Tamoxifen was administered to three strains of female mice (B6C3F1, C57BL/6 and DBA/2) in short- and long-term studies to determine their ability to activate tamoxifen and cause hepatic DNA damage. 32P-Postlabelling of liver DNA from mice treated for 4 days showed a group of major adducts that increased in a dose-dependent manner and co-chromatographed with the major adducts detected in rat liver. On cessation of dosing, the majority of adducts were cleared within 3 days. Binding of [14C]tamoxifen to DNA nucleotides was demonstrated by the use of accelerator mass spectrometry. In long-term studies of 12 months to 2 years duration, dependent on strain, tamoxifen was administered continuously in the diet to give a daily dose of approximately 40 mg/kg. DNA adducts were detected after 3 months, although the number of adducts decreased with time and by 2 years were not detectable in the tamoxifen treated mice. None of the treated groups showed a significantly increased incidence of liver tumours, with or without phenobarbital promotion and there was no sustained liver cell proliferation. Tamoxifen was detected in the mouse livers, but at levels 50 times lower than those reported in a comparable rat study. These results suggest that, in contrast to the rat, tamoxifen is non-carcinogenic in mice because it does not cause sufficient cumulative DNA damage, or act as a promoter by causing cell proliferation.