ages, from the old "sawyers" that have bowed before the current with rhythmic regularity perhaps for centuries, to the freshly-felled monarchs still bearing their green leaves of the season.
But the fact of chief interest is the presence of trees on the brink of these eroded walls, whose still living and healthy trunks are laid bare to a depth of several feet below the present surface of the ground. In some cases the subterranean portion occupies as many as four or five feet of the base of the trunk, descending through a number of distinct strata. But even at much greater depth there are frequent and unmistakable relics of ancient forests long since destroyed, or, as it were, buried alive. At depths of ten or twelve feet below the present surface, old stumps, with roots and remains of trunks, are brought to light by the inroads of the river. The trees which these represented must have been buried deeper and deeper, in the same manner as existing ones are proved to be undergoing burial, until, unable longer to perform the functions of circulation, they died, and all decayed except these deeply buried parts. Sometimes even these are gone, and naught remains beyond a reddish stain against the vertical wall to mark the spot where once there flourished upon the then surface of the valley a large and healthy tree (see Diagram No. IV).
The method thus far described of studying the mode of formation of the river-valley is that of analysis—the observation of the action of the water in disintegrating it. But we may also employ the method of synthesis, and study the manifest process of valley-building which takes place simultaneously. The river is always loading up on one side, and unloading on the other. The deepest part of the river near the high banks, as it sweeps round the great bends, is also the swiftest. The current grows slower and slower in the direction of the opposite shore, and at the same time the water grows more and more shallow, until at last a sand-bar is reached gradually rising out of it. If this proves to be the mainland, the case is simple, and we will first consider this simple case. This sand-bar was formed at the last period of high water in the spring and early summer. It therefore consists of sand only, without vegetation. It may have a width of fifty or a hundred feet when it ceases, and a distinct rise occurs, with a little terrace of sand, thickly covered with seedling willows, all belonging to one species (Salix longifolia, Muhl.), and bearing no other vegetation. The sand is still damp, being saturated with water from the river. This land is two years old. A short distance farther back another similar terrace is reached, bearing a thicket of this same willow, but it is now two to four feet high, and fruit-bearing. The land is here three years old. Another remove brings us to a third terrace, having larger willows and some other vegetation, such as is not injured by periodical floods flowing over it. This four-year-old soil is darker in color and firmer. It may complete the river-bed proper, or there may be still another terrace. As we recede from the river, these old river-bed marks become
