A reduction in the ratio of tetraploid to diploid liver nuclei has been investigated as an early indicator of hepatocarcinogenesis in the rat using the liver carcinogen 3'-methyl-4-dimethylaminoazobenzene (3'M). In a dose ranging study 3'M was administered by gavage to rats at 5, 12.5 and 25 mg/kg for up to 10 weeks and the following parameters studied: bodyweight gain, dye binding to hepatic protein, nuclear ploidy in liver and histopathology. Significant reduction in bodyweight occurred only with 25 mg/kg; dye binding to protein occurred in a dose-related manner; depression of the percentage of tetraploid nuclei compared with diploids was dose-related and effects were detected even at the lowest dose. These observations were consistent with those from previous studies by other investigators. In a separate experiment 3'M was administered at the maximum tolerated dose (MTD) of 25 mg/kg for 3 weeks, during which time there was a significant reduction in bodyweight gain and a reduction in the ratio of tetraploid:diploid liver nuclei. After cessation of dosing the rate of bodyweight gain returned to normal but there was no corresponding recovery of the ratio of tetraploid:diploid nuclei in the liver. A long-term continuous gavage study at 2.5 mg/kg revealed a time dependent reduction in the ratio of tetraploid:diploid liver hepatocyte nuclei and histopathological changes that included hepatocarcinoma were also observed. There was no correlation between the severity of pathological changes and the change in nuclear ploidy ratio in this experiment and it is concluded that the changes in ploidy ratio are related to the carcinogenic effect of 3'M and are independent of its gross toxicity.