Genetic variation and environmental effects on agronomical and commercial quality traits in the main European market classes of dry bean

Repeated testing of diverse commercial classes of beans over time and space and selection for a minimal degree of genotype × environment interaction (GEI) is a common feature of all plant breeding programs. The GEI effect limits the accuracy of yield estimates and complicates the identification of specific genotypes suited for specific environments. The purpose of this work was to study GEI on yield of the main European dry bean market classes by the site regression (SREG) and multiple trait data by genotype–trait (GT) methods, which graphically displayed the interrelationships among traits and facilitated visual comparison of genotypes. Sixty-seven genotypes of common bean, grown in three different sites in northwestern Spain during the 2001 and 2002 growing seasons, were evaluated for yield, two phenological and four commercial seed traits. Interactions between GEI and yield were established using a SREG analysis model to generate a genotype–GEI (GGEI) biplot. The GGEI biplot revealed GEI as a major source of bean yield variation and the different growing sites served to discriminate among the genotypes. This method provided information on the three growing sites: Lugo was identified as the location that best represents the target environment for seed yield; Pontevedra was the location showing the greatest yield stability and León separated the genotypes clearly although as this was not consistent over other sites, León was not representative of an average environment. Each site was represented by a group of genotypes, which showed a superior performance. Large-seed genotypes of the favada market class were best suited to the Lugo site. Commercial seed traits (seed coat fraction, water absorption, crude protein content and seed weight) and days to maturity showed wide variation, as indicated by the relative length of their vectors in the GT plot. Genotypes with the highest yield showed the highest protein content and the poorest seed coat quality and were the latest to flower, while the genotypes that exhibited a high seed coat fraction had the poorest water absorption capacity. The results presented in this work permitted the identification of optimal adapted dry bean genotypes for each bean producing area. These high-yielding genotypes with a good commercial seed quality merit special attention as they could have potential applications for the development of breeding strategies.