22 CLASSIFICATION OF FISHES. 



cited as a native example. The whole form a group re- 

 presenting the Canthileptes, the SUuridcE, the Plecto- 

 gnathes, and the tortoises, or chelonian reptiles. 



(21.) The last suh-family of the Helotincs, although 

 small and unattractive in appearance, is one of the most 

 singularly interesting groups in the whole family. Cuvier 

 has said of these, that "^they constitute a group formed, as 

 it were, to make naturalists despair, by showing how Na- 

 ture laughs at w^hat w^e deem characteristic combinations. 

 The genera Therapon, Datnia, Pelates, and Helotes, pos- 

 sessing a multitude of mutual relations, as well interior 

 as exterior, sufficient to forbid their distant separation, 

 and bearing a great resemblance to the entire percoid 

 family, at the same time combine species furnished with 

 palatine teeth, along with other species which seem to 

 be constantly deprived of these organs ; they also possess 

 close-set teeth in the jaws, and dentations on the subor- 

 bital, the pre-opercules, and not unfrequently on the 

 shoulder bone, while none have more than six branchial 

 rays : no scales are visible on the cranium, muzzle, or 

 maxillae ; the dorsal spines are folded back into a groove 

 of the back, and the swimming bladder is constantly 

 divided by a restriction into two distinct sacks, as in 

 Cyprin us, Choracinus, and Myripristis." We have given 

 this passage entire, that we may quote it in support of 

 our own arrangement. In regard to the facts thus elicited 

 no difference of opinion can arise ; but our inferences 

 are totally diflPerent from those of the learned author : the 

 very circumstance of these fishes not exhibiting in aU 

 instances the typical character, as to teeth, of the perches, 

 clearly shows that they stand at the confines of the fa- 

 mily ; while the resemblance they bear — according to 

 Cuvier' s own showing — to Myripristis, and con- 

 sequently to the HolocentrincB in general, as clearly 

 shows, in our opinion, that they are as equally related to 

 them. So far, therefore, from these genera exciting our 

 *' despair*' at their seemingly discordant characters, we 

 should have been utterly at a loss to complete the circle 

 of the PercidtE without them j for none but these — by 



