240 FISHES thus confirming the supposition of many in- telligent fishermen. Internal impregnation is very general in the plagiostomes, and as this is more certain than the indiscriminate spawning of common fishes, the eggs are much fewer in number and of larger size, as in birds ; the egg in its passage through the oviduct receives a dense corneous covering, so that the cases resemble oblong flattened pillows, often with long tendrils at the corners, in which the em- bryo is snugly coiled up ; they become attached to objects floating near the surface, and are there developed by the influence of solar light and heat ; from the researches of Prof. Wyman it appears that in the skates the eggs are fecun- dated in the ovary, and that the egg case is formed in advance to receive it as it descends. From these and other structural peculiarities Agassiz has separated the chimseraa, sharks, and rays from fishes proper, and elevated them into a class, the selachians. Many facts go to show that fishes undergo a kind of metamor- phosis as well as insects. August Muller has proved that the two genera hitherto considered characteristic of the cyclostome fishes are really different stages of the same animal; he has raised ammocetes from the egg of petromyzon, and watched the change of the former into the latter genus. The usual mode of impregnation in osseous fishes, so analogous to the manner in which the fertilizing pollen is brought in con- tact with the stigmata of flowers, naturally suggested the idea of artificial impregnation; and this has been successfully practised both by naturalists for the study of embryology, and by fish breeders as a profitable branch of industry. (See FISH CULTURE.) In most fishes the young when hatched are left to shift for themselves, and of course the greater number are devoured by larger fish, aquatic birds, and reptiles ; many species devour each other ; small mackerel are often found in the stomachs of larger individu- als, when they are abundant ; so that with all their fecundity the class of fishes does not mul- tiply beyond the limits set by nature. Though fishes are cold-blooded, and the watery ele- ment is less affected by sudden changes of tem- perature than the air, there are external cir- cumstances which limit their distribution both in depth and extent of surface. The difference in density and chemical constitution of salt and fresh water draws the line between the marine and the fluviatile faunas ; below a certain depth, probably not far from 120 fathoms, the absence of light and the increase of pressure would prove an insurmountable barrier to most of the class. Fishes are able to resist extreme cold, and to regain vitality after having been appa- rently frozen, but the average of cold has an important influence on their geographical dis- tribution ; the average temperature of the water for the year has been usually taken as the reg- ulator of this distribution, but Dana has shown that the line of temperature established by the average of the 30 coldest days in the year gives the clue to the limits of the marine faunas. A few arctic species are the same in America and Europe, migrating southward from the same northern centre ; but below this region the marine fauna of America is essentially tropical, and that of Europe essentially temperate. In the Atlantic the zones of temperature are re- markably modified by the Arctic, Gulf stream, and African currents; on the American side the temperate zone extends only from Cape Cod to Cape Hatteras, about 10 degrees of lati- tude, while on the eastern it extends from the Swedish coast to the Cape Verd islands, nearly five times as many degrees; while the tropical zone, which in America extends from Cape Hatteras to 25 S., or 60 degrees, on the other side embraces only about 20 degrees on the Guinea coast of Africa. As a few instances of local distribution, in contradistinction to the cosmopolitan scomberoids and cyprinoids, may be mentioned the American cottoids and goni- odonts, the Mediterranean sparoids, the tropi- cal sciaenoids, squammipennce, and mullets ; the pleuronectidoe of the temperate regions; the tropical fresh-water cJiaracini of America and Africa; the true salmons of arctic and cold regions; and the marine labroids, and fresh- water chromids. Estimating the number of vertebrates at 20,000, the number of living spe- cies of fish may be set down at 10,000, of which more than 6,000 are described. Of all the ver- tebrata, fishes are by far the most numerous and widely distributed in the earth's strata ; their remains are found from the Silurian to the tertiary formations, and are of great aid in determining the changes of the surface of our planet during successive and long geologi- cal periods. The first great geological division, the primary age, comprises the lower and up- per Silurian and the Devonian ; till the close of this age there were no air-breathing ani- mals, and in the Devonian period fishes were the lords of creation ; the latter has, there- fore, been very properly called the "age of fishes." Agassiz, in his RecJierches sur lespois- sons fossiles (1833-'43), laid the foundation of fossil ichthyology ; 1,000 species are described in the most complete and scientific manner, with superb illustrations. He divides fossil fishes, as he afterward did the recent ones, into four orders, according to the form and structure of their scales ; these orders, ganoids, placoids, ctenoids, and cycloids, have been sufficiently described in the article COMPARATIVE ANATO- MY (vol. v., p. 172). Three fourths of all known fossil fishes belong to the ctenoids and cycloids, which occur in all formations from the chalk upward; the remaining fourth belong chiefly to the ganoids (with enamelled scales like the garpike and sturgeon) and the placoids (like sharks and rays), and extend through all the fossiliferous strata, but are most numerous in the coal, Jurassic, chalk, and tertiary forma- tions ; no fish with ctenoid scales (like the perch) or cycloid (like the cod) is found below the chalk. The forms of the earlier fishes were many of them very strange ; the pectorals were
