EFFECTS OF SELECTION ON THE YIELD OF CORN. 17 
for stand. These corrected yields were calculated to relative per- 
centage yields in terms of the average of the two adjacent check 
plats. The averages of the three or four relative yields of each cross 
are used as the basis of comparisons between the crosses. The data 
as summarized in Table 4 are relative yields in bushels per acre. 
These were calculated by multiplying the average relative percentage 
yields by 60.26. which is the average yield per acre of the 135 check 
plats. The data in this summary (Table 4) are arranged to show 
the productiveness of the crosses as influenced by the parent strains 
involved. The relative values of the strains as pistillate parents, as 
staminate parents, and as both pistillate and staminate parents also 
are given. 
DISCUSSION OF DATA. 
The only definite promise of increased yields of corn from breeding 
methods involving selection within self-fertilized lines is the utiliza- 
tion of F, crosses or double crosses between such lines. Other possi- 
bilities suggest themselves and should be investigated, but immediate 
efforts to obtain practical results must be based upon the utilization 
of crosses. Obviously, from this point of view the true value of a 
pure line lies in the productiveness of its crosses rather than in its 
own inherent vigor and yield. With this in mind, the crosses here 
considered were compared in order to measure the value of the parent 
strains, and the relative productiveness of the crosses inter se is the 
point of chief importance. The relative yields of the crosses and of 
comparable noninbred strains also are of some interest. The lines 
had been self-fertilized only four generations prior to crossing, how- 
ever, and selection on the basis of vigor and yield had tended to retain 
the more heterozygous lines. The results consequently are not offered 
as any measure of" the ultimate possibilities of this method of breeding. 
The average yield of the 135 check plats was 60.26 bushels per acre 
in the comparison of crosses. The same strain, Xo. 201. check, 
yielded 5S.49 = 1.21 bushels per acre, relative yield, in the compari- 
son of generations. The relative yields of the crosses, therefore, may 
be compared directly with the relative yields of the noninbred gen- 
erations. The yield of Whatley X St. Charles White was 72.90, and 
the average of Xo. 201. F 4 , F 5 , and F 6 was 66.71 bushels per acre. 
There were four crosses that approximated the production of What- 
ley x St. Charles White and four more that approximated the aver- 
age vield of the three later generations of Xo. 201. On the other 
hand, 33 of the crosses yielded from 4.6 to 20.1 bushels less than the 
later noninbred generations of Xo. 201. 
These relations are shown graphically in Figure 2. The horizontal 
broken lines in this figure represent, respectively, the yield of What- 
ley X St. Charles White and the average production of Xo. 201, F 4 , 
F 5 . and F 6 . The yields of the individual crosses between the selfed 
strains are shown by the location of the short horizontal lines project- 
ing from the heavy vertical bars. A bar connects the yields of all 
crosses in which one strain was involved as a parent; consequently 
each cross is represented twice, once in connection witli each parent. 
The mean yields of the groups of crosses are shown by the points on 
the bars that are connected by the solid line. 
Because of the fact that all the parent strains are not represented 
equally in the crosses, the mean yields are inconclusive. However, 
