Lateral mass redistribution in soils is the key to understanding field-scale solute transport, but the underlying assumptions of common transport theories are violated under transient-flow conditions. We tested the applicability of two limiting cases for solute spreading under transient-flow conditions in the field after an appropriate time coordinate transformation: no lateral mass redistribution, described by the convective-lognormal transfer function (CLT), vs. perfect lateral mass redistribution, described by the convection–dispersion equation (CDE). A Br− transport experiment was performed in six zero-tension lysimeters and in the field for almost 3 yr under atmospheric conditions. Sampling in the field was performed by extracting soil cores during seven campaigns. According to time-domain reflectometry readings, only slight variations in water content were measured in space and time in the lysimeters. In contrast, water content was lower and more variable in the plow layer in the field. The variance of solute spreading was better predicted by the CDE assuming perfect lateral mass redistribution. This hints at the importance of molecular diffusion. Both models have the flexibility to fit the flux-averaged breakthrough curve in the lysimeters and the averaged concentration profiles in the field, but not with one set of parameters. The CLT parameters obtained from the lysimeter experiment better predicted the measured concentration profiles in the field for shorter times, but both models failed for longer times. Due to the occurrence of local saturation at the lower boundary of zero-potential lysimeters, differences in water regimes hamper the transferability of transport parameters from lysimeters to the field.