344 MR. E. B. POULTON ON THE COLOUR-RELATION BETWEEN EXPOSED 
investigation, but (probably on account of disturbance) it again wandered for a short 
time, passing through another Stage I. of about 4\ hours, and then again an 
exceedingly short Stage II. of only a little over 2k hours, the final stage being, perhaps, a 
little less than normal. These facts throw further light on the stages of certain larvae 
which appear to be most abnormally short, and seem at first sight to show that it is 
impossible to obtain an average duration for the stages which would be of any 
practical value for the investigation of any particular larva. Such a conclusion might 
be arrived at by comparing the exceedingly divergent estimates of Stages I. and II. in 
Divisions B. and C. of this series, and also by comparing the estimated lengths of the 
total periods in these two divisions respectively. But the above-described larva 
shows us how after disturbance the whole preparatory period may begin again and all 
its stages may be passed through, but that under these circumstances the stages, 
and especially the two initial ones, are considerably abbreviated. It has been 
remarked above that the larva may even feed again after disturbance and before the 
recommencement of Stage I., so that the resemblance of the abbreviated period to one 
of normal length may be very close. And there is independent evidence for this 
explanation of the abnormal shortness of the stages and periods in Division C., for in 
the introductory sentences of Division B. it is shown that six larvae were already in 
Stage II. on the morning of the 23rd, and of these only two were made use of for 
Division B. It is certain that many others had also entered this stage when the larvae 
were shorn, and the experiments described in Division C. began in the evening of the 
same day, although disturbance had led the larvae to quit the roof and in many cases 
to feed again both before and after they had been placed in the globes. Further 
data will confirm this explanation of the apparent abnormality in Division C. 
Before proceeding to the next series, it will be advisable to recapitulate the results 
of the experiments upon the shorn larvae, and then the subject need not be further 
alluded to. 
Division. 
Surroundings. 
Dc 
grees of colour in pupae 
larvae. 
of shorn 
Degrees of colour in 
larvae 
pupce of normal 
1 
2 
Dark 
3 
3 
Light 
3 
4 
5 
1 
2 
Dark 
3 
3 
Light 
3 
4 5 j 
A. 
Gilt box .... 
i 
2 
1 
i 
3 .. j 
c. { 
White “opal” 1 
globes J 
•• 
2 
3 
1 
3 
i 
2 
! 
B. { 
“ Opal ” covered 1 
cylinders J 
•• 
i 
1 
1 
2 
.. .. 
Totals. 
•• 
V 
i 
2 
5 
4 
1 
= 13 
3 
4 
5 .. =12 
The above Table shows how completely the suggestion was negatived by the 
experiments ; it also indicates that the gilt surroundings were much more powerful 
