In August 1996 the U.S. Geological Survey initiated the Sediment Laboratory Quality-Assurance project. The Sediment Laboratory Quality Assurance project is part of the National Sediment Laboratory Quality-Assurance program. This paper addresses the fmdings of the sand/fme separation analysis completed for the single-blind reference sediment-sample project and differences in reported results between two different analytical procedures.

From the results it is evident that an incomplete separation of fme- and sand-size material commonly occurs resulting in the classification of some of the fme-size material as sand-size material. Electron microscopy analysis supported the hypothesis that the negative bias for fme-size material and the positive bias for sand-size material is largely due to aggregation of some of the fine-size material into sand-size particles and adherence of fine-size material to the sand-size grains. Electron microscopy analysis showed that preserved river water, which was low in dissolved solids, specific conductance, and neutral pH, showed less aggregation and adhesion than preserved river water that was higher in dissolved solids and specific conductance with a basic pH. Bacteria were also found growing in the matrix, which may enhance fme-size material aggregation through their adhesive properties.

Differences between sediment-analysis methods were also investigated as pan of this study. Suspended-sediment concentration results obtained from one participating laboratory that used a total-suspended solids (TSS) method had greater variability and larger negative biases than results obtained when this laboratory used a suspended-sediment concentration method. When TSS methods were used to analyze the reference samples, the median suspended sediment concentration percent difference was -18.04 percent. When the laboratory used a suspended-sediment concentration method, the median suspended-sediment concentration percent difference was -2.74 percent. The percent difference was calculated as follows: Percent difference = (( reported mass - known mass)/known mass ) X 100.