Geologic ultrafiltration—slowing of solutes with respect to flowing groundwater—poses a conundrum: it is consistently observed experimentally in clay-rich lithologies, but has been difficult to identify in subsurface data. Resolving this could be important for clarifying clay and shale transport properties at large scales as well as interpreting solute and isotope patterns for applications ranging from nuclear waste repository siting to understanding fluid transport in tectonically active environments. Simulations of one-dimensional NaCl transport across ultrafiltering clay membrane strata constrained by emerging data on geologic membrane properties showed different ultrafiltration effects than have often been envisioned. In relatively high-permeability advection-dominated regimes, salinity increases occurred mostly within membrane units while their effluent salinity initially fell and then rose to match solute delivery. In relatively low-permeability diffusion-dominated regimes, salinity peaked at the membrane upstream boundary and effluent salinity remained low. In both scenarios, however, only modest salinity changes (up to ∼3 g L⁻¹) occurred because of self-limiting tendencies; membrane efficiency declines as salinity rises, and although sediment compaction increases efficiency, it is also decreases permeability and allows diffusive transport to dominate. It appears difficult for ultrafiltration to generate brines as speculated, but widespread and less extreme ultrafiltration effects in the subsurface could be unrecognized. Conditions needed for ultrafiltration are present in settings that include topographically-driven flow systems, confined aquifer systems subjected to injection or withdrawal, compacting basins, and accretionary complexes.