BEGINNINGS OF THE HETEROSIS CONCEPT 31 



if the advantage consisted solely of the effects of heterozygosity, both Strain 

 A and Strain B were still a good way from being homozygous, Strain B being 

 as yet more effectively heterozygous than Strain A. 



In the reciprocal crosses between these nearly homozygous strains A 

 and B, we have our first opportunity to arrive at an approximation to the 

 actual amount of heterosis. The most important new discoveries these 

 crosses made possible were: (1) As a result of such a cross it is possible to 

 completely cancel in a single year the accumulated deterioration which 

 had gradually accrued, although with lessening annual increments, over a 

 period of several years; and (2) the approximate identity of the results of the 

 reciprocal crosses gave assurance that the amount of heterosis resulting from 

 a given hybridization is a specific function of the particular genetical combi- 

 nation involved in the cross. 



Several new cultures of yellow- and red-grained corn were added to my 

 experimental field in 1908, but these will not be followed here. They are 

 mentioned only because they were included in my numbered pedigrees, and 

 their omission in the following tables leaves a break in the series of numbered 

 families which might lead to some question as to the reason for the apparent 

 vacancies. The data from the 1909 cultures of white dent corn are presented 

 in Table 2.5. 



The families grown in 1909, as tabulated in Table 2.5, fall into three major 

 classes: (1) Twelve families involve continuations of the original self- fer- 

 tilized lines, whose average yields range from 18.8 to 41.2 bushels per acre, 

 with the average for all twelve at 32.8 bushels per acre; (2) Twelve are con- 

 tinuations of crossbred families in which strictly controlled cross-fertiliza- 

 tions were made with mixtures of pollen taken from the other plants in the 

 same crossbred strain. These yielded from 58.1 to 83.3 bushels per acre with 

 the average of all at 73.3 bushels per acre; and (3) there were fourteen Fi hy- 

 brid families from crosses between pairs of individuals representing two dif- 

 ferent selfed lines. The yields of these range from 60.3 to 87.5 bushels per 

 acre, the average for all fourteen being 78.6 bushels per acre. As stated in my 

 1910 paper, the three highest yields of any of these cultures were from the 

 families produced by crossing representatives of different selfed strains (see 

 D8.13, D8.16, andD11.21). 



Besides these, there were two cousin crosses involving matings between 

 different families of the same selfed line. These produced, respectively, 27.1 

 and 44.6 bushels per acre. One cross between two sibs in Strain A gave 26.0 

 bushels per acre. The other cross was two F2 families, each from crosses with 

 mixed pollen within one of the Fi families of my 1908 cultures. These Fo 

 families yielded 54.2 bushels per acre from the (A X B)F2, and 70.6 from 

 the (B X A)F2. These yields should be compared with those of the corre- 

 sponding Fi families grown in the same season, in which (A X B)Fi yielded 

 74.9 and 83.5 bushels in two different families, and (B X A)Fi produced 

 82.6 bushels per acre. 



