184 LEPIDOPTERA INDICA. 



interesting description of liis and Mrs. Wylly's observations on this subject in the 

 Journal of the Bombay Natural History Society, vol. iii. p. 164 (1888). The larvas of 

 some species are carnivorous. In a note by Thwaites, he says it is difficult to realise 

 that the larvae of some of these lovely Lycsenidse, such as Amblypodia, etc., are car- 

 nivorous or even cannibal in their habits, and do not hesitate to eat their own brethren 

 of the same brood when any of the latter are commencing their change into the 

 inactive chrysalis state, with their consequent inability to protect themselves from 

 their voracious kindred, who devour them with avidity, de Niceville says (p. 55) that 

 Green (who is a very careful observer) states that the larvae of Spalgis epius are 

 carnivorous; and James L. Courtice, in Ent. Mo. Mag. 1865, p. 45, gives an account of 

 the cannibalism of Zephyrus quercus ; he says, " On the 27th May, one had changed to 

 chrysalis in a corner of the shade, and I observed a larva attached to the tail end of it, 

 and evidently very busily engaged. I disturbed him, and found that he had not only 

 eaten off the end of the chrysalis, but had cleaned the contents of it right out ; and 

 what still more astonished me was that, after being disturbed, he returned resolutely 

 to the attack, and finished up the greater part of the pupa shell." 



With reference to the pupae, de Nice\'ille says they are usually attached by the 

 cremaster which forms the terminal portion of the pupa, and is furnished with minute 

 hooks for attachment to the pad of silk previously spun for that purpose by the larva, 

 and by a silken girth round the middle of the body to whatever surface the larvae 

 choose on which to perform their transformation ; in some forms, however, the pupae 

 are freely suspended by the tail, and Trimen notes the saure, remarking that the 

 pupae of some Lycsenidae are attached by the tail only in a fixed horizontal or slightly 

 inclined position. Some forms of pupae are hidden in the ground, Trimen says, but 

 de Niceville remarks that this is certainly true in the case of some Indian species, the 

 larvae of which are attended by ants, as the latter drive the larvae into their nest, when 

 they turn to pupae in the usual way, being attached to the trunk of the tree by a tail 

 or a girdle, the ants having constructed a temporary nest around the base of the tree. 



The butterflies of this family are often found in open meadow or grass land, but 

 by far the greater number of forms frequent trees and bushes, especially open paths 

 and the edges of forests ; the males are sometimes to be seen in great numbers on 

 damp sandy ground and the sides of streams, sucking up the moisture, and they have 

 the habit of rubbing the hindwings one over the other, when first settling, the motion 

 being apparently more or less rotary. Trimen notes on the subject: "This curious 

 habit is practised by every member of the family that I have watched when settled, 

 and it seems not improbable that the movement may serve to accentuate their orna- 

 ments, either in rivalry or menace." Scudder thinks that the action may cause a 

 stridulation which, though inaudible to our ears, may be heard by insects ; most of 

 the Lycsenidae close their wings when at rest, as do other butterflies, but this is not 



