In a large sample (N = 234), we tested effects of 24-hr of sleep deprivation on error rates in a procedural task that requires memory maintenance of task-relevant information. In the evening, participants completed the task under double-blind conditions and then either stayed awake in the lab overnight or slept at home. In the morning, participants completed the task again. Sleep-deprived participants were more likely to suffer a general breakdown in ability (or willingness) to meet a modest accuracy criterion they had met the night before. Among sleep-deprived participants who could still perform the task, error rates were elevated, and errors reflecting memory failures increased with time-on-task. The results suggest that sleep-deprived individuals should not perform procedural tasks associated with interruptions and costly errors-or, if they must, they should perform such tasks only for short periods. (PsycINFO Database Record (c) 2019 APA, all rights reserved).