Before the introduction of tamoxifen, diethylstilbestrol (DES) was widely considered to be the hormonal treatment of choice in postmenopausal women with advanced breast cancer. We performed a randomized clinical trial of these two agents to determine their relative efficacy and toxicity. The trial involved 143 evaluable patients, of whom 99 had received no prior systemic therapy and 44 had received previous chemotherapy. The regression rates (complete plus partial) were higher in patients receiving DES (41 per cent) than in those receiving tamoxifen (33 per cent), but not significantly so (P = 0.37). In patients who had had no prior systemic therapy, the rates were 44 per cent and 38 per cent, respectively (P = 0.55), and in those who had had previous chemotherapy, 32 per cent vs. 23 per cent (P = 0.50). Analysis of the time until treatment failure for the two treatment groups showed no significant difference (medians: DES, 142 days; tamoxifen, 171 days). Toxicity was greater in patients receiving DES; nine of 74 patients (12 per cent) discontinued therapy solely because of adverse reactions. Since there was no statistically significant difference in efficacy and since tamoxifen was less toxic, tamoxifen appears to be the preferred agent.