Interação genótipos x ambientes e implicações na associação entre caracteres em soja

Abstract

The path analysis aids in understanding the relations between a set of traits and a principle variable. However, genotype x environment interactions (GE) can alter that pattern or magnitude of the correlations between quantitative traits, which can influence selection process, especially indirect selection and/or selection for multiple traits through indexes. Thus, this study aimed at measuring the effect of the GE interaction in the association between agronomic traits related to soybean productivity, concerning the direct and indirect effects obtained with the path analysis. Its implication in the selection for mu ltiple traits was also evaluated. The experiments were conducted in the agricultural years of 2014/15 and 2015/16, in three municipalities of the state of Minas Gerais, Brazil, using a completely randomized blocks design, with 35 commercial cultivars. The traits evaluated were number of grains, number of legumes, number of grains per legume, grain productivity, weight of 100 grains, lodging, plant height, height of the insertion of the first legume, and absolute maturation. Individual analysis of variance (local and agricultural year), as well as joint analysis per location and agricultural year, and general joint analysis involving all locations and agricultural years, were performed. Path analysis was done in all cases. For simultaneous trait selection, the sum of post index was employed using as importance, in the case of grain productivity, the correlation estimates and estimates of the direct effects of the explanatory traits over the main trait. Selection efficiency was calculated through the coincidenc e between cultivar rankings based on grain productivity and the selection index. The effect of GE interaction was expressive for all traits. For grain productivity, 31.7% of phenotypical variation was attributed to GE interaction. The correlation between grain productivity and the remaining traits, as well as its ramifications in direct and indirect effects, presented expressive change regarding magnitude and direction, considering individual and multi-environmental analyses, depicting the influence of GE interaction. The highest values of positive and significant correlation with grain productivity were observed for height of insertion of the first legume, and absolute maturation. The highest values of the direct effects over grain productivity were verified for absolute maturation and insertion of the first legume, as well as for weight of 100 grains and plant height. When comparing cultivar ranking with productivity and index, considering different economic importance, great similarity was verified between line classification based on productivity and on the index that uses the direct effect as economic importance. The use of the direct effect can be more efficient in selecting superior cultivars when compared to the use of simple correlation between cultivars. Thus, given the GE interaction, the employment of relative importance, as occurred with the direct effect, constitutes a recommendable strategy for the selection of multiple traits.