Simulated populations were used to verify the effect of errors in the selection of males in the estimates of genetic gains along ten generations of selection, using BLUP. Five levels of errors in the selection of males (0, 5, 10, 15 and 20%) and three heritabilites (.10, .30 e .60) were used. Twenty males and 200 females were selected by BLUP. After the selection of 20 males in each generation, 1, 2, 3 and 4 sires were changed by others randomly. The phenotypic values and efficiency of BLUP with errors in relation to the BLUP without errors reduced with the increase in the error level. The decrease of the genetic gains in function of the inclusion of this type of error did not, apparently, present a relationship association with heritability of the trait.
BLUP; genetic gain; selection; simulation
