The experimental driving task was designed to rule out possible effects of momentary compensation for the impairing effects of alcohol, which is suspected of biasing tests in which the subjects know the precise point or instance of measurement of response to a driving emergency. Two groups composed of ten subjects each and with targeted blood alcohol concentrations (BAC) of 50mg percent and 30 mg percent, were each administered a driving test consisting of braking and making an evasive maneuver in response to a simulated highway emergency situation. At the conclusion of two experimental sessions on the driving course--one session without alcohol and one session after the consumption of alcohol--each subject was additionally subjected to a sudden and unexpected test of response to a simulated highway emergency requiring quick evasive action. In both experimental conditions, subjects in both groups, in comparison to those with the no-alcohol condition, were significantly affected by alcohol in their responses, as measured by increased stopping distance and the greater number of pylons hit, which bordered the emergency driving course. Ten subjects under alcoholic influence who were successfully tested in the surprise driving situation 'collided' with a simulated obstacle a much greater number of times than the ten controls tested under surprise conditions. Differences in performance between ten experimental subjects whose BAC's ranged between 24 and 40 mg percent were insignificant. Although this study has demonstrated the detrimental effects of relatively low BAC's in driving situations demanding fast reactions, it is contended that the influence of alcohol in real conditions is even greater because emergency situations occur very infrequently, and because the subjects, unlike those in experimental conditions, have little chance to practice and no anticipation of being tested. Six graphs and 25 references are provided.