Groups of male Alderley Park rats were dosed concomitantly with 2-acetylaminofluorene (2AAF) by gavage at doses between 0.01 mg/kg and 40 mg/kg, and livers sampled 2-72 h later. The liver of one group of animals was perfused to yield hepatocytes which were assayed in vitro for unscheduled DNA synthesis (UDS) via incorporation of tritiated thymidine and autoradiography. DNA was extracted from the livers of the other group and DNA adduct levels determined using the 32P-postlabelling technique. The major C-8 2-aminofluorene/guanosine adduct and 3 minor adducts were quantitated, enabling the relative sensitivity of the 2 techniques to be compared. A dose- and time-related UDS response was observed, which, at the most sensitive time-point (12 h) enabled DNA repair to be discerned at a dose level of 0.1-1 mg/kg of 2AAF, a response classified as formally positive at 5 mg/kg 2AAF. Only the C-8 adduct, as determined by 32P-postlabelling, was discernible at 0.01 mg/kg of 2AAF, although other adducts were visible on autoradiograms at higher dose levels. It is concluded that as part of a well-defined dose response, UDS can be discerned with confidence for doses of 2AAF between approximately 0.1 and 5 mg/kg, and DNA adducts for doses of 2AAF between approximately 0.01 and 1 mg/kg. Discernible UDS for 2AAF in the rat liver is apparent at approximately 13 DNA (total) adducts/10(8) nucleotides, or approximately 8 DNA (C-8) adducts/10(8) nucleotides. The presumed C-8 2-acetylaminofluorene/guanosine adduct, prepared by reaction of 2-acetoxy-2-acetylaminofluorene (2AAAF) with DNA, was a significant but unreliable marker of 2AAF/DNA adducts in the rat liver in vivo. DNA repair did not appear to remove DNA adducts selectively, and adducts remained in DNA when discernible DNA repair had ceased.