Eight normal, healthy, male volunteers each received four triazolam doses (0, 2, 4, and 8 micrograms/kg) and four ethanol doses (0, 0.25, 0.5 and 1.0 g/kg) in a double-blind, double-dummy experiment in which within-subject dose sequence was determined by a balanced Latin square design. Triazolam and ethanol produced dose-related and time-related effects on subject ratings of mood and perceived drug effects and objective measures of memory and psychomotor performance. Dose-response curves for the two drugs were not parallel, and therefore, comparisons of the two drugs were based upon comparisons of the high dose of each drug. Although the two high-dose conditions generally were not different from one another, there were differences in their relative effect sizes, which were important. The high dose of each drug produced comparable degrees of impairment on two different psychomotor tasks. Triazolam, but not ethanol, produced significant impairment on two different memory tasks. The relative effects of each drug on subject ratings of mood and perceived drug effects varied across different subject-rated measures. Only ethanol significantly increased subject ratings of alcohol strength and feeling drunk. In comparison to ethanol, triazolam tended to produce less-pronounced subject ratings of drug effect magnitude, drug liking, and estimated performance impairment. However, less-pronounced subjective effects of triazolam were not universally observed on all subject ratings. Triazolam produced greater effects on several sedative symptoms and produced comparable effects on several mood factor scales.