Rhinoceroses, but the knowledge already gained of the extinct representatives of the suborder shows such a close alliance between these groups that it is exceedingly difficult to make any satisfactory classification of the whole. This is of course exactly what might have been expected; and the same would doubtless be the case with all other groups if we knew as much of their past history as we do of that of the Perissodactyles.

The detailed account of the anatomy of the Horse given in the sequel will afford much information as to the general structure of the members of the suborder.

Family Tapiridæ.

Both upper and lower cheek-teeth brachydont and simply bilophodont; hinder premolars as complex as the molars; last lower molar without third lobe; first upper cheek-tooth with a milk-predecessor.[254] Outer columns of upper molars conical. Four digits in the manus, and three in the pes.

Tapirus.[255]—Dentition i ³⁄₃, c ¹⁄₁, p ⁴⁄₃, m ³⁄₃; total 42. Of the upper incisors, the first and second are nearly equal, with short, broad crowns; the third is large and conical, considerably larger than the canine, which is separated from it by an interval. Lower incisors diminishing in size from the first to the third; the canine, which is in contact with the third incisor, large and conical, working against (and behind) the canine-like third upper incisor. In both jaws there is a diastema between the canines and the commencement of the teeth of the cheek-series, which are all in contact. First upper premolar with a triangular crown, narrow in front owing to the absence of the anterior inner cusp. The other upper premolars and molars all formed on the same plan and of nearly the same size, with four roots and quadrate crowns, rather wider transversely than from before backwards, each having four cusps, connected by a pair of transverse ridges, anterior and posterior. The first lower premolar compressed in front; the others composed of a simple pair of transverse crests, with a small anterior and posterior circular ridge.

Skull elevated and compressed. Orbit and temporal fossa widely continuous, there being no true postorbital process from the frontal bone. Anterior narial apertures very large, and extending high on the face between the orbits; nasal bones short, elevated, triangular, and pointed in front. Vertebræ: C 7, D 18, L 5, S 6, C about 12. Limbs short and stout. Forefeet with four toes, having distinct hoofs: the first is absent, the third the longest, the second and fourth nearly equal, the fifth the shortest and scarcely reaching the ground in the ordinary standing position. Hind feet with the typical Perissodactyle arrangement of three toes,—the middle one being the largest, the two others nearly equal. Nose and upper lip elongated into a flexible, mobile snout or short proboscis, near the end of which the nostrils are situated. Eyes rather small. Ears of moderate size, ovate, erect. Tail very short. Skin thick and but scantily covered with hair.

The existing species of Tapir may be grouped into two sections, the distinctive characters of which are only recognisable in the skeleton. (A) With a great anterior prolongation of the ossification of the nasal septum (mesethmoid), extending in the adult far beyond the nasal bones, and supported and embraced at the base by ascending plates from the maxillæ (genus Elasmognathus, Gill). Two species, both from Central America, Tapirus bairdi and T. dowi. The former is found in Mexico, Honduras, Nicaragua, Costa Rica, and Panama; the latter in Guatemala, Nicaragua, and Costa Rica. (B) With ossification of the septum not extending farther forward than the nasal bones (Tapirus proper). Three species, T. indicus, the largest of the genus, from the Malay Peninsula (as far north as Tavoy and Mergui), Sumatra, and Borneo, distinguished by its peculiar coloration, the head, neck, fore and hind limbs, being glossy black, and the intermediate part of the body white; T. americanus (T. terrestris, Linn.), the common Tapir of the forests and lowlands of Brazil and Paraguay (Fig. 152); and T. roulini, the Pinchaque Tapir of the high regions of the Andes. All the American species are of a nearly uniform dark brown or blackish colour when adult; but it is a curious circumstance that when young (and in this the Malay species conforms with the others) they are conspicuously marked with spots and longitudinal stripes of white or fawn colour on a darker ground.

