In fire-prone southwestern ecosystems, wildfire is a significant disturbance event that promotes accelerated erosion in upland watersheds. This accelerated erosion can cause environmental site degradation and may threaten life, property, and infrastructure in human communities downstream at the wildland/urban interface. The San Dimas Experimental Forest (SDEF) was established in the early 1930s to document and quantify wildland hydrology in the semiarid chaparral-covered steeplands of southern California. Concomitantly, the nearly seventy years of accumulated watershed research, that has included both wildfires and prescribed burning, has produced invaluable information on post-fire erosion. Management treatments following a wildfire in 1960 involved the vegetation type conversion of some native chaparral watersheds to a mixture of perennial grasses. In September 2002 virtually the entire SDEF burned in the Williams Fire, including several small (2-3 ha) headwater catchments in both brush and grass fuel types. Four of these watersheds were instrumented nearly a decade ago to measure hillslope erosion using 30 cm aperture sediment collector traps. In each watershed, 75 collectors were installed on unbounded plots located both along hillslope fall-line transects and at the hillslope/channel interface. One of these catchments with chaparral vegetation was burned in a prescribed fire in May 2001, and did not re-burn during the Williams Fire. Both dry season hillslope erosion, governed by the gravitational process of dry ravel, and wet season erosion, dominated by the hydrologic processes of rainsplash and overland flow, were compared for several years prior to the fire and through the first post-fire winter. Because of differences in the nature of ground surface vegetation, pre-fire hillslope erosion was an order of magnitude less in grass compared to chaparral. In chaparral catchments, post-fire dry season erosion was 2-3 times greater and post-fire wet season erosion was 10-17 times greater than unburned levels. In a grass watershed, post-fire dry season erosion was 9 times greater and post-fire wet season erosion was 375 times greater than comparable unburned values, the latter figure actually exceeding the activity rates from chaparral catchments. The prescribed burn produced pre-fire/post-fire ratios for dry season erosion that were roughly equal to the wildfire, but generated only one third to one half the comparable wet season erosion during the first winter after burning. While this relationship may reflect lower fire severities, more probably this response stems from the record low rainfall following the prescribed burn. Results of this study indicate that wildfire can dramatically alter the erosion response of upland landscapes, but that prescribed burning can have similar effects in chaparral ecosystems.