From these considerations we perceive how important the movement of waterfalls may be in determining the water level of extensive areas. Not only may their retreat lead to the drainage of extensive inland seas, but as they move

up stream, the drainage of all the tributary rivers, the mouths of which are in turn passed by, have their systems of flow changed in an important manner. Thus, when Lake Erie is drained away, a number of subordinate waterfalls will be developed along the streams which now empty into that basin. Each of these in turn will take up its march toward the head-waters of the river in which it forms; and so the effect of the retreat of one great waterfall may be propagated over the whole surface of the land which is drained by a great stream.

of the surface. The erosive action of the water which passes out of a river is determined by the height through which this water descends in every part of its course. Whatever tends to increase the speed of fall in the particular portion of the basin serves to magnify the erosive work in that region. Thus, when a fall disappears, the energy which was ineffectively applied at the base of its cliff may become distributed over a wide surface in the upper portion of the valley in which it lay.

So, too, in a larger way, as the continents sink down into or rise above the level of the sea, in their ceaseless oscillations, each movement is attended by a great variation in the energy with which the streams act upon their surface. If our continent should rise a hundred feet in its southern parts, the Mississippi River would immediately begin to flow with greater swiftness,

The effect of a retreating waterfall deserves to be considered with some attention, for the reason that it will afford the student the means of understanding how far the structure of the rocks in a country may influence the erosion which water brings to its surface. Each of these hard layers of rocks, as well as the other classes of dams which create waterfalls, tends, by determining the rate of flow of the streams, to fix the rate of erosion in all parts of the river-basin above the point where they occur. Whenever such obstructions are cut away, 'hey increase the rate of fall in the waters above them; and so this may greatly enhance the rate of down-wearing

We thus see

and so too all the streams which are tributary to it would have their energy enhanced up to the foot of their mountain-torrents. On the other hand, if the continent sank down a hundred feet, all these streams would at once become less effective agents of erosion and transportation. that all the erosive work of the land is to a greater or less extent determined by what is called the principle of base level of erosion. This principle, first distinctly suggested by J. W. Powell, has been amplified by other American geologists and has served to bring into clear light the peculiar sensitiveness of our streams to the position of the sea or of hard layers in the rocks

which control the inclination of their stream-beds.

We must now turn our attention to another mode in which water wears away the valleys of streams. So far, we have considered only that portion of the rain which flows over the surface of the ground, but it needs only a moment's notice to show us that this is only one

derground water emerges into the open air and journeys through the streams to the sea, conveying much dissolved matter taken from the rocks through which it passes. Through the action of this underground water, all the rocks for a considerable depth below the surface are constantly diminishing in volume, tiny crevices are formed between

their grains, and the weight of the superincumbent matter in most cases causes the strata to press these crevices together almost as fast as they are formed. This action is particularly conspicuous near the surface of the ground, within the limits of a few score feet in depth. The result is that in

element of the rainfall. If we watch any ordinary soil-covered portion of the earth's surface in a time of rain, we observe that a considerable portion of the water, an amount which varies with the amount of water which falls in a given time and the porosity of the surface, enters into the ground. This subterranean or soil water passes for a great distance beneath the surface of the earth. In this journey, the underground water plays a very different part from that performed by the superficial streams. Except in the rare cases where it forms distinct caverns, it slowly creeps on its way downward to the sea, never attaining a speed of motion which gives it any cutting power whatsoever; but in this underground journey it becomes in most cases charged with carbonic-acid gas and is thus enabled to dissolve more or less of the rocks through which it passes. Finally this un

ley we have the whole area gradually down-sinking by subterranean erosion. A portion of this matter, broken up by the action of penetrating water, remains as the soil-covering, but the interstitial decay and the removal of the matter go on for great depths beneath the soil. So hidden is this process that even those well trained in such observations may not note its effects, but careful inquiry exhibits some very conspicuous results of its operation. In the Southern States of this country, it is often possible to observe a layer of limestone, say five feet in thickness, which at one point has, by some impervious overlying deposit, been protected from the action of penetrating waters. A few hundred feet away we may find the same bed exposed to this percolating erosion of water. At such points we observe that the limy matter has been to a great extent removed from the layer of rock, leaving only the clay or

sand which may have been commingled with it. In this case, the layer will always be greatly diminished in thickness; what was originally a bed five feet thick may become a layer not more than one foot in depth, though the bed may in other respects retain its original

form.

