Characterization of mango germplasm in North Karnataka, India: 2. cluster analysis

Numerical taxonomic approach of unweighted pair group method using arithmetic average (UPGMA) in one of SAHN (sequential, aglomerative, hierarchical and non-overlapping) techniques of clustering methods, has resulted in 11 clusters amongst 67 genotypes. The canonical roots analysis after standardization of the data matrix for 37 quantitative and 6 qualitative characters has resulted in 78.1 per cent of the variation depicted in first 3 vectors. The major contributors to the diversity were the fruit characters upto 49.0 per cent, followed by the yield components (18.4%), panicle characters (11.5%) leaf characters (8.5%), tree size and shoot characters (8.4%) and the phenological characters (4.2%). Clusters 2, 4, 3 and 10 were most homogenous, whereas cluster 9, 7 and 11 were highly heterogeneous. Cluster 7 (Neelum, Baramasi, Kalepad) and cluster 11 (Badimavu and Cowasji patel) were the most divergent, followed by cluster 10 (Dophasla, nl.him-46, Neeluddin, Local-4, KO-11, Creeping) and cluster 11, while the cluster 3 (Dashehari, Pahutan, nl.him-32, nl.him-33, Local-1, csr.nl, Nekkare-2, Nekkare 1) and cluster 10 were the least divergent.