The relationship between occupational or environmental exposure to asbestos and the development of mesothelioma, typically after prolonged latency, has been accepted as one of cause and effect. Most studies have concluded that asbestos is not mutagenic to mammalian cells in vitro. We have studied the potential of crocidolite asbestos to induce mutations in a stable mesothelioma cell line, using a mutation assay that measures mutation at the autosomal HLA-A locus and permits clonal growth of mutant cells. The mesothelioma cell line chosen is more akin to the in vivo target cells of asbestos than human peripheral blood lymphocytes used in previous studies. Exposure of mesothelioma cells in culture to both 200 micrograms/ml and 50 micrograms/ml crocidolite for 72 hr did not result in a statistically significant difference in the mutation frequency (MF) in the HLA-A assay when compared to the spontaneous MF in these cells. Mutations in the mesothelioma cells were classified according to their molecular basis. Notwithstanding the lack of statistically significant change in overall MF, molecular analysis of mutants obtained following exposure of mesothelioma cells to crocidolite demonstrated a statistically significant increase in the class of mutations arising from loss of heterozygosity (LOH) events involving the selection locus (HLA-A) and more distal loci. Mutations following exposure to 200 micrograms/ml and 50 micrograms/ml crocidolite showed a greater frequency of LOH than did spontaneous mutants (P < 0.01 and P < 0.001, respectively). These results correlate with those obtained in an earlier study using lymphocytes. The mesothelioma cell-based assay may be useful in detecting the mutagenicity of other asbestiform fibers and man-made fibers.