The Descriptions and Figures of Two dHundred' Fishes collected at Vizagapatam on the Coast of Coroman el (Lond., 1803, 2 vols.) b Patrick Russel, and An ccount of the Fishes found in the River Giznges and its Branches (Edin., 1822, 2 vols.) by F. Hamilton (formerly Buchanan), were works distinguished by greater accuracy of the drawings (especially the latter) than was ever attained before. A Natural History of British Fishes was published by E. Donovan (Lond., 1802-1808); and the Mediterranean fauna formed the study of the lifetime of A. Riseo, Ichlhyologie de Nice (Paris, 1810); and Histoire naturelle de l'Eurvpe méridionale (Paris, 1827). A slight beginning in the description of the fishes of the United Statesswas made by Samuel Lat am Mitchell (1764-1831), who published, besides various papers, a Memoir on the Ichlhyologyof Iilew York, in 1815.
G. Cuvier (1760~183z) devoted himself to the study of fishes with particular predilection. The investigation of their anatomy, CWI" and especially of their skeleton, was continued until he had succeeded in completing so perfect a framework of the system of the whole class that his immediate successors required only to fill up those details for which their master had had no leisure. He ascertained the natural affinities of the infinite variety of forms, and accurately defined the divisions, orders, families and genera ~of fthe class, as they appear in the various editions of the Régne Animal.” His industry equalled his genius; he' formed- connections with almost every accessible part of the globe; and for many years the museum of the Iardin des Plantes was- the centre where all ichthyological treasures were deposited. Thus Cuvier brought together a collection which, as it contains all the materials on which his labours were based, must still be considered as the most important. Soon after the 'year 1820, Cuvier, assisted vm by one of his pupils, A. Valenciennes, commenced c, e, ,:es his great work' on fishes, istorie naturelle des Poissons, of which the first volume appeared in 1828; After Cuvier's death in 1832 the Work was left entirely in the hands of Valenciennes, whose energy and interest gradually slackened, rising to their former pitch in some parts only, as, for instance, in the treatise, on the herring. He left the work unfinished with the twenty-second volume (1848), which treats of the Salmonoids. Yet, incomplete as it is, it is indispensable to the student.
The system finally adopted by Cuvier is the followings- » A. POISSONS OSSEUX.
I. A BRANCHIES EN PEIGNES on EN LAMES.
1. A Mdchoire Supérieure. Libre.
a. A canthopléry iens.
Percoides. Sparoides. iiranchies labyrinthii ues. Polynemes. Chétodonoides. Lophioides.
M ulles. Scombéroides. Gobioides.
joues cuirassées Muges. Labroides.
b. M alacoplérygiens.
A bdominaux. Subbrachiens. A fades.
Cyprinoides. Cyadoides. M urénoides.
2. A Mdchaire. Supérienre Fixée.
Selérodermes., Y Gymnodontes.
II. A BRANCHIES EN FORME DE HoUPPEs.
B. CARTILAGINEUX OU CHONDROPTERYGIENS. E
Sturioniens. Plagiostomes. CyclostomesL
We have only to compare this system with that of Linnaeus if we wish to measure the gigantic stride made by ichthyology during the intervening period of seventy years. The various characters employed for classification have -been examined throughout the whole class, and their relative- importance has been duly weighed and understood. The important category of “ family ” appears now in Cuvier's system fully established as intermediate between genus and order. Important changes in Cuvier's system have been made and proposed by his successors, but in the main it is still that of the present day.
Cuvier had extended his researches beyond the living forms, into the field of palaeontology; he was the first to observe the close resemblance of the scales of the fossil Palaeoniscus to those of the living Palyplerus and Lepidostens, the prolongation and identity of structure of the upper caudal lobe in Palaeoniscus and the sturgeons, the presence of peculiar' “ fulcra ” on the anterior margin of the dorsal -fm in Palaeoniscns and Lepidosteus, and inferred from these facts that the fossil genus was allied either to the sturgeons or to Lepidosteus. But it did not occur to him that there was a close relationship between those recent fishes. Lepidostens and, with it, the fossil genus remained in- his system a member of the order of .M alacoplerygii abdominal es. .
It was left to L. Agassiz (1807-1873) 'to point out the importance of the structure of the scales as a characteristic, and to open a. path towards the knowledge of, a whole new subclass A sh of fishes, the Ganoidei. Impressed with the fact that gas the peculiar scales of Polypterus and Lepidoslens 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.'Placoid.v.—Without scales. proper, but with scales of enamel, sometimes large, sometimes small, and reduced to mere points (Rays, Sharks and Cyclostomigwith. the fossil Hybodontes). 2. Ganoids.-With angular bony scales, covered with ia thick stratum of enamel: to this order belong the fossil Lepidoides, Sa-uroidcs, Pycnodontes and Coelacanthi; the recent Polypterus, Lepidosteus, Sclerodermi, Gymnodontes, Lopbobranches and Siluroicles; also the Sturgeons. 3. Ctenoids.-With rough scales, which have their free margins denticulated: Chaetodontidae, Pleuronectidae, Percidae, Polacanthi, Sciaenidae, Sparidae, Scorpaenidae, Aulostomi. 4. Cycloiali. -With smooth scales, thehind margin of which lacks denticulation; Labridae, Mugilidae, Scombridae, Gadoidei, Gobiidae, Muraenidae, Lucioidei, Salmonidae, Clupeidae, Cyprinidae. 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 offossil remains, he would doubtless have recognized the artificial character of his classiQ fication. ' The distinctions between cycloid and ctenoid scales, between placoid and ganoid fishes, are vague, and can hardly be maintained. So far as the living and post-Cretacean forms are concerned, he abandoned the vantage-ground gained by Cuvier; and therefore his system could never supersede that of his predecessor, and finally shared the fate of every classification based on the modifications of one organ only. But Agassiz opened an immense new field of research by his study of the infinite variety of fossil forms. In his principal work, Recherches snr les poissons fossilcs, Neuchatel, 1833-1843, 4to, atlas in fol., he placed them before the world arranged in a methodical manner, with excellent descriptions and illustrations. His power of discernment and penetration in determining even the most fragmentary remains is astonishing; and, if his order of Ganoids is an assemblage of forms very different from what is now understood by that term, he was the first who recognized that such an order of fishes exists. ' ' V The discoverer of the Ganoidei was succeeded by their explorer Iohannes Müller (1801-1858). In his classical memoir Uber den Ban und die Grenzen der Ganoiden (Berl., 1846) he showed that the Ganoids differ from all the other osseous fishesyand agree with the Plagiostomes, in the structure of the heart. ' 'By this primary character, , all heterogeneous elements, as Siluroids, Osteoglossidae, &c., were eliminated from the order as understood by Agassiz. On the other hand, he did not recognise the affinity of Lepidosiren to the Ganoids, but established for it a distinct subclass, Dipnoi, which he placed at the opposite end of the system. By his researches into the anatomy of the lampreys and Amphibxus, their typical distinctness from other cartilaginous fishes was proved; they became the types of two other subclasses, Cyclostorni and Leptocardii. I Müller proposed several other modifications of the Cuvierian system; and, although all cannot, be maintained as the most natural arrangements, yet his researches have given us a mucli more complete knowledge of the organization of the Teleostean fishes, and later inquiries have shown that, on the whole, the combinations proposed by him(require only some furtheif modification and another definition to render them perfectly*