Psychomotor effects of alprazolam (AZ) and diazepam (DZ) were compared in a controlled double-blind and cross-over trial in 24 student volunteers, who received, three times daily, placebo for the first 4 days, then an active drug (AZ) 0.25 mg or DZ 5 mg) for 7 days, and again placebo for 3 days. After a 3 week wash-out period the procedure was repeated with the comparative drug. Objective and subjective measurements were made on day 3 (placebo), on days 4 and 10 (active drug) and on days 11, 12 and 13 (follow-up placebo). Baseline levels were measured in the morning, a capsule was taken, and the tests were repeated 2 hours and 8 hours later. Blood samples were taken in the afternoon of day 10 for the bioassay of serum benzodiazepine concentrations. Single doses of both AZ 0.25 mg and DZ 5 mg showed similar degree of impairment in several objective tests. At the end of the 7-day maintenance DZ (15 mg daily) proved significantly more sedative and impaired more psychomotor performance than AZ (0.75 mg daily) did. Improvement of the complex test performances (a learning effect) was counteracted by DZ more than by AZ. No actual development of tolerance was demonstrated after either drug. During the follow-up placebo phase, DZ showed more objective residual effects than AZ. The Maddox wing test (exophoria) showed a significant DZ effect to be still present on the third post-treatment day. However, during the follow-up period subjects reported subjectively more sedation and clumsiness after AZ than after DZ, but this was not associated with impairment of objective skills.