AbstractA series of experiments designed to demonstrate the potential of using managed, attached algal production to permanently remove excess phosphorus from agricultural run‐off is described. The experiments were carried out on a secondary canal in the New Hope South region of the Florida Everglades Agricultural Area from October, 1991, to May, 1992. Natural algal populations of periphyton, including species of the generaCladophora, Spirogyra, Enteromorpha, Stigeoclonium, and a variety of filamentous diatoms such asEunotiaandMelosira, were grown on plastic screens in raceways, under a wave surge regime. Considerable biomass production of algae occurred, and the resulting algal canopy also trapped plankton and organic particulates from the water column. A seven‐ to eight‐day harvest interval was determined to be optimal, and both hand harvesting and vacuum harvesting were employed. The vacuum device is applicable to large scale‐up. In source water having total phosphorus concentrations of 0.012–0.148 ppm, mean macro‐recovery dry biomass production levels of 15–27 g/m2/day were achieved. The lower rates occurred in the winter, the higher rates in the late spring. Two techniques were employed to reduce losses of fine material at harvest during the March to May period. Gravity sieving increased mean dry production levels to 33–39 g/m2/day. The mean phosphorus content of harvested biomass ranged from 0.34% to 0.43%. Total phosphorus removal rates during the spring period of average solar intensity and low nutrient supply, by methods demonstrated in this study, ranged from 104 to 139 mgTP/m2/day (380–507 kgP/ha/year). Over the incoming nutrient range studied, phosphorus removal was independent of concentration and was 16.3% of total phosphorus for 15 m of raceway. Up‐stream‐downstream studies of overflowing water chemistry (total P, total dissolved ‐P, orthophosphate ‐P) showed highly ‐significant reductions of all phosphorus species. Total phosphorus reduction closely correlated with phosphorus yield from biomass removal. Yearly, minimum phosphorus removal rates are predicted that are 100–250 times that achieved both experimentally and in long‐term, large‐area wetland systems. Engineering scale‐up to systems of