We observe that this interstitial erosion of rocks goes on in a greater or less measure over all parts of the rivervalley. Thus, while a stream-bed is exposed to the actual cutting which the superficial portions of the river may bring about, all portions of its valley are wearing down by the interstitial decay. It will be observed in the cut on page 147, which shows a section crossing a river-valley, that we have in such a basin two distinct topographic features. There is a channel, which, as we readily see, was carved by the flowing stream. On either side, leading up to the divide which separates the river from the next stream, is a more or less gentle slope across a wide field of country. In the main, the downward wearing of this side slope is accomplished by the percolating waters in the manner before noted. To conceive the formation of a river-valley, the observer must in his imagination combine the action of these erosive agents working on the surface and in the under earth. He must imagine an ordinary river to consist not only of the main channel, but of many tributary streams branching like the limbs of a great fan-shaped tree. Each of these branches is slowly swinging to and fro, driven about by the wrestle with its alluvial material. In time, every portion of the valley is crossed again and again by the bed of some stream in its serpentine swings to the right and left of its present path.

It will be well for the student, when standing in some river-valley of normal structure, such as that of the Ohio, or in other river-valleys south of the glacial belt, to imagine a vertical line extending from the present surface to the height of a mile above that level. He should then try to imagine the endless wandering of the streams in their conflict with the detritus which encumbers their beds. He must conceive that the brooks or rivers which are nearest the vertical line

have again and again swung to and fro across its path. If he could restore to the surface, layer by layer, every part of material which had been taken away, and bring to their ancient positions all the several stream-beds he would find his line again and again intersected by them. The time in which the streambeds lay over the given vertical would be but brief. Perhaps, if it were possible to make an actual diagram of their position and duration, with reference to the given vertical line, we should find that not more than one-fiftieth of its space was occupied by the channels of the old brooks or rivers. All the intermediate space not so occupied by the channels indicates the interstitial erosion effected by underground water.

In order to aid the reader in forming this conception as to the history of a river-valley, a cut is given [p. 147] which shows in a diagrammatic way the process by which a river-valley wears downward. On the basis of fact presented in this figure, it will be well for the observer, by the use of his constructive imagination, to frame a picture of the past history of any considerable system of land waters. If this image is well brought to mind, he will have attained one of the greatest conceptions which geology offers to its votaries.

The foregoing considerations will enable the reader, in a general way, to conceive the laws under which a river-system is developed and maintained. It is necessary, however, in order to complete the picture, to set before him certain accidents which may happen in the history of a stream. In the case of a riverbasin such as that of the Ohio, a basin which we frequently take for illustration, for the reason that it is one of the most normal of all those on the American continent, the natural history of the stream is as When the land which now constitutes this great valley first came above the ocean, it was a region of great plains, on which flourished the dense swamps of the Carboniferous era. Through this plain, the streams seem for a time to have wandered deviously, with undetermined channels. Gradually, as the Appalachian and other mountains developed, and the slopes of the streams increased,

follows:

they carved themselves channels; the general course of these channels being determined to a certain extent by the inclination of the rocks. As the Alleghanies rose higher and the table-lands on their banks came to a greater elevation above the sea, the organization of the main river and its tributaries was made more and more complete. If the continent should continue for some geological periods without any change in the level of the sea, the mountain brooks would gradually carve down the hills in which they lie, the table-lands would slowly disappear, and the surface would return to its primeval state of a great swamp. The rocks beneath this swamp would be subjected only to interstitial or corrosive decay, for the reason that the streams would not have fall enough to work upon their beds by mechanical erosion. In proportion as the lands of the valley were high above the sea, the erosive effect of their waters would have great effect. With every foot of diminished height above the ocean-level, the energy of erosion would decrease, while the corrosive, or underground, wearing would remain more nearly steadfast.

It is, from the foregoing considerations, easy to see that the ratio between the erosion and corrosion effected by the rain-fall in a river-basin determines, in a very important way, the aspect of that region. Whereas, in the Ohio, the total descent of the waters in their great distance of flow is relatively small, corrosion may nearly overtake the erosive downwearing, and so the general level of the country will be brought down almost to the river-channel, the main stream being bordered by a line of low escarpments on the margin of its alluvial plains.

For a contrast with the conditions presented by the Ohio, where the rainfall throughout the valley is great, where the elevation of the region is slowly brought about, and therefore the corrosion relatively considerable, let us turn to the case of the lower Colorado, where the stream flows, for some hundreds of miles, through a country which has a very small supply of rain and where it receives very trifling tributaries and where the surface of the country has risen rapidly from the sea. The head-waters of the Colorado in the Rocky Mountains are

fed by the considerable snow-fall of that region; these melting snows maintain a powerful current through the channel of the stream at all seasons of the year. The result is that, while the region on either side of the Colorado has been rapidly elevated during the last geological periods, there has been no proportionate corrosion of the rocks on either side of that stream. The bordering lands have remained for many geological ages little affected by underground water or the to and fro swingings of the lesser streams. The consequences of this peculiar position is that the Colorado flows through a great cañon, which, in places, has the depth of a mile and has the aspect indicated in the picture on page 137.

