Seventeen water-quality constituents were analyzed in samples collected from 21 surface-water sampling sites in the Raritan River Basin during water years 1991-97. Loads were computed for seven constituents. Thirteen constituents have associated instream water-quality standards that are used as reference levels when evaluating the data. Nine of the 13 constituents did not meet water-quality reference levels in all samples at all sites. The constituents that most commonly failed to meet the water-quality reference levels in the 801 samples analyzed were total phosphorus (greater than 0.1 mg/L (milligrams per liter) in 32 percent of samples), fecal coliform bacteria (greater than 400 counts/100 milliliters in 29 percent), hardness (less than 50 mg/L in 21 percent), pH (greater than 8.5 or less than 6.5 in 17 percent), and water temperature in designated trout waters (greater than 20 degrees Celsius in 12 percent of samples). Concentrations of chloride, total dissolved solids, nitrate plus nitrite, and sulfate did not exceed water-quality reference levels in any sample. Results from previous studies on pesticides and volatile organic compounds in streamwater during 1996-98, and organic compounds and trace elements in sediments during 1976-93, were summarized for this study. Concentrations of pesticides in some samples exceeded the relevant standards. Water-quality data varied significantly as season and streamflow changed. Concentrations or values of 12 constituents were significantly higher in the growing season than in the nongrowing season at 1 to 21 sites, and concentrations of 6 constituents were significantly higher in the nongrowing season at 1 to 21 sites. Concentrations or values of seven constituents decreased significantly with increased streamflow, indicating a more significant contribution from base flow or permitted sources than from runoff. Concentrations or values of four constituents increased with increased flow, indicating a more significant contribution from runoff than from base flow or permitted sources. Phosphorus concentrations increased with flow at two sites with no point sources and decreased with flow at five sites with four or more permitted point sources. Concentrations of five constituents did not vary significantly with changes in streamflow at any of the sites. Concentrations of constituents differed significantly between sites. The sites with the most desirable values for the most constituents were Mulhockaway Creek, Spruce Run, Millstone River at Manalapan, Manalapan Brook, and Lamington River at Pottersville. The sites with the least desirable values for the most constituents were Millstone River at Blackwells Mills, Matchaponix Brook, Raritan River at Bound Brook, Neshanic River, and Millstone River at Grovers Mill. The total instream loads of seven constituents - total ammonia plus organic nitrogen (TKN), biochemical oxygen demand (BOD), total dissolved solids (TDS), nitrate plus nitrite (NO3+NO2), total organic carbon (TOC), total phosphorus, and total suspended solids (TSS) - were analyzed at low, median, and high flows. The quantities of total instream load that originated from facilities with permits issued by the New Jersey Department of Environmental Protection to discharge effluent to streams (permitted sources) and from other sources (nonpermitted sources) were estimated for each sampling site. TOC and TSS loads primarily were contributed by nonpermitted sources at all flows. BOD and TDS loads primarily were contributed by nonpermitted sources at median and high flows. At low flow, permitted sources contributed more than one-third of the TDS load at 10 sites and more than one-third of the BOD load at 3 sites. Permitted sources contributed more than one-third of the total phosphorus load at 15 and 14 sites at low and median flows, respectively. Permitted sources accounted for more than one-third of total instream load of NO3+NO2 at low- and median-flow conditions at nearly