Adaptabilidade e estabilidade de cultivares de soja RR nas regiões Sudeste, Centro-Oeste e Nordeste do Brasil

Abstract

Soy (Glycine max) is an important oleaginous widely adapted to tropical and subtropical climates of Brazil, with high protein rate. It is noted the expansion to the areas of regions from Southeast and Midwest, as well as the increasing usage in Northeast. The selection process is often used to evaluate the performance of genotypes (G) in different environments (E). Therefore, it is possible to detect a G x E interaction, due to the differential responses of genotypes to environmental variation, which can complicate the recommendation of cultivars. Thus, the objective of this study was to evaluate the magnitude of the interaction for grain productiviy and its consequence in the adaptability and phenotypic stability of soybean RR cultivars in Southeast, Midwest and Northeast Brazil, based on methodologies involving linear regression, non-parametric and multivariate analysis, aiming to a complementarity between the methods. The grain yield data (kg.ha-1) were obtained from six cultivars (BRS VALIOSA, P98Y11, P98Y51, SYN1080, SYN1279 e SYN9070) evaluated in 12 environments. The experiments were conducted in a randomized block design with three replications in season 2012/13. Analyses were performed for individual and joint variances, and, once detected the interaction, some aspects were studied. They are the decomposition of this interaction into simple parts; the adaptability and stability by methods like Traditional, Plaisted and Peterson, Wricke, Eberhart and Russell, Additive Main Effect and Multiplicative Interaction (AMMI) and Lin and Binns modified by Carneiro; and the environmental stratification by analysis of factors and AMMI. Wricke, Plaisted and Peterson, Eberhart and Russel and Lin and Binns methods were consistent with each other in identifying stable soybean cultivars grown in 12 environments. Multivariate analysis allowed grouping the environments in a similar way, in which AMMI method was higher than the model of analysis of factors, as it explains a greater proportion of the sum of squares of the interaction. It was concluded that the interaction for the soybean productivity was from a complex nature, in which the genotype SYN1279 showed superior performance and stability, according to the analysis of variance, linear regression, non-parametric and multivariate investigations.