Mar., '07] ENTOMOLOGICAL NEWS. 75 



peduncle just below the bud. Fig. i shows the usual situation 

 of the egg. 



It is important to the young- larva that these young central 

 buds be chosen rather than the outer ones which bloom first. 

 For the sepals of A. lyrata are caducous ; and if the eggs were 

 placed on the older buds, the sepals would fall before the egg 

 hatched, and the young larva would die of starvation. This 

 happened in one instance in the laboratory. 



Just before the egg hatches, it changes from orange yellow 

 to a dull yellowish brown. The larva, measuring a little 

 over i mm. in length, eats away one side of the shell, creeps 

 out, and usually begins to feed at once, leaving the shell as 

 when it emerged. My observations on this point are at vari- 

 ance with the statement commonly made, that butterfly larvae 

 devour the shells from which they emerge before beginning 

 to feed upon the food plant. 



The duration of the instars varies with different individuals, 

 and the moults occur more rapidly indoors than out, especially 

 if an abundant food supply is kept on hand. The larva 

 watched most closely as regards the time of moulting, hatched 

 on May I3th. The first moult occurred at 9.30 A. M., May 

 i6th; the second on the i8th, the third on the morning of 

 the 2Oth, and the last early on the 22d. The adult larva, 

 28 mm. in length, left its food plant near noon on the 25th, 

 and began to crawl about on the sides of the bell jar which 

 was kept over the vessels containing the food, as if seeking a 

 suitable place for pupation. It continued this for more than 

 four hours, and then became very quiet. At six o'clock it was 

 spinning the silken carpet on which they always pupate, and a 

 half hour later it had begun the "button" to which the posterior 

 end is finally attached. 



It is interesting to note that this mass of silk is nearly 

 always spun while the larva hangs head downward. There 

 seems to be a geotactic stimulus of some kind, for in one case 

 the larva was carefully inverted four times successively after 

 the carpet had been spun, and each time after about ten 

 minutes it turned slowly around without leaving the web al- 

 ready spun, and assumed this inverted position. After turning 



