372 BRITISH LEPIDOPTERA. 



transverse axis. No such change occurs in Phtsia where one might 

 expect it, if the long proboscis were the sole cause. In Phtsia, 

 however, the extra space required in the pupa by the long proboscis 

 is obtained not close to the head, but at and beyond the wing 

 apices, whilst in Sphingids, apart from the keel, the central base 

 of the proboscis extends very far forwards, which is much the same 

 as the head being thrown backwards, and the labrum, usually on 

 quite the ventral aspect of a pupa, becomes situated on 

 the anterior apex of the pupa. There is one other characteristic 

 of a Sphingid pupa to be noticed, that is the specialisation 

 of the lateral surfaces of the 5th, 6th and 7th abdominal seg- 

 ments in front of the spiracles. If we examine a pupa 

 of a Citheroniid, say atheroma regalis, we find the anterior 

 margin of the segments presents a raised line, equally developed 

 in all aspects. The other segments are similarly constructed, but the 

 fact is obscured by their not having these margins free as in the 5th, 6th 

 and 7th. In an Amorphid pupa, as that of Smerinthus ocellata, this line 

 is much less definite, but is very distinct as a special ridge just in front 

 of the spiracle. In many Sphingids it is but little different from this. 

 Even in Manduca there is a strong blunt ridge just in front of the 

 spiracles that might easily be taken as the ridge in ,S. ocellata, a 

 little more developed. It has here, however, a strong margin opposed 

 to it on the adjacent segment, and is clearly the product of a fuller 

 and longer development than in Smerinthus. If, as an example of the 

 Citheroniids, we had chosen Eacles imperialis, we should have found no 

 regular line, but, in the prespiracular region of these segments, we should 

 have observed that the general sculpture ol the segment, which is else- 

 where pitting, had here arranged itself as several parallel transverse 

 fine sharp ridges. A precisely similar sculpturing occurs in many 

 Eumorphids, and, progressing by the gradual increase of one ridge 

 at the expense of the others, culminates in one strong sharp ridge 

 anterior and slightly dorsal to the spiracle. In Manduca, this single 

 ridge remains short, sharp and strong, and with still a row of minute 

 ridges in front of it. In Sphinx the condition is not very different 

 from that in Manduca. There are some Eumorphid pupae in which 

 this ridge is developed into a strong flange, and in one, at least, of 

 these species, which escapes from the ground as a pupa before 

 emergence, it is this flange that is probably the active agent in 

 enabling the pupa to force itself upwards. It is very possible that 

 similar structures to these prespiracular flanges of Sphingid pupae 

 exist here and there in other lepidoptera, but I have not met with 

 them, and doubt their existing throughout whole families as in the 

 Sphingids. Their use no doubt is to assist the pupa in getting a 

 firm basis to resist the pressure necessary to enable the imago to 

 break through the earth above it. Most Sphingids bury themselves 

 in the earth for pupation, probably all do when a favourable place 

 is obtainable, but some are not very averse to pupating under a stone 

 or amongst surface debris, and some will even make something like 

 a silken cocoon. That all prefer to "go down," if possible, is 

 probably true. It is noted that some use these flanges to emerge 

 as pupae. In Manduca and some others the flanges very decidedly 

 face forwards, so that their function is here probably somewhat altered, 

 and they serve to hold the empty pupa-case back, rather than as 



