498 BRITISH LEPIDOPTERA. 



necessarily expecting any other ancestral features. Such a pupa 

 we appear to have in Ceratomia amyntor. After the division 

 into Sphingids and Amorphids occurred, one of the first variations 

 to take place in the latter was in the ist stage larva acquiring 

 (independently) conjoined tubercles. The extraordinary resemblance 

 of Deidamia, a Sphingid, to Smerinthns, both in imago and in pupa, 

 may be merely a result of convergence, but it at least gives one a 

 suspicion that some such form of imago and of pupa obtained in 

 the primitive Sphingid. The essential difference between the 

 Amorphids and Sphingids in the pupal state appears to be the 

 existence in the Amorphids of a dorsal thoracic suture opening 

 on dehiscence, and which is absent on the Sphingid side. This 

 is a definite structural difference, a generalised character in the 

 Amorphids, specialised in the Sphingids. The broad descriptions 

 given of the pupae of the several groups are applicable to the 

 great mass in each group, and are not upset by exceptions, which 

 are, on the contrary, most useful as indications of the points at 

 which the several groups parted their ways. One such pupa is 

 that of Daremma undulosa, whose pupa, at first sight, is distinctly 

 an Amorphid — eyes frontal, wings meeting ; it even has the free pupal 

 surface scale often seen in Amorphid pupae, and so frequent, for 

 example, in the pupa of Mimas tiliae. Not only is the pupa very 

 Amorphid in appearance, but so unquestionably is the moth in several 

 particulars. We find, however, on closer examination, that the 

 dorsal suture of the thorax is absent, showing it to be definitely 

 on the Sphingid side of the dividing line, even if very close to it. 

 Ceratomia amyntor is a very similar pupa, but further from the 

 Amorphids in having the proboscis extended to the wing-apices. It is 

 a Sphinx in everything except that the dorsal movement of the head 

 had not commenced ; this is never so pronounced in moths be- 

 longing to the subfamily Sphinginae as in other Sphingidae, the pupal 

 tongue-horn relieving pressure before it had advanced much, whilst 

 in Eumorphids no pupal tongue-horn appeared until the labrum had 

 been forced to a dorsal position. Everyx myron, E. choerilus and 

 Daraspa versicolor have pupae with head-parts ventrad as in Amorphids; 

 they are, however, without dorsal suture, and, in most other respects, 

 are Eumorphids. The imagines, nevertheless, have a very Amorphid 

 facies, with Amorphid antennae, in full accordance with the 

 Amorphid characters of the pupa. We cannot doubt that here 

 again, this time on the Eumorphid branch, we are somewhere 

 near the common base. When we come to the Sesiid (Macroglossid) 

 division we find three branches that we may call the Eterogoninae, 

 Hcmarinae and Sesiinae ( Macrog/ossinae), which have almost equal 

 claim to separate rank with the Sphinginae and Eumorphinae. The 

 Pterogonids again, as in Deidamia, present many Amorphid features. 

 The Pterogoninae are certainly the most basal of these divisions, 

 with the Hcmarinae as a possible branch, quite separate from and 

 lower than the Sesiid (Macroglossidj division, which presents so 

 many features parallel to the Eumorphids that one is not at all 

 sure that it did not arise from that instead of the Pterogonid 

 division. Further, that group of Sphingids of which we may take 

 Philampelus as the type, and which is, in many ways, very similar 

 to and parallel with the Eumorphids, seems to be a distinct branch, 



