Chapter III: The Work of Running Water (3)

The importance of weathering in the general processes of erosion is shown in many ways. In regions where the mantle rock is the product of the decay of the solid rock beneath, and such regions constitute a large portion of the earth’s surface, the soil and subsoil represent the excess of weathering over transportation. Since most of the earth’s surface is covered with soil to a greater or less depth, it is clear that, on the whole, weathering keeps ahead of transportation. Again, it is clear that the loosening of rock by weathering greatly increases the erosion which a given amount of moving water can accomplish. Not only this, but weathering plays a much more important rôle in the development of valleys than is commonly realized. This is best illustrated by the valleys of young swift streams. The valley which is not at its top ten times as wide as its stream is rare. The stream which has such a canyon has been cutting chiefly at its bottom. Ignoring its lateral corrasion, which is slight, the valley which it would cut would have a width equal to its own. This is illustrated by Fig. 96. Weathering in its broadest sense is largely responsible for the width of such a valley, in so far as it exceeds the width of the stream. The work of weathering, slope wash, etc., has been to get the material which originally lay between _a_, _b_, and _c_ down to the stream. The stream has then carried it away. The above illustration would not apply to old and sluggish streams, for they, by their meandering, widen their valleys independently of weathering.

Weathering is a part of erosion, but only a part. In so far as it is effected by solution the process involves the transportation of that which is dissolved to some other point. Transportation is also involved to some extent in the other processes of weathering, but the central idea of the processes embraced under this term is the loosening and disrupting of rock by which it is prepared for transportation.

_Transportation._

The second element of erosion is transportation. The transportation of mechanical sediment is to be distinguished from the transportation of materials in solution. In so far as mineral matter is dissolved it becomes, so far as flowage is concerned, a part of the stream. If the quantity dissolved were large it might influence the mobility of the water, but the amount is usually too slight to influence the flow sensibly.

The sediment transported by a stream is either rolled along its bottom or carried in suspension at some higher level. The coarser materials (gravel and sand) are carried chiefly in the former position, and the finer (silt and mud) largely in the latter.

=Transporting power and velocity.=—The transporting power of running water depends on its velocity. The formula expressing the relations between them is as follows: Transporting power, _t_, varies as the sixth power of velocity, _v_, (_tαv⁶_); that is, doubling the velocity of the stream increases its transporting power 64-fold. Strictly speaking, this means that if a stream of given velocity is just able to move a stone of a given size, a stream with double that velocity will be just able to move a stone of the same shape 64 times as large as the first. This may be graphically illustrated as follows: Let a current be supposed just able to move the cube _a_ (Fig. 97). If the current be doubled, twice as much water will strike the same surface with twice the force in the same time; that is, the force exerted on the cube _a_ will be quadrupled. It will, therefore, be able not only to move the one cube, but it will be able to move three other cubes (_b_, _c_, and _d_) besides (Fig. 98). The same current against any other equal surface would also be able to move four small cubes, and there are sixteen such surfaces on the face of the large cube (Fig. 99). It follows that the dimension of the cube which the stream with the doubled velocity can move is four times as great as that of the cube which the original current could move, and the cubical contents of such a cube is 64 times as great as that of the first (64 = 2⁶) (Fig. 99). Swift streams, therefore, have enormously greater power of transportation than sluggish ones. It does not necessarily follow that transportation keeps pace with transporting power; that depends on the accessibility of materials suitable for transportation. A stream of great transporting power, like the Niagara at its rapids, may carry little sediment, because there is little to be had.

The velocity of a stream depends chiefly on three elements—its gradient, its volume, and its load, (i.e., the sediment it is moving). The higher the gradient the greater the volume, and the less the load the greater the velocity. The relation between gradient and velocity is evident; that between volume and velocity is illustrated by every stream in time of flood, when its rate of flow is greatly increased. The relation between velocity and load is less obvious, but none the less definite. Every particle of sediment carried by a stream makes a draught on its energy, and energy expended in this way reduces the velocity. The draught on a stream’s energy of a particle carried in suspension is measured by its mass into the distance it would fall in a unit of time in still water. It follows that a large particle makes a stronger draught on a stream’s energy than the same amount of material in smaller pieces. It follows also that the comminution of sediment facilitates transportation in much more than a simple ratio, for not only can a given amount of energy carry more fine material than coarse, but a larger proportion of a stream’s energy can be utilized in the transportation of the fine.

=How sediment is carried.=—Coarse materials, such as gravel stones, are rolled along the bottoms of the swift streams which carry them. Their movement is effected by the impact of water. The same is true to a large extent of sand grains, especially if they be coarse. So far as concerns the material rolled along the bottom it is to be noted that a stream’s transporting power is dependent on the velocity of the water at its bottom. This is much less than the surface, or even the average velocity. The particles of fine sediments, such as silt and mud, are frequently carried by streams quite above their bottoms, as shown by the roiliness of many streams. A particle of mud is usually a small bit of mineral matter, the specific gravity of which is two or three times that of water. Why does it not sink through the water and come to rest at the bottom of the stream, or suffer transportation as the gravel does?

