146 Mr. E. B. Poulton. Colour-relation between the [Feb. 4, 



larva ceased feeding before its last ecdjsis, and it died on August 20th 

 when advanced in the last stage, and an intermediate variety. Thus 

 it is quite certain that the spots can appear on other than yellowish 

 varieties. 



18. Salix rubra. — Three larvae (hatched July 10th) were well in 

 the last stage on August 12th, when they were examined. Like those 

 upon 8. babylonica they had been fed for one day upon 8. triandra. 

 One was decidedly on the yellow side of intermediate, one les3 

 markedly so, and one was intermediate. On August 20th the two 

 former were decidedly on the yellowish side, and I have a note to the 

 effect that I was sure that they were yellower than the larvae fed upon 

 this tree last year (1884), and of which an account is given in the 

 paper already alluded to. At this time the two yellower larvae ceased 

 feeding, while the third was still intermediate, and it ceased feeding 

 about August 25th without further change. 



19. 8. rubra. — One larva (hatched July 10th) which had been fed 

 for one day as above described upon 8. triandra was examined on 

 August 16th, when it was in the last stage and apparently on the 

 yellowish side of intermediate. On August 20th it was advanced in 

 the last stage and unchanged in colour, and on August 27th it. was 

 about full fed and slightly on the yellowish side of intermediate, and 

 there is no doubt that this result was final, for the larva could not 

 have undergone further change when it was so mature, this being the 

 last note I have about it. 



Reviewing these sets of experiments and comparing them with those 

 of Series I and II, we find upon the whole considerable evidence for 

 the existence of a hereditary force which influences the larval colour 

 in this species. 



Ordinary apple (Nos. 1 and 2) produces a maximum effect, as might 

 be expected from previous experiment and observation. It would 

 probably do so even if there existed a strong hereditary tendency 

 towards yellow, and in this case the transmitted influence deviated 

 but little from the direction of the typical white variety (as indicated 

 by the life histories of the present larvae). There is no doubt that a 

 similar effect would have been produced in the other two series if the 

 larvae fed upon this food-plant had lived long enough to enable me to 

 take reliable observations. (As this was not the case, such experi- 

 ments were not alluded to in either series.) Since ordinary apple 

 produced maximum effect, it was practically certain that the same 

 result would follow the use of leaves which were sewn so as to show 

 the white under sides only (No. 3). No. 4, in which the leaves of 

 apple were sewn so as to show the upper sides only, did not terminate 

 as I should have expected, as far as I could judge of the effects in the 

 immature larvae. The process of sewing and paring causes injury to 

 the leaves, so that the larvae did not thrive upon them (being less 



