Workers in the nuclear power industry and natural gas utility allocation stations who worked 8 and 12 hour rotating shifts were examined for the effects of self selected napping on alertness and performance. Data from subjects working the day shift were not analyzed because there very few naps observed. There were 93 subjects from the nuclear powerplant and 19 from the gas utility. Worksite evaluations of performance and sleepiness included tests of cognitive, perceptual motor, and motor skills, and scales for measuring perceived fatigue and sleepiness administered by computer. Nap frequency, sleep quality, and time of sleep was determined by questionnaire. The results indicated that the majority of naps were taken before the first day of the workweek, with the most (50%) occurring prior to the 8 hour night shift. Both quality and depth of sleep generally were higher on no nap days compared to nap days, and both were higher before the first workday compared to the rest of the week. On tests of mental performance, workers committed more errors on nap days compared to no nap days. The author concludes that the reduced mental performance and alertness on nap days were not consistent with a compensatory view of napping. Diminished alertness on nap days may be explained by changes in circadian rhythms produced by napping or poor sleep quality.