PRODUCTIVENESS OF SELF-FERTILIZED CORN 17 
based finally on the performance of the lines in combination rather 
than in the self -fertilized condition. Obviously, however, it is neces- 
sary for practical reasons to have lines that are productive in them- 
selves. There is nothing in these experiments to indicate that just 
as high-yielding crosses can not be obtained from lines which are 
themselves productive as can be had from self-fertilized lines that are 
.low yielding and undesirable. 
Considering the investigation as a whole, the data indicate that 
significantly larger yields of corn can be obtained from ¥ 1 crosses 
between self-fertilized lines. Much attention heretofore has been 
focused upon what might be expected when corn was self-fertilized 
continuously with a minimum of selection. This was natural and 
desirable. The newer sj^stems of corn breeding are distinguished 
chiefly in the utilization of inbreeding, which was carefully avoided 
in the older systems. The attention that has been devoted to the 
expectations under self-fertilization without selection seems to have 
been unfortunate also in some respects. It has tended to emphasize 
the importance of self-fertilization to such an extent as to minimize 
the importance of selection. The present data suggest that the 
increased yields of the crosses are due to selection rather than self- 
fertilization, the function of the latter being principally to obtain 
definite entities from among which to select. Finally, although it is 
desirable for practical reasons to have self-f ertilized lines which are 
as productive as may be, the present investigation indicates that the 
final selection of lines for use in crosses must be based upon their 
performance in crosses. 
SUMMARY 
The yield of 70 F x crosses between lines of corn se'f-fertilized for 
six generations before crossing ranged from considerably less to con- 
siderably more than the yield of the parent variety. 
The average yield of three of these crosses was 30 per cent more 
than that of the parent variety, and the consistency of the data 
showed clearly that this superiority was not due to chance. 
This 30 per cent increased yield indicates that these crosses are 
significantly more productive than the Fx varietal cross Whatley xSt. 
Charles White, which has been outstanding in a number of varietal 
comparisons in northeastern Arkansas. 
Comparisons between successive generations of self-fertilized lines 
and between crosses following self-fertilization for different numbers 
of generations show the importance of selection in obtaining larger 
yields by the methods followed and indicate that the principal role of 
self-fertilization is to isolate definite lines differing from each other 
among which selection may be practiced. 
It is necessary for practical reasons to have inbred lines that are 
productive in themselves. The data indicate, however, that there is 
little or no relation between the productiveness of the self-fertilized 
lines and that of their crosses and that the final value of the lines for 
crossing must be determined by comparisons of the productiveness 
of their crosses. 
