Although the presence of extensive gas hydrate on the Cascadia margin, offshore from the western U.S. and Canada, has been inferred from marine seismic records and pore water chemistry, solid gas hydrate has only been found at one location. At Ocean Drilling Program (ODP) Site 892, offshore from central Oregon, gas hydrate was recovered close to the sediment - water interface at 2-19 m below the seafloor, (mbsf) at 670 m water depth. The gas hydrate occurs as elongated platy crystals or crystal aggregates, mostly disseminated irregularly, with higher concentrations occurring in discrete zones, thin layers, and/or veinlets parallel or oblique to the bedding. A 2-to 3-cm thick massive gas hydrate layer, parallel to bedding, was recovered at ???17 mbsf. Gas from a sample of this layer was composed of both CH4 and H2S. This sample is the first mixed-gas hydrate of CH4-H2S documented in ODP; it also contains ethane and minor amounts of CO2. Measured temperature of the recovered core ranged from 2 to - 18??C and are 6 to 8 degrees lower than in-situ temperatures. These temperature anomalies were caused by the partial dissociation of the CH4-H2S hydrate during recovery without a pressure core sampler. During this dissociation, toxic levels of H2S (??34S, +27.4???) were released. The ??13C values of the CH4 in the gas hydrate, -64.5 to -67.5???(PDB), together with ??D values of - 197 to - 199???(SMOW) indicate a primarily microbial source for the CH4. The ??18O value of the hydrate H2O is +2.9???(SMOW), comparable with the experimental fractionation factor for sea-ice. The unusual composition (CH4-H2S) and depth distribution (2-19 mbsf) of this gas hydrate indicate mixing between a methane-rich fluid with a pore fluid enriched in sulfide; at this site the former is advecting along an inclined fault into the active sulfate reduction zone. The facts that the CH4-H2S hydrate is primarily confined to the present day active sulfate reduction zone (2-19 mbsf), and that from here down to the BSR depth (19-68 mbsf) the gas hydrate inferred to exist is a ???99% CH4 hydrate, suggest that the mixing of CH4 and H2S is a geologically young process. Because the existence of a mixed CH4-H2S hydrate is indicative of moderate to intense advection of a methane-rich fluid into a near surface active sulfate reduction zone, technically active (faulted) margins with organic-rich sediments and moderate to high sedimentation rates are the most likely regions of occurrence. The extension of such a mixed hydrate below the sulfate reduction zone should reflect the time-span of methane advection into the sulfate reduction zone. ?? 1998 Elsevier Science B.V. All rights reserved.