[Footnote 2: The huge fossil known as "Irish elk" is really a fallow deer and in no way nearly related to the moose.]
Even to an uncritical eye, the differences between ungulates and carnivores of to-day are many and obvious, but as we trace them back into the past we follow on converging lines, and in our search for the prototypes of the carnivora we are led to the Creodonta, contemporary with Condylarthra, which we have seen giving origin to hoofed beasts, but outlasting them into the succeeding age. These two groups of generalized mammals approached each other so nearly in structure, that it is even doubtful to which of them certain outlying fossils should be referred, and the assumption is quite justified that they had a common ancestor in the preceding period, of which no record is yet known.
The most evident points in which Carnivora differ from Ungulata are their possession of at least four and frequently five digits, which always bear claws and never hoofs; all but the sea otter have six small incisor teeth in each jaw; the canines are large; the molars never show flattened, curved crests after the ruminant pattern, but are more or less tubercular, and one tooth in the hinder part of each jaw becomes blade-like, for shearing off lumps of flesh. This tooth is called the sectorial, or carnassial.
Existing carnivores are conveniently divided into three sections: Arctoidea—bears, raccoons, otters, skunks, weasels, etc.; Canoidea—dogs, wolves and foxes; Aeluroidea—cats, civets, ichneumons and hyaenas.
It is highly probable that these three chief types have descended in as many distinct lines from the Creodonta, and that they were differentiated as early as the middle Eocene, but their exact degree of affinity is uncertain; bears and dogs are certainly closer together than either of them are to cats, and it is questionable if otters and weasels—the Mustelidae, as they are termed—and raccoons are really near of kin to bears.
Seals are often regarded as belonging to this order, but their relation to the rest of the carnivores is very doubtful. Many of their characters are suggestive of Arctoidea, but it is an open question if their ancestors were bear or otter-like animals which took to an aquatic life, or whether they may not have had a long and independent descent. At all events, doubt is cast upon the proposition that they are descended from anything nearly like present land forms by the fact that seals of already high development are known as early as the later Miocene.
The difficulty so constantly met with in attempting to state concisely the details of classification, is well shown in this order, for its subdivisions rest less upon a few well defined characters than upon complex associations of a number of lesser and more obscure ones, a recapitulation of which would be tedious beyond the endurance of all but practiced anatomists. For the present purposes it must be enough to say that bears and dogs have forty-two teeth in the complete set, of which four on each side above and below are premolars, and two above, with three below, are molars, but these teeth in bears have flatter crowns and more rounded tubercles than those of dogs, and the sectorial teeth are much less blade-like, this style of tooth being better adapted to their omnivorous food habits. Bears, furthermore, have five digits on each foot and are plantigrade, while dogs have but four toes behind and are digitigrade. These differences are less marked in some of the smaller arctoids, which may have as few as thirty-two teeth, and come very near to dogs in the extent of the digital surface which rests upon the ground in walking.
In distinction from these, Aeluroidea never have more than two true molars below, and the cusps of their teeth are much more sharply edged, reaching in the sectorials the extreme of scissor-like specialization. In all of them the claws are more or less retractile, and they walk on the ends of their fingers and toes.
Cats are distinguished from the remainder of this section by the shortness of the skull, and reduction of the teeth to thirty, there being but one true molar on each side, that of the upper jaw being so minute that it is probably getting ready to disappear.
Civets, genets, and ichneumons are small as compared with most cats; they are fairly well distinguished by skull and tooth characters; their claws are never fully retractile, and many have scent glands, as in the civets. No member of this family is American.