1893. COLOUR CHANGESVIN INSECTS. 289 
light reflected from them. The suppression of the superficial dark 
pigment thus leads to the green colour of the caterpillar. 
From these experiments, the fact that geometrid larve have a 
very considerable power of colour-adaptation to their environment 
appears to be beyond dispute. There seems no reason to doubt that 
this power is of considerable protective value to the insects, especially 
when we consider that the larvae in which it is developed are all of 
a form closely resembling the twigs of their food-plants, but that 
they need some power of colour-adaptation to render the resemblance 
perfect. Poulton lays stress on the fact that all the numerous 
varieties of A. betulavia produced by the experiments would be in 
correspondence: with some possible natural environment. The ex- 
periment of moving caterpillars from one set of conditions to another 
showed that the early larval stages only are susceptible to any great 
extent; after the last moult but one, little or no change was produced. 
Hence, the most sensitive caterpillar has not that power of changing 
and rechanging its hues which is possessed by many fishes and 
reptiles; but then, in natural conditions, the environment of cater- 
pillars will very rarely be changed. 
In ‘**The Colours of Animals,’ Poulton gave several instances 
of variation in the colour of the cocoons spun by caterpillars before 
pupation, which appeared to correspond to changes in the environ- 
ment. The cocoons of Saturnia pavonia (carpim), Eniogaster lanestris, 
and Halias prasinana were found to be brown among leaves, and white in 
paper bags. Bateson (3,4) has, however, now shown that the white 
cocoons of the two former species are not due to the colour of the 
surroundings, but to the disturbance of the larve before spinning. 
Caterpillars which were removed from their food, when ready to spin, 
produced white cocoons whether placed in dark or light bags, while 
when white strips of paper were placed about the food, without 
disturbing the larve, they spun dark cocoons. Bateson has further 
established that the dark hue of the cocoons is due to a brown 
fluid contained in the alimentary canal, with which the caterpillar 
colours the silk as it is discharged from the mouth, perhaps, also, 
staining the finished cocoon by an ejection from the intestine. 
When a caterpillar is disturbed, this fluid is voided before spin- 
ning, and hence the cocoon is white. Tutt has discovered (5) a 
similar cause for the varying colours of the cocoons of Halias chlorana, 
which, however, do vary in hue with their surroundings to some 
extent. Poulton has now made further experiments on H. prasinana, 
and, carefully avoiding disturbing the larve, finds that the cocoons of 
this species do vary in accord with the colour of the surroundings. 
The caterpillars seem to exercise choice in the colour of the silk they 
produce, and to be less irritable than those of Saturnia pavonia and 
Eviogaster lanestvis. A larva which had begun to spin a white cocoon 
in a chip box was removed, an attempt was made to cut out the 
egg of an ichneumon, and the caterpillar was injured thereby. 
U 