A particle of sediment in running water is obviously subject to two forces, that of the current which tends to move it nearly horizontally down-stream, and that of gravity which tends to carry it to the bed of the stream. In Fig. 100, the arrows _ab_ and _ac_ represent respectively the relative force of gravity and a current of 5 miles per hour. As a result of these two forces the particle would tend to descend in the general direction of _ad_, a line which represents the resultant of these forces, though not the exact path which a particle acted on by them would take in water. If a river were the simple straightforward current which it is popularly thought to be, a particle in suspension would reach its bottom in the time it would take to sink through an equal depth of still water, for the descent would be none the less certain and none the less prompt because of the forward movement of the water. The current would simply be a factor in determining the position of the particle when it reached the bottom, not the time of reaching it. Very fine particles, like those of clay, though having the same specific gravity as grains of sand, would sink less readily than coarser ones, because they expose larger surfaces, relative to their mass, to the water through which they sink. But even such particles, unless of extraordinary fineness, would presently reach the bottom if acted on only by a horizontal current and gravity. Since even sediment which is not of exceeding fineness is kept in suspension it is clear that some other factor is involved. This is found, in part at least, in the subordinate upward currents in a stream.

Where a bowlder occurs in the bed of a stream (Fig. 101) the water which strikes it is in part forced up over it. If there be many bowlders the process is frequently repeated, and the number of upward currents is great. Any roughness will serve the same purpose, and every stream’s bed is rough to a greater or less extent. Where there are roughnesses at the sides of a channel, currents are started which flow from them toward the center. The varying velocities of the different parts of a stream serve a similar purpose. The curves in a river tend to give the water a rotatory movement. A river is therefore to be looked upon not as a single straightforward current, but as a multitude of currents, some rising from the bottom toward the top, some descending from top to bottom, some diverging from the center toward the sides, and some converging from the sides toward the center. The existence of these subordinate currents is often evident from the boiling and eddying readily seen in many streams. It is, of course, true that the sum of the upward currents is always less than the sum of the downward, so that the aggregate motion of the water is down slope; but it is also true that minor upward currents are common. Sediment in suspension is held up chiefly by such currents, which, locally and temporarily, overcome the effect of gravity. The particles in suspension are constantly tending to fall, and frequently falling; but before they reach the bottom many of them are seized and carried upward by the subordinate currents, only to sink and be carried up again. Even if they reach the bottom, as they frequently do, they may be picked up again. It is probable that every particle of sediment of such size that it would sink readily in still water is dropped and picked up many times in the course of any long river journey, and its periods of rest often exceed its periods of movement.

Independently of the subordinate currents, the different velocities of the different parts of a stream tend to keep materials in suspension by exerting different pressures on the different sides of suspended particles.[46]

River ice sometimes facilitates the transportation of débris which the water alone could not carry. The ice freezes to bowlders in the banks of the streams, to those which are partially submerged, and sometimes to those altogether submerged beneath slight depths of water. When the ice breaks up in the spring such bowlders, buoyed up by the ice, may be floated far down the stream. The influence of ice in this connection is most considerable in high latitudes, but it is of consequence as far south as Virginia, where the river deposits sometimes contain bowlders which the unaided streams could not have carried. _Ground ice_ sometimes forms about bowlders in the bottoms of streams, especially in the quiet pools of turbulent rivers, and floats them to the surface before the surface itself is frozen.[47] In the floods of spring rivers often spread their ice widely over their flood-plains. It is sometimes massed in constricted portions of valleys so as to form great dams, the breaking of which is attended with great destruction.

_Corrasion._

=Abrasion.=—The wear effected by running water is _corrasion_. So long as the materials to be carried away are incoherent it is easy to see how running water picks them up and carries them forward. The water which gathers in the depressions on the slope of a cultivated field gathers earthy matter from the surface over which it passes, even before it is concentrated into rills, and the rills continue the process. Thus the loose materials of the surface are gathered at the very sources of the streams, and the amount of sediment in the water after a heavy shower, even at the head of the stream, may be great. The run-off from the slopes of any valley in any part of its course likewise brings sediment to the stream, which gathers more from its bed whereever it flows with sufficient velocity over incoherent material. Streams also undercut their banks, and receive new load from the fall of the overhanging material.

By far the larger part of the sediment acquired by a normal stream is made up of material loosened in advance by the processes of weathering. The stream, or the waters which get together to make the stream, find them ready-made; but rivers frequently wear rock which is not weathered, for the principal valleys of the earth’s surface are cut in solid rock, and many of them in rock of exceeding hardness. How does the stream wear the solid rock?

When a stream flows over a rock bed, the wear which it accomplishes depends chiefly on the character of the rock, the velocity of the stream, and the load it carries. If the rock be stratified and in thin layers, and if these thin layers be broken by numerous joints at high angles to the stratification planes, the impact of the water of a clear stream of even moderate strength may be effective in dislodging bits of the rock. This condition of things is often seen where streams run on beds of shale or slate. If the rock be hard and without bedding-planes and joints, or if its layers be thick and its joints few, clear water will be much less effective. If the surface of the rock be rough, the mechanical action of a swift stream of clear water might still produce some effect on it; but if massive hard rock presents a smooth surface to a clear stream, the mechanical effect of even a swift current is slight.

