Acoustic emission monitored during the hydraulic fracturing of two intact samples of Weber sandstone was used to determine the location and orientation of macroscopic fracture planes caused by the fracturing process. The locations of microfractures that were determined from the acoustic emission data corresponded closely with observed fractures. Pore fluid was injected into the two samples at different rates (3.3¿105 cc/s and 3.3¿104 cc/s) until failure occurred. The sample injected at the slower rate failed in shear, whereas the sample injected at the faster rate failed in tension. Both samples were subjected to 1000-bars confining pressure and 4000-bars differential stress. To further study how tension and shear failure depend on pore fluid injection rate and differential stress, a series of 16 hydraulic fracture experiments were run on smaller 2.54-cm-diameter samples without monitoring acoustic emission. It was found that over a wide range of stresses, either shear or tension failure could be produced simply by varying the rate at which fluid was injected into the sample. A theoretical model was developed to calculate the pore pressure distribution in the samples as a function of time, borechole pressure, and differential stress. The model was used to relate the failure mechanism (tension or shear failure) to the pore pressure distribution in each sample by analysis that required knowledge of the porosity and permeability of Weber sandstone. To supply the input data necessary for the numerical model, a series of permeability measurements were conducted. These measurements show that permeability is dependent upon effective pressure and differential stress. The principal goals of these experiments were to determine whether acoustic emission could be used to locate hydraulic fractures and to study the factors which control hydraulic fracture initiation, either in tension of in shear.