Genotype‐Environment Interactions in Two Wheat Crosses

Genotype‐environment interactions can arise from heterogeneity of variances or change in genotype ranks (or both) from one environment to another (genetic correlation < 1). The purpose of this study was to characterize the interactions for five quantitative traits in two spring wheat (Triticum aestivum L.) crosses, ‘Potam’ ✕ ‘Ingal’ (P ✕ I) and RL3147 ✕ Ingal (R ✕ I). Three F3‐derived lines in each of 48 F2 families were evaluated in the F5 (1986) and F6 (1987) generations at the University of Saskatchewan. A nested split‐plot design was used with four groups of 12 families as whole plots and 36 F3‐derived F5 or F6 lines in each group as subplots. Days to heading, plant height, kernels per spike, kernel weight, and kernel hardness were examined. Variation and covariation among‐ and within‐F2 families were significant for all traits in both crosses. Significant family ✕ year interactions were detected for all traits except kernel hardness in P ✕ I and for days to heading, kernels per spike, and kernel hardness in R ✕ I. Significant line ✕ year interactions were detected for kernel weight in hardness in R ✕ I. These interactions were primarily due to heterogeneity of variances, but not to change in genotype ranks, for all traits. In contrast to crossover interactions, these noncrossover interactions should have little impact on selection; however, the noncrossover interactions may seriously limit the estimation of linkage and epistasis from variances and covariances between relatives.