The diversity in the form of river valleys is exceedingly great. Almost all the variety of the landscape is due to this impress of water action which has operated on the surface in past ages. When first elevated above the sea, the surface of the land is but little varied; at this stage in the development the rivers have but shallow valleys, which generally cut rather straight away over the plain toward the sea. It is when the surface has been uplifted to a considerable height, and especially when, as is usually the case, this uplifting action has been associated with mountain-building, that valleys take on their accented and picturesque form. The reason for this is easily perceived: it lies in the fact that the rocks over which the stream flows are guided in the cutting which they effect by the diversities of hardness in the strata that they encounter. The work which it does is performed by the hard substances that are impelled by the current, principally by the sand and pebbles. These materials, driven along by the stream, become eroding tools of very considerable energy. As will be seen when we shortly come to describe waterfalls, the potholes formed at those points afford excellent evidence as to the capacity of stream-impelled bits of stone to cut away the firmest bed rocks. Naturally the ease with which this carving work is done is proportionate to the energy of the currents, and also to the relative hardness of the moving bits and the rocks over which they are driven.

So long as the rocks lie horizontally in their natural construction attitude the course of the stream is not much influenced by the variations in hardness which the bed exhibits. Where the strata are very firm there is likely to be a narrow gorge, the steeps of which rise on either side with but slight alluvial plains; where the beds are soft the valley widens, perhaps again to contract where in the course of its descent it encounters another hard layer. Where, however, the beds have been subjected to mountain-building, and have been thrown into very varied attitudes by folding and faulting, the stream now here and now there encounters beds which either restrain its flow or give it freedom. The stream is then forced to cut its way according to the positions of the various underlying strata. This effect upon its course is not only due to the peculiarities of uplifted rocks, but to manifold accidents of other nature: veins and dikes, which often interlace the beds with harder or softer partitions than the country rock; local hardenings in the materials, due to crystallization and other chemical processes, often create indescribable variations which are more or less completely expressed in the path of the stream.

When a land has been newly elevated above the sea there is often—we may say, indeed, generally—a very great difference between the height of its head waters and the ocean level. In this condition of a country the rivers have what we may call a new aspect; their valleys are commonly narrow and rather steep, waterfalls are apt to abound, and the alluvial terraces are relatively small in extent. Stage by stage the torrents cut deeper; the waste which they make embarrasses the course of the lower waters, where no great amount of down-cutting is possible for the reason that the bed of the stream is near sea level. At the same time the alluvial materials, building out to sea, thus diminish the slope of the stream. In the extreme old age of the river system the mountains are eaten down so that the torrent section disappears, and the stream becomes of something like a uniform slope; the higher alluvial plains gradually waste away, until in the end the valley has no salient features. At this stage in the process, or even before it is attained, the valley is likely to be submerged beneath the sea, where it is buried beneath the deposits formed on the floor; or a further uplift of the land may occur with the result that the stream is rejuvenated; or once more endowed with the power to create torrents, build alluvial plains, and do the other interesting work of a normal river.

It rarely, if ever, happens that a river valley attains old age before it has sunk beneath the sea or been refreshed by further upliftings. In the unstable conditions of the continents, one or the other of these processes, sometimes in different places both together, is apt to be going on. Thus if we take the case of the Mississippi and its principal tributaries, the Ohio and Missouri, we find that for many geological ages the mountains about their sources have frequently, if not constantly, grown upward, so that their torrent sections, though they have worn down tens of thousands of feet, are still high above the sea level, perhaps on the average as high as they have ever been. At the same time the slight up-and-down swayings of the shore lands, amounting in general to less than five hundred feet, have greatly affected the channels of the main river and its tributaries in their lower parts. Not long ago the Mississippi between Cairo and the Gulf flowed in a rather steep-sided valley probably some hundreds of feet in depth, which had a width of many miles. Then at the close of the last Glacial period the region sank down so that the sea flooded the valley to a point above the present junction of the Ohio River with the main stream. Since then alluvial plains have filled this estuary to even beyond the original mouth. In many other of our Southern rivers, as along the shore from the Mississippi to the Hudson, the streams have not brought in enough detritus to fill their drowned valleys, which have now the name of bays, of which the Delaware and Chesapeake on the Atlantic coast, and Mobile Bay on the Gulf of Mexico, are good examples. The failure of Chesapeake and Delaware Bays to fill with debris in the measure exhibited by the more southern valleys is due to the fact that the streams which flow into them to a great extent drain from a region thickly covered with glacial waste, a mass which holds the flood waters, yielding the supply but slowly to the torrents, which there have but a slight cutting power.

In our sketch of river valleys no attention has been given to the phenomena of waterfalls, those accidents of the flow which, as we have noted, are particularly apt to characterize rivers which have not yet cut down to near the sea level. Where the normal uniform descent which is characteristic of a river's bed is interrupted by a sudden steep, the fact always indicates the occurrence of one of a number of geological actions. The commonest cause of waterfalls is due to a sudden change in the character of horizontal or at least nearly level beds over which the stream may flow. Where after coursing for a distance over a hard layer the stream comes to its edge and drops on a soft or easily eroded stratum, it will cut this latter bed away, and create a more or less characteristic waterfall. Tumbling down the face of the hard layer, the stream acquires velocity; the debris which it conveys is hurled against the bottom, and therefore cuts powerfully, while before, being only rubbed over the stone as it moved along, it cut but slightly. Masses of ice have the same effect as stones. Bits dropping from the ledge are often swept round and round by the eddies, so that they excavate an opening which prevents their chance escape. In these confined spaces they work like augers, boring a deep, well-like cavity. As the bits of stone wear out they are replaced by others, which fall in from above. Working in this way, the fragments often develop regular well-like depressions, the cavities of which work back under the cliffs, and by the undermining process deprive the face of the wall of its support, so that it tumbles in ruin to the base, there to supply more material for the potholing action.

Waterfalls of the type above described are by far the commonest of those which occur out of the torrent districts of a great river system. That of Niagara is an excellent specimen of the type, which, though rarely manifested in anything like the dignity of the great fall, is plentifully shown throughout the Mississippi Valley and the basin of the Great Lakes. Within a hundred miles of Niagara there are at least a hundred small waterfalls of the same type. Probably three quarters of all the larger accidents of this nature are due to the conditions of a hard bed overlying softer strata.

Falls are also produced in very many instances by dikes which cross the stream. So, too, though rarely, only one striking instance being known, an ancient coral reef which has become buried in strata may afford rock of such hardness that when the river comes to cross it it forms a cascade, as at the Falls of the Ohio, at Louisville, Ky. It is a characteristic of all other falls, except those first mentioned, that they rarely plunge with a clean downward leap over the face of a precipice which recedes at its base, but move downward over an irregular sloping surface.

In the torrent district of rivers waterfalls are commonly very numerous, and are generally due to the varying hardness in the rocks which the streams encounter. Here, where the cutting action is going on with great rapidity, slight differences in the resistance which the rocks make to the work will lead to great variations in the form of the bed over which they flow, while on the more gently sloping bottoms of the rivers, where the debris moves slowly, such variations would be unimportant in their effect. When the torrents escape into the main river valleys, in regions where the great streams have cut deep gorges, they often descend from a great vertical height, forming wonderful waterfalls, such as those which occur in the famous Lauterbrunnen Valley of Switzerland or in that of the Yosemite in California. This group of cascades is peculiar in that the steep of the fall is made not by the stream itself, but by the action of a greater river or of a glacier which may have some time taken its place.

Waterfalls have an economic as well as a picturesque interest in that they afford sources of power which may be a very great advantage to manufacturers. Thus along the Atlantic coast the streams which come from the Appalachian highlands, and which have hardly escaped from their torrent section before they attain the sea, afford numerous cataracts which have been developed so that they afford a vast amount of power. Between the James on the south and the Ste. Croix on the north more than a hundred of these Appalachian rivers have been turned to economic use. The industrial arts of this part of the country depend much upon them for the power which drives their machinery. The whole of the United States, because of the considerable size of its rivers and their relatively rapid fall, is richly endowed with this source of energy, which, originating in the sun's heat and conveyed through the rain, may be made to serve the needs of man. In view of the fact that recent inventions have made it possible to convert this energy of falling water into the form of electricity, which may be conveyed to great distances, it seems likely that our rivers will in the future be a great source of national wealth.

We must turn again to river valleys, there to trace certain actions less evident than those already noted, but of great importance in determining these features of the land. First, we have to note that in the valley or region drained by a river there is another degrading or down-wearing action than that which is accomplished by the direct work of the visible stream. All over such a valley the underground waters, soaking through the soil and penetrating through the underlying rock, are constantly removing a portion of the mineral matter which they take into solution and bear away to the sea. In this way, deprived of a part of their substance, the rocks are continually settling down by underwear throughout the whole basin, while they are locally being cut down by the action of the stream. Hence in part it comes about that in a river basin we find two contrasted features—the general and often slight slope of a country toward the main stream and its greater tributaries, and the sharp indentation of the gorge in which the streams flow, these latter caused by the immediate and recent action of the streams.

