as a krantz, the split portion will fall down in an avalanche of angular blocks. Each block now, at the foot of the hill, will be exposed to the same expansion and contraction as before; but as the corners expose a larger surface to the sun than the flat sides the heat will penetrate farther into the rock at the corners than at the sides. Consequently, when the expansion and contraction has gone on long enough to produce a crack, this will take on a curved surface, cutting off the angles of the block, and after the flakes have fallen away the angular block will have changed to a rounded one. The large rounded boulder now undergoes expansion and contraction as before; the spherical contour allows equal heating on all sides, and eventually the block will split into two or more portions. The angular portions of the boulder will now scale off, and smaller boulders will be produced, and so on. The process is repeated till the original angular block has broken down into small sand grains which can be caught up and carried by the wind, but the sand will consist of all the minerals of the original block quite unaltered. The sharp reports like cannon shots which are often heard in the desert at sundown are caused by the splitting of blocks as the heat absorbed during the day leaves them.
Suppose, now, that there is a little moisture. The fallen blocks become rounded, but in some cases they become soaked with water from dew or by fine rain. The water is absorbed in the pores of the rock, and, lying in the crevices, dissolves any of the substances in the rock which are readily soluble, the iron being usually the substance selected for solution in dolerite. When the sun comes out again the water is drawn to the surface of the boulder and evaporates, leaving behind a thin