A number of investigators have suggested that complexes may exist as discrete species on the exchange phase and, thus, be important to a complete description of exchange reactions. Little evidence establishing the presence of complexes on the exchange phase has hitherto been provided, however. Adsorption experiments in which mu-M F solutions were equilibrated with K-, Ca-, and Al-saturated cation-exchange resins demonstrated substantial F removal to the Al-saturated resin, with the adsorbed F quantitatively recovered with a single wash with 1.0 M NH4Cl, whereas no F removal was observed when either K or Ca occupied the exchange sites. Removal of both Al and F from Al-F solutions equilibrated with K- and Ca-saturated resins provided additional evidence for adsorption as fluoro-Al complexes. Modeling of the competitive complexation-exchange reactions with MINTEQA2 yielded results that under-predicted net Al sorption in the presence of F, and thus also inaccurately predicted solution composition, by neglecting potential adsorption of fluoro-Al complexes. Kaolinite (KGa-2) samples also sorbed appreciable F, though both surface OH displacement and sorption to exchange-phase Al were apparently responsible for the observed decrease in F solution concentrations. In the absence of exchange-phase Al, however, all sorbed F was fixed against displacement by 1.0 M NH4Cl. Samples of a Fuquay soil (loamy, siliceous, thermic Arenic Plinthic Paleudult) profile from South Carolina were also extracted with water, 0.01 M CaCl2, and 1.0 M NH4Cl; no detectable F was found in the water extracts, while considerable F was solubilized in the CaCl2 and NH4Cl extracts, with NH4Cl-extractable F being highly correlated with exchangeable Al (r2 = 0.98). These results suggest that part of the Al on the soil exchange phase may in fact exist as fluoro-Al complexes, and also imply a role of exchange-phase Al in regulating apparent F solubility within acid soils.