=Total amount of underground waters.= In order to realize the vast work in solution and cementation which underground waters are now doing and have done in all geological ages, we must gain some conception of their amount. At a certain depth, estimated at about six miles, the weight of the crust becomes greater than the rocks can bear, and all cavities and pores in them must be completely closed by the enormous pressure which they sustain. Below a depth, therefore, water cannot go. Above it all rocks are water-soaked, up to the limit of their capacity, to within a few feet of the surface. Estimating the average pore space of the rocks above a depth of six miles at from two and a half per cent to five per cent of their volume, it is found that the total amount of ground water may be great enough to cover the entire surface of the earth to a depth of from eight hundred to sixteen hundred feet. CHAPTER III RIVERS AND VALLEYS =The run-off.= We have traced the history of that portion of the rainfall which soaks into the ground; let us now return to that part which washes along the surface and is known as the _run-off_. Fed by rains and melting snows, the run-off gathers into courses, perhaps but faintly marked at first, which join more definite and deeply cut channels, as twigs their stems. In a humid climate the larger ravines through which the run-off flows soon descend below the ground-water surface. Here springs discharge along the sides of the little valleys and permanent streams begin. The water supplied by the run-off here joins that part of the rainfall which had soaked into the soil, and both now proceed together by way of the stream to the sea. =River floods.= Streams vary greatly in volume during the year. At stages of flood they fill their immediate banks, or overrun them and inundate any low lands adjacent to the channel; at stages of low water they diminish to but a fraction of their volume when at flood. At times of flood, rivers are fed chiefly by the run-off; at times of low water, largely or even wholly by springs. How, then, will the water of streams differ at these times in turbidity and in the relative amount of solids carried in solution? In parts of England streams have been known to continue flowing after eighteen months of local drought, so great is the volume of water which in humid climates is stored in the rocks above the drainage level, and so slowly is it given off in springs. In Illinois and the states adjacent, rivers remain low in winter and a "spring freshet" follows the melting of the winter's snows. A "June rise" is produced by the heavy rains of early summer. Low water follows in July and August, and streams are again swollen to a moderate degree under the rains of autumn.