Quantitative descriptors to estimative genetic divergence in castor bean genotypes based on multivariate analysis

Knowledge of the diversity is of fundamental importance for plant breeding, as well as for the identification of combinations that can express high heterosis levels, increase the expectation of selection of superior segregating genotypes, based on the accumulation of varibility by crosses between dissimilar genotypes. Several multivariate methods can be used to investigate diversity. The choice is made according to the accuracy desired by the researcher, the ease of the analysis and how the data were obtained. The purpose of this study was to evaluate the diversity of 11 castor bean genotypes in the north and northwest of Rio de Janeiro, Brazil and to investigate the potential of each genotype for future breeding programs. Six quantitative traits (plant height, total raceme length, length of the region with fruiting raceme, length of the region without fruiting raceme, number of fruits per raceme and total yield) were evaluated in a randomized complete block design with four replications. The differences between treatments were significant for all traits evaluated. Results of multivariate analyses (UPGMA, Tocher and canonical variables) showed similar results forming three groups each. The genotypes Nordestina, IAC 80, A1 Guarany, Paraguacu, Mirante and IAC 226 proved promising for the studied regions. For crossing or breeding programs, the combinations A1 Guarany x Nordestina, A1 Guarany x IAC 80 and Nordestina x IAC 80 are recommended.