The hepatocarcinogenic effect of Clophen A 30 and Clophen A 60 was tested in male weanling rats by long-term feeding over a period of 832 days. The mortality rate was investigated in 100-day intervals. In the first 800 days liver carcinoma accounted for 21% of necropsies in the Clophen A 60 group but only 2% of the necropsies in the Clophen A 30 group and none in the control animals. The tumors were first observed after 700 days. After 800 days hepatocellular carcinoma was the most common lesion observed in the Clophen A 60 animals (61%) whereas it was only observed in 3% of animals in the Clophen A 30 group and 2% in the controls. Preneoplastic lesions, such as foci of hepatocellular alterations and neoplastic nodules, were first observed after Day 500. The incidence of foci predominated in all time intervals, but an increase in neoplastic nodules and hepatocellular carcinomas was observed with increased time. There was a marked trend from foci to neoplastic nodule to hepatocellular carcinoma with time. The total mortality rate and the incidence of thymoma, inflammatory lesions of the urogenital tract, in the experiment were significantly reduced by Clophen administration. Whether this protective effect could be induced by polychlorinated biphenyls (PCBs) is discussed.