The relationship between the level of illumination and the prescription-dispensing error rate in a high-volume Army outpatient pharmacy was investigated. The prescription error rate was determined by direct, undisguised observation and retrospective prescription review under three levels of illumination (45, 102, and 146 foot-candles) during 21 consecutive weekdays. Illumination was controlled in the prescription-checking area of the pharmacy by using additional fluorescent lamps and filters. The three levels of illumination were randomly assigned to the 21 days to provide a total of 7 days of observations per level. The final sample consisted of 10,888 prescriptions dispensed by five pharmacists. The overall prescription error rate (including both content and labeling errors) was 3.39% (369 prescriptions). An illumination level of 146 foot-candles was associated with a significantly lower error rate (2.6%) than the baseline level of 45 foot-candles (3.8%). There was a linear relationship between each pharmacist's error rate and that pharmacist's corresponding daily prescription workload for all three illumination levels. The effect of the observer was minimal. The rate of prescription-dispensing errors was associated with the level of illumination. Ergonomics can affect the performance of professional tasks.