330 Mr. 'Poulton sfartlter experiments upon 



A. Dark Surroundings. 



Cylinder with abundant dark twigs 

 intermixed with food. 



II. 



Same as I. 



July 15. — 31 young larvro intro- 

 duced. 



July 24. — Compared on this and 

 following dates. Besting (by day) 

 on the darii tvi'igs were 16 dark, 

 1 green, and 2 intermediate larva?. 

 Resting on leaves and green shoots 

 were 3 dark, 3 green, and 6 inter- 

 mediate larvae. 



Aug. 5. — 3 dark larvas on leaves ; 

 25 dark larvae on dark twigs ; 1 inter- 

 mediate larva on dark twigs ; 2 

 brownish intermediate larvro on 

 dark twigs. 



Aug. 10 and onwards. — The larvre 

 sought pupation without further 

 change of colour. 



July 15. — 30 introduced. 



July 23. — Compared. Nearly all 

 dark brown ; at later dates this ten- 

 dency became more marked, and 

 finally only 2 exceptions remained, 

 1 being bright green, and 1 inter- 

 mediate. 



Aug. 11. — At this date and on- 

 wards the larva? began to seek 

 pupation. 



The extreme susceptibility to dark surroundings is clear 

 from these results. The fact that green surroundings 

 cannot be excluded is no doubt the explanation of the 

 few exceptions. In the converse experiments with green 

 surrounclings everything dark can be excluded, and hence 

 exceptions did not occur. The exceptions undoubtedly 

 show individual differences in the degree of susceptibility 

 to green and brown surroundings respectively, although 

 the final result — 1 marked exception (the single green 

 larva) — out of 61 individuals shows that such differences 

 are of no great numerical importance in determining 

 the colours of this species. The proportions of light and 

 dark larvae found July 24, in I., on dark and green 

 surfaces respectively, seem at first sight to suggest the 

 existence of a tendency to seek an environment with the 

 corresponding colour. On Aug. 5, however, nearly all 

 were on the dark twigs ; and in 1892 very inappropriate 

 situations were often observed (Experiment XXVI.). 



It is also seen that a period of 8 days (July 15 to 23) 

 is sufficient to produce marked effects on the majority 

 of the larva3. 



