After a long period of frustration, many components of the mammalian plasma membrane redox system are now being identified at the molecular level. Some are apparently ubiquitous but are necessary only for a subset of electron donors or acceptors; some are present only in certain cell types; some appear to be associated with proton extrusion; some appear to be capable of superoxide production. The volume and variety of data now available have begun to allow the formulation of tentative models for the overall network of interactions of enzymes and substrates that together make up the plasma membrane redox system. Such a model is presented here. The structure discussed here is of the mammalian system, though parts of it may apply more or less accurately to fungal and plant cells too. Judging from the history of mitochondrial oxidative phosphorylation, it may be hoped that the development of models of the whole system - even if they undergo substantial revision thereafter - will markedly accelerate the pace of research in plasma membrane redox, by providing a coherent basis for the design of future experiments.