If now the reader will conceive himself standing at any point in a river basin, preferably beyond the realms of the torrents, he may with the guidance of the facts previously noted, with a little use of the imagination, behold the vast perceptive which the history of the river valley may unfold to him. He stands on the surface of the soil, that debris of the rocks which is just entering on its way to the ocean. In the same region ten thousand years ago he would have stood upon a surface from one to ten feet higher than the present soil covering. A million years ago his station would have been perhaps five hundred feet higher than the surface. Ten million years in the past, a period less than the lifetime of certain rivers, such as the French Broad River in North Carolina, the soil was probably five thousand feet or more above its present plane. There are, indeed, cases where river valleys appear to have worked down without interruption from the subsidence of the land beneath the sea to the depth of at least two miles. Looking upward through the space which the rocks once occupied, we can conceive the action of the forces in their harmonious co-operation which have brought the surface slowly downward. We can imagine the ceaseless corrosion due to the ground water, bringing about a constant though slow descent of the whole surface. Again and again the streams, swinging to and fro under the guidance of the underlying rock, or from the obstacles which the debris they carried imposed upon them, have crossed the surface. Now and then perhaps the wearing was intensified by glacial action, for an ice sheet often cuts with a speed many times as great as that which fluid water can accomplish. On the whole, this exercise of the constructive imagination in conceiving the history of a river valley is one of the most enlarging tasks which the geologist can undertake.

Where in a river valley there are many lateral streams, and especially where the process of solution carried on by the underground waters is most effective, as compared with erosive work done in the bed of the main river, we commonly find the valley sloping gently toward its centre, the rivers having but slight steeps near their banks. On the other hand, where, as occasionally happens, a considerable stream fed by the rain and snow fall in its torrent section courses for a great distance over high, arid plains, on which the ground water and the tributaries do but little work, the basin may slope with very slight declivity to the river margins, and there descend to great depths, forming very deep gorges, of which the Colorado Canon is the most perfect type. As instances of these contrasted conditions, we may take, on the one hand, the upper Mississippi, where the grades toward the main stream are gentle and the valley gorge but slightly exhibited; on the other, the above-mentioned Colorado, which bears a great tide of waters drawn from the high and relatively rainy region of the Rocky Mountains across the vast plateau lying in an almost rainless country. In this section nearly all the down-wearing has been brought about in the direct path of the stream, which has worn the elevated plain into a deep gorge during the slow uprising of the table-land to its present height. In this way a defile nearly a mile in depth has been created in a prevailingly rather flat country. This gorge has embranchments where the few great tributaries have done like work, but, on the whole, this river flows in an almost unbroken channel, the excavation of which has been due to its swift, pebble-bearing waters.

The tendency of a newly formed river is to cut a more or less distinct canon. As the basin becomes ancient, this element of the gorge tends to disappear, the reason for this being that, while the river bed is high above the sea, the current is swift and the down-cutting rapid, while the slow subsidence of the country on either side—a process which goes on at a uniform rate—causes the surface of that region to be left behind in the race for the sea level. As the stream bed comes nearer the sea level its rate of descent is diminished, and so the outlying country gradually overtakes it.

In regions where the winters are very cold the effect of ice on the development of the stream beds both in the torrent and river sections of the valley is important. This work is accomplished in several diverse ways. In the first place, where the stream is clear and the current does not flow too swiftly, the stones on the bottom radiate their heat through the water, and thus form ice on their surfaces, which may attain considerable thickness. As ice is considerably lighter than water, the effect is often to lift up the stones of the bed if they be not too large; when thus detached from the bottom, they are easily floated down stream until the ice melts away. The ice which forms on the surface of the water likewise imprisons the pebbles along the banks, and during the subsequent thaw may carry them hundreds of miles toward the sea. It seems likely, from certain observations made by the writer, that considerable stones may thus be carried from the Alleghany River to the main Mississippi.

Perhaps the most important effect of ice on river channels is accomplished when in a time of flood the ice field which covered the stream, perhaps to the depth of some feet, is broken up into vast floes, which drift downward with the current. When, as on the Ohio, these fields sometimes have the area of several hundred acres, they often collide with the shores, especially where the stream makes a sharp bend. Urged by their momentum, these ice floes pack into the semblance of a dam, which may have a thickness of twenty, thirty, or even fifty feet. Beginning on the shore, where the collision takes place, the dam may swiftly develop clear across the stream, so that in a few minutes the way of the waters is completely blocked. The on-coming ice shoots up upon the accumulation, increases its height, and extends it up stream, so that in an hour the mass completely bars the current. The waters then heap up until they break their way over the obstacle, washing its top away, until the whole is light enough to be forced down the stream, where, by the friction it encounters on the bottom and sides of the channel, it is broken to pieces. It is easy to see that such moving dams of ice may sweep the bed of a river as with a great broom.

Sometimes where the gorges do not form a stationary dam large cakes of ice become turned on edge and pack together so that they roll down the stream like great wheels, grinding the bed rock as they go.

In high northern countries, as in Siberia, the rivers, even the deepest, often become so far frozen that their channels are entirely obstructed. Where, as in the case of these Siberian rivers, the flow is from south to north, it often happens that the spring thaw sets in before the more northern beds of the main stream are released from their bondage of frost. In this case the inundations have to find new paths on either side of the obstructed way. The result is a type of valleys characterized by very irregular and changeable stream beds, the rivers having no chance to organize themselves into the shapely curves which they ordinarily follow.

The supply which finds its way to a river is composed, as has been already incidentally noted, in part of the water which courses underground for a greater or less distance before it emerges to the surface, and in part of that which moves directly over the ground. These two shares of water have somewhat different histories. On the share of these two depends the stability of the flow. Where, as in New England and other glaciated countries, the surface of the earth is covered with a thick layer of sand and gravel, which, except when frozen, readily admits the water; the rainfall is to a very great extent absorbed by the earth, and only yielded slowly to the streams. In these cases floods are rare and of no great destructive power. Again, where also the river basin is covered by a dense mantle of forests, the ground beneath which is coated, as is the case in primeval woods, with a layer of decomposing vegetation a foot or more in depth, this spongy mass retains the water even more effectively than the open-textured glacial deposits above referred to. When the woods, however, are removed from such an area, the rain may descend to the streams almost as speedily as it finds its way to the gutters from the house roofs. It thus comes about that all regions, when reduced to tillage, and where the rainfall is enough to maintain a good agriculture, are, except when they have a coating of glacial waste, exceedingly liable to destructive inundations.

Unhappily, the risk of river floods is peculiarly great in all the regions of the United States lying much to the east of the Rocky Mountains, except in the basin of the Great Lakes and in the district of New England, where the prevalence of glacial sands and gravels affords the protection which we have noted. Throughout this region the rainfall is heavy, and the larger part of it is apt to come after the ground has become deeply snow-covered. The result is a succession of devastating floods which already are very damaging to the works of man, and promise to become more destructive as time goes on. More than in any other country, we need the protection which forests can give us against these disastrous outgoings of our streams.

LAKES.

In considering the journey of water from the hilltops to the sea, we should take some account of those pauses which it makes on its way when for a time it falls into the basin of a lake. These arrests in the downward motion of water, which we term lakes, are exceedingly numerous; their proper discussion would, indeed, require a considerable volume. We shall here note only the more important of their features, those which are of interest to the general student.

The first and most noteworthy difference in lakes is that which separates the group of dead seas from the living basins of fresh water. When a stream attains a place where its waters have to expand into the lakelike form, the current moves in a slow manner, and the broad surface exposed to the air permits a large amount of evaporation. If the basin be large in proportion to the amount of the incurrent water, this evaporation may exceed the supply, and produce a sea with no outlet, such as we find in the Dead Sea of Judea, in that at Salt Lake, Utah, and in a host of other less important basins. If the rate of evaporation be yet greater in proportion to the flow, the lake may altogether dry away, and the river be evaporated before it attains the basin where it might accumulate. In that case the river is said to sink, but, in place of sinking into the earth, its waters really rise into the air. Many such sinks occur in the central portion of the Rocky Mountain district. It is important to note that the process of evaporation we are describing takes place in the case of all lakes, though only here and there is the air so dry that the evaporation prevents the basin from overflowing at the lowest point on its rim, forming a river which goes thence to the sea. Even in the case of the Great Lakes of North America a considerable part of the water which flows into them does not go to the St. Lawrence and thence to the sea. As long as the lake finds an outlet to the sea its waters contain but little more dissolved mineral matter than that we find in the rivers. But because all water which has been in contact with the earth has some dissolved mineral substances, while that which goes away by evaporation is pure water, a lake without an outlet gradually becomes so charged with these materials that it can hold no more in solution, but proceeds to lay them down in deposits of that compound substance which from its principal ingredient we name salt. The water of dead seas, because of the additional weight of the substances which it holds, is extraordinarily buoyant. The swimmer notes a difference in this regard in the waters of rivers and fresh-water lakes and those of the sea, due to this same cause. But in those of dead seas, saturated with saline materials, the human body can not sink as it does in the ordinary conditions of immersion. It is easy to understand how the salt deposits which are mined in many parts of the world have generally, if not in all cases, been formed in such dead seas.[5]

[Footnote 5: In some relatively rare cases salt deposits are formed in lagoons along the shores of arid lands, where the sea occasionally breaks over the beach into the basin, affording waters which are evaporated, leaving their salt behind them.]

