The purpose of the present investigation was to determine the reliability of the Borg (RPE) Scale under a variety of experimental conditions. Fourteen sedentary female subjects performed six repeated work tasks (12 in all) distributed over four experimental testing sessions and presented in a randomized order. Work tasks involved treadmill walking and jogging, cycling, and stool stepping. Work loads were presented in progressive (step-wise), oscillating, or single load submaximal intensities. During work, heart rate (HR) was recorded at regular intervals and ratings of perceived exertion (RPE) were elicited either at regular intervals (interval) or only during the final minute of work (terminal). RPE responses were found to be highly reproducible whether interval or terminal. High correlations were also found between interval and terminal RPE responses. RPE demonstrated a strong relationship with HR and work intensity. It was concluded that category ratings of perceived exertion according to the Borg (RPE) Scale offer a sensitive and reliable measure of stress encountered during work.