THE CANADIAN ENTOMOIX>GIST. 421 



are plainly present in the fourth and fifth instars of the flat larva shows 

 that these are not Gracilarian instars. 



However, even conceding that there is no great structural difference 

 between the types of larva?, the very appreciable modification of form in 

 the fourth and fifth instars and the non-functional character of the mouth- 

 parts in the sixth and seventh instars of the flat group still await explana- 

 tion. Remembering that the imagoes of LitJiocolletis (typical) and 

 Cameraria are structurally identical, the question resolves itself into a 

 consideration of how much reliance should be placed upon these larval 

 characters in determining the phylogeny. It is true that in the absence of 

 imaginal characters, larval characters may furnish a basis of classification, 

 but before accepting the testimony they afford as final, we should examine 

 them critically to determine whether they represent the phylogenetic 

 divergence of the group or are merely cenogenetic larval modifications 

 adapting that group to different life conditions. 



A phylogenetic tree which shows the independent origin and parallel 

 descent of two groups, distinguished by the larva of one being flattened, 

 the other cylindrical, must be based on the assumption that, e. g., those 

 characterized by a flattened larva are descended from genera or groups, 

 now extinct, which possessed this characteristic. This line of reasoning 

 rests on the hypothesis, which has repeatedly been shown to be unreliable 

 for free-living larval forms,* that the individual recapitulates in its 

 ontogeny, the history of the race. This, it seems to me, is the fundamen- 

 tal weak point in such a phylogenetic tree as that proposed for the group 

 under discussion by Mr August Busck (Proc. Ent. Soc. Wash., XI, ioo, 

 1909). On the other hand, we are justified in concluding that the 

 common possession of at least two Gracilarian instars is proof, additional 

 to that furnished by the imagoes, of the common descent of the group 

 from .Gracilariad ancestry, because this characteristic has been handed 

 down through so many modifications of imaginal structure and environ- 

 mental conditions that it may well be assumed to be conservative. In the 

 fiat-larval group, we have no such basis of comparison to determine 

 whether the later two flattened stages constitute such a conservative char- 

 acter or not, and hence can not accept the evidence afforded by the 

 ontogeny. 



If the flattened form of the larva in the fourth and fifth instars and 

 the slight modification of mouth-parts in the sixth and seventh instars, 



*See Montgomery, "The Analysis of Racial Descent in Animals. 



