This research examined the behavioral effect of alcohol during rising and declining blood alcohol concentrations (BACs) when the dose was repeated three times (mean peak BAC = 78 mg/100 ml). A total of 36 male social drinkers learned a complex psychomotor task and subsequently performed it at intervals after alcohol was received. Subjects performed under one of two conditions: an experimental (E) treatment associated drug-compensatory (nonimpaired) task performance with a positively reinforcing outcome or a control (C) treatment associated no environmental consequence with performance. E treatment diminished impairment at all positions on the BAC curve and carried over to result in progressively less impairment when the dose was repeated. In contrast, C treatment tended to increase impairment around the BAC curve with repeated doses. The rate of recovery during declining BACs remained stable across sessions and was not altered by the treatments. The results imply that impairment under a dose of alcohol is governed by two processes: response-outcome associations that determine the amount of impairment displayed under a dose and some adaptive process that determines the rate of recovery with time during exposure to a dose.