It is an interesting fact that almost all the known dead seas have in recent geological times been living lakes—that is, they poured over their brims. In the Cordilleras from the line between Canada and the United States to central Mexico there are several of these basins. All of those which have been studied show by their old shore lines that they were once brimful, and have only shrunk away in modern times. These conditions point to the conclusion that the rainfall in different regions varies greatly in the course of the geologic ages. Further confirmation of this is found in the fact that very great salt deposits exist on the coast of Louisiana and in northern Europe—regions in which the rainfall is now so great in proportion to the evaporation that dead seas are impossible.

Turning now to the question of how lake basins are formed, we note a great variety in the conditions which may bring about their construction. The greatest agent, or at least that which operates in the construction of the largest basins, are the irregular movements of the earth, due to the mountain-building forces. Where this work goes on on a large scale, basin-shaped depressions are inevitably formed. If all those which have existed remained, the large part of the lands would be covered by them. In most cases, however, the cutting action of the streams has been sufficient to bring the drainage channels down to the bottom of the trough, while the influx of sediments has served to further the work by filling up the cavities. Thus at the close of the Cretaceous period there was a chain of lakes extending along the eastern base of the Rocky Mountains, constituting fresh-water seas probably as large as the so-called Great Lakes of North America. But the rivers, by cutting down and tilling up, have long since obliterated these water areas. In other cases the tiltings of the continent, which sometimes oppose the flow of the streams, may for a time convert the upper part of a river basin which originally sloped gently toward the sea into a cavity. Several cases of this description occurred in New England in the closing stages of the Glacial period, when the ground rose up to the northward.

We have already noted the fact that the basin of a dead sea becomes in course of time the seat of extensive salt deposits. These may, indeed, attain a thickness of many hundred feet. If now in the later history of the country the tract of land with the salt beneath it were traversed by a stream, its underground waters may dissolve out the salt and in a way restore the basin to its original unfilled condition, though in the second state that of a living lake. It seems very probable that a portion at least of the areas of Lakes Ontario, Erie, and Huron may be due to this removal of ancient salt deposits, remains of which lie buried in the earth in the region bordering these basins.

By far the commonest cause of lake basins is found in the irregularities of the surface which are produced by the occupation of the country by glaciers. When these great sheets of ice lie over a land, they are in motion down the slopes on which they rest; they wear the bed rocks in a vigorous manner, cutting them down in proportion to their hardness. As these rocks generally vary in the resistance which they oppose to the ice, the result is that when the glacier passes away the surface no longer exhibits the continued down slope which the rivers develop, but is warped in a very complicated way. These depressions afford natural basins in which lakes gather; they may vary in extent from a few square feet to many square miles. When a glacier occupies a country, the melting ice deposits on the surface of the earth a vast quantity of rocky debris, which was contained in its mass. This detritus is irregularly accumulated; in part it is disposed in the form of moraines or rude mounds made at the margin of the glacier, in part as an irregular sheet, now thick, now thin, which covers the whole of the field over which the ice lay. The result of this action is the formation of innumerable pools, which continue to exist until the streams have cut channels through which their waters may drain away, or the basins have become filled with detritus imported from the surrounding country or by peat accumulations which the plants form in such places.

Doubtless more than nine tenths of all the lake basins, especially those of small size, which exist in the world are due to irregularities of the land surface which are brought about by glacial action. Although the greater part of these small basins have been obliterated since the ice left this country, the number still remaining of sufficient size to be marked on a good map is inconceivably great. In North America alone there are probably over a hundred and fifty thousand of these glacial lakes, although by far the greater part of those which existed when the glacial sheet disappeared have been obliterated.

Yet another interesting group of fresh-water lakes, or rather we should call them lakelets from their small size, owes its origin to the curious underground excavations or caverns which are formed in limestone countries. The water enters these caverns through what are termed "sink holes"—basins in the surface which slope gently toward a central opening through which the water flows into the depths below. The cups of the sink holes rarely exceed half a mile in diameter, and are usually much smaller. Their basins have been excavated by the solvent and cutting actions of the rain water which gathers in them to be discharged into the cavern below. It often happens that after a sink hole is formed some slight accident closes the downward-leading shaft, so that the basin holds water; thus in parts of the United States there are thousands of these nearly circular pools, which in certain districts, as in southern Kentucky, serve to vary the landscape in much the same manner as the glacial lakes of more northern countries.

Some of the most beautiful lakes in the world, though none more than a few miles in diameter, occupy the craters of extinct volcanoes. When for a time, or permanently, a volcano ceases to do its appointed work of pouring forth steam and molten rock from the depths of the earth, the pit in the centre of the cone gathers the rain water, forming a deep circular lake, which is walled round by the precipitous faces of the crater. If the volcano reawakens, the water which blocks its passage may be blown out in a moment, the discharge spreading in some cases to a great distance from the cone, to be accumulated again when the vent ceases to be open. The most beautiful of these volcanic lakes are to be found in the region to the north and south of Rome. The original seat of the Latin state was on the shores of one of these crater pools, south of the Eternal City. Lago Bolsena, which lies to the northward, and is one of the largest known basins of this nature, having a diameter of about eight miles, is a crater lake. The volcanic cone to which it belongs, though low, is of great size, showing that in its time of activity, which did not endure very long, this crater was the seat of mighty ejections. The noblest specimen of this group of basins is found in Crater Lake, Oregon, now contained in one of the national parks of the United States.

Inclosed bodies of water are formed in other ways than those described; the list above given includes all the important classes of action which produce these interesting features. We should now note the fact that, unlike the seas, the lakes are to be regarded as temporary features in the physiography of the land. One and all, they endure for but brief geologic time, for the reason that the streams work to destroy them by filling them with sediment and by carving out channels through which their waters drain away. The nature of this action can well be conceived by considering what will take place in the course of time in the Great Lakes of North America. As Niagara Falls cut back at the average rate of several feet a year, it will be but a brief geologic period before they begin to lower the waters of Lake Erie. It is very probable, indeed, that in twenty thousand years the waters of that basin will be to a great extent drained away. When this occurs, another fall or rapid will be produced in the channel which leads from Lake Huron to Lake Erie. This in turn will go through its process of retreat until the former expanse of waters disappears. The action will then be continued at the outlets of Lakes Michigan and Superior, and in time, but for the interposition of some actions which recreate these basins, their floors will be converted into dry land.

It is interesting to note that lakes owe in a manner the preservation of their basins to an action which they bring about on the waters that flow into them. These rivers or torrents commonly convey great quantities of sediment, which serve to rasp their beds and thus to lower their channels. In all but the smaller lakelets these turbid waters lay down all their sediment before they attain the outlet of the basin. Thus they flow away over the rim rock in a perfectly pure state—a state in which, as we have noted before, water has no capacity for abrading firm rock. Thus where the Niagara River passes from Lake Erie its clean water hardly affects the stone over which it flows. It only begins to do cutting work where it plunges down the precipice of the Falls and sets in motion the fragments which are constantly falling from that rocky face. These Falls could not have begun as they did on the margin of Lake Ontario except for the fact that when the Niagara River began to flow, as in relatively modern times, it found an old precipice on the margin of Lake Ontario, formed by the waves of the lake, down which the waters fell, and where they obtained cutting tools with which to undermine the steep which forms the Falls.

Many great lakes, particularly those which we have just been considering, have repeatedly changed their outlets, according as the surface of the land on which they lie has swayed up and down in various directions, or as glacial sheets have barred or unbarred the original outlets of the basins. Thus in the Laurentian Lakes above Ontario the geologist finds evidence that the drainage lines have again and again been changed. For a time during the Glacial period, when Lake Ontario and the valley of the St. Lawrence was possessed by the ice, the discharge was southward into the upper Mississippi or the Ohio. At a later stage channels were formed leading from Georgian Bay to the eastern part of Ontario. Yet later, when the last-named lake was bared, an ice dam appears to have remained in the St. Lawrence, which held back the waters to such a height that they discharged through the valley of the Mohawk into the Hudson. Furthermore, at some time before the Glacial period, we do not know just when, there appears to have been an old Niagara River, now filled with drift, which ran from Lake Erie to Ontario, a different channel from that occupied by the present stream.

