When we come to examine more closely the work which rivers do in removing mineral substances from the land by washing particles of them from the surface, we find that the records they leave in geological history must be plainly marked. Every stream, large or small, is always busy carrying mud, sand, or gravel. Rivers are the "navvies" of geology. When they are swollen by rain they sweep large stones away with them. If we look at the bed of a mountain torrent we shall often see huge blocks of stone that have fallen from the cliffs on either side blocking the pathway of the stream. To all appearance the stream is quite powerless to remove these blocks, and has to circumnavigate them. But visit such a torrent when the snows are melting, or heavy rain has fallen, and you will hear the stones knocking against each other or on the rocky bottom as they are driven downwards by the flood. It is not easy to estimate the driving power of water. M. Gustave le Bon has furnished an illustration of its power which is very curious. In the south of France a stream is led downwards from the mountains to drive the turbine of some machinery at a manufactory. It comes down several thousand feet. In the manufactory there is a vent-hole, out of which the water can be allowed to shoot. The vent-hole is about an inch in diameter; and the water rushes out with such swiftness and force that the water-jet becomes as rigid as steel. It is impossible to cut through this water-jet; and if any one were to try to do so with a sword, the sword might break but it could never pierce or pass through those swiftly moving particles of water. A more commonplace illustration is the use that is sometimes made of water-jets to break up the surfaces of rock in quarries; nor must it be forgotten that horse-power of great value and extent for electric lighting and other purposes is always being drawn from waterfalls. Thus as a mechanical force merely the river can be immensely powerful; and must leave marks of its power on the rocks. The aspect of its force with which we are, however, most concerned is that which is directed to lowering gradually the surface of the land. In the last chapter we showed how much mineral might be dissolved in the waters of rivers. If we are to include also the amount of mud, sand, and other things classed altogether as silt which a river carries down, the figures become much more imposing. Sir Archibald Geikie says that, taking the Mississippi as a typical river (it is as good an example as would be found, because in its great length it passes through many different kinds of land, soil, and climate), we may assume that the average amount of sediment carried down by a river is one part of sediment to every 1500 parts of water.