98 Mr. E. B. Poulton. On a special Colour-Relation [Feb. 10, 



that the pupee are on the whole darker in the latter circumstances, 

 although dark under both conditions. 



In another experiment some larvee were suspended in a strong direct 

 light without any coloured background sufficiently near to affect them, 

 and as far as the experiment went it indicated that there was an 

 influence in the direction of the lighter varieties, but in this case the 

 numbers employed were too small to be convincing. 



5. The Time during which the Larvee are Sensitive. — The whole 

 period preparatory to pupation, intervening between the cessation of 

 feeding and pupation itself, may be divided into three stages : (i) in 

 which the larva descends from the food-plant and wanders about in 

 search of some (generally mineral) surface upon which to pupate ; 

 (ii) in which it rests motionless, usually in a curved position, upon 

 the surface selected; (iii) in which it hangs head downwards sus- 

 pended by its posterior claspers from a boss of silk spun at the close 

 of the last stage. The duration of Stage (i) depends upon the 

 varying proximity of suitable surfaces, and it was always greatly 

 curtailed in confinement, because such surfaces were close at hand. 

 If the larva is sensitive during this stage, the influences cannot 

 generally contribute towards the result, because the larva is wandering 

 over surfaces of various colours. It is also very improbable that the 

 larva can be sensitive after the first few hours, or at any rate the first 

 half of Stage (iii), because rapid changes are taking place under the 

 larval skin, and it is even likely that processes are already on their 

 way towards completion which will result in the formation of pig- 

 ment or other substances, which will many hours later deepen into 

 the effective causes of pupal colour. The length of Stage (iii) did not 

 vary very much in different larvee, and in the shortest case observed 

 the length was about 14 hours, while 20 hours was an unusually long 

 period, but the majority of larvee passed about 17 or 18 hours in this 

 stage. Stage (ii) was more variable, but about 15 hours was a common 

 length, while 36 hours is a fair estimate of the length of the whole 

 preparatory period. In the majority of cases a larva is probably 

 sensitive to the colour of surrounding surfaces for about 20 hours 

 preceding the last 12 hours of the whole preparatory period. Thus 

 the length is amply sufficient to include many hours of daylight during 

 which the surrounding surfaces are illuminated. If a larva be dis- 

 turbed when Stage (ii) is far advanced the whole period begins again, 

 and all the three stages are again passed through, but they are 

 all abbreviated, including Stage (iii), which had not previously com- 

 menced. Many experiments indicated that darkness may increase 

 the length of the stages ; but my observations were not specially 

 directed towards the settlement of this question, which only occurred 

 to me when the notes were tabulated. Therefore I propose to 

 specially investigate this point in the next season. Such prolongation, 



