=Barriers to distribution.=--As is indicated in the paragraph on the modes of migration, a considerable sheet of water is obviously a barrier to the further travelling of a walking or crawling land-animal, although no barrier to a winged form. Similarly a strip of land is a barrier to a strictly aquatic animal as a fish. Or a high fall in the stream may serve as an insuperable barrier, making it impossible for any fish below the fall to reach the upper part of the stream. Numerous cases of this kind are known in the Rocky Mountains and Sierra Nevada, where a stream may be well supplied with trout below a fall, and quite bare of these fish above the fall. In the Yosemite Valley in California trout live in the Merced River below the great Vernal and Nevada falls, but above these falls the Merced contains no trout. To fresh-water swimming animals salt water may be a barrier; thus some kinds of fresh-water fishes are limited to one of two near-by streams although the mouths of these streams empty near each other into the ocean. The amphibious batrachians, at home in fresh water and on land, are killed by contact with sea-water. Earthworms also are killed by salt water. Thus the narrowest ocean strait is as effective a barrier to these animals as a whole sea. High mountain ranges and broad deserts are barriers to many land-animals, partly because of the physical obstacles, partly because of the differences in temperature and in food-supply. Temperature and climate (as distinct from temperature) and the ocean are the three great barriers when we consider the animal kingdom as a whole, and look for the causes which determine the chief zoogeographical divisions of the earth's surface. Most of the tropical animals cannot endure frost, hence the isothermal line of frost is a line across which few tropical animals venture. Most arctic animals are enfeebled by heat, and the isothermal line which marks off the region in which frost occurs the year round is another great zoogeographical boundary. But while these lines are limits for localized species, some animals, as birds, especially, keep within a relatively uniform temperature by migrations across these lines. It should be borne in mind that the gradual decrease in temperature met with in going north or south from the tropics is also met in the ascent of high mountains. The summits of lofty peaks, even in the tropics, are truly arctic in character; they are snow-covered, and the animals and plants on them are truly arctic. Thus in the ascent of a single mountain a whole series of life-zones from tropical to arctic can be traversed.