Glutathione S-transferase placental form (GST-P) positive foci development and its expression in liver exposed by nongenotoxic carcinogens phenobarbital (PB) and clofibrate (CF), and genotoxic carcinogen 2-amino-3-methylimidazo[4,5-f] quinoline (IQ) were investigated as a measure of carcinogenic potential of these chemicals. Male F344 rats were initially given a single intraperitioneal injection of diethylnitrosamine (200 mg/kg), and 2 weeks later, animals were fed diets containing 0.03% IQ or 0.5% CF or 0.05% PB or basal diet as a control for 6 weeks. All rats were subjected to two-thirds partial hepatectomy (PH) at week 3. Sequential sacrifice of rats was performed at 8 weeks or 52 weeks, and liver tissues were examined for immunohistochemical staining of GST-P positive foci. The numbers (No./cm2) and areas (mm2/cm2) of GST-P positive foci were increased by IQ or PB, but were decreased by CF compare to the control. Consistent with the development of GST-P positive foci, a time-related increase in the expression of GST-P mRNA was found in the rats treated with IQ, whereas CF decreased it. The incidence of hepatocellular carcinoma at 52 weeks was increased by all three chemicals. These results show that PB and IQ induced GST-P positive foci, but the peroxisome proliferator CF did not, which suggest that the prediction of carcinogenic potency based on the development of prenoplastic foci may cause false negative in a particular category compounds like peroxisome proliferators.