BRISTLE INHERITANCE IN DROSOPHILA 141 



change in any way during the many generations of unsuccessful 

 selection. However, in spite of the strong influence of environ- 

 ment, some conclusions that do bear on the problem of the 

 germ plasm are to be drawn. 



It may be concluded that environment was not responsible 

 for the initial rise in the high race. Environment was as potent 

 at this time as at any other, yet only at this time were the 

 curves of the parents and offspring parallel. At the beginning 

 low selection was successful, showing that genetic differences 

 between low and high grade flies did exist. After this rise, the 

 race as a whole was changed. This is shown by the failure 

 of the return selection; at first, selecting these same low grades 

 of parents established a low race. Crosses result in more 

 modification after this rise than before. The analysis of indi- 

 vidual families shows that before this rise the parents with 

 higher grades produced children with higher mean grades, while 

 after this rise, similar analysis shows that the highest grade 

 parents were no more likely to produce offspring with especially 

 high than with especially low mean grades. 



It is certain that there were differences already present in the 

 germ plasma of the flies first bred; that these same sorts of dif- 

 ferences were not found among the flies whose ancestors had been 

 selected for several generations; in other words, in regard to 

 those germinal differences that account for the parallelism be- 

 tween parents and offspring in the early generations, that deter- 

 mine the difference between the successful low selections at the 

 beginning and the unsuccessful return selections later on, that 

 make the difference between an uncrossed selected race and a 

 selected race that has been recovered from the dominance of a 

 cross by normal, in regard to all such demonstrable differences, 

 one finds that selection has sorted out a uniform race. Such a 

 conclusion will stand however completely the environment miay 

 be proved to control the fluctuations in all but the early gener- 

 ations of the high race. 



THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 23, NO. 1 



