Fifty-five patients with Stage II (36 patients) or Stage III (19 patients) malignant melanoma confirmed histologically received adjuvant immunotherapy with a polyvalent melanoma antigen vaccine to evaluate toxicity and immunogenicity. There was no toxicity. Antibody and/or cellular immune responses to melanoma were induced more frequently in Stage II (36 patients [69%]) than Stage III (19 patients [53%]) disease. The ability of different immunization schedules, alum, or pretreatment with low-dose cyclophosphamide to potentiate immunogenicity was compared after 2 months of immunization. Immunization biweekly with a fixed intermediate dose of vaccine was more immunogenic than immunization weekly with escalating vaccine doses. Alum increased the intensity of cellular responses slightly, whereas pretreatment with cyclophosphamide augmented both the incidence and intensity of cellular immune responses slightly. However, these changes did not reach statistical significance. There was a reciprocal relationship between the induction of humoral and cellular immune responses. These results show that (1) active immunotherapy with a polyvalent melanoma vaccine is safe in patients with minimal disease, (2) the vaccine augments immunity to melanoma in many, but not all, patients, and (3) several immunization strategies failed to potentiate immunogenicity significantly.