J. R. Mather (1981) observed that runoff (streamflow) constitutes a significant phase of the hydrologic cycle. He also noted that it takes at least 15-25 years of systematic observations to characterize statistically the spatial and temporal patterns in streamflow. With this in mind, a recent assessment of temporal trends in streamflow (Lins and Slack, 1999) is updated to encompass the 60-year period 1940-1999, using data from 435 climate-sensitive stream-gauging stations and expanded to include regional and seasonal characteristics. The previously documented pattern of increasing discharge in the low to moderate range of flows is corroborated, with this pattern being most pronounced in the central two-thirds of the U.S. and to a lesser extent in the eastern coastal regions and in the Great Basin. Relatively few trends are observed in the annual maximum flow. No systematic shift in the timing of the annual minimum, median, or maximum flow is detected in any region on a monthly time scale. The observed increases in low to moderate streamflows, typical of the warm and transitional seasons, are consistent with documented trends in warm and transition season precipitation, and indicate that natural U.S. surface water supply has increased without a concomitant increase in flooding.