The objective of this study was to perform the environmental stratification in the states of Goiás and Distrito Federal for grain yield evaluation of Carioca common bean genotypes, and to compare the efficiency of different methods of environmental stratification. Twenty-five trials were carried out in a randomized block design with three replications, both in the rainy and in the winter sowing seasons, in 2003 and 2004. The trials were composed of 16 genotypes with Carioca grain type. The environmental stratification was performed through four methodologies: traditional Lin; Cruz & Castoldi's estimate of the simple percentage of the genotypes by location interactions; estimate of Pearson's correlation for pairs of locals; and the Wricke's ecovalence method. Locations identified as similar by at least three of the methodologies were considered poor-informative, and they were the following Goiás state municipalities: Anápolis, Santo Antônio de Goiás, Goiatuba and Urutaí, in the rainy/2003, rainy/2004, winter/2003 and winter/2004 growing seasons, respectively. None of the locations was identified as poor-informative in at least two seasons and, therefore, none of them was eliminated from the evaluation network of Carioca common bean genotypes. The locations used in all growing seasons (Rio Verde, GO, and Planaltina, DF) were not identified as poor-informative in none of them and, therefore, they were indicated for evaluation of Carioca common bean genotypes. The methods of Lin and Wricke were the most indicated to identify poor-informative locations
Phaseolus vulgaris; ecovalence; genotype by environment interaction; test location
