Abstract

---

Yield stability is an interesting feature of today’s soybean breeding programs, due to the high annual variation in mean yield, particularly in the areas across North West of Ethiopia. Nineteen soybean (Glycine max. L Merrill) genotypes sourced from Pawe Agricultural Research Center were tested for yield stability and performance in four environments between 2014 and 2016 using various stability statistics. The experiment of each environment was laid out in a randomized complete block design with four replications. Combined analysis of variance of grain yield showed highly significant differences among genotypes and environments. Significant GEI indicated differential performance of genotypes across environments. Considering coefficient of several linear regression models, including conventional, adjusted independent and Tai models as well as deviation variance from these models, genotype G18 was the most stable genotype. Stability analysis in basis of parameters like environmental variance, coefficient of variation, stability variance, genotypic stability and Superiority index, genotypes G10 and G18 were the most stable genotypes. The result of principal component analysis of stability statistics and mean yield indicated that slope of linear regression of both conventional and independent models were useful for simultaneously selecting for high yield and stability. The plot of the first two principal components also showed that the stability statistics could be grouped as two distinct classes that corresponded to different static and dynamic concepts of stability. Finally, regarding both mean yield and most of stability characteristics, genotypes G10 and G18 were found to be the most stable genotypes. Such an outcome could be employed in the future to delineate rigorous recommendation strategies as well as to help define stability concepts for other crops.

---