the colours of certain Lepidoptera. 459 



the true animal pigments, leading to various shades of 

 brown, grey, &c. ; (/>) the change to a green colour 

 modified from plant pigments in the food. When such 

 a change of colour is possible, the true pigments are 

 always superficial to the green, and cannot be retained 

 without concealing the latter, the degree of concealment 

 depeuding on the amount and distribution of pigment. 

 Thus in A. hetalaria the true pigments are chiefly placed 

 in the epidermic cells, the green in the subjacent fat, 

 while in many others the former are in the superficial 

 layer of the cuticle, the latter in the blood or sometimes 

 in the lower layers of the cuticle. But the appearance 

 of the green is not merely the removal of a screen, 

 although this must occur ; in some cases, at any rate, 

 it also means the formation of the green colouring 

 matter itself. 



In discussing the effects of light it will be important, 

 therefore, to discriminate between {a) modifications of 

 true pigment ; {h) its disappearance, accompanied by a 

 change to green. 



I propose to tabulate all the coloured backgrounds 

 made use of in these experiments, and briefly to compare 

 their effects on the species subjected to them in 1886 and 

 in subsequent years. We shall thus be able to form a 

 sound conclusion as to the constituents of a mixed 

 reflected light (like that from leaves) which effect the 

 change, and as to the existence of any common sus- 

 ceptibility on the part of such Lepidoptera to light from 

 a particular part of the spectrum. 



I wish to express my warm thanks to my friend Sir 

 John Conroy for his great kindness in helping me to 

 make an accurate statement of the quality of the light 

 reflected from backgrounds and transmitted through 

 screens. His well-known researches in this region of 

 physics rendered his kind assistance invaluable. 



The method we employed with the backgrounds was 

 as follows : — The spectrum of lime-light, obtained by the 

 use of a bisulphide of carbon prism, was projected on a 

 white paper screen. The coloured backgrounds were 

 then held so as partly to cover the spectrum, when the 

 rays absorbed and reflected could be determined by 

 comparing the covered with the uncovered part. In 

 many cases, two backgrounds were placed in the spec- 

 trum together so that they could be accurately compared. 



2 l2 



