We have evaluated three methods of assessing patients with cyclical breast symptoms (cyclical mastopathy--CM). One method, a screening questionnaire, was used to identify women with CM, and two others, a symptom severity questionnaire and a daily diary, were used prospectively to record the severity of symptoms in relation to the menstrual cycle. The screening questionnaire was assessed for test-retest reliability and for agreement between the recalled severity of symptoms and those recorded during prospective measurement. The symptom severity questionnaire and diary were assessed by examining their ability to discriminate between pre- and postmenstrual phases of the menstrual cycle, differences in symptom severity at these times being the cardinal feature of CM. Test-retest reliability of premenstrual symptoms by the screening questionnaire gave correlation coefficients over 0.70 for 9 of the 11 items assessed. The symptom severity questionnaire showed significant differences between pre- and post-menstrual scores for 8 of the 11 items assessed and the diary for each of the 2 items assessed. Comparison of symptom severity as recalled and as prospectively recorded showed only modest agreement. These results show that the instruments used to assess symptom severity performed satisfactorily but that the screening questionnaire used to identify women with CM, although reliable, correlated only moderately well with prospective measurement.