54 CLASSIFICATION OF FISHES. 



being covered with close-set^ compact^ spinous scales, or 

 large plates^, by wbicli the body becomes nearly as much 

 mailed as in the file fish (^Balistidte), and the mailed silui-es 

 {LoricanincF). Of three of these genera, which are 

 aberrant, hardly more than one species is at present 

 known; but in the two others there are several sub-genera^ 

 more especially in that of Trigla, composed of the true 

 gurnards, where, from the number of species now de- 

 termined, all the sub-2;enera, would seem to exist. Hav- 

 ingpaid particular attention to this group, which abounds 

 in the Mediterranean, we shall now give the result of 

 theh analysis. 



(54.) The genus Trigia is composed only of those 

 gurnards which have the lower rays of the pectoral 

 fins assuming the form of detached finger-hke pro- 

 fesses, unconnected by a membrane, and furnished with 

 a shshtlv lunated caudal fin. This latter character, 

 indeed, has been so httle regarded in ichthyological 

 groups, that the form of the caudal is often not even 

 noticed by our best authors. Its great importance, 

 however, has been already shown, and in the present case 

 its value becomes apparent, because it is the only de- 

 cisive distinction between the typical gurnards and the 

 ScorpcBuidcE : two gi-oups, abundandy different, indeed, 

 but which have several characters in common ; both, in 

 fact, give us examples of very large pectorals and naked 

 processes ; but the caudal fin of the ScorpcEnidce is always 

 rounded and the body more or less naked. Restricting, 

 therefore, the genus Trigia by these characters, we may 

 trace a gradual progression in the developement of the 

 pectoral fin, which, in the tidbe before us, is of great 

 importance. The nearest approach to Bynchytys is 

 seen in the subgenus Peristedion, where the naked pro- 

 cesses are only two, and the snout is so greatly developed, 

 that the lower jaw is nearly as much beneath it as in 

 Rynchytys. That several forms must intervene between 

 these two cannot, indeed, be doubted ; and, so long as 

 this interval exists, we may even suspect that this rela- 

 tion is merely analogical ; but that question, however im- 