The effects of lakes on the river systems with which they are connected is in many ways most important. Where they are of considerable extent, or where even small they are very numerous, they serve to retain the flood waters, delivering them slowly to the excurrent streams. In rising one foot a lake may store away more water than the river by its consequent rise at the point of outflow will carry away in many months, and this for the simple reason that the lake may be many hundred or even thousand times as wide as the stream. Moreover, as before noted, the sediment gathered by the stream above the level of the lake is deposited in its basin, and does not affect the lower reaches of the river. The result is that great rivers, such as drain from the Laurentian Lakes, flow clear water, are exempt from floods, are essentially without alluvial plains or terraces, and form no delta deposits. In all these features the St. Lawrence River affords a wonderful contrast to the Mississippi. Moreover, owing to the clear waters, though it has flowed for a long time, it has never been able to cut away the slight obstructions which form its rapids, barriers which probably would have been removed if its waters had been charged with sediment.



CHAPTER VI.

GLACIERS.

We have already noted the fact that the water in the clouds is very commonly in the frozen state; a large part of that fluid which is evaporated from the sea attains the solid form before it returns to the earth. Nevertheless, in descending, at least nine tenths of the precipitation returns to the fluid state, and does the kind of work which we have noted in our account of water. Where, however, the water arrives on the earth in the frozen condition, it enters on a role totally different from that followed by the fluid material.

Beginning its descent to the earth in a snowflake, the little mass falls slowly, so that when it comes against the earth the blow which it strikes is so slight that it does no effective work. In the state of snow, even in the separate flakes, the frozen water contains a relatively large amount of air. It is this air indeed, which, by dividing the ice into many flakes that reflect the light, gives it the white colour. This important point can be demonstrated by breaking transparent ice into small bits, when we perceive that it has the hue of snow. Much the same effect is given where glass is powdered, and for the same reason.

As the snowflakes accumulate layer on layer they imbed air between them, so that when the material falls in a feathery shape—say to the depth of a foot—more than nine tenths of the mass is taken up by the air-containing spaces. As these cells are very small, the circulation in them is slight, and so the layer becomes an admirable non-conductor, having this quality for the same reason that feathers have it—i.e., because the cells are small enough to prevent the circulation of the air, so that the heat which passes has to go by conduction, and all gases are very poor conductors. The result is that a snow coating is in effect an admirable blanket. When the sun shines upon it, much of the heat is reflected, and as the temperature does not penetrate it to any depth, only the superficial part is melted. This molten water takes up in the process of melting a great deal of heat, so that when it trickles down into the mass it readily refreezes. On the other hand, the heat going out from the earth, the store accumulated in its superficial parts in the last warm season, together with the small share which flows out from the earth's interior, is held in by this blanket, which it melts but slowly. Thus it comes about that in regions of long-enduring snowfall the ground, though frozen to the depth of a foot or more at the time when the accumulation took place, may be thawed out and so far warmed that the vegetation begins to grow before the protecting envelope of snow has melted away. Certain of the early flowers of high latitudes, indeed, begin to blossom beneath the mantle of finely divided ice.

In those parts of the earth which for the most part receive only a temporary coating of snow the effect of this covering is inconsiderable. The snow water is yielded to the earth, from which it has helped to withdraw the frost, so that in the springtime, the growing season of plants, the ground contains an ample store of moisture for their development. Where the snowfall accumulates to a great thickness, especially where it lodges in forests, the influence of the icy covering is somewhat to protract the winter and thus to abbreviate the growing season.

Where snow rests upon a steep slope, and gathers to the depth of several feet, it begins to creep slowly down the declivity in a manner which we may often note on house roofs. This motion is favoured by the gradual though incomplete melting of the flakes as the heat penetrates the mass. Making a section through a mass of snow which has accumulated in many successive falls, we note that the top may still have the flaky character, but that as we go down the flakes are replaced by adherent shotlike bodies, which have arisen from the partial melting and gathering to their centres of the original expanded crystalline bits. In this process of change the mass can move particle by particle in the direction in which gravity impels it. The energy of its motion, however, is slight, yet it can urge loose stones and forest waste down hill. Sometimes, as in the cemetery at Augusta, Me., where stone monuments or other structures, such as iron railings, are entangled in the moving mass, it may break them off and convey them a little distance down the slope.

So long as the summer sun melts the winter's snow, even if the ground be bare but for a day, the role of action accomplished by the snowfall is of little geological consequence. When it happens that a portion of the deposit holds through the summer, the region enters on the glacial state, and its conditions undergo a great revolution, the consequences of which are so momentous that we shall have to trace them in some detail. Fortunately, the considerations which are necessary are not recondite, and all the facts are of an extremely picturesque nature.

Taking such a region as New England, where all the earth is life-bearing in the summer season, and where the glacial period of the winter continues but for a short time, we find that here and there on the high mountains the snow endures throughout most of the summer, but that all parts of the surface have a season when life springs into activity. On the top of Mount Washington, in the White Mountains of New Hampshire, in a cleft known as Tuckerman's Ravine, where the deposit accumulates to a great depth, the snow-ice remains until midsummer. It is, indeed, evident that a very slight change in the climatal conditions of this locality would establish a permanent accumulation of frozen water upon the summit of the mountain. If the crest were lifted a thousand feet higher, without any general change in the heat or rainfall of the district, this effect would be produced. If with the same amount of rainfall as now comes to the earth in that region more of it fell as snow, a like condition would be established. Furthermore, with an increase of rainfall to something like double that which now descends the snow bore the same proportion to the precipitation which it does at present, we should almost certainly have the peak above the permanent snow line, that level below which all the winter's fall melts away. These propositions are stated with some care, for the reason that the student should perceive how delicate may be—indeed, commonly is—the balance of forces which make the difference between a seasonal and a perennial snow covering.

As soon as the snow outlasts the summer, the region which it occupies is sterilized to life. From the time the snow begins to hold over the warm period until it finally disappears, that field has to be reckoned out of the habitable earth, not only to man, but to the lowliest organisms.[6]

[Footnote 6: In certain fields of permanent snow, particularly near their boundaries, some very lowly forms of vegetable life may develop on a frozen surface, drawing their sustenance from the air, and supplied with water by the melting which takes place during the summertime. These forms include the rare phenomenon termed red snow.]

If the snow in a glaciated region lay where it fell, the result would be a constant elevation of the deposit year by year in proportion to the annual excess of deposition over the melting or evaporation of the material. But no sooner does the deposit attain any considerable thickness than it begins to move in the directions of least resistance, in accordance with laws which the students of glaciers are just beginning to discern. In small part this motion is accomplished by avalanches or snow slides, phenomena which are in a way important, and therefore merit description. Immediately after a heavy snowfall, in regions where the slopes are steep, it often happens that the deposit which at first clung to the surface on which it lay becomes so heavy that it tends to slide down the slope; a trifling action, the slipping, indeed, of a single flake, may begin the movement, which at first is gradual and only involves a little of the snow. Gathering velocity, and with the materials heaped together from the junction of that already in motion with that about to be moved, the avalanche in sliding a few hundred feet down the slope may become a deep stream of snow-ice, moving with great celerity. At this stage it begins to break off masses of ice from the glaciers over which it may flow, or even to move large stones. Armed with these, it rends the underlying earth. After it has flowed a mile it may have taken up so much earth and material that it appears like a river of mud. Owing to the fact that the energy which bears it downward is through friction converted into heat, a partial melting of the mass may take place, which converts it into what we call slush, or a mixture of snow and water. Finally, the torrent is precipitated into the bottom of a valley, where in time the frozen water melts away, leaving only the stony matter which it bore as a monument to show the termination of its flow.

It was the good fortune of the writer to see in the Swiss Oberland one very great avalanche, which came from the high country through a descent of several thousand feet to the surface of the Upper Grindelwald Glacier. The first sign of the action was a vague tremor of the air, like that of a great organ pipe when it begins to vibrate, but before the pulsations come swiftly enough to make an audible note. It was impossible to tell when this tremor came, but the wary guide, noting it before his charge could perceive anything unusual, made haste for the middle of the glacier. The vibration swelled to a roar, but the seat of the sound amid the echoing cliffs was indeterminable. Finally, from a valley high up on the southern face of the glacier, there leaped forth first a great stone, which sprang with successive rebounds to the floor of ice. Then in succession other stones and masses of ice which had outrun the flood came thicker and thicker, until at the end of about thirty seconds the steep front of the avalanche appeared like a swift-moving wall. Attaining the cliffs, it shot forth as a great cataract, which during the continuance of the flow—which lasted for several minutes—heaped a great mound of commingled stones and ice upon the surface of the glacier. The mass thus brought down the steep was estimated at about three thousand cubic yards, of which probably the fiftieth part was rock material. An avalanche of this volume is unusual, and the proportion of stony matter borne down exceptionally great; but by these sudden motions of the frozen water a large part of the snow deposited above the zone of complete melting is taken to the lower valleys, where it may disappear in the summer season, and much of the erosion accomplished in the mountains is brought about by these falls.

