Seleção multivariada entre e dentro de populações mutantes de feijão

Induction of mutation is a strategy widely used in plant breeding programs, providing the creation of genetic variability for many characteristics of agronomic importance. The aim of this study was to select promising genotypes within mutant populations of agronomically promising beans ( Phaseolus vulgaris L.), in the agricultural year 2009/10 and 2010/11, in Lages, SC. Forty mutant bean populations derived from 60 Co gamma irradiation at doses of 100 and 200 Grays were evaluated. The genotypes subjected to induced mutation were Perola, Iapar_81, IPR Uirapuru and IPR Chopim. This study used a randomized block design, with three replicates. Multivariate analysis of variance and canonical discriminant analysis were performed to explore the hypotheses. Multivariate contrasts between the mutant populations and their original populations were conducted to test the hypotheses. Multivariate analysis of variance showed significant effect of mutant populations according to crop year. Variability was also detected within the mutant population. These results show the efficiency of the mutagen in creating variability. The PMP_100 and PMC_200 mutant populations were selected due to showing a promising performance for the characteristics stem diameter, thousand-grain mass and grain yield. Within the selected mutant populations was detected genetic variability that can contribute to effective selection. The PMP_100 and PMC_200 populations must be conducted and further selected plant-by-plant to obtain promising genotypes.