BRISTLE INHERITANCE IN DROSOPHILA 449 



of the same bristle grade, there were many generations in which 

 the coefficients for the means of the parents would be the same 

 as the coefficients for the parents taken separately. The avail- 

 able generations were further limited by the lack of correlation; 

 when coefficients were too small to be significant, there would 

 naturally be no chance to compare the different amounts of 

 correlation when the parents were taken separately. Five gen- 

 erations have been used in making these tests, namely, genera- 

 tions 6, 27, 28, 29, and 30. The coefficients are given in table 6. 

 If any sex linkage is involved, mothers by sons might be ex- 

 pected to give higher values than mothers by daughters. The 

 table shows that when the daughters correlated to the average 

 of the parents, give higher coefficients than the sons similarly 

 correlated, they also give higher coefficients than the sons, when 

 both are correlated to the mothers and fathers separately. In 

 other words, the same relationship holds between sons and 

 daughters whether they are correlated to the average of their 

 parents or to their mothers and fathers separately. Mothers 

 and daughters are more closely correlated than mothers and 

 sons in four of the five generations. Fathers by daughters 

 are more closely correlated than fathers by sons in four cases. 

 The averages of the coefficients when both relatives are females 

 (mothers by daughters), when one relative is female (mothers 

 by sons and fathers by daughters), and when neither relative is 

 female (fathers by sons) are, respectively, 9X9= 0.1685, 

 9 X cf = 0.1103, (^ X c^ = 0.0998. These statements do not 

 take into account the fact that the coefficients in generation 29 

 are negative; yet the agreement of these coefficients with the 

 other cases rather strengthens the following interpretation. The 

 coefficients indicate that no sex linkage is apparent. The 

 females, whether mothers or daughters, raise the degree of cor- 

 relation; the differences in the amounts of correlation seem to 

 be due to the differences in the variability in the sexes, rather 

 than to any genetic differences in the relationships of the fathers 

 and mothers to their sons and daughters. The difference in the 

 amount of variability in the two sexes appears to be due to 

 some restriction of the higher bristle grades in the males. The 



THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 30, NO. 4 



