Abstract
Among the four cultivated cotton species, G. hirsutum (allotetraploid) presently holds a primary place in cultivation. Efforts to further improve this primary cotton face the constraints of its narrow genetic base due to repeated selective breeding and hence demands enrichment of diversity in the gene pool. G. arboreum (diploid species) is an invaluable genetic resource with great potential in this direction. Based on the dispersal and domestication in different directions from Indus valley, different races of G. arboreum have evolved, each having certain traits like drought and disease resistance, which the tetraploid cotton lack. Due to lack of systematic, race wise characterization of G. arboreum germplasm, it has not been explored fully. During the present study, 100 polymorphic SSR loci were used to genotype 95 accessions belonging to 6 races of G. arboreum producing 246 polymorphic alleles; mean number of effective alleles was 1.505. AMOVA showed 14 % of molecular variance among population groups, 34 % among individuals and remaining 52 % within individuals. UPGMA dendrogram, based on Nei's genetic distance, distributed the six populations in two major clusters of 3 populations each; race 'bengalense' was found more close to 'cernuum' than the others. The clustering of 95 genotypes by UPGMA tree generation as well as PCoA analysis clustered 'bengalense' genotypes into one group along with some genotypes of 'cernuum', while rest of the genotypes made separate clusters. Outcomes of this research should be helpful in identifying the genotypes for their further utilization in hybridization program to obtain high level of germplasm diversity.