thickly, it tillers but little and produces but few spikes per plant If sown thinly, it stools more and the spikes are larger, ofter sufficiently to counterbalance the thin seeding.

In climates where the winters are uniformly mild, much thinner seeding may be practiced than where the winters are severe. The fact seems to be that when the winters are mild the plant largely adjusts itself to its surroundings, so that it makes but little difference how much seed is sown within reasonable limits, but when the winter is severe and the wheat partly killed, if the wheat is sown thickly there may still be wheat enough left to raise a fair crop.

The Statistician of the United States Department of Agricul ture estimates the average quantity of winter wheat sown at 1 3-8 bushels per acre, and of spring wheat at 1 1-2 bushels per acre. Professor Brewer found by means of circular letters sent to representative farmers throughout the country that the amount sown in the Middle Atlantic States was seven to nine pecks, in the Mississippi and Ohio Valleys six to eight pecks, and in California three to eight pecks, the smaller amount being used in the drier regions.

Experiments have been carried on in the experiment stations of Ohio, Kentucky, Indiana, Illinois, Kansas and Oklahoma for periods ranging from three to eleven years, aggregating thirtythree years' trials. In no case was the largest average yield at any of these stations made with less than six pecks of seed per acre, or more than eight pecks. Two stations report in favor of six pecks, one in favor of seven pecks, and three in favor of eight pecks. The Ohio Station not only reports in favor of eight pecks, but also states that with the thicker seeding the weight per bushel is greater, and consequently the quality of seed better.1 In some cases, on moderately fertile soil, better results were obtained with nine to ten pecks. In experiments of all the stations the variation in yield between five and

1 Ohio Bul. 118.

eight pecks was not usually large. In ordinary practice the ten dency seems to be to use too little rather than too much seed.

Indiana

133. Influence of Size of Seed.-Ontario Agricultural College, by selecting seed of winter and spring wheat, oats, barley and peas during five to eight years, found the average yield of grain and straw and the weight of grain per measured bushel to be in favor of large, plump seed as compared with medium-sized or small seed.1 found an average gain during three years of 2.5 bushels in favor of large seed. Kansas Station found on an average of four years a slightly higher yield from wheat with high weight per bushel.2 Nebraska Station found that large heavy seed gave much better yields than unselected seed.3 North Dakota Station concludes as the result of four years' tests that perfect grains of large size and greatest weight produce better plants than perfect grains of smaller size and weight, even if the grains come from the same spike.1 A summary of nine years' results at the Ohio Station with selected seed, second grade and unscreened seed, shows that neither the quantity nor the quality of the crop was varied by the seed used. No marked difference was obtained at Penn

5

1 Ont. Agr. Col. Expt. Farms Rpt. 1901, pp. 82-111. 2 Kan. Bul. 59, pp. 89-105.

8 Neb. Bul. 72.

4 N. Dak. Rpt. 1901, pp. 30-44

5 Ohio Bul. 29, p. 25.

000

0000 8900

4000

0000

0000

10000

00000

0000

0000

Spike of wheat grown in

New South Wales, onehalf natural size, showing relative size of grains as extracted from spikelets on one side only of the spike (After Cobb.)

sylvania Station between seed from threshing machines and that selected by hand.1 At the Tennessee Station, with two varieties, while in general the yield was in favor of the larger seed, it was not uniformly so. The evidence showed that the largest grains usually came from the largest spikes, but the seed from the largest spikes did not always give the largest yield.2 Middleton, at the University College of Wales, obtained nearly double the yield of wheat from large seed than from small seed.3 Lubanski has experimented in Russia with winter wheat, barley, oats and sugar beets, and finds the yield, and to some extent the quality, influenced in favor of large seed. Desprez, at Grignon, France, has conducted experiments with several varieties for several years, the general results being in favor of the large seed. Different weights of seed were sown with each variety, but the same weights of large and small seed were sown: thus no two plats received the same number of seeds.5 In 1900, Deherain reports from the same station but slightly better results from large seed. Cobb reports tests of various sizes of wheat grains and concludes that the superior yield from large, plump grain is sufficient to justify the cost of first-class cleaning machinery.'

