During the Workshop in North Carolina, the in vivo metabolism, adduct formation and genotoxicity data available from rodent and human exposure to 1,3-butadiente (BD) were reviewed and they are summarized in the present report. BD is metabolized by cytochrome P-450-dependent monoxygenases to the primary metabolite 1,2-epoxybutene-3 (epoxybutene, EB). EB is subjected to further metabolism: oxidation to 1,2:3,4-diepoxybutane (DEB), hydrolysis to 3-butene-1,2-diol and conjugation to glutathione. The first pathway seems to prevail in mice while the latter is characteristic for rats and possibly for humans. Species differences exist in adduct formation of the monoepoxide to hemoglobin, for which the following pattern has been found: mice > rats > humans. Genotoxity of BD was found in mice with all applied tests; however, negative results were obtained in rats. In exposed humans, the cytogenetic studies in peripheral blood lymphocytes did not show genotoxic effects, although one report described elevated hprt variant levels in peripheral blood lymphocytes of exposed workers. It was concluded that the presently available data are insufficient for the application of the parallelogram model to estimate genetic risk for humans. As an alternative approach, a tentative estimate of the doubling dose for induction of hprt mutations in somatic cells of mice and men was performed and the calculated values were surprisingly similar, i.e. 9000 ppmh. However, this estimate is burdened with a number of caveats which were discussed in detail. The working group identified a series of urgent research needs to provide the appropriate data for the application of the parallelogram model, such as identification of metabolic pathways in different rodent species and humans, metabolic studies in mice, rats and humans considering metabolic polymorphisms, studies of adducts to DNA and hemoglobin especially of DEB and other butadiene metabolites in rodents and humans, studies of mutational spectra (mutational fingerprinting) in somatic and germinal cells, confirmation of the human hprt mutation data, conformation of the rodent malformation data, dose-response studies in rodent germ cell tests and studies on repair kinetics of mono-adducts induced by EB as opposed to repair of cross-links produced by DEB. Finally, it was suggested that the original parallelogram consisting of data from somatic cell studies in rodents and humans plus studies of heritable effects in rodents to extrapolate to germ cell risk for humans should be supplemented with studies in sperm of experimental animals and exposed men.