36 Mr. E, B. Poulton's notes upon the colours 



Contrasting the conditions of these larvae, which 

 behave so differently before pupation, we observe that 

 the chief food-plants of S. ocellatus and S. iwiynli are 

 sallow and poplar respectively, and that these almost 

 invariably grow in damp situations, and are surrounded 

 by grass or other green vegetation. Thus the original 

 ground colour of the larvae is the best protection under 

 such circumstances, and, in fact, it is very slightly 

 altered. But the food-plants of the other two larvae 

 (elm and lime for S. tilice, many plants for Sphinx 

 ligustri, e. g., privet, lilac, laurustinus, syringa, ash, 

 holly, &c.) do not grow in damj) places, and are generally 

 surrounded by the bare ground or short turf. Hence, I 

 believe, the importance of the darkening in these 

 instances. In the case of a very large brightly coloured 

 larva like Sp)hinx ligustri, which would show up as an 

 extremely conspicuous object against the ground, this 

 time of moving rapidly about before burying would 

 indeed be fatal. 



It is probable that the brown or dark shades so often 

 met with before pupation are due to changes taking 

 place in the larval colours, and are merely incidental to 

 their destruction. Natural selection would then in 

 certain cases seize upon the slight incidental darkening, 

 and would give it such determinate direction as we see 

 in Sphinx ligustri, &c. It is probable, too, that the 

 same agency would diminish and retard the darkening 

 in cases where it is unfavourable. 



This explanation is probably correct, if we grant the 

 extreme danger of this period to larvae when they 

 descend from their food-plant and come within the reach 

 of new foes as well as old. 



7. The relation of the markings of Sphinx to those of 

 Smerinthns. — There is one fact in the ontogenj^ of S. 

 ligustri which I think is not recorded by Weismann, that 

 the larva is covered with white shagreen dots, until 

 (I believe) the last ecdysis. I am almost certain that 

 the sha greening is retained up to this point, although I 

 have not seen it for many years. The shagreen is 

 exactly like that on Smerinthns larvae, and the retention 

 of this character in the earlier stages of Sjjhinx is a 

 very strong argument for the later origin of the latter, 

 and from a form possessing likeness to Smerinthns. I 

 believe that in the younger stages the oblique stripes 

 also resemble those of Smerinthus in the character of 