This general principle is illustrated by the Niagara River. Just above the falls the current is swift. When the river is essentially free from sediment, the surface of the limestone near the bank beneath it is sometimes distinctly green from the presence of the one-celled plants (fresh-water algæ) which grow upon it. The whole force of the mighty torrent is not able to sweep them from their moorings. Were the stream supplied with a tithe of the sand which it is capable of carrying, it would not take many hours, and perhaps not many minutes, to remove the last trace of vegetation. This illustration furnishes a clue to the method by which the erosion of solid rock in a stream’s bed is effected.

It has been seen that the ingathering waters which make a stream often have abundant sediment before they reach well-defined stream channels, and that the streams continue to gather sediment whereever their beds are composed of material which is readily detached. The sediments which the stream carries are the tools with which it works. Without them it is relatively impotent, so far as the abrasion of solid rock is concerned; with them, it may wear any rock over which it passes (Fig. 102).

We have next to inquire the methods by which running water uses its tools in the excavation of valleys. When gravel is rolled along in the channel of a stream there is friction between it and the bed over which it moves. If the pebbles be as hard as the bed over which they are rolled their movement must result in its wear, and even if they be softer more or less wear takes place. As the moving stones wear the rock of the stream’s bed they are themselves worn by impact with it and with one another. In all cases the softer material suffers the more rapid wear. The first effect of wear on materials in transportation is the reduction of their rugosities of surface. The projecting points and sharp angles are worn off, and the stones are reduced to rounded _water-worn_ forms. The particles broken off make grains of sand, or, if very fine, particles of silt or mud. Even after a stone has been rounded it is subject to further wear and reduction, and in the course of time may be literally worn out.

The sediment carried in suspension, as well as that rolled along the bottom, may wear the rock bed of a stream. When a grain of sand in suspension escapes from an upward moving current it may not sink quietly. If it be caught by a downward current it may be made to strike a blow on the bed of the stream, and the effect of the blow is to wear the surface which receives it. The larger the grain and the stronger the current the greater the wear.

The ceaseless repetition of the blows struck by the material in suspension, or rolled on its bottom, hour after hour, day after day, and year after year, will accomplish sensible results. In the long course of the ages this process has excavated deep valleys. Concomitant processes are largely concerned in making valleys wide, but the depth of valleys cut in solid rock is chiefly the result of the impact and friction of the sediment in transportation.

The wear effected in this way is not proportional to the number of blows struck. Since every pebble and every grain of sand carried diminishes the velocity of a stream, and since with diminished velocity the force of the blows struck is diminished, it follows that the blows may become so weak, as the result of their multiplication, as to be ineffective. The larger the load, therefore, which the stream carries, the more the tools with which it has to work, but the less effectively can it use them; and the load may be so far increased as to destroy its corrasive power altogether. On the other hand, the smaller the load of the stream the greater its velocity and the more effectively will its tools be used; but their number may be so far reduced that their aggregate effect is slight. To accomplish the greatest results on a bed of solid rock a stream must have tools to work with, but must not be so heavily burdened as to interfere with its effective use of them.

Whatever the cause of their unequal velocities swift and slow streams corrade their valleys differently. The erosion of a swift stream is chiefly at the bottom of its channel. The sluggish stream lowers its channel less rapidly, while lateral erosion is relatively more important. The result is that slow streams increase the width of their valleys more than the depth, while swift streams increase the depth more than the width. It follows that slow streams develop flats, while swift ones do not. Not only is a slow stream more likely to have a flat, and therefore a better chance to meander, but it is more likely to take advantage of opportunities in this line, for a slow stream gets out of the way for such obstacles as it may encounter, while a swift stream is much more likely to get obstacles out of its way.

Special phases of corrasion are introduced where waterfalls and other peculiarities dependent on inequalities of rock resistance occur.

=Solution.=—In most cases the solution effected by a stream is much less important than its mechanical work. Only when conditions are unfavorable to the latter, is solution the chief factor in the excavation of a valley. This may be the case where a stream’s bed is over soluble rock, such as limestone, and where the stream is clear, or its gradient so low that its current is sluggish. The solvent power of water is not influenced by the presence of sediment, though the presence of sediment offers the water a greater surface on which to work.

CONDITIONS AFFECTING THE RATE OF EROSION.

In considering the rate of erosion, both the work of the stream in its valley and that of the general run-off are to be considered. The conditions which favor the most rapid erosion in a stream’s channel are not necessarily those which determine most rapid degradation in the basin outside of the valley.

_The Influence of Declivity._

In general the greater the declivity the more rapid the rate of erosion, whether in the stream’s channel or on the slopes above it. The truth of this conclusion is illustrated by the great erosive power of swift streams as compared with slow ones.

