RECIPROCAL FULL-SIB RECURRENT SELECTION IN MAIZE .2. CONTRIBUTIONS OF ADDITIVE, DOMINANCE, AND GENETIC DRIFT EFFECTS

Reciprocal full-sib recurrent selection emphasizes selection for additive and nonadditive genetic effects, and direct response is measured in the population crosses. Our objectives were to determine the contributions of additive and dominance effects to the response of selection and to estimate effects of genetic drift after eight cycles of reciprocal full-sib recurrent selection (for grain yield) in maize (Zea mays L.) populations BS10 and BS11. The BS10 and BS11 populations and their crosses corresponding to Cycles 0, 2, 4, 6, and 8 were evaluated at two levels of inbreeding (F = 0 and F = 0.5) across eight environments for grain yield and moisture, stalk and root lodging, ear height, prolificacy, and days to anthesis. Population means were compared to estimate the contributions of genetic effects. The contributions of heterozygotes to the means of BS10C0 and BS11C0 were more important than homozygotes for grain yield, whereas the contributions of homozygotes were more important for grain moisture, prolificacy, standability, days to anthesis, and ear height. Significant heterosis was detected between BS10C0 and BS11C0 for grain yield. Heterosis for grain yield of the interpopulation crosses increased with selection and was caused by the accumulation of favorable alleles with additive and dominant effects from both populations and the heterozygous condition at loci for which genetic drift had caused fixation of alleles in one of the parental populations. Effects of genetic drift were significant for grain yield (BS10 and BS11), days to anthesis (BS10), and ear height (BS11). The improved grain yield of the parent populations over cycles of selection was attributed mainly to alleles with additive (BS11) and dominant effects (BS10). Means of selected BS10 and BS11 populations, adjusted for genetic drift, indicate that the difference between direct and indirect rates of response to selection for grain yield was not significant. The effects of assortative mating during progeny formation and small effective population sizes during recombination of selected progenies should be considered in long-term reciprocal full-sib recurrent selection programs, particularly if improved performance of the parent populations is desired.