A treatment with the GnRH-agonist, buserelin, was given intranasally in a dosage of 400 micrograms once daily, to induce anovulation in 26 women with premenstrual tension syndrome; 23 patients completed the study course. The design was double-blind and cross-over. Daily symptom ratings were made for two pretreatment, diagnostic cycles and continued for up to six cycles or 6 months. The rating scale used was an earlier described visual analogue scale. Blood samples for estradiol and progesterone radio-immunoassay were taken once weekly throughout the study. Results show beneficial effects of both placebo and GnRH-agonist, compared with the pretreatment situation. The GnRH-agonist was, however, significantly better than placebo. At the end of the treatment periods the patients while still taking placebo, still showed cyclical symptom changes, whereas during the GnRH-agonist treatment the cyclical changes had disappeared. The results indicate that a factor from the corpus luteum must be involved in the etiology of cyclical mood changes. The results also show that inhibition of ovulation by mean of GnRH-agonists is one possible way to treat premenstrual tension syndrome.