As previously shown with isolated oxyhemoglobin, methemoglobin formation can also be induced in intact erythrocytes by hydroxylamine compounds and substituted phenols such as butylated hydroxyanisole (BHA). Electron spin resonance investigations revealed that, accordingly, free radical intermediates were formed in erythrocytes from hydroxylamine, N,N-dimethylhydroxylamine, and N-hydroxyurea. Due to the low stability of the dihydronitroxyl radicals, their detection required the use of a continuous flow system and relatively high amounts of the reactants. As has already been demonstrated with the solubilized hemoglobin system, hemoglobin of intact erythrocytes also reacts with the more hydrophilic xenobiotics such as hydroxylamine. However, the reaction rate was slightly reduced, indicating the existence of an incomplete permeability barrier for these compounds. The limited solubility of phenolic compounds in the aqueous buffer of suspended erythrocytes (in combination with the strict requirement of osmolarity in order to prevent hemolysis) impeded the direct detection of the respective phenoxyl radicals previously reported in hemoglobin solutions. However, in accordance with earlier findings in homogeneous reaction systems, chemiluminescence was observed as well, indicating the existence of a further reaction intermediate, which was also obtained in pure hemoglobin solutions when mixed with the respective reactants. As has recently been demonstrated, this light emission is indicative of the existence of highly prooxidative compound I intermediates during methemoglobin formation. Prooxidant formation in erythrocytes is reflected by a significant decrease in thiol levels even with those compounds where free radical formation was not directly detectable by ESR spectroscopy. The use of the spin-labeling technique revealed membrane effects as a result of oxidative stress. Oxidative metabolism of hemoglobin with hydroxylamine caused a release of low molecular weight iron. The marked hemolysis observed in the presence of BHA results from a direct membrane effect of this compound rather than a consequence of free radical-induced oxidative stress. A correlation of the different results is discussed in terms of possible toxicological consequences.