Hemoglobin (Hb) adduct determination by the N-alkyl Edman method was used for in vivo dosimetry of endogenously formed malonaldehyde (MA) and ethene in mice fed diets with different fatty acid composition and induced for lipid peroxidation with carbon tetrachloride (CCl4). In order to amplify lipid peroxidation animals were pretreated with phenobarbital (PB) and the glutathione-depleting agent DL-buthionine-(S,R)-sulfoximine (BSO). Non-treated animals raised on different diets were used as controls. Lipid peroxidation products in liver were measured as 2-thiobarbituric acid reactive compounds (TBA-C). Livers from control mice fed a soya oil based diet (rich in polyunsaturated fatty acids, diet S) showed approximately 6.5-fold higher levels of TBA-C than those from animals raised on a coconut oil based diet (mostly saturated fatty acids, diet C). The level of adducts of MA to Hb, determined as N-(3-hydroxypropyl)valine, was approximately 1.5-fold higher in animals from diet S than in animals raised on diet C. The highest increases in the levels of TBA-C and MA adducts were obtained after a simultaneous treatment of the animals with PB, BSO and CCl4. The increases of TBA-C were 1.3-fold (diet C) and 1.7-fold (diet S). The corresponding increases of MA-Hb adduct levels were 1.3- and 1.6-fold respectively, indicating an increased susceptibility of mice fed diet S to lipid peroxidation. The level of adducts from ethene, determined as N-(2-hydroxyethyl)valine, was also higher in mice from diet S than in animals fed diet C, when all treatment groups were considered. The difference was, however, only slightly significant (P less than 0.02). No difference between control and CCl4-treated animals, with regard to the ethene-Hb adduct, was found. Our results validate the use of Hb dosimetry for monitoring the effects of factors known to influence lipid peroxidation induced in vivo.