Submerged aquatic vegetation (SAV) communities exhibit phosphorus (P) removal mechanisms not found in wetlands dominated by emergent macrophytes. This includes direct assimilation of water column P by the plants and pH-mediated P coprecipitation with calcium carbonate (CaCO3). Recognizing that SAV might be employed to increase the performance of treatment wetlands, we investigated P removal in mesocosms (3.7 m2) stocked with a mixture of taxa common to the region: Najas guadalupensis, Ceratophyllum demersum, Chara spp. and Potamogeton illinoensis. Three sets of triplicate mesocosms received agricultural runoff from June 1998 to February 2000 at nominal hydraulic retention times (HRTs) of 1.5, 3.5 or 7.0 days. Mean total P (TP) loading rates were 19.7. 8.3 and 4.5 g/m2/yr. After eight months of operation. N. guadalupensis dominated the standing crop biomass and P storage, whereas C. demersum exhibited the highest tissue P content. Chara spp. was prominent only in the 7.0)-day HRT treatments while P. illinoensis largely disappeared. Inflow soluble reactive phosphorus (SRP) (10 163 microg/L) was reduced consistently to near the detection limit (2 microg/L) in the 3.5- and 7.0-day HRT treatments, and to a mean of 9 microg/L in the 1.5-day HRT treatment. The mean inflow TP concentration (10(7) microg/L) was reduced to 52, 29 and 23 microg/L in the 1.5-, 3.5- and 7.0-day HRT treatments, respectively. Total P concentrations in new sediment (mean= 641, 408 and 459 mg/kg in the 1.5-. 3.5-, and 7.0-day HRT mesocosms, respectively) were much higher than in the muck soil used to stock the mesocosms (236 mg/ kg). The calcium content of new sediment was twice that of the muck soil (16.5% vs. 7.6%), demonstrating that CaCO3 production and, perhaps, coprecipitation of P occurred. We observed no nocturnal remobilization of SRP despite diel fluctuations in pH and dissolved oxygen. Mean outflow TP (21 microg/L) from a 147 ha SAV wetland (4-day nominal HRT) was similar to mean outflow TP in the 3.5-day and 7.0-day HRT treatments. The mesocosms adequately mimicked P removal and other important characteristics of the larger system and can be used to address research questions regarding treatment performance of full-scale SAV wetlands. Available data suggest that the incorporation of SAV communities into the stormwater treatment areas may benefit Everglades restoration.