HYBRIDIZATION METHODS IN CORN BREEDING. 



105 



TABLE 1. Variations in number of rows of grains in self-fertilised strains, and 



in F l and F 2 hybrids. 



It is clearly seen that the F 2 ears show a greater range of varia- 

 tion in number of rows on the ear than those of the F!. Since the 

 empirical range of variation is capricious, and therefore is not a satis- 

 factory measure of variability, I have calculated the mean, standard 

 deviation, and coefficient of variation, for the numbers of rows on 

 the ears of each of these six families. The constants thus derived are 

 arranged for comparison in the following table: 



TABLE 2. Mean, standard deviation, and coefficient of variation for number 



of roivs on ear. 



It will be seen by noting the numbers in the last column that the 

 inferences as to the relative variability of F x and F 2 , drawn from the 

 range of variation in the several families, were correct. This fact will 

 be even more obvious if the coefficients of variation are averaged in 

 pairs. In this way it may be found that these pure strains had an 

 average variability of 9.081 per cent ; their Fj hybrids had an average 

 variability of 9.063 per cent ; and the F 2 hybrids an average variability 

 of 12.63 per cent. My hypothesis requires that in the fertilization 

 of the pure homozygous strains and in the production of the F x hybrids 



