We evaluated the effect of four hydraulic retention times (HRT, 0.3, 0.8, 2.3, and 9.3 days) on pollutant removal in a surface-flow wetland system for polishing tertiary effluent from a sewage treatment plant (STP). The removal efficiency of pollutants at these HRTs was based on mass budgets of the water inputs and outputs in parallel ditches, which together with a presettling basin, made up the wetland system. Fecal coliform and N-removal efficiencies in the ditches were enhanced by increasing the HRT, with only little removal of fecal coliforms during spring-summer at a HRT of 0.3 days. A HRT of 4 days turned out to be required to meet the desired bathing water standard for fecal coliforms (10³ cfu 100 ml⁻¹) and the future standard of ammonium (1 mg N l⁻¹) all year. An annual N-removal efficiency of approximately 45% can be accomplished in the ditches at this HRT, corresponding to an annual N mass loading rate of 150 g N m⁻² yr⁻¹. Annual P removal was not improved by increasing the HRT even up to 9.3 days, largely because of the still high P mass loading rate (14 g P m⁻² yr⁻¹) in combination with relatively low P input concentrations. Substantial P removal can probably only be achieved at HRTs longer than 15 days, which will not be feasible for the situation investigated because of the large land area that would be required to reach such long HRTs. The future P standard (1 mg P l⁻¹) can therefore only be met by additional chemical P removal. In a densely populated country such as the Netherlands, adequate polishing of tertiary STP effluent in surface-flow wetlands with similar goals as for this wetland is restricted to small and medium-sized STPs. The simultaneous use of these treatment wetlands for other functions, such as nature conservation, recreation, and flood control, however, would permit the use of relatively larger land areas.