2. The fishes are equally peculiar and characteristic of the era upon which we enter. Nothing can more strikingly show the violence and universality of the change that was cotemporaneous with the tertiary arrangements, than the total disappearance of the old tribes of fishes, and their replacement by entirely new specific, and a large infusion likewise of new generic, types. The change is no less remarkable when viewed in its relation to existing races, every one of which, with the solitary exception of the salmon family, have here their representatives. Perch, cod, herring, mackerel, eels, had all become occupants of the seas of this period, and their remains deposited in the clays of Sheppey are in the greatest profusion. “The number of fossil fish from the London clay,” says Agassiz, “amounts to ninety-two in the one single locality of Sheppey, without counting ten species to which I have not yet assigned names, not having hitherto been able to characterize them in a satisfactory manner.” The difficulty arises from two causes—the imperfect and fragmentary state of the fossils themselves, and the new principle adopted by him for their classification. Most of the fishes belonging to the tertiary era are of the Cycloid and Ctenoid orders, with thin fragile scales, which, unlike the Ganoids whose cuirassed bodies were protected by a thick covering of plates, have been unable to preserve the integrity of their form and outline. The greater number of these interesting remains, accordingly, have rotted in the matrix, their bones separating, and the soft parts all replaced by clay. The scales are _disaggregated_ (leur sécailles désagrégées), and the cranium alone of the osseous structure remaining entire, owing to the soldering of the pieces composing it, the ingenious naturalist has adopted this single organ as the basis of the new system. “The characteristic features of the skulls of the mammalia and reptilia are known; the variations which such a bone, such a crest, such a groove may undergo in such and such a family are understood; and already, at the first glance, it is possible to ascertain whether the animal under consideration is carnivorous, ruminant, or solipedal. But nothing is more variable than the forms of the cranium and of the heads of fish. The multitude of bones and of spines which serve for the attachment of the muscles, the infinite variety of forms in the families themselves, impart such a diversity to the crania of the fish, that the ichthyologist frequently despairs of being able to reduce them to their respective types, and in fact a comparative craniology of fish does not exist. There is no one, that I know, who can tell at first sight, whether such and such a cranium belongs to a percoid, to a sparoid, or to a chetodontal type.”[8]