A randomized clinical trial involving postmenopausal patients with estrogen receptor positive recurrent breast cancer is reported. Of 168 patients entered, 156 were evaluable, of whom 79 received oral tamoxifen citrate 10 mg twice daily and 77 oral megestrol acetate 40 mg four times a day. Partial response (PR) plus complete response (CR) rates (both arms, 34%) and time-to-disease progression were similar in both arms. Side effects and toxicity were minimal with both regimens, although more patients who received tamoxifen complained of hot flushes (33% v 11%) and more patients who received megestrol acetate had a 10% or greater weight gain at 6 months from baseline (51% v 19%). On progression of disease, 73 patients who had achieved a CR, PR, or stable response received the alternative hormonal treatment in addition to the original hormonal therapy. Ten of 40 patients (25%) who began treatment with megestrol acetate had a further CR or PR; none of 33 patients originally receiving tamoxifen had a response when megestrol acetate was added. Similarly, patients who received tamoxifen as an addition to their original megestrol acetate treatment also had a significantly longer time to second progression than did those in the comparative arm. It was concluded that as initial hormonal therapy for relapsed patients, either tamoxifen or megestrol acetate can be used with confidence. However, it is suggested that tamoxifen and megestrol acetate should not be used in combination, except for those few occasions when tamoxifen is added as second-line therapy following a completed megestrol acetate response, and the megestrol acetate is continued for its palliative effects on appetite and weight gain. Possible mechanisms behind these results are discussed.