In all Alpine regions avalanches are among the most dreaded accidents. Their occurrence, however, being dependent upon the shape of the surface, it is generally possible to determine in an accurate way the liability of their happening in any particular field. The Swiss take precaution to protect themselves from their ravages as other folk do to procure immunity from floods. Thus the authorities of many of the mountain hamlets maintain extensive forests on the sides of the villages whence the downfall may be expected, experience having shown that there is no other means so well calculated to break the blow which these great snowfalls can deliver, as thick-set trees which, though they are broken down for some distance, gradually arrest the stream.

As long as the region occupied by permanent snow is limited to sharp mountain peaks, relief by the precipitation of large masses to the level below the snow line is easily accomplished, but manifestly this kind of a discharge can only be effective from a very small field. Where the relief is not brought about by these tumbles of snow, another mode of gravitative action accomplishes the result, though in a more roundabout way, through the mechanism of glaciers.

We have already noted the fact that the winter's snow upon our hillsides undergoes a movement in the direction of the slope. What we have now to describe in a rather long story concerning glaciers rests upon movements of the same nature, though they are in certain features peculiarly dependent on the continuity of the action from year to year. It is desirable, however, that the student should see that there is at the foundation no more mystery in glacial motion than there is in the gradual descent of the snow after it has lain a week on a hillside. It is only in the scale and continuity of the action that the greatest glacial envelope exceeds those of our temporary winters—in fact, whenever the snow falls the earth it covers enters upon an ice period which differs only in degree from that from which our hemisphere is just escaping.

Where the reader is so fortunate as to be able to visit a region of glaciers, he had best begin his study of their majestic phenomena by ascending to those upper realms where the snow accumulates from year to year. He will there find the natural irregularities of the rock surface in a measure evened over by a vast sheet of snow, from which only the summits of the greater mountains rise. He may soon satisfy himself that this sheet is of great depth, for here and there it is intersected by profound crevices. If the visit is made in the season when snow falls, which is commonly during most of the year, he may observe, as before noted in our winter's snow, that the deposit, though at first flaky, attains at a short distance below the surface a somewhat granular character, though the shotlike grains fall apart when disturbed. Yet deeper, ordinarily a few feet below the surface, these granules are more or less cemented together; the mass thus loses the quality of snow, and begins to appear like a whitish ice. Looking down one of the crevices, where the light penetrates to the depth of a hundred feet or more, he may see that the bluish hue somewhat increases with the depth. A trace of this colour is often visible even in the surface snow on the glacier, and sometimes also in our ordinary winter fields. In a hole made with a stick a foot or more in depth a faint cerulean glimmer may generally be discerned; but the increased blueness of the ice as we go down is conspicuous, and readily leads us to the conclusion that the air, to which, as we before noted, the whiteness of the snow is due, is working out of the mass as the process of compaction goes on. In a glacial district this snow mass above the melting line is called the neve.

Remembering that the excess of snow beyond the melting in a neve district amounts, it may be, to some feet of material each year, we easily come to the conclusion that the mass works down the slope in the manner which it does even where the coating is impermanent. This supposition is easily confirmed: by observing the field we find that the sheet is everywhere drawing away from the cliffs, leaving a deep fissure between the neve and the precipices. This crevice is called by the German-Swiss guides the Bergschrund. Passage over it is often one of the most difficult feats to accomplish which the Alpine explorer has to undertake. In fact, the very appearance of the surface, which is that of a river with continuous down slopes, is sufficient evidence that the mass is slowly flowing toward the valleys. Following it down, we almost always come to a place where it passes from the upper valleys to the deeper gorges which pierce the skirts of the mountain. In going over this projection the mass of snow-ice breaks to pieces, forming a crowd of blocks which march down the slope with much more speed than they journeyed when united in the higher-lying fields. In this condition and in this part of the movement the snow-ice forms what are called the seracs, or curds, as the word means in the French-Swiss dialect. Slipping and tumbling down the steep slope on which the seracs develop, the ice becomes broken into bits, often of small size. These fragments are quickly reknit into the body of ice, which we shall hereafter term the glacier, and in this process the expulsion of the air goes on more rapidly than before, and the mass assumes a more transparent icelike quality.

The action of the ice in the pressures and strains to which it is subjected in joining the main glacier and in the further part of its course demand for their understanding a revision of those notions as to rigidity and plasticity which we derive from our common experience with objects. It is hard to believe that ice can be moulded by pressure into any shape without fracturing, provided the motion is slowly effected, while at the same time it is as brittle as ice to a sudden blow. We see, however, a similar instance of contrasted properties in the confection known as molasses candy, a stick of which may be indefinitely bent if the flexure is slowly made, but will fly to pieces like glass if sharply struck. Ice differs from the sugary substance in many ways; especially we should note that while it may be squeezed into any form, it can not be drawn out, but fractures on the application of a very slight tension. The conditions of its movement we will inquire into further on, when we have seen more of its action.

Entering on the lower part of its course, that where it flows into the region below the snow line, the ice stream is now confined between the walls of the valley, a channel which in most cases has been shaped before the ice time, by a mountain torrent, or perhaps by a slower flowing river. In this part of its course the likeness of a glacial stream to one of fluid water is manifest. We see that it twists with the turn of the gorge, widens where the confining walls are far apart, and narrows where the space is constricted. Although the surface is here and there broken by fractures, it is evident that the movement of the frozen current, though slow, is tolerably free. By placing stakes in a row across the axis of a glacier, and observing their movement from day to day, or even from hour to hour if a good theodolite is used for the purpose, we note that the movement of the stream is fastest in the middle parts, as in the case of a river, and that it slows toward either shore, though it often happens, as in a stream of molten water, that the speediest part of the current is near one side. Further observations have indicated that the movement is most rapid on the surface and least at the bottom, in which the stream is also riverlike. It is evident, in a word, that though the ice is not fluid in strict sense, the bits of which it is made up move in substantially the manner of fluids—that is, they freely slip over each other. We will now turn our attention to some important features of a detailed sort which glaciers exhibit.

If we visit a glacier during the part of the year when the winter snows are upon it, it may appear to have a very uninterrupted surface. But as the summer heat advances, the mask of the winter coating goes away, and we may then see the structure of the ice. First of all we note in all valley glaciers such as we are observing that the stream is overlaid by a quantity of rocky waste, the greater part of which has come down with the avalanches in the manner before described, though a small part may have been worn from the bed over which the ice flows. In many glaciers, particularly as we approach their termination, this sheet of earth and rock materials often covers the ice so completely that the novice in such regions finds it difficult to believe that the ice is under his feet. If the explorer is minded to take the rough scramble, he can often walk for miles on these masses of stone without seeing, much less setting foot on any frozen water. In some of the Alaskan glaciers this coating may bear a forest growth. In general, this material, which is called moraine, is distributed in bands parallel to the sides of the glaciers, and the strips may amount to a half dozen or more. Those on the sides of the ice have evidently been derived from the precipices which they have passed. Those in the middle have arisen from the union of the moraines formed in two or more tributary valleys.



Where the avalanches fall most plentifully, the stones lie buried with the snow, and only melt out when the stream attains the region where the annual waste of its surface exceeds the snowfall. In this section we can see how the progressive melting gradually brings the rocky debris into plain view. Here and there we will find a boulder perched on a pedestal of ice, which indicates a recent down-wearing of the field. A frequent sound in these regions arises from the tumble of the stones from their pedestals or the slipping of the masses from the sharp ridge which is formed by the protection given to the ice through the thick coating of detritus on its surface. These movements of the moraines often distribute their waste over the glacier, so that in its lower part we can no longer trace the contributions from the several valleys, the whole area being covered by the debris. At the end of the ice stream, where its forward motion is finally overcome by the warmth which it encounters, it leaves in a rude heap, extending often like a wall across the valley, all the coarse fragments which it conveys. This accumulation, composed of all the lateral moraines which have gathered on the ice by the fall of avalanches, is called the terminal moraine. As the ice stream itself shrinks, a portion of the detritus next the boundary wall is apt to be left clinging against those slopes. It is from the presence of these heaps in valleys now abandoned by glaciers that we obtain some information as to the former greater extent of glacial action.

The next most noticeable feature is the crevasse. These fractures often exist in very great numbers, and constitute a formidable barrier in the explorer's way. The greater part of these ruptures below the serac zone run from the sides of the stream toward the centre without attaining that region. These are commonly pointed up stream; their formation is due to the fact that, owing to the swifter motion in the central parts of the stream, the ice in that section draws away from the material which is moving more slowly next the shore. As before noted, these ice fractures when drawn out naturally form fissures at right angles to the direction of the strain. In the middle portions of the ice other fissures form, though more rarely, which appear to depend on local strains brought about through the irregularity of the surface over which the ice is flowing.

