It has long been believed that the toxicity of autoxidized oils was due to hydropcroxides. On the other hand, there are several reports that some other secondary oxidation products such as hydroxyl and carbonyl compounds make greater contribution to the toxicity than hydroperoxides. However, the major toxic compound still remains to be elucidated and despite of much research on the mechanism of the toxicity, it has not been clarified yet. Authors made some attempts in this paper to cast light on the toxic compounds and their mechanism of the toxicity. Methyl linoleate was autoxidized at 60°C and the toxicity of this deteriorated ester on mice was examined. It was found that the longer the ester was oxidized, the more toxic it became. In order to fractionate the toxic substance, this oxidized ester was molecular-distilled first for the separation of low molecular weight compounds (I). The residue was washed with dil. alkali to remove the free fatty acids (II) and then column chromatographed on silicic acid to fractionate into four fractions (III-A, B, C, D). Fractions (I), (II.), (III-A, B, C, D) were fed through stomach tube on mice to investigate the degree of toxicity. Fr. (I) was found the most toxic followed by (II), (III-D) in the order. Fr. (I) was further fractionated into six fractions (I-A, B, C, D, E, F) on silica gel. Fr. (I-D) appeared the most toxic on mice. Fr. (I-D) was identified as 4-hydroperoxy-2-en-l-al with carbon chain ranging from five to nine by means of IR, UV, elementary analysis, melting point, MS and NMR.