oct. ja, 1917 Indirect Effects of Selections in Breeding Indian Corn 107 
Table I. —Frequency distributions with respect to physical characters of the ears of four 
strains of corn bred for chemical composition—crop of 1914 —Continued 
LOW-OIL STRAIN 
Length of ears. 
Weight of ears. 
Circumference of ears. 
Rows of kernels. 
Inches. 
Frequency. 
Ounces. 
Frequency. 
Inches. 
Frequency. 
Number. 
Frequency. 
2. 5 
2 
4. 00 
1 
8 
ty 
3-0 
3 
3 
4 
4-25 
O 
10 
59 
3-5 
2 
4 
19 
4. 50 
3 
12 
287 
4. 0 
10 
5 
40 
4*75 
7 
14 
1S5 
4-5 
22 
6 
S 3 
5- 00 
I 9 
16 
3<3 
S-o 
25 
7 
126 
5*25 
20 
18 
4 
5 * 5 
29 
8 
97 
5. 50 
46 
6. 0 
62 
9 
100 
5. 75 
77 
6. 5 
106 
10 
55 
6. 00 
no 
7.0 
100 
11 
2 3 
6. 2% 
117 
7 * 5 
116 
12 
9 
6. 50 
130 
8.0 
95 
13 
4 
6. 75 
•j 
8.5 
60 
7. 00 
0 1 
9. 0 
18 
t 
7. 21; 
9 - 5 
7 
7 - So 
■ L tj 
10. 0 
2 
7*75 
0 
x 
The character of the frequency distributions may be described in a 
general way as not far from a symmetrical form. Occasionally a distri¬ 
bution has presented an appearance that seemed pretty decidedly skew, 
and we have followed such cases into succeeding generations without good 
evidence that such decided skew form persisted in later generations. On 
this account it seems likely that the decided skew condition was due 
mainly to soil differences on a given plot. There is, however, some tend¬ 
ency for the distributions as to length and weight of ears to be slightly 
skew, with the mode greater than the mean. 
We have tested a number of the distributions, and found some of the 
normal type, and some differing significantly from the normal type. 
Taken as a whole, it seems, however, that these distributions are suffi¬ 
ciently near the normal type, to be well described for our purposes by the 
means, standard deviations, and coefficients of variability. 
THE FOUR STRAINS WITH RESPECT TO CHEMICAL COMPOSITION 
These four strains were produced from the same original stock (163 
ears) and each seed selection was made from the highest (or lowest) ears 
within the strain with respect to the character in question. For example, 
the high-protein strain is the offspring of those ears showing the highest 
protein content in the original stock, and was developed by the selection, 
for successive generations of planting, of the highest protein ears, always 
within the high-protein strain. In a similar manner, each of the other 
three strains was produced. Attention 1 has been called to some indi- 
1 Davenport, Eugene, principles op breeding, p. 447. Boston, 1907. 
7768°—17-2 