If the observer is fortunate, he may in his journey over the glacier have a chance to see and hear what goes on when crevasses are formed. First he will hear a deep, booming sound beneath his feet, which merges into a more splintering note as the crevice, which begins at the bottom or in the distance, comes upward or toward him. When the sound is over, he may not be able to see a trace of the fracture, which at first is very narrow. But if the break intersect any of the numerous shallow pools which in a warm summer's day are apt to cover a large part of the surface, he may note a line of bubbles rushing up through the water, marking the escape of the air from the glacier, some remnant of that which is imprisoned in the original snow. Even where this indication is wanting, he can sometimes trace the crevice by the hissing sound of the air streams where they issue from the ice. If he will take time to note what goes on, he can usually in an hour or two behold the first invisible crack widen until it may be half an inch across. He may see how the surface water hastens down the opening, a little river system being developed on the surface of the ice as the streams make their way to one or more points of descent. In doing this work they excavate a shaft which often becomes many feet in diameter, down which their waters thunder to the base of the glacier. This well-like opening is called a moulin, or mill, a name which, as we shall see, is well deserved from the work which falling waters accomplish. Although the institution of the moulin shaft depends upon the formation of a crevice, it often happens that as the ice moves farther on its journey its walls are again thrust together, soldered in the manner peculiar to ice, so that no trace of the rupture remains except the shaft which it permitted to form. Like everything else in the glacier, the moulin slowly moves down the slope, and remains open as long as it is the seat of descending waters produced by the summer melting. When it ceases to be kept open from the summer, its walls are squeezed together in the fashion that the crevices are closed.

Forming here and there, and generally in considerable numbers, the crevices of a glacier entrap a good deal of the morainal debris, which falls through them to the bottom of the glacier. Smaller bits are washed into the moulin, by the streams arising from the melting ice, which is brought about by the warm sun of the summer, and particularly by the warm rains of that season. On those glaciers where, owing to the irregularity of the bottom over which the ice flows, these fractures are very numerous, it may happen that all the detritus brought upon the surface of the glacier by avalanches finds its way to the floor of the ice.

Although it is difficult to learn what is going on at the under surface of the glacier, it is possible directly and indirectly to ascertain much concerning the peculiar and important work which is there done. The intrepid explorer may work his way in through the lateral fissures, and even with care safely descend some of the fissures which penetrate the central parts of a shallow ice stream. There, it may be at the depth of a hundred feet or more, he will find a quantity of stones, some of which may be in size like to a small house held in the body of the ice, but with one side resting upon the bed rock. He may be so fortunate as to see the stone actually in process of cutting a groove in the bed rock as it is urged forward by the motion of the glacier. The cutting is not altogether in the fixed material, for the boulder itself is also worn and scored in the work. Smaller pebbles are caught in the space between the erratic and the motionless rock and ground to bits. If in his explorations the student finds his way to the part of the floor on which the waters of a moulin fall, he may have a chance to observe how the stones set in motion serve to cut the bed rock, forming elongated potholes much as in the case of ordinary waterfalls, or at the base of those shafts which afford the beginnings of limestone caverns.

The best way to penetrate beneath the glacier is through the arch of the stream which always flows from the terminal face of the ice river. Even in winter time every large glacier discharges at its end a considerable brook, the waters of which have been melted from the ice in small part by the outflow of the earth's heat; mainly, however, by the warmth produced in the friction of the ice on itself and on its bottom—in other words, by the conversion of that energy of position, of which we have often to speak, into heat. In the summer time this subglacial stream is swollen by the surface waters descending through the crevices and the moulins which come from them, so that the outflow often forms a considerable river, and thus excavates in the ice a large or at least a long cavern, the base of which is the bed rock. In the autumn, when the superficial melting ceases, this gallery can often be penetrated for a considerable distance, and affords an excellent way to the secrets of the under ice. The observer may here see quantities of the rock material held in the grip of the ice, and forced to a rude journey over the bare foundation stones. Now and then he may find the glacial mass in large measure made up of stones, the admixture extending many feet above the bottom of the cavern, perhaps to the very top of the arch. He may perchance find that these stones are crushing each other where they are in contact. The result will be brought about by the difference in the rate of advance of the ice, which moves the faster the higher it is above the surface over which it drags, and thus forces the stones on one level over those below. Where the waters of the subglacial stream have swept the bed rock clean of debris its surface is scored, grooved, and here and there polished in a manner which is accomplished only by ice action, though some likeness to it is afforded where stones have been swept over for ages by blowing sand. Here and there, often in a way which interrupts the cavern journey, the shrunken stream, unable to carry forward the debris, deposits the material in the chamber, sometimes filling the arch so completely that the waters are forced to make a detour. This action is particularly interesting, for the reason that in regions whence glaciers have disappeared the deposits formed in the old ice arches often afford singularly perfect moulds of those caverns which were produced by the ancient subglacial streams. These moulds are termed eskers.

If the observer be attentive, he will note the fact that the waters emerging from beneath the considerable glacier are very much charged with mud. If he will take a glass of the water at the point of escape, he will often find, on permitting it to settle, that the sediment amounts to as much as one twentieth of the volume. While the greater part of this detritus will descend to the bottom of the vessel in the course of a day, a portion of it does not thus fall. He may also note that this mud is not of the yellowish hue which he is accustomed to behold in the materials laid down by ordinary rivers, but has a whitish colour. Further study will reveal the fact that the difference is due to the lack of oxidation in the case of the glacial detritus. River muds forming slowly and during long-continued exposure to the action of the air have their contained iron much oxidized, which gives them a part of their darkened appearance. Moreover, they are somewhat coloured with decayed vegetable matter. The waste from beneath the glacier has been quickly separated from the bed rock, all the faces of the grains are freshly fractured, and there is no admixture of organic matter. The faces of the particles thus reflect light in substantially the same way as powdered glass or pulverized ice, and consequently appear white.

A little observation will show the student that this very muddy character of waters emerging from beneath the glacier is essentially peculiar to such streams as we have described. Ascending any of the principal valleys of Switzerland, he may note that some of the streams flow waters which carry little sediment even in times when they are much swollen, while others at all seasons have the whitish colour. A little further exploration, or the use of a good map, will show him that the pellucid streams receive no contributions of glacial water, while those which look as if they were charged with milk come, in part at least, from the ice arches. From some studies which the writer has made in Swiss valleys, it appears that the amount of erosion accomplished on equal areas of similar rock by the descent of the waters in the form of a glacier or in that of ordinary torrents differs greatly. Moving in the form of ice, or in the state of ice-confined streams, the mass of water applies very many times as much of its energy of position to grinding and bearing away the rocks as is accomplished where the water descends in its fluid state.

The effect of the intense ice action above noted is rapidly to wear away the rocks of the valley in which the glacier is situated. This work is done not only in a larger measure but in a different way from that accomplished by torrents. In the case of the latter, the stream bed is embarrassed by the rubbish which comes into it; only here and there can it attack the bed rock by forcing the stones over its surface. Only in a few days of heavy rain each year is its work at all effective; the greater part of the energy of position of its waters is expended in the endless twistings and turnings of its stream, which result only in the development of heat which flies away into the atmosphere. In the ice stream, owing to its slow movement and to the detritus which it forces along the bottom, a vastly greater part of the energy which impels it down the slope is applied to rock cutting. None of the boulders, even if they are yards in diameter, obstruct its motion; small and great alike are to it good instruments wherewith to attack the bed rocks. The fragments are never left to waste by atmospheric decay, but are to a very great extent used up in mechanical work, while the most of the detritus which comes to a torrent is left in a coarse state when it is delivered to the stream; the larger part of that which the glacier transports is worn out in its journey. To a great extent it is used up in attacking the bed rock. In most cases the debris in the terminal moraine is evidently but a small part of what entered the ice during its journey from the uplands; the greater part has been worn out in the rude experiences to which it has been subjected.

It is evident that even in the regions now most extensively occupied by glaciers the drainage systems have been shaped by the movement of ordinary streams—in other words, ice action is almost everywhere, even in the regions about the poles, an incidental feature in the work of water, coming in only to modify the topography, which is mainly moulded by the action of fluid water. When, owing to climatal changes, a valley such as those of the Alps is occupied by a glacial stream, the new current proceeds at once, according to its evident needs, to modify the shape of its channel. An ordinary torrent, because of the swiftness of its motion, which may, in general, be estimated at from three to five miles an hour, can convey away the precipitation over a very narrow bed. Therefore its channel is usually not a hundredth part as wide as the gorge or valley in which it lies. But when the discharge takes place by a glacier, the speed of which rarely exceeds four or five feet a day, the ice stream because of its slow motion has to fill the trough from side to side, it has to be some thousand times as deep and wide as the torrent. The result is that as soon as the glacial condition arises in a country the ice streams proceed to change the old V-shaped torrent beds into those which have a broad U-like form. The practised eye can in a way judge how long a valley has been subjected to glacial action by the extent to which it has been widened by this process.

