CONTINUOUS SELECTION FOR EAR TYPE IN CORN • t> 
DATA ON NUMBER OF KERNEL ROWS 
Data prior to 1923 were based on 1,000 ears of each Idnd grown 
in the seed plats. The data for 1923 were based on the ears pro- 
duced in the yield comparison. About 340 ears of each strain and 
cross were available except in strain Xo. 1, some of the ears of 
which became mixed with other corn before counting. Ears that 
could not be classified definitely because of irregularity were in- 
cluded in the counts, but the distributions shown here are for ears 
per 1,000 classifiable ears. The slight discrepancies in the totals 
arise from rounding off decimals. The data showing the number 
of ears with dilferent numbers of kernel rows, the mean, the stand- 
ard deviation, and the coefficient of variatiofi for each strain and 
cross throughout the experiment are shown in Table 1. 
No data are available regarding the number of kernel rows on 
the ears of C. I. Xo. 119 at the beginning of the experiment. The 
621 ears used to plant the ear-row plats from 1907 to 1918, inclusive, 
had an average of 16.6±0.05 kernel rows,^ and the mean for the 
1,475 ears of the 1923 crop was 16.1 ±0.04. Beginning with about 
16.0 to 16.5, therefore, the mean num^ber of kernel rows was shifted 
upward or downward by selection, so that the means ranged from 
22.9±0.07 for strain Xo. 1 to 10.9±0.04 for strain Xo. 6 in the crop 
of 1922. This is a spread of 12 ro^vs, or only 2 rows less than that 
of the entire distribution of C. I. Xo. 119 in 1923. 
The seed planted to represent the different strains in 1923 was not 
selected for number of rows. Consequently the general tendency of 
the distributions in 1923 to approximate those in 1922 affords some 
indication of the degree to which the characteristic number of rows 
was fixed in the different strains. The standard deviations and 
coefficients of variation, in connection with the means, give a better 
measure of conformity to the characteristic number of rows. Be- 
cause of the large influence of the means on the coefficients of varia- 
tion and the slightly greater variation to be expected in strains 
with larger row numbers, both constants should be considered to- 
gether. On this basis each of the strains, except perhaps Xo. 1, was 
practically as uniform by 1923 as a standard varietv of corn, C. I. 
Xo. 119. 
The numbers of kernel rows on the ears from the crosses were 
mtermediate, in general, between those of the two parents. The 
distributions of the ears of three crosses and their parent strains are 
shown graphically in Figure 2. 
There are some interesting differences in the distributions for the 
reciprocal crosses between strains Xos. 1 and 2, 1 and 6, and 3 and 6 
(Table 1). In each case the distribution of the ears of the cross 
tends to be more like that of the pistillate parent strain. The dif- 
ference of 2.2 ±0.08 rows in the reciprocal crosses between strains 
Xos. 1 and 6 is particularly striking. Whether this was due to some 
influence of the methods used or to inherent tendencies can not be 
determined. In view of the facts that the pollinations were not 
strictly controlled by bagging, etc., and that the seed ears used to 
represent the different sorts were not selected, too much importance 
probably should not be attached to this difference. 
3 Riehey, F. D., and Willier, J. G. A statistical study of the relation between seed-ear 
characters and productiveness in corn. U. S. Dept. Agr, Bui. 1321, 20 p. 1925. 
43522—25 2 
