One hundred volunteers completed a modified version of the Moos Menstrual Distress Questionnaire daily for 35 days. The purpose of the study was disguised. None of the participants was taking oral contraceptives. Significantly more symptoms were reported in the premenstrual and menstrual phases and fewer during the follicular phase of the cycle, but the pattern of response varied considerably between subjects. Symptom reporting was higher, in all phases of the cycle, in women with high trait anxiety or psychiatric morbidity (indexed by the General Health Questionnaire) but these measures did not relate to fluctuations of symptoms around the menstrual cycle. The need for prospective longitudinal studies of menstrual cycle symptomatology is emphasised.