Subsurface-flow constructed wetlands, sand filters, and peat filters near Duluth, Minnesota, were studied to determine their seasonal performance for removing pathogens from wastewater. Influent was a high-strength septic tank effluent (mean values of 5-day biochemical oxygen demand, total nitrogen, and total phosphorus were 294, 96, and 15 mg/L, respectively) at the Natural Resources Research Institute's alternative treatment system test facility in northern Minnesota. Each treatment system was inoculated with cultures of Salmonella choleraesuis (serotype typhimurium) for 5 to 7 consecutive days in summer and winter during 1998 to 1999. After the seeding, outflow samples were taken until Salmonella counts were sustained at background levels. The removal of Salmonella was calculated for each system, although the exact removal mechanisms were not determined. During the summer, the wetlands removed 99.6 to 99.999 4% (2.4 to 5.3 log 10 reduction) of the culturable Salmonella. The sand filters demonstrated a greater than 7 log 10 removal of Salmonella cells, whereas the peat filters were responsible for a greater than 8 log 10 loss of cells. Fewer Salmonella cells were removed by all of these systems during the winter, although the pattern of removal was similar to their summer operation. During the winter, the wetlands and sand filters removed greater than 1 log 10 of culturable cells, but the peat filters were responsible for a greater than 5 log 10 loss of cells. Fecal coliform removal patterns reflected those for Salmonella by treatment systems for summer and winter periods. Based on Salmonella and fecal coliform removal, the peat filters operated most effectively followed by the sand filters and the constructed wetlands. Water Environ. Res., 73, 204 (2001).