386 



Mr. Voiilton's further' experiments upon 



be made out, for the percentages of /? are either practi- 

 cally the same as either ^ or £, or intermediate between 

 them. The wide difference between the (2)s in the 

 averages of these large numbers justifies the conclusion 

 that darkness tends rather less towards the appearance 

 of dark pupae, than dark surroundings in strong light. 



So far as it is possible to judge from the small 

 numbers, gilt surfaces in complete darkness (y) produce, 

 as we should expect, the same effect as black surfaces 

 subjected to the same condition. 



The larger numbers attached to a darkish surface of 

 zinc (0 tend in the same direction as those exposed to 

 black surfaces, but are, as might be anticipated, stronger 

 in the intermediate forms, and weaker in the darkest. 



The influence of various bbight metallic surfaces. 

 The following is a summary of all experiments of this 

 kind in 1887 and 1888. In those of the latter year the 

 crowded are carefully separated from the uncrowded 

 pupae, and those which pupated on the food-plant or 

 floor are excluded. 



{See Table, pages 387, 388.) 



The percentages are now placed under one another to 

 facilitate comparison, the results obtained by "gilt" 

 paper in 1886 being also added. 



Degrees of Colour. 



Embossed Dutch metal, not crowded 

 Embossed Dutch metal, crowded 



Dutch leaf, 1886 



Dutch leaf, 1887 



Polished Dutch metal, not crowded 

 Polished Dutch metal, crowded 



Silver paper (compartment), not crowded 

 Silver paper (compartment), crowded . . 



Silver-paper (cylinder), not crowded 

 Silver-paper (cylinder), crowded 



Tin-plate, 1887 



Tin-paper, not crowded 



Tin-paper, crowded 



Numbers 

 of Pupae. 



35 

 46 



67 



10 



84 

 58 



24 

 10 



3 

 44 



12 



11 



57 



