The effects of different patterns of alcohol administration on hepatocarcinogenesis induced by diethylnitrosamine (DEN) in male Wistar rats were assessed using a modified Ito's medium-term bioassay system. Carcinogenic potential was scored by comparing numbers and areas of glutathione S transferase placental form (GST-P)-positive foci. The activity of ornithine decarboxylase (ODC), the rate-limiting enzyme for polyamine synthesis, was also measured as a parameter of cell proliferation. Rats were given a single i.p. injection of DEN (200 mg/kg body weight), maintained on basal solid diet for two weeks, then divided into five groups: group A maintained on liquid diet in which 36% of total calories were provided by ethanol (diet Al) for 24 weeks; group B maintained on diet Al for 12 weeks and subsequently on control diet (diet C) for 12 weeks; group C maintained on diet C for 24 weeks; group D maintained on a cycle of two days on diet Al followed by two days on diet C; group E maintained on another liquid diet in which 18% of total calories were provided by ethanol for 24 weeks. The numbers and areas per cm2 of GST-P positive foci in group B were highest and in group D were the lowest among the five groups. ODC activities in groups A and E were significantly lower than in group C, that for group B was intermediate. These results suggest that the intermittent intake of alcohol exerts preventive potential on hepatocellular lesion development, and that interruption of long-term alcohol administration enhances hepatocarcinogenesis.