POTATO BREEDING AND SELECTION. 25 
fall of 1903 a field of potatoes was selected the variety of which was 
known to be pure, and in a place where the soil was uniform 100 
consecutive hills were dug and weighed separately. From these 100 
hills the 25 heaviest and 25 lightest yielding hills were selected. 
This process was repeated until 125 hills of each had been secured. 
In the following spring 10 rows of 232 plants each were planted from 
the heaviest yielding hills and 5 from the light-yielding ones. The 
resultant yields averaged 362J bushels per acre from the productive 
hills and 339^ bushels from the unproductive, the gain in favor of 
seed from the heaviest yielding hills being at the rate of 23 -^ bushels 
of marketable tubers per acre. 
It was found that the amount of variation in the yields of adjacent 
hills in the 1904 crop was almost as great as in that of the original 
stock. The 1904 variations were 11.9 ounces, or 39.18 per cent, as 
against 9.37 ounces, or 39.44 per cent, in the 1903 crop. In this 
connection Eustace says: 
That the variation was not materially reduced by the uniform conditions under 
which the experiment was made was a surprise. The conclusion is that factors which 
are apparently unimportant produce wide differences in yield. 
From our present knowledge of the behavior of .individual hills of 
potatoes, it seems quite certain that Eustace would have secured 
much more uniform results had the progeny of each individual hill 
been planted separately. In mass plantings individual variations 
are obscured, rather than emphasized, as in the tuber-unit or progeny- 
row method. 
At the annual meeting of the xlmerican Breeders' Association, in 
1907, Waid * reported the results of studies which he had undertaken 
at the Ohio Agricultural Experiment Station, in which the progeny 
from productive and unproductive plants in 1903 had been carefully 
studied during 1904, 1905, and 1906. The results of this work showed 
quite clearly that with few exceptions low-yielding plants remained 
unproductive. 
The 3-year average from high and low yielding plants was found to 
be 1 .38 pounds for the former and 0.73 pound for the latter, or a differ- 
ence of over 89 per cent. A comparison of yields from the productive 
plants and plants from common stock showed a gain of over 25 per 
cent for the former. 
A further study of Waid's data reveals the fact that the average 
weight of the 10 original high-yielding hill selections was 2.38 pounds 
per plant, whereas the 3-year average of their progeny was only 1.38 
pounds per plant. This suggests that in the selection of high-yielding 
hills one is not at all certain what proportion of the hills is likely to 
maintain their seemingly productive character. 
1 Waid, C W. Results of hill selection of seed potatoes. In Amer. Breeders' Assoc, 3d Ann. Rpt., 
1907, p. 191-198. 
