The experiments described in this paper were undertaken to show the effect of continuous selection of large or small seed wheat of a small-kernelled variety on varietal purity and behavior. Purplestraw wheat was grown in 1921 between plats of Abruzzes rye and Fulcaster (Bearded Purplestraw) wheat and the resulting crop of Purplestraw seed was sorted into large and small sizes. In the following years of the experiment the large seed was sorted annually from the crop of the large-seed plats of the previous harvest and the small seed from the crop of the small-seed plats. Increased grain yields were obtained in all years but the first from the use of large wheat kernels. The gains from using large seed in comparison with small seed varied from approximately 0.3% in 1924 to 18.7% in 1926. The bushel weight of the grain produced by large seed was heavier in four of the six years of the experiment, but the differences were small and not conclusive. Large seed produced a higher percentage of large kernels than small seed in each of the five years that data were recorded. The five-year average percentage of large kernels in the large-seed plats was 35.5 as compared to 17.8% in the small-seed plats. Hybrid kernels of natural crosses between wheat and rye, being small, were sorted automatically into the small seed in 1922 and 1923, the years following the growing of the seed plats in the vicinity of rye. In the crop of 1922 there were an average of 19 F1 wheat-rye hybrids in each of the seven small-seed plats as compared to 1 wheat-rye hybrid in all seven large-seed plats. Seed from unsorted Purplestraw straw wheat produced an average of one wheat-rye hybrid to the 40th-acre plat. In 1923, an average of six wheat-rye hybrids were present in each of the small-seed plats and two hybrids in all seven large-seed plats. Extra small seed from the threshing machine taken from unsorted Purplestraw seed wheat grown near rye produced an average of about 21 wheat-rye hybrids in each of two 40th-acre plats in 1923. The Purplestraw seed from which the extra small seed was obtained produced only two wheat-rye hybrids in 110 40th-acre plats. Seeds of fully awned wheat rogues, which may be due to natural crossing between Purplestraw X Fulcaster, were sorted into the large seed, being themselves large, and the number present in the large-seed plats increased rapidly in each of the four years in which data were recorded. In 1924 there were 13 times as many bearded heads in the large-seed plats as in the small-seed plats, and in 1927 there were over 64 times as many. Unsorted Purplestraw seed in 1927 produced over twice as many bearded heads as small seed, but the large seed produced 29 times as many as the unsorted seed. Three times as many loose-smutted heads occurred in the plats sown with small seed as in the plats sown with large seed in 1926 and over five times as many in 1927. Small-seed plats contained four times as many loose-smutted heads as unsorted Purplestraw seed.

A large variety of soils occur in the Fort Collins region. These vary from a variety of alluvial deposits, residual sandstones, shales, and limestones to wind blown material. The tall grasses such as porcupine grass, sleepy grass, wild rye, as well as western wheat grass, grow well on deep loam and clay soils, often containing considerable sand or gravel, provided the moisture penetrates sufficiently deep for the growth of the roots. Such conditions occur in the valleys between the foothills and on the plains immediately adjacent to the foothills. The short grasses, grama grass, buffalo grass, and ring muhlen-bergia, are dominant on compact soils that contain available moisture during part of the growing season in only the surface 6 to 24 inches. This kind of soil is found on the plains a few miles east of the foothills. Taller grasses do not occur except in depressions where western wheat grass and porcupine grass may be found. The so-called hardpan layer was found to impede or prevent downward growth of roots on the plains in places. The depth and hardness of the cemented layer was indicated by the character and density of the vegetation. A thin stand of grama grass and ring muhlenbergia indicated shallow well-developed hardpan-like structure. Denser stands of grama with little or no muhlenbergia indicated a deeper and often more weakly cemented layer.

This paper reports studies of the germination of different varieties of three major farm crops--wheat, soybeans, and oats. The wheat varieties tested were Turkey, Red Rock, and Marquis, representatives of the three most important market classes, viz., hard red winter, soft red winter, and hard red spring, respectively. Soybean varieties tested were Manchu, Virginia, and Wilson. Only one variety of oats, the Silvermine, was included. Samples of each variety were obtained from various agricultural experiment stations and germinated on plaster of paris blocks at five constant temperatures of 10, 15, 20, 25, and 30 degrees C. Not less than 100 seeds were used in making a germination test, and every germination test was repeated. Each of the two tests was performed at different periods of time and not concomitantly. Duplicate tests made at the same time and under identically the same conditions are of value only in so far as the total number of seeds is increased. The same results would be obtained by doubling the number of seeds and performing a single germination test.

1. This paper reports 1927 results from 21 cooperative rod-row wheat trials variously located in North Dakota and also furnishes an estimate of 1927 common wheat rust losses in the state. 2. Judged by the size of the probable errors, by the rank in yield of certain wheats in comparison to what was expected of them from their previous behavior, and judged by the general consistency of the results among themselves, the outcome of these rod-row trials is considered satisfactory. 3. From 19 comparisons Garnet outyielded Marquis by a significant amount. 4. In regions I and II, in eastern North Dakota, the rust-resistant hybrid selection 1656.85 significantly outyielded all wheats except that in region I it was outyielded by Mindum (a durum) but not significantly. The excess of yield of 1656.85 over Marquis for 12 localities in the eastern section of the state was 16.0 +/- 0.64 bushels. 5. Six comparisons in eastern North Dakota between the essentially immune variety Hope and other wheats show that it yields significantly more than Marquis, not significantly different from Ceres, and significantly less than 1656.84 and 1656.85. Hope stands nearly midway between Marquis and 1656.85. 6. In western North Dakota, with less rust, 1656.85 showed less advantage over Marquis than it did farther east. 7. Judging by the behavior of 1656.85 from region to region and by Marquis and 1656.85 in the four regions, factors other than rust brought about losses equal to or surpassing losses due to rust. 8. An estimate based on available acreage and yield statistics indicated a loss to common wheat in North Dakota in 1927 of 28.6% which is equal to approximately 26 million bushels.

