The relationship between rate of sediment discharge and rate of water discharge at a cross section of a stream is frequently expressed by an average curve. This curve is the sediment rating curve. It has been widely used in the computation of average sediment discharge from water discharge for periods when sediment samples were not collected. This report discusses primarily the applications of sediment rating curves for periods during which at least occasional sediment samples were collected. Because sediment rating curves are of many kinds, the selection of the correct kind for each use is important. Each curve should be carefully prepared. In particular, the correct dependent variable must be used or the slope of the sediment rating curve may be incorrect for computing sediment discharges. Sediment rating curves and their applications were studied for the following gaging stations: 1. Niobrara River near Cody, Nebr. 2. Colorado River near Grand Canyon, Ariz. 3. Rio Grande at San Martial, N. Mex. 4. Rio Puerto near Bernardo, N. Mex. 5. White River near Kadoka, S. Dak. 6. Sandusky River near Fremont, Ohio Except for the Sandusky River and the Rio Puerco, which transport mostly fine sediment, one instantaneous sediment rating curve was prepared for the discharge of suspended sands, at each station, and another for the discharge of sediment finer than 0.082 millimeter. Each curve was studied separately, and by trial-end-error multiple correlation some of the factors that cause scatter from the sediment rating curves were determined. Average velocity at the cross section, Water temperature, and erratic fluctuations in concentration seemed to be the three major factors that caused departures from the sediment rating curves for suspended sands. The concentration of suspended sands varied with about the 2.8 power of the mean velocity for the four sediment, rating curves for suspended sands. The effect of water temperature was not so consistent as that of velocity and theoretically should vary considerably with differences in the size composition of the suspended sands. Scatter from the sediment rating curves for sediments finer than 0.082 millimeter seemed to be caused by changes in supply of these sediments. Some of the scatter could be explained by seasonal variations, by a pattern of change in concentration of fine sediment following a rise, or by source of the runoff as indicated by the measured relative flows of certain tributaries. Daily or instantaneous sediment rating curves adjusted for factors that account for some of the scatter from an average curve often can be used to compute approximate daily, monthly, and annual sediment discharges. Accuracy of the computed sediment discharges should be better than average for streams that transport mostly sands rather than fine sediments and for some ephemeral or intermittent streams, such as Rio Puerco, in semiarid regions. Accuracy of computed sediment discharges can be much improved for many streams by shifting the sediment rating curve on the basis of 2 or 4 measurements of sediment discharge per month. Of 26 annual sediment discharges that were computed by shifting sediment rating curves to either 2 or 4 measured sediment discharges per month, 18 were within I0 percent of the annual-sediment discharges that were computed on the basis of a daily sampling program. Monthly and daily sediment discharges computed from daily or instantaneous sediment rating curves, either shifted or unshifted, were less accurate than similarly computed annual sediment discharges. Even so, the difference in cost between occasional sediment samples and daily samples is so great that the added accuracy from daily sampling may not Justify the added cost. Monthly and annual sediment-rating curves can be applied simply, with adjustments if required, to compute monthly and annual sediment discharges with reasonably good accuracy for gaging stations like the Rio Puerco near Bernardo,