Later on, in a note to his translation of Weissmann's _Theory of Descent_, Professor Meldola seemed disposed to think that the variations of colour of some of the species might be phytophagic--that is, due to the direct action of the differently coloured leaves on which the insect fed.

Mr.
Poulton's experiments have thrown much light on this question, since he has conclusively proved that, in the case of the sphinx caterpillar of Smerinthus ocellatus, the change of colour is not due to the food but to the coloured light reflected from the leaves.
This was shown by feeding two sets of larvae on the same plant but exposed to differently coloured surroundings, obtained by sewing the leaves together, so that in one case only the dark upper surface, in the other the whitish under surface was exposed to view.

The result in each case was a corresponding change of colour in the larvae, confirming the experiments on different individuals of the same batch of larvae which had been supplied with different food-plants or exposed to a different coloured light.
An even more interesting series of experiments was made on the colours of pupae, which in many cases were known to be affected by the material on which they underwent their transformations.

The late Mr.T.W.

Wood proved, in 1867, that the pupae of the common cabbage butterflies (Pieris brassicae and P.rapae) were either light, or dark, or green, according to the coloured boxes they were kept in, or the colours of the fences, walls, etc., against which they were suspended.