202 A Classification of Lejndojyterous Larvse. 



from all heretofore proposed, namely, the division into Jugatse and 

 Frenatae is abundantly sustained.^ The divisions of the FrenatiB are 

 not strictly sustained ; but the general idea holds good, of general- 

 ized and specialized Frenatae. More than this we seem to have two 

 lines of descent indicated by these terms. I have not found full 

 justification of Prof. Comstock's further division of the Specialized 

 Frenatae into "Frenulum conservers" and "Frenulum losers." In 

 fact, the use of these characters as a division of the Specialized 

 Frenatte is not sustained by a study of the larvse. The loss of the 

 frenulum is rather to be considered as a secondary adaption for 

 flight with broad wings, and it has arisen separately in several dis- 

 tinct lines of descent. 



The two points in which my classification contradicts that of 

 Prof. Comstock now require attention. 



1. Superfamily Zygeenidse. — So far as the North American fauna 

 is concerned there are but two families which claim notice, the 

 Euchromiidae and Pyromorphidas. According to my views, these 

 belong to the two separate lines of descent, the specialized frenatae 

 and generalized frenatae respectively. I will not do more here than 

 call attention to this difference, as the Pyromorphidae are one of the 

 families about which I have been in doubt. 



2. The Family Lacosomidae. — This is placed among the Saturnlna 

 by Prof. Comstock, but from larval characters it belongs with the 

 generalized frenatas. It is true that the moths bear a close resem- 

 blance to the Saturnina; but I believe that this will be found to 

 be fortuitous. I regard it as an instance where specialization has- 

 taken the same form in different lines of descent, and has led to a 

 misinterpretation of the characters. The larv® have retained a 

 generalized condition on account of their secluded life, and give u& 

 truer characters for the classification of the family. It is a case 

 the converse of that of the Eucleidte, where specialization has 

 taken place in the larva, and where the generalized adult gives the 

 best indication of the relationship of these curious insects. 



To show how completely my classification of the larvae contra- 

 dicts the generally accepted classification of Lepidoptera, it is only 

 necessary to briefly consider the "Superfamily Bombyces." No 

 one has prepared a satisfactory definition of this group so far as I 



1 This same division of the Lepidoptera has been made by G. F. Hampson 

 (Fauna of British India : Moths, vol. I, p. 8), but apparently without recog- 

 nition of the real distinctive character. 



