It is evident that the accidental means of transport mentioned above are insufficient to account for the facts as we find them; we are therefore compelled to seek for further explanation. Anything in the nature of a current furnishes a ready means of transport for Mollusca which have obtained a footing in the upper waters of a river, and there is no difficulty in imagining the gradual spread of species, through the agency of floods or otherwise, over a whole river system, when once established at any point upon it. The feeble clinging power of newly-hatched Limnaea has often been noticed as contributing to the chances of their range of distribution becoming extended. Fresh-water Mollusca, too, or their ova, are exceedingly likely, from their extreme abundance, to be transported by water-birds, which fly without alighting from one piece of water to another. Again, the isolation of one river system from another is, in many instances, by no means well marked or permanent, and a very slight alteration of level will frequently have the effect of diverting the supplies of one watershed into another. When we know what enormous oscillations in level have taken place over practically the whole surface of the globe, we can recognise the probability that the whole river system of the earth has been mixed up and reconstructed again and again, with a very thorough blending of adjacent fauna.
It is possible that the very uniform conditions under which fresh-water Mollusca live may have something to do with the uniformity of their distribution and the comparative sameness in their development. There can scarcely be any question that the environments of fresh-water species are in themselves less varied and less liable to fluctuation than those of species whose home is the land. Water is very like water, all the world over; it may be running or motionless, warm or cold, clear or muddy, but the general tendency is for it to be free from extremes of any kind. Even if the surface water of a lake or river freezes, or becomes unusually hot, there is generally plenty of water at a lower stratum which maintains a less extreme temperature, and to which creatures can retire on the first symptoms of a change. From this two results will follow. Not only will the inhabitants of a piece of water not be inclined to vary much from the type, since their whole surroundings, food, etc., continue very much the same, but, if transported by any accident or cataclysm elsewhere, they will be exceedingly likely to arrive at a place which closely resembles their former home in all essentials. Thus the tendency for new types to be formed would be constantly checked, or rather would very seldom arise.
Mr. Belt, while recognising the importance of changes of level as affecting the distribution of fresh-water species, appears to regard the operations of such changes from a rather different point of view to that described above. “I think it probable,” he writes,[359] “that the variation of fresh-water species of animals and plants has been constantly checked by the want of continuity of lakes and rivers in time and space. In the great oscillation of the surface of the earth, of which geologists find so many proofs, every fresh-water area has again and again been destroyed.... Thus species of restricted range were always exposed to destruction, because their habitat was temporary and their retreat impossible, and only families of wide distribution could be preserved.”
The terrestrial surface of the globe has been divided, as indicating the facts of geographical distribution, into six regions—the Palaearctic, Oriental, Australasian, Ethiopian, Nearctic, and Neotropical. To these is sometimes added a seventh, the Neantarctic, consisting of Chili and Patagonia (and certain islands of the south Atlantic); but since the Mollusca of Chili unmistakably form a part of the Neotropical fauna, it seems hardly worth while to recognise a separate region for those of the extreme south of South America, which have no peculiar characteristics.
In certain points the exact limits of these regions, as indicated by the Mollusca, will probably not correspond to those which are marked out by other zoological classes. Wallace’s line, for instance, does not exist, as far as the Mollusca are concerned.
These regions may be further subdivided into sub-regions, thus:—
| Regions | Sub-regions | |
| Palaearctic |  | Septentrional |
| Mediterranean | ||
| Central Asiatic | ||
| Oriental | | Indo-Malay |
| Chinese | ||
| Australasian | | Papuan |
| Australian | ||
| Polynesian | ||
| Ethiopian | | Central African |
| South African | ||
| Malagasy | ||
| Nearctic | | American |
| Californian | ||
| Neotropical | | Antillean |
| Central American | ||
| Colombian | ||
| Brazilian | ||
| Chilian |
A. The Palaearctic Region
The southern boundary of this region is the northern limit of the African Sahara, the Mediterranean forming no break whatever in its continuity. In Asia this boundary is less well marked, but roughly corresponds to the southernmost of the vast ranges of mountains which border the great tablelands of central Asia. Across Africa the line of desert is well defined; but in the north-east, as the desert approaches more nearly to the sea, the African extent of the region is correspondingly narrowed until it becomes little more than a strip of coast land, scarcely widening even in Lower Egypt. On the Morocco coast, Palaearctic land forms penetrate as far south as Cape Nun.[360] At its eastern extremity the line becomes less well defined, but probably proceeds along the snowy mountains west of Setchouan, the Pe-ling and Tan-sia-shan ranges, so as to include all the high ground of Thibet and of the upper waters of the Hoang-ho, and ultimately reaches its eastern limit at some point on the shores of the Sea of Japan.