Proc. Am. Soc. of Agronomy 1: 131-135 (1910)
Experiment Station, Pullman, Wash.
Presented at the Omaha Meeting, 1909.

That the chemical composition of wheat varies extremely has long been recognized. The causes for these variations have been the subject of much conjecture and study. The, general conclusion seems to have been that wheat is of all the cereals the one most affected by changes in its environment, but that the particular features of environmental conditions which have the greatest influence in causing variations in composition have not yet been determined. An observation that high-nitrogen wheat grew, as a rule, in those States or sections where the soil was comparatively rich in nitrogen led to the belief by some investigators that the chemical composition of the soil has a close and definite relation to the quality of the grain growing upon it. Later researches, however, not only fail to confirm this view, but cast considerable doubt upon its accuracy.

The chemical composition of wheat, in its relation to other properties of the grain, and the causes for variations in it are now under investigation in our chemical laboratory at the Washington Experiment Station. My remarks today are in the nature of a report of progress of these investigations. The general problem is by no means solved as yet, but some positive evidence has been secured which appears to me to point to pretty definite conclusions in the near future.

As a first proposition, I may say that there are easily to he recognized certain fairly regular and constant differences in composition between different varieties of wheat. In any given wheat-growing section, certain varieties are very generally recognized as "harder," or as yielding flour of greater "strength" than others. Such differences have been accepted, even by careful scientists, as definite variety characteristics, and hence not to be explained by the same causes which produce observed differences between separate samples of the same variety. My observations lead me to believe, however, that this is a mistaken idea, and that the so-called "variety-differences" are really to be attributed largely, if not solely, to the same causes which produce variations within any given variety. Different varieties have, of course, different habits of growth (for example, some are fall and others spring varieties, some mature earlier than others under the same conditions, etc., etc.), and I believe that these different habits of growth permit different environmental conditions during the development or ripening of the seed which lead to variations in composition of the grain that are really due to exactly the same fundamental causes as are changes in composition within a single variety. To illustrate this point and at the same time to call attention to what I believe to be the fundamental cause for all changes in composition of the wheat kernel, I wish to cite the following facts: We are engaged at the Washington Station in a series of selections of high and low nitrogen wheats for breeding purposes. In these selections we are using four varieties, two of them, Jones' winter Fife and Red Russian, being fall wheats, and the other two, Bluestem and Little Club, spring varieties. It is a matter of common knowledge that in any wheat-growing section the fall-sown grains almost invariably average lower in percentage of nitrogen or protein and richer in starch than the spring-sown varieties. Spring varieties under all normal conditions blossom at a later period in the summer and ripen up more rapidly than the winter wheats which blossom several days or weeks earlier. This phenomenon is in harmony with the view recently advanced by Hall and others that ripening is essentially a process of desiccation and that conditions which favor rapid evaporation or desiccation result, therefore, in rapid ripening of grain. During the summer of 1907 we had in Washington for the first two weeks of July a period of unusually hot, dry weather, during which our fall wheats ripened up very rapidly, followed by five weeks of cool, cloudy weather, resulting in very slow ripening of our spring-sown grains, which did not fully mature until late in August. When the grains were analyzed for seed selection that season, it was found that the 320 samples from the two fall-sown varieties carried an average of 2.56 per cent nitrogen or 16 per cent protein, while an equal number of samples from the spring-sown varieties showed an average of only 1.32 per cent nitrogen or only a little over half as much protein as the average of the fall-sown wheats, and that scarcely a single sample of the spring-sown grains could be found which carried as high a percentage of protein as the lowest one among the fall-sown varieties. This was a complete reversal of the supposed "variety-characteristics" of the fall and spring grains, due apparently to no other possible cause than the unusual climatic conditions which resulted in a reverse to the usual comparative rapidity of ripening of these varieties.

In the meantime other experiments which we were conducting were giving additional evidence pointing in the same direction.

