EFFECTS OF SELECTION" ON THE YIELD OF CORN. 11 
rows "containing the deficient stand. Comparisons based partly on 
these adjacent rows, therefore, should tend to reduce any distortion 
caused by the correction made for stand. 
It seems desirable to consider first the relations between the three 
first-generation crosses. The difference between the yield of Whatley 
X St. Charles White and that of its reciprocal cross is negligible, 
and the reciprocal crosses may be considered essentially equal. 
No. 201, F lf is the cross Whatley X St. Charles White, as made in 
1914. The object of remaking this cross was to provide seed grown 
under the same conditions and of the same age as the seed of the 
later generations. Can Whatley X St. Charles White be substituted 
for No. 201, F 17 in a comparison of generations? It would seem so. 
Reference to the results in 1915 (p. 1) shows that seed from the 
same lot as that used to plant the rows of No. 201, F x , in the present 
experiment produced 81.9 bushels per acre, in comparison with a 
yield of 67 bushels for St. Charles White. This relation agrees suffi- 
ciently with that between Whatley X St. Charles White and the St. 
Charles White in the experiment in 1921 to warrant substituting 
Whatley x St. Charles White for No. 201, F x , in comparing the 
generations. 
The productiveness of No. 201, F^ unquestionably was higher than 
is shown by the actual-yield data, as indicated both by its relative 
yield and by a knowledge of the experimental conditions. Never- 
theless, it was not as productive as the new seed of the same cross. 
Its earlier ratio to St. Charles White shows that this loss was due to 
deterioration and not to any effect of the season during which the 
seed was grown. This marked reduction in yield, definitely due to 
age, seems worthy of note. 
If the mass selection practiced in the successive generations had 
any effect on yield, the results fail to show it. The explanation that 
heterosis is due to the complementary action of dominant growth 
factors assumes that there are genetic differences between varieties 
that produce an increased yield when crossed, with the entire F x 
population heterozygous for such factors. Following random mat- 
ing there would be a reduction in the degree of heterozygosis in the 
F 2 generation, but no further reduction would occur in later genera- 
tions. 4 The comparative yields of the different generations are in 
fair accord with this theory. The F x generation of the cross What- 
ley X St. Charles White was markedly superior to the average of the 
parents and slightly better than the better parent, Whatley. This 
superiority was lost in the F 2 generation, which produced somewhat 
less than the average of the parents. The yield of the F 4 , F 5 , and F 6 
generations, however, was approximately the same as that of No. 
201, F 3 . 
The very low yield of No. 201, F 3 , probably was caused by close 
breeding. As noted, the original F 2 generation was produced in an 
isolated plat of some 20 plants. The F 3 seed planted in the present 
experiment was obtained from this F 2 generation, with a high proba- 
bility of close breeding. It should be noted, however, that the F 3 
population would be as heterozygous as a whole as the F 2 , barring 
such factors as may have been eliminated accidentally. It was the 
individuals that were close bred, and intercrossing among these indi- 
4 Pearson, Karl. A generalized theory of alternative inheritance, with special reference to M 
laws. In Phil. Trans. Roy. Soc. London, ser. A. vol. •_':;. p. 53 
