Experiments on several quartzites at 300 MPa confirming pressure show them to behave as if near the brittle-ductile transition at 900¿C when oven-dry and in a more ductile fashion at 1000 ¿C when extra water is added. In both cases the stress-strain curves are substantially higher than would be expected on the basis of the strength of hydrolytically weakened synthetic crystals containing less ''water'' than the rocks, as measured by the broad infrared absorption at low temperature. Thus, would appear that, relative to the single crystals, the ''effective'' concentration of the water-related species that is active in mechanical weakening within the grains is only a small fraction of the ''water'' present. The presence of glass in the grain boundaries of ''wet'' deformed specimens suggest that grain boundary weakening may also be significant in these experiments. Questions of equilibrium and of geological extrapolation are discussed.