This study examined whether prolonged exercise employing upper or lower body muscle groups led to significant alterations in three differentiated ratings of perceived exertion (RPE). Multiple regression analyses were used to identify those physiological responses which accounted for the greatest variability in these three RPE. Nine volunteer males performed 60 min of arm crank and cycle exercise at similar absolute and at similar relative exercise intensities. There were no significant differences (P greater than 0.05) between arm and leg exercise for oxygen uptake (VO2) during the absolute tests (approximately 1.60 1 . min-1) or during the relative tests (approximately 60% peak VO2). The RPE included local RPE (muscle and joint exertion), central RPE (ventilatory and circulatory exertion), and overall RPE. During the absolute tests, the final means for all three RPE were lower (P less than 0.05) for leg than arm exercise. No differences (P greater than 0.05) were found during the relative tests between arm and leg exercise for any of the three RPE. Local RPE was generally higher than central RPE. The various physiological responses accounted for more (total) variance in all three RPE for arm than leg exercise (absolute and relative arm exercise: median R2 = 0.99; absolute and relative leg exercise: median R2 = 0.75). Lactate and the ventilatory equivalent of oxygen made the greatest contribution to R2. These data suggest that RPE may be more closely related to relative exercise intensity, and perceptual cues may be more readily monitored from smaller muscle masses such as the upper body. Further, steady-state RPE was not attained during prolonged upper body exercise.