HEREDITY 495 



10.92 per cent, with a mode of 1 1, and an upper limit (one ear) 

 of 14 per cent, more accurately 13.87. 



The result of one year's selection made little evident impres- 

 sion. The second year's crop was not much better. The mode 

 was the same, the average slightly lower, but the distribution 

 became somewhat extended, with the appearance of two higher 

 values represented by one ear each. 



In the third year, with still better seed (13.06), the distribu- 

 tion extends still farther, but the new value is down, not up. 

 We note, however, a general increase of the higher values, and 

 the mode has moved up a notch. 1 



In the fourth year (1900), with still better seed (13.74), the 

 lower values lessen and some drop off entirely. One new value 

 appears. All the upper frequencies are increased ; the mode 

 has gone up two notches, and we now have no less than 28 ears 

 as good or better than the extreme ear of 1896. 



The same principle continues in 1901, which was an excep- 

 tionally good year for protein, and the theoretical mode goes up 

 nearly three points, reaching a par with the single exceptional 

 ear of the foundation stock. It so happens that this year the 

 number of ears examined was the same as that of the foundation 

 stock (163), and of these ears no less than 47, or 28 per cent, were 

 the equal of the exceptional first ear. The effects of selection 

 settle back slightly the next year, but in the succeeding season 

 lost ground is recovered and there is produced the exceedingly 

 exceptional ear, 17.33, which has since proved a remarkable 

 parent. 



The principle of progression is still further illustrated by the 

 next table, showing the effect of selection both ways. This is 

 taken from Dr. Hopkins's original data in breeding for high oil 

 and for low oil from the same foundation stock. The student 

 should note the decisive manner in which these generations 

 separate themselves from the foundation stock and from each 

 other as selection goes on. Nothing could show more conclu- 

 sively that under intense and persistent selection new values 

 appear freely as the race becomes liberated from the heavy drag of 



1 It can be seen by inspection that the theoretical mode is not so high as the 

 empirical mode, 12. 



