The in vivo micronucleus assay working group of the International Workshop on Genotoxicity Testing (IWGT) discussed new aspects in the in vivo micronucleus (MN) test, including the regulatory acceptance of data derived from automated scoring, especially with regard to the use of flow cytometry, the suitability of rat peripheral blood reticulocytes to serve as the principal cell population for analysis, the establishment of in vivo MN assays in tissues other than bone marrow and blood (for example liver, skin, colon, germ cells), and the biological relevance of the single-dose-level test. Our group members agreed that flow cytometric systems to detect induction of micronucleated immature erythrocytes have advantages based on the presented data, e.g., they give good reproducibility compared to manual scoring, are rapid, and require only small quantities of peripheral blood. Flow cytometric analysis of peripheral blood reticulocytes has the potential to allow monitoring of chromosome damage in rodents and also other species as part of routine toxicology studies. It appears that it will be applicable to humans as well, although in this case the possible confounding effects of splenic activity will need to be considered closely. Also, the consensus of the group was that any system that meets the validation criteria recommended by the IWGT (2000) should be acceptable. A number of different flow cytometric-based micronucleus assays have been developed, but at the present time the validation data are most extensive for the flow cytometric method using anti-CD71 fluorescent staining especially in terms of inter-laboratory collaborative data. Whichever method is chosen, it is desirable that each laboratory should determine the minimum sample size required to ensure that scoring error is maintained below the level of animal-to-animal variation. In the second IWGT, the potential to use rat peripheral blood reticulocytes as target cells for the micronucleus assay was discussed, but a consensus regarding acceptability for regulatory purposes could not be reached at that time. Subsequent validation efforts, combined with accumulated published data, demonstrate that blood-derived reticulocytes from rats as well as mice are acceptable when young reticulocytes are analyzed under proper assay protocol and sample size. The working group reviewed the results of micronucleus assays using target cells/tissues other than hematopoietic cells. We also discussed the relevance of the liver micronucleus assay using young rats, and the importance of understanding the maturation of enzyme systems involved in the processes of metabolic activation in the liver of young rats. Although the consensus of the group was that the more information with regard to the metabolic capabilities of young rats would be useful, the published literature shows that young rats have sufficient metabolic capacity for the purposes of this assay. The use of young rats as a model for detecting MN induction in the liver offers a good alternative methodology to the use of partial hepatectomy or mitogenic stimulation. Additional data obtained from colon and skin MN models have been integrated into the data bases, enhancing confidence in the utility of these models. A fourth topic discussed by the working group was the regulatory acceptance of the single-dose-level assay. There was no consensus regarding the acceptability of a single dose level protocol when dose-limiting toxicity occurs. The use of a single dose level can lead to problems in data interpretation or to the loss of animals due to unexpected toxicity, making it necessary to repeat the study with additional doses. A limit test at a single dose level is currently accepted when toxicity is not dose-limiting.