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



of dots. These facts seem to indicate that the white 

 lines of the larva first arose as enlargements of the 

 shagreen dots, and that the effect has been increased by 

 the ground colour becoming gradually lighter along the 

 same lines. The extreme anterior end of the subdorsal 

 seems to be made up of a row of dots only, without any 

 whitening of the ground colour. The dots on each side 

 of the dorsal line doubtless serve to render the latter 

 more distinct. The importance of the dorsal line is 

 very great for protective purposes, as it represents the 

 midrib of a leaf. It is possible that the ground colour 

 may now be in process of change along these two lines. 

 At the same time the changes of colour before pupation 

 afford evidence that the dark anterior borders to the 

 oblique stripes have risen to the position of distinct and 

 independent markings, and are not merely local deepen- 

 ings of the ground colour. At the time I speak of, the 

 ground colour fades and is replaced by dirty brown, but 

 the green borders appear more distinct than ever (in 

 some specimens at least, and in none do they share the 

 fate of the ground colour). It is very probable that the 

 origin of the white markings from the shagreen dots can 

 be proved in the ontogeny. 



3. The use of the remains of the subdorsal in the last 

 stage of Smerinthus. — At this stage the subdorsal line 

 remains distinct (as Weismann points out) in front of the 

 oblique stripes, and also intersects the first two or three of 

 these, gradually disappearing posteriorly. I believe that 

 it is not generally known that there is the beginning of an 

 8th oblique stripe, slightly marked, on the 5th segment 

 (see figs. 1, 3, and 4). This stripe (which possesses a 

 dark coloured anterior border remaining after the changes 

 before pupation) begins superiorly at almost the same 

 level as the others ; inferiorly it reaches, but does not 

 cross, the subdorsal. If the oblique stripes were 

 repeated anteriorly, with the same relation to the seg- 

 ments that they bear to those where they are present, 

 they would entirely mar the effect of a series, because the 

 anterior segments are so much smaller than the others, 

 and are, further, much contracted when the larva most 

 needs the protective resemblance, i. e., when it is at rest 

 and assuming the Sphinx attitude. But, if a larva be 

 watched in this position, it will be seen that the sub- 

 dorsal line, following the curved anterior segments, 

 becomes approximately parallel to the oblique stripes 



