STUDIES ON GENETIC DIVERGENCE OF SEEDED BANANA IN WEST BENGAL

Abstract

Thirty four (34) types of seeded banana with B genome collected from different agroclimatic zones of West Bengal, India were characterised at the Horticultural Research Station, West Bengal, India during the years 2005-2008 by using 123 plant morphological characters with different multivariate techniques. The proximity matrix, both by squared Euclidean and cophenetic correlation, between types indicated high closeness/similarity among ‘Attiakala’, ‘Bichkela-1’, ‘Bichkela-2’ and ‘Hill Banana’. The highest proximity value of 20.62 with ‘Kalyani Local-3’ showed maximum dissimilarity with ‘Maricha’ and ‘Jhama Diara’. A dendrogram using the single linkage clustering technique on squared Euclidean distance matrix and cophenetic correlation matrix showed 13 and 14 clusters, respectively. PCA was used by considering 13 factors on the basis of variance explained, i.e. more than 3%, and total explained variation was confined to 68%. Thus considering the dominant characters with positive loading under Factor 1, such as bract scar on rachis, pollen sacs colour, leaf corrugation, general fruit shape, fruits and compound tepal, the positively loaded types were ‘Bichkela-1’, ‘Hill Banana’, ‘Attiakala’, ‘Bichkela-2’, ‘Kalyani Local-1’ and ‘Jhama Daira’. Factor 1 thus explained 14.21% of the total variance. In general, PCA results agreed with the results obtained by cluster analysis. PCA gave a better picture of the relationship between seeded banana types than cluster analysis and was useful in confirming group(s) obtained through cluster analysis. MDS group plots also clearly indicated the clustering of homogeneous types. Among the 34 seeded banana types, 32 were assessed as parthenocarpic. The two non-parthenocarpic types identified were ‘Baruipur’ and ‘Seed Banana-15’.