Data collected at 152 sampling sites during 1970-79 were used to characterize fluvial sediment in North Carolina streams. On the basis of predominant land use in individual basins, sites were categorized into one of five groups: forested (7 sites), forested and affected by minor development (7 sites), rural affected by agriculture (83 sites), rural affected by nonagricultural activities (38 sites), and urban (17 sites). Results of >13,000 suspended sediment samples collected during the study were used to determine sediment yield, sediment discharge, concentrations, and other site and basin characteristics. Fluvial sediment characteristics, such as yields, are regionalized, with lower values occurring in the Coastal Plain Province. Statewide, when compared by predominant land use, minimum annual yields occur in forested basins and range from 5-88 tons/sq mi; ratios of average annual yields for forested, rural-agricultural, and urban sites in the Piedmont Province are approximately 1:6:14, respectively. During high flow (0.1% flow duration) in Piedmont basins, the mean suspended sediment concentration for large urban streams is about 1,600 mg/L as compared with 870 mg/L for rural-agricultural sites and 100 mg/L for forested sites. Maximum sediment yields of rural-agricultural basins occur in predominantly clay soil areas of the western Piedmont, with annual values of as much as 470 tons/sq mi, whereas minimum yields as small as 7 tons/sq mi occur in the sandy soil of the Coastal Plain Province. Considerable amounts of fluvial sediment are deposited on flood plains and streambeds as major rivers flow from the rolling Piedmont Province into the flat Coastal Plain Province. For example, >130,000 tons are deposited annually in an 85 mi stretch of the Neuse River between stations at Smithfield and Kinston. Mathematical relations were developed for estimating suspended sediment transport characteristics at unmeasured rural-agricultural sites and urban sites in the Piedmont. Correlation coefficients for the relations range from 0.75 to 0.98, and standard errors of estimate range from 25% to 74%. The best single-variable equation used log-transformed values of drainage area.