It does not follow, however, that high declivity favors each element of erosion. The effect of declivity on weathering is far from simple. For example, great declivity, by allowing more of the rainfall to flow off over the surface, and by causing it to flow off more promptly, restricts the work of solution, and therefore of decomposition, both at the surface and beneath it. High declivity is also unfavorable to the growth of vegetation, and so to the wedge-work of roots. On the other hand, a given amount of wedge-work of roots and ice is more effective where the slope is steep than where it is gentle, for such materials as are loosened descend the slopes more readily. The prompt removal of weathered materials, by exposing fresh surfaces of rock, accelerates weathering. The total amount of weathering may therefore not be diminished by the increase of slope, even though certain of its processes are hindered.

The effect of high declivity on transportation, the second element of erosion, is too patent to need explanation.

Corrasion likewise is favored by high declivity, for the abrasive power of a stream increases as the square of its velocity. With corrasive power increased, corrasion will also be increased if the water has tools to work with. Since high declivity greatly increases both the transporting and the corrasive power of running water, and favors certain elements of weathering, it is clear that the aggregate effect of high declivity is to favor erosion, whether in the channel of the stream or on the general surface of its drainage basin.

_The Influence of Rock._

The physical constitution, the chemical composition, and the stratigraphy of a rock formation, influence the rate at which it may be broken up and carried away. Clastic or fragmental rocks are usually stratified and made up of cemented pebbles (conglomerate), sand grains (sandstone), or particles of mud (shale). Igneous rocks, such as granite, are massive instead of stratified, and are usually made up of great numbers of interlocking crystals which bind one another together. Some crystalline rocks, such as schists, though not stratified, possess cleavage, which has much the effect of stratification, so far as erosion is concerned. All rocks are affected by systems of more or less nearly vertical cracks called _joints_. All these structures have their influence upon the rate of degradation.

=Physical constitution.=—Clastic rocks may be firmly cemented, or their constituents may be loosely bound together. The less the coherence the more ready the disintegration, and the finer the particles the more easily are they carried away. When the particles in transportation are angular they effect more wear on the bed over which they move, and on one another, than when they are round. The difference is great where the particles are large, and little where they are very small. If the materials carried be harder than the bed over which they pass, corrasion of the latter is favored.

=Chemical composition.=—Something also depends on the chemical composition of the rock, since this affects its solubility, and therefore its rate of decomposition. The more soluble the rock the larger the proportion of it which will be taken away in solution; but it does not follow that the most soluble rock will be most rapidly eroded, since the rate of erosion depends on abrasion as well as solution, and a rock which is readily soluble, as rocks go, may be less easily abraded than a rock which is made of discrete and insoluble particles bound together by a soluble cement. In such rocks, for example a sandstone in which the grains are cemented together by lime carbonate, the solution of the cement sets free a considerable quantity of sand, so that a small amount of solution prepares a large amount of sediment for removal. A stream might cut its valley much more rapidly in such a sandstone than in a compact limestone, though the latter is, as a whole, the more soluble. The constituent minerals of crystalline rocks resist solution and decay unequally, and when any one is dissolved or decomposed the rock crumbles and the less soluble constituents are ready for removal by mechanical means. So long as the material loosened by disintegration is removed, chemical heterogeneity favors erosion; but if the loosened débris is not removed erosion is not favored by chemical heterogeneity. In such a case erosion would be most rapid where the rock was most soluble.

=Structure.=—The structure of the rock has much to do with the rate of its erosion. Other things being equal, stratified rock is more readily eroded than massive rock, since stratification-planes are planes of cleavage, and therefore of weakness. Taking advantage of these planes the water has less breaking to perform to reduce the material to a transportable condition. For the same reason a thin-bedded formation is more easily eroded than a thick-bedded one.

The beds of stratified rock may be horizontal, vertical, or inclined, and inclined strata may stand at any angle between horizontality and verticality. In indurated formations the rate of erosion is influenced both by the position of the strata and by the relation of the direction of the flowing water to their dip and strike. On the whole the strata which are horizontal, or but slightly inclined, are probably less favorable for rapid erosion than those which are vertical or inclined at considerable angles. This is at least true where the layers are of uniform hardness and the joints infrequent.

Horizontal strata expose fewer cleavage planes to the water flowing over them than strata in any other position. In Fig. 103 the stream which has the profile ad crosses bedding-planes at _b_ and _c_. In Fig. 104, where the beds dip up-stream, many more division-planes are crossed in the same distance. Since bedding-planes are planes of weakness, it follows that horizontal and nearly horizontal strata are not, under ordinary conditions of erosion, in a position favorable for most rapid wear. When strata are horizontal, it makes no difference which way the stream runs, for the current sustains the same relation to the cleavage-planes whatever its course.

In the case of incoherent material the position of the beds, or even their existence, has little influence on the rate of erosion. Such formations are weak in all directions, not simply along bedding-planes.

