EYE COLOR IN DROSOPHILA MELANOGASTER 379 



the proportion of 60 per cent of general or non-specific modifiers 

 of eosin, such as vermihon and pink, 22 per cent specific modifiers 

 of eosin, and 18 per cent allelomorphs of eosin. The proportion 

 in which mutations are found should, of course, be distinguished 

 from the proportion in which mutations arise. In this particular 

 example the difference should be considerable, for since all the 

 flies are eosin, the situation is particularly favorable for the 

 detection of specific and disproportionate modifiers of eosin, and 

 the percentage found should be correspondingly higher than in 

 our general work where eosin flies constitute only a small fraction. 

 It is probable that mutation is very much more frequent than 

 appears, since a great many mutations are of very shght somatic 

 effect and would pass undetected except that certain characters, 

 such as eosin eye color, truncate wings, beaded wings, and a few 

 others, are pecuharly sensitive differentiators for eye-color and 

 wing-shape genes, etc. Mutations capable of affecting eye color, 

 wing shape, etc., are presumably not less frequent of origin in 

 ordinary stocks, such as pink eye or rudimentary wings, but 

 these latter are poor reagents for the detection of mutations. 

 Whether or not a specific modifier should arise during the course 

 of selection depends on the length of the selection, that is, on the 

 number of individuals, as well as on the frequency of that type 

 of mutation. 



The previous eye colors (all types) may be roughly classified 

 as 90 per cent dilutors and 10 per cent darkeners of eye color 

 which is about the same proportion as that found among the 

 specific modifiers themselves. Cases hke that of eosin sepia 

 show that the ratio of darkeners to dilutors considered from the 

 basis of eosin may be far different from that shown by the pri- 

 mary effects of these same genes (differences from the wild-type) . 

 Whatever this normal ratio for eosin is, it should obtain in the 

 selection experiment as well. When mutation in the direction 

 of selection occurs, there should be a jump in the speed of pro- 

 gression, and the final grade should be correspondingly more 

 extreme. However, the modifier might be a darkener, but in 

 that case it would not be selected and consequently would have 

 only neghgible effect upon the speed and the resultant grade. 



