Stable cracks (crack velocity of 1 m s-1) were propagated in 20 rectangular parallepipeds of identical orientation cut from a synthetic quartz crystal, having and initial dislocation density of 103 mm-2 or less and a water content of 5¿10-2 wt%. Five tests each were conducted dry (ambient atmosphere PH2O) in distilled water (&zgr;=-40 mV, where &zgr; is a type of electrostatic potential), and in two aqueous solutions of dodecyl trimethyl ammonium bromide (DTAB) (&zgr;=0 and &zgr;= +70 mV). Mean values of the crack propagation stress &sgr;c were 69.7 MPa dry, 60.5 MPa in water, 48.6 MPa in DTAB at &zgr;=0, and 45.3 MPa in DTAB at &zgr;=+70 mV. All cracks were mapped by phase contrast and transmission electron microscopy, from open crack throats to tips. Each crack consists of straight segments parallel to the basal plane and zigzag segments composed of r(1011) and z(0111) cleavage steps. There is no evidence of dislocation nucleation or mobility influencing crack growth. These observations, the strength minimum at (&zgr;≠O), and recent data from the surface chemistry literature strongly suggest that the chemomechanical models of Westwood (1974) and Rehbinder et al. (1944) cannot be applied to quartz and silicate geologic materials. The data appear to support a model presented here which relates crack propagation stress to the availability of adsorbable species.