contain indigenous mammalia or batrachia, and generally a much greater variety of birds, reptiles, insects, and plants, than do the oceanic islands. From these various characteristics we conclude that they have all once formed parts of continents, or at all events of much larger land areas, and have become isolated, either by subsidence of the intervening land or by the effects of long-continued marine denudation.
Now, if we trace the thousand-fathom line around all our existing continents we find that, with only two exceptions, every island which can be classed as "continental" falls within this line, while all that lie beyond it have the undoubted characteristics of "oceanic" islands. We, therefore, conclude that the thousand-fathom line marks out, approximately, the "continental area,"—that is, the limits within which continental development and change throughout known geological time have gone on. There may, of course, have been some extensions of land beyond this limit, while some areas within it may always have been ocean; but so far as we have any direct evidence, this line may be taken to mark out, approximately, the most probable boundary between the "continental area," which has always consisted of land and shallow sea in varying proportions, and the great oceanic basins, within the limits of which volcanic activity has been building up numerous islands, but whose profound depths have apparently undergone little change.
Madagascar and New Zealand.
The two exceptions just referred to are Madagascar and New Zealand, and all the evidence goes to show that in these cases the land connection with the nearest continental area was very remote in time. The extraordinary isolation of the productions of Madagascar—almost all the most characteristic forms of mammalia, birds, and reptiles of Africa being absent from it—renders it certain that it must have been separated from that continent very early in the Tertiary, if not as far back as the latter part of the Secondary period; and this extreme antiquity is indicated by a depth of considerably more than a thousand fathoms in the Mozambique Channel, though this deep portion is less than a hundred miles wide between the Comoro Islands and the main-