First among these was the fact, which was repeatedly observed both in our own experimental fields and in samples taken from fields in many different parts of the State, that when (as is often done in Southeastern Washington) Little Club or Bluestem wheat was sown both in the fall and in the spring in the same field, the spring-sown grain, ripening later and more rapidly, invariably carried a considerably higher percentage of protein than the fall-sown wheat grown side by side with it, thus exhibiting the same variations between fall-sown and spring-sown grain of the same varieties that are observed in comparing true fail varieties with spring varieties in the same section.

Again, we found by repeated analyses that wheat which was cut early and cured or ripened in the shock was invariably richer in protein and lower in percentage of starch than the same grain if allowed to stand until thoroughly mature before cutting. This appears to indicate that starch continues to be elaborated and transferred to the seed kernel very nearly up to the period of full maturity. This is in harmony with the observations of Fleurent and others in France some three or four years ago, that while the total protein of the plant does not increase very much after the blossoming period, starch is elaborated as long as any green coloring matter remains in the upper leaves or stems.

Further, we found that the grain growing in the hotter, drier sections of the State, where the ripening period is shorter, was invariably higher in percentage of protein than the same varieties when grown in the sections of heavier rainfall and cooler harvest weather, where ripening is slower. But when we transferred seed of any variety from one of these regions to the other, the resultant grain was not like the original seed in composition, but precisely like the grain from local seed, except that in one year when an unusually dry, hot season made our local harvest weather resemble closely that of the "dry-belt" region, the wheat grown from the seed brought from the dry belt did not carry so high a percentage of protein as that from home-grown seed, because the dry-belt seed was "acclimated" to such harvest weather and did not ripen up so fast as the strains which were accustomed to cooler and moister conditions.

Finally, in order to determine what influence, if any, variations in soil have upon the composition of the wheat growing upon it, we transferred from our "dry belt" station, where our wheats were always comparatively high in protein, to our home farm, in the district of heaviest rainfall, in the wheat-growing part of the State, where our wheats are always softer and more starchy, and vice versa, a considerable quantity of soil taken to the depth of three feet in each case. After carefully replacing the soil with respect to the first, second, and third foot layers, and digging up and replacing the untransferred soil from an adjoining plot at each station, in order to, get identical conditions of aeration, drainage, tillage, etc., each type of soil at both stations was seeded with the same variety of wheat, one-half of the plot in each case being seeded with grain grown the previous year at each of the two localities. For each of three succeeding years the same plan of seeding was followed, and the grain harvested from each half plot separately and analyzed. In no instance was there found any variation in composition greater than that normally found on adjoining plots of uniform soil. This leads us inevitably to the conclusion that the soil is not the factor which causes the differences in composition of grain which we have always found at the two stations, but that climatic influences are responsible for these differences.

We have also investigated the effect of certain different features which are parts of climatic influence, as for example, variations in composition caused by direct sunlight during ripening as compared with diffused light produced by shading parts of the rows of grain with cloths of different densities and colors. Our experiments along this line are not yet completed, but indicate that shading always results in an increased percentage of protein.

Briefly summarized, our conclusions from our work thus far are that differences in chemical composition of wheat, either as between different varieties or different samples of the same variety, are due chiefly, if not wholly, to differences in the rapidity of ripening of the grain; or that any climatic condition which results in shortening the time between the blossoming of the plant and the complete maturity of the seed produces grain having a higher percentage of protein and a correspondingly lower relative proportion of starch than will be present in the same or any other grain which ripens more slowly.

The discussion thus far has had reference to the average composition of bulk samples of grain from different fields or plots. In any such bulk samples the individual kernels show as wide, or usually much wider, variations in composition than do the averages of the lots. Some of the reasons for these variations, as related to, the part of the individual plant or spike on which the different kernels grew, were pointed out in the paper which I read before the Washington meeting of this association last year. Whether the differences which we have found to exist between the kernels growing on different parts of the same plant are due solely to difference in the rapidity of ripening of the grains in these different parts of the individual plant has not yet been determined, but is being investigated. Results of these investigations, so far as they have gone, seem to confirm the same general relationship of rapidity of ripening to composition of the grain that I have already pointed out. We feel, therefore, that we are now ready to state it to be our conclusion from the results of the first four years' work on this problem, that differences in chemical composition of wheat are due primarily to differences in the rapidity of ripening of the grain.