84 JOHN B. SMITH. 



our species of Hemaris. The discussion of further differences be- 

 longs to another place, and the subfjimily character only need be 

 noted here. 



The Ch(crocampin(e have species the most diverse in habitus. 

 None of them, however, have distinctly clavate antenn:e, and none 

 of them, unless the abdomen is tufted anally or at sides, have entire 

 or even outer margins. There is no strong structural character iso- 

 lating the subfamily. I have here placed all those genera eliminated 

 from the Macroglossince, forming of them a small group of aberi-ant 

 forms without definite or close relation to other forms in our fauna 

 and which must be placed in connection with work on other faunas. 

 8uch forms as Enyo and Aellopos are spurs from a tropical fauna, 

 showing numerous similar forms and perhaps entitled to subfamily 

 rank when j)roperly associated. 



A tendency throughout this family is to bright colors and distinct 

 shades and bands in wing maculation. The abdomen is rarely 

 banded, and there are no sober gray forms represented. 



A very large proportion of our sjjccies in the larval state will feed 

 on gra])e, or Ampelopsis. In fact, there are but three exceptions to 

 this rule, though many of those on grape will also feed on other 

 plants. 



This subfamily finds its greatest representation more southwardly, 

 where many of the forms are of the most beautiful and graceful. 



In contrast to the above the Sphingithe are, as a rule, sober gray 

 or brown insects of large or moderate size, with pointed wings and 

 even outer margin, not sinuate or angulated as in the Choei^ocampince. 

 The maculation consists either of simple, undulated, transverse lines 

 like that in so many Bo)ubi/ci(he, or of longitudinal interspaceal black 

 dashes. Sometimes the primaries are a mottled surface of gray and 

 black streaks without distinct pattern. Rarely the secondaries are 

 red. brown, or yellow with bhack margins. 



In this group there is a very general tendency to spinose, or armed 

 tibii^ and tarsi, which in tlie former fjimily is barely indicated in 

 DeUephila. There is a gradual decrease in the length of tongue from 

 forms like eeleus, where it is five ov six inches long, to forms like 

 Ellona harrisii, where it is wanting, or a mere membraneous rudi- 

 ment. 



At the foot of the series are the Smerinthina' with angulated wings, 

 small, retracted head and obsolete tongue. Insects thoroughly bom- 

 biciform in habit and appearance, but completely Sphingiform in 

 larval and inuiffinal characters. 



