The role of macropores in infiltration through a sandy loam was studied using laboratory columns pretreated with water possessing a δ 1 8 O signature of -8.7‰ and 15 mg l - 1 Cl - . A simulated snowmelt pulse of δ 1 8 O-depleted water containing 1100 mg l - 1 Cl - was added to a control column and two columns containing a single vertical macropore, one continuous and the other discontinuous. Macropores were formed in situ by disintegration of a biodegradable foam thread inserted during column packing. Macropores were 2 mm in diameter, which has been suggested to be the threshold for significant macropore flow given the soil's mean textural pore diameter of 0.41 mm. Meltwater was flushed from the columns by adding two pore volumes of isotopically enriched water containing 15 mg l - 1 Cl - at a rate of 17.2 mm day - 1 . Tensiometers, time domain reflectometry probes and suction samplers were used to monitor matric potential, soil water content and soil water chemistry at 0.1 m intervals down the columns. Column effluent was sampled daily for δ 1 8 O and Cl - . Mobile soil water contents (θ m ) and dispersivities (ε) were estimated by fitting a one-dimensional analytical solution of the convection-dispersion equation to Cl - breakthrough curves (BTCs). θ m increased with depth in all columns, whereas only the discontinuous macropore column showed an increase in ε with flow length. Cl - and δ 1 8 O breakthrough occurred earlier at all depths in the macropore columns relative to the control, resulting in larger ε values for macroporous soil. ε for a given flow length tended to be greatest in the discontinuous macropore column, reflecting the role of internal catchment processes. Macropore presence was associated with decreased θ m during infiltration and bimodal BTCs in column effluent. The threshold ratio of macropore-to-micropore diameters at which macropores exert a detectable influence on water transport must be less than that examined here.