Since faults are not plane, but undulating and often highly irregular, fractures, the walls will not coincide after slipping; and if the rocks are hard enough to resist the enormous pressure, the cavities or fissures produced in this way may remain open. Now faults are continuous fractures of the earth’s crust, reaching down to an unknown but very great depth; and hence they afford the best outlets for the heated subterranean waters; so that it is common to find an important fault marked on the surface by a line of springs, and these are often thermal. The warm mineral waters on their way to the surface deposit part of the dissolved minerals in the irregular fissures along the fault, which are thus changed to mineral veins. This agrees with the fact that the walls of veins usually show faulting as well as crushed rock, slickensides, and other evidences of slipping. If the earth’s surface were not subject to erosion, every fault would be marked on the surface by an escarpment equal in height to the throw of the fault; and, notwithstanding the powerful tendency of erosion to obliterate them, these escarpments are sometimes observed, although of diminished height. Thus, according to Gilbert, the Zandia Mountains in New Mexico are due to a fault of 11,000 feet, leaving an escarpment still 7000 feet high. But, as a rule, there is no escarpment or marked inequality of the surface, the fault, like the fold, not being distinctly indicated in the topography. In all such cases we must conclude either that the faults were made a very long time ago, or that they have been formed with extreme slowness, so slowly that erosion has kept pace with the displacement, the escarpments being worn away as fast as formed. These and other considerations make it quite certain that extensive displacements are not produced suddenly, but either grow by a slow, creeping motion, or by small slips many times repeated at long intervals of time. JOINTS AND JOINT-STRUCTURE.—This is the most universal of all rock-structures, since all hard rocks and many imperfectly consolidated kinds, like clay, are jointed. Joints are cracks or planes of division which are usually approximately vertical and traverse the same mass of rock in several different directions. They are distinguished from stratification planes by being rarely horizontal, and from both stratification and cleavage planes by being actual cracks or fractures, and by dividing the rock into blocks instead of sheets or layers. The art of quarrying consists in removing these natural blocks; and most of the broad flat surfaces of rock exposed in quarries, are the joint-planes (Fig. 36). Some of the most familiar features of rock-scenery are also due to this structure, cliffs, ravines, etc., being largely determined in form and direction by the principal systems of joints; and we have already seen that the same is true of veins and dikes.