Abstract Interpretation of regional two-dimensional seismic surveys and three-dimensional seismic surveys in the central North Sea has demonstrated the existence of a pervasive polygonal network of normal faults affecting Tertiary shale-dominated slope and basin-floor depositional systems. The area affected by the faulting encompasses most areas of hydrocarbon production from Tertiary sandstone reservoirs. The polygonal fault networks were active during sedimentation and early burial. Throws measured on individual faults range from 8 to 100m, with average fault plane dips of 45°. Lengths of individual fault segments range from 80 to 1400m, and average fault spacings range from 100 to 500m. The high density and the polygonal geometry of the faults make seismic interpretation of the Lower Tertiary interval problematic, and can lead to misinterpretation of faults as apparent seismic-stratigraphic features. The recognition of this fault system has consequences for the development of of fields in Tertiary reservoirs. The remaining hydrocarbon potential in the Lower Tertiary could be considered predominantly to be stratigraphic plays, but the presence of faults active during sedimentation and early burial implies that this play concept is too simplistic, particularly for Eocene reservoirs. The role of both syn-depositional and early-burial normal faulting on original reservoir distribution and post-depositional modification should be considered further. The presence of this fault system may also be important for secondary migration into Lower Tertiary reservoirs from Kimmeridge Clay Formation source rocks. The existence of an interconnected fault and fracture network in the low-permeability mudrocks may have provided an efficient vertical migration pathway for charging isolated lower Tertiary sandstone reservoirs. Finally, the maximum fault throws of between 50 and 100m are large enough to represent potential barriers for lateral communication in sandstone reservoirs where individual sand bodies are commonly 25–100m thick.