The liver is, under normal conditions, mitotically inactive and relatively resistant to chemical carcinogenesis. When rat or mouse hepatocytes are stimulated to divide, however, the liver becomes exquisitely sensitive to carcinogenesis. The heightened sensitivity of dividing liver cells to carcinogens is one of the most dramatic phenomena in the field of experimental chemical carcinogenesis and is reproducible with a wide variety of chemical agents and experimental conditions. This same phenomenon seems to apply to humans, as circumstances that produce a sustained hepatocellular proliferation in man are associated with an increased risk of hepatocellular carcinoma (HCC). These include inborn errors of metabolism (e.g., haemochromatosis, Wilson's disease, hereditary tyrosinaemia) as well as alcoholism. A recent editorial in this Journal suggested that any condition resulting in cirrhosis is also associated with an increased risk of HCC, and this may in turn be due to regenerative hyperplasia always present in cirrhotic liver (Johnson PJ, Williams R. J Hepatol 1987; 4: 140-147). In the case of HCC associated with hepatitis B virus (HBV) infection, the possibility must be entertained that chronic HBV infection serves to produce a sustained hepatonecrosis with concurrent (regenerative) hyperplasia. This proliferative state would in theory serve to increase the liver's susceptibility to environmental dietary carcinogens and may tend to increase the risk of HCC by this indirect mechanism. Until a molecular mechanisms is demonstrated whereby HBV produces a defined cellular lesion that endows hepatocytes with a malignant phenotype, it should not be assumed that HBV is a direct cause of HCC.