Stable isotope studies of Cl, O, and H carried out in addition to routine shipboard chemical analyses performed during Ocean Drilling Program (ODP) Leg 164 aid in constraining estimates of hydrate concentration and mechanisms of hydrate formation and dissociation in the submarine hydrate zone of the Blake Ridge in the West Atlantic. Chlorine isotope ratios show a steady downward decrease from a shallow maximum of about 0 per mil near the roof of the hydrate zone (postulated to occur at 24 mbsf) to a d37Cl value of -3.68 per mil near the bottom of Site 997, one of the more negative Cl-isotope values measured in marine pore waters. The downward depletion in the heavy isotope is coupled with a chlorinity decrease, determined by the shipboard measurements. Modeling of the chlorine isotopic data provides an independent test of the hydrochemical model developed by Egeberg and Dickens to explain the chlorinity profile of Site 997. According to that model, the downward chlorinity decrease is largely due to advection of low-chlorinity water from below the drilled section, the difference to the shallow maximum being bridged by diffusion. Using the Egeberg and Dickens advection-diffusion model for the Cl-isotope data gives the best fit between measured trend and simulation for an upward advection rate of 0.18 mm/yr and an assumed ratio of 1.0023 for the diffusion coefficients of the light and heavy Cl isotopes. The low-chlorinity water advected from below the drilled section carries the 37Cl-depleted isotopic signature, but the source of this water is unknown, as are the mechanisms for the heavy isotope depletion. Fractionation of the chlorine isotopes due to different mobilities during downward diffusion from the chloride maximum can only explain a minor portion of the 37Cl depletion. Positive d18O and dD spikes in the upper part of the section drilled at Site 997 are within the range of isotopic excursions during the Quaternary glaciations and can be explained by inheritance from buried connate waters, like the chloride maximum in the upper 24 mbsf, whereas the effects of hydrate formation (salt exclusion and preferential uptake of the heavy isotopes) are minor due to the low hydrate concentrations at the site. Negative d18O values <= 0.6 per mil in the middle of the section between 159 and 532 mbsf largely reflect authigenic carbonate formation, whereas the solid reaction partners that cause negative dD spikes below 200 mbsf remain unknown. Heavy oxygen and hydrogen isotope spikes between 200 mbsf and the base of the hydrate zone at 452 mbsf correspond to low-chlorinity spikes, indicating hydrate dissociation in local, hydrate-rich sediment layers. Overall, oxygen isotope ratios increase below 300 mbsf, whereas hydrogen isotope ratios show a general downward decrease for the site.