This study was made to define the relation between floods in a humid region and the causative factors that account for their variability. The study covered the basic relationships between peak discharges and hydrologic factors as well as the practical working methods for generalizing the results on a regional basis. Statistical multiple-correlation techniques were applied to hydrologic data in New England. A field and library investigation of historical flood data extended the period of flood knowledge to as much as 300 years. The relation of many topographic and climatic factors to flood peaks was tested. Flood peaks with recurrence intervals at 9 levels, from 1.2 to 300 years, were related to 6 independent variables, 3 of which are topographic, 2 climatic, and 1 orographic.