BRISTLE INHERITANCE IN DROSOPHILA 125 



grade race in comparison with the means of the high selected 

 race. The flies selected as parents in the low race are shown, 

 as well as the flies selected as parents in the high race. The 

 clear cut divergence of the curves of the two sets of means, 

 whether males or females are considered, leaves no question 

 as to the interpretation to be accepted. The means of the low 

 race are lower than those of the high race because the germ 

 plasm of the low-selected flies determined lower bristle numbers 

 than did the germ plasm of the high-selected flies. The estab- 

 lishment of this race has considerable importance; it proves 

 that there are differences in the germ plasm that differently 

 influence the numbers of extra bristles; it proves that inheritance 

 plays a part in the variation of the numbers of extra bristles. 



RETURN SELECTED RACE 



If the continuous appearance of germinal variations be as- 

 sumed to be the cause of the progress of selection that was 

 demonstrated in the early generations, it would be expected that 

 the reverse movement' of the means could be effected at any time 

 by selecting the low grade flies. Whether or not the advance 

 found in the early generations can be reversed becomes a test 

 of the hypothesis that variations in the germ plasm are continu- 

 ously taking place. If it be found that the return selection does 

 not lower the means as fast as the advance selection raised 

 them it would seem that the germ plasm, having passed through 

 a series of selections, had become changed in regard to its prop- 

 erties of variation. In other words, this would mean that 

 selection, in reducing the variability of the germ plasm, had 

 limited the possibilities of further successful selection; that cer- 

 tain specific properties being sorted out of the germ plasm, one 

 could not again obtain from such a ine the complete array of 

 somatic variations previously obtainable. 



Means 



From the 15th generation of the high race, four pairs of low 

 grade flies (averaging 1.12) were mated to start an inbred race 



THE JOUR^fAL OF EXPERIMENTAL ZOOLOGY, VOL. 23, NO. 1 



