"Prior to the elevation of a given mountain chain, the courses of the rivers may generally have been entirely different from their later ones. Subsequent to this period, they can only have undergone partial modifications. As subsequent submergences can rarely have extended to the highlands where such streams originate, the fishes of such rivers can only have been destroyed so far as they were unable to reach those elevated regions, and preserve themselves from destruction from salt water by sheltering themselves in mountain streams. On the other hand, a period of greater elevation of the land, and of consequent greater cold, would congeal the waters and cover their courses with glaciers. The fishes would be driven to the neighborhood of the coast, though no doubt in more southern latitudes a sufficient extent of uncongealed fresh waters would flow by a short course into the ocean, to preserve from destruction many forms of fresh-water fishes. Thus, through many vicissitudes, the fauna of a given system of rivers has had opportunity of uninterrupted descent, from the time of the elevation of the mountain range, in which it has its sources.... "As regards the distinction of species in the disconnected basins of different rivers, which have been separated from an early geologic period, if species occur which are common to any two or more of them, the supporter of the theory of distinct creations must suppose that such species have been twice created, once for each hydrographic basin, or that waters flowing into the one basin have been transferred to another. The developmentalist, on the other hand, will accept the last proposition, or else suppose that time has seen an identical process and similar result of modification in these distinct regions. [Illustration: FIG. 186.--Chub of the Great Basin, _Leuciscus lineatus_ (Girard). Heart Lake, Yellowstone Park. Family _Cyprinidæ_.] "Facts of distribution in the eastern district of North America are these. Several species of fresh-water fishes occur at the same time in many Atlantic basins from the Merrimac or from the Hudson to the James, and throughout the Mississippi Valley, and in the tributaries of the Great Lakes. On the other hand, the species of each river may be regarded as pertaining to four classes, whose distribution has direct reference to the character of the water and the food it offers: first, those of the tide-waters, of the river channels, bayous, and sluggish waters near them, or in the flat lands near the coast; second, those of the river channels of its upper course, where the currents are more distinct; third, those of the creeks of the hill country; fourth, those of the elevated mountain streams which are subject to falls and rapids." In the same paper Prof. Cope reaches two important general conclusions, thus stated by him: "I. That species not generally distributed exist in waters on different sides of the great watershed.