Transitions between tasks can produce behavioral disruptions that are characterized as problematic. Advance notice, a procedure designed to reduce disruptions, involves presenting a stimulus to warn of the end of the ongoing activity and the nature of the upcoming activity. Clinical evaluations of advance notice have produced mixed results. We studied advance notice in a controlled laboratory setting. Pigeons' keypecking was maintained on a multiple schedule with 2 fixed-ratio components. In the lean component, completing the ratio produced brief access to food; in the rich component, completing the ratio produced longer access. Disruptions in operant behavior, measured as pauses in pecking, were reliably produced in the transition from a rich component to a lean one. Advance notice was provided by flashing the houselight before transitions to lean components. Advance notice did not reduce pausing in the rich–lean transition; instead, it tended to increase it. When the flashing houselight warned of a transition but was equally likely to be followed by the rich component as the lean one, the stimulus had no reliable effect on pausing. Despite its limitations as a translational model of clinical settings, this experiment suggests that clinical use of advance notice should be approached with caution.