Saline groundwaters with 4.7 to 26-parts per thousands total dissolved solids were sampled from springs and artesian wells in Mississippian and Ordovician carbonates and sandstones in central Missouri. U-Th isotopic variations provide a means of evaluating processes of water-rock interaction and fluid mixing and estimating the time scales of element transport. Recently developed mass spectrometric techniques are used to make isotopic measurements on small-volume groundwater samples (0.1-4 1) with high precision (e.g., < +/- 5-parts-per-thousands for U-234/U-238 activity ratios). The groundwaters have a wide range of U-238 concentrations, 50 x 10(-12) to 200 x 10(-12) g/g; U-234/U-238 activity ratios, 2.15-16.0; Th-232 concentrations, 0.10 x 10(-12) to 33 x 10(-12) g/g; and Th-230 concentrations, 0.91 x 10(-17) to 26 x 10(-17) g/g. Unfiltered and filtered (0.4-mu-m, 0.1-mu-m) aliquots of a saline sample have the same isotopic composition and concentration of U, indicating that U-234 and U-238 occur almost entirely as dissolved species. The concentration of Th-232 is up to seven times lower in filtered vs. unfiltered aliquots, indicating that Th-232 is predominantly associated with particulates in the groundwaters. In contrast, most of the Th-230 is in solution. Previous geochemical studies indicate that: (1) the saline waters originated as meteoric recharge and evolved through halite dissolution, reactions with silicates and saline-dilute mixing processes during a long-distance flow history; and (2) interaction with limestone and dolomite aquifer rocks in central Missouri has been limited. A consistent relationship between U/Ca and U-234/U-238 activity ratio is observed in the groundwaters and provides constraints on the U/Ca ratios and U-234/U-238 activity ratios of end-member reservoirs and on the processes of isotopic exchange in this water-rock system. Model calculations that simulate (1) saline-dilute groundwater mixing and (2) limited extents of dissolution of carbonate aquifer minerals by the groundwaters can account for the variations in U/Ca, U-234/U-238 and O-18/O-16 in the suite of water samples. The model calculations demonstrate that dissolved U isotopic compositions can be a sensitive indicator of water-rock interaction, which in turn limits the usefulness of U-234-U-238 disequilibria for groundwater age determinations. The concentration of dissolved Th-230 in the groundwaters is (1) two to three orders of magnitude below Th-230-U-234 equilibrium activity levels, and (2) significantly in excess of concentrations estimated for the supply of Th to solution via desorption and dissolution. A model involving the derivation of the excess Th-230 from the in situ decay of dissolved U-234 in the groundwaters indicates the operation of an adsorption mechanism on the time scale of 10-10(3) years. The results reported here may have broader application to the assessment and management of hazardous chemical species in natural environments.