Hydrous pyrolysis experiments at 200 to 365??C were carried out on a thermally immature organic-rich limestone containing Type-IIS kerogen from the Ghareb Limestone in North Negev, Israel. This work focuses on the thermal behavior of both organic and inorganic sulfur species and the partitioning of their stable sulfur isotopes among organic and inorganic phases generated during hydrous pyrolyses. Most of the sulfur in the rock (85%) is organic sulfur. The most dominant sulfur transformation is cleavage of organic-bound sulfur to form H2 S(gas). Up to 70% of this organic sulfur is released as H2S(gas) that is isotopically lighter than the sulfur in the kerogen. Organic sulfur is enriched by up to 2??? in 34S during thermal maturation compared with the initial ??34S values. The ??34S values of the three main organic fractions (kerogen, bitumen and expelled oil) are within 1??? of one another. No thermochemical sulfate reduction or sulfate formation was observed during the experiments. The early released sulfur reacted with available iron to form secondary pyrite and is the most 34S depleted phase, which is 21??? lighter than the bulk organic sulfur. The large isotopic fractionation for the early formed H2S is a result of the system not being in equilibrium. As partial pressure of H2S(gas) increases, retro reactions with the organic sulfur in the closed system may cause isotope exchange and isotopic homogenization. Part of the ??34S-enriched secondary pyrite decomposes above 300??C resulting in a corresponding decrease in the ??34S of the remaining pyrite. These results are relevant to interpreting thermal maturation processes and their effect on kerogen-oil-H2S-pyrite correlations. In particular, the use of pyrite-kerogen ??34S relations in reconstructing diagenetic conditions of thermally mature rocks is questionable because formation of secondary pyrite during thermal maturation can mask the isotopic signature and quantity of the original diagenetic pyrite. The main transformations of kerogen to bitumen and bitumen to oil can be recorded by using both sulfur content and ??34S of each phase including the H2S(gas). H2S generated in association with oil should be isotopically lighter or similar to oil. It is concluded that small isotopic differentiation obtained between organic and inorganic sulfur species suggests closed-system conditions. Conversely, open-system conditions may cause significant isotopic discrimination between the oil and its source kerogen. The magnitude of this discrimination is suggested to be highly dependent on the availability of iron in a source rock resulting in secondary formation of pyrite. Copyright ?? 2005 Elsevier Ltd.