A dog’s fore-paw possesses normally five well-formed toes, while the hind-paws have only four with a rudimentary fifth. This latter disappears in some races, mostly of a diminutive size. In certain large races, on the contrary, it is developed, and becomes equal to the other four. There must be then a formation of bone corresponding to the tarsus and metatarsus. Something analogous to the appearance we have just remarked may be observed in the pig, complicated, however, by a fresh phenomenon. Here the normal foot bears two small rudimentary lateral toes, and two medial toes, each with its own hoof. Now in certain races, already known to the ancients, a third medial toe is developed, and the whole is enveloped in a single hoof. Instead of being _cloven-footed_, which is the normal type of the _species_, the race becomes _solidungulate_. Nothing of this kind is ever seen in man. In every race the feet maintain their ordinary composition, in the Bosjesman as in the Patagonian. Some teratological exceptions with a tendency to heredity are nevertheless occasionally displayed, of which we shall speak in another chapter. VII. The vertebral column is, so to speak, the fundamental portion of the skeleton, and yet it does not vary the less on that account. I shall not insist upon the differences presented by its caudal portion, merely remarking that there are races of dogs, sheep, and goats, in which the tail is so reduced as to be nothing more than a short coccyx. The central portions themselves are known to be liable to change. Philippi tells us that the oxen of Piacentino had thirteen ribs instead of twelve, and, consequently, an extra dorsal vertebra. In the pig Eyton has observed the dorsal vertebræ vary from thirteen to fifteen, the lumbar from four to six, the sacral from four to five, and the caudal from thirteen to twenty-three, so that the total is forty-four in the African pig and fifty-four in the English pig. In man, the presence of one extra vertebra has occasionally been observed. These have always been isolated cases, except in one Dutch family, quoted by Vrolich. But it does not approximate to a constant character in any human group, and if such a group did exist, it is evident that the variation would here again be less than in animals, for without even reckoning the tail, it is three times stronger in the latter. Of course, I do not take into consideration what has been so often said of men asserted to have tails. We now know better how much credit to attach to this statement. But the variations which take place in the caudal region among animals teaches us that even a considerable elongation of the coccyx in a human group, and the multiplication of the vertebræ which compose it, must not be considered _à priori_ as a specific character.