FAMILY HABITS IN BUTTERFLY LARVAE THE LYCJENIDS. 75 



refused to feed on, and commenced to hybernate when quite small 

 instead of producing a second brood. On the other hand, he notes 

 that he has bred many Agriades bellargus, and never knew summer- 

 feeding larvae, from eggs laid by $ s of the early emergence, fail to 

 feed on as " forwards," and produce autumnal imagines. 



The Palaearctic and Nearctic Plebeiids, as a rule, then, have winter- 

 feeding larva?, i.e., larvae that commence feeding in August and 

 September, and continue for some time, then rest, and begin again in 

 the early spring. Some species do not seem to rest very completely 

 in the winter, for Buckler says that small larvae of A. bellargus are 

 to be found in October on the underside of Hippocrepis comosa, that 

 they feed slowly through the winter, and make much larger blotches 

 on the undersurface of the leaves by early February, whilst in March 

 the leaves are eaten from the edge, and often demolished entirely. 

 Even the Plebeiid species that hybernate in the egg-stage disclose their 

 larvae so early — end of February or early March — that they commence 

 under almost winter conditions, whilst all feed on through the spring- 

 months, until at least May, although the absolutely single-brooded 

 species continue until June, e.g., Plebeius aegon, Agriades corydon, etc., 

 but, in addition, many species produce summer-feeding larvae that go 

 through their larval life quickly in July- August, and produce 

 imagines in autumn, e.g., Aricia astrarche, Agriades bellargus, Poly- 

 ommatus icarus, Cyaniris semiargus, etc. The Nomiadid species — N. me- 

 lanops, N. cyllarus, etc. — have purely summer-feeding larvae. One 

 suspects that Polyommatus donzelii has a purely spring-feeding larva 

 (it hybernates as larva inside egg), but a note by McDunnough (Ent. 

 Rec, xviii., p. 313) states how two, out of a large number of ova found 

 at Pontresina at about 6200ft., disclosed their larvae during very hot 

 weather towards the end of August, 1905, one of which died, whilst 

 the other fed up well to the 3rd instar (the hybernating stage for 

 most Plebeiid larvae), and then prepared to hybernate under a dead 

 leaf ; the specimen was not observed further. 



The most remarkable feature, however, in the habits of the larvae 

 of the Lycaenids, is their association with ants. It is almost safe to 

 say that it is only our ignorance that prevents us from asserting that 

 the larvae of nearly every Lycaenid species is, more or less, so associated. 

 How ancient a habit this is one dare hardly hazard a guess, for the 

 habit is characteristic of the " blue" larvae in all parts of the world — 

 Australia, Asia, Africa, Europe and America — and, so similar in all 

 cases, that one suspects for it a vast antiquity. The connection 

 between the ants and the Lycaenid larvae is generally supposed to be 

 two-fold and advantageous to both, the ants obtaining from the larvae 

 a fluid they love, whilst their presence is supposed to keep away 

 enemies that might otherwise attack the larvae. The general character 

 of the structures, that appear to be intimately connected with this 

 symbiosis, seems to be practically identical in all Lycaenid larvae, (1) a 

 pair of retractile organs or eversible caruncles, one behind each of the 

 spiracles of the 8th abdominal segments, which are, when protruded, 

 surmounted by a series of spiny or feathery branches, and which are 

 supposed to attract the ants ; (2) the other, an eversible honey-gland 

 with somewhat thickened lips, on the dorsum of the 7th abdominal 

 segment, from which a drop of sweet fluid is emitted on the solicita- 

 tion of the ants. The fact that ants consorted with Lycaenid larvae, 

 and were symbiotic therewith, has been known for at least 130 years ; 



