Now suppose that the river has dug itself in as far as it can go. There must be a limit, and the limit is reached when the slope of the bed has been made so slight that the current can only go on languidly. In that case it cannot sweep along stones, or shingle, or even coarse gravel; and then the river so far from deepening its channel begins to raise it by allowing more of the transported sediment to settle down. If a fast stream meets a slower one deposition of material will take place; and the same thing will occur when the rivers meet a lake or a sea. Whatever checks the swiftness of a current weakens its carrying power and causes it to drop some of its sediment to the bottom. Therefore accumulations of sediment occur at the foot of torrent slopes along the lower and more level ground. These deposits we call alluvium, and sometimes when the mountain torrent ends abruptly in the plain they may stand up in cones of silt. They are sometimes called _alluvium cones_ or _fans_. Quitting the steep descents, and reinforced by tributaries on either side, the stream ceases to be a torrent and becomes a river. It goes fast enough at first to carry still coarse gravel; but the big angular blocks of rock have been dropped, and the stones it now leaves in its bed are smaller, and become rounded and smoothed as it goes farther and farther across the plain. At many places it deposits gravel or sand, more especially at the inner side of the curves which the stream makes as it winds down the valley. When the stream runs low in summer, strips of bare sand and shingle are seen at each of these bends; and the stones are always well smoothed and lie on the whole regularly. Those that are oblong are so placed that the greater length of the stone points across the stream; those that are flat usually slope upstream. These facts, though apparently insignificant, are really of importance, because they point to us a method by which geology can determine, after a river has disappeared, the slope of the bed and the direction of the curves which once it had. If we examine the steep banks or cliffs by the side of a river the layers of gravel or shingle in the strata may be found to lie not flat on one another but in sloping planes. That at once will furnish a clue to the direction of the river. Another thing of great importance are the terraces which a river forms by the side of itself. When it overflows in floods it deposits mud on either side, and when after the flood it subsides the mud is left. If the reader will imagine the river in the course of ages sinking lower into its bed he will see that successive eras of flood-levels will leave their mark in a series of steps, or _river terraces_ as they are called, on either side of the channel.