Eleven healthy men (M age = 27 years, SD = 4) completed three cycling and three walking trials in an alternating order. During each trial, participants were allowed, within 3 min, to adjust the work rate to correspond to given rating of perceived exertion (RPE) values according to the following order: RPE 11, 13, and 15. For cycling as well as walking, at each RPE there were no significant differences between mean heart rate responses across the three trials (p greater than 0.05). Mode-specific estimates for heart rate intraclass correlation coefficient and coefficient of variation ranged between 0.80 and 0.91, and 5.6% and 8.3%, respectively. This study provides absolute reliability estimates for heart rate responses when using RPE in a production format and suggests there may be RPE- (and mode) specific practice requirements for achieving a reliable heart rate response at a given RPE. (Contains 1 figure and 3 tables.)