362 CALVIN B. BRIDGES 



separations were attempted it was found that this color was not 

 sufficiently distinct from the red to make classification accurate. 

 In the eosin half of the flies on the other hand, the separation was 

 easy and entirely accurate. That is, while the so-called 'cream 

 Iir gene by itself gave a certain effect, it gave so much more 

 marked an effect in the presence of eosin that it was decided to 

 retain the name 'cream III' instead of renaming the mutant 

 'magenta.' The separations in the not-eosin half of table 19 

 correspond roughly to those of the eosin half, but the large size 

 of the ebony not-magenta class is due to the impossibility of 

 distinguishing the 'magenta' character, even though it was un- 

 doubtedly present in most of these twenty-six flies. If this 

 character were accurately classifiable (as magenta) without first 

 laboriously combining eosin with all the flies used in the matings 

 of each experiment, it would be incomparably more useful even 

 if slightly less interesting from our present viewpoint. However, 

 an experiment planned through eight generations had to be 

 abandoned because it was found impractical to distinguish be- 

 tween the cream III and the not-cream III where eosin was not 

 present as a sensitizer. But this aborted experiment revealed 

 that the ebony was a disturbing factor — that in the not-ebony 

 flies the distinction was sharper than in ebony flies. The new 

 possibility arose that the mutation could be used (as magenta) 

 by excluding ebony from all the experiments. While this has 

 not been adequately tested, it seems hopeful that with experience 

 one may be able to use this eye-color without a preliminary 

 eosinization of all the stocks, though it is not to be denied that 

 in the presence of eosin the ease and speed of classification would 

 be greater. 



CREAM B 



An eosin female from a stock of non-disjunction when mated 

 to a bar male gave (culture 82, March 10, 1914) among the 

 eosin sons one whose eye color was as light as that of cream II 

 or cream III. This male was outcrossed to a wild female and 

 in F2 gave creams among the eosin sons, but no disturbance of 

 the color of the not-eosin flies (cultures 183, 184, 185). The 



