264 BRITISH LEPIDOPTERA. 



them in alliance with Amorphids, of which, possibly, one of the 

 most cogent is the remarkable similarity of the larval markings. 

 It is, indeed, difficult to place some of the more primitive groups, 

 which might with equal propriety be placed in either of these 

 divisions, and yet are not clearly ot sufficient value to be given 

 separate rank — e.g., Darapsa, Ceratomia, Deidamia, Gicrelca, Daremma, 

 &c, of which Ceratomia and Daremma incline to the Sphingid 

 and the others to the Eumorphid side. Chapman states (in lift.) 

 that the pupas of the lower Sphingids are not always- easy to 

 differentiate from those of the lower Eumorphids or Amorphids, 

 a statement that we have already noted as being probably, in some 

 degree, true of them in other stages. This difficulty, he says, " is increased 

 when one has only dead pupae to deal with, but, if we accept 

 Ceratomia amyntor as a Sphingid (sens, restr.), and, in this, we see no 

 difficulty, then we may have the labrum ventral and no tongue- 

 horn present. If Daremma undulosa be a Sphingid (sens, restr.) 

 which is perhaps more doubtful, then we may have a Sphingid (sens, 

 restr.) pupa hardly to be differentiated from that of an Amorphid. 

 The primitive Sphingid (sens, restr.) must, however, have had a pupa 

 almost identical with that of Amorpha. If D. cataipae be a Sphingid (sens, 

 restr.) and our pupae are correctly named, then a Sphingid (sens, restr.) 

 pupa may be, in all respects, like that of an Amorphid, with the wings 

 meeting in front. As our specimens are not dehisced, we cannot say 

 whether it has attained the true Sphinx character, of not slitting the 

 mesothorax on dehiscence, but there is no sign of a suture. The 

 more developed Sphingid (sens, restr.) pupa appears to present two 

 forms, one of which may be typified by those of Hyloicus pinastri and 

 Sphinx ligustri, and is the less specialised. This pupa is fairly straight 

 and fairly cylindrical, and the prespiracular flange of the 5th, 6th 

 and 7th abdominal segments is fairly developed. The proboscis-horn 

 or -case is generally short and simple and pressed close to the front of 

 the pupa. The pupa of Agrius convolvuli tends rather to belong to the 

 following section, but is somewhat intermediate. This second form, the 

 more developed Sphingid (sens, restr.) pupa, may be represented by 

 Phlegethontius, Macrosila or Amphonyx. In these, there is a good deal of 

 antero-posterior flattening and a certain amount of the S-curving 

 characteristic of the similarly flattened pupae of Sesia ( Macroglossa) 

 and some Eumorphids. The flanges of abdominal segments 5, 6, 

 7 are well developed and the horn of the proboscis is largely 

 developed and may be variously curved or coiled. The prothoracic 

 callosity is well-developed and similar roughnesses may be present 

 on the abdominal segments. In the most developed of these, 

 the labrum is quite anterior, but never makes any approach to the 

 fully dorsal situation it occupies in fully evolved Eumorphid pupae. 

 The pupa suggests that Manduca is an AmpJionyx in which the 

 proboscis-case has receded again. In form and outline, callosities 

 and spiracular flanges, it fully agrees with the higher Sphingids. 

 Looked at from this point of view, one may doubt the right of 

 Manduca to rank as a separate subfamily. Amorphines, Hemarines, 

 Pterogonines, Sphingines, Eumorphines, Philampelines, Sesiines, 

 and probably others, all left the original stirps very early and 

 nearly together, Eumorphines, Philampelines and Sesiines (with some 

 other of the old Macroglossid and Chcerocampid families) probably 



