Rhizoxin is a new anti-tumour agent isolated from the pathogenic fungus Rhizopus chinensis. It has shown broad activity against murine tumour models and is also active against vinca alkaloid-resistant cells. The purpose of our studies was to determine the clinical activity of this compound in patients with advanced breast cancer and melanoma. Based on the results of a phase I study, 2.0 mg m-2 was administered as intravenous infusion over 5 min every 21 days. Nineteen patients were entered into the breast cancer phase II trial and received a total of 50 courses (median 2, range 1-6). Of these, dose reductions were performed in three courses because of leucopenia or stomatitis (1.5 mg m-2, one course; 1.45 mg m-2, two courses). Twenty-six patients were entered into the melanoma trial and received a total of 70 courses (median 2, range 1-12). No dose reductions were required. All patients were eligible for toxicity. Haematological toxicity included neutropenia CTC grade 3 (29/120 courses, 24.2%) and grade 4 (11/20 courses, 9.2%). Only drug-related CTC grade 1 thrombocytopenia was observed. Non-haematological toxicity included alopecia in all patients after two courses of treatment as well as CTC grade 3/4 stomatitis and asthenia. In the breast cancer study, one patient achieved a more than 50% tumour reduction after six cycles but was progressing after 6 weeks. Another patient showed a partial remission after the first course but was taken off the study because of CTC grade 3 skin toxicity. One patient was not evaluable for response (early death). No objective remissions were observed in 15 evaluable patients. In melanoma, no objective remissions were observed. We conclude that rhizoxin can be safely administered at 2.0 mg m-2 every 3 weeks. However, it has little activity in patients with advanced breast cancer and melanoma.