“…The second approach is to select independently for each trait (combining both stability and overall performance using, for example, the biplots presented in Figure 4) and then to cross the selections by controlled pollinations, trying to combine both improvements in the offspring (WRIGHT, 1976, p. 315). Nevertheless, the lack of a significant negative correlation between the stability parameters facilitate the selection based on several parameters simultaneously (DUREL and ROMAN AMAT, 1987).…”