This report describes the experimental details and interprets results from a study conducted by the U.S. Geological Survey (USGS) in 1992 to assess the effect of different sample-processing treatments on the stability of eight nutrient species in samples of surface-, ground-, and municipal-supply water during storage at 4 degrees Celsius for about 30 days. Over a 7-week period, splits of filtered- and whole-water samples from 15 stations in the continental United States were preserved at collection sites with sulfuric acid (U.S. Environmental Protection Agency protocol), mercury (II) chloride (former U.S. Geological Survey protocol), and ASTM (American Society for Testing and Materials) Type I deionized water (control) and then shipped by overnight express to the USGS National Water Quality Laboratory (NWQL). At the NWQL, the eight nutrient species were determined in splits from each of the 15 stations, typically, within 24 hours of collection and at intervals of 3, 7, 14, 22, and 35 days thereafter. Ammonium, nitrate plus nitrite, nitrite, and orthophosphate were determined only in filtered-water splits. Kjeldahl nitrogen and phosphorus were determined in both filtered-water and whole-water splits. Data on which this report is based, including nutrient concentrations in synthetic reference samples determined concurrently with those in real samples, are extensive (greater than 20,000 determinations) and have been published separately. In addition to confirming the well-documented instability of nitrite in acidified samples, this study also demonstrates that when biota are removed from samples at collection sites by 0.45-micrometer membrane filtration, subsequent preservation with sulfuric acid or mercury (II) provides no statistically significant improvement in nutrient concentration stability during storage at 4 degrees Celsius for 30 days. Biocide preservation had no statistically significant effect on the 30-day stability of phosphorus concentrations in whole-water splits from any of the 15 stations, but did stabilize Kjeldahl nitrogen concentrations in whole-water splits from three data-collection stations where ammonium accounted for at least half of the measured Kjeldahl nitrogen.