The habits of all the kinds of Tapirs appear to be very similar. They are solitary, nocturnal, shy, and inoffensive, chiefly frequenting the depths of shady forests and the neighbourhood of water, to which they frequently resort for the purpose of bathing, and in which they often take refuge when pursued. They feed on various vegetable substances, as shoots of trees and bushes, buds, and leaves. They are hunted by the natives of the lands in which they live for the sake of their hides and flesh.

The singular fact of the existence of so closely allied animals as the Malayan and the American Tapirs in such distant regions of the earth, and in no intervening places, is accounted for by what is known of the geological history of the race; for the Tapirs must once have had a very wide distribution. There is no proof of their having lived in the Eocene epoch, but in deposits of Miocene and Pliocene date remains undistinguishable generically from the modern Tapirs, and described as T. priscus, T. arvernensis, etc., have been found in France, Germany, and in the Red Crag of Suffolk. Tapirs appear, however, to have become extinct in Europe before the Pleistocene period, since none of their bones or teeth have been found in any of the caverns or alluvial deposits in which those of Elephants, Rhinoceroses, and Hippopotamuses occur in abundance; but in other regions their distribution at this age was far wider than at present, as they are known to have extended eastward to China (T. sinensis, Owen) and westwards over the greater part of the southern United States of America, from South Carolina to California. Lund also distinguished two species or varieties from the caves of Brazil, one of which appears identical with T. americanus. Thus we have no difficulty in tracing the common origin in the Miocene Tapirs of Europe of the now widely separated American and Asiatic species. It is, moreover, interesting to observe how very slight an amount of variation has taken place in forms isolated during such an enormous period of time.

The anatomy of the soft parts of the Tapirs[256] conforms to the general Perissodactyle type, as exemplified in the Rhinoceros and the Horse, although on the whole (as might have been expected) presenting a closer resemblance to the former. T. americanus differs from T. indicus by the absence, or at any rate the less development, of the intestinal valvulæ conniventes, the presence of a moderator band in the heart, the shape of the glans penis, and the more elongated cæcum, which is sacculated by four distinct longitudinal fibrous bands. The convolutions of the hemispheres of the brain of the Tapirs are simpler than in other Perissodactyles, thus tending to confirm the inferences which may be drawn from the skeleton and teeth as to the comparatively low or generalised organisation of these animals.

Palæotapirus.—This name has been applied to an imperfectly known form from the Upper Eocene Phosphorites of Central France, which is regarded by Dr. Filhol as referable to this family.

Family Lophiodontidæ.

Molars brachydont and bilophodont, those of the lower jaw with either straight or imperfectly crescentoid ridges; premolars smaller and usually simpler than the molars; last lower molar generally with a third lobe. Outer columns of upper molars conical or flattened. Digits usually as in the preceding family.

