Fe and Sr bearing phases were thoroughly investigated by means of spectrometric and microscopic techniques in Callovian-Oxfordian (COX) samples originating from the ANDRA Underground Research Laboratory (URL) in Bure (France). Sr was found to be essentially associated with celestite, whereas Fe was found to be distributed over a wide range of mineral phases. Fe was mainly in its +II oxidation state in the studied samples (~93 % from Mössbauer results). Most of the Fe(II) was found to be present in pyrite, sideroplesite/ankerite and clay minerals. Iron +III, if present, was associated with clay minerals (probably illite, illite-smectite mixed layer minerals and chlorite). No Fe(III) oxi(hydro)xide could be detected in the samples. Strontianite was not observed either. Based on these observations, it is likely that the COX porewater is in equilibrium with the following carbonate minerals, calcite, dolomite and ankerite/sideroplesite, but not with strontianite. It is then shown that this equilibrium information can be combined with clay cation exchange composition information in order to give direct estimates or constraints on the solubility products of these carbonate minerals: dolomite, siderite and strontianite. As a consequence, an experimental method was developed to retrieve the cation exchanged Fe(II) in very well preserved COX samples. The very homogeneous cation exchange composition of the formation is completely in agreement with a homogeneous presence of calcite and dolomite minerals whose equilibrium reactions control part of the porewater composition. Amongst the broad range of values available for dolomite solubility products in thermodynamic databases, the value of log Kdolomite = -3.57 is the most appropriate for the one present in the COX formation. As regards strontianite, it appears that the equilibrium constant tabulated in the Llnl database is not valid for natural clay systems. The value given in Busenberg et al. Geochimica et Cosmochimica Acta, 48, 2021-2035, used by most of the other available thermodynamic databases seems to be far more appropriate. Concerning Fe(II) and siderite/ankerite equilibrium, the measured Fe/Ca ratio on the clay exchanger (~0.01) could only be considered as a maximum value due to possible experimental bias, leading to the following constraint for the solubility of the sideroplesite phase present in the COX: Fe0.7Mg0.2Ca0.1CO3 + H+  0.7 Fe2+ + 0.2 Mg2+ + 0.1 Ca2+ + HCO3- log Ksid_COX < 0.42. This constraint is in agreement with the estimated solubility of this phase based on a solid solution model: log Ksid_COX = 0.23.