In an attempt to clarify the controversy about sodium fluoride (NaF) clastogenicity, the induction of chromosome aberrations in Chinese hamster ovary cells (CHO) by NaF was investigated. Following a protocol used for screening chemicals for clastogenic activity, significant increases of aberrant cells were observed when cells were exposed to NaF for 4 h and harvested 8 h later. Cell-cycle kinetic studies demonstrated most cells were exposed in G2 of the cell cycle. Smaller increases in aberrant cells were observed when cells were harvested 20 h later (most cells were exposed in G1/S). The sensitivity of G2 cells to NaF was investigated further, along with the induction of aberrations at low doses. The results indicated that G2 cells are sensitive to NaF and the percent of aberrant cells increased with dose and length of exposure. With a 3-h exposure until harvest, no statistically significant increase in aberrant cells was observed at doses below 10 micrograms/ml NaF. These data are consistent with a threshold for NaF-induced clastogenicity around 10 micrograms/ml, as has been proposed previously (Scott and Roberts, 1987). It thus may be predicted that clastogenic effects would not occur in humans exposed to the levels of fluoride that are present in drinking water or dentifrices. An understanding of the mechanism of NaF-induced clastogenicity would help to clarify this point. It has previously been reported that NaF inhibits DNA synthesis/repair. The types of aberrations, mostly deletions and gaps, the induction of endoreduplicated cells, the cell-cycle delay and the sensitivity of G2 cells to NaF observed are similar to that reported in the literature for DNA synthesis/repair inhibitors like aphidicolin (APC). Similarities in the induction of aberrations by NaF and APC were confirmed in experiments with G2 cells. Based on these results and those previously reported for NaF and APC, it is proposed that NaF-induced aberrations may occur by an indirect mechanism involving the inhibition of DNA synthesis/repair.