BRISTLE INHERITANCE IN DROSOPHILA 



133 



ther. But no single proof of such a conclusion seems even 

 theoretically obtainable. However, in the accumulation of 

 evidence from different directions, the basis for a conclusion 

 may be found. 



The conclusion that high and low grade flies seem to have simi- 

 lar germ plasm sounds quite the reverse of the concluding state- 

 ment of the preceding section, namely, that there are different 

 sorts of germ plasm that differently influence the numbers of 



STANDARD DEVIATIONS 



MALES 



1.6 _ 



STANDARD DEVIATIONS 



FEMALES 



RETURN RACE 

 HIGH RACE-—- 



I I I 



J \ ! L 



2 4 6 8 10 12 14 16 18 



A 



10 12 14 16 IB 



B 



Fig. 8 Standard deviations in the return race compared with the stand- 

 ard deviations in the high race. The close agreement between the standard 

 deviations in these two races supports the conclusion that the races are not 

 differentiated. 



extra bristles. The explanation of these rather conflicting state- 

 ments will be found in the first section, where the course of the 

 high selection was described showing a strong rise in the early 

 generations. The conclusion that different grades do not have 

 uniform germ plasm applied to the flies after the advance had 

 occurred; the conclusion that different grades do have different 

 germ plasm, applied to the flies before the advance had occurred. 

 In view of the conclusions reached in the second section, the 

 failure of the return selections to establish a low grade race 

 can not be claimed to depend on the concealing influence of 