Between the conditions of the Colorado cañon and those of a valley such as the southern part of the Ohio basin exhibits, we have every degree of divergence of aspect, and the slope of the drainage-basin toward the gorge of the stream indicates in a general way the relative intensity of the erosive and corrosive forces. There is a peculiar effect arising from the diverse hardness of horizontal strata in a river valley, which deserves note in this part of our inquiry. Wherever it has a very hard bed underlaid by softer strata, this hard bed at first makes a precipice next the bank of the stream. If the underlying bed be so little resisting that the weather wears it rapidly away, it will often decay with such speed that the steep face will be driven backward across the country until it finally appears in the form of an isolated tableland as is shown in the cut. Finally, when this table-land, decaying on its several sides, has been reduced much in area, it may appear in the form of what is called in the Mississippi Valley a butte. Such retreat-escarpments are often very conspicuous and beautiful features in the landscape. Excellent examples of such structures occur in horizontally disposed strata on both sides of the Mississippi and in the Saxon Switzerland, where they afford the table-like rocks of that beautiful district-isolated eminences, which, in that region of ancient warfare, are often crowned by fortresses. Such buttes, or tables of rock, only occur where the

strata of a river valley lie in a horizontal attitude and where hard beds and soft are intermingled. Where the rocks of varied hardness depart very much in their attitudes from the horizontal, they greatly affect the flow of the stream as it wears down its bed, in the manner indicated by the accompanying figures. Thus the position of a stream in a valley where the rocks are steeply inclined is determined by the various inclinations of the strata. [P. 146.]

So far we have considered the history of a stream where it has been left free from all natural interference to development. In such conditions, its basin is shaped as the concurrence of the erosive and corrosive forces may determine. In fact, few river-basins enjoy any such immunity from disturbing conditions. Their sensitive streams are variously affected by geological influences of an external sort. As these invading forces profoundly affect the form of river-valleys we may take a glance at their nature. The most common disturbing influence which may affect a river-valley of considerable area arises from the construction of mountain-ridges across the path of its streams. It was once supposed that mountains were suddenly formed. It is now clear that in most, if not in all, cases they have gradually grown to their present height. Now, as the greater number of our mountains lie in the paths of streams which existed before the elevations were formed, it follows that our rivers which intersect mountainridges have had to wrestle with the barriers produced by the elevations. It may in cases have happened that the ridge or wall of a mountain has been suddenly uplifted across the path of a stream, but in most of the cases where we can trace the history of the contention between ridge and stream, we find that the elevation has been formed with such slowness that the river has kept open its channel across the line of the developing obstruction. This leads us to the conclusion that mountains are never, to any extent, barriers to the path of rivers; they probably, in most cases, grow so gradually that the streams may keep their ways open through the obstacle which they tend to interpose. The part which mountains play in the history of

rivers is thus limited to a narrower field than we should at first suppose. They affect the path of rivers by changing the inclination of rocks and so directing the swing of the streams. They also serve to maintain the torrential portion of a river-system, and so afford a ground whence the stream may obtain the alluvium necessary to make the plains which border the lower part of its course. As we have seen, a chemical action which goes on in the material of these deltadistricts serves an important purpose in the economy of the earth's surface. Were it not for the continuance of the mountain-building forces, the torrents, owing to the rapid down-wearing of their beds, would soon cease to afford such detrital material. The combined machinery of torrent and mountain so operates as to maintain the supply of detritus required by the needs of the sea for the maintenance of organic life in its depths and for the deposition of strata on its floor.

There are other and more formidable geologic agents tending to modify riverbasins; the chief of these are glaciers. When a glacial period comes upon a country, the sheets of ice are first imposed upon the mountain tops, and thence the ice creeps down the torrent and riverbeds far below the snow-line, in a manner now seen in Switzerland and Norway. As long as the ice-streams follow the old torrent-channels, they act in something like the fashions of the flowing waters, to gouge out the rocks and deepen the valleys; but as the glacial period advances and the ice-sheet spreads beyond the mountains, enveloping the plains as well; when the glacier attains the thickness of thousands of feet, it disregards the valleys in its movement and sweeps on in majestic march across the surface of the country. As long as the continental glacier remains, its tendency is to destroy the river-valleys. The result of this action is to plane down the whole land and, to a certain extent, to destroy all pre-existing river-systems. During the last glacial period, the old river-valleys were, to a great degree, worn away and the remaining portion of their troughs was, to a considerable extent, buried beneath a thick coating of débris which the ice had worn from the surface