'36 BULLETIN 1489, U. S. DEPARTMENT OF AGRICULTURE 
At the time that Shull's suggestion was made for using selfed lines 
in corn breeding, it was thought that hybrid vigor was a phenome- 
non of physiologic stimulation. Under such a hypothesis selfed lines 
could be used only as a basis for obtaining hybrid combinations of 
one kind or another. The Mendelian interpretation of hybrid vigor 
as due to the complementary action of favorable dominant factors, 
which already has been discussed (p. 22), was not suggested until 
1910 (7, 44)- It was not generally accepted until after Jones (38) 
had shown in 1917 that the existence of linkage overcame certain the- 
oretical objections that had been raised. More recently Collins (12) 
has demonstrated that these objections were invalid in any event. 
Under this interpretation of hybrid vigor it should be possible to 
obtain high-yielding selfed lines for direct utilization. These could 
be obtained either by extensive selection or by systematically build- 
ing up better lines by selection with alternate periods of selling and 
cross-fertilization. Thus, a productive F x hybrid could be mated back 
to its better inbred parent for several generations, selecting seed only 
from the most vigorous plants each year. In this way a strain could 
be produced that would be homozygous for most of the favorable fac- 
tors of the better parent and heterozygous for at least many of the 
favorable dominant factors of the other parent. Selection in such a 
strain should produce much better selfed lines than those now 
available. 
Although there is no direct evidence as yet, there is some indirect 
evidence, as well as good theoretical reason, for believing that high- 
yielding selfed lines can be obtained. If so, their utilization will be 
simple and needs no specific consideration. Selection within selfed 
lines will be discussed here, therefore, primarily from the standpoint 
of obtaining lines for use in hybrid combination. 
In practice, a number of desirable plants are self -pollinated. The 
seed from the better plants is planted an ear to a row. Self-pollina- 
tions then are made among the progeny plants, from among which 
selection is continued in the same way. After the various lines begin 
to breed relatively true, crosses are made between them and compared 
for productiveness. The lines that produce the higher yielding and 
more desirable crosses then may be combined into single or double 
crosses or into synthetic varieties for commercial utilization. The 
better crosses also may be used as a basis for further selection in at- 
tempts to build up more desirable selfed lines. 
No definite rules can be given for the number of lines to be carried, 
pollinations to be made, or crosses to be compared. The question is 
one of selection, and the larger the quantity of material from which 
to select the better are the chances of success. An outline showing 
the pedigrees of the different lines of a family from a single open- 
pollinated kernel through a 6-year period is shown in Figure 7. 
This will give an idea of the way that selection may be practiced 
within a single family in different years. 
The corn-breeding program being carried cooperatively by the 
Office of Cereal Crops and Diseases of the Bureau of Plant Industry 
and the Iowa Agricultural Experiment Station may be used to show 
the numbers involved in a comprehensive breeding program. From 
2,000 to 3,000 ear rows have been grown each year in this experiment 
from selfed seed tracing back to 500 original open-fertilized ears 
