306 Mr. Poulton's notes in 1886 



constructed a perfect oval cocoon. This entirely confirms 

 my conclusions of last year, that the larvae fail to make 

 cocoons in a comparatively large space, because all their 

 silk is exhausted in the scaffolding. 



Thus larva No. (2) fed during a period of about four- 

 teen days, and the other for about seventeen days ; so 

 that the duration of growth is precisely similar to that 

 of the larvae described last year (also extending from 

 fourteen to seventeen days), although the size attained 

 was far less in the latter case. The conclusion that the 

 amount of growth is independent of the duration of the 

 whole feeding period is also shown in the fact that the 

 larger of the two larvae (1886) ceased to feed in the 

 shorter time. It is also interesting in relation to the 

 same conclusion that larva No. (2) killed its host in 

 eleven days, while in 1885 the Dicranura larva lasted 

 through the attack of seven parasites during twelve days. 

 On the subject of the size of these two larvae, Mr. E. A. 

 Fitch informs me that size is so entirely irregular 

 among the imagos of the Ichnemnonidce that it cannot 

 be adopted as any criterion of specific identity. It is 

 exceedingly interesting to find so perfect an adaptation 

 to the necessarily uncertain amount of food obtainable 

 in the larval stage. I look forward with interest to the 

 appearance of the imagos of my larvae, for they have 

 probably reached the upward limit of attainable size, 

 inasmuch as the Dicranura larvae did not seem to be 

 completely eaten when the Paniscus larvae ceased to feed. 



I am now able to add the duration of the pupal period 

 in Paniscus. In my last paper I described six larvae 

 (1885) as pupating May 17th— 25th. On June 15th 

 three imagos emerged, another appeared on the following 

 day, and the remainder a few days later. Thus the 

 pupal period lasts for about four weeks. Each imago 

 directly after emergence ejects short white cylinders of 

 excreta, which from their appearance are probably some 

 product of nitrogenous metabolism. It is thus seen that 

 these imagos appear some weeks before the time at 

 which the larvae of D. vinula have arrived at the stages of 

 growth in which they are liable to the attacks of these 

 parasites. A number of cocoons of J>. vinula were kept 

 under the same conditions as the Paniscus larvae and 

 pupae, with the result that the eggs laid by the imagos 

 of the former hatched at the time when the Paniscus 



