Adriamycin‐Fe³⁺ caused lipid peroxidation of erythrocyte membrane in relation to its concentration. Adriamycin‐Fe³⁺ had a high affinity for membrane and the adriamycin‐Fe³⁺‐binding membranes was also found to cause lipid peroxidation. Under aerobic conditions, adriamycin‐Fe³⁺ caused a reduction of cytochrome c and ferrous iron formed spontaneously. Superoxide dismutase (EC 1.15.1.1) (SOD) strongly inhibited the reduction of cytochrome c; however, the enzyme promoted formation of ferrous iron independent of enzymatic action. These results suggest that cytochrome c was reduced by superoxide radical (O₂ ⁻) or an adriamycin‐iron‐O₂ complex such as adriamycin‐Fe³⁺–O₂ ⁻, but not by adriamycin‐Fe²⁺. The ferrous iron chelator bathophenanthroline sulfonate (BPS) completely inhibited oxygen consumption caused by adriamycin‐Fe³⁺, indicating that ferrous iron is absolutely required for the lipid peroxidation. SOD and hydroxyl radical scanvengers did not inhibit the lipid peroxidation, indicating that O₂ ⁻ and hydroxyl radical were not involved in membrane peroxidation. The peroxidation reaction was dramatically inhibited by Tris buffer (2‐amino‐2‐hydroxymethyl‐1,3‐propanediol). However, hydroxyl radical generation and lipid peroxidation in Tris buffer were not related obviously, indicating that Tris did not act as a hydroxyl radical scavenger. The initial rate of TBARS (thiobarbituric acid reactive substances) formation induced by a mixture of adriamycin‐Fe³⁺ and adriamycin‐Fe²⁺ was much faster than that indcued by adriamycin‐Fe²⁺ or adriamycin‐Fe³⁺ alone. These results made it became possible to speculate that the lipid peroxidation might be initiated by an adriamycin‐Fe³⁺‐oxygen‐adriamycin‐Fe²⁺ complex.