Clustering of rice mutants by different methods of analysis

Forty genotypes of rice comprising 34 mutants of Kalakeri, the parent variety, and five standard varieties were evaluated in two seuons with a view to group the entries into different clusters following various methods of analysis. Based on 13 characters, the genotypes were grouped into different clusters following D² analysis, canonical analysis, metroglyph analysis and numerical classificatory analysis. Clustering pattern in different methods indicated that out of 40 entries, 30 tall genotypes were grouped into 3 clusters in D² analysis, 6 clusters in metroglyph analysis, and 5 and 9 clusters in UPGMA method of numerical taxonomic approach at 80% and 85% phenons, respectively. The remaining 10 semidwarf strains were placed in 4 clusters by D² analysis, 6 clusters in metroglyph analysis, and 5 and 9 clusters in numerical taxonomic approach at 80% and 85% phenon levels, respectively. The study of clusters formed by different methods indicated that the numerical taxonomic approach was more potent for classificatory analysis of biological populations compared to other methods.