This family includes a number of more or less imperfectly known forms, all of which are extinct and apparently confined to the Eocene period, and ranging from the size of a Rabbit to that of a Rhinoceros. Although some of these appear to have died out without giving rise to more specialised forms, it is probable that this family contained the ancestral types from which most or all of the modern Perissodactyles have been derived. Only very brief mention can be made here of some of the leading genera. Lophiodon, of the Middle and Upper Eocene of Europe, with the dental formula, i ³⁄₃, c ¹⁄₁, p ³⁄₃, m ³⁄₃, includes the largest representatives of the family, and is generally regarded as a stock which has died out without giving rise to later forms. The ridges of the lower molars are straight, and the last of these teeth has a third lobe; while the second transverse ridge of the last upper premolar is usually incomplete; the outer columns of the upper molars are flattened, as in the next genus. Hyrachyus, of the Upper Eocene of the United States, and probably also occurring in the French Eocenes, is an allied genus, with four premolars and no third lobe to the last lower molar; the fourth upper premolar having the two ridges uniting internally to form a crescent. This genus has been regarded as the ancestor of the Rhinocerotic Hyracodon. The genus Hyracotherium was established in 1839 by Owen for a small animal no larger than a Hare, the skull of which was found in the London Clay at Herne Bay. A more nearly perfect specimen, apparently of the same species, was afterwards (in 1857) described under the name of Pliolophus vulpiceps, of which the skull is figured in the accompanying woodcut. Other forms referable to the same genus have been obtained from the Wasatch Eocene of the United States, and were described by Professor Marsh under the name of Eohippus. There were four premolars, the fourth being unlike the molars, and in the upper jaw having only one inner cusp. The upper molars are of the general type of those of Lophiodon, but have conical outer columns, and the anterior transverse ridge imperfect, while the ridges of the lower molars are crescentoid. Systemodon differs from Hyracotherium by the absence of a diastema between the first and second premolars; it occurs in the Wasatch Lower Eocene of the United States. In Pachynolophus (Lophiotherium, Orotherium, or Orohippus), which is common to the Middle and Upper Eocene of Europe and the Bridger Eocene of North America, the outer columns of the upper molars are flattened, and in some cases, at least, the last premolar resembles the molars, that of the upper jaw having two inner cusps.[257] This genus, indeed, so closely connects Hyracotherium with the genera Epihippus and Anchilophus as to show that the distinction between the Lophiodontidæ and Palæotheriidæ is really an arbitrary one. Epihippus, of the Upper Eocene of the United States, has both the third and fourth upper premolars as complex in the molars, and is distinguished from Anchilophus by the lower cusps and more imperfect transverse ridges of these teeth. The so-called Orohippus agilis belongs to this genus. Isectolophus is another American Eocene genus which may be provisionally placed in this family; it is regarded by Professors Scott and Osborn as connecting Systemodon with the Tapiridæ; the fourth and probably the third upper premolar approximating in structure to the molars; the upper molars have conical outer columns. Helaletes is another closely allied form, with similar premolars, but with the outer columns of the upper molars flattened.

Family Palæotheriidæ.

Molars (Fig. 155) brachydont, with the valleys between the ridges never filled with cement; upper premolars either simpler than or as complex as the molars; lower molars with crescentoid ridges, and the last of the series with or without a third lobe. Outer columns of upper molars flattened. Orbit (at least usually) confluent with temporal fossa. Three digits on each foot. This family includes extinct genera ranging from the Middle and Upper Eocene to the Miocene, and passes so gradually into the following one that the maintenance of the two can only be supported on the ground of convenience. The typical genus, Palæotherium, was made known to science in the early part of the present century by Cuvier, who restored the skeleton (Fig. 154) with a short neck like that of the Tapirs, although it has been subsequently found that the neck was considerably longer. This genus (which may be taken to include Paloplotherium) ranges from the Middle to the Upper Eocene of Europe, and usually has the full typical dentition, although the first premolar may disappear. The last lower molar has a third lobe; and in the typical forms the last premolar is as complex as the molars, the diastema is short, and the canines are not large. In other forms, however, the hinder ridge of the fourth upper premolar may be aborted. The first upper cheek-tooth is generally a well-developed tooth, which may have a deciduous predecessor. Anchilophus, of the Upper Eocene of Europe, and Anchitherium, of the Miocene of Europe and North America, connect the preceding forms with the Equidæ. In the latter genus there is the full number of teeth, the last lower molar has almost completely lost the third lobe of Anchilophus, and the surfaces of the two outer lobes of the upper molars (Figs. 157, 158) lack the median vertical ridges of that genus. In the American species of Anchitherium (which have been described as Mesohippus and Miohippus) the lateral digits are larger than in the European Middle Miocene Anchitherium aurelianense; a mere splint represents the fifth metacarpal, and the meso- and ento-cuneiform of the tarsus do not unite as they do in the latter.

Family Equidæ.

Molars hypsodont, with the outer columns of the upper ones flattened, the valleys completely filled with cement, and the enamel thrown into folds and plications; upper premolars as complex as molars, which they slightly exceed in size; ridges of lower molars crescentoid, and complicated by enamel-foldings; no distinct third lobe to last lower molar; summits of incisors with a central infolding of enamel. Orbit completely surrounded by bone. Digits three or one, but in the former case the median one is alone of functional importance; ulna and fibula incomplete; meso- and ento-cuneiform of tarsus united.

