This study was designed to confirm the hypothetical dose-dependent effect of alprazolam on memory and to compare its effects on tests measuring different aspects of cognitive and psychomotor functioning. A secondary purpose was to compare the sensitivity of newly developed telephone tests with a standard laboratory test of memory. Twenty healthy male volunteers (aged 18-35 years) participated in a five-way double-blind crossover design. Treatments were single oral doses of alprazolam (0.25, 0.5 and 1.0 mg), placebo and lorazepam 2.0 mg. The latter was included as an internal control for demonstrating sensitivity of the method. Lorazepam significantly impaired all performance measures except two: verbal fluency and accuracy of music recognition were insensitive to drug effects. Alprazolam had clear dose-dependent effects on different aspects of memory: 0.25 mg did not affect any parameter; 0.5 mg impaired performance in a Word Learning Test and a Spatial Memory Test and also psychomotor performance; 1.0 mg additionally impaired Syntactic Reasoning, Semantic Verification, and Critical Flicker Fusion performance. Alprazolam's effects were not selective for any of the cognitive functions as measured in this study. The telephone tests were clearly less sensitive than the standard test, though sufficient for showing significant effects of alprazolam 1 mg and lorazepam 2 mg.