When the strata are vertical, three distinct cases may arise (Fig. 105). The stream may flow (1) with the strike (_aa_); (2) at right angles to the strike (_bb_); or (3) oblique to it (_cc_) at any angle whatsoever. It is perhaps not possible to say which of these positions is most favorable for erosion, for the character of the rock, the thickness of its layers, its ability to stand with steep slopes, and the strength of the currents concerned, would influence the result. A stream which flows at right angles to the strike (_bb_, Fig. 105) would cross more cleavage-planes in a given distance than a stream flowing in any other direction, and would strike the outcropping edges of layers at the angle of greatest advantage. A stream flowing along the strike (_aa_), on the other hand, has better opportunity to sink its channel on cleavage-planes, and the current oblique to the strike (_cc_), has some of the advantages of each of the others.

When the strata are inclined five cases may arise. (1) The stream may be parallel to the strike (_aa_, Fig. 106), when it makes no difference which way the current flows; it may be at right angles to the strike (_bb′_), and (2) flowing with the dip (toward _b′_), or (3) against it (toward _b_); it may be oblique to the strike, and flowing (4) in the general direction of dip (toward _c′_); or (5) in the opposite direction (toward _c_). As before, the stream flowing at right angles to the strike would cross the largest number of layers in a given distance, and so have an opportunity to take advantage of more cleavage-planes than a stream in any other position. But in the case of inclined strata a new element enters into the problem. When the stream flows parallel to the strike, the valley which is in process of deepening is not sunk vertically, but is shifted more or less in the direction of the dip (Fig. 107). This is called _monoclinal shifting_. The result is that there is a constant tendency to undermine (sap) the valley bluff on the down-dip side, and this process of sapping will, according to its rate, accelerate the growth of the valley, especially in width. Monoclinal shifting is favored by the presence of a hard layer (_H_), as shown in Fig. 107, if this stratum is the bed of the stream.

In the second and third cases mentioned above, the only difference is in the angle at which the current strikes the outcropping edges of layers and laminæ. The mechanical advantage is with the stream which flows with the dip. In the fourth and fifth cases something will depend on the angle which the stream’s course makes with the strike. In all these cases, as in those where the strata are vertical, much will depend on the thickness and resistance of the layers and on the strength of the currents concerned.

_The Influence of Climate._

Climate has both a direct and an indirect effect on erosion. Its direct influence is through precipitation, evaporation, changes of temperature, and wind; its indirect, through vegetation. Like declivity and rock structure, climate does not affect all elements of erosion equally.

The chief elements of climate are temperature, moisture, and atmospheric movements; the principal factors which influence it are latitude, altitude, distance from the sea, direction of prevailing winds, and topographic relations.

The effects of variations in temperature on rock weathering have already been discussed (p. 43). They are chiefly mechanical, and are seen at their best where the daily range is great.

High temperature favors chemical action, and the weathering of rock by decomposition is at its best in the presence of abundant moisture in regions where the temperature is uniformly high. Furthermore, a warm moist climate favors the growth of vegetation, the decay of which supplies the water with organic acids which greatly increase its solvent power. The climatic conditions favoring mechanical weathering are therefore different from those favoring chemical weathering. High temperature and abundant moisture and vegetation are found in many tropical regions, and here the rock is often decomposed to greater depths, on the whole, than in high latitudes. How far this is the result of rapid weathering, and how far of slow removal, due in part to the protective influence of the plants, cannot be affirmed. If the weathered material is not removed, it will presently become a mantle thick enough to retard the processes which brought it into existence.

So long as the water of the surface and that in the soil remains unfrozen, temperature affects neither corrasion nor transportation. But in middle and high latitudes the surface is frozen for some part of each year. During this time corrasion is at a minimum, for although the streams continue to flow there is relatively little water running over the surface outside the drainage channels, and that little is relatively ineffective. Under some conditions, therefore, temperature affects both corrasion and transportation.

The humidity of the atmosphere has an influence even more important than that of temperature on the rate of erosion, and its influence is exerted on each of the elements of that complex process. A moist atmosphere favors oxidation, carbonation, hydration, and the growth of vegetation, all of which promote certain phases of rock weathering. On the other hand, humidity tends to prevent sudden and considerable variations in temperature, thus checking the weathering effected by this means. Precipitation, the most important single factor in determining the rate of erosion, is dependent on atmospheric humidity. Its amount, its kind (rain or snow), and its distribution in time, are the elements which determine its effectiveness in any given place.

