The in vitro hepatocyte micronucleus assay was tested for its practicability and its usefulness in detecting mutagens. The assay protocol developed by Alati et al. (1989) was shown to give reproducible levels of proliferating hepatocytes and the formation of micronuclei could be readily assessed by fluorescence microscopy. Epidermal growth factor and insulin were used as mitogens, yielding mitotic indices of 2.4 +/- 0.74% after 72 h of culture. The high number of 8.0 +/- 3.33% micronucleated hepatocytes in control cultures at that time, typically for in vitro stimulated hepatocytes, is probably due to disordered mitoses frequently leading to chromosome loss. The direct acting mutagen N-methyl-N'-nitro-N-nitrosoguanidine and the clastogens cyclophosphamide and retrorsine, which require metabolic activation, induced dose dependent increases in the frequencies of micronucleated hepatocytes. The carcinogen 2-AA.F also yielded significantly enhanced rates of micronuclei. The non-mutagen KCI as well as the peroxisome proliferator clofibrate, which is considered to be a non-genotoxic hepatocarcinogen, yielded consistently negative results. Problems occurred when chemicals exerting strong cytotoxic effects were tested in this assay. The mutagen and hepatocarcinogen aflatoxin B1 did not enhance the number of micronucleated hepatocytes. Rather a reduction of micronuclei and of mitoses was observed at AFB1 concentrations considered positive in other genotoxicity assays. Hepatocyte proliferation seems to be highly susceptible to the cytotoxic action of chemicals. A decrease in the proliferating activity of hepatocytes can obviously prevent the detection of mutagenic effects. Further studies on the in vitro hepatocyte micronucleus assay are necessary to clarify its role in mutagenicity testing.