42 Mr. E. B. Poulton's notes upon the colours 



cling to the branches. They are at this stage dimorphic, 

 the more numerous being green, while the others are 

 brown. The latter are protected by their resemblance to 

 twigs, the former by their likeness to young green shoots, 

 and by being invisible against the background of leaves. 



The larvae of E. pendularia emerged from the eggs 

 before those of the other two sj^ecies, and fed up far 

 more rapidly. I was therefore able to watch the other 

 larvae during the whole period in which protective atti- 

 tudes are assumed ; for the chief protection at the earliest 

 stage (I mean before the first ecdysis) must be the 

 extremely small size and the absence of conspicuous 

 colours or markings. I have already mentioned that I 

 never observed the green larvae of E. omicronaria in any 

 other than the normal attitude. After the last change 

 of skin these larvae also become dimorphic, but the 

 brown variety forms a very small proportion in this case 

 (one out of twelve larvae). Before the last stage all these 

 larvae were green. 



The larvae of E. orhicularia were dark-coloured when 

 young, and very early took up a position in which they 

 exactly resembled snail's excrement. The body was 

 twisted into a very irregular spiral (far more irregular 

 than in the case of E. jyendularia when observed). The 

 illusion was all the more complete, because at this stage 

 the leaf was eaten from the surface and not from the 

 edge, and thus the chief strands of the fibro-vascular 

 framework were left, exactly as they are when a snail 

 has been at work. When the larva becomes larger it 

 presents, when at rest, a striking resemblance to the 

 excrement of birds. It is still dark-coloured, but 

 possesses a series of white markings along the sides, 

 which are shown irregularly in the twists and curves of 

 the attitude assumed. A common position at this stage 

 (when about half-grown) is a spiral twist of about one 

 turn, from the posterior end as far as the last pair of 

 true legs, and then at this point a sudden bend back- 

 wards of the anterior segments and head to a right angle 

 with the proximal posterior segments. The effect is 

 highly irregular. From some points of view the appear- 

 ance is that of another smaller turn added to the first. 



In larvae which habitually assume an apparently 

 angular position it is very common for the bends to be 

 marked by a tubercle or prominence on their convex 

 side. Thus the bends, pointed in this way, appear to be 

 much sharper than they really are, — in fact they look 



