Experiments wpon the colour-relation, ce. 259 
3 pupe were light (3), 2 typical and greenish, faintly coloured; 
1 typical and yellowish, patches very 
slightly developed. 
1, 4, dark (4), typical and faintly pinkish. 
dead . (4), 2typical and pinkish; 1 faint and 1 distinct, 
_ 2 typical and faintly yellowish. 
POTS », light (4), typical and faintly pinkish. 
16 
The results contrast very strongly with those recorded 
by Mr. Poulton as produced by deep red, for the pupex 
were very dark in tint, while these are on the whole very 
light. The explanation is no doubt due to the extreme 
paleness and transparency of the tissue-paper employed 
by Mr. Griffiths, so that the results approximate to 
those which might be expected from white paper, and, 
in fact, the spectrum of the former paper was barely 
distinguishable from that of white light. At first sight 
it may appear as if the pinkish ground tint of many of 
the pupe may have been determined by the pink 
surroundings, but this is extremely improbable, because 
the pink shade of ground colour is always and in all 
grades of pupal colour (except (5) ) by far the commonest, 
and it will be found to be equally predominant when 
other colours were employed. 
B. Another experiment was intended as a test of the 
effects of conflicting colours, but it really belongs here 
for the reasons given above. In this experiment the 
larve were treated exactly in the manner described above 
under Pink until Stage III., when other colours were 
added, as described below. 
Data.—Three larve, which must be described sepa- 
rately. (a), when girdled Green tissue-paper was in- 
serted anteriorly, under about one-fourth of the body, 
the rest of the body remaining on the Pink paper. 
(b), Blue paper was similarly placed anteriorly, and gilt 
paper posteriorly, the colours occupying equal areas. 
(c), to the anterior half Gold paper was added, and 
Black posteriorly, about equally. 
Results.— 
(a) was a dark (3), typical and yellowish. 
(d) ” light (3), ” ” 
(c) ,, dark (4), 4; pinkish. 
These results harmonise completely with those described 
above, no effect being produced by the colours which 
