BRISTLE INHERITANCE IN DROSOPHILA 455 



Two investigators have used the linkage method in connection 

 with the genetics of bristle characters, Payne ('18) and Sturte- 

 vant ('18). Payne selected for increased numbers of bristles on 

 the scutellum; for twenty-nine generations the means seemed to 

 respond to selection. During this time there were two periods, 

 generations 11 to 14, 18 to 25, during which there was no ad- 

 vance. These are thought to be due to the attainment of homo- 

 zygosis; the subsequent advances are taken to indicate the ap- 

 pearance of new mutations. Unfortunately, no special study 

 was made of the environmental influences, and in their absence 

 the fluctuations in the curve of the means can have but little 

 weight in indicating • the genetic constitution of the race (Mac- 

 Dowell, '17, p. 139). 



The linkage relationships indicate that there is a bristle factor 

 in the sex chromosome and probably one in the third chromo- 

 some; but the relation of these factors to the advance of the 

 selected means has not been satisfactorily demonstrated. The 

 suggestion that only one factor was concerned at the beginning 

 of the experiment does not find support in the ratios (p. 7). It 

 is to be regretted that neither of these factors was isolated and 

 its Mendelian nature proved by crosses. 



Payne mentions two races that were separated by selection 

 from a low-grade mutant; the complete factorial analysis of these 

 races in connection with the high-selected race should yield inter- 

 esting results. Having shown that extra bristles in his flies de- 

 pend upon more than one factor, Payne applied the test of cross- 

 ing the selected race repeatedly to normals; if multiple factors 

 are involved, the successive crosses should show increasing re- 

 gression. However, this result was not found. In view of the 

 uncertainty as to the environmental influence and the relation- 

 ship and generic homogeneity of the wild males used in the crosses, 

 this result can hardly invalidate the method. This is the test 

 that Castle has accepted as crucial and has occasioned his final 

 adoption of the mutliple-factor hypothesis. 



Sturtevant has analyzed the effect of selecting the variable 

 numbers of dorsocentral and scutellar bristles that appear asso- 

 ciated with the mutation called dichaet ; besides influencing the 



