The distribution of free and bound acetaldehyde in human blood was studied. Fresh whole blood was precipitated with a perchloric acid (PCA) in saline solution and an aliquot of the crude sample was taken for determination of 'total' acetaldehyde. The remaining sample was centrifuged and the clear supernatant taken for analysis of 'soluble' acetaldehyde. 'Bound' acetaldehyde was calculated by subtracting soluble from total amounts. In samples collected from healthy control subjects, the acetaldehyde level in separated plasma was usually below the limit of detection of the method (0.2 uM), while much higher concentrations (> 2.5 uM) were detected when analyses were carried out on whole blood. In whole blood, about 70% was recovered as bound (i.e. PCA-insoluble) acetaldehyde. The soluble (i.e. free + PCA-soluble) level was higher than that found in separated plasma, suggesting that some acetaldehyde was liberated from the blood cells by PCA treatment. In blood spiked with ethanol, a spontaneous formation of acetaldehyde occurred during the analytical procedure. The artefactual formation increased only the soluble amount, while the bound level remained unchanged. Likewise, in samples drawn from intoxicated subjects, artefactual formation of acetaldehyde was observed in the soluble fraction, while the bound amount was not significantly increased. No significant differences in acetaldehyde levels were found between males and females, nor between healthy control subjects and alcoholic patients undergoing treatment with the aldehyde dehydrogenase inhibitor disulfiram (Antabuse*). However, some of the Antabuse patients possessed elevated levels of bound acetaldehyde.