Thirty one commercial maize hybrids were evaluated in the southern Mato Grosso State, Brazil, 16 being regular crop (1996/97) and 15 in the late of sowing (February ending and the beginnig of March, 1997) making possible analyses of stability and adaptability for each period. Each experiment was carried out in randomized block design with three repetitions. The metodologies of EBERHART & RUSSELL (1966) (univariate analysis), LIN & BINNS (1988) (univariate analysis) and modifications by CARNEIRO (1998) (univariate and multivariate analysis) were applied. Significant effects were verified for environment, hybrids, hybrids x environment interactions and combined deviations. The majority of the hybrids showed wide adaptability and good stability according to EBERHART & RUSSELL (1966) methodology. The methodology of b1iN & BINNS (1988) modified by CARNEIRO (1998) was more discriminant than EBERHART & RUSSELL (1966) and because of that, more efficient and suitable for analysis of stability and adaptability. The multivariate analysis allowed to understand better the performance of the genotypes being an important complemental tool in the genotype indication in agreement with their performance.
stability; methodology; maize
