At 80 K, exposure of the Cu(100) surface to oxygen results in a disordered (2 × 2) structure, followed by a (√2 ×√2) R45° mesh at higher coverage. The FWHM value of the O(1s) peak at 80 K is large (4.5 eV) indicating a range of metal-oxygen bond strengths. On warming to 290 K about half of the oxygen adsorbed at 80 K is desorbed below 150 K and there is a narrowing of the O(1s) peak to 2.3 eV. The (√2 ×√2) R45° mesh is, however, retained but with apparently improved ordering of the surface structure. A model for the adlayer is suggested. At 290 K, with increasing exposure to oxygen at low pressure (10–6 Torr), three diffraction patterns are observed; an oblique “4 spot” mesh, followed by a (√2 ×√2) R45° and finally a (√2 × 2 √2) R45°. We suggest that the (√2 ×√2) R45° structure, the stable structure at 80 K, involves only chemisorbed oxygen while the (√2 ×2 √2) R45° reflects oxygen incorporation. At 290 K, exposure of the Cu(100) surface to 5 Torr of oxygen leads to substantial oxygen uptake, a disordered diffraction pattern, the copper 2p peaks showing for the first time the satellites known to be characteristic of copper (II). Heating to 520 K in vacuo causes the reappearance of the (√2 × 2 √2) R45° mesh and the disappearance of the satellites. Heating to 570 K in 0.5 Torr of oxygen and subsequent cooling in oxygen results in the appearance of diffraction rings which are compatible with the formation of CuO, which XPS shows to be present. Interpretation is facilitated by recourse to thermodynamic data and to results already published for polycrystalline copper surfaces.