For cancer risk assessment of 1,3-butadiene from rodent cancer test data, the in vivo doses of formed 1,2:3,4-diepoxybutane (DEB) should be known. In vivo doses of DEB were measured through a specific reaction product with hemoglobin (Hb), a ring-closed adduct, N,N-(2,3-dihydroxy-1,4-butadiyl)valine (Pyr-Val), to N-terminal valines. An analytical method based on tryptic digestion of Hb and quantification of Pyr-modified heptapeptides by LC-MS/MS has been further developed and applied in vivo to DEB-treated rats. Furthermore, N-(2,3,4-trihydroxybutyl)valine adducts (THB-Val) to the N-terminal valine in Hb were measured in rats and mice treated with DEB and in a complementary experiment with 1,2-epoxy-3,4-butanediol (EBdiol), using a modified Edman degradation method and GC-MS/MS. In vitro reactions of hemolysate with DEB and EBdiol were used to measure reaction rates for adduct formation needed for calculation of doses and rates elimination in vivo. The results showed that the level of the Pyr-Val adduct per administered dose of DEB was approximately the same in rats as had earlier been observed in mice [Kautiainen et al. (2000) Rapid Commun. Mass Spectrom. 14, 1848-1853]. Levels of the THB-Val adduct after DEB treatment were 3-4 times higher in rat than in mouse, probably reflecting an enhanced hydrolysis of DEB to EBdiol catalyzed by epoxide hydrolase. After EBdiol treatment, the THB-Val adduct levels were about the same in rat and mouse. Calculations from in vitro data show that the Pyr-Val adduct is a relevant monitor for the in vivo dose of DEB and that THB-Val primarily reflects doses to EBdiol. The calculated rates of formation of adducts and rates of elimination agree with expectations. Procedures for quantification of Hb adducts as modified peptides as well as preparation and characterization of peptide standards have been evaluated.