Fault dip measurements (n = 3291) from 122 faults in British Coalfields are presented. Two classes of normal fault are recognized: (a) gently dipping (<55°) synsedimentary faults of gravity driven or, less commonly, tectonic origin and (b) steeply dipping (>55°) post-coalification tectonic faults formed at depths of about 2.5 km which comprise 90% of the faults analysed. Dip data from the steeply dipping faults are variable (standard deviation = 9°), both on single faults and between faults, and 70° is a better approximation of fault dip than the commonly accepted value of 60°. Data on tectonic normal faults, with comparable depths of formation (<3–4 km), from other sedimentary sequences show that dips of greater than 60° are quite usual. Systematic changes of dip on single fault surfaces have predictable effects on the curvature of fault traces.
Using the modified Griffith fracture criterion the dips of faults at varying depths (up to 4 km) are calculated for both unsaturated (dry) and saturated rocks. The dips of normal faults are expected to vary from 90° at the surface to 60° at 3–4 km and this variation is consistent with the limited field data available. The coalfield data are consistent with fault formation at standard hydrostatic pore fluid pressures at depths of about 2.5 km. Below depths of a few kilometres normal faults are expected to flatten out gradually to a dip of 45°, but such flattening is not necessarily an indication of an underlying detachment.