Anecdotal accounts and unsystematic reports have suggested that women who exercise experience fewer premenstrual symptoms and less severe dysmenorrhoea than women who are sedentary. The present study tested this belief. Mood states and physical symptoms of 143 women were monitored for five days in each of the three phases of the cycle (mid-cycle, premenstrual and menstrual). The women were 35 competitive sportswomen, two groups of exercisers (33 high exercisers and 36 low exercisers) and 39 sedentary women. Principal component analysis of their responses to a mood and physical symptom checklist revealed five dimensions: positive affect, negative affect, physical symptoms, fatigue and irritability. Component-based subscale scores were calculated for all dimensions. In general mood and symptoms were worse menstrually and premenstrually than mid-cycle. The high exercisers experienced the greater positive affect and sedentary women the least. The high exercisers also reported the least negative affect. The differences between exercise groups were greatest during the premenstrual and menstrual phases. These results are consistent with the belief that women who frequently exercise may be to some extent protected from deterioration of mood before and during menstruation. This, however, is not the case for competitive sportswomen.