Ninety-four patients with metastatic breast cancer were entered in a prospective randomised trial comparing 2 schedules of the same combination chemotherapy. Group I consisted of 46 patients, treated in a monthly three day course with adriamycin (ADM), cyclophosphamide (CPM), vindesine (VDS) and 5 fluoro-uracil (5 FU). Group II included 48 patients who received the same total monthly doses in 4 injections of ADM, CPM, VDS on days 2, 5, 16, 19 of each month together with a continuous infusion of 5 FU over a total of 10 days (days 1-5 and 15 and 19). Patient characteristics in the 2 groups as regards essential prognostic parameters were identical. We observed no difference in myelosuppression between the 2 groups. There were fewer gastrointestinal side effects in group II. Whereas complete alopecia occurred in 100% of patients in group I, only 33% of patients with the fractionated schedule totally lost their hair. The objective response rates at 8 months were 75 and 74% respectively, but the complete response rates (17.5 versus 28%) showed an advantage for group II. This advantage was not statistically significant. Durations of response were 15 months (group I) and 18 months (group II) and the median survival was 27 months in both. We conclude that prolonged low dose and fractionated administration of chemotherapy over 2 five day courses per month improves the therapeutic index and diminishes side effects.