The induction and persistence of sister chromatid exchanges (SCEs) were studied in peripheral blood and spleen lymphocytes of mice given a single i.p. injection of ethylnitrosourea (ENU) of 100, 350, or 600 muMoles ENU/kg. SCE frequencies were measured on days 1, 3, 5, and 7, and at seven additional times up to 172 days post-injection. SCEs were analyzed statistically by comparing the mean frequencies as well as the distribution of SCEs per cell at each time. The latter approach was based on a non-parametric method of identifying high frequency cells (HFCs). The SCE frequencies and proportion of HFCs in each dose and tissue remained elevated for up to 172 days following treatment, although the degree and statistical significance of the increase varied according to the tissue, dose, and statistical test employed. The SCE frequencies were found to oscillate during the first week. Following this, however, the return of the SCE frequencies to control levels was fit to a linear regression model with time as the only independent variable. The persistence of SCE-forming lesions was found to be dose-dependent for the spleen but not for blood. Within each dose the persistence of SCE-forming lesions was significantly greater for the blood relative to the spleen. The results emphasize that tissue, dose, and time since exposure are important factors to consider when quantifying SCEs in vivo; analysis of high frequency cells may be a more sensitive method of detecting exposure than the t-test; and a single determination of SCE frequencies may not be sufficient to quantitatively assess genotoxic damage in the first week following exposure.