The applicability of cation exchange equations and the Nutrient Cycling Model (NuCM) to a beech ( Fagus grandifolia Ehrh.) forest soil in the Great Smoky Mountains, North Carolina, was tested by examining temporal variations in soil solution chemical composition. Both field data and NuCM output implied that short‐term, local variations in exchangeable Ca 2+ and Al 3+ occurred, causing non‐ideal plots of pAl vs. pCa in soil solution. Thus, plots of pCa vs. pAl, with their inherent assumptions of constant exchangeable cations, do not provide an adequate test of the applicability of cation exchange equations to field conditions. A large NO 3− pulse caused sequential peaks in soil solution Mg 2+ , Ca 2+ , and Al 3+ , which became more separated with soil depth. The NuCM model did not reproduce this pattern, showing instead simultaneous peaks of all cations in response to NO 3− pulses. As a consequence, NuCM underestimated peak Al/Ca ratios. The disparity was probably due to the coarse scale at which NuCM and models like it represent the soil. A chromatographic model may be more appropriate for this time scale.