The escarpments due to earthquake faults in loose materials at the earth’s surface can obviously retain their steepness for a few years or decades at the most; for because of their verticality they must gradually disappear in rounded slopes under the action of the elements. Smaller displacements within a rock which rapidly disintegrates under the action of frost and sun will likewise before long be effaced. In those exceptional instances where a resistant rock type has had all altered upper layers planed away until a fresh and hard surface is exposed, and has further been protected from the frost and sun beneath a thin layer of soil, its original surface may be retained unaltered for many centuries. Upon such a surface the lightest of sensible shocks, or even the smaller earth movements which are not perceived at the time, may leave an almost indelible record. Such records particularly show that the movements which they register occur upon the planes of jointing within the rock, and that these ready formed cracks have probably been the seats of repeated and cumulative adjustments (Fig. 62). [Illustration: FIG. 62.—Post-glacial earthquake faults of small but cumulative displacement, eastern New York (after Woodworth).] [Illustration: FIG. 63.—Earthquake cracks in Colorado desert (after a photograph by Sauerven).] =Contraction of the earth’s surface during earthquakes.=—The wide variations in the amount of the lateral displacement upon earthquake faults, like those opened in California in 1906, show that at the time of a heavy earthquake there must be large local changes in the density of the surface materials. Literally, thousands of fissures may appear in the lowlands, many of them no doubt a secondary effect of the shaking, but others, like the _quebradas_ of the southern Andes or the “earthquake cracks” in the Colorado desert (Fig. 63), may have a deeper-seated origin. Many facts go to show, however, that though local expansion does occur in some localities, a surface contraction is a far more general consequence of earth movement. In civilized countries of high industrial development, where lines of metal of one kind or another run for long distances beneath or upon the surface of the ground, such general contraction of the surface may be easily proven. Comparatively seldom are lines of metal pulled apart in such a way as to show an expansion of the surface; whereas bucklings and kinkings of the lines appear in many places to prove that the area within which they are found has, as a whole, been reduced. [Illustration: FIG. 64.—Diagrams to show how railway tracks are either broken or buckled locally within the district visited by an earthquake.] [Illustration: FIG. 65.—The Biwajima railroad bridge in Japan after the earthquake of 1891 (after Milne and Burton).] [Illustration: