282 BRITISH LEPIDOPTERA. 



adaptive character. This difficulty is, perhaps, as great in the 

 Attacids, as anywhere, though the thickness, straightness and circular 

 section, as though turned in a lathe, are usually characters at once 

 recognisable in many genera both of Citheroniids and Saturnians, 

 the large antennae being also very obvious facts. Dimorpha is a 

 generalised Attacid in many respects. Its relations are nearer to 

 Citheroniids than to Saturnians, but its pupa is curved, whilst 

 those of Citheroniids are straight, and its armature everywhere it 

 spiculate, whilst that of Citheroniids possesses pits. The arrange- 

 ment of face-, wing- and mouthparts is very much the 

 same. There are certain resemblances but it is impossible to 

 say whether they do or do not indicate alliance" (in litt.). 

 Attention has already been drawn to the sexual differences exhibited 

 by the pupal antennae ot the Attacids (antea, vol. ii., pp. 46, 47), 

 and Poulton gives (Morph. Lep. Pupa, pi. xxvi., figs. 1 — 13) an 

 excellent series of drawings of the pupal antennae of Saturnia pavonia, 

 Aglia tan, etc., comparing them with the corresponding imaginal 

 antennae. Mosley, in 187 1, made some investigations into the 

 relationship, observing that, in Saturnia pavonia, the sheaths of 

 the antennae in the female pupa are large and inflated, with traces 

 of pectination, resembling in this respect those of the male pupa, 

 but in a reduced degree, although the antennae of the female imago 

 are as is well-known merely filiform. Mosley concluded from this that 

 " in the ancestral Saturniids the imagines of both sexes must have 

 had large pectinated antennae, and that they had not been developed 

 as such only in the male for sexual purposes, but must have been retained 

 in the male and degenerated in the female." He considers that " the 

 pupal integument, requiring for purposes of protection to be hard and 

 rigid, when once it had assumed a particular shape, suited to the 

 contours of the body of the original ancestral insect, would retain 

 that form indefinitely, although the shape of the several parts of the 

 imago formed within it might, by the action of natural selection 

 on it when in the free moving condition, alter considerably. It 

 appears to be of no detriment to the development of the legs and 

 the antennae of the imago that their pupa-cases are far too big and 

 do not fit." He concluded from these observations that probably 

 members of the Saturniids existed with well-marked pectinated 

 antennae in both sexes, and soon found on looking at Westwood's 

 drawers of insects that this was the case. Keeping to the Attacids, 

 Poulton finds that there is an immense difference between the imaginal 

 antennae of the two sexes of Aglia tan, while the corresponding pupal 

 organs are not widely different. Although the antennae of the female 

 imago are extremely degenerate, a careful examination, he says, reveals 

 traces of the structure, which is so elaborately developed in the male 

 (Joe. rit., p. 247). Poulton thinks that Mosley's conclusions require some 

 modification, and he believes that " the particular size and shape 

 of a pupal organ, which, at any one time, fits an imaginal organ 

 developed within it, will not be retained indefinitely upon the 

 shrinkage of the latter. . . but the pupal organ will also eventually 

 become smaller." Poulton further considers it probable that " the 

 male imaginal antennae have increased as the female organs have 

 diminished, so that the pupal organs of the former must have under- 

 gone recent increase, while the female pupal antennae may indicate 



