342 CLASSIFICATION OF FISHES. 



milton are to the same effect. The teeth, in fact, 

 throughout the whole of the SiluridcE, are so minute, 

 that their structure can seldom be determined, except 

 under the power of a high magnifier. The value of a 

 character can only be determined by its extent ; and, 

 according to this rule, we find that it either serves to 

 designate a family, or a species. The fins of fishes, as 

 every one knows, are analogous to the wings of birds ; 

 and we were the first, in conjunction with professor 

 Temminek, who employed the modifications of the pri- 

 mary quills as characters for the discrimination of the 

 natural groups among birds. There is consequently 

 good reason to suppose that the same importance should 

 be attached to the structure of these organs among fishes; 

 and the proofs offered in this volume, for such an hypo- 

 thesis, are not few. The number of the cirri among the 

 SiluridcE seem more constant than the characteristics 

 of which we have just spoken : but they do not appear 

 to be sufficient for sub-generic distinctions, when not 

 supported by other and more constant characters. 



(281.) Leaving now the two typical genera of the 

 PimelodintEj we may particularly mention three others, 

 each of which presents some very prominent and pecu- 

 liar characters ; these we shall look on in the light of 

 generic types — not so much from our belief that the two 

 first are really so, but that they may be kept for the 

 present distinct from the two presumed typical groups. 

 The first of these is Sisor*, which has the head broad, 

 depressed, covered with bony plates marked with tu- 

 bercles and ridges, and terminating behind in three nar- 

 row processes. The cirri are more numerous than in 

 any other Siluri yet discovered — amounting to no less 

 than fourteen, all of which, however, are shorter than 

 the head; the lips fleshy; and there are no teeth : the 

 eyes are small, and very high. Between the head and 

 first dorsal fin is a bony plate; and near the latter are 

 five small plates, nearly united, and disposed in a double 



* Sisor rabdophoruSt Hamilt, p. 208- 



