Ro 15-1 788, a selective 'benzodiazepine antagonist', was recently recognized as possessing either intrinsic pharmacological activity or the capability of blocking activities of one or more hypothetical endogenous ligands at the benzodiazepine receptor. A study was designed to test the hypothesis that Ro 15-1788 effects will reverse those of sleep deprivation on the vigilance performance of healthy volunteers. Eight subjects were trained and then tested in five trials at hourly intervals over a morning following normal sleep (baseline session). On the latter occasions the subjects were treated 15 min before the 1st trial with Ro 15-1788, 3 mg, i.v. and placebo according to a double-blind, counterbalanced, crossover design. Changes in the subjects' vigilance performance from baseline indicated first the expected overall effect of sleep deprivation; and secondly, a significant (p<0.05 or lower, depending upon the parameter and statistical test) interaction of treatments and trials. Placebo was followed by consistently poor performance over all trials. Drug was followed by superior performance in the first trial (nearly the same as baseline) and increasingly poorer performance on subsequent trials. The results of the trial-by-trial comparisons indicated that the significant interaction was attributable to a biphasic drug effect.