42 GROTE — SPECIALIZATION'S OF LEPIDOPTEROUS WING. [April?, 



larvae are neither homologous nor dependent upon the specializa- 

 tions of the imago. The external influences by which the difl"erent 

 stages are surrounded are radically diverse. It is demonstrable 

 that in Apaiela the larvae are more specialized, as larvae, than are 

 the moths, as moths. These latter are simple Agrotids, or 

 Hadenids. The larvae rank with the Arctians in specialization. 

 Generic differences between images are not necessarily shown also, 

 but may be displayed independently in the earlier stages of the 

 insect. A specialized chrysalis may be attained by a form which, 

 in the imago state, lags behind its fellows. I ventured first to give 

 this view of the independence in specialization of the stages as 

 early as 1876. Mr. Butler's paper on Apatela remains, at least, an 

 exquisite satire on a generic classification from larval characters 

 alone. 



However, I seem to differ from Mr. Scudder (^Hist. Sketch, 103), 

 who holds that generic distinctions are as easily traced in the larva 

 as in the imago, thus assuming a parity in specialization. Never- 

 theless, in the case of the forms of Agrotis Led., we may have 

 moths which offer characters upon which generic distinctions have 

 been founded, while the larvae are so much alike that no such 

 characters appear with them. And again we find species of Apatela, 

 feebly differentiated in the imago state, proceeding from strongly 

 diverging larvae. The whole group of Acronyctid genera is held 

 together by specializations of the larva alone. No intimate char- 

 acters hold Fanthea and Apatela united as moths, and here it seems 

 possible that the larval specializations common to both are non- 

 phyletic, convergent, they have been acquired along different 

 routes, and thus the basis of the family Apatelidae would be artificial. 

 Where no such contradiction is offered the development may be 

 assumed as monophyletic, the classification as natural. This view 

 does not militate against the validity of Dyar's general classification 

 as based on the larval tubercles in position. This character, as 

 pointed out by me in '97, is valuable from its indifference to 

 external influences. 



In the case of the Papilionides there appears to be the alternative 

 that either vein '' ix " has developed subsequently to the disappear- 

 ance of viii or before its appearance. If we accept the latter, then 

 the Papilionides have branched off, as Prof. Comstock says, long 

 before butterflies assumed their present form {^Evolution and Tax- 

 onomy, 112). In this case all traces of an immediate ancestry 



