ON THE TEETH OF THE SALMON. 243 



relation of analogy, while the latter is obviously one of 

 affinity. 



(217.) The foregoing observations will give the 

 general reader some idea of the most remarkable fish 

 included in the two typical divisions of the Salmonidce : 

 but this manner of treating the subject is so dry, and 

 so little calculated to excite the attention of the natural- 

 ist, that we shall now attempt to investigate the whole, 

 with reference to the natural series, and to the different 

 relations of the minor groups. True it is, that our ana- 

 lysis has not been carried so far into this family as into 

 many others ; and therefore we must have recourse to 

 that mode of investigation already explained.* This 

 has partly arisen from the insufficiency of those charac- 

 ters which have been given to the numerous sub- genera 

 recently proposed, the majority of which are made to 

 rest entirely on the shape of the teeth : thus it is that 

 natural groups have been broken up into smaller ones; 

 and these being all considered of the same rank, subor- 

 dination in their value has been lost sight of. If the 

 teeth are really of such primary importance in classifi- 

 cation, why are they viewed so in one instance, and not 

 in another — not in different orders and families, but in 

 the same genus ? Why, for instance, is Myletes to be 

 separated from Serrasalmo solely " on account of their 

 teeth," when, in the very next sub-genus, Hydrocyon, 

 M. Cuvier states that ee some have a crowded range of 

 small teeth on the maxillaries and palatines ; others, a 

 double range on the intermaxillaries and lower jaw, and 

 none on the palatines ; others, a single range on the 

 maxillaries and lower jaw ; " and, finally, ei others have 

 absolutely no teeth whatever, except on the intermaxil- 

 laries and lower jaw." Here, then, is a group confess- 

 edly varying in the teeth of almost every species, yet to 

 which no definite characters, taken from the shape or 

 fins, are given, by which the naturalist can possibly com- 

 prehend the extent of the author's meaning. For our 

 own parts, we confess our perfect inability to compre- 



* Page 3. (2.) 

 R 2 



