A 10-mM LiCl solution was hydraulically forced through a 0.3-cm-thick smectite membrane at a constant solution flux in order to depict the chemical and isotopic evolution of analogous hyperfiltration cells in the subsurface. At steady state (~10 days) there was, relative to the input solution, a measured 12% buildup of solute within the entire cell and a calculated 248% solute increase at the clay/solution interface. The maximum depletion of 7Li (about 13% relative to the input solution) was calculated to occur about 1 cm away from the membrane out into the high-pressure reservoir. When scaled to realistic solution fluxes in natural systems, the position of maximum fractionation would occur meters away from a shale/aquifer contact. A steady state hyperfiltration model indicates that a solute's heavy isotopic species will always be depleted on the high-pressure side of a clay membrane, a prediction corroborated not only by the Li experiment but also by previous hyperfiltration studies involving stable isotopes of C and Cl. The extent of fractionation increases with higher values of the membrane's reflection coefficient, &sgr;. In the experiment a &sgr; of 0.55 for the 90% porosity clay membrane yielded these results. Because shales have substantially lower porosities with attendant higher values of &sgr;, we conclude that the hyperfiltration phenomenon in the subsurface may result in sampling of waters that are not chemically and isotopically representative of the formation penetrated by the well.