The effects of nonpoint-source contamination on the water quality of four tributaries to the Toms River in Ocean County, New Jersey, have been investigated in a 5-year study by the U.S. Geological Survey (USGS), in cooperation with the New Jersey Department of Environmental Protection (NJDEP). The purpose of the study was to relate the extent of land development to loads of nutrients and other contaminants to these streams, and ultimately to Barnegat Bay. Volumetric streamflow (discharge) was measured at 6 monitoring sites during 37 stormflow and base-flow sampling events over a 5-year period (May 1994-September 1999). Concentrations and yields (area-normalized instantaneous load values) of nitrogen and phosphorus species, total suspended solids, and fecal coliform bacteria were quantified, and pH, dissolved oxygen, and stream stage were monitored during base-flow conditions and storms. Sufficient data were collected to allow for a statistical evaluation of differences in water quality among streams in subbasins with high, medium, and low levels of land development. Long Swamp Creek, in a highly developed subbasin (64.2 percent developed); Wrangle Brook, in a moderately developed subbasin (34.5 percent); Davenport Branch, in a slightly developed subbasin (22.8 percent); and Jakes Branch, in an undeveloped subbasin (0 percent) are the subbasins selected for this study. No point-source discharges are known to be present on these streams. Water samples were collected and analyzed by the NJDEP, and discharge measurements and data analysis were conducted by the USGS. Total nitrogen concentrations were lower in Davenport Branch than in Long Swamp Creek and Wrangle Brook during base flow and stormflow. Concentrations of total nitrogen and nitrate were highest in Wrangle Brook (as high as 3.0 mg/L and 1.6 mg/L, respectively) as a result of high concentrations of nitrate in samples collected during base flow; nitrate loading from ground-water discharge is much higher in Wrangle Brook than in any of the other streams, possibly as a result of an experimental wastewater-(secondary effluent) disposal site that was in operation during the 1980's. Ammonia concentrations were higher in samples from Long Swamp Creek than in those from the other two monitoring sites under all flow conditions, and ammonia yields were higher during stormflow than base flow at all monitoring sites. Concentrations and yields of fecal coliform bacteria and total suspended solids were higher during stormflow than during base flow at all monitoring sites. Concentrations and yields were significantly higher in Long Swamp Creek, a highly developed subbasin and Wrangle Brook, a moderately developed subbasin than in Davenport Branch, a slightly developed subbasin. Concentrations and yields of phosphate species, which also are strongly related to stormflow, were higher during stormflow in Long Swamp Creek than in the other subbasins. Base-flow separation techniques were used on hydrographs generated for storms to distinguish the fraction of discharge and constituent loading attributable to storm runoff (overland flow) from the fraction contributed by ground-water discharge. Precipitation records were used to determine the total annual volumes of ground-water discharge and runoff at each monitoring site. These volumes were used in conjunction with water-quality data to calculate total annual loads of each constituent at each monitoring site, separated into ground-water discharge and runoff fractions. It was determined that loads of ammonia, nitrate, organic nitrogen, total nitrogen, and orthophosphate in ground-water discharge were significantly higher in the moderately developed Wrangle Brook subbasin than in the highly developed Long Swamp Creek subbasin, and that no relation was apparent between the percent of land development and constituent loads from ground-water discharge. The loading of each constituent contributed by ground-water discharge is specific