Forty-five patients who had primary cancer of the breast had a laparotomy when a second cancer appeared in the ovaries. There were no other peritoneal metastases to suggest that it was disseminated breast cancer. Twenty-eight had primary ovarian cancer and 17 metastases from breast cancer. The study comparing the two groups shows that: there are few clinical or paraclinical features that make it possible to arrive at a differential diagnosis; ovarian metastases occurring in patients who have primary breast cancer carry a more serious prognosis; if the cancer is a primary ovarian cancer (and not a secondary) the chances of survival are less poor; in primary ovarian cancer the results justify our attitude for carrying out surgery; screening systematically for ovarian involvement by using CA 125 or by routine ultrasound does not seem to us to be effective.