340 MR. E. B. POULTON OX THE COLOUR-RELATION BETWEEN EXPOSED 
employed, together with their results, are insufficient evidence from which to conclude 
that the larvae are not sensitive during this stage, although it appears almost certain 
that the earlier Stage II. is the time of chief susceptibility to surrounding influences. 
It will be shown that the larva can hardly be susceptible after the first part of 
Stage III., and it is very likely that in two out of three instances the most important 
and earliest hours had already elapsed when the larvae were transferred ; this suggestion 
cannot, however, explain the case of No. 1, for this larva must have been shifted very 
soon after the beginning of suspension. 
C. Another division of larvae was taken from the “ opal’’-roofed cylinders and made 
use of in order to test whether the spines contain the terminal organ which receives 
impressions from coloured surroundings. Nine larvae wmre carefully shorn of their 
bristles on the evening of August 23, and two of them were left in one of the above- 
described “ opal ’’-covered cylinders as they had quitted the food and were resting on 
its roof, while the other seven, being rather less advanced, were placed one in each of 
seven “opal” globes similar to those previously described, except that a glazed white- 
paper roof was, in nearly all instances, fixed to the smaller opening, while the edge of 
the larger opening of the globe rested on a floor of ordinary white paper, upon which 
never more than two leaves w T ere placed at one time, renewed when necessary, and 
removed when it was obvious that the larva had ceased to feed. Seven exactly 
similar, but normal, larvae were placed in seven other globes, all conditions being 
identical, except the following unimportant differences, which were compensated as far 
as possible :—The globe numbered 1 below was lower and smaller than the others; 
that numbered 2 was turned with the larger end uppermost and covered by the roof; 
5 and 13 had been broken and were mended with white glazed paper, so that there 
was more surface of paper in 5 than in the others, the loose piece in 13 being merely 
retained in its place by paper glued on to the outside of the globe. The 14 globes 
were arranged on two shelves, one above the other, in a strong north light close to a 
large window, and the two series of larvse were placed in alternate globes : the shorn 
larvae occupying globes with the even numbers 2-14, the others occupying those with 
the odd numbers 1-13 ; and Globes 1-7 occupied the upper shelf, 8-14 being below. 
By this arrangement all possible errors due to differences of illumination were com¬ 
pensated, for as soon as a larva had ascended the side or had suspended itself the 
globe was always turned round, so that the larva was next to the window. Such 
an experiment was certain to yield useful results, apart from the main question of a 
possible terminal organ in the bristles ; for, by noting the results of frequent exami¬ 
nation, it was possible to gain further knowledge of the three stages of the preparatory 
period ; and, as in the larvse of Division B., the action of white surroundings upon the 
larvse was tested in the most satisfactory way by the elimination of the sources of 
error present in the otherwise similar experiments of the preceding series. The results 
of examination are expressed in a tabular form below. The larvse numbered 4 and 5 
died, and are not further alluded to. 
