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



be a result of the same conditions. There is no doubt 

 that this outhne is protective, or has been descended 

 from a protective ancestral character, and such protection 

 by an angular outline, of course, follows from the fact 

 that the pupte are freely exposed. These facts are true 

 of the pupae of Ehopalocera as a whole, but there are 

 exceptions, just as in the imaginal characters. 



Explanation of Plate I. 



Fig. 1. — Natitral size. Adult larva of Smerinthus ocellatus, seen 

 from the left side. The ground colour is bright yellowish gi'een, 

 and the three rows of red spots are seen to form an alternate 

 pattern, the spots of the highest row being placed nearly vertically 

 over those of the lowest ; while the middle spots are approximately 

 intermediate. The anterior remnant of the subdorsal is normal. 

 The sHght obhque stripe anterior to the seven-stripe system is indi- 

 cated. The shagreen dots are arranged in rings parallel with the 

 intersegmental furrows, and there are generally eight rings in the 

 larger segments. The anterior borders of the obhque white stripes 

 are darker green than the ground colom*. 



Fig. 2. — Natural size. Apparently adult larva of 8. jpoindi, seen 

 from the left side ; from a painting by Mr. G. C. Bignell. The larva 

 is remarkable in its light ground colour, and especially in the 

 extreme development of the red spots, which are continued anteriorly, 

 where there are no obhque stripes, and the spots have no tendency 

 to become drawn otit into colom'ed borders. They are rather feebly 

 developed on the claspers. Anteriorly they are not distributed 

 regularly on the segments, since they exceed the latter in number. 

 The alternate arrangement is less marked than in the last figure. 

 This unusually-marked variety does not bear out Prof. Weismann's 

 theory that the sjiots tend to become drawn out into coloured bor- 

 ders to the light stripes, and yet here the system of spots reaches a 

 far higher development than in S. tilice (as far as I have seen figures 

 of the variety of this species) . 



Fig. 3. — Natm-al size. Eather exceptionally large adult larva of 

 S. ocellatus, from the left side. The ground coloiu- is light bluish 

 green. No red spots are present, but there is the normal red line 

 round each spiracle. The subdorsal is retained for its whole length, 

 although but faintly, posteriorly to the normal limits. It ends pos- 

 teriorly in the last oblique stripe (the limit shown in an earlier 



