We determined the genotoxicity of synthetic red tar dyes currently used as food color additives in many countries, including Japan. For the preliminary assessment, we treated groups of 4 pregnant mice (gestational day 11) once orally at the limit dose (2000 mg/kg) of amaranth (food red No. 2), allura red (food red No. 40), or acid red (food red No. 106), and we sampled brain, lung, liver, kidney, glandular stomach, colon, urinary bladder, and embryo 3, 6, and 24 h after treatment. We used the comet (alkaline single cell gel electrophoresis) assay to measure DNA damage. The assay was positive in the colon 3 h after the administration of amaranth and allura red and weakly positive in the lung 6 h after the administration of amaranth. Acid red did not induce DNA damage in any sample at any sampling time. None of the dyes damaged DNA in other organs or the embryo. We then tested male mice with amaranth, allura red, and a related color additive, new coccine (food red No. 18). The 3 dyes induced DNA damage in the colon starting at 10 mg/kg. Twenty ml/kg of soaking liquid from commercial red ginger pickles, which contained 6.5 mg/10 ml of new coccine, induced DNA damage in colon, glandular stomach, and bladder. The potencies were compared to those of other rodent carcinogens. The rodent hepatocarcinogen p-dimethylaminoazobenzene induced colon DNA damage at 1 mg/kg, whereas it damaged liver DNA only at 500 mg/kg. Although 1 mg/kg of N-nitrosodimethylamine induced DNA damage in liver and bladder, it did not induce colon DNA damage. N-nitrosodiethylamine at 14 mg/kg did not induce DNA damage in any organs examined. Because the 3 azo additives we examined induced colon DNA damage at a very low dose, more extensive assessment of azo additives is warranted.