A multidisciplinary study of precipitation and streamflow data and paleoflood studies of channel features was made to analyze the flood hydrology of foothill and mountain streams in the Front Range of Colorado (with emphasis on the Big Thompson River basin) because conventional flood-frequency analyses do not adequately characterize the flood hydrology. In the foothills of Colorado, annual floodflows are derived from snowmelt at higher elevations in the mountain regions, from rainfall at lower elevations in the plains or plateau regions, or from a combination of rain falling on snow. Above approximately 7,500 feet snowmelt dominates; rain does not contribute to the flood potential. Regression analyses were done to determine flood characteristics at ungaged sites. These study results helped identify a relatively homogeneous hydrologic foothill region in the South Platte River basin. When the .m4r^ina£e_,are^a below_ 8,000_feet w?s Ubed in the regional flood-prediction equations rather than"the total drainage area, the standard error of estimate improved from 142 to 44 percent for the regional flood-prediction equations. These regression relationsNand study results indicate that methods of computing flood characteristics ,~^4>ased on rainfall-runoff modeling, overestimate flood magnitude in the foothills and mountains of Colorado. Regional flood-frequency relations were ecwnpared with rainfallrunoff flood-estimating technique results, which included an evaluation of the magnitude and frequency of the probable maximum flood. The study demonstrated that the concept of storm transposition from lower elevations to higher elevations, that is the basis of the rainfall-runoff method, is not supported by meteorological, hydrological, and paleoflood data. Regional-regression relations were used to compute the recurrence interval of selected large floods in the study area. Regional flood-frequency equations, combined with paleoflood investigations, provide more reliable estimates of both common and rare floods. This technique improved flood estimates beyond the 100-year recurrence interval. These regional analyses, supported by radiocarbon dating, indicate that the 1976 Big Thompson flood, in the area of most intense rainfall, had a recurrence interval of about 10,000 years. Evaluation of streamflow data and paleoflood investigations provide an alternative for evaluating flood hydrology and the safety of dams. The study indicates the need for additional data collection and research to understand the complexities of the flood hydrology in mountainous regions, especially its effects on flood-plain management and design of structures in the flood plain.