Capacidade combinatória e estratégias de seleção em Brachiaria ssp.

Abstract

The aims of this study were: i) to estimate the general (GCA) and specific (SCA) combining ability of apomictic and sexual parents of Brachiaria humidicola, ii) to evaluate the early selection in Brachiaria ssp. breeding program on initial and final stages of evaluation and iii) to verify the efficiency of visual selection (VS) practiced in individual plants of B. decumbens compared with the sward plots. In B. humidicola, 71 full-sib progenies were evaluated in seven harvests to obtain the GCA and SCA for apomictic and sexual parents. For Brachiaria spp., the evaluation in initial stages included 50 hybrids of B. humidicola in nine haversts, 50 and 324 hybrids of B. decumbens in six and seven harvests, respectively. While for the final stages, four potential genotypes to cultivar were compared with four cultivars of B. brizantha in two sites, as well as on the combined analysis of sites, considering the evaluation of 16 harvests. To check the efficiency of visual selection, we used the repeatability coefficient for accumulated harvests and the Spearman's rank correlation between the BLUPs of optimum and the total number of harvests based on satisfactory repeatability. For VS, 50 hybrids were selected visually in a population of 457 hybrids, which were evaluated in plots with repetitions and after this, 324 hybrid of the population were also evaluated in repetitions for agronomic traits. The efficiency of the selection (ES) and genetic gains with selection (GS) based in BLUPs were estimated. There was a predominance of genes with additive effects for agronomic traits and nutritional value in parental sexual of B. humidicola, while for apomictic the GCA and SCA were zero for almost all of the characters. Considering the early selection based on repeatability estimates in initial stages of evaluation within the Brachiaria ssp. program, 18 months are enough to select with determination coefficients above 80%, as 6 harvests in B. humidicola. For the final stages, in which a smaller number of genotypes are evaluated, one year was sufficient to select genotypes, with eight harvests. The ES was low to SV evaluated on individual plants and GS were 10 times lower than selection based on the sward plots.