CLASSIFICATION OF AFRICAN PLANTAIN LANDRACES AND BANANA CULTIVARS USING A PHENOTYPIC DISTANCE INDEX OF QUANTITATIVE DESCRIPTORS

Proper classification and establishment of relationships between and w ithin Musa taxonomic clusters will be important tools for the genetic improvement of plantain and banana. This paper assesses the value of a phenotypic diversity index, based on 16 quantitative descriptors, for germplasm clustering and for identification of duplicates among 92 tr iploid plantain and banana accessions. Data were recorded during the p lant and ratoon crops at Onne, a humid forest location in southeastern Nigeria. The phenotypic distance matrix was developed by calculating the average difference between each pair of accessions for all quantit ative descriptors. Significant differences were observed for this phen otypic distance index between Musa taxonomic clusters. The between-clu ster variance was larger (0.001779) than the within-cluster variance ( 0.001380). Wright's phi(FS), which measures the overall diversity, was 0.5663. This value suggested little gene flow among triploid taxonomi c clusters via pollen, which explains the higher population differenti ation exhibited by this vegetatively propagated crop with very low mal e fertility. The results also suggested that variation observed within each Musa taxonomic cluster arose from mutations accumulated througho ut the history of cultivation of this crop. Some putative duplicates b ased on qualitative descriptors were not regarded as the same accessio n according to the phenotypic diversity index based on quantitative de scriptors. Hence, gene-bank curators should assess quantitative descri ptors for the identification of duplicate accessions in Musa.