The influence of different sources of nitrogen on the availability of phosphate in soils from plats of the sources of nitrogen experiments of the Alabama, Rhode Island, and New Jersey Experiment Stations has been studied by the following methods: (a) Concentration of inorganic and total phosphates in the displaced solution. (b) Concentration of inorganic and total phosphates in 1:5 water extracts. (c) Concentration of inorganic phosphate in 1:5 carbonic acid extracts. In addition, the soils from the New Jersey Experiment Station were studied by other methods, as follows: (a) Amount of inorganic phosphate removed from the soils by continued leaching. (b) Amount of phosphate absorbed by the soils from a solution of mono-potassium phosphate. (c) Amount of phosphate removed from the soil by, wheat seedlings, using the Neubauer method. All methods have shown that physiologically basic nitrogen fertilizers increase and acid-forming fertilizers decrease phosphate availability, as compared with the check plats, when lime is not applied. Liming greatly increased the availability of the phosphate on all plats, and corrected the detrimental effects of the acid-forming nitrogenous fertilizers.

Various factors which affect the inoculation of soybeans were studied and the following results were secured. Considerable difference exists in the ability of various cultures of soybean bacteria to produce nodules on soybean plants. A low soil moisture content at time of planting may be an important factor in reducing nodule formation. This is probably due to the fact that the root-hairs on the main stem disappear before the bacteria have an opportunity to come in contact with them. Keeping a soil at a high moisture content after it was held at a low moisture content for 10 days after soybeans were planted produced a lower percentage of inoculated plants and fewer nodules per plant in 9 out of 13 varieties of soybeans studied. Preparing a milk suspension of soybean bacteria before adding the culture to the seed did not give any better inoculation than when a water suspension was used. The addition of superphosphate (acid phosphate) and potash to the soil used in these experiments did not affect nodule formation. The use of chlorinated water from the city water supply instead of distilled water to keep the soil at an optimum moisture content resulted in a considerable depression in nodule formation but did not completely prevent inoculation. Field and greenhouse studies on the inoculation of a large number of soybean varieties, using both pure and mixed cultures, indicate that there is considerable variation in the ability of different varieties of soybeans to resist infection by a particular strain of soybean bacteria. If the nitrogen content of plants can be used as a measure of the benefits derived from the association of soybean bacteria with the soybean plant, data are presented to show that under certain conditions uninoculated plants may contain a higher percentage of total nitrogen than inoculated plants. This would indicate that some soybean bacteria may be of less value than others from the standpoint of nitrogen fixation and the subsequent utilization of this nitrogen by the soybean plant. Well-inoculated soybean plants grown on soil treated with 2 tons of wheat straw per acre contained less nitrogen than inoculated plants grown on untreated soil. Various factors which might cause such a condition were considered.

1. Nitrate of soda applications of 100, 200, and 400 pounds per acre to replicated plats at Davis gave significantly increased yields of barley following White Durra. The average yields were 810, 1,125, and 1,495 pounds per acre, respectively,, over 1,225 pounds, the average yield of the unfertilized check plats. 2. Significant increases in yield were secured with 200- and 400-pound applications of sulfate of ammonia to barley and wheat from replicated plats in the Imperial Valley following Hegari. All 400-pound applications and some 200-pound applications more than doubled the yields of barley. In general, higher yields were secured by applying the fertilizer 35 and 67 days after planting than at planting time. 3. In triplicate pot experiments where different weights of sorghum roots, corn roots, and sucrose were added to the soil, barley decreased in yield with an increase in sorghum roots and sucrose added; nitrates in uncropped pots were progressively depressed by increased additions of sorghum roots and sucrose; and fenugreek, a legume naturally inoculated at Davis, made normal growth. A close relationship is indicated between the depression of nitrates by sorghum roots and that by an equivalent amount of sucrose. Likewise, a relationship between barley yields and nitrates thus depressed is indicated. 4. In replicated field trials barley was depressed in yield progressively by following millet, corn, broom corn, and sweet sorghum below the yield following fallow, while fenugreek made normal growth on all of the plats previously in crop, in fact yielding higher than those following fallow. 5. The theory of sorghum injury as due to a heavy draft on available essential elements cannot satisfactorily explain observations under 3 above, while the toxin theory cannot explain 1, 2, and 4. The theory of competition between micro-organisms and the crop plants for nitrogen and possibly other essential elements very satisfactorily explains all of these observations. 6. The chance of profit from the use of some rates of application of nitrogenous fertilizers where small grains normally follow sorghums seems hopeful on the basis of increased yields secured, present prices, and increased costs of production involved. Where small grains do not normally follow sorghums because of poor yield, this initial loss must be less than the profit shown for the fertilizer applications themselves to make the sequence of sorghums and fertilized small grains profitable. 7. The practically normal growth of alfalfa and fenugreek following sorghums suggests that planting a legume may be the best way to dodge "the injurious after-effects of sorghum."