Salivary melatonin concentration is an established marker of human circadian rhythmicity. It is thought that melatonin is relatively robust to the masking effects of exercise. Nevertheless, the extent and even the direction of exercise-related change is unclear, possibly due to between-study differences in the time of day exercise is completed. Therefore, we aimed to compare melatonin responses between morning and afternoon exercise, and explore the relationships between exercise-related changes in melatonin and heart rate. At 08:00 and 17:00 hours, seven male subjects (mean ± SD age, 27 ± 5 years) completed 30 min of cycling at 70% peak oxygen uptake followed by 30 min of rest. Light intensity was maintained at ~150 lx. Salivary melatonin (ELISA) and heart rate were measured at baseline, 15 min during exercise, immediately post-exercise and following 30 min recovery. Melatonin was ≈15 pg ml(-1) higher in the morning trials compared with the afternoon (P = 0.030). The exercise-related increase in melatonin was more pronounced (P = 0.024) in the morning (11.1 ± 8.7 pg ml(-1)) than in the afternoon (5.1 ± 5.7 pg ml(-1)). The slope of the heart rate-melatonin relationship was significantly (P = 0.020) steeper in the morning (0.12 pg ml(-1) beats(-1 )min(-1)) than in the afternoon (0.03 pg ml(-1) beats(-1 )min(-1)). In conclusion, we report for the first time that the masking effect of moderate-intensity exercise on melatonin is approximately twice as high in the morning than the afternoon. The much steeper relationship between heart rate and melatonin changes in the morning raises the possibility that time of day alters the relationships between exercise-mediated sympathetic nervous activity and melatonin secretion.