The problem of correlating unit hydrograph features to topographic and man-made basin characteristics received attention in this report. The unit graph features considered herein were the peak discharge and the time lag of basin response. In order to facilitate the desired regression analysis, the two-parameter gamma model proposed by Edson was utilized in the investigation. The parameters of the model were obtained by the simultaneous solution of the equations for unit graph peak and lag using observed unit hydrographs for 16 basins in the Piedmont region of North Carolina and 14 basins located in Northern Virginia. In the opinion of many, these parameters are a better measure of the complex relationship which exists between the runoff from a basin and the topographic features of that basin than are the values of the unit graph peak and lag time themselves. The basin characteristics utilized in the investigation were: basin area, length of the longest streamcourse in the basin, average stream slope between points 10 percent and 85 percent downstream of the headwaters, and the percent of the impervious area contained in the basin. This last factor served as a measure of the amount of urban development present in the watershed. The investigation was hampered by a regrettable lack of sufficient data to derive regression equations of good reliability. This fact was due to the reduction of the data into groups by narrow geographical ranges. Thus, the number of stations available for analysis in anyone group was insufficient for purposes of a reliable regression analysis. From the investigation, it appears that the most significant basin characteristics affecting runoff are length, slope, and urban development. The strongest regression equations were derived using those three characteristics. It appears that the length and slope factors give better results when combined in the form (L/√S). / Master of Science