LEPIDOPTEROUS PUP^ AND THEIR SURROUNDING SURFACES. 
431 
The results in the case of P. brcissicce are probably more trustworthy, because the 
numbers of pupae employed in the various experiments were more uniformly large. 
The similarity of the results in the two species is very striking, and it would have 
been even greater but for the fact that the palest varieties of P. brassicce retain more 
pigment than those of P. ra/pce. Looking at the colours which retard the formation 
of pigment when they preponderate in the incident light, we see that they contain 
certain rays in common which are probably highly efficient in this respect, i. e ., the 
rays from 57 to 59 or 60. The immense difference between the action of red and 
orange corresponds to the fact that these active rays are present in the latter, and 
they are also present in the highly efficient yellow and bright green—that is, in all 
the colours which retard the formation of pigment, except white, which, of course, 
contains these rays in addition to the others, although this is not necessarily the 
case, and it would be extremely interesting to experiment with whites from which 
this part of the spectrum is absent. Concerning such an experiment, it may be 
argued that this relation of colour to pigment formation is essentially protective, and 
is, therefore, concerned with the visual perception of insect-eating animals, and 
especially the Vertebrata; and if these latter cannot with the unassisted eye dis¬ 
tinguish between a pure and an impure colour, or between a white which contains 
all the colours of the spectrum and one which contains only some of them, it would 
seem that the pupa would lose immensely if it were influenced by the one and not by 
the other, or in different directions by the two. It is, however, clear that we must 
only expect perfect parallelism between the sensitiveness of such widely separated 
animals as far as the stimulus is provided by colours which form the natural environ¬ 
ment of the pupse. 
