Increasing the groat-oil content (GO) of oat (Avena sativa L.) could increase the profitability of oat when grown as an oilseed crop or for high-energy feed grain. In this study, three recurrent selection regimes for increasing GO of oat were evaluated and compared with predicted responses. For Regime 1, selection for GO was practiced among S0 plants grown in the field; for Regime 2, selection occurred among S0 plants grown in the greenhouse; and for Regime 3, selection among field-grown So plants was followed by selection among and within S0:1 progenies grown in the greenhouse. Gains from selection for GO via the three regimes were evaluated in field experiments by using random lines from each selection cycle of each regime. Groat-oil contents increased significantly in Regimes 1 and 3, but not in Regime 2. Annual gains in GO were 6.04, 5.85, and 11.98 g kg-1 in Regimes 1, 2, and 3, respectively, whereas the predicted gains were 9.69, 4.92, and 11.60 g kg-1, respectively. Selection among S, plants in the field and among and within S0:1 progenies in the greenhouse both contributed to the increased GO in Regime 3. In each regime, the genetic variance of GO remained significant. Line M022-4-5 with a mean GO of 162.85 g kg-1 had the largest GO in the evaluation experiment. Selection for groat-oil content on the basis of data from greenhouse-grown oat plants can be successfully incorporated into a recurrent selection program.