Diallel analysis of the number of grains per ear in spring barley

Summary

Diallel analysis was applied to untransformed data for number of grains per ear in spring barley. Heterogeneity of variances was detected when the 13-parent diallel containing 2-row and 6-row genotypes, was analysed. This heterogeneity was found to be much reduced when the two subsets, containing only 2-row and only 6-row genotypes, were analysed. Analyses of variance from 9-parent full diallels at F₁ for two seasons indicated significant differences between reciprocal crosses. Large and significant items for additive and non-additive genetic variance were detected. Significant effects for additive and non-additive variation were also detected when data from 13-parent half-diallels at F₁ and F₂ were analysed in both seasons. Graphical analysis of the 13-parent diallels revealed a high level of dominance with arrays of 2-row parents showing most dominance. Analysis of the 2-row populations revealed complete dominance acting in the direction of greater number of grains per ear. Tests for epistasis mostly failed to detect it but those interactions which did reach significance were generally of the duplicate type. Estimation of genetic components confirmed that a high level of dominance was operating in the inheritance of this character. Rather small effects for general and specific combining ability both for high and low numbers of grains per ear were detected, but significant positive correlations were found between the g.c.a. effects and the corresponding parental expression of the character. Significant interactions were detected between additive and non-additive effects at F₁ and years in analyses conducted over both seasons.