In a case-control study of the risk of adenocarcinoma of the endometrium in relation to conjugated-estrogen use, we found that 31 per cent of 425 women with endometrial cancer and 15 per cent of 792 controls reported having used conjugated estrogens; the rate-ratio estimate was 3.5 with a 95 per cent confidence interval of 2.6 to 4.7. For use that lasted at least one year, the rate-ratio estimate for Stage I or II cancer was 5.2 (95 per cent confidence interval, 3.7 to 7.2), and for Stages III and IV combined it was 3.1 (1.5 to 6.4). Among women who had used estrogen for at least one year and then discontinued it, the risk of endometrial cancer remained significantly elevated even after estrogen-free intervals of over 10 years. The findings suggest that long-term use of conjugated estrogen increases the risk of both localized and widespread endometrial cancer. The data also suggest that women who have taken conjugated estrogen for one or more years remain at increased risk for at least 10 years after they discontinue use. Such women should be considered for long-term gynecologic surveillance.