locomotion. But if the digitigrade posture be extreme, the fate of the digits will depend upon the habitat of the animal. Should the region ordinarily traversed offer smooth, firm footing, an animal running upon the tips of its toes would use chiefly the longer third digit; this digit would naturally be strengthened and increased in size, while the other digits, being useless, might gradually disappear. Such changes have taken place in the foot of the horse, and in this way the perissodactyl, or odd-toed, type of foot has arisen. As is well known, the habitat of the horse family is the dry rolling plain, and what evidence we have goes to show -that the remote ancestors of the horse ranged a similar country. The fact that the digits of the horse were first reduced at their distal extremities points to the same conclusion. For, as we shall see, digits which do not support the weight of the body. may. if the animal frequents swampy regions, keep the foot from sinking too deeply. In such cases the reduction of the digits begins always at the upper or proximal end, just the reverse of conditions in the foot of the horse. We may conclude then that the odd-toed foot resulted from digitigrade locomotion over firm, comparatively level ground.
For rapid progression over swampy ground, the structure of the foot must conform to two requirements. It must be prevented from sinking too deeply and must be easily withdrawn. Any one who has attempted to walk across a mud flat can appreciate the importance of these two factors. It is in adaptation to these requirements that the artiodactyl, or even-toed, type of foot has evidently been developed. If an ungulate was of semi-aquatic habits, or attached to swampy places, all four digits, by spreading, would prevent the sinking of the foot; its withdrawal would be facilitated by making the foot occupy as little space as possible, and this could be accomplished by shifting the proximal articulations of the outer digits inward and posterior to the middle digits. The toes would then be arranged in pairs, the outer pair lying somewhat behind the other. Now in a semi-aquatic animal each pair of digits would be subjected to the same usage in walking or running through boggy ground; the digits of each pair would tend toward the same structure on this account. But as the middle digits would support the greater part of the strain brought to bear upon the foot, this pair of digits would naturally become larger and stronger than the other. Should our hypothetical ungulate change its habitat from the swamp to the firmer footing of the plain or upland, the outer pair of digits would not reach the ground, and unless they proved of use to the animal in some other way we should expect them to speedily disappear.
This theory as to the origin of the artiodactyl foot is supported when we examine into the habits of the even-toed ungulates. The more primitive forms are attached to the water. The amphibious hippo-
