the streams are larger, and where the erosive action can not be modified by enlarging the channel, as is the case in many gorges, it is necessary to make an artificial bed for the stream and at times to supplement this by masonry dams. The dams are not permanent in their effects, for as soon as the basin immediately above the dam has been filled with deposits and the original slope of the stream has re-established itself, the products of the erosion pass over the dam. They prove, however, of great utility at times of large freshets by causing deposits which are subsequently carried down in much smaller quantities. This prevents the disasters that would be caused by sudden enormous deposits when the streams are swollen.
The prospect for many a narrow valley would be a sad one indeed if means had not been taken to prevent the lateral erosions so common with the mountain torrents and so productive of landslips. Where the stream has provided itself with a stony bed on which to roll, it often tries to do damage by leaving the stones and attacking the earthy banks. It is in these cases that the danger has proved so insidious; for until disasters actually occurred, in many instances, the undermining effect of the water was not suspected, being entirely hidden from view.
The most economical way of combating such cases is by making use of those points that by their formation arrest the erosive action. This is done by re-enforcing them in such manner as to produce a series of natural steps. The upward march of the excavating action of the water is stopped at least partially and temporarily at each step. Such a method is of great advantage when it is necessary to delay a more costly correction for financial reasons.
The experience gained since a connected system was begun in the management of the water courses has been of incalculable value, and many have been the landslips arrested and prevented by seeking their causes in the hidden erosive action of a small stream. In still another manner does the water threaten the mountain sides, and that is by permeating the soil (which is thus rendered much heavier), until it reaches a bed of rock or other layer that it can not penetrate. There it forms a layer of slippery mud on which the soil above slides bodily down. The method pursued in such cases is sometimes that of a ditch dug to the impermeable layer, sometimes that of a drain. In both cases there may or may not be small feeding ditches. Another plan which appears very contradictory of what has just been said is also employed. This consists in preventing, by means of horizontal trenches, the water from flowing off on the surface in the regions above the timber line. The water is forced to percolate through the soil and so reaches the wooded portions, where it would permeate anyhow, much more gradually than otherwise. The soil thus