18 BULLETIN 1396, U. S. DEPARTMENT OF AGRICULTURE 
that most of the characters would cause a deviation from normal 
too slight to be recognized as individual characters, a choice of lines 
or crosses would have to be made on the comparative vigor in combi- 
nation. Uncertainties thus introduced would lessen the percentage 
of recessive characters eliminated to something less than one-quarter. 
The crossing and reselection necessary to eliminate multiple-factor 
characters is a very laborious operation; and unless methods are 
devised that will greatly increase the number of comparisons that 
can be made the elimination of multiple-factor characters as an 
object in breeding appears very unpromising. 
Crosses of inbred strains that are outstanding in vigor may owe 
their vigor to a small number of relatively important factors rather 
than to the chance conjunction of a large number of insignificant 
factors. If this is the case, the chances are better for finding any 
desired character among crosses of the original stocks than among a 
similar number of crosses of inbred lines, and in addition there is a 
saving of time. 
This line of reasoning assumes that we are seeking exceptional, 
favorable, and dominant variations rather than the elimination of 
recessive and deleterious variations. 
In the present experiments the failure to obtain significant increases. 
by the elimination through selfing of the poorest yielding strains, 
together with the numerous instances in which individual crosses of 
selfed lines have shown exceptional vigor, suggests the greater impor- 
tance of dominant variations. Against this view is the failure of 
high-performing selfed lines to show themselves uniformly superior 
as the parents of crosses. This failure may be due to the fact that 
in most of the reported experiments the parents of the crosses were 
closely related. There would be the liability for a favorable domi- 
nant variation to carry with it deleterious recessive characters. Or, 
to put it another way, the deleterious recessive characters would 
survive when offset by some accompanying favorable character. 
If the parents of a cross are unrelated, most of the deleterious. 
recessive characters will be kept out of expression. We may gain the 
advantage of the favorable factors and suppress the accompanying 
recessive characters by crossing unrelated high-yielding selfed strains. 
But there is great need for more information regarding the value as. 
parents of high and low performing selfed lines in intervarietal and 
intravarietal crosses. 
CONCLUSIONS 
from a practical standpoint 1t appears that both the selfed and 
the crossbred breeding methods gave substantial increases over the 
original variety. Excepting the 1924 F,, which had but two-years’ 
selection before the crosses were made and was represented in the 
test by a rather small number of crosses, all seed stocks were sig- 
nificantly more productive than the crib-selected Sacaton June corn 
which was taken to represent the origina! variety. 
The 1924 F, progenies outyielded the crossbred seed by an amount. 
that is statistically significant, but it scarcely can be said that the 
superiority is agriculturally important. The other seed stocks repre- 
senting the selfed experiment and more nearly corresponding to an 
improved open-pollinated variety were not superior to the crossbred. 
seed. 