Such are the leading characters which serve to distinguish the existing Horses and their nearest fossil allies from the Palæotheriidæ. The Horse, as being the best known of the Perissodactyle Ungulates, is selected for a somewhat detailed description; but before proceeding to this it will be advisable to take a brief survey of the relations of the Equidæ to the extinct forms already noticed, and also of the modifications of the family at present existing.

The earliest form which can be certainly included in this line of descent is the American Lower Eocene genus Phenacodus (noticed below under the head of the suborder Condylarthra), in which there were five complete digits to the feet. From this form there is but a step to Systemodon and Hyracotherium, in which the functional digits of the manus were reduced to four, as in Pachynolophus (Fig. 156, a), although one species retained a rudiment of the metacarpal of the pollex.

The transition from these animals of the Eocene period to the Horses of modern times has been accompanied by a gradual increase in size. The diminutive Hyracotherium of the Lower, and Pachynolophus of the Middle and Upper Eocene were succeeded in the Miocene period by the forms to which the name of Anchitherium has been given, of the size of sheep; these again in Pliocene times by Hipparion and Protohippus, as large as the modern donkeys; and it is mainly in the Pleistocene period that Equidæ occur which approach in size the existing Horse. Important structural modifications have also taken place, with corresponding changes in the mode of life of the animal. Thus the neck has become elongated, the skull altered in form, the teeth greatly modified, and the limbs have undergone remarkable changes. The last two require to be described more in detail.

The teeth in the Eocene forms had, as mentioned above, the characteristic number of forty-four. This number has been retained throughout the series, at least theoretically; but one tooth on either side of each jaw, the anterior premolar, which in all the Eocene and Miocene species was well developed, persisting through the lifetime of the animal, is in all modern Horses rudimentary, functionless, and generally lost at an early period of life, evidently passing through a stage which must soon lead to its complete disappearance. The canines have also greatly diminished in size, and are rarely present in the female sex, so that practically a very large number of adult Horses of the present day have eight teeth less than the number possessed by their predecessors. The diastema or interval between the incisor and premolar teeth (of essential importance in the domesticated Horse to his master, as without it there would be no room for inserting the special instrument of subjugation to his commands, the bit) already existed in the earliest known forms, but has gradually increased in length. The incisors have undergone in comparatively recent times that curious change producing the structure more fully described hereafter, which distinguishes the Horse’s incisors from those of all other known animals, with the exception of the extinct Macrauchenia. Lastly, the molars have undergone a remarkable series of modifications, much resembling in principle those that have taken place in several other groups of herbivorous animals. Distinctions in form which existed between the premolars, at least in the anterior part of the series, and the true molars have gradually disappeared, the teeth becoming all very uniform in the shape and structure of their grinding surface. The crowns of all these teeth in the early forms were very short (see Fig. 158, a); there was a distinct constriction, or neck, between the crown and roots; and when the tooth was developing, as soon as the neck once rose fairly above the alveolar margin, the tooth remained permanently in this position. The term “brachydont” expresses this condition of teeth, the mode of growth of which may be illustrated by those of man. The free surface had two nearly transverse curved ridges, with valleys between (Fig. 157, a); but the valleys were shallow and had no deposit of cement filling them, the whole exposed surface of the unworn tooth being formed of enamel. When the ridges became worn down the dentine of the interior was exposed, forming islands surrounded by enamel. With the progress of time the crowns of the teeth gradually became longer, the valleys deeper, and the ridges not only more elevated but more curved and complex in arrangement. To give support to these high ridges and save them from breaking in use, the valleys or cavities between them became filled up to the top with cement, and as the crown wore down an admirable grinding surface consisting of patches and islands of the two softer substances, dentine and cement, separated by variously reduplicated and contorted lines of intensely hard enamel, resulted (Fig. 157, c). The crown continued lengthening until in the modern Horses it has assumed the form called “hypsodont” (Fig. 158, b). Instead of contracting into a neck, and forming roots, its sides continue parallel for a considerable depth in the socket, and as the surface wears away, the whole tooth slowly pushes up, and maintains the grinding edge constantly at the same level above the alveolus, much as in the perpetually growing Rodent’s teeth. But in existing Horses there is still a limit to the growth of the molar. After a length is attained which in normal conditions supplies sufficient grinding surface for the lifetime of the animal, a neck and roots are formed, and the tooth is reduced to the condition of that of the brachydont ancestor. It is perfectly clear that this lengthening of the crown adds greatly to the power of the teeth as organs of mastication, and enables the animals in which it has taken place to find their sustenance among the comparatively dry and harsh herbage of the open plains, instead of being limited to the more succulent vegetable productions of the marshes and forests in which their predecessors probably dwelt.

