Eight cycles of half-sib family BSK(HI) and S(,1) progeny BSK(S) recurrent selection have been completed in Krug (BSK), a strain of the open-pollinated maize (Zea mays L.) variety 'Krug Yellow Dent'. In each cycle, 100 families were evaluated and 10 selected lines were recombined for an equal selection intensity (10%) and effective population size (N(,e) (TURNEQ) N = 10). Experiments designed to evaluate and compare the changes in grain yield and other agronomically important traits in BSK included BSKCO, the C4 and C8 populations of BSK(S) and BSK(HI) per se and selfed; a diallel among those five populations and the diallel crosses selfed. Also included were the BSKCO, BSK(S)C8 and BSK(HI)C8 populations in crosses with testers used in the BSK(HI) program and the same populations in crosses with eight unrelated tester populations. All yield trials were grown at 4 Iowa locations for 2 years (8 environments). The genetic gains per cycle for grain yield in the populations per se for cycles 0-4 adjusted for the effects of finite populations size were significantly larger for the BSK(S) program. In cycles 4-8, the genetic gains (adjusted for drift) in the BSK(S) program were lower (not significantly) and higher for BSK(HI), compared with their respective rates in cycles 0-4. This resulted in a nonsignificant difference between methods for cycles 4-8. In the populations selfed and in crosses with testers used in the BSK(HI) program (averaged over three testers), the genetic gains were parallel to the gains in the populations per se, except for cycles 4-8 in the BSK(HI) program, where the gain in the population selfed was equal to the gain in cycles 0-4. In crosses with unrelated tester populations, the average response over eight cycles for grain yield was not significantly different between methods. The effects of finite population size (genetic drift) were significant for grain yield for both methods and confounded the observed response to selection, especially in cycles 4-8. The effects of drift were larger in the BSK(S) program. The genetic gains, adjusted for drift, were smaller than the predicted gains for both methods. A significant linear decrease in root lodging was observed in the BSK(S) program. This decrease was evident in all groups of entries. Changes in stalk lodging are not consistent, but a tendency for higher stalk breakage is evident for BSK(S). Percentage grain moisture at harvest increased in the BSK(HI) program, especially in cycles 4-8.