42 



only much smaller than the wing in the pupa, but than the portion 

 of the wing within Poulton's line. 



It follows, therefore, that, admitting for the moment that Poulton's 

 line marks off the imaginal wing, and still more if we do not admit 

 it, the imaginal wings contain parts of the wing in the pupa lying 

 outside the position to which it reaches when mature. 



In Orgyia antiqua the female pupa shows Poulton's line tolerably 

 close to its margin ; it has retained it as a normal and constant 

 structure of the wing in the pupal state, and not as marking off 

 the imaginal wing. Were that its real meaning we should expect 

 it to coincide with the limits of the mature imaginal wing, and that 

 the space outside it would be enlarged. The wing of the pupa does 

 not diminish so rapidly as that of the imago, as it atrophies in these 

 apterous female moths ; therefore the margin of the pupal wing 

 ought to be a long way outside that of the imaginal wing. There- 

 fore, if Poulton's line were that margin, we should find it far inside 

 its present position, and the total pupal wing would probably have its 

 margin outside its present position. 



It is not, however, in these pupae of apterous moths that we find 

 a great space between Poulton's line and the margin of the pupal 

 wing, but in certain butterflies which have very ample wings and in 

 which there is no reason at all to suppose that the wings have 

 dwindled in size in the imagines. Perhaps the most remarkable of 

 these are in various Pieridce. In Euchloeihe wing beyond Poulton's 

 line extends to a great distance, and helps to form the point at the 

 angle where the pupa appears to be so curiously bent. It has been 

 suggested that this extended area beyond Poulton's line represents 

 the wing of some ancestor who had very prolonged hooked tips to 

 the wings; if this be so, we shall of course find in C. rham/ii con- 

 siderable correspondence of Poulton's line with the wing margin. 

 As a matter of fact it is much the contrary ; we have in the pupa of 

 C. rhamni a very similar general form to that of Eitcliloc, and a 

 very similar prolongation of the wing to form the angle. Poulton's 

 line is, however, very distant ; it marks out the form of the wing as 

 we see it in the imago, but the point of the hooked tip in no way 

 corresponds with the apex of the pupal wing. 



With regard to the butterflies, we must recognise that their pupae 

 have varied immensely and developed various remarkable spines, 

 processes, &c., entirely in view of the exigencies of their environ- 

 ment ; and the immense development of this extreme portion of the 

 pupal wing in many species is no doubt more related to this circum- 

 stance than to any condition directly concerning the wing itself. In 

 any case, in order to fall smoothly on to the rounded surface of the 

 fourth abdominal segment, it was necessary that the wing margin 

 should preserve a regular curve and avoid all angulation, and so we 

 never in any pupa find a sharp angle on the hind margin of the 

 wing. So far as I am able to suggest, we must take it as an ultimate 

 fact that this wing margin was able to vary for pupal purposes to the 



