Methodology to estimate the soil heterogeneity coefficient, the number of repetitions and the optimum plot size for common bean (Phaseolus vulgaris L.) research

摘要

With the aim to estimate the coefficient of soil heterogeneity (b) without performing tests of uniformity, we proposed a methodology to use data from yield experiments insolating the treatment effect of the response variable. To identify methodological issues and illustrate the statistical proceeding management, data from one of the yield trials conducted in common bean by the Andean Breeding Program of CIAT were used. The coefficient of soil heterogeneity was estimated using the law of variance of Smith (1938) and the equation proposed by Federer (1963). Values of 0.59 and 0.66 respectively were obtained. Finally, we used the methodology of Hatheway (1961) and the ?“b?” value estimated based on Federer (1963) to find the best combination of plot size, number of repetitions and difference to be detected as a mean percentage.