We must not forget, however, that the faunas of the two continents were always to a great extent distinct and contrasted--such important Old-World groups as the civets, hyænas, giraffes, and hippopotami, never passing to America, while the extinct Oreodontidæ, Brontotheridæ, and many others are equally unknown in the Old World. This renders it probable that the communication even in the north was never of long continuance; while it wholly negatives the theory of an Atlantis bridging over the Atlantic Ocean in the Temperate Zone at any time during the whole Tertiary period. But the past history of the North American fauna is complicated by another set of migrations from South America, which, like those from the Old World, appear to have occurred at distant intervals, and to have continued for limited periods. In the Post-Pliocene epoch, along with elephants and horses from Europe or Asia, we find a host of huge sloths and other Edentata, as well as llamas, capybaras, tapirs, and peccaries, all characteristic of South America. Some of these were identical with living species, while others are closely allied to those found fossil in Brazilian caves and other deposits of about the same age, while nothing like them inhabited the Old World at the same period. We are therefore quite sure that they came from some part of the Neotropical region; but the singular fact is, that in the preceding Pliocene epoch none of them are found in North America. We conclude, therefore, that their migration took place at the end of the Pliocene or beginning of the Post-Pliocene epoch, owing to some specially favourable conditions, but that they rapidly disappeared, having left no survivors. We must, however, study the past history of South America in order to ascertain how far it has been isolated from or connected with the northern continent. Abundant remains of the Post-Pliocene epoch from Brazilian caves show us that the fauna of South America which immediately preceded that now existing had the same general characteristics, but was much richer in large mammalia and probably in many other forms of life. Edentata formed the most prominent feature; but instead of the existing sloths, armadillos, and ant-eaters, there were an immense variety of these animals, some of living genera, others altogether different, and many of them of enormous size. There were armadillos as large as the rhinoceros, while the megatherium and several other genera of extinct sloths were of elephantine bulk. The peculiar families of South American rodents--cavies, spiny-rats, and chinchillas--were represented by other species and genera, some of large size; and the same may be said of the monkeys, bats, and carnivora. Among Ungulata, however, we find, in addition to the living tapirs, llamas, peccaries, and deer, several species of horse and antelope, as well as a mastodon, all three forms due probably to recent immigration from the northern continent.