GENETIC-VARIATION OF DEVELOPMENTAL STABILITY IN BARLEY

Developmental stability of Scandinavian barley lines and varieties over years and locations within years was studied. In 5 series of experiments highly significant interactions between genotypes and years and between genotypes and locations were found in all cases for grain yield. The interaction component of variation was split up into individual genotype variance components and ecovalences, linear regressions and deviations from linearity, and in principal components. Because of great deviations from linearity, the linear regression coefficient was in general inadequate as a stability parameter. Variance components and ecovalences are useful for ranking of genotypes, but cannot be related to special environmental factors. Principal components can divide the interactions into orthogonal components, while a related procedure called PLS can relate the components to special environmental factors. The results indicate similar genotypic stability reactions over years as between locations within years. Differences in stability can be induced by mutagenic treatments. Stability in one character is not always associated with stability in other characters. There was no close genotypic relationship between average grain yield over a range of environments and stability.