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



It is possible that the same causes have helped to produce yellower 

 results than are normal in the breeding experiments with Salix alba 

 and 8. Smithiana, for these leaves have been somewhat disturbed, 

 although not nearly so much as in the case of S. viminalis. I do not 

 doubt the validity of this explanation for the latter plant, and the 

 results of experiments are thus satisfactorily interpreted which have 

 been sources of difficulty and uncertainty since the summer of 1884. 

 This explanation also clears up what I felt to be a great difficulty in 

 my former paper when I wrote the words (page 314) " it is only the 

 part of the environment imitated which produces any effect, e.g., the 

 under sides of the leaves in the case of S. ocellatus, and yet the 

 environment, of course, includes both surfaces." I have shown above 

 that the effective part of the environment — the immediate environ- 

 ment — does not in many cases include both surfaces, but either 

 entirely or chiefly the under surface, i.e., that which is ipso facto 

 imitated, and when it does include the other surface for a sufficiently 

 long period, a different effect is produced (in the case of leaves with 

 differently coloured sides). It is therefore obvious that when we speak 

 of the tendency of a plant to produce a certain colour, we mean a 

 tendency from the size and arrangement of the leaves to encourage a 

 larval position in which the effective colour of the environment is 

 only contributed by one leaf surface, or, on the other hand, a tendency 

 to change the larval position into one in which both surfaces may 

 become equally effective, or, again, into one in which either of them 

 may predominate. In this explanation of what is meant by the 

 tendency of a plant, I am, of course, especially referring to those 

 with leaves having white under sides ; but it will probably apply to 

 some extent in nearly all cases, for there is always some difference 

 between the two sides of the leaves. 



There is one other comparison between the captured and bred 

 larvae which is a source of difficulty. The exceedingly uniform 

 results upon Salix rubra in the field (I have found altogether twenty- 

 two larvae, of which eighteen were yellow, three yellowish interme- 

 diate, and one intermediate) render it more than probable that the 

 plant has possessed an influence sufficiently powerful to reverse a 

 larval tendency in the direction of white (for it is very unlikely that 

 in all the eighteen instances of a maximum result the larva happened 

 to tend in the same direction as the food-plant) ; and yet in the 

 breeding experiments, out of nine larvae eight were yellowish interme- 

 diate and one intermediate ; and thus in no case has the food-plant 

 completely overcome a strong tendency (1884) or a somewhat modified 

 tendency (1885) towards white. 



I have thought that part of this difference (also observable in the 

 cases of 8. babylonica and S. triandra) may be due to the fact that the 

 tops of the glass cylinders in which the larvae are bred, are covered 



