Non-DNA binding genotoxins (e.g., aneugens), unlike DNA-binding genotoxins, are theoretically expected to show thresholded concentration-effect response curves. This is a major issue in genetic toxicology testing because the identification of thresholds in vivo can provide a safety margin for exposure to a particular compound. In the current study we measured micronucleus induction by flow cytometry to determine the dose-response curves for tubulin interacting agents, a specific class of aneugens. All experiments with aneugens, which include colchicine, vinblastine, vincristine, as well as the clastogen cyclophosphamide (CP) were performed in mice to avoid the splenic elimination of micronucleated reticulocytes, which has been described in rats. Flow cytometry analysis revealed a non-linear dose-dependent increase in micronuclei frequencies for all tested aneugens, and a linear dose response curve for the clastogen, CP. To determine whether micronucleus induction at higher doses was due to chromosome loss (aneuploidy) or chromosome breakage (clastogenicity), flow sorting of the micronucleated reticulocytes and fluorescent in situ hybridization (FISH) with a mouse pan centromeric probe were performed for vinblastine, vincristine, and colchicine. Statistical evaluation of the flow cytometry and FISH data was performed to determine the threshold levels for chromosome loss in vivo. The threshold concentrations for vinblastine, vincristine, and colchicine were found at 0.35, 0.017, and 0.49 mg kg(-1), respectively.