During a 5-month study, we examined the effects of the relaxation response on premenstrual syndrome in 46 women who were randomly assigned to one of three groups: a charting group, a reading group, and a relaxation response group. The relaxation response group showed significantly greater improvement than the charting and the reading groups on physical symptoms (P less than .025 for both comparisons). There was a significant group-by-severity effect for charting versus relaxation response and for reading versus relaxation response on symptoms measured daily (P less than .01 for both comparisons), on emotional symptoms measured retrospectively (P less than .001 and P less than .025, respectively), and on symptoms of social withdrawal measured retrospectively (P less than .01 and P less than .025, respectively). Women with severe symptoms in the relaxation response group showed a 58.0% improvement, compared with a 27.2% improvement for the reading group and a 17.0% improvement for the charting group. We conclude that regular elicitation of the relaxation response is an effective treatment for physical and emotional premenstrual symptoms, and is most effective in women with severe symptoms.