The modifications of the limbs which took place pari passu with those of the teeth must have been associated with increased speed, especially over firm and unyielding ground. Short, stout legs, and broad feet, with numerous toes, spreading apart from each other when the weight of the creature is borne on them, are sufficiently well adapted for plodding deliberately over marshy and yielding surfaces, and the Tapirs and the Rhinoceroses, which in the structure of the limbs have altered but little from the primitive Eocene forms, still haunt the borders of streams and lakes and the shady depths of the forests, as was probably the habit of their ancient representatives, while the Horses are all inhabitants of the open plains, for life in which their whole organisation is in the most eminent degree adapted. The length and mobility of the neck, position of the eye and ear, and great development of the organ of smell, give them ample means of becoming aware of the approach of enemies, while the length of their limbs, the angles the different segments form with each other, and especially the combination of firmness, stability, and lightness in the reduction of all the toes to a single one, upon which the whole weight of the body and all the muscular power are concentrated, give them speed and endurance surpassing that of almost any other animal. When surprised, however, they are by no means helpless, both fore and hind feet becoming at need powerful weapons of defence.

If we were not so habituated to the sight of the Horse as hardly ever to consider its structure, we should greatly marvel at being told of a mammal so strangely constructed that it had but a single toe on each extremity, on the end of the nail of which it walked or galloped. Such a conformation is without a parallel in the vertebrate series, and is one of the most remarkable instances of specialisation, or deviation from the usual type, in accordance with particular conditions of life. It is clear, both from the structure of the foot itself, and also by an examination of the intermediate forms, that this toe corresponds to the middle or third digit of the complete typical or pentadactyle foot; and there is very strong evidence to show that by a gradual concentration of all the power of the limb upon this toe, and the concurrent dwindling away and final disappearance of all the others, the present condition of the Horse’s foot has been produced.

Protohippus.[258]—In this Lower Pliocene North American genus (also described as Merychippus) the cheek-teeth resemble those of the generalised species of Equus, but have shorter crowns; while the milk-molars approximate to the permanent molars of Anchitherium. Each foot has three digits.

Hipparion.[259]—Upper cheek-teeth (Fig. 159), with the antero-internal column, or anterior pillar as it may be conveniently termed in this family, detached throughout the greater part of its height from the adjacent column. Either a single or three digits in each foot. First upper premolar large and persistent. This genus was very widely distributed in the Pliocene, occurring in Europe, Asia, and North America. In the typical European forms, and also in those of North America, there were three digits in the feet (Fig. 156, d); but in the Indian H. antilopinum (separated by Cope as Hippodactylus) the lateral digits seem to have disappeared. There is some doubt whether or no Hipparion should occupy a place in the direct ancestry of the Horse, and Professor Cope suggests that while in America the intermediate place between Anchitherium and Equus was held by Protohippus, in Europe the same position was occupied by Hipparion—a view which involves the dual origin of the Horses of the New and Old Worlds.

Equus.[260]—Upper cheek-teeth with the anterior pillar (except in a very early stage of wear) joined by a narrow neck to the adjacent column (Fig. 157, c). Each foot with a single complete digit, but with remnants of the proximal portions of the second and fourth metapodials (Fig. 156, e); some extinct forms having claw-like rudiments of the terminal phalangeals of the lateral digits. First upper premolar very small or altogether absent in existing species, but in some fossil species larger and persistent; first lower premolar only occasionally developed in some fossil forms. Ears long. Tail long, with long hairs either at the end or throughout. A callosity on the inner side of the fore limb above the carpus.

