This phase II study evaluates the efficacy and toxicity of a prolonged schedule of oral etoposide in patients with measurable advanced ovarian cancer resistant to, or relapsed following, platinum-based chemotherapy. Forty-seven patients participated, 20 of whom had received more than one prior treatment. Seventy-seven per cent had evidence of disease progression during or within 6 months of the previous chemotherapy. Initially, oral etoposide, 50 mg b.d. (regardless of patient size), was given for 14 days on a 21-day cycle. However, after encountering toxicity, the schedule was modified to 7 days' treatment escalating to 10 then 14 days if well tolerated. Among 41 assessable patients there were two complete and eight partial objective responses (24% response rate; 95% confidence interval 12-41%). Nine further patients (22%) had stable disease, four with a sustained fall of > 50% in CA-125. Median duration of response or stable disease was 35 weeks (range 21-49). Overall median survival was 41 weeks from study entry (range 2 to 96+ weeks). Toxicity for most patients was mild, but sporadic severe myelotoxicity occurred, with two treatment-related deaths. Risk factors for severe toxicity were: performance status 3; hepatic impairment; renal impairment. We conclude that oral etoposide has activity in platinum-resistant ovarian cancer and that it is a useful palliative therapy. It has significant toxicity which may be avoided by appropriate patient selection and an escalating-duration schedule.