190 ROSCOE R. HYDE 



In this experiment the entire output of eggs of the truncates 

 was collected. The count from the Woods Hole stock does not 

 represent the entire output since these flies were about ten days 

 old when the experiment began. It is true that it is practically 

 impossible to get the entire output of eggs by relying upon an 

 ordinary dissecting microscope without spending an undue amount 

 of time, but as the same method was applied throughout this 

 source of error as it enters into a comparison of the output of 

 eggs of the different races is about constant. I feel quite sure that 

 the counts, as here shown, represent over 80 per cent of all the 

 eggs laid. That many were missed is certain as the food from 

 which the eggs were picked was often kept and later larvae were 

 seen to emerge. In regard to the conditions of the experiment as a 

 test of viability, no objection on the grounds of personal selection 

 can be made, as it is impossible to tell beforehand which eggs are 

 going to hatch. Care must be taken not to allow the eggs to 

 stand too long before picking them as the larva will emerge from 

 the fertile ones while the sterile eggs are left behind to be picked 

 up and considered in the counts. 



The question may arise in the mind of the reader as to why 626 

 eggs are not considered in table 14 b in determining the percentage 

 of eggs which hatched. The explanation is given here since the 

 question recurs in some of the other tables and in succeeding 

 experiments. It is evident from the method used in making the 

 test that a picture is given of the pr.oductivity of any single female 

 day by day. The 120 recorded in column "No. not considered " — 

 No. 18 (table 14 b) represents the total number of eggs laid by the 

 female through a number of consecutive days in which none of her 

 eggs hatched. The eggs that failed to hatch were in practically all 

 cases those isolated during the first few days of the experiment, 

 although in a few cases they came in the middle of the count, 

 more frequently however toward the close of the count. This 

 seems a fair way to treat the facts as this may be due to other 

 causes operating other than eggs meeting sperm. 



In this case the benefit of the doubt is given to the truncates for 

 if these are considered 749 in 3643, or 21 per cent, hatched. In 

 reality the percentage is not changed much and had this been 



