Water quality in the headwaters of the Little Conestoga Creek, Lancaster County, Pa., was investigated from April 1986 through September 1989 to determine possible effects of agricultural nutrient management on water quality. Nutrient management, an agricultural Best-Management Practice, was promoted in the 5.8-square-mile watershed by the U.S. Department of Agriculture Rural Clean Water Program. Nonpoint-source- agricultural contamination was evident in surface water and ground water in the watershed; the greatest contamination was in areas underlain by carbonate rock and with intensive row-crop and animal production. Initial implementation of nutrient management covered about 30 percent of applicable land and was concentrated in the Nutrient-Management Subbasin. By 1989, nutrient management covered about 45 percent of the entire Small Watershed, about 85 percent of the Nutrient- Management Subbasin, and less than 10 percent of the Nonnutrient-Management Subbasin. The number of farms implementing nutrient management increased from 14 in 1986 to 25 by 1989. Nutrient applications to cropland in the Nutrient- Management Subbasin decreased by an average of 35 percent after implementation. Comparison of base- flow surface-water quality from before and after implementation suggests that nutrient management was effective in slowing or reversing increases in concentrations of dissolved nitrate plus nitrite in the Nutrient-Management Subbasin. Although not statistically significant, the Mann-Whitney step-trend coefficient for the Nutrient-Management Subbasin was 0.8 milligram per liter, whereas trend coefficients for the Nonnutrient-Management Subbasin and the Small Watershed were 0.4 and 1.4 milligrams per liter, respectively, for the period of study. Analysis of covariance comparison of concurrent concentrations from the two sub- basins showed a significant decrease in concen- trations from the Nutrient-Management Subbasin compared to the Nonnutrient-Management Subbasin. The small, positive effect of nutrient management on base-flow water quality should be interpreted with caution. Lack of statistical significance for most tests, short-term variation in climate and agricultural activities, unknown ground-water flow rates, and insufficient agricultural-activity data for farms outside of the Nutrient-Management Subbasin were potential problems. A regression model relating nutrient applications to concen- trations of dissolved nitrate plus nitrite showed no significant explanatory relation.