The extent to which caffeine antagonizes alcohol-induced impairment of simulated automobile driving at the current lowest legal American limit (0.08% BrAC) was the focus of this study. Fifteen adults swallowed a capsule (0, 200, or 400 mg caffeine) then drank a beverage (0.0 or 0.6 g/kg ethanol) in a within-subject, double-blind, randomized procedure. Forty-five minutes later, participants completed a test battery of subjective effects scales, dynamic posturography, critical flicker fusion (CFF), choice reaction time (CRT), divided attention (Stroop test), and simulated driving. Alcohol alone increased ratings of 'dizzy', 'drug effect', and 'high', slowed CRT and brake latency, and increased body sway. Caffeine alone increased ratings of 'alert' and 'jittery', but did not significantly affect body sway or psychomotor performance. Both caffeine doses comparably counteracted alcohol impairment of brake latency but not CRT or body sway. Brake latency with either alcohol-caffeine combination remained significantly longer than that with placebo. Stroop and CFF performance were unaffected by any drug condition. The results suggest that caffeine may increase alertness and improve reaction time after alcohol use but will not completely counteract alcohol impairment in a driver.