Morphology based multivariate analysis on rice [Oryza sativa]

Rice is grown in a variety of agroclimatic zones throughout the world, particularly in Asia. Several breeding programmes are undertaken to develop new varieties to respond to the rising needs of society. Since variation is a prerequisite for any breeding programme, it is exploited from the existing diversity within the crop. This study evaluated forty rice genotypes for nine yield-related attributes. An eigenvalue-based multivariate approach (PCA) was used to assess relationships among traits and to classify genotypes, while D² analysis identified the variability among them. PCA revealed nine components, of which three had eigenvalues greater than one and contributed most to the total variance. Traits such as panicle numbers per plant, tillers per plant, grains per panicle along with per plant grain yield revealed notable variability. D² analysis showed that plant height (39.10%), panicle numbers per plant (33.08%), per plant grain yield, and spikelet fertility (10.51%) contributed most to genetic divergence. The genotypes were grouped into six clusters, with cluster II being the largest, followed by cluster I. The maximum inter-cluster distance occurred between clusters V and VI, indicating substantial diversity. Hybrids derived from such diverse clusters are likely to exhibit substantial heterosis and offer useful segregants for future breeding programs.