Multivariate analysis and clustering of Cuphea procumbens inbred lines
15 C. procumbens inbred lines of different eco-geographical origin maintained at National Botanical Research Institute, Lucknow, India were evaluated for seed yield and its 6 component traits to assess genetic divergence among them. The standardized mean values of different traits were subjected to principal component analysis and cluster analysis was performed based on two different clustering strategies i.e. UPGMA and Wards. Ward's method that showed relatively high cophenetic correlation coefficient and significant Wilk's Lambda was identified as the best clustering solution. The first four principal components (PC) with eigenvalues >1 contributed 91.56% of variability among the inbred. First PC was related with fruits/plant and branches/plant; second PC with yield/ plant, seeds/fruit and test weight third PC with plant height; and fourth PC with days to flowering. The genotypes were grouped into five clusters and cluster II was largest with 5 genotypes followed by clusters I, III and Clusters IV, V. Cluster IV exhibited highest mean for seed yield (14.77g) followed by cluster III (14.53g) and the former incorporated inbred lines, NBCP-53 and NBCP-58 that were highly divergent among themselves and from genotypes in other clusters. The inbred in cluster IV and cluster III with good amount of genetic divergence and superior agronomic traits were identified as promising inbred to develop superior recombinants with desirable agronomic traits.