Experimental pain thresholds (electrical intracutaneous finger and dental pulp stimulation) and plasma hormone levels (beta-endorphin, cortisol, and catecholamines) were measured in ten healthy sportive men before, during, and after progressively more strenuous physical exercise. In a double-blind study conducted on two different days, 20 mg of the opioid-antagonist naloxone or placebo was administered prior to exercise. A significant pain threshold elevation was found during exercise for finger (ANOVA, P less than 0.004) and dental pulp stimulation (P less than 0.01). Pain threshold elevation was most pronounced during maximal exertion, at which time the subjects reported the greatest subjective fatigue. Thresholds remained elevated 10-15 min after the end of exercise, and, 60 min after exercise, thresholds returned to baseline values. The subjective magnitude estimation of suprathreshold stimuli was significantly reduced (P less than 0.0001) 5-10 min after exercise. Plasma beta-endorphin, cortisol, and catecholamines increased significantly (P less than 0.0005, all values) during exercise. Plasma beta-endorphin levels did not correlate significantly with pain thresholds (r = -0.37, NS). Naloxone failed to affect pain thresholds, although beta-endorphin and cortisol increased significantly more (P less than 0.02) during exercise after naloxone. It is concluded that short-term, exhaustive physical exercise can evoke a transient elevation in pain thresholds. This exercise-induced elevation in pain threshold does not, however, appear to be directly related to plasma endorphin levels.