188 CONTRIBUTIONS TO NORTH AMERICAN ICHTHYOLOGY III. 



himself even seen the lisli upon which the genus is founded, and refers to another 

 genus a species which cannot be separated from this. Moreover, the characteristics of 

 the genus, as given hy Rafinesqne, are not true to nature. Yet, notwithstanding these 

 objections, I do not feel at libertj' to reject his generic name, since it is possible to 

 identify the fish he meant by the vernacular name under which it is known in the 

 West. There is another reason why Rafinesque's description of our western lishea 

 oui;ht to be carefully considered and every possible effort made to identify his genera 

 and species, the fact that he was the first to investigate the fishes of the Oliio and its 

 tributaries upon a large scale, and that notwithstanding the looseness with which he per- 

 formed the task and the lamentable inaccuracies of his too short descriptions, his works 

 bear almost upon every page the imprint of his keen perception of the natural affinities 

 of species, and their intimate relations to one another ; so much so, that even where he 

 has failed to assign his genera atiy characters by which they may be recognized, yet, 

 when the species upon which ihey were founded can be identified, we usually lind that 

 there are good reasons for considering them as forming distinct genera. 



"The trouble with Rafinesqne is, that he too often introduced in his works species 

 •which he had not always seen himself, and which he referred almost at random among 

 his genera, thus difacing his well characterised groups, or that he went so far as to 

 found genera upon species which he had never seen, overlooking perhaps that ho had 

 already described such types under other names. 



"The genus Cychptus affords a striking example of all these mistakes combined 

 together. In his remarkable paper upon the genus Catostomus, Lesueur describes aud 

 figures one species from the Ohio River, under the name of C. elovgattts, peculiar for 

 its elongated cylindrical body, aud for its long dorsal fiu beginning half way between 

 the pectorals and ventrals, and extending as far back as the insertion of the anal. 

 The species Rafinesqne introduces in his subgenus Decactijhis among the genuine Ca- 

 tohtomi, without perceiving that it belongs to his own genus Cydfpiua. This mistake 

 arises undoubtedly from his belief that in Cydeptns there are two dorsals, which indeed 

 he mentions as characteristics of this genus; but this statement is erroneous: the 

 rays of the dorsal are, in fact, enclosed in a continuous membrane, the anterior rays 

 only being much longer than thoee of the middle and posterior portion of the fin ; oc- 

 casionally these long rays split, and accidentally separate from the following ones, 

 when they seem to form two dorsals. 



"The character of this genus, so far as the dorsal is concerned, consists in r»ality not 

 in its division, but in its great extension along the back, and the elongation of its 

 anterior rays. The anal is very long in proportion to the size of the fish, and inserted 

 far back, so that the leiigtli of the abdominal cavity is greater than in tiie genera Car- 

 piodvH, Ichlhyohux, and lUihaUchthys, with which Cythptus is closely allied by the pecu- 

 liar form of its dorsal. Again, Rafinesqne remarks that the mouth is terminal, round 

 and small. This requires also to be qualified. The mouth apjiears terminal and round 

 only when the jaws are protruded to their utmost extent ; wlien closed, it is rather 

 crescent-shaped and entirely retracted under the projecting, pointed snout ; the lips 

 are covered with numerous projecting i)apillie and spread horizontally, — these are 

 moreover, continuous around tlie angles of tlie mouth, so that the upper an<l lower 

 1 ps are hardly separated by a small fold, and the lower lip is slightly cmarginate in 

 the middle, while in other genera of this tribe it is actually bilobod. 



