BRISTLE INHERITANCE IN DROSOPHILA 443 



progeny they respectively produce. Instead of carrying an 

 experiment over a long period of time and through many genera- 

 tions, a similar opportunity for the study of germinal phenomena 

 is afforded by raising large numbers of families in the same 

 generation at the same time. Indeed, the more uniform environ- 

 ment so secured will tend to make the germinal influence more 

 clear. If there is no difference between the germ plasm of 

 high- and low-grade parents, their children will all be alike, 

 that is, there will be no tendency for high-grade parents to 

 produce offspring of any higher grade than are produced by 

 low parents. The correlation coefficients in this case would 

 show no correlation at all, and it could be safely concluded that 

 selection will not have an effect upon the racial mean. This is 

 exactly the same reasoning as applied to the previous genera- 

 tions, but in this case the answer is more critical. Fifty gen- 

 erations of the closest inbreeding and selection must have re- 

 moved all traces of heterozygosity so completely that the 

 discovery of positive correlation in these last generations, with 

 their environmental uniformity, would admit no alternative to 

 the conclusion that genetic changes must have occurred dur- 

 ing the generations in question, and that selection accordingly 

 could occasion further advance in the means of the race. 



This experiment, even standing alone, supplies sufficient data 

 for the establishment of a general conclusion regarding the pos- 

 sibility of securing, through selection, a racial modification in 

 the bristle number of this stock after the homogeniety of the 

 race has been unquestionably established. 



Experiment 



Starting from one pair of flies that came from the 49th gen- 

 eration of continuous inbreeding and high selection, all the 

 progeny were graded and mated with no further selection. As 

 in all other cases, only virgin females were mated, always with 

 their brothers, one pair of flies per bottle. The only other 

 restriction imposed was that the two flies mated should not 

 differ by more than one bristle. 



