Working memory for spatial information was studied in an eight-arm radial maze by imposing a 15 min delay following an injection of either saline (0.9%) or 10% alcohol (1.25 g/kg) between an animal's fourth and fifth arm choices. Exposure to intoxicating doses of alcohol did not significantly influence the number of errors for the first 4 arm choices, but did increase the errors on the final set of arm choices following the delay and treatment. Moreover, alcohol injections significantly increased the mean response time to complete the first 4 arm choices over the 10 days of testing, suggesting a conditioned aversion to the maze as a consequence of the noxious effects of the alcohol treatments.