Male Wistar-Furth rats were fed 0.02% 2-acetylaminofluorene (AAF) for 3 days or 0.02% AAF for 25 days followed by 0.02% [ring-3H]AAF for an additional 3 days. The concentration of hepatic DNA adducts was then monitored by both radioimmunoassay and radiolabeling during 28 days of control diet. This approach allowed comparisons to be made of adduct accumulation, removal and persistence at both the beginning and end of a four week carcinogen feeding period. DNA adduct formation remained constant during the month of AAF administration with an accumulation rate of 157 fmol adduct/micrograms DNA during days 1-3 and days 25-28 of the experiment. Furthermore, the rate of removal of adducts formed during these three day periods was similar when both groups were fed control diets for 28 additional days. Continued AAF administration resulted in a slow accumulation of persistent adducts; thus, 91 +/- 6% of the adducts detected after 3 days of AAF feeding were removed during a subsequent month of control diet, while only 65 +/- 11% of the adducts detected after 28 days of AAF diet were removed when rats were fed control diet for an additional 28 days. In a second experiment, the removal of adducts was compared in animals fed control or AAF diet after previously being fed 0.02% AAF for 17 days. Similar removal curves were observed in both groups; therefore, continued ingestion of AAF did not affect the rate of adduct removal. In both experiments, biphasic repair curves were observed. These data were used to develop a pharmacokinetic model. Two genomic regions were postulated, an area susceptible to fast repair and a region more resistant to the removal of AAF adducts. At equilibrium, which was reached after 2-3 weeks of AAF feeding, the concentration of adducts in each region was similar with approximately 150 fmol adduct/micrograms DNA. Although the total number of adducts formed in the fast repair region during one month of AAF administration was five times greater than in the resistant region, the model predicted that the adducts localized in regions resistant to repair were the persistent adducts detected after one month of control diet. Overall, the removal of adducts formed during chronic AAF feeding was very efficient since greater than 93% of the adducts were removed by the end of a subsequent month of control diet.