Vol. 9, Special Issue 9, Part A (2025)

The present study assessed phenotypic variability among 250 M₄ generation kodo millet (Paspalum scrobiculatum L.) mutants developed through gamma irradiation, along with three checks, using multivariate approaches. The experiment was undertaken at Shaheed Gundadhoor College of Agriculture and Research Station Jagdalpur during Rabi-2024 utilizing augmented block design as layout. Principal Component Analysis (PCA) revealed four principal components explaining 71.5% of total variability, with grain yield per plant, number of grains per raceme, biological yield, and flowering traits emerging as major contributors. The first principal component (PC1) explained the largest share of variation (26.9%), mainly contributed by grain yield per plant (0.54), number of grains per raceme (0.492), and biological yield (0.46). Genetic divergence analysis grouped the mutants into ten clusters, with the largest being Cluster VI including 42 lines. The highest intra-cluster distance was observed in Cluster Ⅲ (2.886), followed by Cluster I (2.671), reflecting the existence of considerable genetic divergence within these groups, while maximum inter-cluster distance was observed between Cluster X and Cluster I, indicating the presence of highly divergent genotypes suitable for future hybridization. Regression analysis highlighted grains per raceme, biological yield, and harvest index as key predictors of grain yield per raceme. The model showed a high coefficient of determination, with grains per raceme (75.47%) and biological yield (71.09%) explaining maximum variability. These findings demonstrate the utility of multivariate approaches in identifying superior mutant lines for breeding programs.