Groups of male B6C3F1 mice (N = 50) were provided drinking water containing 2 g/liter sodium chloride (control) and 0.05, 0.5, and 5 g/liter dichloroacetic acid (DCA). Treatment of 30 animals in each group was carried out to 60 or 75 weeks. In a separate experiment, mice exposed to 3.5 g/liter DCA and the corresponding acetic acid control group were killed at 60 weeks. Groups of 5 mice were killed at 4, 15, 30, and 45 weeks. Time-weighted mean daily doses of 7.6, 77, 410, and 486 mg/kg/day were calculated for 0.05, 0.5, 3.5, and 5 g/liter DCA treatments. Animals exposed to 3.5 and 5 g/liter DCA had final body weights that were 87 and 83%, respectively, of the control value. Relative liver weights of 136, 230, and 351% of the control value were measured for 0.5, 3.5, and 5 g/liter, respectively. At 60 weeks mice receiving 5.0 g/liter DCA had a 90% prevalence of liver neoplasia with a mean multiplicity of 4.50 tumors/animal. Exposure to 3.5 g/liter DCA for 60 weeks resulted in a 100% tumor prevalence with an average of 4.0 tumors/animal. The prevalence of liver neoplasia and tumor multiplicity at 60 and 75 weeks in the 0.05 g/liter DCA (24.1%; 0.31 tumors/animal) and in the 0.5 g/liter group (11.1%; 0.11 tumors/animal) did not differ significantly from the control value (7.1% and 0.07 tumors/animal). No liver tumors were found in the group treated with acetic acid. Hyperplastic nodules were seen in the 3.5 (58%; 0.92/animal) and 5 g/liter DCA groups (83%; 1.27/animal). There was a significant positive dose-related trend in the age-adjusted prevalence of liver tumors. These data confirm the hepatocarcinogenicity of DCA administered in the drinking water to male B6C3F1 mice for 60 weeks. The results together with those in an earlier report from this laboratory suggest, for the conditions under which these assays were conducted, a threshold concentration of at least 0.5 g/liter followed by a steep rise to a maximum tumor incidence at 2 g/liter DCA.