Triaxial compression tests were run on seven orientations of three rocks with pervasive planar anisotropy—slate, phyllite, and schist—to evaluate the effects of cohesion and granularity on anisotropic strength variation and deformational mode. The rocks show a gradation in the percentage of platy minerals, and in the size, shape, and percentage of quartz grains present. Experiments were carried out at confining pressures of 500, 1,000, and 2,000 bars, and all tests were duplicated. The tests on slate were duplicated in two different apparatuses, with and without specimen end lubrication, to evaluate the effects of testing conditions on the experimental results.
The strength data obtained were analyzed in terms of cohesive strength and internal friction. Cohesive strength increases with increasing granularity, but the coefficient of internal friction becomes the principal contributor to the strength variation. The coefficient of internal friction does not appear to change significantly with increased granularity; moreover, it appears to be less sensitive to material variability than does the cohesive strength. Cohesive strength becomes the more significant contributor to strength variation with increased percentage of platy minerals. Cohesion can have a pronounced effect on the mode of deformation.
Curves of differential stress at failure versus anisotropy orientation are concave upward with pronounced strength minima for maximum compression at 30° to 45° to the anisotropy, depending upon the granularity and percentage of platy minerals. Maximum strength occurs in the 90° or 0° orientation, also as a function of material characteristics. The observed strength variation among different orientations is in excellent accord with the extended variable cohesive strength theory of Jaeger (1960). Strength for a given orientation increases linearly for phyllite at pressures up to 2,000 bars; the increase is nonlinear for slate above 1,000 bars.
Significant shear stress developed on the ends of specimens under certain testing conditions that, for one orientation, resulted in the rotation of the maximum principal stress from the vertical loading direction by as much as 30°. As a consequence, for these test conditions measured values of strength for this orientation can be about 25 percent lower than the true strength. Lubrication of specimen ends and a rigid loading piston are believed to give the most accurate results.