BEGINNINGS OF THE HETEROSIS CONCEPT 37 



years (sections A and C, Table 2.6) shows the latter in excess of the former 

 by 34.0 per cent in grain-row number, 22.1 per cent in height of stalks, and 

 154.2 per cent in per acre yields of ears. The superiority of the Fi hybrids 

 between different inbreds and the families in which selfing had been pre- 

 vented during six generations of controlled breeding (sections D and C, 

 Table 2.6), is indicated by an excess in heights of stalks of the Fi families 

 over the crossbreds, of 9.4 per cent, and in yields of ear-corn per acre of 12.3 

 per cent. But here there is a notable reversal in grain-row numbers. Not- 

 withstanding these proofs of the superior vigor of the F/s over the cross- 

 breds, the latter exceed the former in grain-row number by 10.8 per cent. 



The reason for this reversal is easily recognized when we consider that 

 parents were selected in these studies for their grain-row numbers, with no 

 noticeable selection for heights and yields. In section D of Table 2.6, we note 

 that only one parent of any of the Fi families had a grain-row number in 

 excess of 18. The crossbred families ranged in parental grain-row numbers 

 from 8 to 32. Five of the families came from parents having more than 18 

 rows of grains. 



To make a fair comparison between the two types of breeding in their re- 

 lation to grain-row number, it is necessary to use only the crossbred families 

 having parents with no more than 18 grain-rows. When we make such a limi- 

 tation, we find the average grain-row number for the remaining six crossbred 

 families is only 12.9. The grain-row average for the six Fi families, namely, 

 15.2, exceeds the crossbreds by 17.1 per cent. Limiting the other indicators 

 of physiological vigor to the same six crossbred families, we find that the F/s 

 exceed the corresponding crossbreds on the average by 6.3 per cent in height 

 of stalks and 7.0 per cent in yield of ear-corn. 



In 1911 I was again in full personal charge of the corn experiments at 

 the Station for Experimental Evolution, and was able to expand the work 

 considerably, both quantitatively and in the types of matings studied. 

 We planted 84 cultures in the white dent series as well as 25 cultures of 

 other types of corn. The total number of white dent ears of which the grain- 

 rows were counted was 6,508 which showed the following frequencies: 



Grain-rows 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 Total 



Frequencies 267 767 1725 1298 931 683 363 164 114 95 65 23 7 3 3 6508 



Percentages 4.1 11.8 26.5 19.9 14.3 10.5 5.6 25 1.8 1.5 0.9 0.4 0.1 0.1 0.1 99 9 



In Table 2.7 the 1911 results are presented in condensed form. Families 

 are grouped in eleven sections representing fairly homogeneous groups, 

 mostly based on the types of matings involved. Sections D and E are both 

 made up of the same five families of F2 hybrids produced by selfing the same 

 number of different Fi's. For these families each seed ear was used to plant 

 two rows. The one row of each such family was grown with the other cultures, 

 as usual, in the East Garden. The second row of each of these families was 



