The use of the number of families not superior to what may be considered as sufficient for a good representation of a certain crossing is recommended, but taking into account the relationship between cost of experiment implementation and the progress obtained from the breeding program. Thus, the aim of this work was to estimate the minimum number of half-sibling families of carrot necessary to represent the genetic characteristics of a population, as well as to identify the best families and the best plants inside families as to color features. The experiment was carried out at Embrapa Vegetable Crops, Gama-DF, during the summer of 2004. A hybrid population with 71 half-sibling families of carrot plus two standard cultivars were evaluated in a experiment arranged in a complete randomized block design with 2 replications, with plots measuring 2,0 m². Ninety days after planting, 10 plants per plot were harvested and evaluated as to xylem and phloem L* a* b* parameters. The analysis of variance was performed and the minimum number of families to represent the population as well as the gains from the selection of 10% of the best families and plants inside families were estimated. It was observed that the selection among and inside families provided similar βcarotene level increases for faboth root parts. Higher gains with the selection were obtained for the xylem characteristics. For xylem L* a* b*, both standard cultivars were surpassed by the selected families. Only families 71, 7, 61 and 20 should be selected for phloem L* a* b*. For the evaluation of xylem and phloem color parameters, 52 and 42 families are necessary, respectively.
Daucus carota L.; β-carotene; gains with selection