Habits of the larva of Lycaena arion. 321 



It may be that all the larvae reach the fourth instar (like 

 arion) before hibernation ; but since all the third-instar 

 larvae had gone, presumably to hibernate, and only this 

 one taking another moult was left, I take it to have been 

 an exceptional individual taking an additional moult (as 

 occurs in pheretes). So exceptional, however, as to be 

 usually a failure, as in this case. Such a variation occurring 

 in (the original) arion would obviously facilitate the 

 acquisition of its peculiarities as regards moulting. Of 

 course L. alcon may have the same habits as L. arion ; we do 

 not certainly know, but the mature larva is known and has 

 a green coloration, like most of our other Lycaenids. The 

 swampy places where G. pneumonanthe grows are not 

 suitable for abundant ants' nests, and they would have to 

 be abundant, as the butterfly swarms in the locality 

 whence Mr. Powell sent me the flowers, and many flower- 

 heads had a large number of eggs on them, probably each 

 laid separately ; the larva has a well-developed honey- 

 gland. The flower-feeding larvae of L. arion are fiercely 

 cannibalistic ; the numbers of L. alcon that live amicably 

 together in the same flower show that it is entirely free 

 from such a habit. 



If arion, then, like L. alcon, occasionally entered the 

 fourth instar before hibernating, the question as to moults 

 resolves itself merely into losing the fourth moult ; but 

 not really quite so simply as this, since the fact is not so 

 much the loss of a moult as the loss of an instar, viz. the 

 fourth, the third moult not resulting in a penultimate, 

 but in the last instar. This is the fact in Acronycta alni, 

 and should be as easily obtained in L. arion. 



