THE EVALUATION OF GENOTYPE × ENVIRONMENT INTERACTIONS OF DURUM WHEAT’S YIELD USING OF THE AMMI MODEL

In order to study GE interaction and determine winning durum wheat genotypes, field experiments were conducted with 20 genotypes for 3 years (2007-2009) in Iran. Highly significant differences were found for the GE interaction, indicating the possibility of selection for the most stable genotypes. F-test Gollob determined the first five axes and FGH1 and FGH2 tests identified the first three axes as significant AMMI (additive main effect and multiplicative interaction) components. According to the test F Ratio and cross validation results, the first two axes were significant. The illustrated yearly AMMI2 biplots showed that mega-environments patterns varied across years. This could be summarized as indicating that the test locations Gachsaran (with winning G15), Khoramabad (with winning G4) and Gonbad (with winning G19) can be considered as the individual mega-environments but the behaviour of the other test locations (Moghan and Ilam) was variable across years. In any case, it seems that, regarding both mega-environments, the winning genotypes G3 and G5 could be suitable for these test locations. The AMMI stability parameters (EV, AMGE, SIPC and D) according to the first two numbers of the IPCs indicated that genotype G15, followed by genotype G4, were as the most stable genotypes, but none of these genotypes were grouped as the winning genotypes. Therefore, evaluation of genotypes based on graphic presentation of AMMI biplot appears to be superior to the AMMI stability parameters.