Filling and cleaning of pesticide sprayers presents a potential risk of pollution of soil and water. Three different solutions for handling sprayers have been suggested: Filling and cleaning in the field, filling and cleaning on hard surfaces with collection of the waste water, and filling and cleaning on a biobed, which is an excavation lined with clay and filled with a mixture of chopped straw, sphagnum and soil with turf on top, and with increased sorption capacity and microbial activity for degradation of the pesticides. In the present study the degradation and leaching of 21 pesticides (5 g of each) was followed in an established full-scale model biobed. Percolate was collected and analysed for pesticide residues, and the biobed material was sampled at three different depths and analysed by liquid chromatography double mass spectrometry (LC-MSMS). During the total study period of 563 days, no traces of 10 out of 21 applied pesticides were detected in the percolate (detection limits between 0.02 and 0.9 microg l(-1)) and three pesticides were only detected once and at concentrations below 2 microg l(-1). During the first 198 days before second application, 14% of the applied herbicide bentazone was detected in the leachate with maximum and mean concentrations of 445 and 172 microg l(-1), respectively. About 2% of the initial mecoprop and fluazifop dose was detected in the percolate, with mean concentrations of 23 microg l(-1), while MCPA and dimethoate had mean concentrations of 3.5 and 4.7 microg l(-1), respectively. Leachate concentrations for the remaining pesticides were generally below the detection limit (0.02-0.9 microg l(-1), below 1% of applied). Sorption studies of five pesticides showed that compounds with a low K(d) value appeared in the leachate. After 169 days, all pesticides in the biobed profile were degraded to a level below 50% of the calculated initial dose. Pesticides with K(oc) values above 100 were primarily found in the uppermost 10 cm and degraded slowest due to the low bioavailability. The 11 most degradable pesticides were all degraded such that less than 3% remained in the biobed after 169 days. Following second pesticide application of the biobed, leachate was sampled 215 and 365 days after the treatment. This showed the same pesticides to be leached out and at concentrations comparable to those of the first treatment. The same pesticides as after the first treatment were retained in the biobed.