Petrological and geochemical studies have been carried out on the chalk matrix and the fracture filling calcite from several oil fields in the Great Ekofisk area of the Central Graben, Norwegian North Sea, in order to understand the mechanism of chalk cementation and fracture healing. A hypothesis involving an episode of ascending hot water from the underlying Permian evaporites has been postulated by some workers to explain the healing of similar fractures in the chalk reservoirs of the Norwegian and Danish North Sea and the Danish Sub Basin. However, the results presented in this paper (including cathodoluminescence microscopy; major and trace element geochemistry; C, O, and Sr isotope composition) and modelling of heating and compactional water flow rates suggest that the ascending hot water hypothesis does not satisfactorily explain the required water budget, rate of flow, and the source of the calcite cement. Heat flow simulations demonstrate that a fluid flow rate of 7.5 cm/yr is required by the ascending hot water hypothesis. This is much higher than the flow rates calculated for this and other sedimentary basins. An alternative model is proposed which can account for the observed delta O-18 values of the fracture filling calcite (-12.7 to -8.5 parts per thousand, PDB), the chalk matrix (-7.4 to -4.1 parts per thousand, PDB) and the formation water (-4 parts per thousand, SMOW) in terms of closed to semi-closed system diagenesis. In this model calcite from the chalk matrix dissolves congruently and cement precipitates with an isotope ratio governed by the geothermal gradient. Similarity in the cathodoluminescence, trace element composition, and carbon isotope values between the fracture filling calcite and the chalk matrix also suggests that the cements in the fractures and the chalk matrix are similar and were probably derived locally by pressure solution processes.