44 Journal New York Entomological Society. [Voi. v. 



Cutting open the other side of the cocoon, I found that the pupa shell 

 was sucked nearly dry of its contents. The Cecropia cocoons occur 

 commonly on white maples and are generally placed near the ends of 

 the long drooping branches, and it will be seen from the foregoing that 

 it is probably the safest situation afforded by the tree. If a woodpecker 

 is successful in making a hole into a cocoon, it is, nevertheless, some- 

 times disappointed at its contents. I have found a cocoon that con- 

 tained the tough pupa case of the Ophion ichneumon fly, that had been 

 drilled in the side by a woodpecker, and then abandoned, leaving the 

 parasite unharmed. 



THE CLASSIFICATION OF THE SATURNIIDES. 



By a. Radcliffe Grote, A. M. 



The publication by Dr. Dyar* of a critical notice of my recent 

 paper (June, 1896)"!' on the Satiimiides, affords me, in replying, the 

 opportunity of briefly stating the characters which I found in the group. 

 I founded the two families into which the superfamily naturally divides 

 (any other division being in my opinion unnatural) as follows : 



Vein IV, anastomosing with IVj Saturniid.^. 



Vein IV2 out of the cross-vein AgliiD/E. 



Perhaps some reason should have been given by Dr. Dyar for call- 

 ing this fundamental difference in the neuration "artificial," while con- 

 trasting it with a "natural classification which should combine several 

 such special ones." But this combination does not exist; it remains 

 ideal. It reminds one of the hazy statement, that we must take charac- 

 ters from all parts of the insect, which procedure, without a strict 

 weighing of values, would lead us nowhere. But the fact is, that al- 

 though I have taken the structure of the Radius as the principal charac- 

 ter, determining as it does the dichotomous division of the superfamily, 

 I have not left out of sight the characters of differentiation offered by the 

 larvae and cocoons. I have worked out the gradual modifications of 

 the Radius in the highest of the two families. I have not "selected " a 

 random or arbitrary character, which would in the end fail. I have 

 been obliged to take the fundamental character which carries with it all 



*Can. Ent. XXVIII, 270. 



t Miltheilungen aus dem Roemer Museum zu Hildesheim, No. 6. 



