XLI 



rarely, also in Java. The habits of this species in Oueenslands (Australia) are 

 fully described by T. F. Dodd in The Enlomolooisi ]\xr\e and July 1902; it 

 seems that the larvae, and also the pupae, live in the nests of the ant-species 

 Oecophylla Smaragdina F. found also in Java, and that the former feeds 

 upon the larvae of the ant, while caterpillar, pupae and butterfly are protected by 

 sticky threads and other means of defense, against attacks from the ants. Further, 

 I have reason to believe that the larvae of various Gerydinae species feed upon 

 Aphididae. Such larvae were repeately brought to me as having been found 

 upon the leaves of certain plants, but which I could never induce to eat these 

 leaves. This leads me to suppose, that these larvae do not live upon vegetable 

 food, but upon Aphididae, and all the more, because I found that these 

 caterpillars, when kept together in one box, devoured eachother, a cannibalism, 

 which points rather to carnivorous habits than to those of vegetarian diet. 

 Cannibalism is not unknown amongst Lycaenidae larvae; Swinhoe tells of it 

 for instance in Iraota Timoleon Stoll, and on the authority of other naturalists, 

 in Spalgis Epius Westw. and Zephyrus Ouercus L. In the Lycaenidae-larvae 

 observed by me, the tendency exhibited Itself especially when one of them in 

 passing into the chrysalis stage had stripped off the caterpillar skin, so that it 

 was in a very soft condition ; in this state it was attacked by other caterpillars 

 and eaten up. Others, as quoted by Swinhoe have observed the same thing. 

 An observation made by Dr. Chr. Schroder forms an interesting parallel to 

 this ; he noticed exactly the same thing in the breeding of the larvae of 

 Coccinellidae, these Coleoptera also feeding principally upon Aphididae. Con- 

 cerning this symbiosis of Lycaenidae with ants in Java many and accurate 

 observations are very desirable ; what we know about it is still very incomplete. 

 For the rest there is not much to say about the larvae of Lycaenidae 

 beyond what is found in other works. I will not again enter upon the 

 supposed seasonal varieties. In the introductions to my earlier monographs 

 I fully explained, that I cannot accept such forms, as regards Java. In the 

 dry as well as in the rainy season, individuals are found more or less advanced 

 in the evolutionary process to which they are subjected ; the most that can be 

 said is, that the former are somewhat more numerous in the rainy reason. 

 I only wish to recall to memory a few facts which I have observed and 

 already published elsewhere. I once found some larvae of a Lycaena species 

 presumably Lycaena Cnejus F. inside the pods of a wild plant. These 

 larvae, living thus shut off from the day-light, were very faint in colour ; in 

 one the vas dorsalis showed dark through the skin, seeming to form a dorsal 

 line ; the skin itself was however not marked, but in another there was actually a 

 darker dorsal stripe, just above the vas orsalis. This would lead us to suppose, 



6 



