The effects of oral fructose on hepatocarcinogenesis were investigated with cytomorphological, cytochemical and stereological methods. Carcinogenesis was induced in male Sprague-Dawley rats by application of N-nitrosomorpholine (NNM) for 7 weeks. Afterwards, the animals received fructose in the drinking water (120 g/l) and food ad libitum (group I) or tap water and food ad libitum (group II). The incidence of hepatocellular carcinoma in rats treated with NNM plus fructose was 46% as compared to 24% in animals receiving NNM alone (P less than 0.05). There was no difference in the incidences of other malignancies between the groups (group I: 32.1%, group II: 32.0%). Morphometric evaluation of preneoplastic liver lesions indicated the enhancing effect of the fructose treatment several months before malignant tumors appeared. As early as 6 weeks after treatment the hepatic parenchyma occupied by focal lesions was increased from 6.7% in the animals which had received NNM alone to 8.5% (P less than 0.05) in animals having received NNM plus fructose. This increase was predominantly caused by an increase in glycogen storing foci (P less than 0.0005). In addition, the fructose treatment caused a histochemically detectable increase in the activity of glucose-6-phosphatase and glucose-6-phosphate dehydrogenase in both the hepatocytes of the focal lesions and the surrounding parenchyma. In the NNM plus fructose group the activity of the glucose-6-phosphatase in the foci was frequently approximately equal to the activity in the parenchyma of untreated controls. The striking increase in the activity of this enzyme in the surrounding hepatocytes, however, still sharply demarcated the lesions. The potential mechanisms by which fructose enhances hepatocarcinogenesis are discussed.