In October 2000, a study was undertaken by the U.S. Geological Survey (USGS) in cooperation with the Arkansas Department of Health to determine the hydrogeologic characteristics, including the extent of the recharge areas, for Hughes Spring, Stark Spring, Evening Shade Spring, and Roaring Spring, which are used for public-water supply in northern Arkansas. Information pertaining to each spring can be used to enable development of effective management plans to protect these water resources and public health. An integrated approach to determine the ground-water characteristics and the extent of the local recharge areas of the four springs incorporated tools and methods of hydrology, structural geology, geomorphology, geophysics, and geochemistry. Analyses of discharge, temperature, and water quality were completed to describe ground-water flow characteristics, source-water characteristics, and connectivity of the ground-water system with surface runoff. Water-level contour maps were constructed to determine ground-water flow directions and ground-water tracer tests were conducted to determine the extent of the recharge areas and ground-water flow velocities. Hughes Spring supplies water for the city of Marshall, Arkansas, and the surrounding area. The mean annual discharge for Hughes Spring was 2.9 and 5.2 cubic feet per second for water years 2001 and 2002, respectively. Recharge to the spring occurs mainly from the Boone Formation (Springfield Plateau aquifer). Ground-water tracer tests indicate the recharge area for Hughes Spring generally coincides with the surface drainage area (15.8 square miles) and that Hughes Spring is connected directly to the surface flow in Brush Creek. The geochemistry of Hughes Spring demonstrated variations with flow conditions and the influence of surface-runoff in the recharge area. Calcite saturation indices, total dissolved solids concentrations, and hardness demonstrate noticeable differences with flow conditions reflecting the reduced residence time and interaction of water with the source rock within the ground-water system at higher discharges for Hughes Spring. Concentrations of fecal indicator bacteria also demonstrated a substantial increase during high-flow conditions, suggesting that a non-point source of bacteria possibly from livestock may enter the system. Conversely, nutrient concentrations did not vary with flow and were similar to concentrations reported for undeveloped sites in the Springfield Plateau and Ozark aquifers in northern Arkansas and southern Missouri. Deuterium and oxygen-18 data show that the Hughes Spring discharge is representative of direct precipitation and not influenced by water enriched in oxygen-18 through evaporation. Discharge data show that Hughes Spring is dominated by conduit type ground-water flow, but a considerable component of diffuse flow also exists in the ground-water system. Carbon-13 data indicate a substantial component of the recharge water interacts with the surface material (soil and regolith) in the recharge area before entering the ground-water system for Hughes Spring. Tritium data for Hughes Spring indicate that the discharge water is a mixture of recent recharge and sub-modern water (recharged prior to 1952). Stark Spring supplies water for the city of Cushman, Arkansas, and the surrounding area. 2 Hydrogeologic Characteristics of Four Public Drinking-Water Supply Springs in Northern Arkansas The mean annual discharge for Stark Spring was 0.5 and 1.5 cubic feet per second for water years 2001 and 2002, respectively. The discharge and water-quality data show the ground-water system for Stark Spring is dominated by rapid recharge from surface runoff and mainly consists of a conduit- type flow system with little diffuse-type flow. Analyses of discharge data show that the estimated recharge area (0.79 square mile) is larger than the surface drainage area (0.34 square mile). Ground-water tracer tests and the outcrop of the