The magnitude of groundwater recharge to coal aquifers in a 150-km2 area in west-central North Dakota was determined using three separate approaches: (1) the net water level rise in water-table wells; (2) calculations of the fluid flux between nested piezometers, using the Darcy equation and measured values of hydraulic conductivity and vertical gradients; and (3) evaluation of the inputs to and outputs from the coal aquifer, using a steady-state control volume approach in which the aquifer was divided into semi-rectangular cells bounded by equipotential lines and flow lines. Measurements of potential gradients and hydraulic conductivity permitted indirect determination of all components of flow into and out of the cell except the recharge input, which was determined by difference. All methods yielded consistent results on the order of 0.04-0.01 m yr.-1 These values, which represent 2-9% of the annual precipitation, are consistent with results of other studies on recharge throughout the prairies of North America. Evaluation of site hydrology and stable-isotope data indicates that recharge is restricted in both time and place. Most recharge occurs in late spring and in the fall following heavy rainfall events. During these seasons the ground is not frozen and vegetation is not transpiring large amounts of water. Some recharge may occur during very heavy localized summer storms, but it is not considered volumetrically significant. Major permanent depressions on the site are a source of significant recharge. In addition, the extensive area of ephemeral standing water bodies that result from snowmelt can produce significant amounts of infiltration over the entire site. ?? 1982.