Cu-thionein from yeast was investigated by EPR spectroscopy to probe the oxidation state of copper, and the effects on it of oxidizing and reducing agents. At pH 0.2 the copper was released, but no EPR signal from Cu(II) was observed, unless air was present. Optical experiments did not detect any disulphide groups which might have been formed during anaerobic release of copper. The mercurial, p-hydroxymercuribenzoate caused the release of EPR-detectable copper only under aerobic conditions, and EDTA caused release of Cu(II) on heating. No reduction of the copper-thiolate units in Cu-thionein by ascorbate was detected. Potentiometric titrations with hexachloroiridate(IV) or hexacyanoferrate(III) produced several different Cu(II) EPR signals at various stages of oxidation. The former oxidizing agent required a lower oxidation-reduction potential (+350 mV) to oxidize the copper, than the latter (+410 mV) and neither titration was fully reversible. The EPR signal from Cu(II) oxidized by hexachloroiridate(IV) resembled that produced by p-hydroxy-mercuribenzoate in air, suggesting that the copper was released from its thiolate ligands. It is concluded that the EPR non-detectable copper in the native protein is Cu(I). Oxidation-reduction of the copper-thiolate clusters of Cu-thionein is proposed to be decisive for controlling storage and transport of cellular copper.