The efficacy and feasibility of annual harvesting of Phragmites australis and Typha latifolia shoots in autumn for nutrient removal was evaluated in a wetland system used for polishing sewage treatment plant (STP) effluent. Aboveground biomass and nutrient dynamics nutrient removal through harvest were studied in parallel ditches with stands of Phragmites or Typha that were mown in October during two successive years. The inflow rate of STP effluent to the ditches was experimentally varied, resulting in pairs of ditches with mean hydraulic retention times (HRT) of 0.3, 0.8, 2.3, and 9.3 days, corresponding to N and P mass loading rates of 122-4190 g N m(-2) yr(-1) and 28.3-994 g P m(-2) yr(-1). Nitrogen and P removal efficiency by harvest of Phragmites and Typha shoots in October increased with increasing HRT, despite the opposite HRT effect on N and P standing stocks. This removal through harvest appeared to be useful in treatment wetlands with N and P mass loading rates lower than approximately 120 g N m(-2) yr(-1) and 30 g P m(-2) yr(-1), corresponding to a HRT of roughly 9 days in the ditches of this wetland system. At the HRT of 9.3 days, the annual mass input to the ditches was reduced through the harvest by 7.0-11% and 4.5 -9.2% for N and P, respectively. At the higher nutrient mass loading rates, the nutrient removal through harvest was insignificant compared to the mass inputs. The vitality of Phragmites and Typha, measured as maximum aboveground biomass, was not affected by the annual cutting of the shoots in autumn over two years. The Typha stands yielded higher N and P removal efficiencies through shoot harvest than the Phragmites stands, which was largely the result of lower decreases in N and P standing stocks between August and October. This difference in nutrient standing stocks between the two species was caused by a combined effect of greater decreases in nutrient concentrations largely due to higher nutrient retranslocation efficiencies of Phragmites plants and greater reductions in shoot Phragmites biomass because of leaf fall and mass resorption. Nutrient removal by harvesting Phragmites shoots can probably be doubled without a reduction in vitality of the stands by advancing the harvest date to mid-September, which would at least approach the nutrient removal by harvesting Typha shoots in October. Phragmites also may be more profitable in very low-loaded wetland systems because the vigor of Typha stands seemed to be more sensitive to a lower nutrient availability at N and P mass input rates lower than the range indicated.