Second-look laparotomy has been performed on 76 patients with ovarian carcinoma each of whom was clinically free of tumour after completing a programme of postoperative chemotherapy (60 patients), pelvic radiotherapy followed by chemotherapy (14 patients) or observation (2 patients). The initial stage, bulk of residual tumour, completion of hysterectomy and bilateral salpingo-oophorectomy and an interval of 11 months or more between diagnosis and second-look surgery were all significantly (p less than .05) associated with negative findings. Recurrences occurred after a negative operation in 10 of 42 (23.8%) patients and were more frequent amongst patients with grade 3 tumours (p = .03). Macroscopic cancer was found in 23 (30.3%) patients and microscopic cancer in 11 (14.8%) patients. Survival was related to the surgical findings and to the amount of residual tumour at the conclusion of the second-look operation. Survival of patients with no macroscopic residue (those found to have microscopic tumour and those who had complete excision of macroscopic tumour) was significantly longer than that of patients with macroscopic residue (median survival 38.2 versus 19.0 months, p = .01). These results suggest that benefit from second-look laparotomy is confined to patients with microscopic or totally resectable tumour for whom second-line therapy may still be curative.