This report by the U.S. Geological Survey, prepared in cooperation with the Pennsylvania Department of Environmental Protection, Bureau of Mining and Reclamation, evaluates factors affecting the application of specific-capacity tests in six low-yielding water wells in areas of coal mining or quarrying in Pennsylvania. Factors such as pumping rate, duration of pumping, aquifer properties, wellbore storage, and turbulent flow were assessed by theoretical analysis and by completing multiple well tests, selected to be representative of low-yielding household-supply wells in areas of active coal mining or quarrying. All six wells were completed in fractured-bedrock aquifers--five in coal-bearing shale, siltstone, sandstone, limestone, and coal of Pennsylvanian and Permian age and one in limestone of Cambrian age. The wells were pumped 24 times during 2007-09 at rates from 0.57 to 14 gallons per minute during tests lasting from 22 to 240 minutes. Geophysical logging and video surveys also were completed to determine the depth, casing length, and location of water-yielding zones in each of the test wells, and seasonal water-level changes were measured during 2007-09 by continuous monitoring at each well. The tests indicated that specific-capacity values were reproducible within about ? 20 percent if the tests were completed at the same pumping rate and duration. A change in pumping duration, pumping rate, or saturated aquifer thickness can have a substantial effect on the comparability of repeated tests. The largest effect was caused by a change in aquifer thickness in well YO 1222 causing specific capacity from repeated tests to vary by a factor of about 50. An increase in the duration of pumping from 60 to 180 minutes caused as much as a 62 percent decrease in specific capacity. The effect of differing pumping rates on specific capacity depends on whether or not the larger rate causes the water level in the well to fall below a major water-yielding zone; when this decline happened at well CA 462, specific capacity was reduced by about 63 percent. Estimates of the maximum yield for low-yielding wells that are computed by multiplying the available drawdown by the specific-capacity value may contain large errors if the wells were pumped at low rates that do not cause much water-level drawdown. The estimates of yield are likely to be too large because the effects of lowering the water level in the well below water-yielding zones have not been incorporated. Better yield estimates can be made by the use of step-drawdown tests or by over-pumping at a rate large enough to dewater most of the wellbore. The maximum well yield, after overpumping, can be estimated from the rate of water-level recovery or by subtracting the incremental rate of change of borehole storage at the end of the test from the pumping rate.