Quantitative analysis of the mutagenicity and carcinogenicity of the low doses of genotoxic carcinogens present in food is of pressing concern. The purpose of the present study was to determine the mutagenicity and carcinogenicity of low doses of the dietary genotoxic carcinogen 2-amino-3-methylimidazo[4,5-f]quinoline (IQ). Male F344 gpt delta transgenic rats were fed diets supplemented with 0, 0.1, 1, 10 or 100 ppm IQ for 4 weeks. The frequencies of gpt transgene mutations in the liver were significantly increased in the 10 and 100 ppm groups. In addition, the mutation spectra was altered in the 1, 10 and 100 ppm groups: frequencies of G:C to T:A transversion were significantly increased in groups administered 1, 10 and 100 ppm IQ in a dose-dependent manner, and the frequencies of G:C to A:T transitions, A:T to T:A transversions and A:T to C:G transversions were significantly increased in the 100 ppm group. Increased frequencies of single base pair deletions and Spi- mutants in the liver, and an increase in glutathione S-transferase placental form (GST-P)-positive foci, a preneoplastic lesion of the liver in rats, was also observed in the 100 ppm group. In contrast, neither mutations nor mutation spectra or GST-P-positive foci were statistically altered by administration of IQ at 0.1 ppm. We estimated the point of departure for the mutagenicity and carcinogenicity of IQ using the no-observed-effect level approach and the Benchmark dose approach to characterise the dose-response relationship of low doses of IQ. Our findings demonstrate the existence of no effect levels of IQ for both in vivo mutagenicity and hepatocarcinogenicity. The findings of the present study will facilitate an understanding of the carcinogenic effects of low doses of IQ and help to determine a margin of exposure that may be useful for practical human risk assessment.