Studies of overtime have pointed to fatigue as a potential factor producing, for example, a three-fold increase in accident rate after 16 h of work, increases in back injuries, hospital outbreaks of bacterial infection, or nuclear-power plant safety compromises. Fatigue has been measured more directly in studies of scheduled long workshifts, where performance decrements in both work-related tasks and laboratory-type behavioural tests have been observed, and significant loss of sleep and increases in subjective sleepiness have been reported. Analyses of accidents or injuries during scheduled extended workshifts, however, have produced equivocal results. Factors which could compound the fatiguing effects of extended workshifts, such as workload, noise, chemical exposure, or duties and responsibilities outside of the workplace, rarely have been studied systematically. It is concluded that extended workshift schedules should be instituted cautiously and evaluated carefully, with appropriate attention given to staffing levels, workload, job rotation, environmental exposures, emergency contingencies, rest breaks, commuting time, and social or domestic responsibilities.