Impressed with the fact that the peculiar scales of Polypterus and Lepidosteus are common to all fossil osseous fishes down to the chalk, he takes the structure of the scales generally as the base for an ichthyological system, and distinguishes four orders:—
1. Placoids.—Without scales proper, but with scales of enamel, sometimes large, sometimes small and reduced to mere points (Rays, Sharks, and Cyclostomi, with the fossil Hybodontes).
2. Ganoids.—With angular bony scales, covered with a thick stratum of enamel: to this order belong the fossil Lepidoides, Sauroides, Pycnodontes, and Coelacanthi; the recent Polypterus, Lepidosteus, Sclerodermi, Gymnodontes, Lophobranches, and Siluroides; also the Sturgeons.
3. Ctenoids.—With rough scales, which have their free margins denticulated: Chætodontidæ, Pleuronectidæ, Percidæ, Polyacanthi, Sciænidse, Sparidæ, Scorpænidæ, Aulostomi.
4. Cycloids.—With smooth scales, the hind margin of which lacks denticulation: Labridæ, Mugilidæ, Scombridæ, Gadoidei, Gobiidæ, Murænidæ, Lucioidei, Salmonidæ, Clupeidæ, Cyprinidæ.
We have no hesitation in affirming that if Agassiz had had an opportunity of acquiring a more extensive and intimate knowledge of existing fishes before his energies were absorbed in the study of their fossil remains, he himself would have recognised the artificial character of his classification. The distinctions between cycloid and ctenoid scales, between placoid and ganoid fishes are vague, and can hardly be maintained. As far as the living and post-cretacean forms are concerned, the vantage-ground gained by Cuvier was abandoned by him; and therefore his system could never supersede that of his predecessors, and finally shared the fate of every classification based on the modifications of one organ only. But Agassiz has the merit of having opened an immense new