Other things being equal the greater the amount of precipitation the more rapid the corrasion and transportation. Much, however, depends on its distribution in time. A given amount of rainfall may be distributed equally through the year, or it may fall during a wet season only. The maximum inequality of distribution would occur if all the rainfall of a given period were concentrated in a single shower. With such concentration the volume of water flowing off over the surface immediately after the down-pour would be greater than under any other conditions of precipitation, and since velocity is increased with volume, and erosive power with velocity, it follows that the erosive power of a given amount of water would be greater under these circumstances than under any other. Furthermore, a larger proportion of the precipitation would run off over the surface under these circumstances than under any other, for less of it would sink beneath the surface and less would be evaporated. If erosive power and rate of erosion were equal terms, this would therefore be the condition for greatest erosion; but erosive power and rate of erosion do not always correspond. If the water falling in this way could get hold of all the material it could carry, extreme concentration of precipitation would be the condition favorable for most rapid erosion. But if the amount of available material for transportation is slight, a large part of the force of the water could not be utilized in erosion. It follows that if there were a large amount of disintegrated material on the surface, erosion would be greater the greater the concentration of precipitation. If, on the other hand, there were but little disintegrated material on the surface, frequent showers, with intervening periods when conditions were favorable for weathering, that is, for preparing material for transportation, might be more favorable for rapid erosion. While the total energy of running water available for erosion under these conditions would be less than before, there might in the long run be more material for transport; for weathering in the presence of moisture, and all that goes with it, might be more effective in preparing material for transportation, than weathering during the long periods of drought which would occur if the precipitation were concentrated to its maximum. Temperature favoring, the uniform distribution of moisture through the year would allow the growth of vegetation, which, although favoring some processes of weathering, retards erosion in general. While therefore it is not possible to say what distribution of rainfall favors most rapid erosion without knowing the nature of the surface on which it is to fall, enough has been said to show that the problem is by no means a simple one. Some of the most striking phases of topography developed by erosion, such as those of the Bad Lands (Figs. 75 to 78, and 108), are developed where the rainfall is unequally distributed in time, and too slight or too infrequent to support abundant vegetation.

During its fall, and immediately after, rain is more effective than an equal amount of snow; but the snow may be accumulated through a considerable period of the year, and then melted rapidly, when it has an effect comparable to that which would be produced by the concentration of the rainfall into a limited period of the year. If the ground beneath be frozen when the snow melts (and this is often the case) the erosion accomplished by the resulting water will be diminished.

Except in dry regions, where wind-work sometimes exceeds water-work, the movements of the atmosphere are of less importance directly than precipitation in determining the rate of erosion. But even in regions which are not arid the winds have much to do with the rate of evaporation and the distribution of rainfall, so that their indirect effect is great. Even their direct effects in moist climates are not to be lost sight of, for even here the surface is sometimes dry enough to yield dust and sand, and the uprooting of trees so disturbs the surface as to make earthy débris more accessible to wind and water. Where trees gain precarious footholds on steep slopes, as they often do, they are likely to be overturned as soon as they are large enough to offer considerable resistance to the wind, and in the overturning, large quantities of rock are sometimes loosened and carried down the slope by gravity. This phase of destructive work is seen at its best on the walls of gorges, where trees often flourish until their tops project above the rim of the valley.

Through vegetation, climate influences erosion in ways which are easily defined qualitatively, but not quantitatively. Both by its growth (wedge-work of roots) and by its decay (supplying CO₂, etc., to descending waters) it favors certain phases of weathering; but, on the other hand, it retards corrasion and transportation both by wind and water. This is well shown along the banks of streams and on the faces of cliffs, in clay, sand, etc. Its aggregate effect is probably unfavorable to erosion by mechanical means, and favorable to that by chemical processes.

Erosion in high arid regions differs from that in regions of abundant rainfall in several ways. It is obvious that the valleys will develop more slowly in the former, that they will remain young longer, that the period necessary for the dissection of the surface is greater, that the watercourses will be less numerous, and that fewer of them will have permanent streams. There are certain other differences which are less obvious. If the arid region be high and composed of heterogeneous strata, the topography which erosion develops is more angular (Fig. 83) than that of the humid region. This is because there is less rock decay, and less vegetation to hold the products of decay. The more resistant beds of rock therefore come into greater prominence, especially on slopes, where they develop cliffs (Figs. 109 and 110). These general principles find abundant illustration in the plateaus of the western part of the United States,[48] where the cliffs are by no means confined to the immediate valleys of the streams (Fig. 1, Pl. XII).

EFFECTS OF UNEQUAL HARDNESS.

In the preceding pages incidental reference has been made to the results of inequalities of rock resistance. This topic will now be considered more fully.

=Rapids and falls.=—Returning for a moment to the hypothetical island with which our study of erosion began, let a horizontal layer of hard rock be assumed to run through it (H, Fig. 111). As the rain falls on the land and runs off over it, wear will be less rapid where the hard layer comes to the surface than at the higher or lower levels. As a result, the slope will become steeper at and below the outcrop of the hard layer, and less steep immediately above it, as shown by _ab_ in Fig. 111. Under these conditions the water passing over the hard ledge constitutes _rapids_. The increased erosion which accompanies the increased velocity makes the rapids more rapid. The process may continue until the water _falls_, rather than _flows_ over the hard layer (_cd_, Fig. 111). With continued rainfall the edges of the hard layer, together with the slopes above and below, would continue to recede toward the center of the island. Under conditions of absolute homogeneity of material, save for the hard layer specified, no valley would be developed, and therefore no stream.

