Pesticides may be released to farmyard surfaces as a result of spillages, leakages, and the decontamination of tractors and sprayers. Biobeds can be used to intercept and treat contaminated runoff, thus minimizing losses to the environment. Previous studies using lined and unlined biobeds showed that water management was the limiting factor for both systems. While lined biobeds effectively retained pesticides, the system rapidly became water logged and degradation was slow. Studies using unlined biobeds showed that >99% of the applied pesticides were removed by the system, with a significant proportion degraded within 9 months. However, peak concentrations of certain pesticides (Koc < 125) were unacceptable to the regulatory authorities. These experiments were designed to optimize the design and management of unlined biobeds. Experiments performed to investigate the relationship between biobed depth and water loading showed that biobeds need to have a minimum depth of 1-1.5 m. The surface area dimension of the biobed depends on the water loading, which is controlled by the nature and frequency of pesticide handling activities on the farm. Leaching losses of all but the most mobile (Koc < 15) pesticides were <0.32% of the applied dose from 1.5 m deep biobeds subject to a water loading of 1175 L m(-2). These were reduced to <0.06% when a water loading of 688 L m(-2) was applied and down to <0.0001% for a water loading of 202 L m(-2). On the basis of these data, a 1.5 m deep biobed, subject to a maximum water loading of 1121 L m(-2) and with a surface area of 40 m(2) should be able to treat < or =44000 L of pesticide waste and washings such that the average concentration of all pesticides, other than those classified as very mobile, does not exceed 5 microg L(-1). This level of treatment can be improved by further reduction in the hydraulic loading.