The oxidation of lipids, lipid peroxidation, is usually assayed with thiobarbituric acid (TBA). We compare the TBA assay measuring TBA-reactive substances (TBARS), and a new gas chromatography-mass spectrometric (GC-MS) assay measuring malondialdehyde (MDA) with unsaturated fatty acids and biological samples. The extent of oxidation to different unsaturated fatty acids is related to the total number of bis-allylic positions, the position of the first double bond from the methyl terminus, and the lipid chain length. The extent of oxidation of different biological samples or organs is related to the component polyunsaturated fatty acids. Both the GC-MS and TBA assays give parallel results for oxidation of unsaturated fatty acids and biological samples. The GC-MS assay is about two- to sixfold more sensitive than the TBA assay for oxidation of unsaturated fatty acids. In contrast, the TBA assay gives about two- to sixfold higher TBARS than MDA by GC-MS assay in biological samples, possibly due to the nonspecificity and artifactual formation of derivatives in the acid-heating step of the TBA assay. The GC-MS assay is shown to be useful in oxidation-related cell culture studies with as few as 250,000 neural cells. These results suggest that the GC-MS assay is a useful, sensitive, and specific assay for lipid peroxidation. The TBA assay is also quite useful because of its sensitivity and simplicity, if one clearly understands its nonspecificity.