If the surface was so changed as to allow of the development of a valley (p. 63) the same principles would be applicable. As an active stream passes from a hard layer to one less resistant, the greater wear on the latter gives origin to rapids. At first the rapids would be slight (_a_, Fig. 112), but would become more considerable (_b_) as time and erosion go on. When the bed of the rapids becomes sufficiently steep, the rapids become _falls_[49] (_cd_). When the water falls rather than flows over the rock surface below the hard layer, erosion assumes a new phase. The hard layer is then undermined, and the undermining causes the falls to recede. This phase of erosion is sometimes called _sapping_.

If the hard layer which occasions a fall dips up-stream (Fig. 112), its outcrop in the stream’s bed becomes lower as the fall recedes (_e_). When it has become so low that the water passing over it no longer reacts effectively against the less resistant material beneath (_f_), sapping ceases, and the point of greatest erosion may be shifted from the soft material beneath the fall to the hard layer itself. The actual rate of erosion at this point may be no greater than before, though the relative rate is. Under these circumstances the vertical edge of the hard layer will presently be converted into an incline (_f_), and as this takes place the fall becomes rapids. The conversion of the falls into the rapids begins about the time the lower edge of the hard stratum in the channel reaches grade. By continuation of the process which transformed the falls into rapids, the rapids become less rapid, and when the upper edge of the hard layer has been brought to grade, the rapids disappear (_h_, Fig. 112). The history of rapids which succeed falls is the reverse of that which preceded. The later rapids are steepest at the beginning of their history, the earlier at their end. Stated in other terms, rapids are steepest when nearest falls in time. Slight differences in hardness in successive layers often occasion successive falls or rapids (Fig. 114).

If the hard layer which occasions the falls be horizontal, instead of dipping up-stream, the general result would be the same; but, other things being equal, the duration of the falls developed under these conditions would be greater, since they must recede farther before becoming rapids.

If the layers of unequal hardness in a stream’s bed be vertical and the course of the stream at right angles to the strike, rapids, and perhaps falls, will develop (Fig. 115). The chances for falls are greater, the greater the difference in hardness. Falls developed under these conditions, as well as the rapids preceding and following, would remain constant in position until the resistant layer was brought to grade, but they would ultimately disappear as in the preceding cases. Falls are not likely to develop where the strata of the stream’s bed dip down-stream, though they may develop even under these conditions if the gradient of the stream is greater than the dip of the strata (Fig. 116).

The inequality of resistance in the rock which occasions a fall may be original or secondary. In the case of Niagara Falls[50] (Fig. 113) relatively resistant limestone overlies relatively weak shale. At the Falls of St. Anthony (Minneapolis) limestone overlies friable sandstone. The falls of the Yellowstone and the Shoshone Falls of the Snake River (Idaho), are in igneous rock. In the former case the unequal resistance is occasioned by unequal decay of the rock, due perhaps to the rise of hot vapors which have decomposed the rock along the lines of their ascent; in the latter, a more resistant sort of igneous rock overlies a less resistant.

Structural features, such as jointing, sometimes give rise to falls, or determine their distinctive features (Fig. 117), even where the formations involved are of uniform hardness. A joint plane has the effect of a weak vertical or highly inclined bed. If an open joint is discovered in a stream’s bed, the water enters it. If it finds an outlet below, a channel is worn along the new line of flow, with rapids or falls where the water descends. Rock originally homogeneous may be much fractured in some parts, while it remains unbroken in others. Where a stream passes from the solid to the broken portion rapids, or even falls, may develop.

Falls may originate in still other ways. If for any reason a stream is forced out of its valley, it may in its flow find entrance to another valley, or to another part of its own valley, over a steep slope. If the structure of the slope favors, a fall may speedily develop. The Falls of St. Anthony are an example, the Mississippi having been turned out of its earlier course by deposits of glacial drift. Again, if an obstruction of any sort, such as a flow of lava, dams a stream, rapids or falls are developed where the water overflows the dam. When a main valley is notably deepened by glaciation the drainage from tributary valleys may fall into it, if the tributaries were not equally deepened. Falls which originated in this way are common in the western mountains of the United States, as well as in most mountain regions recently affected by local glaciers (Fig. 118).

One waterfall often breeds others. Thus where a fall recedes beyond the mouth of a tributary stream, the tributary falls. The Falls of Minnehaha, on a small tributary to the Mississippi, near Minneapolis, may serve as an illustration. In such cases the falls may not develop from rapids. Once in existence, the fall of a tributary follows the same history as that of a main stream.

Streams which have falls are relatively clear.[51] If a stream favorably situated for the development of a fall carried a heavy load, deposition would take place below the rapids, and the tendency would be to aggrade the channel at that point and so to prevent the development of the fall. Falls occur only on streams which have relatively high gradients. This means that the streams which have falls are well above base-level, and streams well above base-level are young. Falls therefore are a mark of topographic youth.

