Experiments upon the colour-relation, &c. 2.63 
predominance of white light may have assisted in the 
production of such light-coloured pupe (with the excep- 
tion of the first described). 
Before proceeding to a few miscellaneous experiments 
made by Mr. Griffiths, it will be well to tabulate all the 
- results recorded above, and to include for comparison 
the results Mr. Poulton obtained with P. rape, as shown 
in the table on p. 428 of his paper (l.c.). It must be 
remembered, however, that Mr. Poulton’s results were 
confirmed by an almost equal number of experiments 
with P. brassice. The conditions of illumination were 
varied greatly in Mr. Poulton’s experiments, but for a 
complete account of the methods, &c., the paper itself 
must be consulted. Owing to the method adopted by 
Mr. Griffiths, and previously described, the colours 
employed must have been always in shadow. Mr. 
Poulton used a somewhat similar method in the case of 
some of the experiments with green, but so arranged 
that the illumination remained moderately strong. 
Two of Mr. Griffiths’ pupe have been omitted from the 
table, the one under I. A., which pupated on the clear 
area, and pupa (b) under II., because it could not be 
placed in any of the degrees of colour with certainty. 
This table brings out the very great uniformity which 
exists between these two independent sets of experiments, 
and the confirmation is most complete and striking just 
where the results are most marked, viz., when yellow 
surroundings were employed. Mr. Poulton has shown the 
results of his experiments graphically (l.c., p. 431) by 
a curve of which the ordinates represent the average 
amounts of pigment obtained in the several experiments, 
while the abscissa line is formed from the scale of wave- 
lengths. As it was impossible to obtain satisfactory 
spectra from Mr. Griffiths’ pale tints and discontinuous 
tissue-paper, it was thought better not to construct a 
curve to express his results; but the ordinates were 
calculated by the use of Mr. Poulton’s scale, and were 
found to correspond with the lengths of those of his 
scheme in a very remarkable way. But this fact is 
apparent by a simple inspection of the table given 
below, which in reality shows far more than could be 
seen by the graphic method. 
