Spectacular production rates have been achieved by certain U.S. longwalls, and yet a large number of longwall operations have failed to meet anticipated production targets. This study attempts to identify the primary factors which contribute to the production shortcomings of many marginal longwall operations. This study presents details of the classification and analysis of delay data for a group of thirty-nine longwall sections located in the eastern and mid-eastern United States. Downtime data correspondIng to over fourteen-thousand shifts were collected and classified according to equipment type, delay type, and specific delay event. A dBase IV-based database was constructed to allow flexible interrogation of the data. The relative downtime contributions of the various equipment components and of the delay types have been determined. Machine availabilities and system availabilities are presented. Probability density functions have been flit to the time-to-failure and to the time-to-repair data sets, both for the principal equipment types and for the longwall system as a whole. Recommendations are made for increasing the availability of longwall systems.