Fossil Species.—In the Pleistocene Horses of South America described as Hippidium, as well as in the closely allied ones from North America for which the name Pliohippus has been proposed, the upper molars are shorter and more curved than in the existing species, while their anterior pillar is not longer antero-posteriorly than in Hipparion; the lateral claw-like hoofs persisting. Some of the European Pliocene species (like E. stenonis) agree with these species in the form of the grinding surface of the anterior pillar of the upper molars. In one of the species from the Lower Pliocene of India (E. sivalensis)—which was a contemporary of Hipparion—and in all the existing species, the grinding surface of the pillar in question is greatly elongated in the antero-posterior direction, as in Fig. 157, c.

Fossil remains of Horses are found abundantly in deposits of the most recent geological age in almost every part in America, from Eschscholtz Bay in the north to Patagonia in the south. In that continent, however, they became quite extinct, and no Horses, either wild or domesticated, existed there at the time of the Spanish conquest, which is the more remarkable as, when introduced from Europe, the Horses that ran wild proved by their rapid multiplication in the plains of South America and Texas that the climate, food, and other circumstances were highly favourable for their existence. The former great abundance of Equidæ in America, their complete extinction, and their perfect acclimatisation when reintroduced by man, form curious but as yet unsolved problems in geographical distribution.

Existing Species.—The existing species of the genus are the following:—

The Horse, Equus caballus, is distinguished from the others by the long hairs of the tail being more abundant and growing quite from the base as well as the end and sides, and also by possessing a small bare callosity on the inner side of the hind leg, just below the “hock” or heel joint, in addition to the one on the inner side of the fore limb above the carpus, common to all the genus. The mane is also longer and more flowing, and the ears are shorter, the limbs longer, the hoofs broader, and the head smaller.

Though the existing Horses are not usually marked in any definite manner, or only irregularly dappled, or spotted with light surrounded by a darker ring, many examples are met with showing a dark median dorsal streak like that found in all the other members of the genus, and even with dark stripes on the shoulders and legs indicating “the probability of the descent of all the existing races from a single dun-coloured, more or less striped, primitive stock, to which our horses still occasionally revert.”[261]

In Europe wild Horses were extremely abundant in the Neolithic or polished-stone period. Judging from the quantity of their remains found associated with those of the men of that time, the chase of these animals must have been among man’s chief occupations, and they must have furnished him with one of his most important food supplies. The characters of the bones preserved, and certain rude but graphic representations carved on bones or reindeers’ antlers, enable us to know that these Horses were rather small in size, and heavy in build, with large heads and rough shaggy manes and tails, much like, in fact, the present wild horses of the steppes of the south of Russia. They were domesticated by the inhabitants of Europe before the dawn of history, but it is doubtful whether the majority of the animals now existing on the Continent are derived directly from them, as it is more probable that they are descendants from Horses imported through Greece and Italy from Asia, derived from a still earlier domestication, followed by gradual improvement through long-continued attention bestowed on their breeding and training. Horses are now diffused by the agency of man throughout almost the whole of the inhabited parts of the globe, and the great modifications they have undergone in consequence of domestication and selective breeding are well exemplified by comparing such extreme forms as the Shetland pony, dwarfed by uncongenial climate, the thoroughbred racer, and the London dray-horse. In Australia, as in America, horses imported by the European settlers have escaped into the unreclaimed lands, and multiplied to a prodigious extent, roaming in vast herds over the plains where no hoofed animal ever trod before.

A wild Horse from Central Asia, named E. prezevalskii,[262] is described as having callosities on both limbs and broad hoofs like E. caballus; but the long hairs of the tail do not begin until about half way down its length. It also differs from E. caballus in having a short erect mane and no forelock; neither is there any dorsal stripe. The ears are of moderate size; the whole body is of a whitish-gray, paler beneath, and reddish on the head and upper parts of the limbs. If rightly described this form would appear to be intermediate between the true Horses and the Asses.

