The descriptive term "rock mass" encompasses individual block dimensions ranging from centimeters to many tens of meters. Strength and deformability vary both qualitatively and quantitatively as a result of this size range. A key issue is therefore the appropriate size of the test sample. A large body of test data was reviewed to determine the influence of block size on the displacement required to mobilize peak strength. It is shown that the shear strength and shear stiffness reduce with increased block size due to reduced effective joint roughness, and due to reduced asperity strength. Both are a function of the delayed mobilization of roughness with increasing block size. A method of scaling shear strength and shear displacement from laboratory to in situ block sizes is suggested. It is based on the assumption that size effects disappear when the natural block size is exceeded. This simplification appears to be justified over a significant range of block sizes, but is invalidated when shearing along individual joints is replaced by rotational or kink-band deformation, as seen in more heavily jointed rock masses. Recent laboratory tests on model block assemblies illustrate some important effects of block size on deformability and Poisson's ratio.
The wide range of natural block sizes found in nature has a strong and obvious influence on the morphology of a landscape. The contrast in natural slope angles and slope heights sustained by a ravelling "sugar cube" quartzite and a monolithic body of granite suggests that block size may be a controlling factor when compressive strength and slake durability are high in each case. In a tunnel, the contrast in behavior may produce more than an order of magnitude change in costs per meter. It is clear that the strength, the deformability and the mode of deformation (ravelling versus elastic) are strongly controlled by relative block size. The mode of deformation cannot, however, be exclusively tied to block size. The loaded volume relative to block size, and the level of stress relative to the yield stress will each tend to control the mode of deformation. The above factors illustrate the difficulty that often arises in selecting the appropriate sizes of test sample. Size effects will then be evident. On occasion, large size cores may be recovered which include a representative number of interlocked blocks, giving presumably a fair approximation to the strength and deformation behavior of a heavy jointed rock mass.
Major through-going joint sets or individual discontinuities often dominate the stability and deformability of engineered structures in rock. Attempts to sample and test these surfaces are less successful than generally appears. This is because the size of the test sample often determines the magnitude of the strength data obtained. Examples of joint sample size effects are illustrated in Figures 1 and 2. These shear tests were performed at such low normal stress (self-weight) that no shearing of asperities occurred. The marked difference in strength is strictly a function of different effective joint roughness.
