entirely excluded. If they thrive there they may establish colonies which will gradually spread all over the isolated area as far as it affords favourable conditions of life. But oceanic islands are not the only cases of isolated regions; mountains and mountain-ranges which rise in the midst of a plain also form isolation-areas for mountain-dwelling plants or animals which have not much power of migrating. In the same way marine animals may be completely isolated from each other by land-barriers, as the inhabitants of the Red Sea, for instance, are from those of the Mediterranean, as has been clearly expounded by Darwin. The idea of an isolated region is always a relative one, and the region which seems absolutely insular for a terrestrial snail is not so at all for a strong-flying sea-bird. There is no such thing as _absolute_ isolation of any existing colony, for otherwise the colony could never have reached the region; but the degree of isolation may be _absolute_ as far as the time of our observation is concerned, if the transportation of the species concerned occurs so rarely that we cannot observe it in centuries, or perhaps in thousands of years, or if the extension of its range could only take place through climatic or geological changes, such as a subsidence of land-barriers between previously separated portions of the sea, or, in the case of land animals such as snails, the elevation of the sea-floor and the filling up of arms of the sea which had separated two land-areas. But even the transportation of a species by the accidental means already indicated will occur so rarely, if the isolated insular region is very distant, that the isolation of a colony by such a chance may be regarded as almost absolute as far as the members of the same species in the original habitat are concerned. If we examine one of these insular regions with reference to the animal inhabitants which live upon it in isolation, we are confronted with the surprising fact that it harbours numerous so-called endemic species, that is to say, species which occur nowhere else upon the earth, and that these species are the more numerous the further the island is removed from the nearest area of related species. It looks at first sight quite as if isolation alone were a direct cause of the transformation of species. The facts which seem to point in this direction are so numerous that I can only select a few of them. The Sandwich Islands, to which we have already referred, possess eighteen endemic land-birds, and no fewer than 400 endemic terrestrial snails, all belonging to the family group of Achatinellinæ, which occurs there alone.