The objective of this work was to determine the degree of correlation among parameters that evaluate oil yield adaptability and stability in soybean. The parameters were estimated starting from data analysis resultants of the evaluation of 93 soybean lines divided in three maturation cycles (28 early, 38 semiearly and 27 intermediate). The trials were carried out in twelve environments of the municipal district of Piracicaba, São Paulo, Brazil. The following parameters were correlated: grand mean (m i), regression coefficient (b i), Wricke ecovalence (omegadi), AMMI ecovalence (omegaAi), regression deviations variance (sigma2di) and determination coefficient (R²i). The parameters were compared using the ranking Sperman correlation of the lines, according to each parameter. The correlations among m i and b i and among the parameters of stability to each other (omegadi,omegaAi, sigma2di and R²i) were positive and significant (0.53<r<0,93). The correlations between the adaptability and stability parameters were low or negatives and not significant (-0.39<r<0.38), except among b i and R²i (r = 0.49 and 0.89, for the early and intermediate lines, respectively). The stability parameters present rank orders of the lines in a similar way and, probably, they belong to the same stabillity concept. The results indicated that is difficult the simultaneous selection for oil yield adaptability and stability in soybean.
Glycine max; association measures; genotype by environment interaction; predictability