In the valleys of Switzerland and other mountain districts which have been attentively studied it is evident that glacial action has played a considerable part in determining their forms. But the work has been limited to that part of the basin in which the ice is abundantly provided with cutting tools in the stone which have found their way to the base of the stream. In the region of the neve, where the contributions of rocky matter to the surface of the deposit made from the few bare cliffs which rise above the sheet of snow is small, the snow-ice does no cutting of any consequence. Where it passes over the steep at the head of the deep valley into which it drains, and is riven into the seracs, such stony matter as it may have gathered is allowed to fall to the bottom, and so comes into a position where it may do effective work. From this serac section downward the now distinct ice river, being in general below the snow line, has everywhere cliffs, on either side from which the contributions of rock material are abundant. Hence this part of the glacier, though it is the wasting portion of its length, does all the cutting work of any consequence which is performed. It is there that the underrunning streams become charged with sediment, which, as we have noted, they bear in surprising quantities, and it is therefore in this section of the valley that the impress of the ice work is the strongest. Its effect is not only to widen the valley and deepen it, but also to advance the deep section farther up the stream and its tributaries. The step in the stream beds which we find at the seracs appears to mark the point in the course of the glacier where, owing to the falling of stones to its base, as well as to its swifter movements and the firmer state of the ice, it does effective wearing.

There are many other features connected with glaciers which richly repay the study of those who have a mind to explore in the manner of the physicist interested in ice actions the difficult problems which they afford; but as these matters are not important from the point of view of this work, no mention of them will here be made. We will now turn our attention to that other group of glaciers commonly termed continental, which now exist about either pole, and which at various times in the earth's history have extended far toward the equator, mantling over vast extents of land and shallow sea. The difference between the ice streams of the mountains and those which we term continental depends solely on the areas of the fields and the depth of the accumulation. In an ordinary Alpine region the neve districts, where the snow gathers, are relatively small. Owing to the rather steep slopes, the frozen water is rapidly discharged into the lower valleys, where it melts away. Both in the neve and in the distinct glacier of the lower grounds there are, particularly in the latter, projecting peaks, from which quantities of stone are brought down by avalanches or in ordinary rock falls, so that the ice is abundantly supplied with cutting tools, which work from its surface down to its depths.

As the glacial accumulation grows in depth there are fewer peaks emerging from it, and the streams which it feeds rise the higher until they mantle over the divides between the valleys. Thus by imperceptible stages valley glaciers pass to the larger form, usually but incorrectly termed continental. We can, indeed, in going from the mountains in the tropics to the poles, note every step in this transition, until in Greenland we attain the greatest ice mass in the world, unless that about the southern pole be more extensive. In the Greenland glacier the ice sheet covers a vast extent of what is probably a mountain country, which is certainly of this nature in the southern part of the island, where alone we find portions of the earth not completely covered by the deep envelope. Thanks to the labours of certain hardy explorers, among whom Nansen deserves the foremost place, we now know something as to the conditions of this vast ice field, for it has been crossed from shore to shore. The results of these studies are most interesting, for they afford us a clew as to the conditions which prevail over a large part of the earth during the Glacial period from which the planet is just escaping, and in the earlier ages when glaciation was likewise extensive. We shall therefore consider in a somewhat detailed way the features which the Greenland glacier presents.

Starting from the eastern shore of that land, if we may thus term a region which presents itself mainly in the form of ice, we find next the shore a coast line not completely covered with ice and snow, but here and there exhibiting peaks which indicate that if the frozen mantle were removed the country would appear deeply intersected with fiords in the manner exhibited in the regions to the south of Greenland or the Scandinavian peninsula. The ice comes down to the sea through the valleys, often facing the ocean for great distances with its frozen cliffs. Entering on this seaward portion of the glacier, the observer finds that for some distance from the coast line the ice is more or less rifted with crevices, the formation of which is doubtless due to irregularities of the rock bottom over which it moves. These ruptures are so frequent that for some miles back it is very difficult to find a safe way. Finally, however, a point is attained where these breaks rather suddenly disappear, and thence inward the ice rises at the rate of upward slope of a few feet to the mile in a broad, nearly smooth incline. In the central portion of the region for a considerable part of the territory the ice has very little slope. Thence it declines toward the other shore, exhibiting the same features as were found on the eastern versant until near the coast, when again the surface is beset with crevices which continue to the margin of the sea.

Although the explorations of the central field of Greenland are as yet incomplete, several of these excursions into or across the interior have been made, and the identity of the observations is such that we can safely assume the whole region to be of one type. We can furthermore run no risk in assuming that what we find in Greenland, at least so far as the unbroken nature of the central ice field is concerned, is what must exist in every land where the glacial envelope becomes very deep. In Greenland it seems likely that the depth of the ice is on the average more than half a mile, and in the central part of the realm the sheet may well have a much greater profundity; it may be nearly a mile deep. The most striking feature—that of a vast unbroken expanse, bordered by a region where the ice is ruptured—is traceable wherever very extensive and presumably deep deposits of ice have been examined. As we shall see hereafter, these features teach us much as to the conditions of glacial action—a matter which we shall have to examine after we have completed our general survey as to the changes which occur during glacial periods.

In the present state of that wonderful complex of actions which we term climate, glaciers are everywhere, so far as our observations enable us to judge, generally in process of decrease. In Switzerland, although the ancients even in Roman days were in contact with the ice, they were so unobservant that they did not even remark that the ice was in motion. Only during the last two centuries have we any observations of a historic sort which are of value to the geologist. Fortunately, however, the signs written on the rock tell the story, except for its measurement in terms of years, as clearly as any records could give it. From this testimony of the rocks we perceive that in the geological yesterday, though it may have been some tens of thousands of years ago, the Swiss glaciers, vastly thickened, and with their horizontal area immensely expanded, stretched over the Alpine country, so that only here and there did any of the sharper peaks rise above the surface. These vast glaciers, almost continually united on their margins, extended so far that every portion of what is now the Swiss Republic was covered by them. Their front lay on the southern lowlands of Germany, on the Jura district of France; on the south, it stretched across the valley of the Po as far as near Milan. We know this old ice front by the accumulations of rock debris which were brought to it from the interior of the mountain realm. We can recognise the peculiar kinds of stone, and with perfect certainty trace them to the bed rock whence they were riven. Moreover, we can follow back through the same evidence the stages of retreat of the glaciers, until they lost their broad continental character and assumed something like their present valley form. Up the valley of any of the great rivers, as, for instance, that of the Rhone above the lake of Geneva, we note successive terminal moraines which clearly indicate stages in the retreat of the ice when for a time it ceased to go backward, or even made a slight temporary readvance. It is easily seen that on such occasions the stones carried to the ice front would be accumulated in a heap, while during the time when day by day the glacier was retreating the rock waste would be left broadcast over the valley.

As we go up from the course of the glacial streams we note that the successive moraines have their materials in a progressively less decayed state. Far away from the heap now forming, and in proportion to the distance, the stones have in a measure rotted, and the heaps which they compose are often covered with soil and occupied by forests. Within a few miles of the ice front the stones still have a fresh aspect. When we arrive within, say, half a mile of the moraine now building, we come to the part of the glacial retreat of which we have some written or traditional account. This is in general to the effect that the wasting of the glaciers is going on in this century as it went on in the past. Occasionally periods of heavy snow would refresh the ice streams, so that for a little time they pushed their fronts farther down the valley. The writer has seen during one of these temporary advances the interesting spectacle of ice destroying and overturning the soil of a small field which had been planted in grain.

It should be noted that these temporary advances of the ice are not due to the snowfall of the winter or winters immediately preceding the forward movement. So slow is the journey of the ice from the neve field to the end of a long glacier that it may require centuries for the store accumulated in the uplands to affect the terminal portion of the stream. We know that the bodies of the unhappy men who have been lost in the crevices of the glacier are borne forward at a uniform and tolerably computable rate until they emerge at the front, where the ice melts away. In at least one case the remains have appeared after many years in the debris which is contributed to the moraine. On account of this slow feeding of the glacial stream, we naturally may expect to find, as we do, in fact, that a great snowfall of many years ago, and likewise a period when the winter's contribution has been slight, would influence the position of the terminal point of the ice stream at different times, according to its length. If the length of the flow be five miles, it may require twenty or thirty years for the effect to be evident; while if the stream be ten miles long, the influence may not be noted in less than threescore years. Thus it comes about that at the present time in the same glacial district some streams may be advancing while others are receding, though, on the whole, the ice is generally in process of shrinkage. If the present rate of retreat should be maintained, it seems certain that at the end of three centuries the Swiss glaciers as a whole will not have anything like their present area, and many of the smaller streams will entirely disappear.

