The presence of progesterone receptors was found to be associated with a favorable prognosis in 98 patients with primary breast cancer. The occurrence of metastases was 3.6 times less probable in patients with progesterone receptor-positive tumors than in patients with progesterone receptor-negative tumors. There was also an inverse relationship between the concentration of progesterone receptor and the frequency of metastases. However, there was no statistical correlation between frequency of local recurrences and progesterone receptor content of the tumor. In patients displaying clinical or histological criteria of gravity, the presence of progesterone receptors allowed us to define subgroups with good prognosis. Thus, in women with progesterone receptor-positive cancers, metastases had occurred at 18 months, in only 5% of the 39 Grade III cancers and in none of the 25 cases with invaded axillary nodes. Measurement of estradiol receptor (105 patients including the previous 98 patients) was found to be less effective for guiding the prognosis of early breast cancer. Combined evaluation of estradiol and progesterone receptors did not provide any more information than did the determination of progesterone receptor alone.