The results of foreign experiments are rather uniformly in favor of large seed: some experiments showing rather striking results. A careful analysis of all American experiments appears to show that where large and small seed are obtained by the use of the ordinary fanning mill the yield has been only slightly if at all increased on account of the seed, while apparently, where greater care is taken in the selection, a moderate increase

1 Penn. Rpt. 1893, p. 112.

2 Tenn. Bul. Vol. XIV, No. 2 (1901), pp. 42-47.

8 University College of Wales Rpt. 1899, pp. 68-70.

4 Selsk. Khoz. i Lyesov. 200 (1901), Mar., pp. 611-617. (E. S. R. XIV, 432 )

5 Jour. Agr. Prat. 2 (1897), No. 37, pp. 416-420.

6 Ann. Agron. 26 (1900), No. 1, pp. 20-23. (E. S. R. XII, 233.)

Seed Wheat, pp. 1-60: Sidney, 1903.

in the yield has been obtained. In a number of experiments the influence of the number of seeds per acre has not been eliminated.

If the grains of the spikelets of wheat be designated by numbers according to the distance from the spikelet, it has been found that grains occupying the second place are the heaviest; that those in the first and the third place are about equal in weight; while grains in the fourth and the fifth place, if any, are still lighter. It is also found that of grains occupying the same relative position, those on the lower half of the spike are the larger. The following table gives results with two varieties of wheat:1

It would thus appear that small and large grains come from the same plant, varying in size because of their position, as do the grains of maize on the ear. If the plant and not the individual seed is the unit of reproduction, small seeds from productive plants will be better than large seeds from unproductive plants, provided productivity is due to heredity and not to environment, except in so far as large seeds may give the plant a more vigorous start in life. (43) It has been shown, however, that on an average, the larger spikes contain the larger grains, so that in selecting the larger grains the larger number of them would come from the larger spikes."

134.

Treatment of Seed. Before sowing, the seed should be carefully screened in a fanning mill, or wheat grader, or

1 Kurt Rumker: Jour. of Landw. 38 (1890), p. 309.

2 Seed Wheat, pp. 1-60: Sidney, 1903.

preferably both, not only to eliminate all small and unde veloped grains, but to remove weed seeds and diseased grains, if any.

If seed comes from plants that have been affected with stinking smut (149), the seed should be immersed in cold water and

seed wheat grader suitable for use by wheat growers. Wheat is sorted according to size of grains and not according to specific gravity. The screen is a cylinder of perforated sheet metal, actuated by the crank E. A brush, AA, an important feature, is held against the screen by the springs, BB. Meshes ranging from two to three millimeters may be used; where only one size is supplied, 2.5 millimeters (one-tenth inch) should be used for American wheat. (After Cobb.)

stirred, when the smut balls will rise to the surface and can be skimmed off. The seed should then be sprinkled or immersed thirty minutes in a solution of formalin mixed at the rate of fifty gallons of water to one pound of formalin (forty per cent solution of formaldehyde). Blue stone solution or hot water may be used in place of the formalin. (149) In case wheat has been affected with the loose smut the wheat may be given the modified hot water treatment. (148) It is necessary in such case to

use one-half more seed to replace seed injured by treatment. Since loose smut is usually not very destructive, it will probably be rarely advisable to resort to treatment of seed for loose smut.

135. Wheat Seeding Machinery. For broadcasting small areas, the hand grass seeder will do satisfactory work when it is not too windy. The usual horse broadcast seeder is not unlike the wheat drill, except the wheat is scattered directly from the hopper onto the surface of the ground instead of being conveyed by means of hoes underground. Standard widths are eight,