3B8 Mr. Voiilion' a fa rtJicr experiments upon 



larvae appeared in a few days. The majority of these 

 were at once placed in a cylinder, and fed upon the leaves 

 and green shoots of Popuhis Jiir/ra, being thus kept in 

 green surroundings, although just before they were re- 

 arranged many of the leaves had become withered and 

 brown. The susceptibility of larvae during these early 

 stages, if any exists, has been shown not to interfere 

 with such experiments. These larvae formed the stock 

 from which, when they were rather older, the majority 

 of the experiments were supplied. They will be alluded 

 to below as " the first stock." 



As soon as the larvae began to hatch, a mass of the 

 eggs was separated, and placed (June 29th) in green 

 surroundings in complete darkness until 11 p.m., when 

 it was exposed to the light of a paraffin lamp until 

 9.10 a.m. the following morning, when it was again 

 placed in darkness. Under these conditions the larvae 

 hatched, and they constitute "the second stock," from 

 which several experiments were supplied. This alterna- 

 tion of darkness and lamp-light was continued in some 

 of the experiments until the evening of August 2nd. 

 The changes were made every day, and the fixed times 

 were never departed from by so much as an hour. 



All the larvae which hatched from the eggs were made 

 use of in the experiments, except 30, which were sent to 

 Mr. Bateson, and were experimented upon by him with 

 results pubhshed in this volume (p. 213), and 80, which 

 were placed on a tree (Populus nigra) in muslin bags. 

 These last were intended for experiments, which, how- 

 ever, I was unable to undertake. 



The sizes of the glass vessels in which the larva) were 

 kept are given, because the amount of crowding is shown 

 to exercise a considerable influence on the colour. 



When measurements are stated, it must be understood 

 that they were taken when the larvae were at rest in 

 the rigidly straight position which is characteristic of 

 Geometrce. 



The experiments are so numerous that it has been 

 necessary to classify them, and treat the various classes 

 separately. The following table indicates the arrange- 

 ment pursued, and serves as a guide to any particular 

 experiment : — 



