Although bulk gold is barely reactive, highly dispersed gold on metal oxides and zeolites is highly active as a catalyst for a variety of reactions, including CO oxidation and various selective oxidations. There is still a controversy about the nature of the catalytically active species. The activity of supported gold has variously been attributed to nano-particles of gold, to isolated cationic gold sites, and to structures present at the gold–support interface. In an attempt to resolve matters, we designed a catalyst containing stable site-isolated cationic gold sites in the absence of zerovalent gold. These were tested as catalysts for CO oxidation. The catalyst was prepared by adsorption of Au(CH₃)₂(acac) [acac is CH₃COCHCOCH₃] in a pentane slurry on calcined nanostructured cerium oxide (Daiichi Genso, 168 m²/g) in the absence of air and moisture. IR and X-ray absorption spectroscopy (XAS) were used to characterize the supported species. The lack of Au–Au contributions in the extended X-ray absorption fine structure (EXAFS) spectra indicate that the gold in the as-prepared sample was present as mononuclear (site-isolated) species. EXAFS and IR data showed that acac ligands were removed and the gold was anchored to the support. The EXAFS data showed that each Au atom was bonded, on average, to approximately 2 oxygen atoms of the support. X-ray absorption near edge structure (XANES) spectra indicated the presence of cationic gold species, indicated by a white line characteristic of Au(III) in the as-prepared sample. CO oxidation catalysis was carried out at 80ºC and atmospheric pressure in a once-through plug-flow reactor fed with a 1:1 molar ratio of CO and O2; the conversions were about 5%. The catalyst was stable for as long as it was tested (4 h on stream). EXAFS spectroscopy provided evidence of the stability of the gold species during the catalysis, as indicated by the lack of Au–Au contributions. Moreover, XANES provided evidence of cationic gold species during catalysis; there was no evidence of Au(0). The data indicate that isolated cationic gold sites on the ceria surface are stable and are the catalytically active species. There is no evidence of a role of gold nano-clusters or particles in the catalyst. Thus, we infer that zerovalent gold is not necessary for catalysis of CO oxidation by supported gold.