The formation of benzo(a) pyrene-deoxyribonucleoside adducts has been studied in several systems of differing structural and cellular integrity. Marked quantitative and qualitative differences were observed between the adducts formed using microsomes and those observed in isolated cells, tissue explants in short-term organ culture and rodent skin in vivo, thus illustrating that one must be very wary when extrapolating results from subcellular studies to isolated cells, tissues in culture or animals in vivo. Hydrocarbon-deoxyribonucleoside adducts formed using microsomal preparations from rat liver and rat lung, in the presence of benzo(a) pyrene and DNA, were mainly derived from 9-hydroxy benzo(a) pyrene whilst similar studies with microsomes from mouse skin showed that the DNA - bound products arose predominantly through reaction of (+/-) 4,5-dihydro-4,5-epoxy-benzo(a) pyrene with DNA. In striking contrast, DNA - adducts formed from benzo(a) pyrene in freshly isolated viable rat hepatocytes, rat tracheal epithelial 2Cl cells and mouse skin in vivo were mainly derived from reaction of (+/-) 7beta, 8alpha-dihydroxy-9alpha,10alpha-epoxy-7,8,9,10-tetrahydrobenzo(a) pyrene with deoxyguanosine. When mice were treated topically with benzo(a) pyrene, deoxyribonucleoside adducts derived from reation of (+/-) 7beta,8alpha-dihydroxy-9alpha,10alpha-epoxy-7,8,9,10-tetrahydrobenzo(a) pyrene with deoxyadenosine were formed in DNA isolated from the treated skin. This is the first study to report the formation of such products in vivo. No qualitative or quantitative differences were observed in the binding of benzo(a) pyrene to DNA in the skin of mice of different strains varying in their susceptibility to polycyclic aromatic hydrocarbon - induced carcinogenesis. However, in rat skin benzo(a) pyrene became bound to DNA to a far lesser extent than in mouse skin, possibly partially explaining the relative resistance of rat skin to carcinogenesis inducedby benzo(a) pyrene.