132 EDWIN CARLETON MACDOWELL 



ence may be understood as the result of the unconscious selec- 

 tion of the small and weakened flies on account of their fewer 

 extra bristles. 



As was the case in the high selected race, the means of the 

 parents and their offspring show great independence in the 

 return selected lines. This is the more noticeable when one 

 compares the close parallelism between the curves of the high 

 and return races which have totally different parentage. 



Standard deviation 



It has already been demonstrated that the means and the 

 standard deviations in the high race form closely parallel curves. 

 It may be supposed from this that finding close agreement n 

 the standard deviations of the two races would argue for the 

 close agreement of their distributions. The standard deviations 

 of the various groups of the return selected race (dotted line) 

 are compared with the standard deviations of the corresponding 

 groups of the high selected race (broken line) in figure 8. It is 

 evident that the variability in the two races is similar. The 

 fluctuations in the return race are more striking, as should be 

 expected with fewer individuals, but the general agreement as 

 to the amounts and directions of the changes is unmistakable. 



Conclusions 



The data from the return selected race g've ample evidence 

 that the advance found in the early generations of the high 

 selection has not been reversed. In the beginning, each high 

 selection immediately raised the means higher; in this return 

 selection during 8, and again during 6 generations, no general 

 reduction of the means is discernible. Furthermore, it even 

 appears that the offspring from the low parents are the same as 

 those from the high selected race, with the exception of a slight 

 owering that seems to be due to the weakened condition of the 

 return race. This similarity of offspring from high and low 

 parents is of course, not incontestably proved by the facts of 

 this race; were this true, it would be unnecessary to go fur- 



