To determine a concentration of 5-bromo-2-deoxyuridine (BrdU) sufficient for sister chromatid differentiation (SCD), and yet having a minimal effect on the number of sister chromatid exchanges (SCEs), we assessed the effect produced on the number of SCEs by low concentrations (1, 3, and 10 micrograms/mL) of BrdU. SCD was not obtained in 19% of the 31 subjects with 1 microgram/mL of BrdU, while the differentiation was adequate for all samples treated with 3 and 10 micrograms/mL. We statistically analysed the effects of these three different doses and found no significant difference in the number of SCEs obtained with the doses of 1 and 3 micrograms/mL, but a significant difference was observed between these two concentrations and 10 micrograms/mL. We therefore suggest that the dose of 3 micrograms/mL, while sufficient to produce reliable differential staining, still permits an adequate evaluation of the base line of SCEs and appears to enhance the sensitivity of the test to evaluate between-individual variations. Our experiments also underline that SCE counts should include the centromere exchanges.