1 The effects of 1.0 mg and 2.5 mg lorazepam on learning performance were examined in a double-blind cross-over study using student volunteers. 2 Test conditions were manipulated to prevent rehearsal and to vary the subjects' motivation to perform well. Self-ratings of alertness, motivation to perform well and state anxiety were obtained prior to each test. 3 Performance in arithmetic tasks of varying difficulty was also studied. 4 Lorazepam produced dose-related deficits in verbal and nonsense-syllable learning tasks. A greater proportion of errors in the number of problems attempted in the arithmetic tests reflected an impairment in cognitive function. Lorazepam reduced the number of arithmetic problems that were correctly solved as well as increasing the percentage of errors in the problems attempted. 5 Lorazepam did not significantly decrease motivation to perform well and the lorazepam impairment was found even when the test conditions were manipulated so as to prevent rehearsal. Therefore the learning deficits cannot be explained solely by changes in motivation or impairments in rehearsal. 6 Performance in the learning tasks correlated with ratings of alertness and therefore the deficits observed after administration of lorazepam seem likely to result from the non-specific sedative effect of the drug.