The effect of rat liver Sg on the mutagenicity of 10 nitrated polycyclic aromatic hydrocarbons (nitro-PAHs) was evaluated with Salmonella typhimurium TA98NR using S9 from phenobarbital-, 3-methylcholanthrene (MC)-, beta-naphthoflavone- and polychlorobiphenyl-treated and untreated rats. 2-Nitrofluorene (2-NFl), 2-nitrofluoren-9-one (2-NFlone), 2-nitrocarbazole (2-NCz), 3-NCz, 2-nitrodibenzothiophene (2-NDBT), 2-nitro-6H-dibenzo[b,d]pyran-6-one (2-NDBP) and 3-NDBP were metabolically activated by one or more of the S9 fractions, and the highest enhancement of the mutagenic potency of nitro-PAHs was observed with 3-MC-induced S9. Only in the case of 3-NFlone was the mutagenicity in strain TA98NR decreased by the addition of S9, regardless of S9 induction. 2-NDBP was most efficiently activated among nitro-PAHs tested by all S9 fractions used. The cytosolic fraction of S9 accounted for more of the activation of 2-NDBP than the microsomal fraction. NADH and NADPH were the most effective electron donors on the activation of 2-NDBP by S9. 2-NDBP was also metabolically activated by NADH plus commercial preparations of xanthine oxidase. These activations of 2-NDBP were inhibited by allopurinol, indicating that cytosolic xanthine oxidase in rat liver S9 participates in the activation of 2-NDBP. The potency of 2- and 3-NDBP isomers as base substitution mutagens was also enhanced by S9. In the presence of S9, both compounds showed the highest mutagenicity in strain TA7005 (C . G --> A . T) followed by strains TA7004 (G . C --> A . T), TA7006 (C . G --> G . C) and TA7002 (T . A --> A . T), and this mutation specificity was similar to that without S9, indicating that the mechanism of mutagenesis caused by NDBP isomers with S9 is similar to that without S9.