O objetivo deste trabalho foi avaliar a eficiência da seleção recorrente fenotípica em relação ao número de dias para o início do florescimento em feijoeiro e verificar se a seleção afeta outros caracteres. Utilizou-se uma população proveniente da mistura das populações F2 (S0) de 11 combinações híbridas, avaliadas para a precocidade. As plantas S0 (ciclo 0), que floresceram mais precocemente, foram intercruzadas para a obtenção do ciclo I. As sementes obtidas foram utilizadas para a continuidade da seleção, e o processo foi repetido até o quinto ciclo seletivo. O progresso genético foi avaliado com cinqüenta e três progênies, provenientes da mistura das sementes das plantas S0 após uma autofecundação (S1:2) de cada ciclo. Os caracteres avaliados foram: número de dias para o florescimento, severidade da mancha-angular (Phaeoisariopsis griseola), número de dias para a maturação, produtividade e tipo de grãos. Constatou-se que o progresso com a seleção foi de 2,2% ao ano, o que indica que a seleção recorrente fenotípica foi efetiva em reduzir o número de dias para o florescimento. Não houve resposta correlacionada à seleção quanto ao número de dias para o florescimento nos demais caracteres avaliados, do que se depreende ser possível a seleção de progênies que associem florescimento precoce à expressão fenotípica dos demais caracteres, conforme o interesse dos melhoristas.

The objective of this work was to evaluate the efficiency of phenotypic recurrent selection for the number of days until blooming in common bean, and to verify if this selection affected other traits. One S0 population derived from the mixture of the F2 populations of 11 hybrid combinations evaluated for earliness was used. The plants of the S0 (cycle 0) which bloomed first were crossed to obtain cycle I. To go on with the selection the obtained seeds were used. The process was repeated until the fifth selection cycle. Fifty-three S1:2 progenies of each cycle were used for evaluation of the genetic progress. The number of days to flowering, severity of angular leaf spot (Phaeoisariopsis griseola), number of days to maturation, yield and grain type were evaluated. A selection gain of 2.2% per year was observed, which indicates that phenotypic recurrent selection reduced the number of days to flowering. The selection for the number of days to flowering had no correlated response with other evaluated traits, indicating that progeny selection for early flowering can be associated with the phenotypic expression of other traits, according to the breeder's interest.

The efficiency of recurrent selection was assessed in obtaining common bean (Phaseolus vulgaris L) plant lines resistant to the phytopathogenic fungi Phaeoisariopsis griseola. The base bean population was obtained from the partial diallel between seven lines with carioca-type grains and 10 sources of resistance to P. griseola. The plants most resistant to the pathogen were selected in the F2 (S0) generation of the populations (C-0). The best S0:1 plants that presented carioca-type grains were intercrossed to obtain cycle I (C-I). The same procedure was adopted to obtain cycles C-II to C-V. In each recurrent selection cycle, S0:1 progenies selected were also assessed in experiments carried out in Lavras, Brazil, always using as check the Carioca MG (susceptible to P. griseola) and Pérola (tolerant) cultivars. The response to selection for resistance to the pathogen was estimated from the general mean of the S0:1 progenies from each selective cycle compared to the susceptible check Carioca MG. The estimated gain was 6.4% per cycle and the indirect response in grain yield by selection for resistance to the pathogen was 8.9% per cycle. The variability detected among the progenies in the last selective cycles enabled the prediction of additional responses to recurrent selection.