The second species is the domestic Ass (E. asinus), and the wild Asses of Africa (E. asinus, var. africanus and var. somalicus[263]). The domestic Ass, which is now nearly as widely diffused and useful to man as the Horse, was known in Egypt long before the latter, and is doubtless of African origin. The ears are long, the mane erect, the tail without long hairs at the base, and there are no callosities on the hind limbs. There is a dark dorsal stripe, and another across the shoulders; while the limbs are frequently banded. Of the wild forms the Nubian race (var. africanus) has distinct dorsal and shoulder stripes, but the rings on the limbs are often very indistinct; while in the Somali race the dorsal stripe is indistinct, and the shoulder stripe wanting, but the rings on the limbs are very boldly marked. Teeth and bones from a Pleistocene cavern deposit in Madras have been referred to E. asinus.

The Asiatic wild Asses, which roam in small herds in the open plains of Syria, of many parts of Persia, of the north-west of India, and the highlands of Tartary and Tibet, from the shores of the Caspian to the frontiers of China, differ from the last in being of a more rufous or isabelline colour, instead of pure gray, in wanting the dark streak across the shoulder, and having smaller ears. They have all a dark-coloured median dorsal stripe. Though it is considered probable by many zoologists that they form but a single species[264] (E. hemionus), they present such marked variations in size and form that they have commonly been divided into three—the Syrian Wild Ass (E. hemippus), the Onager (E. onager) from Persia, Baluchistan, the Punjab, Sind, and the desert of Kach, and the Kiang or Dzeggetai (E. hemionus) of the high table-lands of Tibet, where it is usually met with at an elevation of 15,000 feet and upwards above the sea-level. The last is considerably larger than either of the others, and differs from them in external appearance, having more the aspect of the horse. They are all remarkably swift, having been known to outstrip the fleetest Horse in speed.

Lastly, there are four striped species, all inhabitants of Africa. These constitute the genus Hippotigris of Hamilton-Smith, but they are not separable except by their coloration from the true Asses, and one of them, the Quagga (E. quagga), may be considered as intermediate. This animal was formerly met with in vast herds on the great plains of South Africa, between the Cape Colony and the Vaal River, but now, in common with most of the larger wild animals of that region, is becoming extremely scarce, owing to the encroachments of European civilisation, if, indeed, it is not already extinct. In length of ears and character of tail it more resembles the Horse than it does the Ass, although it agrees with the latter in wanting the callosity on the inner side of the hind leg, just below the hock, characteristic of the Horse. The colour of the head, neck, and upper parts of the body is reddish-brown, irregularly banded and marked with dark brown stripes, stronger on the head and neck and gradually becoming fainter until lost behind the shoulder. There is a broad dark median dorsal stripe. The under surface of the body, the legs, and tail are nearly white, without stripes. The crest is very high, surmounted by a standing mane, banded alternately brown and white. Though never really domesticated, Quaggas have occasionally been trained to harness. The accompanying figure is reduced from a painting made from one of a pair which were driven in Hyde Park in the early part of the present century. The name is an imitation of the shrill barking neigh of the animal—“ouag-ga, ouag-ga,” the last syllable very much prolonged. It must be remembered, however, in reading books of African travel that the same word is very commonly applied by hunters to Burchell’s Zebra.

Of the Zebras proper, the one which was first known to Europeans, and was formerly considered the most common, is the True Zebra (E. zebra), sometimes called the Mountain Zebra. It inhabits the mountainous regions of the Cape Colony; but now, owing to the advances of civilised man into its somewhat restricted range, it has become very scarce, and is even, like the Quagga, threatened with extermination at no distant date. The second species, Burchell’s Zebra (E. burchelli), still roams in large herds over the plains to the north of the Orange River, but in yearly diminishing numbers. Both species are subject to considerable individual variations in marking, but the following are the principal characters by which they can be distinguished.