318 Dr. T. A. Chapman on the Evolution of the 



nests would benefit, and also the ants would more easily 

 transfer them. 



The assumptions here made are not very far-fetched, and 

 if they be granted it becomes possible to understand how 

 such a remarkable life-history became evolved. It is to 

 be remembered that this history refers only to L. arion 

 (probably also to arionides, which though a distinct species 

 is exceedingly close to arion), and that its congeners 

 {ewphemus, areas, alcon) have ordinary Lycaenine life- 

 histories, so that the evolution is comparatively modern and 

 does not involve an ancestry including other species, and 

 therefore does not imply a number of conditions probably 

 unknown to us. That is to say, there can be no doubt that 

 L. arion itself developed these curious habits, and did not 

 inherit them from any other species with ant-hosts. To 

 settle certain points that at present we can only guess, it 

 is necessary to know the precise life-histories of the con- 

 geners of L. arion, especially how many moults they 

 have, and in what instar they do hibernate. Though all 

 those larvae are known, I cannot find definite records on 

 these points. I have reared melanops and cyllarus and 

 have had the early larva of L. alcon, when it mined in the 

 tissues of Gentiana pneumonanthes, but I carried it no 

 further; it is, however, well known that ewphemus, areas, 

 and alcon, which are certainly congeneric with arion, pass 

 the winter as partly-grown larvae, presumably in the 

 normal third instar. 



Cyllarus and melanops are not congeneric with arion, 

 and both, I think, pass the winter as pupae. 



We want actual observation of how the ants collect to 

 one food-plant on the nest the larvae of coridon, thetis, 

 argyrognomon (these, I feel sure, are so collected), and 

 probably others. This habit might easily merge into one 

 of carrying the larva into the nest. In the three species 

 I have mentioned the cannibal habit hardly exists, so that 

 a larva in the nest would almost certainly perish of starva- 

 tion, and thus no arion habit would be initiated. Perhaps 

 the most remarkable fact as regards the hibernation of 

 L. arion is that it does so as a half-grown larva in the last 

 instar. No other Lycaenid is known to do so. This 

 change of habit must have developed after the use of the 

 ants as hosts was established, and is perhaps the most 

 difficult fact in the life-history of which to suggest the 

 development. 



