objective of this work was to identify representative and discriminatory
environments to select rice genotypes using the
Biplot GGE technique. The rice data base of the Rice Project
2001-2009 was used. Garin yield and proportion of whole
grains, individually and by the use of a selection index (yield
+ whole grains) were analyzed using Biplot GGE. Every Biplot
generated was analyzed for distance in mm. between real
localities and the ideal; distances were later standardized. In
addition, the discriminatory capacity and representativity of
each locality were determined. With the exception of Alanje,
the localities more appropriate for higher yields (Soná,
Barú), were not the same for the obtention of more whole
grains (Tonosí, Barú, Divisa). The selection index identifi ed
appropriate locations for select (Tonosí, Alanje, Calabacito,
Soná, Barú). All localities were effective in their discriminatory
capacity for yield. Differences in representativity
were observed, with Calabacito and Changuinola occupying
the highest and lowest positions, respectively. All localities
showed similar discriminatory capacity and representativity
for whole grains. Integrating yield and more whole grains it
became posible to separate more discriminatory (Remedios,
Tanara, Alanje) and more representative (Calabacito, Tonosí,
Barú) locations. The practical implication of this work is that
it allows us to prioritize research in localities more appropriate
for the identifi cation of superior genotypes.
Keywords: Rice selection, genotype x environment interaction, discriminatory and representative environments, ideal environment.