THE ASSOCIATION OF ANTS WITH BUTTERFLY LARV-E. 35 



inhabitants of a Zisyphus tree may be observed. The ants are busy 

 all day long running along the branches and leaves in search of the 

 larvae, and when they meet one fullgrown and ready to pupate, they 

 drive the caterpillar down the stem of the tree towards their nest. As 

 a rule, the larvae are docile and easily led, and, having got him into 

 his proper place he undergoes transformation into a pupa. If one 

 gently scrapes away the loose earth piled up at the base of the tree, 

 one will see some hundreds of larvae and pupae in all stages of develop- 

 ment arranged in a broad even band all round the trunk and lightly 

 covered with earth. The ants object to their being uncovered, and 

 will immediately set to work to recover them, and, if one persists, they 



will remove all the chrysalids and bury them lower down A 



larva of a species of Catopsilia (one of the Pierinae) given to the ants 

 as an experiment, was immediately set upon and torn to pieces in a 

 second by the ants. A larva of T. theophrastus, taken from a tree, 

 was introduced into the pathway of another company of the same 

 species of ants, which lived on our verandah, but kept no "farm." 

 It was odd to see the ants come tumbling over headlong to fight the 

 intruder, and the sudden way they cooled down on investigation of 

 the foe. None attemped to harm him, and he was politely escorted 

 across their boundary, the ants running alongside, and feeling him 

 all over with their antennae. This must have been instinct as they 

 could have had no former knowledge of him as a " milk-giver." The 

 ants distinguish between dead and living pupae, the dead chrysalids 

 being carefully removed and thrown away outside." 



Distant mentions that the larvae of the genus Ambly podia are attended 

 by Formica smaragdina, Fab. Bethune-Baker says that the larvae of almost 

 all the species of the Australian genus Ogyris are probably attended by ants, 

 some apparently by different species in different neighbourhoods, whilst 

 Dodd records that 0. zosine is attended by two species of Camponotus, and 

 also by a small black ant (Trans. Ent. Soc. Lond., 1905, pp. 269-270). 

 Of 0. zosine, Bethune-Baker writes (op. cit., p. 279) that the larvae feed on 

 Loranthus linophyllus and L. celastroides, and hide in the cracks of the 

 bark of the host tree .... coming out at dusk and feeding at 

 night, at which time the ants associated with them are likewise said 

 to be on the move. The species that Dodd has found them with most 

 commonly is JEcophylla virescens, but several other species also 

 associate with them. They evidently protect the larvae, and have been 

 observed to milk them ; in one instance, an ant was observed to 

 approach a larva and wave its antennae over its terminal segments, 

 and then to lightly touch it with its foreleg, when a small globule of 

 liquid was emitted from a small, retractible, nipple-like organ on the 

 dorsum, which was at once sucked up by the ant. Lyell and Fricot 

 (Vic. Nat., xxi., pp. 166-167) have confirmed these observations, and 

 state that, in order to test the action of the attendant ants, one or two 

 larvae were placed a couple of feet or so away from a tree ; they 

 were, however, soon discovered, and dragged carefully back to the 

 tree by the ants, at a pace much more rapid than their own rate ; 

 pupae were likewise carried back to the tree. It is recorded also 

 that " ants are always found with the larvae of 0. abrota. 

 Anderson further observes (Victorian Butterflies, pp. 101-102) that 

 the larvae of the genus Ogyris are greatly attractive to ants, which 

 tend them with great care, never leaving them. Raynor notices (in 



