Genotype x environment interactions and underlying environmental factors for winter wheat in Ontario

Many studies have dealt with the problem of varying genotype performance in different environments (GE interaction). Little is known, however, of the underlying causes of GE interaction. This study was undertaken to identify some of the environmental factors that are associated with GE interactions in winter wheat in Ontario, Canada. Grain yields were obtained from the 1992/1993 and 1993/1994 winter wheat (T riticum aestivum L. em. Thell.) recommendation trials in Ontario. In each y ear, trials were grown at nine sites. Variance components for the second-or der interaction between genotypes x sites x years were considerably larger (22% of the phenotypic variance) than the first-order interactions between genotypes x years and genotypes x sites, which, respectively, accounted for only 5% of the total variation, suggesting that the genotypes were respond ing differentially to environmental factors at the various sites. Both line ar and non-linear components of the regression of genotype grain yields on site mean grain yield were significant. Grain yield of many genotypes was a ssociated with both the lowest and the mean minimum temperature in January and with precipitation during the vegetative stage (May) in 1992/1993 and d uring the grain-filling period (July) in 1993/1994. Interaction principal c omponent analysis scores for the environment were associated with January m inimum temperature in both years although winter survival differences were only apparent in 1992/1993. The results indicate that January temperatures, together with moisture supp ly before anthesis, are associated with some of the GE interaction noted wi th winter wheat in Ontario. Both of these aspects, but in particular Januar y temperatures, should thus be taken into account when evaluating cultivar performance trials. Cultivars should be exposed to low January temperatures at some sites during the evaluation process, and weather records from eval uation sites should be examined to ensure that this has occurred.