Total phosphorus (TP) and nitrogen (TN) mass balances were calculated for Boney Marsh, a subtropical constructed freshwater wetland located along the floodplain of the Kissimmee River in south Florida, USA. River water was diverted through the marsh for a 9-year period (1978)(1979)(1980)(1981)(1982)(1983)(1984)(1985)(1986). Monthly mean retention rates were 0.03 and 0.41 g m -2 month-~ for TP and TN, respectively, for the period of record. Nutrient retention rates and nutrient loading rates were strongly correlated for TP but not for TN. Total phosphorus removal efficiencies were consistently higher than TN removal efficiencies at all times, and remained relatively unchanged during the entire study period. The subtropical marsh was a net positive sink for TP year-round but not for TN. Boney Marsh nutrient assimilation capacity remained high and invariable for the period of record for TP but not for TN. Our analysis showed that studies which derive nutrient removal estimates from reductions in surface water concentrations, rather than mass balances, may under-represent mass retention by as much as 50 and 100% for TP and TN, respectively. Rainfall contributions to Boney Marsh nutrient budgets could be as high as 99 and 94% for TP and TN, respectively, depending on the season. An apparent net settling velocity for total phosphorus of 9.93 m yeart provided independent confirmation of settling rates previously estimated for the Water Conservation Areas of southern Florida.