Following the method of the illustrious Louis Agassiz, who first attentively traced the evidence which shows the geologically recent great extension of glaciers by studying the evidence of the action in fields they no longer occupy, geologists have now inspected a large part of the land areas with a view to finding the proofs of such ice work. So far as these indications are concerned, the indications which they have had to trace are generally of a very unmistakable character. Rarely, indeed, does a skilled student of such phenomena have to search in any region for more than a day before he obtains indubitable evidence which will enable him to determine whether or not the field has recently been occupied by an enduring ice sheet—one which survives the summer season and therefore deserves the name of glacier. The indications which he has to consider consist in the direction and manner in which the surface materials have been carried, the physical conditions of these materials, the shape of the surface of the underlying rock as regards its general contour, and the presence or absence of scratches and groovings on its surface. As these records of ice action are of first importance in dealing with this problem, and as they afford excellent subjects for the study of those who dwell in glaciated regions, we shall note them in some detail.

The geologist recognises several ways in which materials may be transported on the surface of the earth. They may be cast forth by volcanoes, making their journey by being shot through the air, or by flowing in lava streams; it is always easy at a glance, save in very rare instances, to determine whether fragments have thus been conveyed. Again, the detritus may be moved by the wind; this action is limited; it only affects dust, sand, and very small pebbles, and is easily discriminated. The carriage may be effected by river or marine currents; here, again, the size of the fragments moved is small, and the order of their arrangement distinctly traceable. The fragments may be conveyed by ice rafts; here, too, the observer can usually limit the probabilities he has to consider by ascertaining, as he can generally do, whether the region which he is observing has been below a sea or lake. In a word, the before-mentioned agents of transportation are of somewhat exceptional influence, and in most cases can, as explanations of rock transportation, be readily excluded. When, therefore, the geologist finds a country abundantly covered with sand, pebbles, and boulders arranged in an irregular way, he has generally only to inquire whether the material has been carried by rivers or by glaciers. This discrimination can be quickly and critically effected. In the first place, he notes that rivers only in their torrent sections can carry large fragments of rock, and that in all cases the fragments move down hill. Further, that where deposits are formed, they have more or less the form of alluvial deposits. If now the observations show that the rock waste occupying the surface of any region has been carried up hill and down, across the valleys, particularly if there are here and there traces of frontal moraines, the geologist is entitled to suppose—he may, indeed, be sure—that the carriage has been effected by a glacial sheet.

Important corroborative evidence of ice action is generally to be found by inspecting the bed rock below the detritus, which indicates glacial action. Even if it be somewhat decayed, as is apt to be the case where the ice sheet long since passed away, the bed rock is likely to have a warped surface; it is cast into ridges and furrows of a broad, flowing aspect, such as liquid water never produces, which, indeed, can only be created by an ice sheet moving over the surface, cutting its bed in proportion to the hardness of the material. Furthermore, if the bed rock have a firm texture, and be not too much decayed, we almost always find upon it grooves or scratches, channels carved by the stones embedded in the body of the ice, and drawn by its motion over the fixed material. Thus the proof of glacial extension in the last ice epoch is made so clear that accurate maps can be prepared showing the realm of its action. This task is as yet incomplete, although it is already far advanced.

While the study of glaciers began in Europe, inquiries concerning their ancient extension have been carried further and with more accuracy in North America than in any other part of the world. We may therefore well begin our description of the limits of the ice sheets with this continent. Imagining a seafarer to have approached America by the North Atlantic, as did the Scandinavians, and that his voyage came perhaps a hundred thousand years or more before that of Leif Ericsson, he would have found an ice front long before he attained the present shores of the land. This front may have extended from south of Greenland, off the shores of the present Grand Banks of Newfoundland, thence and westward to central or southern New Jersey. This cliff of ice was formed by a sheet which lay on the bottom of the sea. On the New Jersey coast the ice wall left the sea and entered on the body of the continent. We will now suppose that the explorer, animated with the valiant scientific spirit which leads the men of our day to seek the poles, undertook a land journey along the ice front across the continent. From the New Jersey coast the traveller would have passed through central Pennsylvania, where, although there probably detached outlying glaciers lying to the southward as far as central Virginia, the main front extended westward into the Ohio Valley. In southern Ohio a tongue of the ice projected southwardly until it crossed the Ohio River, where Cincinnati now lies, extending a few miles to the southward of the stream. Thence it deflected northwardly, crossing the Mississippi, and again the Missouri, with a tongue or lobe which went far southward in that State. Then again turning to the northwest, it followed in general the northern part of the Missouri basin until it came to within sight of the Rocky Mountains. There the ice front of the main glacier followed the trend of the mountains at some distance from their face for an unknown extent to the northward. In the Cordilleras, as far south as southern Colorado, and probably in the Sierra Nevada to south of San Francisco, the mountain centres developed local glaciers, which in some places were of very great size, perhaps exceeding any of those which now exist in Switzerland. It will thus be seen that nearly one half of the present land area of North America was beneath a glacial covering, though, as before noted, the region about the Gulf of Mexico may have swayed upward when the northern portion of the land was borne down by the vast load of ice which rested upon it. Notwithstanding this possible addition to the land, our imaginary explorer would have found the portion of the continent fit for the occupancy of life not more than half as great as it is at present.

In the Eurasian continent there was no such continuous ice sheet as in North America, but the glaciers developed from a number of different centres, each moving out upon the lowlands, or, if its position was southern, being limited to a particular mountain field. One of these centres included Scandinavia, northern Germany, Great Britain about as far south as London, and a large part of Ireland, the ice covering the intermediate seas and extending to the westward, so that the passage of the North Atlantic was greatly restricted between this ice front and that of North America. Another centre, before noted, was formed in the Alps; yet another, of considerable area, in the Pyrenees; other less studied fields existed in the Apennines, in the Caucasus, the Ural, and the other mountains of northern Asia. Curiously enough, however, the great region of plains in Siberia does not appear to have been occupied by a continuous ice sheet, though the similar region in North America was deeply embedded in a glacier. Coincident with this development of ice in the eastern part of the continent, the ice streams of the Himalayan Mountains, some of which are among the greatest of our upland glaciers, appear to have undergone but a moderate extension. Many other of the Eurasian highlands were probably ice-bound during the last Glacial period, but our knowledge concerning these local fields is as yet imperfect.

In the southern hemisphere the lands are of less extent and, on the whole, less studied than in the northern realm. Here and there where glaciers exist, as in New Zealand and in the southern part of South America, observant travellers have noticed that these ice fields have recently shrunk away. Whether the time of greatest extension and of retreat coincided with that of the ice sheets in the north is not yet determined; the problem, indeed, is one of some difficulty, and may long remain undecided. It seems, however, probable that the glaciers of the southern hemisphere, like those in the north, are in process of retreat. If this be true, then their time of greatest extension was probably the same as that of the ice sheets about the southern pole. From certain imperfect reports which we have concerning evidences of glaciation in Central America and in the Andean district in the northern part of South America, it seems possible that at one time the upland ice along the Cordilleran chain existed from point to point along that system of elevations, so that the widest interval between the fields of permanent snow with their attendant glaciers did not much exceed a thousand miles.

Observing the present gradual retreat of those ice remnants which remain mere shreds and patches of the ancient fields, it seems at first sight likely that the extension and recession of the great glaciers took place with exceeding slowness. Measured in terms of human life, in the manner in which we gauge matters of man's history, this process was doubtless slow. There are reasons, however, to believe that the coming and going were, in a geological sense, swift; they may have, indeed, been for a part of the time of startling rapidity. Going back to the time of geological yesterday, before the ice began its development in the northern hemisphere, all the evidence we can find appears to indicate a temperate climate extending far toward the north pole. The Miocene deposits found within twelve degrees, or a little more than seven hundred miles, of the north pole, and fairly within the realm of lowest temperature which now exists on the earth, show by the plant remains which they contain that the conditions permitted the growth of forests, the plants having a tolerably close resemblance to those which now freely develop in the southern portion of the Mississippi Valley. Among them there are species which had the habit of retaining their broad, rather soft leaves throughout the winter season. The climate appears, in a word, to have been one where the mean annual temperature must have been thirty degrees or more higher than the present average of that realm. Although such conditions near the sea level are not inconsistent with the supposition that glaciers existed in the higher mountains of the north, they clearly deny the possibility of the realm being occupied by continental glaciers.

Although the Pliocene deposits formed in high latitudes have to a great extent been swept away by the subsequent glacial wearing, they indicate by their fossils a climatal change in the direction of greater cold. We trace this change, though obscurely, in a progressive manner to a point where the records are interrupted, and the next interpretable indication we have is that the ice sheet had extended to somewhere near the limits which we have noted. We are then driven to seek what we can concerning the sojourn of the ice on the land by the amount of wearing which it has inflicted upon the areas which it occupied. This evidence has a certain, though, as we shall see, a limited value.