Parametric and nonparametric measures of phenotypic stability in linseed (Linum usitatissimum L.)

A combined analysis with three parametricand two nonparametric measures to assess G × E interactions and stability analyses toidentify stable genotypes of linseed across18 environments in Ethiopia wereundertaken. The combined analysis ofvariance for environments (E), genotypes(G) and G × E interaction was highlysignificant (p<0.01), suggestingdifferential responses of the genotypes andthe need for stability analysis. Theparametric stability measures ofcoefficient of variability and thestability variance showed that R12-N10D wasthe most stable genotype, whereascultivars' superiority measure indicatedChilalo to be the most stable cultivar.Like most of the parametric methods, thenon-parametric measures revealed thatR12-N10D had the smallest changes in ranksand thus was the most stable genotype incontrast to R12-D24C, which was unstableand the lowest yielder. A comparison of thefive stability measures showed that thecoefficient of variability, stabilityvariance and variance of ranks were similarin assessing the relative stability of thegenotypes, whereas cultivars' superioritymeasure deviated from the others. Thestability variance and variance of rankswere significantly rank correlated, andwere the best in determining thecomparative stability of linseed genotypes.The coefficient of variability was alsorelatively better than the cultivar'ssuperiority measure. Further studies ofrepeatability tests are, however, needed todetermine the best methods. The stabilitystatistics generally identified R12-N10D,followed by Chilalo, as the most stablevarieties, whereas R12-D24C and R11-M20Gwere the least stable varieties.