The objective of this work was to compare criteria for selection of sunflower genotypes based on general mean obtained in different locations and its decomposition in favorable and unfavorable environments, as well as other methods of adaptability and stability, such as Eberhart & Russell, Lin & Binns, Carneiro and Carvalho et al. Data from 1999 to 2004 of the Official Sunflower Trials Network, coordinated by Embrapa Soja, were analyzed. The characters evaluated were yield of grains and oil (kg ha-1). The analysis of the decomposition of the general mean in favorable and unfavorable environment means (Indication Method - Partitioning of General Mean) was the most suitable for the indication of genotypes. However, the regression analysis contributed with additional information, indicating responsivity and previsibility of the genotypes to environmental changes.
Helianthus annuus; genotypes x environments interaction; genetic improvement; varieties recommendation
