In a 4-year grain rotation, on a fertile soil, the accumulations of nitrogen during the period of sod, brought about by non-symbiotic fixation, together with that supplied by the clover roots, is sufficient to meet the normal demands of the grain crops. This is true only when lime, phosphorus, and potassium are liberally supplied. Thus in 40 years the application of 1,440 pounds of nitrogen on plat 28 has resulted in a total increased yield of only 62 bushels of corn, 16 bushels of oats, 68 bushels of wheat, and 1,400 pounds of hay in excess of the PK treatment. This is true in spite of the fact that no lime was used during the 40-year period. The fact that nitrogen is not a serious limiting factor in such a rotation is further emphasized when it is realized that a total of 1,440 pounds of nitric nitrogen applied during the 10 rotations has caused total increased yields of only 12 bushels of corn, 3 bushels of oats, 27 bushels of wheat, and 70 pounds of hay in excess of an application of 480 pounds of nitrogen applied to plat 26. The results of a supplementary phosphate experiment suggest that heavier applications of superphosphate on the PK and PKN treatments would have brought about a somewhat higher level of crop yields which, in turn, would have resulted in a more economic utilization of the applied nitrogen. It is true, however, that the major limiting factor for crop production is that of moisture distribution during the growing season rather than the lack of plant food. Thus, in 1928, when ample rain fell during the summer months, plat 28 produced 82.7 bushels of corn compared to an average annual yield of 58.3 bushels.

A yield analysis of three varieties of barley, Pedigree, Velvet, and Spartan, growing in drill plats, was made by taking plants from 50 1-foot sections of drill row located at random from each of two plats of a variety. The plats were approximately 1/24 of an acre in size. The data were reduced to a plant basis. The stands within the sampled sections ranged from 2 to 27 plants per foot. Variation in stand was probably due primarily to the unequal distribution of seed by the drill in addition to other causes which affect seedling survival. Spartan produced an average of 2.34; Wisconsin Pedigree, 1.30; and Velvet, 1.19 heads per plant. This significant varietal difference in average head production per plant was considered to be inherent. Variation in soil fertility was probably not a serious factor, while differences in size of seed and thickness of stand had only slight effect. The rate of yield in bushels per acre was velvet, 46.08; Wisconsin Pedigree, 45.52; and Sparta 42.85. Yield per unit was more closely, associated with the number of heads than with the number of plants per unit area. Average yield per plant depended upon the average number of heads per plant and the average weight of grain produced per head. Average yield of grain per head was Wisconsin Pedigree, 0.99 gram; Velvet, 0.89 gram; and Spartan, 0.57 gram. Average weight of grain produced per head varied with the variety, number of heads per plant, class of tiller, and average kernel weight. In total yield for the entire area studied, Velvet was first, Wisconsin Pedigree second, and Spartan third. The order was the same for total number of plants. Spartan, while superior to the other varieties in average yield per plant, had a smaller rate of yield because of a smaller total number of plants, and for the same reason Wisconsin Pedigree yielded less than Velvet. The yield characteristics of a variety can best be determined by a study of single plants. Information derived from such a study is of value to the plant breeder in choosing varieties for crossing in an attempt to combine all of the desirable plant-yield characters. The analysis showed that if seeded at the same rate (pounds per acre) a small-seeded variety may outyield a large-seeded variety on account of the larger number of plants per unit area rather than because of superior plant-yield characters.