POLYOMMATUS ICARUS. 197 



whole experiment appeared to me to be the patient, persistent manner 

 in which the ant continued its caressings, as if it were perfectly well 

 acquainted with the method to be employed, and absolutely assured of 

 the successful issue of its labours. There was at no time any excited 

 running over the body of the larva by the ant, such as was the case in 

 my experiments with Agriades coridon and A. thetis (bellargus) last year; 

 it remained for nearly an hour on the posterior half of the larva, and, 

 for several minutes after each successful termination to its entreaties, 

 devoted itself to combing and stroking its antennae, as though well aware 

 that the larva required time to recover from the treatment to which it 

 had been subjected. Eayward adds that he has unfortunately not been 

 able to ascertain if any connection exists between the larva of this species 

 and ants before hybernation, and further notes that while his observa- 

 tions of the larvae of P. icarus have, so far, unfortunately been restricted 

 to those reared in confinement, and that he cannot, therefore, state 

 positively that they are attended by ants under natural conditions, 

 there "can, he thinks, be very little doubt that they are, occasionally 

 at least, so attended, as the low-feeding habits of the species must 

 certainly frequently bring them into contact with ants, and the 

 functional development of the gland would appear to show that it is 

 not long allowed to remain unemployed (Ray ward). It may be further 

 noted that Eayward commented (Ent. Rec, xix., p. 219) on the large 

 size and brilliancy of some bred examples he exhibited at a meeting 

 of the South London Entomological Society, in spite of the fact that 

 ants were almost constantly in attendance on the larvae. 



Ontogeny of Larva. — First instar : When newly-hatched not 

 quite 1mm. in length, of dumpy figure, the head very small, the pro- 

 thorax as wide as, and longer than, any of the others, and having a 

 semicircular plate with its rounded edge in front, down the back a 

 row of transverse pits on the front edge of each segment ; the general 

 colour grey, with purplish tinge, the segmental divisions green, the 

 belly yellowish ; the usual tubercular dots black, on grey warts, each 

 furnished w 7 ith a rather long, curved, glistening bristle ; the whole 

 skin besides being finely sprinkled with tiny black dots. The head 

 shining black ; the prothoracic plate rather darker than the ground 

 colour, the spiracles black. It moults w r hen about l|mm. long. 

 Second instar : Entirely pale olive-drab, the usual tubercular dots 

 paler than before, only outlined in black ; the bristles increased in 

 number, some being short and pale, others longer and darker. It 

 moults a second time when 2-|mm. long, and is at this time dull pale 

 green in colour, with a paler double dorsal line, and paler subspiracular 

 line. Third instar: The dorsum pale greyish-green, the lines on the 

 dorsal ridges rather paler ; the spiracles round, pale brownish, incon- 

 spicuous ; the skin set w r ith a number of small w 7 arts ; along the 

 dorsal and subspiracular ridges are rows of long tubercular bristles, 

 and on the sides some shorter ones. Fourth instar : When 9mm. 

 long the larva is dull, full green in colour, with the subspiracular 

 ridge yellowish, the belly and legs yellowish-green ; the head shining- 

 black ; the spiracles green, indistinct, the bristles very pale browmish. 

 Final instar: When full-grown, 12mm. -13mm. in length, nearly 

 4mm. wide, the head small, and under the prothorax, w 7 hich is 

 flattened ; the other segments slope from the subspiracular ridge up 

 to the double dorsal ridge, the last three also sloping dowm to the 



