198 



Mt'inoii's 0/ tlic I ml in It Miiatiiiii. 



[Vol. Ill, 



'iiandibles. In the case of the Aceraiinae it was useless to introduce any reference to 

 this character into the key to the genera, as these were found to be well defined and 

 to form a linear series of increasing specialization.' But in the Gnaphalocneminae it 

 is more useful; for the genera, besides being less sharply defined by other charac- 

 ters, form a branching series with one or more genera in which the dentition is 

 reduced at the end of each branch; and the nature of the reduction , which is the 



•''ifO-MAt 









\°o \ \ 



o \ \ / 



\%' \ ^. i^^/teyt/e-ifftcnai^ 



\ 

 \ 



\ /* / // 



\ I / / ^^ 



■i'' 



I ^ 



I ? 



\ I "a. 



same in all genera in which it occurs in any one branch, is difierent in different 

 branches. 



Zang has attached great importance to the fusion of the (anterior) intermediate 

 and lateral areas of the metasternum in certain genera of Gnaphalocneminae, and here 

 we have another character, usually very sharply defined, which is most useful for the 

 determination of these genera. But it does not appear to me to be of such primary 

 importance as the sculpturing of the mentum and of the anterior margin of the head ; 

 and Zang's " Kmipiolus Group," which includes all genera so characterised, does not 



' I have had to modify my views ou these points, see Appendix I, p. 316. 



