Proc. Am. Soc. of Agronomy 1: 98-104 (1910)
Ontario Agricultural College, Guelph, Canada.
Presented at the Ithaca Meeting, 1908.

It has been truly stated that good seed is at the very foundation of good farming. The question of seeds is both a comprehensive and a complicated problem, which is worthy of careful study at the various institutions organized for the purpose of investigation. Various experiments have been conducted on this subject at different experiment stations. In order to properly comprehend the results of these various experiments, it is necessary to have a comparatively full understanding of the methods used, as well as a knowledge of the conditions under which the experiments were conducted. I wish it clearly understood that this paper deals simply with one phase of this question, viz., the relationship between the sizes of the seeds and the yields of the plants produced from these seeds. I intend to confine myself almost entirely to the work which has been accomplished in this particular line at the Agricultural College at Guelph, Ontario.

Within the past fourteen years a large amount of experimental work has been done to determine the influence of different selections of seed upon the resulting crop. From the numerous experiments which have been conducted at our college on seed selection, only those bearing directly on the size of the seed sown are here presented. The various experiments include grains, potatoes, field roots, and rape.

One Year's Influence of Seed Selection.

For the grain experiments, fresh seed was taken each year from the general crop of grain grown in the large fields, and for the experiments with field roots and rape, seed of the leading varieties, which was obtained from prominent seedsmen, was made use of. It. will therefore he understood that the results represent simply the one year's influence from seed selection, but in order to ascertain the influence from one year's work of this selection the experiments were repeated from season to season in order to secure a good average of conditions of soil, temperature, and rainfall. For the large sample, none but well developed seeds were selected; for the medium sized sample, the seeds selected were of a uniform character, plump, and of medium size; and for the small sample, none but sound, plump, and apparently good seeds of small size were used.

The selections were made with great care by the use of sieves and then by hand-picking the seeds, so that all the seeds of each selection would be very uniform and true to its particular selection.

Grain Crops.— In the selection of large, plump grain, a sufficient quantity of each class to sow a plot twenty-five links square was carefully weighed out, after which the grains were counted. A corresponding number was then taken of the medium-sized seeds of oats and of small, plump seeds of oats, barley, spring wheat, winter wheat, and field peas. The different selections were sown upon plots of similar size.

The following table gives the average results obtained from the various selections of grain crops sown for six, seven, and eight years in succession:

TABLE I.— Results of Seed Selection in Grain Crops.

The average results here presented in tabulated form show that in every instance the largest seed produced the greatest yield of grain. In no case did the smaller seed produce a greater yield of straw or a greater weight of grain per measured bushel than the seed of a larger size, but in three instances the results were similar.

On making an examination of the detailed results, it was found that in forty separate tests made with seed grain of different sizes, the greatest yield of grain per acre was produced from the larger seeds in thirty-seven of the tests, from the smaller seeds in two of the tests, and the yields were equal in one test.

Root Crops.— Four tests were made annually with the different selections of seed of field roots. Duplicate experiments were conducted in which the seeds of the different selections were planted separately, and duplicate experiments were also conducted by planting three large, five medium, and eight small seeds at each place where it was desirable for a root to grow. The plants were afterwards thinned, leaving one plant in each place and having the plants of the different selections of each class at an equal distance apart. When the roots were harvested the yields of the duplicate tests of each method were averaged.

The following table gives the average results of the duplicate tests made by means of each of two methods of comparison, in order to ascertain the amount of the influence of the size of root seeds on the size of the roots produced:

TABLE II.— Results of Selection of Seed of Root Crops.

In the foregoing table there are no less than ninety-four distinct experiments conducted with large, medium-sized, and small-sized seeds of five distinct classes of roots. In fully eighty-five per cent, of the separate tests the larger seeds yielded more heavily than the smaller seeds.

Field Rape.— From good average samples of commercial rape seed large, medium, and small-sized seeds were selected for experimental testing in each of five years. In each of the years, large seed was sown on one plot, medium-sized seed on another plot and small seed on still another plot, and the test was conducted two, three, or four times each season, usually the latter. For experiment, the seed was sown quite thickly in rows three and one-third links apart, and when the plants were about two inches in height they were carefully thinned to equal distances apart, thus leaving exactly the same number of plants in each plot, the object being to ascertain the comparative value of seeds of different sizes for crop production.

The average results in tons of green rape per acre obtained from the different tests in sowing rape seed of different sizes in each of five years are given in the following table:

TABLE III.— Results of Seed Selection in Rape.

In each of the years, the large seed produced the largest crop and the small seed the smallest crop, except in 1896, when the medium-sized seed and the small seed gave equal results. In the average results of all the tests conducted in five years, the large seed gave fully forty per cent, heavier yields of green rape than those produced by the small seed.

Potatoes.— In the case of Irish potatoes, selections were made of the tubers instead of the seeds. Equal numbers of large, medium, and small marketable potatoes were carefully selected and were then planted whole at equal distances apart on plots of similar size. The rows were three and one-third links apart, and the potatoes were planted one foot apart in the rows. The tests were made in duplicate in each of two and in triplicate in each of four years. Both early and late varieties were used, the greater number being late. The average results in yields of potatoes per acre from the different tests made in each of six years are given in the following table:

TABLE IV.— Results of Selecting Potato Tubers.

In each of the sixteen tests made within the six-year period the large potatoes yielded more than the medium, and the medium potatoes yielded more than the small marketable potatoes.

Continuous Selection of Seed Oats.

In 1893 an experiment was started by selecting seed from the general crop of the Joanette black oats of the previous year. The selection made in each of the following years was from the product of the selected seed of the year previous. The selections were large, plump, black seeds and light weighing and light colored seeds. The selections each year were composed of an equal number of grains, and were sown on uniform plots each twenty-five links square. Both the yields per acre and the weights of grain per measured bushel are given for each of the years from 1894 to 1906, inclusive, and also for the average of each of the four-year periods, which are as follows:

TABLE V.— Results of Seed Selection in Joanette Oats.

It will be seen that in every year the large plump seed produced a greater yield and a heavier weight per measured bushel than the light seed, and that the differences between the two selections became greater as the number of years increased in which the selections were continued.


From the numerous experiments which have been conducted at the Ontario Agricultural College within the past fourteen years, it seems very evident that large seeds will give a greater yield than an equal number of small seeds, in the case of each of at least twelve different classes of farm crops.