The coefficient of variation for sulfate determinations = 1% at the 50-mg∙L−1 level. In the absence of sulfate, the chloride conductance response curve is linear for (0 < Cl < 100 mg∙L−1). The sulfate response curve is mildly nonlinear, being, however, adequately modeled using log K1 (SO42−) = 1.97 (25 °C), where K1 is the proton association constant. The temperature-dependent solubility product Ksp (AgCl) affects the blank correction for unbiased SO4 estimates. The bicarbonate conductance effect is effectively modeled using log K1 (HCO3−) = 6.35 (25 °C). The fluoride effect is adequately modeled using log K1 (F−) = 3.18 (25 °C) for aqueous mixtures of chloride, sulfate, and fluoride. In applying the cation-exchange method to sulfate analyses of alkaline high-Cl, high-F, geothermal waters, positively biased sulfate estimates were obtained; the bias depended on the SO4/F equivalence ratio and on the dilution level (of the natural water), and probably resulted from pH-dependent HF complexing with SiO2 or B(OH)3, both constituents being notably high in geothermal waters. Analytical strategies are discussed allowing procedural optimization for many natural waters.Key words: water analysis, cation exchange, sulfate determination, chloride determination, geothermal waters