72 



GENETICS AND EUGENICS 



has become general. Often hundreds of thousands of beets 

 are tested at a single factory. De Vries has plotted a varia- 

 tion curve for forty thousand beets tested in 1896 at a factory 

 in Holland. The result (Fig. 13) was a beautiful frequency 

 of error curve with its mode at 15.5 per cent. The upper 

 limit of variation was 21 per cent, or the same per cent as 



Fig. 13. Graph showing the variation in sugar-content of 40,000 sugar beets tested at a 

 factory in Holland. (After De Vries.) The data are as follows: 



Percent sugar 

 12 12.5 13 13.5 14 14.5 15 15.5 16 16.5 17 17.5 18 18.5 19 



Number 



340 635 1,192 2.205 3,597 5,561 7,178 7,829 6,925 4,458 2,233 692 133 14 5 



The broken line shows the theoretical curve for (a+ b)*". 



Vilmorin obtained after two generations of selection. The 

 general average, to be sure, is considerably higher than when 

 the selection began, but De Vries believes that this is due in 

 part to improved methods of cultivation and more accurate 

 methods of determining the sugar-content. He believes that 

 whatever real improvement has taken place is due largely 

 to the elimination of the poorest sorts through selection, and 

 that these would speedily become reestablished if the selec- 

 tion were discontinued. 



The fact has only recently come to light that sugar beets 

 are regularly cross-pollinated by a minute insect, a species of 



