REMARKS UPON METAMORPHOSIS. 279 



of such as are exclusively founded upon metamor- 

 phosis; and we therefore consider that metamor- 

 phosis, however valuable in helping to distinguish 

 large assemblies, is pushed much further than 

 nature warrants, when it is used as the chief corner 

 stone for the construction of genera and sub-genera ; 

 which groups, if they are really natural, will always 

 be distinguished by other and more intelligible 

 characters. 



(189.) From what has been just said, it must 

 not, however, be inferred, that the metamorphosis 

 of insects is to be disregarded ; or that characters 

 derived therefrom are not to be employed. The 

 value of this and every other character depends 

 upon the judicious skill with which it is used. In 

 a group of unarranged animals, we can never know, 

 a priori, what are truly its essential characters. It 

 therefore becomes necessary to study all, that we 

 may discover their relative prevalence, and there- 

 upon make our selection. Another important ad- 

 vantage will also result from such investigations : 

 some characters will be brought to light, which, 

 although not employed as essential distinctions of 

 the group, will nevertheless throw considerable light 

 on its analogical relations. A more beautiful 

 instance of this cannot possibly be found, than in 

 the subordinate types of the genera Amphrisius 

 and Papilio, detailed at some length in our ZooL 17. 

 second series. Each of these genera forms a circular 

 group ; and the contents of one intimately corre- 

 spond with the contents of the other, not only in the 

 form of the perfect insect, but even in the larva of 

 all such as are yet known. It has been well said, 

 T 4 



