Changes in cortisol concentration in response to exercise at 3 different intensities were quantified. Ten apparently healthy, recreationally active males participated. On 4 separate occasions, subjects were assigned a random order of 1-hour cycle ergometer bouts of exercise at 44.5 +/- 5.5%, 62.3 +/- 3.8%, and 76.0 +/- 6.0% (mean +/- SD) of VO2peak and a resting control session. Saliva samples were collected before exercise at 10, 20, 40, and 59 minutes of exercise and at 20 minutes of recovery. Differences in cortisol concentration were assessed via multivariate analysis of variance (alpha = 0.05) Tukey post hoc analysis when indicated. During the highest-intensity exercise session, cortisol was significantly higher at 59 minutes of exercise (p = 0.004) and at 20 minutes of recovery (p = 0.016) than at those same time points during the resting control session. No significant differences in cortisol concentration were noted among resting, low-, and moderate-intensity exercise. Exercise <40 minutes in duration elicited no significant differences at any intensity. These data indicate that only exercise of high intensity and long duration results in significant elevations of salivary cortisol.