The assays for the detection of unlabeled 1-beta-D-arabinofuranosylcytosine (cytosine arabinoside, Ara-C) incorporation into DNA was simplified. The procedure includes DNA isolation from leukemic cells, quantification of DNA concentrations, breakdown by enzymatic digestion of DNA to nucleosides and a radioimmunoassay (RIA) using an antibody against Ara-C. Different techniques for quantification of DNA concentrations are compared. A fluorimetric technique using Hoechst 33258 is preferred because it is the most specific method. Comparison of this RIA assay with measurement of [3H]-Ara-C/DNA formation under similar conditions in HL-60 cells showed a correlation of 0.99. Ara-C incorporation into DNA of leukemic cells was studied using two rat-leukemia cell lines, one of which is sensitive to Ara-C and the other is an Ara-C-resistant wild type: BNML-Cl/0 and BNML-Cl/Ara-C, respectively. The results showed that Ara-C is incorporated when the cells are incubated at concentrations equal to or higher than the Ara-C concentration that induces 50% growth inhibition after 48 h incubation (IC50). This implies that at lower Ara-C concentration, i.e. levels that do not induce cytotoxicity, Ara-C is not incorporated into DNA. Similar results were obtained with human HL-60 myeloid leukemia cells. The detection limit of this assay is 2 pmol/ml Ara-C; therefore, the assay is more sensitive than measurement of Ara-C triphosphate (Ara-CTP), the only metabolite that can be measured in leukemic cells from patients after in vivo Ara-C administration. On the basis of in vitro studies, the finding of detectable Ara-C/DNA levels in vivo is expected to correlate with cytotoxicity; whether or not the Ara-C/DNA level itself is informative remains to be evaluated.