In Germany, zero-tension lysimeters are used as part of the registration requirements in case pesticides pose a potential threat to contaminate the groundwater. However, the water regime and the method of pesticide sampling differ between the lysimeters and the field. We monitored the transport of the two herbicides ethidimuron [1-(5- ethylsulfonyl-1,3,4-thiadiazol-2-yl)-1,3-dimethylurea] (ETD) and methabenzthiazuron [1-benzothiazol-2-yl-1,3-dimethyl-urea] (MBT) and their main metabolite, accompanied with bromide as conservative tracer, in zero-tension lysimeters filled with undisturbed soil and in the field. The herbicides were applied as a short pulse to the bare soil surface. Herbicide concentrations were analyzed in the drainage water of the 1.2-m-deep lysimeters and from soil cores taken from the field during six campaigns. Soil coring in the field emphasized matrix flow and allowed us to estimate the field-based dissipation and sorption parameters. Based on mass recovery calculations, the field fate half-life was 870 d for ETD compared with 389 d for its main metabolite. The initially fast field-based dissipation of MBT with a half-life value of approximately 1 mo was followed by a much slower dissipation. The retardation factor was estimated from the concentration profiles by inversely solving the convection-dispersion equation and yielded 18.2 +/- 1.3 for ETD and 36.9 +/- 17.5 for MBT. For the lysimeters, a leaching period of 2 1/2 yr was too short to monitor bulk herbicide mass through the soil matrix. Only 1.7% of the applied EDT and 1.4% of the applied MBT were sampled in the drainage water at 1.2 m depth. Despite contrasting sorption and dissipation properties, both herbicides appeared fast and at the same time in the drainage water, hinting at preferential flow phenomena. Compared with field fate of herbicides measured by soil coring, zero-potential lysimeters emphasize the transport of small amounts of herbicides triggered by preferential flow events.