The effects of various oral doses (1, 2, 4 mg) of flunitrazepam on vigilance, attention, immediate memory, short-term memory, learning, non-consolidated and consolidated long-term memory were determined. Twelve healthy young male volunteers were given placebo or flunitrazepam in a double-blind, random latin-square sequence, crossing over every 2 weeks. Volunteers completed a battery of tests at night, 3.5 h after drug administration, and in the morning, 10 h after drug administration. Flunitrazepam 1 mg did not significantly impair any of the functions tested at night, while 4 mg impaired vigilance, attention, immediate memory, short-term verbal memory and learning. The impairments of immediate and short-term memory seem to be related and proportional to reductions in vigilance and attention. Doses of 2 mg and 4 mg impaired the speed of learning but did not decrease the amount of material learned. Flunitrazepam caused dose-related impairment of long-term memory, both consolidated and not. This reduction of long-term memory does not seem to be related to the impairments of vigilance, attention or learning. The lowest dose did not modify vigilance and learning in any subject, improved attention in half of the subjects but reduced long-term memory in a similar number of subjects. Therefore, our results indicate selective impairment of long-term memory. Since there were no differences between the effects on consolidated and non-consolidated memory, the amnesic effect of flunitrazepam seems to be due to a decrease in the storage of memory traces. There were no clear generalized residual effects in the morning after administration.