Most biosphere and contamination assessment models are based on uniform soil conditions, since single coefficients are used to describe the transfer of contaminants to the plant. Indeed, physical and chemical characteristics and root distribution are highly variable in the soil profile. These parameters have to be considered in the formulation of a more realistic soil-plant transfer model for naturally structured soils. The impact of monolith soil structure (repacked and structured) on Zn and Mn uptake by wheat was studied in a controlled tracer application (dye and radioactive) experiment. We used Brilliant Blue and Sulforhodamine B to dye flow lines and 65Zn and 54Mn to trace soil distribution and plant uptake of surface-applied particle-reactive contaminants. Spatial variation of the soil water content during irrigation and plant growth informs indirectly about tracer and root location in the soil profile. In the structured monolith, a till pan at a depth of 30 cm limited vertical water flow and root penetration into deeper soil layers and restricted tracers to the upper third of the monolith. In the repacked monolith, roots were observed at all depths and fingering flow allowed for the fast appearance of all tracers in the outflow. These differences between the two monoliths are reflected by significantly higher 54Mn and 65Zn uptake in wheat grown on the structured monolith. The higher uptake of Mn can be modelled on the basis of radionuclide and root distribution as a function of depth and using a combination of preferential flow and rooting. The considerably higher uptake of Zn requires transfer factors which account for variable biochemical uptake as a function of location.