CORN BREEDING 41 
corn. Of these, 681 (36 per cent) segregated for different seedling 
abnormalities, the most common being some form of chlorophyll 
deficiency (#£). As the plants grow older, other defects become 
evident. Some chlorophyll deficiencies do not develop until after 
the seedling stage. Aberrant conditions of the inflorescences, includ- 
ing complete sterility of the staminate or pistillate parts, appear at 
blossoming time, and there may be evidences of an unbalanced metab- 
olism such as striping or firing of the leaves. All of these condir 
tions — and only a few examples have been mentioned — are undesir- 
able from the standpoint of corn production and are to be eliminated 
from the breeding stocks. Furthermore, although many of them 
would be eliminated automatically sooner or later, it is important 
to weed them out promptly. Their presence tends to result in 
grossly irregular stands of uneven plants and so obscures the less 
striking differences that occur. It is good practice to increase tem- 
porarily the number of lines representing a strain that is segregat- 
ing for some striking abnormality so as to permit eliminating the 
abnormality as quickly as possible if it is desired to maintain the 
strain. 
So much for the elimination of the grossly unfit, or what may 
be called negative selection. The decision as to what direction posi- 
tive selection shall take is more difficult. Selection for productiveness 
among the plants and ears will be on much the same basis as in mass 
selection in open-fertilized corn, namely, the production of sound 
grain from normal plants. The exact type of ear and plant will 
differ with the locality for which the corn is being selected. The 
evidence from experiments with open-fertilized corn that relatively 
long ears with not too many rows of moderately wide, thick kernels 
tend to be more productive suggests this type as desirable also in 
selfed lines. In fact, there already is some evidence that an associa- 
tion exists between this t} T pe of ear and productive selfed lines (30, 
b2). It seems probable that too much attention to type of ear is not 
warranted in selecting within selfed lines any more than in open- 
fertilized corn. When opportunity presents. to select between ears 
from equally desirable plants, however, the longer ears with larger 
kernels probably should be chosen. Where some plant character 
such as height of plant, height of ear, freedom from suckers, husk 
protection against insect attack, or the like is the improvement 
sought, it obviously will be the basis of selection. Otherwise, about 
all that can be done is to select from the more productive plants 
that are of a type and length of season adapted to the conditions 
where the corn is to be grown and as free as possible from disease. 
The final selection is to be based upon the measured yield of the 
lines in hybrid combination, and up to the present no characters 
have been found that can be used to tell which selfed lines will pro- 
duce the highest yielding crosses. 
It is obvious that the foregoing consideration of plant selection 
as an elimination of the worst and a retention of the best is a matter 
of convenience only. Clearly, the dividing line is arbitrary. At the 
same time, this difference occurs naturally in the actual practice of 
selection. Most of the selection in the earlier stages of the breeding 
program of necessity consists in discarding the grossly unfit. So 
many abnormal characters occur that less obvious differences are 
