Selected inbred strains of mice were compared with respect to their susceptibility to two-stage liver carcinogenesis. Five-week-old male mice of strains C57BL/6NCr (C57), C3H/HeNCrMTV- (C3H) and DBA/2NCr (DBA) were given a single i.p. injection of N-nitrosodiethylamine (DEN, 90 mg/kg body weight) or the solvent tricaprylin (10 ml/kg). Beginning 2 weeks later, half of the DEN-treated and half of the control mice were given drinking water containing 0.05% phenobarbital (PB). Ten mice from each treatment group were killed at 12, 24, 36 and 52 weeks of age (5, 17, 29 and 45 weeks exposure to PB). PB significantly increased both the number of hepatocellular foci/cm2 and the incidence of hepatocellular tumors after 17 weeks of treatment in 24-week-old DEN-initiated mice of strains C3H (0.11 +/- 0.07 versus 2.9 +/- 0.3 foci/cm2 and 20 versus 70% incidence of hepatocellular tumors) and DBA (0.09 +/- 0.09 versus 3.72 +/- 0.6 foci/cm2 and 0 versus 90% incidence of hepatocellular tumors) but was ineffective in C57 mice (0.04 +/- 0.04 versus 0.07 +/- 0.07 foci/cm2). At 36 weeks of age the incidence of liver cell tumors in mice given DEN but not PB was 10 (DBA), 10 (C57) and 50% (C3H); the incidence was increased by PB to 90% in DBA and 100% in C3H mice, but there was no increase in C57 mice. Even at 52 weeks, the low incidence of hepatocellular tumors in C57 mice given DEN only (20%) was not significantly increased by subsequent exposure to PB. Serum PB levels observed at 12, 24 and 36 weeks of age were significantly higher in DBA mice than in C57 or C3H mice. Similar results were observed in a separate study in which PB was administered in drinking water to 7-week-old male mice of these three strains for 20 days, during which period serum PB levels were measured at shorter intervals. DBA mice thus appear to be unable to metabolize PB, which itself rather than its metabolites is probably responsible for tumor-promoting effects. DBA mice were especially sensitive, while C57 mice were refractory to promotion of hepatocarcinogenesis by PB. These two strains, which differ with respect to other significant parameters for chemical carcinogenesis including inducibility for aryl hydrocarbon hydroxylase and susceptibility to promotion of hydrocarbon-initiated skin tumors by 12-O-tetradecanoylphorbol-13-acetate, thus also provide a means for analysis of the pharmacogenetics of susceptibility to hepatocellular tumor promotion.