Regression models were developed for the 25th percentile of June and September flows (first quartile of flow) for 47 streamflow-gaging stations (gaging stations) in the Upper Mississippi, Ohio, and Great Lakes drainage basins. The gaging stations that were selected for this analysis are on unregulated rivers, have at least 40 years of record, and have a nearby weather station with at least 70 years of precipitation record. Regression models were developed for each gaging station relating annual 25th percentile of June and September flows to selected precipitation variables. The explanatory variables are monthly precipitation (April-June, July-September) for each year of record, precipitation for the previous year, and average precipitation for the preceding 5-, 10-, 15-, 20-, 25-, and 30-year periods. Short-term precipitation (April-June or July-September monthly precipitation) variables are the most common significant variables in the regression equations for the 25th percentile of June and September streamflows. May and June monthly precipitation are the most common significant variables among the regression models of the 25th percentile of June flows. August and September monthly precipitation are the most common significant variables in the regression models of the 25th percentile of September streamflow. July precipitation also is a significant explanatory variable in regression models of September streamflow. The 25th-percentile flows in this study also are related to intermediate- and long-term precipitation variables. The intermediate-term precipitation variable (previous-year's precipitation) has a more distinct spatial pattern than the long-term precipitation variable (multiyear running averages of annual precipitation) and is more likely to be significant in the western part than in the eastern part of the study area.