374 MR, E. B. POULTON ON THE COLOUR-RELATION BETWEEN EXPOSED 
space between the larva and the perforation in the lower disc, but this was largely 
blocked by the larval bristles. The different sizes of the openings through which the 
compartments were illuminated corresponded to the fact that the light which came 
down from the window into the upper compartment was far stronger than that which 
was reflected up into the lower chamber (the glass sheet with the tubes adhering 
being placed a short distance above an ordinary plain deal table). In order to apply 
the tubes, the larvae were induced to suspend themselves from sheets of glass. A 
number of strips of glass were cut of various lengths, and equal, but narrow, widths 
(about 4 centimetres), and these were placed together so as to form a number of 
separate rectangular frames, the angles being secured with gummed paper, a sheet of 
glass being placed as a roof over the top of each. In this way a number of glass boxes 
were obtained, having very low sides (4 centimetres), and very large covers. As soon 
as the larvae in the stock quitted the food-plant they were turned into these boxes, 
and, the low sides offering little impediment, they soon mounted to the roof, and 
prepared for suspension. Some of the boxes were compartmented by another glass 
strip passing across the centre. In other cases the same results were obtained by 
placing a sheet of clear glass over a low cylinder of great diameter. In all cases the 
larvae of the succeeding subdivisions and those of the next series also were kept in 
separate glass boxes, compartments, or cylinders. A certain number of larvae were 
left free to suspend themselves among the tubes which surrounded the other larvae, in 
order to observe the effect of varying proximity to tubes with gilt or black external 
surfaces as compared with the results produced upon the larvae inside the tubes. At 
the same time the results are not very trustworthy, for there is no note as to the time 
at which the free larvae suspended themselves, relatively to that at which the tubes 
were fixed, although in nearly all cases the latter took place at an earlier date. 
In the succeeding subdivisions C indicates the time when the suspended larva was 
covered with a tube, and r that this took place very shortly after suspension began, 
1 being substituted if suspension had begun some hours previously; P indicates 
pupation, and r that the change had just taken place. In the results, “ black below ” 
and “ gold below ” refer to the lower compartments of the tubes, the upper compart¬ 
ments being always of the opposite colour. Each of the subdivisions below (a, (3, &c.) 
corresponds to a single glass sheet with all the pupae suspended from it either free or 
in tubes. 
(a) As soon as the tubes had been made on the morning of September 4, a few 
larvae, which had not been suspended the night before, were covered (a, b, &cc., in the 
Table below). At a later date larvae were covered very soon after the beginning of 
suspension (1, 2, &c, below) The experiment was conducted as follows :— 
