1918] Pupce of Common SphingidcB 407 



and the Semmiaphorce. These groups may be distinguished in 

 the pupae by a very easy character — the presence of spiracular 

 furrows on three abdominal segments, that is 5-7, in the 

 former; and the absence of the furrows from one, two, or all of 

 the segments in the latter. These divisions are further subdi- 

 vided into subfamilies, and of these we have representatives of 

 five, the Acherontiinas and AmbuHcina^ belonging to the 

 Asemanaphorce; the Sesiinas, Philampelinae, and Chaerocam- 

 pinae belonging to the Semanaphorce. It was found possible to 

 separate the first two subfamiHes and the last by well-defined 

 characters; the Sesiinae and Philampelinae were more difficult. 

 All of our genera of Sesiinae except Pseudosphinx seem to fall 

 into a natural group but any character which could be used to 

 separate it from the Philampelinae is shared by Pholus, and in a 

 less degree by Proserpinus, of that subfamily. The remaining 

 genera of Philampelinae form a very compact group. Roths- 

 child and Jordan found the same condition in their study of the 

 adults. On page 349 we find "Of all these issues from the 

 ancestral Sema?iaphorce only one group of genera is sharply 

 circumscribed; it claims the rank of a subfamily. This group, 

 * * * is termed by us the Chaerocampinae. The remaining 

 groups of Semanaphorce are not so obviously distinguished from 

 one another in all their members, owing to the preservation of 

 generalized forms linking the groups together, and to the 

 differences becoming obscured by the recurrence of similar 

 structures in phylogenetically widely different genera." Owing 

 to this difficulty it seems best not to attempt keys to the sub- 

 famiHes, but to furnish a key to the genera of the family and 

 then discuss the genera under the different subfamilies of 

 Rothschild and Jordan. The synonymy follows, in the main, 

 Barnes and McDunnough's Check List which differs slightly 

 from that of Rothschild and Jordan, but is more easily acces- 

 sible for American students. Some sUght changes have been 

 made necessary where it was impossible to separate, or to 

 connect, genera and species in this study. 



