LIFE-HISTORY OF LIPHYRA BRASSOLIS 155 



nearly two weeks ; they also became sickly, and had to be 

 returned to nests on the trees. However, they deposited frass, 

 showing that they had been feeding ; this is small for such bulky 

 creatures, is grey or greyish white, and is greasy-looking. So, 

 though I cannot positively declare that they exist upon the ant 

 larvsB, I am quite satisfied myself that they do. As five or even 

 six caterpillars may be found in one nest, the quantity of food 

 required by such large creatures must be considerable, though 

 they are very slow growing ; but the masses of ant larvae could 

 be drawn upon without making any great reduction in same. I 

 cannot discern the mandibles of the caterpillar, and incline to 

 the belief that when it comes to be thoroughly examined, the 

 mouth-parts will prove to be of peculiar construction. 



As many readers may be sceptical as to this insect subsisting 

 upon ant larvae, I may mention that at present I have young 

 caterpillars of a moth (female, 20-25 mm.) feeding upon the 

 larv^ of another species of ant which lives in the ground. When 

 given the fresh grubs they soon take a lively interest in them, 

 and affix themselves thereto, and appear to suck their victims. 

 None of the grubs are eaten, but they are considerably smaller 

 and much shrivelled when finished with. The moth is even 

 more greasy than our butterfly. None of the grubs given to 

 L. brassolis were eaten, but some had the same shrivelled 

 appearance as those given to the moth larvae. 



Moulting. — When the caterpillars are about to cast their skins, 

 they spin a webbed footing, nearly their whole length, to which they 

 affix themselves rather firmly ; but in many instances the ants, 

 utilising the tveh of their grubs, secure them still more firmly to 

 their position, the web reaching half-way to, or even up to, the 

 rim. Whether this ant web is added with friendly intent, or 

 with a view to fasten down a larva to get rid of him if possible, 

 is a puzzling matter ; still I have not met with any dead larvae 

 so fastened. Pup^ are occasionally treated in a similar manner, 

 without the escape of the butterflies being prevented. The 

 larger larvae require three to four days after taking up their 

 moulting position to crawl out of their old skins, which become 

 a little darker, and finally split downwards under the front edge, 

 and right and left along the rim. The old skin retains its shape 

 above, and being relieved of its occupant, regains its former 

 colour, so that, viewed casually, it is difficult to believe that it 

 is an empty shell, as it resembles a real larva so closely. 



When the larva is about half grown, the spots so plainly 

 discernible upon young examples become much smaller, but can 

 always be detected, even upon the pupae. I have only observed 

 the moulting process in several specimens about one-third grown, 

 and in a fair number which had attained almost full size. 



Change to the puija. — The larval skin is not cast off, but 

 changes in shape, and becomes a comparatively strong and outer 



n2 



