196 BRITISH BUTTERFLIES. 



searching for cases of Coleophora discordella during the winter. 

 Newman erroneously observes that the species hybernates as egg. 



Symbiosis of larvae of Polyommatus icarus with ants. — In 

 August, 1906, larva? of P. icarus hatched from eggs ( $ captured at 

 Folkestone), and, after growing satisfactorily on Lotus corniculatus for 

 three or four weeks, laid up for hyber nation about September 21st. 

 They were wintered on a growing plant of Lotus corniculatus, and 

 during the second week of March were placed in a warm room on 

 Trifolium repens, one larva moulting on March 26th, the honey-gland 

 on the dorsum of the 7th, and the evaginable tubes on the 8th, 

 abdominal segments, both being well-developed. A few days. later the 

 two most advanced larva? were experimented with, and an example of 

 Lasius flavus was introduced to one of the larva. The ant appeared some- 

 what sluggish, but after a few moments walked slowly over the back of 

 the larva and proceeded to clean itself, and particularly its antenna?, very 

 thoroughly, the operation lasting some four or five minutes, the larva 

 during this time remaining perfectly quiescent, the honey-glands and 

 evaginable tubes showing no sign of movement. After carefully cleaning 

 its antenna?, etc., the ant went straight to the honey-gland, and, in the 

 most methodical and business-like manner, began to caress it with the 

 antenna?, evidently well aware of the delectable drink waiting to be 

 yielded for the asking, yet the result for several minutes longer 

 continued disappointing, for, although the gland showed some signs of 

 contraction, followed by distention, there w r as no indication of the 

 protrusion of the inner membrane, which always precedes the flow in 

 Af/riades coridon and A, thetis. Still the ant persisted in its coaxings 

 and caresses, and at length the evaginable tubes on the 8th abdominal 

 segment showed signs of activity, first one, then the other, and 

 presently both together, being evaginated and withdrawn, the time 

 occupied by the double movement of protrusion and retraction being 

 possibly less than one second. The ant took little notice of the 

 activity of these tubes, but continued its entreaties at the gland, and 

 at last with the desired result, for the greenish inner membrane was 

 slowly protruded, and a bead of viscous-looking fluid was extruded, 

 and was quickly and most eagerly absorbed by the ant. The time 

 occupied by the protrusion of the gland and emission of the fluid, 

 was perhaps three seconds, and about twice that length of time was 

 taken by the ant in drinking the bead of liquid. Four times, in the 

 hour covered by this experiment, did the ant succeed in inducing the 

 larva to yield the fluid, but only after the most persistent and con- 

 tinuous coaxing was success attained, and I do not think the larva 

 yielded voluntarily to the ant's solicitations. The evaginable tubes 

 were most active just before the exclusion of the liquid, and, at the 

 times of their evagination, there was a corresponding pulsation of the 

 lips of the gland ; it appeared to me that the action of the tubes was 

 intended to cause the ant to desist from its excitation of the gland, and 

 that, having failed in inducing it to refrain for more than a moment 

 from its labours, the protrusion of the membrane and emission of the 

 fluid were spasmodic and involuntary. The tubes ^ are white, and 

 have, at their apex, a number of filaments or tentacles, but the evagin- 

 ation and retraction were so quickly effected that I was unable to 

 ascertain their number. The second larva subsequently gave a 

 successful demonstration also, and the most interesting feature of the 



