[Illustration: FIG. 23.—Right hind foot of _Ophiacodon_: _a_, astragalus; _c_, calcaneum; _c1_, _c2_, centralia; 1, 2, 3, 4, 5, tarsalia.] The tarsus of reptiles differs from that of mammals, in that the chief movements of extension and flexion of the foot upon the leg occur within the tarsus rather than between the tarsus and leg bones. Primitively the tarsus of reptiles consisted of nine bones, two in the first row, two in the second, and five in the third, but in all modern reptiles the bones of the middle row and the fifth one in the third row have disappeared; in some lizards and turtles the two of the first row are fused. The two bones of the proximal row correspond quite to the astragalus and calcaneum, the astragalus articulating with both tibia and fibula proximally, the calcaneum with the latter only. The oldest known tarsus of any vertebrated animal, one from the Coal Measures of Ohio, has this structure, while in all the early amphibians there were three bones, the tibiale, intermedium, and fibulare. Some of the later swimming reptiles, like the ichthyosaurs and plesiosaurs, have apparently this amphibian structure, with three bones that are usually called tibiale, intermedium, and fibulare, but it is very doubtful indeed whether they are homologically the same. In the middle row two centralia are known in one or two very ancient reptiles, but for the most part there is only a single centrale, and even that is usually lost in later reptiles. The third row, like the third row of the carpus, had a distinct bone for each digit originally, but the fifth one was very soon lost and has never reappeared. The structure of the digits and number of bones are quite like those of the hands, except that the fifth toe has four bones instead of three, that is, the phalangeal formula was 2, 3, 4, 5, 4. As a rule in terrestrial reptiles, as in terrestrial mammals, the hind foot is more specialized than the front ones. Most reptiles have an external covering or exoskeleton of horny plates or scales or bony scutes. Horny scales are of course not preservable as petrifactions, though in many instances their actual carbonized remains or their impressions have been detected. Such information comes only rarely, though doubtless in the course of time we shall obtain it for most extinct reptiles. In the mosasaurs, for instance, very perfect impressions showing the detailed structure of the scales have been frequently found. Similar impressions were long since observed by Lortet in _Pleurosaurus_, and in not a few dinosaurs impressions of most wonderful perfection have been found. It is only in the water reptiles, probably, that all external coverings tended to disappear.