During storms when rainfall intensity exceeded about 1 inch per hour, peak unit runoff from the Unnamed Tributary (surface-mined and filled) Watershed exceeded peak unit runoff from the Spring Branch (unmined) Watershed in the Ballard Fork Watershed in southern West Virginia. During most storms, those with intensity less than about 1 inch per hour, peak unit (area-normalized) flows were greater from the Spring Branch Watershed than the Unnamed Tributary Watershed. One storm that produced less than an inch of rain before flow from the previous storm had receded caused peak unit flow from the Unnamed Tributary Watershed to exceed peak unit flow from the Spring Branch Watershed. Peak unit flow was usually similar in Spring Branch and Ballard Fork. Peak unit flows are expected to decrease with increasing watershed size in homogeneous watersheds; drainage area and proportion of the three watersheds covered by valley fills are 0.19 square mile (mi?) and 44 percent for the Unnamed Tributary Watershed, 0.53 mi? and 0 percent for the Spring Branch Watershed, and 2.12 mi? and 12 percent for the Ballard Fork Watershed. Following all storms with sufficient rainfall intensity, about 0.25 inches per hour, the storm hydrograph from the Unnamed Tributary Watershed showed a double peak, as a sharp initial rise was followed by a decrease in flow and then a delayed secondary peak of water that had apparently flowed through the valley fill. Hortonian (excess overland) flow may be important in the Unnamed Tributary Watershed during intense storms, and may cause the initial peak on the rising arm of storm hydrographs; the water composing the initial peaks may be conveyed by drainage structures on the mine. Ballard Fork and Spring Branch had hydrographs with single peaks, typical of elsewhere in West Virginia. During all storms with 1-hour rainfall greater than 0.75 inches or 24-hour rainfall greater than 1.75 inches during which all stream gages recorded a complete record, the Unnamed Tributary yielded the most total unit flow. In three selected major storms, total unit flow from the Unnamed Tributary during recessions exceeded storm flow, and its total unit flow was greatest among the streams during all three recessions. Runoff patterns from the mined watershed are influenced by the compaction of soils on the mine, the apparent low maximum rate of infiltration into the valley fill compared to that in the unmined, forested watershed, storage of water in the valley fill, and the absence of interception from trees and leaf litter. No storms during this study produced 1-hour or 24-hour rainfall in excess of the 5-year return period, and streamflow during this study never exceeded a magnitude equivalent to the 1.5-year return period; relative peak unit flow among the three streams in this study could be different in larger storms. Rainfall-runoff relations on altered landscapes are site-specific, and aspects of mining and reclamation practice that affect storm response may vary among mines.