An operant version of the classical delayed alternation task is presented and applied to evaluate the effects of bilateral prefrontal and striatal lesions in rats. Retractable levers in a conventional operant chamber control discrete trial opportunities for making sequential choice responses to the two sides, and the rats are required to maintain repeated nose poke responses to a central panel during the delay interval, which is randomly varied. The operant task provides measures of the speed and accuracy of response alternation and side bias; analysis at different delay intervals provides an index of the memory demands of accurate performance; and analysis of accuracy depending on the response on preceding trials provides measures of proactive interference and perseveration. Following pretraining in the task contingencies, both striatal and prefrontal lesions induced profound deficits in task accuracy, with no change in side bias and only small changes in movement times. The deficit in the prefrontal lesion group recovered more rapidly, neither group showed any change in sensitivity to proactive interference, while the rats with striatal lesions alone exhibited an increased tendency to perseverate incorrect responses on either side. We conclude that the operant delayed alternation task should assist analysis of fronto-striatal function in rats as well as be useful for the analysis of strategies for fronto-striatal repair.