Abstract In the eastern Gulf of Corinth, geological structure is closely reflected in the topography. In the 1981 Corinth earthquakes ground deformation accentuated existing geomorphological patterns. We explore the relationship between seismicity and landforms in space and time. The area examined extends beyond the epicentral region of the 1981 events and its evolution is traced over the last 40000 years. To understand the relation between deep and superficial structures we discuss processes that occurred in the 1981 earthquakes, and go on to propose that surface faulting appears only above linear faults at depth that move in earthquakes with magnitudes (Ms) greater than six. Where faults at depth intersect or bend, the faulting becomes distributed and prim ary faulting does not reach the surface. Here the surface manifestation of faulting at depth is surface folding and secondary faulting. Such regions are identified in maps of the eastern Gulf of Corinth. Evidence for the evolution of the fault systems is provided by eyewitness accounts of shoreline changes following the recent earthquakes, historical and archaeological data pertaining to old sea levels, and 14C dates for molluscs collected from fossil shorelines. The results of these studies broadly confirm an earlier view of J. A. Jackson et al. (Earth planet. Sci. Lett. 57, 377—397 (1982)) that motion has shifted from faulting south of the city of Corinth to faulting in the Perachora peninsula. But the transfer is not complete: areas lying between the two fault systems have moved up and down in historical and archaeological times indicating that both systems are still active. The evidence for activity of the northern system points to a repeat time of 300 years for earthquakes com parable with those of 1981. The work discussed in this paper suggests that some of the features previously described as m arine terraces are fault-controlled continental surfaces. Furtherm ore, beach deposits at a num ber of different levels long regarded as successive shorelines are in fact contemporaneous. The critical ages were obtained by radiocarbon dating of shell carbonate, a technique which yields dependable results if applied to samples which have been carefully selected and pretreated. Palaeontological dating of the terraces, in contrast, is found to be misleading because the faunas are environm entally controlled and consequently diachronous.