the colours of certain Lepidoptera. 327 



In working at experiments such as these, I often note 

 the results in each set of larvae, without paying attention 

 to the conditions (which are generally indicated by a 

 number or letter). Hence the observations are entirely 

 unbiassed, for I do not know the past history of each 

 set. Subsequently the notes are written out and the 

 conditions described, and then only can the value of the 

 experiment be estimated properly. In the case of the 

 experiments just recorded, this has only just been done, 

 more than six years after the experiments were con- 

 ducted. Had I written out the results earlier, I should 

 have seen what admirable material was afforded by the 

 larvae of this species, and should have sought them for 

 more detailed and careful investigation. My general 

 impression at the time the notes were taken was, as is 

 often the case, the reverse of that now gained by a 

 careful study and comparison of the whole course of the 

 experiments. At the time I thought that the larvas were 

 not susceptible, or but slightly so. I had even less time 

 than usual to do more than take the necessary notes, 

 being exceptionally hurried while this work was pro- 

 gressing. 



1889. 



The next investigation of this species took place three 

 years later, in 1889, and was the outcome of the acci- 

 dental capture of a female moth which laid large numbers 

 of eggs. I remembered the tendency of the birch leaves 

 to become brown, and determined chiefly to make use of 

 Popuius nigra, the large bright green leaves of which will 

 keep fresh for a very long time if the twigs are placed in 

 water. 



Most of the experiments were begun July 15, a few 

 days after hatching, when the larvffi were still quite 

 small. The results can be given most concisely in a 

 tabular form. 



{See Table, page 330. 



