Two Types of GGE Biplots for Analyzing Multi‐Environment Trial Data

SA genotype main effect plus genotype × environment interaction (GGE) biplot graphically displays the genotypic main effect (G) and the genotype × environment interaction (GE) of the multienvironment trial (MET) data and facilitates visual evaluation of both the genotypes and the environments. This paper compares the merits of two types of GGE biplots in MET data analysis. The first type is constructed by the least squares solution of the sites regression model (SREG 2 ), with the first two principal components as the primary and secondary effects, respectively. The second type is constructed by Man‐del's solution for sites regression as the primary effect and the first principal component extracted from the regression residual as the secondary effect (SREG M+1 ). Results indicate that both the SREG 2 biplot and the SREG M+1 biplot are equally effective in displaying the “which‐won‐where” pattern of the MET data, although the SREG 2 biplot explains slightly more GGE variation. The SREG M+1 biplot is more desirable, however, in that it always explicitly indicates the average yield and stability of the genotypes and the discriminating ability and representativeness of the test environments.