This study investigated the effect of plant surface area (plant density) on the efficiency of organic carbon removal in a bench-scale constructed wetland. Constructed wetlands are commonly assumed to be biofilm reactors in which organic carbon removal occurs primarily through sedimentation and aerobic degradation by attached microbial biofilms. In conventional biofilm reactors, aerobic degradation of organic carbon is proportional to the amount of surface area for microbial attachment, provided that sufficient oxygen is available. In contrast, current design equations for constructed wetlands assume that the amount of surface area is not an important parameter. A bench-scale simulation of a constructed wetland was conducted, using bulrushes planted at varying plant densities in soil with a free water surface depth of about 0.27 m. The carbon source was diluted ENSUR (TM). Total organic carbon (TOC) removal was measured. Concentration of TOC was correlated with biochemical oxygen demand (BOD). Tests were conducted in conditions of light and dark, and under two different carbon loadings. Performance of bulrushes was compared with that of inert acrylic rods. The rate of carbon removal by mature bulrushes was found to increase with increasing plant density until oxygen became depleted. Higher densities degraded carbon at rates much faster than those predicted by current design equations. Young bulrushes degraded carbon at faster rates than mature bulrushes. Once oxygen was depleted, rates of degradation were reduced to rates anticipated by current models. When plant density was 15% or greater, oxygen became depleted in less than 6 hours. Removal efficiency was greater at higher loadings (70 mg/l BOD) than at lower loadings (25 mg/l BOD). Bulrushes performed significantly better than inert rods, sometimes by a full order of magnitude. The microbial community on the bulrushes appeared to be more complex and robust than that on the rods. Also, the presence of light did not significantly increase degradation rates for the bulrushes but was significant for the rods. The microbial community on the rods contained a larger proportion of epiphytic algae. The presence of light did result is greater overall efficiency of removal for both bulrush and rods. Currently, a major drawback of constructed wetlands in wastewater treatment has been their demand for large areas of land. This study suggests that it would be possible to reduce the land area requirements for constructed wetlands for both carbon removal and nitrification/denitrification provided designs gave more consideration to oxygen supply. Using current designs, a retention time of 4-8 days typically results in 70% BOD removal. This experiment suggests that wetlands with a retention time of about 1 day could provide the same performance if additional oxygen were supplied. / Graduation date: 1995