GENETIC CHARACTERIZATION OF PROMISING PROSOMILLET LINES USING CLUSTER AND PRINCIPAL COMPONENT ANALYSES

Abstract

Prosomillet is important rainfed crop among small millets for which this study was taken to characterize 18 promising prosomillet genotypes using multivariate analysis. High variability observed for most of the characters indicated the scope of improvement of these characters by direct selection. Phenotypic correlation between grain yield and panicle length was highly significant and positively associated. Similarly, fodder yield was also highly significant and positively associated with days to maturity and negatively associated to grain yield. The principal component analysis revealed that the first five components with eigen value >0.40 contributed about 95.47% of total variability. The characters including grain yield, fodder yield, days to maturity, panicle length, number of tillers plant-1, plant height and days to 50% flowering were the most important traits contributing for the overall variability. Cluster analysis grouped 18 genotypes into three different clusters through multivariate hierarchical clustering. Cluster I, II and III formed distinct clusters and genotypes DHP2181, GPUP 23, TNPM 230, DhPrMV 2769, DhPrMV 2164, DhPrMV 2721, DHP 2780, GPUP 24, TNPM 228, TNAU 145, DhPrMV 2710 and TNAU 151 found to be more distinct. Recombination breeding using these parents provides segregants with rare combination of traits of interest towards maximizing grain and fodder yield potential.