The effect of chenodeoxycholic acid, ursodeoxycholic acid and hyodeoxycholic acid on gallstone dissolution was studied in the prairie dog. Cholesterol gallstones were found in all animals after feeding a semipurified diet plus 1.2% cholesterol for six wk. Gallstone regression was examined by feeding a chow diet containing the bile acids (chenodeoxycholic acid, ursodeoxycholic acid or hyodeoxycholic acid) alone (30 mg/kg/day) or in combination (chenodeoxycholic acid plus ursodeoxycholic acid) for an additional six wk. Chenodeoxycholic acid was effective in dissolving established cholesterol gallstones (two out of 16 animals still had stones) and cholesterol crystals (six out of 16 animals had crystals); the hydrophilic bile acids, ursodeoxycholic acid and hyodeoxycholic acid, were ineffective in the six-wk regression study. The lithogenic indices averaged 1.09 at the end of the induction period; all biles became unsaturated with respect to cholesterol after the six-wk regression period (group 1, 0.82; group 2, 0.66; group 3, 0.81; group 4, 0.84; group 5, 0.66). Cholesterol levels in liver, plasma and bile were elevated after the six-wk induction phase (4.59 mg/g, 610 mg/dl and 0.36 mg/ml, respectively) but returned to near normal levels after the six-wk regression period. Biliary bile acids contained increased levels of the dietary bile acid administered to each group. This experiment shows that relatively hydrophobic bile acids may be more effective than hydrophilic bile acids for gallstone dissolution during the period studied.