Comparison of methods for adjustment of plot yields with unequal stands

Abstract

Five methods of analysis of experiments involving corn plot yields with unequal stand were compared. Four of them included yield adjustments: by proportion; using Zuber's (1942) formula; by analysis of covariance and the procedure proposed by Vencovsky & Cruz (1991). The fifth was made without adjustment. These methods of adjustment were applied to the data resulting from 19 experiments, coordinated by "Centro Nacional de Pesquisa de Milho e Sorgo" (National Center for Research of Corn and Sorghum) (CNPMS-Embrapa). The analysis of results allowed the following conclusions: a) no adjustment or adjustment by proportion were inadequate for experimental comparisons of the yields of the genetic material evaluated; b) the correction method proposed by Zuber (1942), even for compensation coefficient near 0.3 was not the best option for adjustment of plot yields with unequal stands; c) the method of adjustment based on analysis of covariance, and that proposed by Venconvsky & Cruz (1991), gave the best results in terms of adequacy of adjustment; d) the capacity of plants, to compensate yield reduction due to stand loss in plots, varied much from one experiment to another. Estimating this compensation is important both as an ancillary measure for the adjustment of plots yields and to provide additional genetic characterization of the material being evaluated.