The aim of this work was to evaluate the viability of using the augmented block designs and spatial analysis methods for early-stage selection in breeding programs for common bean (Phaseolus vulgaris L.). A total of 121 S2 progenies were evaluated in three locations: Lavras, Lambari and Patos de Minas, Brazil. A simple 11 x 11 lattice design was employed by location. Additionally, two controls, Carioca and EMGOPA 201-Ouro, were included. The split plots were composed by two 2 m lines with 15 seeds/m. The data of the production (g/plot) were submitted to a variance analysis, considering the following strategies: lattice design, augmented block design, randomized complete block design, Papadakis' method, moving means method, and check plots method. The comparison between the different strategies was done considering their efficiency in controlling the experimental error and in relation to the precision of the estimates of genetic and phenotypic parameters obtained in each method. General results indicate that augmented block design is useful in earlier stage of selection, when the intensities of selection are moderate to low; however, this design is not useful for estimation of genetic and phenotypic parameters due to lower precision of estimates; the use of check plots method did not improve the experimental precision; the nearest neighbour, Papadakis and moving means methods were efficient in removing the heterogeneity within blocks; this efficiency was equivalent to lattice analysis for estimation of genetic and phenotypic parameters.
breeding methods; genetic parameters; agronomic characters; statistical analysis; plant breeding; Phaseolus vulgaris
