1. We evaluated the feasibility of using a simulated public speaking (SPS) test to assess the activity of anxiolytic drugs. SPS was achieved by requesting subjects to present a speech to an audiocassette recorder. Thirty volunteers were randomly assigned to one of three groups treated with 10 mg diazepam, 10 mg buspirone or placebo, under double-blind conditions. One h after drug administration, subjective states were measured by the Visual Analogue Mood Scale (VAMS), the State-Trait Anxiety Inventory (STAI) and by a Bodily Symptoms Scale (BSS). Heart rate and blood pressure were also recorded. 2. SPS induced both physiological and subjective changes characteristic of anxiety. Moreover, diazepam attenuated experimentally induced increases in excitement (as measured by VAMS) and agitation (as measured by BSS). Therefore, SPS using an audiocassette recorder is sensitive to a prototypical anxiolytic and may thus be a useful test for evaluating putative anxiolytics. 3. No effect was observed with the new anxiolytic drug buspirone. However, the present negative result may be explained by clinical data indicating that patients may experience a longer lag period before the onset of the anxiolytic effect of buspirone.