In this article, the relationship was studied between the in vivo mutagenicity assay of mouse bone marrow micronucleus (MIC), and four in vitro assays:Salmonella typhimurium, chromosomal aberrations in Chinese hamster ovary (CHO) cells, sister chromatid exchanges (SCE) in CHO cells, and mutation in mouse lymphoma L5178Y cells. A comparison with the rodent carcinogenicity data was also undertaken. The MIC data on 49 chemicals were generated by Shelby et al. (1993). The MIC assay system employed three daily exposures by intraperitoneal injection; bone marrow samples were obtained 24 h following the final exposure. A preliminary analysis indicated that the 49 chemicals selected by Shelby et al. (1993) are a representative subset of the National Toxicology Program database. This study showed that MIC has a number of particular characteristics that are not shared by other biological systems. MIC is basically different from rodent carcinogenicity, despite being an in vivo system. At the same time, it responds to the chemicals in a different way from that of the in vitro genotoxicity assays. These in vitro assays mainly differ from each other in their different sensitivities to the genotoxins: MIC gives just a few positives (limited sensitivity), but, at the same time, some of these positives are detected only by the most sensitive assays, like the mouse lymphoma or SCE assays. In terms of risk assessment, MIC does not complementSalmonellafor predicting chemical carcinogenicity, and would be better used to verify if the in vitro positive chemicals are able to exert their genotoxic potential in vivo.