Proc. Society for Horticultural Science (1905)

H. J. Eustace,
N. Y. Agricultural Experiment Station

The wide variation in the yield of adjacent hills of potatoes is a common observation, and was brought to the attention of the writer very forcibly through his experience with many experiments on potato diseases.

To illustrate, ten consecutive hills were found that yielded as follows: 16, 48, 38, 25, 32, 48, 7, 53, 37 ounces. So general and marked did this variation seem to be that it was decided to determine how much of it was due to the inherent productive capacity of the "seed."

Experiments along the line of selection have been recorded by Goff and others, but the method of selection has been different from the one used in this experiment.

By using care, the common factors that might cause variation, such as differences in the productivity of different strains of a variety, different varieties, differences in size of seed piece, variation in depth of planting, variation in the amount of fertilizer, variation in the virulence of disease and insect attacks, whole or cut tubers, partial rotting of some seed pieces, unequal tramping of horses in cultivation (a hill stepped on would be much injured) and better mechanical condition of the soil in some hills than in others, were practically eliminated and the one factor of the inherent productivity of the "seed" was studied. The question being how much can the yield be increased by using "seed" from the most productive hills?

The method of selection was as follows: from a field of potatoes where the variety was known to be pure, and in a place where the soil was uniform 100 consecutive hills were dug and weighed separately, and from this 100 were selected the 25 heaviest and the 25 lightest hills, all the others being rejected. This process was repeated until 125 heavy and 125 light hills were secured.

The tubers were stored in a good cellar during the winter and were in splendid condition in the spring. Just before planting time they were given the formalin treatment for scab, and then each lot was cut by hand as uniformly as possible to a little less than the size of a hen's egg. No whole tubers were used. After cutting, the pieces were thoroughly mixed and then 232 pieces or exactly enough to plant one row equal to 1-50 of an acre were counted out and made to weigh, by substitution, 16 pounds. The pieces were then planted very accurately 15 inches apart, at a uniform depth and each carefully placed. A chemical fertilizer at the rate of 500 lbs. per acre was applied very carefully and as uniformly as possible.

Fifteen rows were planted in this way, 10 from the heavy hills and 5 from the light hills, two of the former alternating with one of the latter.

During the growing season the plants were sprayed very thoroughly four times to protect them from diseases and insects, these sprayings were successful and served to completely eliminate these factors that might have caused a variation. There was no apparent difference in the growth or the size of the vines at any time during the season, those from heavy hills did not differ from those from light hills.

When the crop was harvested the product of each row was carefully weighed and it was found that the results were consistent throughout, that in each set of three rows each of the two rows from heavy hills outweighed the row from the light hills.

The average yield of the 10 rows from the heavy hills was at the rate of 362 bushels 15 lbs. per acre and the average yield of the 5 rows from the light hills was 339 bushels 10 lbs. per acre. The gain in favor of "seed" from the heavy hills being at the rate of 23 bushels 5 lbs. of marketable tubers per acre.

Considering the great difference in the two lots of "seed," being the extreme of the heaviest hills in contrast with the extreme of the lightest hills, this difference in yield is considerably less than was expected.

It would seem that this method of "seed" selection could hardly be recommended for ordinary field conditions, especially when it is remembered the great care used in preparing the "seed" and the planting, to eliminate the common factors of variation. However, by making a practice of continuing this method of selection with one strain of "seed" for a number of years, it is probable that the yield could be considerably increased as compared with the yields obtained by planting tubers selected at random.

An examination of the data shows the interesting fact that the average variation in the yield of adjacent hills in the experiment was 11.9 ounces or 39.18 per cent, while in the original stock from which the "seed" for the experiment was taken the average variation of adjacent hills was 9.37 ounces or 39.44 per cent. 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.