The genotype x environment interaction (GEI) has an influence on the selection and recommendation of genotypes. To this end, G x E interaction and grain yield stability study was conducted for 17 advanced Ethiopian mustard across three districts (Sinana, Adaba and Agarfa) in the highlands of Bale zone during 2014 and 2015 main cropping season. Randomized complete block design with four replications was used. The combined analysis for the mean grain yield was highly significant (p≤0.01) for genotypes, environment and genotype x environment interaction. The mean seed yield of the locations was ranged from 0.9427 t ha-1 for Agarfa to 2.645 t ha-1 for Sinana in 2014. The AMMI analysis indicated that 76.7% of the GE sum squares was justified by the first two AMMI (AMMI1 and AMMI 2) components. The regression coefficient (bi) of genotypes ranged from 0.629 to 1.345. Genotypes G7 was the most stable with optimum grain yield (2.21 t ha-1), bi- value nearer to unity (bi = 1.03) and minimum value of deviation from regression (0.12). Based on the AMMI Stability Value (ASV), G12, G10, G17, G5, G3, G2, G7, and G8 showed the lowest ASV indicating as they are most stable. However the most stable genotypes would not necessarily give the highest yield. Therefore, based on mean grain yield and the result of stability parameters such as ASV, bi and Genotypes Selection Index (GSI), genotype G7 was found the best candidate variety and recommended for possible release for the test environments and similar agro-ecologies.
Keywords: AMMI, AMMI stability value (ASV), biplot, genotypes selection index (GSI), mustard, stability.