The present study comprises 847 women operated upon for invasive breast carcinoma at 19 surgical departments and enrolled in protocol DBCG-82TM from January 1983 to November 1987. Among them 662 (78%) were allocated for breast-preserving therapy or mastectomy by randomization, while 185 patients (22%) did not accept randomization. Within the randomized group 6% could not be entered into adjuvant protocols, i.e. subsequent programmes of postoperative therapy and follow-up. This left 619 evaluable patients. In the non-randomized series 26% did not fulfil the demands for entrance into the adjuvant protocols, leaving 136 evaluable patients, 60 of whom had chosen a breast-preserving operation and 76 mastectomy. In the randomized series the patients in the two treatment arms were comparable in age, menopausal status, site of tumour, pathoanatomical diameter of the tumour, number of removed axillary lymph nodes, number of metastatic axillary lymph nodes, and distribution on adjuvant regimens. Ninety per cent of the patients in the randomized group accepted the method offered, whereas 10% declined and wanted the alternate form of operation. The median follow-up period was approximately 1.75 years. The cumulative recurrence rate in the randomized group was 13% and in the non-randomized group 7%. These results are preliminary. Life-table analyses have not so far demonstrated differences in recurrence-free survival either in the randomized or the non-randomized series.