vi DEFENCES OF INSECTS 239 



colouring matter of the different, kinds of leaves 

 on the tissues of the insects, the more so from the 

 total absence of green among internal feeding larvae, 

 while the number of arboreal feeders so coloured is 

 legion. This cause of colour certainly suffices in 

 many transparent caterpillars. But in other cases the 

 adjustment arises from something very different, in 

 fact the larvae are affected by the colour of the 

 leaves as surroundings, and not by the leaves as food. 

 Having regard to the experiments on pupae, Mr. 

 Poulton considers it probable that reflected light 

 influences the skin. 



The investigations which he undertook in proof of 

 this were laborious, and very numerous. Leaves 

 were sewn together, so that the larvae were exposed 

 to the upper or the under side alone, although they 

 ate the same leaf in both instances. The ' bloom ' 

 was removed from the under sides of some leaves, 

 while some were left normal. The results pro- 

 claimed the surroundings and not the food to 

 be the influencing agent. Species of Geometrae 

 were the subjects of trial. Larvae were selected 

 which resemble the twigs or bark of their food-plant. 

 Some were surrounded solely by the leaves upon 

 which they fed, and by white or green sur- 

 faces ; in other cases numbers of dark twigs were 

 mixed with the leaves of the food-plant. In the 

 former experiments the majority of the larvae be- 

 came very light brown or light gray in colour, and 

 quite different from the darker larvae of the same 

 kinds which were produced when the latter test was 

 resorted to. Mr. Poulton points out that the change 



