It will be observed that this case is somewhat like that of Madagascar, but with a characteristic difference. In the case of Madagascar, the separation has been long. The extreme peculiarity of its fauna is the result partly of progressive divergence and partly of many forms saved by isolation. In the case of the coast-islands of California, the time has not been long enough for any great divergence by modification. The peculiarity of its species is due almost wholly to species saved by isolation.[29] _b. Oceanic Islands._--We have seen that faunas and floras of continental islands are somewhat similar to those of the neighboring continent, though with varying degrees of difference--the amount of difference, or divergence by evolution, being in proportion to the amount of time and the impassableness of the separating barriers. But oceanic islands have never been connected with any continent. They are new land formed in the midst of the ocean by volcanic action. When they first appeared they were, of course, without inhabitants of any kind, animal or vegetal. How were they peopled? We answer by _waifs_ from here and there--by _castaways from other lands_. The dominance of particular kinds will depend on the direction of winds and currents, bringing from some lands more than others, and upon the kinds of animals or seeds of plants most liable to be successfully carried across wide seas. Their faunas and floras, therefore, are characterized by a mixture of species resembling, though not usually identical with, those of various lands, with a predominance of those of some one land, and by the singular and complete absence of mammals and amphibians, these being unlikely to be transported by floating timber, as are small reptiles and insects, etc. Among mammals, however, there is a significant exception in favor of bats, the reason being both their power of flight and their habit of concealment in hollow trees, etc. To this explanation, however, we must add that divergence by isolation will meanwhile go on in proportion to time. The Azores, for example, have been peopled from Europe, Africa, and America, but mostly from Europe, on account of the prevailing winds and currents being favorable to colonization from that direction. There are many curious peculiarities in the species, however, because colonization is very slow, and divergent variation has been going on _pari passu_. The Bermudas, on the other hand, have been colonized mainly from America, because of the current of the Gulf Stream. These few examples are sufficient for our purpose, which is only to illustrate the causes of geographical distribution. If any one desires to pursue this interesting subject, we would refer him to that most fascinating book, Mr. Wallace’s “Island-Life.”