To evaluate surgical staging procedures in women with endometrial carcinoma, we examined the techniques used to assess the peritoneal cavity in 295 clinical stage I patients treated between 1985 and 1993. These patients were felt to be at increased risk for extrauterine disease because of significant myometrial invasion, high-grade (2 or 3), or variant histology (papillary serous, clear cell, or mixed). Patients had a mean of two intraperitoneal samples taken: 224 patients (76%) had at least an omental biopsy and peritoneal cytology. Additional peritoneal biopsy sites included pericolic gutters (50), pelvic peritoneum (45), bowel serosa/mesentery (24), diaphragm (22), appendix (11), and adhesions (7). At the time of staging laparotomy, 22 patients (7.5%) had gross evidence of peritoneal spread, which was readily confirmed by directed biopsy. In the 273 women without gross peritoneal disease, 3 (1%) had occult metastases detected by routine biopsy, 3 (1%) had microscopic metastases in palpably abnormal biopsies, and 22 bad positive cytology as the only evidence of peritoneal disease. Only three operative complications were potentially attributable to peritoneal assessment: cystotomy (1), partial small bowel obstruction (1), and ileus (1). Peritoneal failures have been noted in 12 patients over a mean follow-up interval of 39 months. Seven of these patients had obvious peritoneal disease at laparotomy. Two of the remaining 5 had optimal peritoneal sampling and represent false-negative cases. A staging laparotomy that included total abdominal hysterectomy with adnexal resection, cytology, omental biopsy, and biopsy of grossly abnormal sites would have potentially identified all patients with known peritoneal disease. Routine biopsy of other grossly normal peritoneal sites is associated with extremely low yield and is not recommended.