Analytical modeling has been used to test assumptions on the mode of inheritance of a quantitative trait in the course of diallel crossing between pure strains that are sufficient for adequacy of a simple regression model. This model frequently proved to be adequate in analysis of numerous data on diallel crossings of wheat and maize. An allelism test for quantitative trait genes has been suggested. Computer simulation has been used to estimate the effect of random experimental errors and deviations from the suggested model.