FINLAY AND WILKINSON MODEL VS. AMMI MODEL IN THE ANALYSIS OF GENOTYPE-ENVIRONMENT INTERACTION IN SORGHUM

To assess grain yield and stability of sorghum genotypes of (Sorghum bicolor L. Moench), 44 hybrids were sowed in 16 environments during 2001 and 2002 in the states of Tamaulipas, Nuevo Leon and Coahuila, Mexico. The genotype-environment interaction was estimated by the Finlay and Wilkinson is regression model and by the additive main effects and multiplicative interaction model (AMMI). The AMMI model was more effective for characterizing the behavior of the studied genotypes, than Finlay and Wilkinson is regression analysis. The first four principal components (ACP) of the AMMI model were significant (P < 0.01) and explained 28, 19, 10 and 9 % of the sum of squares of the interaction. In total, the AMMI model retained 75 % of the total sum squares, while the residual only represented 4 %. Thus, the AMMI model effectively explains genotype performance. In this study, the most stable sorghum hybrids were 'RB-119x435', 'Magnum', 'RB-106x25CEA', 'RB-118x430REA', 'RB-119x430CEA', 'Asgrow Coral' and 'WAC-690'. No association was observed between the most productive hybrids and the best environments.