Thiobarbituric acid (TBA) assays which have been modified for detection of lipid hydroperoxides appear to be useful for demonstration of in vivo lipid peroxidation. Since these methods require heating tissue membranes with the buffered TBA, there is a possibility of interference from the detection of autoxidation that occurs during heating. These studies were undertaken to investigate conditions which favor TBA color production from hydroperoxide while limiting autoxidation during the assay. An acetic acid-sodium acetate buffered (pH 3.6) TBA assay was used. Heating linoleic acid hydroperoxide with 50 microM ferric iron or under nitrogen nearly doubled color production compared to heating it with no added iron or under air. The lipid antioxidant butylated hydroxytoluene inhibited color production from fatty acid hydroperoxides. When tissue fractions, including liver and lung microsomes and lung whole membranes, were heated in the assay, color production was greater under air than under nitrogen and was much greater under oxygen. When liver microsomes from carbon tetrachloride-exposed rats were used, color was increased only when oxygen was present in the heating atmosphere. The results with tissue fractions appear to demonstrate autoxidation during color development rather than the presence of preformed hydroperoxides. Finally, it was found that color production from membrane fractions was dependent on the vitamin E content of the membranes. It appears that autoxidation during heating should be limited by heating under nitrogen and not by adding antioxidants, which inhibit color production from hydroperoxides. As the vitamin E effect demonstrates, antioxidant status must be considered, since a change in color production could result from a change in antioxidant content without the accumulation of lipid hydroperoxides.