Attempts to introduce molecular water into dry, natural quartz crystals by diffusive transport and thus weaken them hydrolytically at T=700¿-900¿C and PH2O=400-1550 MPa have failed. Infrared spectroscopy of hydrothermally annealed single crystals of natural quartz reveals the diffusive uptake of interstitial hydrogen (resulting in hydroxyl groups) at rates similar to those previously proposed for intracrystalline water at high water pressures. The solubility of interstitial hydrogen at these conditions is independent of temperature and pressure; instead, it depends upon the initial aluminum concentration by the local charge neutrality condition i>=Si'>. The rate of interstitial hydrogen diffusion parallel to c is given by an Arrhenius relation with D0=1.4¿10-1 m2/s and Q=200¿20 kJ/mol, in close agreement with H diffusivities reported for much lower pressures (PH2O=2.5 MPa). Deformation experiments following hydrothermal annealing show no mechanical weakening, and the lack of any detectable broadband absorption associated with molecular water shows that the diffusion rates of structural water are much lower than those of hydrogen. These results are consistent with the available oxygen diffusion data for quartz and with the failure to observe weakening in previous studies of quartz deformation at pressures of 300-500 MPa; they call into question the rapid rates of diffusion originally suggested for the hydrolytic weakening defect. It is suggested that the observed weakening in many previous experiments was complicated by microcracking processes in response to nonhydrostatic stresses and low effective confining pressures. Extensive microcracking may provide a mechanism for molecular water to enter quartz and allow local plastic deformation to occur. It does not appear that molecular water can diffuse far enough into uncracked quartz to allow hydrolytic weakening over annealing times that are feasible in the laboratory.