THE REVISION OF THE SPHINGIDES. 77 



The Smerinthids are certainly the group retaining the largest 

 number of ancestral characters. They have not gone forward with 

 the others, but have retreated, especially in the matter of proboscis 

 and wing power. They took the Sphinges with them in those 

 characters which make both Axemanophorae. The Choerocampids, 

 however, are more primitive as larvae, so far as having preserved 

 early stages goes, less so in so far that one type of larva was early 

 reached by the Asemanophorae, and was adhered to, whilst the 

 Choerocampids developed in various directions. 



Darapsa, Ceratomia, Smerlnthus are related together in characters 

 that seem to us to be ancestral, but in the Revision are treated as 

 highly evolved by recession. As regards the proboscis, it no doubt 

 relies on these having shorter proboscids than we postulate for the 

 ancestral Sphinx. Granted that these have receded in this matter, and 

 we must go to sundry Sesiads to find genera in which neither recession 

 nor advance has obviously occurred. 



The abdominal spines also seem to us to have originated all over 

 the surface of the segments, to have failed except at the margins, 

 where they persisted in several rows, but as they got larger and 

 stronger one row became sufiicient. In this matter we think Pseudo- 

 sphinx tetrio is probably at the highest stage of development. In this 

 and other characters, to regard this species as ancestral in the sub- 

 family Sesianae seems difficult to accept. Haeinurrhagia, for instance, 

 is in many points more primitive, highly specialised though it is in 

 others. These spines are not, as the Rerision says, altogether special 

 to the Sphingides, but occur in some Noctuids and Geometrids (in 

 Nyssia hispidaria they are rampant). 



We fully agree with the sinking of all subfamily or tribal distinc- 

 tion between the Amorphid and Ambulicid sections of the Sineriuthinae 

 {Aiiihulicinae) unless a number of smaller divisions be made. 



The Seinanophorae (Macro(/lossi)iae and Choeyoccanpinae, Auct.) include 

 some 480 species. Amongst these they describe the subfamily Choero- 

 campina as being sharply circumscribed. It includes 144 species. 

 The majority of our British Choerocampids (Auct.), belong to this sub- 

 family — elpenor, t/allii, eelerio, etc. The genus Xijlophanes, with 50 

 species, is confined to the western hemisphere. In the arrangement 

 in the phylogenetic table, we believe, chiefly on larval characters, 

 that Pe)yesa {elpenor, etc.) is wrongly placed with ('elerio {(jallii, etc.), 

 and away from Hippotion and Theretra {eelerio, etc.). 



"The remaining groups of Semanophorae are not so obviously 

 distinguished from one another in all their members owing to the 

 preservation of generalised forms linking the groups together, and to 

 the dift'erence becoming obscured by the recurrence of similar structures 

 in phylogenetically widely difterent genera." This is so true that it 

 seems impossible to criticise the arrangements of these genera, any 

 other being probably open to at least as much objection as those in 

 the Revision, though there are one or two points, perhaps, worth 

 considering. 



The Dilopkonoticae appear to be a well-defined group, and have 

 such a Sphingid facies that they have usually been classed as Sphinges 

 and not as Choerocampids. The larvae and pupas are definitely Choero- 

 campine. Burmeister separated them as a distinct subfamily without 

 distinctly stating their Chcerocampine affinities, which is first done, 

 we believe, in the Reoision. 



