Alkaline persulfate digestion was evaluated and validated as a more sensitive, accurate, and less toxic alternative to Kjeldahl digestion for routine determination of nitrogen and phosphorus in surface- and ground-water samples in a large-scale and geographically diverse study conducted by U.S. Geological Survey (USGS) between October 1, 2001, and September 30, 2002. Data for this study were obtained from about 2,100 surface- and ground-water samples that were analyzed for Kjeldahl nitrogen and Kjeldahl phosphorus in the course of routine operations at the USGS National Water Quality Laboratory (NWQL). These samples were analyzed independently for total nitrogen and total phosphorus using an alkaline persulfate digestion method developed by the NWQL Methods Research and Development Program. About half of these samples were collected during nominally high-flow (April-June) conditions and the other half were collected during nominally low-flow (August-September) conditions. The number of filtered and whole-water samples analyzed from each flow regime was about equal.By operational definition, Kjeldahl nitrogen (ammonium + organic nitrogen) and alkaline persulfate digestion total nitrogen (ammonium + nitrite + nitrate + organic nitrogen) are not equivalent. It was necessary, therefore, to reconcile this operational difference by subtracting nitrate + nitrite concentra-tions from alkaline persulfate dissolved and total nitrogen concentrations prior to graphical and statistical comparisons with dissolved and total Kjeldahl nitrogen concentrations. On the basis of two-population paired t-test statistics, the means of all nitrate-corrected alkaline persulfate nitrogen and Kjeldahl nitrogen concentrations (2,066 paired results) were significantly different from zero at the p = 0.05 level. Statistically, the means of Kjeldahl nitrogen concentrations were greater than those of nitrate-corrected alkaline persulfate nitrogen concentrations. Experimental evidence strongly suggests, however, that this apparent low bias resulted from nitrate interference in the Kjeldahl digestion method rather than low nitrogen recovery by the alkaline persulfate digestion method. Typically, differences between means of Kjeldahl nitrogen and nitrate-corrected alkaline persulfate nitrogen in low-nitrate concentration (< 0.1 milligram nitrate nitrogen per liter) subsets of filtered surface- and ground-water samples were statistically equivalent to zero at the p =level.Paired analytical results for dissolved and total phosphorus in Kjeldahl and alkaline persulfate digests were directly comparable because both digestion methods convert all forms of phosphorus in water samples to orthophosphate. On the basis of two-population paired t-test statistics, the means of all Kjeldahl phosphorus and alkaline persulfate phosphorus concentrations (2,093 paired results) were not significantly different from zero at the p = 0.05 level. For some subsets of these data, which were grouped according to water type and flow conditions at the time of sample collection, differences between means of Kjeldahl phosphorus and alkaline persulfate phosphorus concentrations were not equivalent to zero at the p = 0.05 level. Differences between means of these subsets, however, were less than the method detection limit for phosphorus (0.007 milligram phosphorus per liter) by the alkaline persulfate digestion method, and were therefore analytically insignificant.This report provides details of the alkaline persulfate digestion procedure, interference studies, recovery of various nitrogen- and phosphorus-containing compounds, and other analytical figures of merit. The automated air-segmented continuous flow methods developed to determine nitrate and orthophosphate in the alkaline persulfate digests also are described. About 125 microliters of digested sample are required to determine nitrogen and phosphorus in parallel at a rate of about 100 samples per hour with less than 1-percent sample in