34 Mr. E. B. Poulton's 7iotes upon the colours 



have tried, but without success, to find figures of the 

 younger stages. I have no doubt that in the early- 

 stages the horizontal system is alone present, and that 

 later these lines coexist with an oblique system, while in 

 the highest stage the former becomes part of the latter. 

 In fig. 5 a copy is given of the larva represented upon 

 page 203 of Newman's ' British Moths.' 



4. The comparatively recent replacement of the sub- 

 dorsal hi) the oblique lines in Smerinthus. — Weismann 

 proves this fact from the ontogeny of the larvae. 



After the formation of the oblique line system it seems 

 likely that the subdorsal somewhat rapidly disappeared 

 from all but the anterior segments of the adult larva, 

 because (as Weismann points out) it would interfere 

 with the protection from the oblique lines, and also 

 because it would spoil the effect of its anterior part, 

 which, as above shown, lends itself to the newer system 

 of markings. But these changes must have been very 

 recent, for traces of the subdorsal can generally be 

 made out by careful observation upon all the usual seg- 

 ments in adult larvae. 



In the gradual deepening of the ground colour to form 

 the borders to the oblique white lines the point of inter- 

 section of the former with the white subdorsal must 

 have been lighter than the ground colour, and has 

 therefore taken longer to darken. Hence in each seg- 

 ment, at the level of the posterior part of the subdorsal 

 remnant, a light linear interruption of the green border can 

 generally be detected. In one adult larva, with the bluish 

 green ground colour (found on Salix riminaUs by the 

 Eiver Cherwell, Sept. 8th, 1883), these light short lines 

 were connected together into a faint representation of a 

 complete subdorsal line, ending posteriorly in the last 

 white stripe (see fig. 3). The line was here most 

 distinct and linear where it crossed the dark borders, 

 but was fairly well-defined inferiorly through its whole 

 length. The upper limits gradually shaded off into the 

 ground colour, and the line seemed to spread upwards 

 in the anterior part of each segment, so that it was of 

 considerable thickness where it joined the white oblique 

 line. But this appearance can be detected in the pos- 

 terior part of the subdorsal remnant in most larvae. 

 The larva which thus retained the subdorsal was full- 

 fed and of large size. This observation shows (together 

 with the traces which can often be detected) tbat the 



