366 



H. G. MAY. 



of reducing them to a minimum. In eyes with 100 facets or 

 less the error is certainly less than one per cent., but it increases 

 rapidly as the facet number increases. Recounts made on eyes 



2 <J 



VBa Stock 



6363d 



62692 



2I(J<3 

 4922 



13 (15) 



55<J< 



4522 



"A A, 



117f3- 

 31d<j 



2492 



9 <S 



hi (15) 

 78d<J 



7122 



hlf 1-5 



34<J<5 



2729 



hlfl-8 (hi) - 



1822 



9 

 r* 



<J 



h2 (15) 

 19<J(5 

 4629 



I / v 

 ' A A 



h3 (15) 



22<JJ 

 3222 



,1 A 



9 <5 



30 70 110 150 190 230 

 16 (h15) 

 23<J< 



!6f1-6 



42<JcJ 

 2522 



I 



30 70 110 150 190 230 270 310 350 390 430 470 

 x. (Stock) <* 



^ 102<J<J 



30 70 110 150 190 230 270 310 350 390 430 470 510 550 590 610 650 690 730 



FIG. 2. Curves showing the effect of selection in individual lines in VBa. All 

 curves are plotted on the scale of fifty. The males are represented by continuous 

 lines and the females by interrupted lines. The parents are indicated at the 

 top and the averages by the black pyramids below. The values plotted are for 

 flies between 20 and 39, 40 and 59, 60 and 79, etc. The number of the mating is 

 given above, followed by the source in brackets unless the parents come from the 

 mating represented by the curve just above. In case of difficulty in determining 

 the relationship of the lines consult Fig. i. 



with as many as 300 facets showed an error of possibly two or 

 three per cent. Of course it takes two or three weeks of practice 

 to reach that point of efficiency. 



