Habits of the larva of Lycaena arion. 319 



In discussing various of those matters later with Mr. 

 Bacot, he makes certain comments that may lead us further ; 

 but, at any rate, a short note of them may clear up some 

 points and present others as yet insoluble. Were the 

 larvae not always carried in in last instar ? This seems very 

 improbable, as the only larvae of Lycaenines that hibernate 

 in last instar do so when full-fed. The carrying in would 

 be done probably in third instar, and at the same time and 

 at any rate later, but at first much more rarely, in fourth 

 instar, but previous to a fourth moult. The ants would 

 be likely to attack a larva when soft just after moult, but 

 probably did so rather infrequently, but sufficiently often 

 to make the dropping of the fourth moult advantageous. 

 All this, however, may be incorrect, and the actual adapta- 

 tion more easy, if it be the case that Lycaena (as now 

 represented by euphemus, etc.) only moult three times, and 

 so forth. 



How does Liphyra brassolis manage its moults ? It is 

 clearly regarded by the ants inimically, and has a suit of 

 very strong armour-plate. 



Would the ants not select the smaller larva, first of third 

 instar, and, as the selection diminished the size of the 

 larvae in each instar, then of fourth instar ? 



In view of the comparatively gigantic larvae of argyro- 

 gnomon I have seen ants carrying, viz. small in last instar, 

 it seems more probable that the ants took larvae of any size, 

 since fourth-instar hibernating larvae would not be too large 

 for them, and that the diminution in size was the result of 

 an effort to reach as early as possible the more nutritive 

 food afforded by the ant larvae, i. e. of course that larvae 

 that were sooner in such better circumstances had a sur- 

 vival advantage. Mr. Bacot asks is there anything known 

 of the danger from ants to the external feeding species 

 at the period of the moult, or, a parallel question, is there 

 any suggestion of a specially quick rehabilitation of the 

 honey-gland in contradistinction to other details after a 

 moult ? So far as I can say I think the answer is that we 

 know nothing on these points. The larvae I have found on 

 plants on ants' nests have been in various instars, so pro- 

 bably quite escape attacks at moulting ; but this is only 

 guessing. The ants also build tents over individual larvae, 

 and obviously regard them as valuable cattle to be care- 

 fully tended and defended, but they may eat a tempting 

 larva nevertheless, but I feel tolerably sure that they do 



