This report presents an assessment made by the U.S. Environmental Protection Agency Gene-Tox Program's Work Group on mammalian cytogenetics of the clastogenic effects of chemicals in in vivo and in vitro mammalian cell assays. This assessment is based on information provided by the Environmental Mutagen Information Center, Oak Ridge National Laboratory, with the proviso that the experimental protocol used in these papers was adjudged to be acceptable by standards outlined by the Work Group. Some data were accepted as "qualitative only" because the protocol used was fairly close to that proposed as suitable. Using these criteria, 177 papers were selected for review. 6 assays were reviewed: bone marrow (32 papers, 31 chemicals), spermatogonial (10 papers, 10 chemicals), spermatocyte (25 papers, 25 chemicals), oocyte or early embryo (18 papers, 19 chemicals), in vitro cell culture (30 papers, 66 chemicals), and leukocyte (66 papers, 53 chemicals). Each assay was considered separately, and comparisons were then made between them for their similarities or differences in producing a positive or negative clastogenic effect of a particular chemical or chemical class. A large proportion of the available cytogenetic data was not suitable for inclusion in the final data base because of poor experimental design or unsatisfactory reporting of the information. It was not possible to recommend any one assay for determining potential clastogenicity because each had its own particular advantages and limitations and provided unique information. For demonstrating in vivo effects, the bone-marrow assay is probably the simplest and most economical. If only in vitro exposures were considered, leukocytes or cultured mammalian cell lines would be suitable. However, there are advantages to using leukocytes because they are a synchronous population, at least through their cell division, and because of the ready availability of human cells. In general, there was good agreement between clastogenicity and carcinogenicity.