Water quality in the interior marsh of the Arthur R. Marshall Loxahatchee National Wildlife Refuge is characterized by low concentrations of major ions, principally sodium and chloride, and is affected primarily by natural seasonal processes, such as evapotranspiration, rainfall, and biological activity. During the dry season, evapotranspiration exceeds precipitation, and specific conductance and conservative ion concentrations at marsh background sites typically increase by 40-70 percent between the end of the rainy season in September and the end of the dry season in May. Water enters the Refuge mainly from rainfall and perimeter canals. Water is pumped into the perimeter canals from large pumping stations, such as S-5A and S-6. In recent years, much of the water pumped into the Refuge passes through Stormwater Treatment Areas (STAs) before being released into the perimeter canals that surround the Refuge. Since 2001, water at S-6 has been diverted south toward STA-2, away from the Refuge perimeter canals. Water from S-5A and S-6 flows through agricultural lands with intense agricultural activity and typically contains relatively high concentrations of major ions, nutrients, and pesticides. Specific conductance, major-ion concentrations, and nutrient concentrations are an order of magnitude higher at S-5A and S-6 canal sites than at interior marsh sites. Water quality in the marsh bordering the canals can be affected substantially by the canal water, and these effects can extend several miles or more into the marsh depending on location in the Refuge and on the water level in the canals. As canal water flows into the marsh, processes such as uptake by periphyton and rooted vegetation and settling of particulate matter reduce the concentrations of nutrients to a greater extent than conservative ions such as chloride. Long- and short-term trends for specific conductance, chloride ion, sulfate ion, total phosphorus, and total nitrogen at five sites were evaluated primarily using an uncensored seasonal Kendall test with a water-level adjustment to reduce the effects of long wet or dry periods. Significant long-term trends (1974-2003) for specific conductance, chloride, total phosphorus, and total nitrogen at canal sites S-5A and S-6 were generally downward. Of the five sites, S-5A had the most pronounced decline for specific conductance at about -340 microsiemens per centimeter (?S/cm), followed by S-6 with a decline of about -280 ?S/cm. The two internal marsh sites, LOX8 and LOX13, had significant long-term trends in specific conductance of about +37 and -36 ?S/cm, respectively. Long-term trends for other constituents at the two internal marsh sites were generally small in magnitude or not measurable between 1978 and 2003. Marsh site LOX15 near the Hillsboro Canal showed no long-term trends, although specific conductance and sulfate concentration increased about 560 ?S/cm and 30 milligrams per liter, respectively, from 1998 to 2002. Site LOX15 is influenced strongly by intrusions of canal water, and increases in specific conductance and sulfate at this site coincided with increased canal-water inflows from STA-1W between 2001 and 2003. Median concentrations at LOX13 and S-5A were used to represent background and canal concentrations, respectively. Based on these values, the median chloride concentration at LOX15 indicates that the water is typically about 31 percent canal water and 69 percent ?natural? background water. Using median sulfate concentrations, similarly to chloride, the fraction of water at LOX15 was estimated to be 17 percent from canals and 83 percent from ?natural? background water. This finding suggests that in the low sulfate environment of the Refuge, sulfate is not conservative and only about half of the sulfate from canal water typically reaches LOX15; the rest presumably is removed by marsh plants, algae, and bottom sediments. Concentrations of pesticides and other organic compounds were measured