HEREDITY 493 



parents, even with the handicap of 0.7 inch, and only 3 of them 

 were born of extreme parentage (row , column 14). Not only 

 that, but the upper limits of height in this table are decidedly 

 with the children rather than with the parents ; in other words, 

 children can be found taller than any parent. 



At the other extreme of the table, but 6 mid-parents are 

 recorded at or below 64.5, but no less than 103 children are 

 recorded at 64.2, or below. 



All this goes to show that, notwithstanding the principle of 

 regression, the child population extends over a much wider 

 range than does the parental, a fact made clearly evident by 

 comparing the offspring (row n), extending from below 62.2 to 

 above 73.2, with the parents (column 17), extending only from 

 below 64.5 to above 72.5. While neither of these ranges fixes 

 the limit, yet the range of the offspring is clearly the greater. 



Viewed from any standpoint, offspring cover a wider range of 

 variability than their parents, and some of them lie quite beyond 

 the limits of parentage. This is progression, and its effect is, 

 especially in the case of extreme parents, to send some indi- 

 viduals not only beyond the limits of parents but well beyond 

 the former limits of the race. The practical effect is that the 

 population of any race can be moved either up or down through 

 a large range, and either limit extended at will by the use of 

 highly selected parents. 



The behavior of a race under rigid selection and the operation 

 of the principle of progression are well illustrated by the table 

 on the next page, the data for which came from the records of 

 ten years of corn breeding by Dr. C. G. Hopkins, of the Agri- 

 cultural Experiment Station of the University of Illinois. In 

 these experiments the purpose was to influence the protein 

 and the oil content of corn by selection. It is notable in this 

 connection that all four strains, high-protein, low-protein, 

 high-oil, and low-oil, were developed from the same founda- 

 tion stock. 



This table is taken from the experiments in breeding for 

 increased protein in corn. 1 It will be noted that the foundation 

 stock of corn (1896), mixed ancestry, showed an average of 



1 Bulletin No. 7/9, Agricultural Experiment Station, University of Illinois. 



