SELECTION IN BAR-EYED RACE OF DROSOPHILA. 369 



All of the offspring from the parents placed into the new stock 

 bottles were counted, including the ones that were used to con- 

 tinue the stock, and selections were made during the entire 

 period. From the preliminary counts it was concluded that 

 flies above a certain limit could be selected as high and those 

 below another limit as low; but when the selections w^ere begun 

 it was found that no flies appeared that approached the lower 

 limit set for high selections and many flies were below the limit 

 set for low selections. As a result a new standard had to be 

 established. Later the stock became more variable and some 

 higher flies appeared together with some very low ones; indeed 

 the lowest female obtained during these experiments was taken 

 from one of these bottles. In all 370 males gave an average of 

 96 facets and 364 females gave an average of 54 facets. More 

 counts on the stock were made March 17 to April 2, and this 

 time 197 males gave an average of 82 facets and 192 females, 

 an average of 44. This variability of the stock was contrary 

 to the conditions found by Zeleny and Mattoon and seriously 

 interfered with the success of the present experiments. It was 

 necessary to pay as much attention to the nature and possible 

 cause of the variation as to the effect of selection. 



The greatest difficulty encountered in the VBa selections, 

 however, was sterility and low production in the single lines. 

 Of 42 matings made from the stock only 6 were fertile and of 

 these only 2 gave a sufficient number of offspring for selection. 

 On the whole about 80 per cent of the matings were sterile, 

 and in many cases even lot matings failed to produce an}- off- 

 spring. By the end of the fifth generation all lines had died 

 out except h/ and hi 5, and these were saved only through 

 lots. For that reason the experiment was discontinued at that 

 point. 



No special experiments were made to determine the cause of 

 sterility. Obviously it was not due to inbreeding, for it was as 

 great in the first generation as in succeeding ones. In the 

 majority of cases it was also not due to the inability of the 

 females to deposit eggs as was the case in the experiments of 

 Hyde. A few of such abnormal females were observed, but in 

 most cases eggs appeared in the food and failed to hatch. The 



