1. Pigeon heart mitochondria produce H₂O₂ at a maximal rate of about 20nmol/min per mg of protein. 2. Succinate–glutamate and malate–glutamate are substrates which are able to support maximal H₂O₂ production rates. With malate–glutamate, H₂O₂ formation is sensitive to rotenone. Endogenous substrate, octanoate, stearoyl-CoA and palmitoyl-carnitine are by far less efficient substrates. 3. Antimycin A exerts a very pronounced effect in enhancing H₂O₂ production in pigeon heart mitochondria; 0.26nmol of antimycin A/mg of protein and the addition of an uncoupler are required for maximal H₂O₂ formation. 4. In the presence of endogenous substrate and of antimycin A, ATP decreases and uncoupler restores the rates of H₂O₂ formation. 5. Reincorporation of ubiquinone-10 and ubiquinone-3 to ubiquinone-depleted pigeon heart mitochondria gives a system in which H₂O₂ production is linearly related to the incorporated ubiquinone. 6. The generation of H₂O₂ by pigeon heart mitochondria in the presence of succinate–glutamate and in metabolic state 4 has an optimum pH value of 7.5. In states 1 and 3u, and in the presence of antimycin A and uncoupler, the optimum pH value is shifted towards more alkaline values. 7. With increase of the partial pressure of O₂ to the hyperbaric region the formation of H₂O₂ is markedly increased in pigeon heart mitochondria and in rat liver mitochondria. With rat liver mitochondria and succinate as substrate in state 4, an increase in the pO₂ up to 1.97MPa (19.5atm) increases H₂O₂ formation 10–15-fold. Similar pO₂ profiles were observed when rat liver mitochondria were supplemented either with antimycin A or with antimycin A and uncoupler. No saturation of the system with O₂ was observed up to 1.97MPa (19.5atm). By increasing the pO₂ to 1.97MPa (19.5atm), H₂O₂ formation in pigeon heart mitochondria with succinate as substrate increased fourfold in metabolic state 4, with antimycin A added the increase was threefold and with antimycin A and uncoupler it was 2.5-fold. In the last two saturation of the system with oxygen was observed, with an apparent K_m of about 71kPa (0.7–0.8atm) and a V_max. of 12 and 20nmol of H₂O₂/min per mg of protein. 8. It is postulated that in addition to the well-known flavin reaction, formation of H₂O₂ may be due to interaction with an energy-dependent component of the respiratory chain at the cytochrome b level.