SELECTION IN BAR-EYED RACE OF DROSOPHILA. 363 



decrease as the part of the character under observation decreased 

 in comparison with the part not under observation. 



THE PRESENT EXPERIMENTS. 



Since the experiments of Zeleny and Mattoon gave such clear 

 results, but were interrupted before they led to a final con- 

 clusion, it seemed desirable to repeat them on a larger scale and 

 continue them for a greater number of generations to determine 

 if a pure line could be established and to study the changes in 

 the existing factor or factor complex. 



MATERIAL AND METHODS. 



In order to have a check on any possible contamination a stock 

 was selected which had a second recessive character, the vestigial 

 wing, in addition to the bar eye. This stock also had the ad- 

 vantage of having on the average a lower facet number than the 

 long-winged stock. But sterility in the race and the indefinite 

 character of the results made it necessary, after a few generations, 

 to return to the long-winged stock. In both cases the eye 

 color was that of the normal wild fly. 



The vestigial-winged, bar-eyed stock was designated as VBa; 

 the long-winged, bar-eyed stock as Ba. The downward and 

 upward selected lines in VBa were distinguished by 1 and h, 

 in Ba by d and u respectively. Individual lines were distin- 

 guished by numbers. Any given mating received the number 

 of the generation to which it belonged and a serial number corre- 

 sponding to the number of matings made from that line in the 

 given generation. In that way a number like Bau4f2~5, although 

 cumbersome, gives nearly the whole pedigree of the mating 

 involved. Since the lines are now clearly distinguished without 

 the stock designations, Ba and VBa, those designations will 

 frequently be omitted in this paper. 



The material for the present set of experiments was obtained 

 from Professor Charles Zeleny on whose advice the work was 

 undertaken. To him the author is also indebted for a keen 

 interest in the work and for many helpful suggestions. 



All stocks were kept in large bottles while eight dram vials 

 were used for individual pairs. The vials were plugged with 



