1. We report some pitfalls in the measurement of whole blood and plasma glutathione in man. 2. Using a simple tourniquet to the forearm and a 21-gauge needle, blood samples were collected by brachial vein puncture from healthy subjects. Whole blood and plasma were analysed for total glutathione, including the reduced and oxidized forms, by a spectrophotometric recycling method involving the glutathione reductase/NADPH couple. 3. The concentration of oxidized glutathione was determined after treatment of aliquots with either 2-vinylpyridine or N -ethylmaleimide to trap reduced glutathione. Reduced glutathione in the native samples could then be obtained by subtraction. 4. When the reagents were added to separated plasma, 2-vinylpyridine yielded values for oxidized glutathione that were twice as high as with N -ethylmaleimide. In whole blood studies the discrepancy was even greater, and the problem was not resolved by deproteinization of samples with 5-sulphosalicylic acid. Using N -ethylmaleimide, levels of oxidized glutathione were less than 1% of total glutathione in whole blood. 5. Despite attempts to minimize haemolysis, lysed erythrocytes contributed on average 25% to the ‘plasma’ glutathione concentration.