A distinction must be maintained between the crushing strength of a cube of rock or soil, and the bearing power of that soil when it lies as a mass of indefinite extent under some structure. A soil can fail only by being actually displaced by the load above it, or because it has been undermined, perhaps by a stream of water. A sample of rock which might crush with comparative ease when tested as a six-inch cube in a testing machine, will probably withstand as great a concentration of load as it is practicable to put upon it by any engineering structure. Even a gravel which would have absolutely no strength if it were attempted to place a cube of it in a testing machine, will be practically immovable when lying in a pit where it is confined laterally in all directions.