2 AMERICAN LEPIDOPTERA. 



allel wart formation in the Arctiid allies, another great branch of 

 the Bombyees. At first these warts are in line, but soon we reach 

 forms (e. g. Apatelodes) in which the central wart is moved back 

 out of line. At this point a large group of moths in India has di- 

 verged from the type in losing one vein on the fore wings. These 

 are the true Eupterotidre, and form the highest group of the Ptilo- 

 dont allies. The line is, however, almost directly continued by the 

 European genus Lemonia (frenulum gone) into the Lachneidse 

 (cubitus 4-branched), the larva remaining true to type, but becoming 

 gradually more specialized to culminate in the Lachneids. 



Therefore, I would make three groups or families, from the moths 

 most nearly related to the Ptilodonts as follows : 



Fore wings 12-veined. 



Tongue present; larvae simple haired Ptilodonticlse. 



Tongue often absent : larvae with warts Nelalopliidie. 



Fore wings 11-veined, tongue absent ; larvae with warts Eupterotidse. 



The first two families will be treated here ; the third will be found 

 in Hampson's " Moths of India," vol. i, and I would only change 

 his account by removing the genus Cnethocampa. I shall include 

 the genera of the Ptilodonts and Melalophids in a single synoptic 

 table, since the families present no sharp characters in the moths. 

 Dr. Packard includes then both in his Notodontidse, and describes 

 the Melalophid larva as a " hairy Notodontian." This is really what 

 they are, and I differ from Dr. Packard only in making the character 

 of family importance. 



The Melalophidte will thus stand on the same footing as the Apa- 

 telid^e, and it is open to those who so prefer to give both these groups 

 less than family rank. I should probably do so myself were I not 

 especially interested in the larval characters. 



The characters heretofore used for genera are in so far arbitrary 

 that they often separate closely allied species. However, in default 

 of a natural grouping I adopt the characters used by Hampson. 

 They are not more arbitrary than those which I used in the paper 

 with Mr. Neumoegen, while they allow of more easily correlating 

 the Indian genera. 



I am indebted to Mr. A. R. Grote for kindly confirming the types 

 of many of the genera. 



