A COMPARISON OF MAIZE-BREEDING METHODS 15 
The hope of improvement through the elimination of recognizable 
variations by selfing is based on the idea that these unproductive 
variations appear in commercial varieties with sufficient frequency 
to depress the yield materially. Although seed from the unproduc- 
tive variations is not planted, the characters persist in the hetero- 
zygous state and by means of the pollen from the unproductive plants. 
Careful detasseling would, of course, eliminate the last source of 
infection; but so long as the plants are cross-pollinated, heterozygous 
individuals can not be distinguished from homozygous dominants, 
and recessive characters wil! persist. 
On the other hand, selfing brings simple heterozygous characters 
into expression, and failure to appear in even a small population 
makes it reasonably certain that the line carries only dominant fac- 
tors. 
So far as simple characters are concerned there would seem to be 
no value in selfing for more than one generation, for with a popula- 
tion of only 50 F, individuals the chance of a simple heterozygous 
character not coming 
into expression is less 1924 1'923 1922 92! 1920 1919 1918 
than one in a million. aise. \ 
It probably is safe 
to conclude that the Az 
great majority of the 
conspicuous Varia- 
tions of maize are less t7 
productive than the 
normal type. The 4x 
question is: Do these 
appear in commercial hip caces 
varieties in sufficient sin rane * 
numbers to reduce the 
yield appreciably ? 
2-4 = SS IV 2 
it In the course of these 
experiments 400 selfed 
lines have been grown ae ee of Rromeuses in aie e pe 
of maize tested in 1924. Numbers indicate individual plants 
and the characters of Selfed generations are shown by double lines. F=45.3 
the individuals noted. 
These lines exhibit the usual diversity, including numerous minor 
chlorophyll disorders, but only three conspicuous variations of a 
clearly deleterious nature have been observed. These are ‘ ‘ golden,” 
“silkless,” and ‘‘zigzag’’ stalks. In a careful examination of the 
commercial plantings of Sacaton June corn, golden and _ silkless 
plants were found, but with a frequency something less than one in 
a thousand, and no zigzag plants were observed. 
It is obvious that in this variety at least the elimination of con- 
spicuous variations can not be expected to effect any material 
improvement in yield. 
Breeding with simple characters, the individual effects of which 
are too insignificant to be recognized, is the essence of the dominant 
growth-factor idea. When this theory was formulated, it was 
assumed that although the individual characters were not recogniz- 
able, the stocks most desirable in combination might be selected from 
their behavior as selfed lines. The work of Richey (5) has shown 
