Patients with a clinically palpable breast carcinoma, 4 cm or less in diameter, and with no palpable nodes in the axilla were prospectively entered into a randomized clinical trial. A total of 713 patients were registered between November 1982 and December 1987, of whom 708 are evaluable at a median follow-up of 37 months. Following excision of the primary tumour, patients were randomly allocated either to have radiotherapy to the affected quadrant only (LF group) or to the whole breast and regional lymph node areas (WF group). No adjuvant hormone or chemotherapy was prescribed. The primary tumour was reported as completely excised histologically in 80% of cases, incompletely excised in 10%, and no estimate was possible in 10%. At six years from first randomization, 96% of the WF group and 92% of the LF group have remained free of breast recurrence (94% and 87% actuarial breast recurrence-free survival at 5 years). Part of the difference may be explained by the 20% recurrence rate in the breast for lobular carcinomas treated within the LF group. Of the WF group 14 patients (4%) developed recurrent disease in the axilla, compared to 50 patients (14%) in the LF group (95% and 86% actuarial axillary recurrent-free survival at 5 years). Patients with primary tumours histologically 1 cm or less in diameter had a 98% actuarial 5-year survival compared with 74% for those with tumours measuring 2 cm or more in diameter (P = 0.003). Continued follow-up of these patients will provide further information on the factors governing local/regional recurrence.