L5178Y wild-type and TK+/- (3.7.2c) cells were treated with sodium fluoride over a range of concentrations (10-500 micrograms ml-1) and treatment times (4, 16 and 48 h) covering less than 10-100% survival. The mutant frequency at five genetic loci (resistance to ouabain, 6-thioguanine, excess thymidine, methotrexate and 1-beta-D-arabinofuranosyl cytosine) was assayed in wild-type cells and trifluorothymidine in TK+/- cells. No significant induced mutation at any locus was observed after 4 h of treatment. Sixteen hours of treatment with high concentrations of sodium fluoride did not induce resistance to ouabain, but resulted in some significant induction of 6-thioguanine, 1-beta-D-arabinofuranosyl cytosine and methotrexate resistance, although the results were variable between experiments and no dose-response was observed. At the thymidine kinase locus, a dose-related increase in mutant frequency to excess thymidine and trifluorothymidine resistance was observed. The maximum induction was approximately eight times the control frequency after TK+/- cells were treated with the highly toxic concentration of 500 micrograms ml-1 of sodium fluoride for 16 h. These observations, and an analysis of the colony size of trifluorothymidine-resistant mutants in TK+/- cells, suggest that sodium fluoride is clastogenic to dividing cultured mammalian cells at high, toxic concentrations. Further work is desirable to investigate the mechanism by which chromosomes are damaged at high concentrations of fluoride, since without such a mechanistic understanding, extrapolation of our data to the human situation must be insecure. Nevertheless, the knowledge available at present gives no reason to expect any genotoxic effects in human tissues at levels of fluoride ions to which they are currently exposed in the general population.