TiiK Kcc: ix'ri:i{\AL ('iiAX(;i:s. 7 



this period the lar\!i attains its finhryonic juatm-ity hy excee(liii<;ly ra])i(l 

 iii-owth ; the firoiiit of tlie (\i<'^ no longer suffiees for its more extended hody 

 and the head Li,ra<hially moves a little u[)\vards and inwards, the hnver por- 

 tions of the anterior segments slide over the upper portions of the posterior 

 ones and the head occupies the centre of the summit of the structure ; in 

 tliis position the larva is to be found just before emerging. 



Twenty-four hotu's before liateliing, the c<j!;<i; of Euphoeades troilus shows 

 the beginning of this stage; the mandibles of the larva may then be [)lainly 

 seen, hanging separated from each other like ordinary a|)pendages, while 

 the head is just beginning to twist inwards ; sul)se(piently the animal is so 

 coiled in the cg,g; that the middle line of the mandibles (their sei-rated edges 

 rolling over each other like cog-v^dieels ) lies directly over the suture Avliich 

 separates the fourth and fifth segments of the abdomen ; the thoracic seg- 

 ments are so bent as to bring the head against the abdominal rings, the 

 apex of the triangle on the front of the head lying just at or slightly in ad- 

 vance of the sunnnit of the egg, the hairs of the subdorsal abdominal series 

 are directed toward and interlap those of the opposite sides, excepting on 

 the two terminal segments where they are directed posteriorly. All this 

 growth is effected in a single day, during which the larva frequently changes 

 its position by twirling in the shell, a movement probably produced by the 

 aid of the dermal a})pendages of the body ; these, as I shall hereafter show, 

 are peculiar to this stage of the insect's life, being lost, not indeed at the 

 exit from the egg-envelope, but at the first moulting of the larva. 



In Erynnis metea, Limochores taumas, and doubtless all the other Pam- 

 philidi, where the third period is marked by a lateral blotch, the change 

 from the third to the fourth period is indicated by a removal of this patch 

 to the sunnnit of the egg and the appearance within it of a darker smaller 

 spot, — the mandibles. In Erynnis metea this patch occupies the whole 

 of the u})per third of one side of the egg and the upper half of the other 

 side ; the darker spot included in it is transverse and reniform and occurs on 

 that side where the patch is largest, a little above the midtUe of the egg. 

 In Eurymus i)hilodice the change is indicated by the assumption of a plum- 

 beous hue and in Euphydryas phaeton by a growing paleness at the base 

 and of dinginess at the summit of the egg. In nearly all these instances 

 the change occurs within a day of hatching, but in those butterflies which 

 remain a long while in the egg state (excluding of course those which hiber- 

 nate, and which have not been brought at all under consideration) , this 

 period may last for two or three days. In butterflies which I have studied 

 the duration of the egg-state in the summer varies from five to twenty-seven 

 days. 



