But if we look to the origin of a continent; almost every geologist will admit that in most cases it will have first existed as separate islands which gradually increased in size{409}; and therefore all that which has been said concerning the probable changes of the forms tenanting a small archipelago is applicable to a continent in its early state. Furthermore, a geologist who reflects on the geological history of Europe (the only region well known) will admit that it has been many times depressed, raised and left stationary. During the sinking of a continent and the probable generally accompanying changes of climate the effect would be little, _except_ on the numerical proportions and in the extinction (from the lessening of rivers, the drying of marshes and the conversion of high-lands into low &c.) of some or of many of the species. As soon however as the continent became divided into many isolated portions or islands, preventing free immigration from one part to another, the effect of climatic and other changes on the species would be greater. But let the now broken continent, forming isolated islands, begin to rise and new stations thus to be formed, exactly as in the first case of the upheaved volcanic islet, and we shall have equally favourable conditions for the modification of old forms, that is the formation of new races or species. Let the islands become reunited into a continent; and then the new and old forms would all spread, as far as barriers, the means of transportal, and the preoccupation of the land by other species, would permit. Some of the new species or races would probably become extinct, and some perhaps would cross and blend together. We should thus have a multitude of forms, adapted to all kinds of slightly different stations, and to diverse groups of either antagonist or food-serving species. The oftener these oscillations of level had taken place (and therefore generally the older the land) the greater the number of species <which> would tend to be formed. The inhabitants of a continent being thus derived in the first stage from the same original parents, and subsequently from the inhabitants of one wide area, since often broken up and reunited, all would be obviously related together and the inhabitants of the most _dissimilar_ stations on the same continent would be more closely allied than the inhabitants of two very _similar_ stations on two of the main divisions of the world{410}. {409} See the comparison between the Malay Archipelago and the probable former state of Europe, _Origin_, Ed. i. p. 299, vi. p. 438, also _Origin_, Ed. i. p. 292, vi. p. 429. {410} _Origin_, Ed. i. p. 349, vi. p. 496. The arrangement of the argument in the present Essay leads to repetition of statements made in the earlier part of the book: in the _Origin_ this is avoided.