The influence of season and location on the grain of severalwheat varieties

The study throws some light on the advisability of using composite samples for the evaluation of new strains or varieties. By compositing grain from several locations, the power to measure differences between locations is lost, and the effect of location on anything being studied cannot be estimated. The question then resolves itself to whether or not these effects can be disregarded without leading to unjustified conclusions. By considering the 4-year combined analysis, it is seen that the variety-location interaction, which gives a measure of the differential response of varieties to locations, is in no case of either considerable size or great mathematical significance. The greatest effect was on yield, which is obtained by locations in any case. This study indicates that serious errors would not be introduced by using composites to rank varieties for weight per bushel, wheat ash, or wheat protein, at least if samples were not drawn from a more diverse set of environmental growing conditions. It must be remembered, however, that the varieties included in this series were all of relatively good quality and of proved adaptability to a much larger area than that covered in this study. From the point of view of the plant breeder or cereal chemist, it seems justifiable to conclude that when strains are tested for several-seasons over a range of soil types representative of the area where a new variety is to be recommended, differences in yielding capacity, weight per bushel, wheat ash, and wheat protein can be measured. The differences need not be very large to be significant. In spite of a considerable variety-season interaction for yield, superior sorts should be distinguished without too much difficulty. In the present study, 4-year averages were clearly significant in spite of marked change in rank between seasons. The standard deviations for a single plot, calculated from the triple interaction, were 3.32 bushels per acre, 0.70 pound per bushel, 0.056% ash, and 0.41% protein. The grower can feel assured that when new varieties are released after several years of intensive testing at representative points in the area considered, their properties have been determined with considerable accuracy and the chances are slight that an undesirable sort will be recommended. The length of time necessary to reach a conclusion will be greater if some seasons vary considerably from normal. This statement assumes, of course, study of more than purely agronomic characteristics.