Feb. 28, 1916 
IO47 
Correcting for Soil Heterogeneity 
In this field there were tested 11 commercial varieties and 12 pure- 
line varieties in quadruplicate one-fortieth acre plots. In addition, 
seven other pure lines were tested in single plots. It will be noted that 
in the lower row of the figures there are five plots not planted. In order 
to use this method of correction, it is necessary to assign values to these 
plots. The best method of doing this is to assign as the observed yield 
of each such plot the mean yield of the field. This method does not 
bias the results in either direction. 
Table I shows the average yield, both observed and corrected, for the 
four plots of each commercial variety and for the 12 pure-line varieties. 
These corrected yields have been obtained by the percentage method 
described above. 
Table I .—Variation constants for the observed and corrected average yields of commercial 
and pure-line varieties of oats tested in 1915 
COMMERCIAL VARIETIES 
Variety. 
Observed 
yield 
(bushels per 
acre). 
Standard 
deviation. 
Coefficient 
of 
variation. 
Corrected 
yield 
(bushels per 
acre). 
Standard 
deviation. 
Coefficient 
of 
variation. 
Minnesota No. 26. 
Early Pearl. 
Banner. 
Gold Rain. 
Siberian. 
Irish Victor. 
Prosperity. 
Swedish Select. 
Kherson. 
Imported Scotch. 
Senator.. 
Average. 
8r. 70*2.00 
86.61 ±2.80 
83. 28*2.03 
79- 83*1.84 
77 - 78 ±i. 4 S 
75 * 39 * 3 -06 
74 * 50*2-14 
68.34* -8o 
64.68*2. 50 
64.68* .63 
55.84*1.62 
5.94*1*41 
8.31*1*98 
6.04*1.44 
5.48*1.30 
4.33*1.03 
9.09*2.16 
6.37*1.51 
2 * 39 * • 56 
7 * 43 * 1*77 
i.88± .44 
4.81*1.14 
7.27*1.74 
9 * 59 * 2-31 
7.25*1.73 
6.86±i. 64 
5 * 57 ±i *33 
12.06*2.91 
8.55*2.05 
3 * 50 ± *83 
11.46*2. 76 
2.91* .69 
8.61 * 2.06 
82.75*1.46 
82.27*1.61 
81.77*1.77 
77*90*1.50 
77 * I 4 ±i *34 
76.19*1.80 
74.13*1.56 
68.29*1.41 
66. 79*2.10 
64. 80*1. 83 
57.92*1-11 
4.34*1.03 
4. 79 ±i. 15 
5. 26*1. 25 
4.48*1.06 
4.00* .95 
5.35*1.27 
4.65*1.10 
4.20*1.00 
6.25*1.49 
5.43*1.29 
3 * 32 * -79 
5-24*1.25 
5*82*1.39 
6. 43 * 1*53 
5 - 75 * 1-37 
5.18*1. 23 
7.02*1.68 
6. 27*1. 50 
6 . 15 * 1-47 
9.36*2. 25 
8. 38*2.01 
5 - 73 * 1*37 
73*89 
5*64 
7*63 
73*63 
4 - 73 
6.48 
PURE-LINE VARIETIES 
No. 340. 
82.77*1.65 
4.90*1.16 
5.92*1.41 
83 .SS±i *94 
5.76*1.37 
6.89*1.65 
No. 355. 
82.19*2.44 
7.24*1.72 
8-81*2.11 
83. 55*1.26 
3.76* .89 
4.50*1.07 
No. 281. 
81.31*2.78 
8.27*1.97 
10.17*2.45 
81. i6± 2.22 
6.61*1.57 
8.14* 1.94 
No. 337. 
78.83*2.27 
6.76*1.61 
8.58*2.06 
79.17*1.79 
5 . 34 ±i *27 
6. 74*1.61 
No. 247. 
80.15*2.66 
7.92*1.88 
9.88*2.16 
79.11*1.84 
5.47*1*30 
6.91*1.65 
No. 357. 
79*67*1.58 
4.70*1.12 
5.90*1.41 
77 - 58*1.16 
3.47* .82 
4.47*1*06 
No. 230. 
77.94*1.50 
4.47*1.06 
5 * 73 ±i *37 
77. 58*1.05 
3.12* .74 
4-02* .95 
No. 346. 
79.69*2.54 
7.56*1.80 
9.49*2. 28 
77.04*2.16 
6.41*1. 52 
8.32*1.99 
No. 307. 
76.94*1*30 
3.86* .92 
5.02*1. 20 
77. 00 * 1.13 
3.36* .80 
4 - 36 * I- 04 
No. 286. 
77 * 73±3*26 
9.69*2.31 
12.47*3.01 
75.65*2.86 
8.49*2.02 
11.22*2. 70 
No. 351. 
77.47* .91 
2.73* .65 
3.52* .84 
75 . 54 ±I *04 
3-io± .73 
4.10* .97 
No. 336. 
73.99*2.19 
6. 5 i±i >55 
8,79*2.11 
75 . 05 * .58 
1 - 74 * * 4 i 
2.32* .55 
Average. 
79.06 
6. 22 
7.86 
78.50 
4 * 72 
6.00 
From figure 4 it is seen that in many plots the corrected yield varies 
quite widely from the observed. However, Table I shows that when the 
four plots of each variety are averaged there are in most cases compara¬ 
tively slight differences between the two. This point is a strong argu¬ 
ment for the efficiency of four systematically repeated plots in reducing 
the experimental error. There are, however, a few cases in the table 