The fall of the Niagara[52] (Pl. IX) is one of the most remarkable known, both because of its large volume of water and its great descent, between 160 and 170 feet. The rate at which the fall is receding is a matter of interest not only in itself, but because, once determined, it may be made to serve as a unit of measurement for certain important events in geological history. It was formerly conjectured that this fall was receding at the rate of one to three feet per century, but it was not until recent years that its actual rate of recession was approximately fixed. By surveys executed in 1842 and 1890 it has been determined that its average rate of recession between those dates was something like 4½ feet per year, or about 150 times as great as the highest estimate stated above. It is to be noted that this is the _average_ rate of recession, for all parts of the ledge over which the water falls are not receding at the same rate. The point of the “Horseshoe” has, during the same time, gone back at more than twice this rate.[53]

Rapids and falls sometimes occasion the development of _pot-holes_ (Fig. 119), a peculiar rather than important erosion feature. The holes are excavated in part by the falling and eddying of silt-charged water, but chiefly by stones which the eddies move. Pot-holes which are not now in immediate association with rapids or falls often point to the former existence of rapids or falls.

=Rock terraces.=—The tendency to sapping shown in many waterfalls is also shown in the weathering and erosion of the sides of a valley where a hard layer outcrops above the bottom, and the profile of the side slopes of the valley simulates that of the stream; that is, the slope becomes gentle just above the hard layer, and steep, or even vertical, at and below its outcrop. This is illustrated by Fig. 120, where the hard layer through which the stream has sunk its valley stands out as a rock terrace on either side of the valley. Such terraces are not rare and are popularly believed to be old “water-lines”; that is, to represent the height at which the water once stood. In one sense this interpretation is correct, since a river has stood at all levels between that of the surface in which its valley started, and its present channel, but the shelf of hard rock does not mean that the river, after attaining its present channel, was ever so large as to fill the valley to the level of the terrace. Rock terraces may also result from changes of level.

=Narrows.=—Inequalities in hardness occasion another peculiarity common to valleys. If a stream crosses vertical or highly inclined strata of unequal hardness, its valley is usually constricted at the crossing of the harder layers. If such a constriction be notable it is called a _narrows_, or sometimes a _water-gap_ (Figs. 121, 159, and Fig. 2, Pl. XII). The Appalachian Mountains afford numerous examples. The constriction arises because the processes which widen the valley are less effective on the hard layer than on the less resistant ones on either hand. Though most narrows are due to the superior resistance of the rock where they occur, they are sometimes the result of other causes.

Narrows are much more conspicuous in certain stages of erosion than in others. While a valley is still so young as to be narrow at all points, no narrows will be conspicuous; but at a later stage in its history, when the valley is otherwise wide, narrows are more pronounced. At a still later stage, when the hard strata themselves approach base-level, the narrows again become inconspicuous.

From what has preceded it is clear that rapids or falls are likely to occur at narrows, especially in the early part of their history.

=Other effects on topography.=—Inequalities in the hardness of rock develop certain peculiarities of topography other than those of valleys. The less resistant portions of a land area more or less distant from streams are worn down more readily than those which are more resistant. If great areas of high land be capped with hard rock they are likely to remain as plateaus after surrounding areas of less resistance are brought low. If the hard capping affects a small area instead of a large one, the elevation is a butte, a hill, or a mountain, instead of a plateau (Fig. 110). Many buttes and small mesas are but remnants of former plateaus (Mesa Lauriano, N. M., Fig. 1, Pl. XII). A feature of buttes and mesas capped by hard rock is the steep slope or cliff corresponding to the edge of the hard bed (Figs. 78 and 109).

If the rock of a region be stratified and the layers tilted, the removal of the softer beds leaves the harder ones projecting above the general level in the form of ridges or “hog-backs” (Figs. 122 and 123). Dikes of igneous rock, harder than the beds which they intersect, likewise become ridges after the degradation of their surroundings. The plugs of old volcanic vents and other igneous intrusions of limited area often constitute conspicuous hills or mountains after erosion has removed their less resistant surroundings (Fig. 124). Inequalities of hardness are therefore responsible for many hills and ridges. In the isolation of the hills and ridges picturesque coves are developed, where the attitude and distribution of the weak and strong rocks are propitious. The bottoms of the coves are located on the weak rocks, and above them rise the precipitous slopes of the resistant ones. Round valley (Fig. 1, Pl. XVII, High Bridge, N. J., quadrangle, U. S. Geol. Surv.) and the coves about the head of Hiawassee River (Dahlonega, Ga., quadrangle) are examples.

Ridges and hills resulting from the unequal degradation of unequally resistant terranes are not equally prominent at all stages in an erosion cycle. In early youth the material surrounding the hard bodies of rock has not been removed; in early maturity considerable portions of their surroundings still remain about them; but in late maturity or early old age the outcropping masses of hard rock have been more perfectly isolated and are most conspicuous. Most of the even-crested ridges of the Appalachian system, as well as many others which might be mentioned, became ridges in this way. In the final stages of an erosion cycle the ridges of hard rock are themselves brought low. Isolated remnants of hard rock which remain distinctly above their surroundings in the late stages of an erosion cycle (Fig. 124) are known as _Monadnocks_, the name being derived from Mount Monadnock, N. H., an elevation of this sort developed in a cycle antedating the present.