Reliable experimental data required to test hydrodynamic dispersion/chemical reaction models are scarce. This paper provides such a dataset based on absorption of gypsum solution by horizontal columns of relatively dry soil with an initially high exchangeable potassium ratio. Initial and boundary conditions are well defined. Water and cation concentration profiles measured after 200 and 400 min lay on single curves when graphed in terms of distance divided by the square root of time. Cation exchange occurred close to the intake surface and calcium derived from the gypsum was confined to a narrow band well behind the notional piston front that separates the absorbed solution from that originally present. Anion exchange was negligible and the solution concentration up to the piston front approximated the anion concentration of the invading solution. The interval between the region of cation exchange and the piston front maintained the original cation adsorption ratios but at a total cation solution concentration approximating that of saturated gypsum (~25 mmol_c/L). Some implications of this phenomenon are discussed. Comparison of cation exchange isotherms observed when the virgin soil absorbs effluent-like solutions and when effluent-irrigated soil absorbs saturated gypsum suggest that, operationally, these isotherms may be considered to be unaffected by hysteresis in the exchange reactions.