Surveys in eight paired urban and rural watersheds illustrate how urbanization changes fluvial morphology and processes. Our data also provide quantitative criteria for evaluating stream-restoration projects in urban areas. Bankfull depth, reach-averaged bed slope, and median grain size are similar in urban and rural watersheds. The median width of urban channels is 26% larger than the median width of rural channels. The median sinuosity is 8% lower in urban channels and pools are 31% shallower. The median composite Manning's n based on median grain diameter, pool depth, and channel sinuosity is 10% lower in urban streams, while the median bankfull discharge per unit drainage basin area is 131% higher in urban channels. Histograms of bed sediment-size distributions in urban channels lack a secondary mode in the size range 2–64 mm characteristic of rural channels, indicating that these sizes tend to be selectively removed from urban channels. However, bankfull Shields stresses in urban and rural channels exceed typical threshold values at most sites, indicating significant bedload transport at bankfull stage. Apparently, increased peak discharges caused by decades of urbanization have not removed all the transportable sediment from these urban stream channels. We speculate that the supply of sediment to urban channels from hillslope processes and channel erosion remains significant, even though much of the upland surfaces of these urban catchments are covered with nonerodible impervious surfaces.