Six watersheds in New Hampshire were studied to determine the effects of road salt on stream water quality. Specific conductance in streams was monitored every 15 min for one year using dataloggers. Chloride concentrations were calculated from specific conductance using empirical relationships. Stream chloride concentrations were directly correlated with development in the watersheds and were inversely related to streamflow. Exceedances of the EPA water quality standard for chloride were detected in the four watersheds with the most development. The number of exceedances during a year was linearly related to the annual average concentration of chloride. Exceedances of the water quality standard were not predicted for streams with annual average concentrations less than 102 mg L(-1). Chloride was imported into three of the watersheds at rates ranging from 45 to 98 Mg Cl km(-2) yr(-1). Ninety-one percent of the chloride imported was road salt for deicing roadways and parking lots. A simple, mass balance equation was shown to predict annual average chloride concentrations from streamflow and chloride import rates to the watershed. This equation, combined with the apparent threshold for exceedances of the water quality standard, can be used for screening-level TMDLs for road salt in impaired watersheds.