GGE Biplot Analysis of Oil Yield in Lemongrass (<i>Cymbopogon</i> spp.)

Abstract

ABSTRACTSix varieties/strains belonging to three species of Cymbopogon, viz., C. flexuosus, C. khasianus and C. pendulus, or their hybrids, were evaluated for oil yield across four years (1997-98-2000-01) as very limited information is available on variety-by-environment interaction from multi-environment trials. The main effects, varieties and years, were significant, but variety-by-year interaction was not significant. Log-transformation of data reduced coefficients of variation from 45.52% to 16.7% and increased R-square for the linear model from 70.8% to 81.9%. A traditional analysis of variance indicated that CKP25 and F2-38 were not different in oil yield and were superior to other varieties. Principal component analysis of GGE biplot methodology also revealed CKP25 to be the highest yielding genotype in all four years, followed by F2-38. Variety CKP25 could be used in a breeding program to develop new high yielding, consistent-performing varieties. Both untransformed and log-transformed data analyses revealed that 1998-99 and 2000-01 were closest to the ideal environment, i.e., they were most discriminating of genotypes and most representative of all environments. GGE biplot methodology can be used effectively in variety trials, whether or not genotype-by-environment interaction is significant.Key Words: Cymbopogon (aromatic grasses)genotype-by-environment interactionGGE biplotmulti-environment trialsoil yieldprincipal component analysis