The British Geological Survey installed a seismometer network to monitor earthquakes around Rosslyn Chapel in the Midlothian Coalfield from November 1987 until January 1990. Accurate locations were obtained for 247 events and a close spatial and temporal association with concurrent coal mining, with a rapid decay of earthquake activity following pit closure, was demonstrated, indicating a mining-induced cause. Residual stress from past mining appears to have been an important factor in generating seismicity, and observations indicate that limiting the width of the workings or rate of extraction may significantly reduce or eliminate mining-induced earthquake activity, an extremely desirable prospect which warrants further investigation. A frequency-magnitude analysis indicates a relatively high abundance of small events in this coalfield area. The maximum magnitude of a mining-induced earthquake likely to have been experienced during the life of the coalfield (maximum credible magnitude) was 3.0 ML, although an extreme event (maximum possible magnitude) as large as 3.4 ML was remotely possible. Significant seismic amplification was observed at Rosslyn Chapel, which is founded on sand and gravel, compared with a nearby bedrock site. As a consequence, relatively small magnitude events caused high, and occasionally damaging, seismic intensities at the chapel. This is likely to be an important effect at similar sites elsewhere.