The objetive of this study was to evaluate the stability and adaptability of soybean (Glycine max (L.) Merrill) cultivars in five different planting dates. The tests were composed by 21 cultivars outlined in a random blocks design, with three replicates, conduced in 1996/97 by Embrapa-Centro de Pesquisa Agroflorestal de Rondônia, in a field of Vilhena, RO, Brazil. The used methods showed accordance between the achieved results, being the behavior of cultivars in different times of sowing represented by a linear model. The coefficient of linear regression beta1i did not show any significant difference of the unit by the t test for all the tested cultivars (P<0.05). The cultivars which presented significant regression deviation sigma²(d i) (P<0.05) were MT/BR50 e EMBRAPA 31, R² = 58.56% e 74.43%, respectively. The other cultivars did not present any significant regression deviation.
Glycine max; genotype environment interaction; breeding methods
