Now it might have been expected, that instead of the floor rising up, the ceiling would sink down, and this effect, called a "Thrust," does, in fact, take place where the pavement is more solid than the roof. But it usually happens, in coal-mines, that the roof is composed of hard shale, or occasionally of sandstone, more unyielding than the foundation, which often consists of clay. Even where the argillaceous substrata are hard at first, they soon become softened and reduced to a plastic state when exposed to the contact of air and water in the floor of a mine. The first symptom of a "creep," says Mr. Buddle, is a slight curvature at the bottom of each gallery, as at _a_, fig. 66.: then the pavement continuing to rise, begins to open with a longitudinal crack, as at _b_: then the points of the fractured ridge reach the roof, as at _c_; and, lastly, the upraised beds close up the whole gallery, and the broken portions of the ridge are re-united and flattened at the top, exhibiting the flexure seen at _d_. Meanwhile the coal in the props has become crushed and cracked by pressure. It is also found, that below the creeps _a_, _b_, _c_, _d_, an inferior stratum, called the "metal coal," which is 3 feet thick, has been fractured at the points _e_, _f_, _g_, _h_, and has risen, so as to prove that the upward movement, caused by the working out of the "main coal," has been propagated through a thickness of 54 feet of argillaceous beds, which intervene between the two coal seams. This same displacement has also been traced downwards more than 150 feet below the metal coal, but it grows continually less and less until it becomes imperceptible. No part of the process above described is more deserving of our notice than the slowness with which the change in the arrangement of the beds is brought about. Days, months, or even years, will sometimes elapse between the first bending of the pavement and the time of its reaching the roof. Where the movement has been most rapid, the curvature of the beds is most regular, and the reunion of the fractured ends most complete; whereas the signs of displacement or violence are greatest in those creeps which have required months or years for their entire accomplishment. Hence we may conclude that similar changes may have been wrought on a larger scale in the earth's crust by partial and gradual subsidences, especially where the ground has been undermined throughout long periods of time; and we must be on our guard against inferring sudden violence, simply because the distortion of the beds is excessive.