The general structure of the enzyme, its polypeptide composition, and a proposal for a rational nomenclature are discussed. The mitochondrially coded and bacterial cytochrome c oxidase subunits have been analyzed with more attention focused on elucidating the number of metals present in the enzyme and the ligands available for their coordination. The picture of a 2 Cu/2 Fe enzyme has been compared with that of a 3 Cu/2 Fe enzyme and a new model is proposed for the location of the metal centers in the enzyme.