458 Mr. Poulton's farther experiments upon 



It has already baen stated that there was not saffisienfc 

 evidence that the colours of the larvae were modified by 

 the conditions of the experiments. 



The colour of silk was variable, being sometimes white 

 and sometimes brown, but there was no evidence for the 

 existence of any power of adjustment. Adventitious 

 particles were so freely used as to render the cocoon very 

 hard to detect. In soft peat the larvae made their 

 cocoons at some considerable depth, but they also freely 

 spun among loose blocks of chalk on the surface. 



The imagines, when carefully compared on a white 

 background, showed not the least tendency towards 

 colour- variation in accordance with the environments 

 which had been employed. Indeed they were, for this 

 species, remarkably uniform, being light grey forms, 

 but none of them extremely white, like the var. pnllata. 



It is clear that there is no susceptibility during the 

 period over which the experiments extended. Either the 

 period was not long enough, or the species is not sensi- 

 tive in this way. If the latter be the true explanition, 

 it is probable that the local races are to be accounted 

 for by natural selection, the lighter varieties being more 

 conspicuous, and therefore exterminated, in peaty dis- 

 tricts, and the darker ones on chalk. Before seriously 

 considering this suggestion, I should much like to repeat 

 the experiments, keeping up the conditions for nearly 

 the whole life of the larvaG. In the EphijricUe we meet 

 with pupae the colours of which are determined by those 

 of the larvae. It is conceivable that, in this case, the 

 colours of the larvae may be modified by environment 

 acting in the usual way during the early stages, and 

 that the imaginal colours may follow those of the larva). 

 I should also very much like to know the result of 

 exposing the pupae to different temperatures, as in Mr. 

 Merrifield's most interesting researches. 



F. Conclusions. 



1. The light ivhich effects the chief colour -changes in 

 larvce and pupce. — Great interest attaches to the attempt 

 to define the light rays which constitute the stimulus 

 leading to the colour-changes. Of these we must dis- 

 tinguish two main kinds : (a) changes in the colour of 



