THE ASSOCIATION OF ANTS WITH BUTTERFLY LARVJE. 33 



segment, with lips like a mouth. These lips can, at the will 

 of the larvae, be somewhat protruded, and a drop of sweet liquid 

 exuded. The larvae possessing this gland are greatly affected by ants 

 of different species, which, in return for the food they obtain from the 

 larvae, act as their most efficient guardians. He says that he has found 

 as many as four species of ants attending one species of larva. Ant- 

 tended larvae are most easily found by looking for the ants. The larvae 

 are usually coloured like the leaves, buds, flowers and seedpods on which 

 they feed, and, for other reasons, are not easily seen ; but the restless 

 red or black ants are very conspicuous. Curetis larvae, which are not 

 attended by ants, have a highly-developed eversible organ on either 

 side of the 8th abdominal segment, apparently for protective purposes ; 

 in other larvae, attended by ants, the organs on the 8th abdominal 

 segment are smaller than in Curetis, and are, one supposes, gradually 

 becoming aborted, probably because, the ants having constituted them- 

 selves their defenders, there is no further use for them for defence, but 

 Edw^ards possibly correctly surmises that in their aborted condition they 

 serve as signals to the ants to examine the 7th abdominal segment 

 for the sweet fluid emitted by the larvae. Doherty has recorded (Journal 

 As. Soc. Be?ig., h\,])t. 2, p. 122) some interesting observations on the same 

 subject ; so also has Mrs. Wylly (Jo urn. Bomb. Nat. Hist. Soc, iii., p. 

 164). Not only do the ants attend the larvae from their very first and 

 smallest stages (some ants were found attending larvae of Rapala 

 schistacea only i in. long), until they are fullgrown, but they often 

 cause the larvae to change to pupae within their nests, in this manner 

 protecting them from harm from the time they emerge as minute 

 caterpillars from the egg, until they assume the pupal stage. Niceville 

 also mentions that Aphnaeus vulcanus is attended by the black ants, 

 Pheidole quadrispinosa and P. cremastog aster, and that Gerydus symethus 

 and Tarucus theophrastus are also attended by ants. Green says, " The 

 larvae of a Cingalese Lycaenid, Aphnaeus lohita, Horsf. (=lazularia, 

 Moore), frequent the nests of Cremastog aster, on Acacia and Grevillea 

 trees, upon the foliage of which they feed. These larvae carry a dorsal 

 honey-gland near the posterior extremity of the body (7th abdominal 

 segment), and are cultivated by the ants on this account. They are 

 herded in special shelters built by the ants, are driven out at night to 

 feed, and brought back to their shelters each morning" (Ent., xxxv., 

 p. 202). 



On this subject, Doherty writes (Journ. As. Soc. Beng., lv., pp. 122- 

 173) : " Dr. Thwaites (in Moore's Lepidoptera of Ceylon) says, ' Nature, 

 however, finds a protection for these helpless Lycaenid larvae, in the 

 instincts of an ant, Formica smaragdina, Fab., which, finding a sub- 

 stance most palatable to it, secreted naturally from a glandular defined 

 spot upon the body of the larvae, takes possession of them as cows, 

 surrounding each separate one, and the leaf on which it feeds, protect- 

 ing them jealously and attacking most fiercely any living thing- 

 intruding upon them.' Besides a remark of Herrich-Schaffer's, quoted 

 in Distant's Rhopalocera malayana, that Gerydus symethus inhabits 

 ants' nests, I have met with no other mention of this singular habit. 

 I have, however, myself observed it in quite a number of Indian 

 Lycaenidae, belonging to several distinct groups, and feeding on the 

 leaves of various trees and herbs. The larvae in question, are all very 

 helpless and inactive grubs, slug-like in shape, tapering at both ends, 



