the colours of certain Lepidoptera. 391 



darker as a whole, although including 6*5 % of the 

 lightest varieties. It will probably be found that larvae 

 suspended from threads at a distance from any back- 

 ground would tend, like the isolated ones on clear glass, 

 to produce intermediate forms. 



The effect of light transmitted through deep green 

 glass was, with a single exception, to produce dark or 

 intermediate pupae. This will be alluded to further on 

 in discussing the very different effect upon V. io of the 

 same light reflected from a white background. 



The few transference experiments need not be extracted 

 from the descriptive table. They quite confirm, although 

 they add nothing to, the results obtained with larger and 

 more carefully conducted experiments in 1886. 



The general result of the whole series of 68 experi- 

 ments conducted, in 1888, upon many hundreds of pupae 

 is to afford abundant confirmation of the earlier work, 

 at the same time extending it in many directions. 



1892. 



The only experiment upon V. urticce in this year was 

 one with conflicting colours. The results of such experi- 

 ments have so important a bearing upon the physiology 

 of the adjustment of pupal colours that I was anxious to 

 repeat them, if possible, in a more searching manner. 



In 1886 I devoted a great deal of time and attention 

 to the subject (Phil. Trans. 1887, pp. 368—392), exposing 

 the larvae during Stage III. to gilt and black surroundings 

 in compartmented tubes, and frames with perforated 

 shelves between the contrasted colours. 



The following questions as to the physiology of the 

 process are answered by the results of these experi- 

 ments : — (a) The possible injiuence of colour upon the 

 larval eyes. Blinding the larva had ifailed to affect the 

 power of adjustment, and this experiment would apply 

 a valuable test to the conclusion that the eyes are of 

 no importance in the matter. If the colour surrounding 

 the anterior part of the larva had no more influence 

 than that surrounding the posterior part, the conclusions 

 from the blinding experiments receive strong con- 

 firmation, (b) The direct photographic effect of light upon 

 the skin. Although the earlier view that the pupal tints 

 are determined in this way after the last ecdysis, has 

 been completely upset by the results of transference ex- 



